Output Contexts

Generated from C++ output definitions and io/output_contexts.cpp. Required BP attributes are <variable>/description, <variable>/unit and <variable>/plot_label. Optional attributes identify coordinates, topology, basis, normalization, species and time indexing. IMAS candidates are semantic hints, with dictionary version and confidence; they do not convert native data.

Plot labels support $...$ math with single backslashes in the editor. Units remain separate. The literal {name} expands to the native variable name only outside math delimiters, retaining species/channel qualifiers. Rendering depends on the plotting application.

To edit documentation, run python3 utils/python/documentation/xgc_doc_editor.py. Review JSON is temporary; only accepted/amended edits explicitly applied to C++ affect this reference or future BP files. Rebuild XGC after applying source edits.

./xgc.saved_volumes.bp: nnode (VolumesAndAreasNnode)

Native units: Positive integer.

Number of 2D mesh nodes.

  • plot_label: $N_n$

Source: XGC_core/cpp/grid/volumes_and_areas.cpp:203.

./xgc.saved_volumes.bp: node_area (VolumesAndAreasNodeArea)

Native units: m^2.

Poloidal-plane area associated with each mesh node.

  • plot_label: $\Delta A_{\mathrm{node,lin}}$

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • centering: node

Source: XGC_core/cpp/grid/volumes_and_areas.cpp:241.

./xgc.saved_volumes.bp: node_vol (VolumesAndAreasNodeVol)

Native units: m^3.

Nodal volume weight for linear barycentric interpolation over each local toroidal slab. For axisymmetric geometry the full-torus volume is wedge_n*nphi*sum(node_vol[0,:]); for stellarator geometry it is wedge_n*sum(node_vol over all stored planes and nodes).

  • plot_label: $\Delta V_{\mathrm{lin}}$

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • centering: node

Source: XGC_core/cpp/grid/volumes_and_areas.cpp:232.

./xgc.saved_volumes.bp: node_vol_ff_h (VolumesAndAreasNodeVolFfH)

Native units: m^3.

Field-following spatial volume associated with each mesh node for the two neighboring field-following planes.

  • plot_label: $\Delta V_{\mathrm{ff}}$ (midplane)

  • axes: toroidal_plane,field_following_side,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Middle axis selects the two linear toroidal deposition sides (0 left, 1 right) of each local slab. Midplane/left/right geometry is identified by the variable suffix.

  • centering: node

Source: XGC_core/cpp/grid/volumes_and_areas.cpp:258.

./xgc.saved_volumes.bp: node_vol_ff_l_h (VolumesAndAreasNodeVolFfLH)

Native units: m^3.

Left-side field-following spatial volume associated with each mesh node for the two neighboring field-following planes.

  • plot_label: $\Delta V_{\mathrm{ff}}$ (left plane)

  • axes: toroidal_plane,field_following_side,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Middle axis selects the two linear toroidal deposition sides (0 left, 1 right) of each local slab. Midplane/left/right geometry is identified by the variable suffix.

  • centering: node

Source: XGC_core/cpp/grid/volumes_and_areas.cpp:270.

./xgc.saved_volumes.bp: node_vol_ff_r_h (VolumesAndAreasNodeVolFfRH)

Native units: m^3.

Right-side field-following spatial volume associated with each mesh node for the two neighboring field-following planes.

  • plot_label: $\Delta V_{\mathrm{ff}}$ (right plane)

  • axes: toroidal_plane,field_following_side,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Middle axis selects the two linear toroidal deposition sides (0 left, 1 right) of each local slab. Midplane/left/right geometry is identified by the variable suffix.

  • centering: node

Source: XGC_core/cpp/grid/volumes_and_areas.cpp:277.

./xgc.saved_volumes.bp: node_vol_nearest (VolumesAndAreasNodeVolNearest)

Native units: m^3.

Spatial volume assigned to each mesh node using nearest-node volume accumulation.

  • plot_label: $\Delta V_{\mathrm{nearest}}$

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • centering: node

Source: XGC_core/cpp/grid/volumes_and_areas.cpp:226.

./xgc.saved_volumes.bp: npsi00 (VolumesAndAreasNpsi00)

Native units: Positive integer.

Number of points in the uniformly spaced psi00 flux grid used for saved volume data.

  • plot_label: $N_{\mathrm{surf}}$

Source: XGC_core/cpp/grid/volumes_and_areas.cpp:208.

./xgc.saved_volumes.bp: ntriangle (VolumesAndAreasNtriangle)

Native units: Positive integer.

Number of triangles in the planar mesh.

  • plot_label: $N_{\mathrm{tr}}$

Source: XGC_core/cpp/grid/volumes_and_areas.cpp:198.

./xgc.saved_volumes.bp: tr_area_h (VolumesAndAreasTrAreaH)

Native units: m^2.

Area of each triangle in the 2D plane mesh.

  • plot_label: $\Delta A_{\mathrm{tr}}$

  • axes: toroidal_plane,triangle

  • mesh_context: xgc.mesh.bp: rz, nd_connect_list (zero-based). One geometric volume/area weight per triangle; use cell-centered visualization.

  • centering: triangle

Source: XGC_core/cpp/grid/volumes_and_areas.cpp:213.

./xgc.saved_volumes.bp: tr_vol_h (VolumesAndAreasTrVolH)

Native units: m^3.

Straight-toroidal-extrusion volume of each triangle over one local toroidal slab of angular width delta_phi.

  • plot_label: $\Delta V_{\mathrm{tr,slab}}$

  • axes: toroidal_plane,triangle

  • mesh_context: xgc.mesh.bp: rz, nd_connect_list (zero-based). One geometric volume/area weight per triangle; use cell-centered visualization.

  • centering: triangle

Source: XGC_core/cpp/grid/volumes_and_areas.cpp:218.

./xgc.saved_volumes.bp: vol00 (VolumesAndAreasVol00)

Native units: m^3.

Volume weights on the uniform psi00 grid, accumulated with linear flux-coordinate weights in included main-plasma/SOL regions over the simulated toroidal domain.

  • plot_label: $\Delta V(\psi)$

  • axes: toroidal_plane,flux_surface

  • mesh_context: Uniform psi00 grid from the simulation setup; corresponding physical flux coordinates are also written as xgc.oneddiag.bp:psi00 when that diagnostic is available. This static file does not contain a same-step oneddiag slice. The grid differs from the particle-moment psi/psi_mks coordinate.

  • time_context: Static volume data saved during mesh setup and reused on later starts; not a diagnostic time slice.

  • centering: flux_surface

  • normalization: Monte Carlo vol00 uses the full simulated toroidal domain rather than one local toroidal slab. It is not normalized like node_vol; do not multiply it by nphi again.

Source: XGC_core/cpp/grid/volumes_and_areas.cpp:246.

field-following projection matrix .bp files: dl_par (FfProjectionDlPar)

Native units: m.

Parallel distance along the magnetic field associated with the field-following projection.

  • plot_label: Parallel projection distance

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz, nd_connect_list and plane geometry; the filename xgc.ff_[1dp/hdp/mdp]_[fwd/rev].bp identifies full/half/midplane projection and direction. eindex/value are the interpolation operator; apply weights only in each row’s first nelement valid slots. dl_par is distance along that field-line projection.

  • coordinate_context: Nonnegative approximate field-line path length per projection source node. Direction and full/half/midplane pairing are identified by the filename. Use this array directly with the corresponding source-node rows; it is not assembled as a sparse matrix.

  • related_variables: Same projection file: eindex, value, nelement specify interpolation; dl_par uses their source-row node association.

Source: XGC_core/cpp/grid/ff_projection.cpp:366.

field-following projection matrix .bp files: eindex (FfProjectionEindex)

Native units: index.

Column indices for stored entries of the field-following projection matrix.

  • plot_label: Projection column index

  • mesh_context: xgc.mesh.bp: rz, nd_connect_list and plane geometry; the filename xgc.ff_[1dp/hdp/mdp]_[fwd/rev].bp identifies full/half/midplane projection and direction. eindex/value are the interpolation operator; apply weights only in each row’s first nelement valid slots. dl_par is distance along that field-line projection.

  • coordinate_context: Sparse linear operator, not a scalar mesh field. Preserve row/column associations of the operator named by the filename; ignore padded fixed-width slots beyond nelement per row. These arrays must be assembled as a sparse matrix before applying interpolation. Fixed-width eindex[row,slot] is ONE-based; subtract one only in slots 0:nelement[row]. CSR ridx and cidx are ZERO-based; row r uses v[ridx[r]:ridx[r+1]] at columns cidx[ridx[r]:ridx[r+1]].

  • related_variables: Same matrix BP file: nrows, ncols, width; fixed-width nelement,eindex,value or CSR ridx,cidx,v,nnz.

Source: XGC_core/cpp/grid/ff_projection.cpp:356.

field-following projection matrix .bp files: ncols (FfProjectionNcols)

Native units: Positive integer.

Number of columns in the field-following projection matrix.

  • plot_label: Matrix column count

Source: XGC_core/cpp/grid/ff_projection.cpp:341.

field-following projection matrix .bp files: nelement (FfProjectionNelement)

Native units: Non-negative integer.

Number of stored nonzero entries in each row of the field-following projection matrix.

  • plot_label: Projection row entry count

  • mesh_context: xgc.mesh.bp: rz, nd_connect_list and plane geometry; the filename xgc.ff_[1dp/hdp/mdp]_[fwd/rev].bp identifies full/half/midplane projection and direction. eindex/value are the interpolation operator; apply weights only in each row’s first nelement valid slots. dl_par is distance along that field-line projection.

  • coordinate_context: Sparse linear operator, not a scalar mesh field. Preserve row/column associations of the operator named by the filename; ignore padded fixed-width slots beyond nelement per row. These arrays must be assembled as a sparse matrix before applying interpolation. Fixed-width eindex[row,slot] is ONE-based; subtract one only in slots 0:nelement[row]. CSR ridx and cidx are ZERO-based; row r uses v[ridx[r]:ridx[r+1]] at columns cidx[ridx[r]:ridx[r+1]].

  • related_variables: Same matrix BP file: nrows, ncols, width; fixed-width nelement,eindex,value or CSR ridx,cidx,v,nnz.

Source: XGC_core/cpp/grid/ff_projection.cpp:351.

field-following projection matrix .bp files: nrows (FfProjectionNrows)

Native units: Positive integer.

Number of rows in the field-following projection matrix.

  • plot_label: Matrix row count

Source: XGC_core/cpp/grid/ff_projection.cpp:336.

field-following projection matrix .bp files: value (FfProjectionValue)

Native units: dimensionless.

Interpolation weights for stored entries of the field-following projection matrix.

  • plot_label: Field-following interpolation weight

  • mesh_context: xgc.mesh.bp: rz, nd_connect_list and plane geometry; the filename xgc.ff_[1dp/hdp/mdp]_[fwd/rev].bp identifies full/half/midplane projection and direction. eindex/value are the interpolation operator; apply weights only in each row’s first nelement valid slots. dl_par is distance along that field-line projection.

  • coordinate_context: Sparse linear operator, not a scalar mesh field. Preserve row/column associations of the operator named by the filename; ignore padded fixed-width slots beyond nelement per row. These arrays must be assembled as a sparse matrix before applying interpolation. Fixed-width eindex[row,slot] is ONE-based; subtract one only in slots 0:nelement[row]. CSR ridx and cidx are ZERO-based; row r uses v[ridx[r]:ridx[r+1]] at columns cidx[ridx[r]:ridx[r+1]].

  • related_variables: Same matrix BP file: nrows, ncols, width; fixed-width nelement,eindex,value or CSR ridx,cidx,v,nnz.

Source: XGC_core/cpp/grid/ff_projection.cpp:361.

field-following projection matrix .bp files: width (FfProjectionWidth)

Native units: Positive integer.

Maximum number of stored nonzero entries per row in the field-following projection matrix.

  • plot_label: Matrix row capacity

Source: XGC_core/cpp/grid/ff_projection.cpp:346.

matrix .bp files: cidx (MatrixCidx)

Native units: index.

CSR column indices for the stored matrix.

  • plot_label: CSR column index

  • mesh_context: xgc.mesh.bp:rz,nd_connect_list define mesh nodes and zero-based triangle connectivity. xgc.smooth_pol.bp and xgc.hyp_vis_rad.bp are node-by-node operators. xgc.cnv_to_surf.bp and xgc.cnv_from_surf.bp both store matrices with mesh-node rows and flux-surface columns; to_surf is applied TRANSPOSED for surface averages, from_surf is applied normally for surface-to-node interpolation. Surface coordinates/membership are xgc.mesh.bp:psi_surf,surf_idx,surf_len. Matched counterparts xgc.cnv_matched_to_surf.bp/xgc.cnv_matched_from_surf.bp use the same axis associations; matched restriction also needs the surface Gram-system inverse and is not a row-normalized average. These matrices are operators, not mesh fields.

  • coordinate_context: Sparse linear operator, not a scalar mesh field. Preserve row/column associations of the operator named by the filename; ignore padded fixed-width slots beyond nelement per row. These arrays must be assembled as a sparse matrix before applying interpolation. Fixed-width eindex[row,slot] is ONE-based; subtract one only in slots 0:nelement[row]. CSR ridx and cidx are ZERO-based; row r uses v[ridx[r]:ridx[r+1]] at columns cidx[ridx[r]:ridx[r+1]].

  • related_variables: Same matrix BP file: nrows, ncols, width; fixed-width nelement,eindex,value or CSR ridx,cidx,v,nnz.

Source: XGC_core/cpp/utils/matrix.cpp:376.

matrix .bp files: eindex (MatrixEindex)

Native units: index.

Column indices for stored entries of the fixed-width matrix representation.

  • plot_label: Matrix column index

  • mesh_context: xgc.mesh.bp:rz,nd_connect_list define mesh nodes and zero-based triangle connectivity. xgc.smooth_pol.bp and xgc.hyp_vis_rad.bp are node-by-node operators. xgc.cnv_to_surf.bp and xgc.cnv_from_surf.bp both store matrices with mesh-node rows and flux-surface columns; to_surf is applied TRANSPOSED for surface averages, from_surf is applied normally for surface-to-node interpolation. Surface coordinates/membership are xgc.mesh.bp:psi_surf,surf_idx,surf_len. Matched counterparts xgc.cnv_matched_to_surf.bp/xgc.cnv_matched_from_surf.bp use the same axis associations; matched restriction also needs the surface Gram-system inverse and is not a row-normalized average. These matrices are operators, not mesh fields.

  • coordinate_context: Sparse linear operator, not a scalar mesh field. Preserve row/column associations of the operator named by the filename; ignore padded fixed-width slots beyond nelement per row. These arrays must be assembled as a sparse matrix before applying interpolation. Fixed-width eindex[row,slot] is ONE-based; subtract one only in slots 0:nelement[row]. CSR ridx and cidx are ZERO-based; row r uses v[ridx[r]:ridx[r+1]] at columns cidx[ridx[r]:ridx[r+1]].

  • related_variables: Same matrix BP file: nrows, ncols, width; fixed-width nelement,eindex,value or CSR ridx,cidx,v,nnz.

Source: XGC_core/cpp/utils/matrix.cpp:389.

matrix .bp files: ncols (MatrixNcols)

Native units: Positive integer.

Number of columns in the stored matrix.

  • plot_label: Matrix column count

Source: XGC_core/cpp/utils/matrix.cpp:347.

matrix .bp files: nelement (MatrixNelement)

Native units: Non-negative integer.

Number of stored nonzero entries in each row of the fixed-width matrix representation.

  • plot_label: Matrix row entry count

  • mesh_context: xgc.mesh.bp:rz,nd_connect_list define mesh nodes and zero-based triangle connectivity. xgc.smooth_pol.bp and xgc.hyp_vis_rad.bp are node-by-node operators. xgc.cnv_to_surf.bp and xgc.cnv_from_surf.bp both store matrices with mesh-node rows and flux-surface columns; to_surf is applied TRANSPOSED for surface averages, from_surf is applied normally for surface-to-node interpolation. Surface coordinates/membership are xgc.mesh.bp:psi_surf,surf_idx,surf_len. Matched counterparts xgc.cnv_matched_to_surf.bp/xgc.cnv_matched_from_surf.bp use the same axis associations; matched restriction also needs the surface Gram-system inverse and is not a row-normalized average. These matrices are operators, not mesh fields.

  • coordinate_context: Sparse linear operator, not a scalar mesh field. Preserve row/column associations of the operator named by the filename; ignore padded fixed-width slots beyond nelement per row. These arrays must be assembled as a sparse matrix before applying interpolation. Fixed-width eindex[row,slot] is ONE-based; subtract one only in slots 0:nelement[row]. CSR ridx and cidx are ZERO-based; row r uses v[ridx[r]:ridx[r+1]] at columns cidx[ridx[r]:ridx[r+1]].

  • related_variables: Same matrix BP file: nrows, ncols, width; fixed-width nelement,eindex,value or CSR ridx,cidx,v,nnz.

Source: XGC_core/cpp/utils/matrix.cpp:384.

matrix .bp files: nnz (MatrixNnz)

Native units: Non-negative integer.

Total number of nonzero entries in the CSR matrix.

  • plot_label: Matrix nonzero count

Source: XGC_core/cpp/utils/matrix.cpp:361.

matrix .bp files: nrows (MatrixNrows)

Native units: Positive integer.

Number of rows in the stored matrix.

  • plot_label: Matrix row count

Source: XGC_core/cpp/utils/matrix.cpp:342.

matrix .bp files: ridx (MatrixRidx)

Native units: index.

CSR row-pointer array for the stored matrix.

  • plot_label: CSR row pointer

  • mesh_context: xgc.mesh.bp:rz,nd_connect_list define mesh nodes and zero-based triangle connectivity. xgc.smooth_pol.bp and xgc.hyp_vis_rad.bp are node-by-node operators. xgc.cnv_to_surf.bp and xgc.cnv_from_surf.bp both store matrices with mesh-node rows and flux-surface columns; to_surf is applied TRANSPOSED for surface averages, from_surf is applied normally for surface-to-node interpolation. Surface coordinates/membership are xgc.mesh.bp:psi_surf,surf_idx,surf_len. Matched counterparts xgc.cnv_matched_to_surf.bp/xgc.cnv_matched_from_surf.bp use the same axis associations; matched restriction also needs the surface Gram-system inverse and is not a row-normalized average. These matrices are operators, not mesh fields.

  • coordinate_context: Sparse linear operator, not a scalar mesh field. Preserve row/column associations of the operator named by the filename; ignore padded fixed-width slots beyond nelement per row. These arrays must be assembled as a sparse matrix before applying interpolation. Fixed-width eindex[row,slot] is ONE-based; subtract one only in slots 0:nelement[row]. CSR ridx and cidx are ZERO-based; row r uses v[ridx[r]:ridx[r+1]] at columns cidx[ridx[r]:ridx[r+1]].

  • related_variables: Same matrix BP file: nrows, ncols, width; fixed-width nelement,eindex,value or CSR ridx,cidx,v,nnz.

Source: XGC_core/cpp/utils/matrix.cpp:366.

matrix .bp files: v (MatrixV)

Native units: matrix-dependent.

CSR nonzero values for the stored matrix.

  • plot_label: CSR matrix coefficient

  • mesh_context: xgc.mesh.bp:rz,nd_connect_list define mesh nodes and zero-based triangle connectivity. xgc.smooth_pol.bp and xgc.hyp_vis_rad.bp are node-by-node operators. xgc.cnv_to_surf.bp and xgc.cnv_from_surf.bp both store matrices with mesh-node rows and flux-surface columns; to_surf is applied TRANSPOSED for surface averages, from_surf is applied normally for surface-to-node interpolation. Surface coordinates/membership are xgc.mesh.bp:psi_surf,surf_idx,surf_len. Matched counterparts xgc.cnv_matched_to_surf.bp/xgc.cnv_matched_from_surf.bp use the same axis associations; matched restriction also needs the surface Gram-system inverse and is not a row-normalized average. These matrices are operators, not mesh fields.

  • coordinate_context: Sparse linear operator, not a scalar mesh field. Preserve row/column associations of the operator named by the filename; ignore padded fixed-width slots beyond nelement per row. These arrays must be assembled as a sparse matrix before applying interpolation. Fixed-width eindex[row,slot] is ONE-based; subtract one only in slots 0:nelement[row]. CSR ridx and cidx are ZERO-based; row r uses v[ridx[r]:ridx[r+1]] at columns cidx[ridx[r]:ridx[r+1]].

  • related_variables: Same matrix BP file: nrows, ncols, width; fixed-width nelement,eindex,value or CSR ridx,cidx,v,nnz.

Source: XGC_core/cpp/utils/matrix.cpp:371.

matrix .bp files: value (MatrixValue)

Native units: matrix-dependent.

Stored matrix coefficients in fixed-width row representation.

  • plot_label: Matrix coefficient

  • mesh_context: xgc.mesh.bp:rz,nd_connect_list define mesh nodes and zero-based triangle connectivity. xgc.smooth_pol.bp and xgc.hyp_vis_rad.bp are node-by-node operators. xgc.cnv_to_surf.bp and xgc.cnv_from_surf.bp both store matrices with mesh-node rows and flux-surface columns; to_surf is applied TRANSPOSED for surface averages, from_surf is applied normally for surface-to-node interpolation. Surface coordinates/membership are xgc.mesh.bp:psi_surf,surf_idx,surf_len. Matched counterparts xgc.cnv_matched_to_surf.bp/xgc.cnv_matched_from_surf.bp use the same axis associations; matched restriction also needs the surface Gram-system inverse and is not a row-normalized average. These matrices are operators, not mesh fields.

  • coordinate_context: Sparse linear operator, not a scalar mesh field. Preserve row/column associations of the operator named by the filename; ignore padded fixed-width slots beyond nelement per row. These arrays must be assembled as a sparse matrix before applying interpolation. Fixed-width eindex[row,slot] is ONE-based; subtract one only in slots 0:nelement[row]. CSR ridx and cidx are ZERO-based; row r uses v[ridx[r]:ridx[r+1]] at columns cidx[ridx[r]:ridx[r+1]].

  • related_variables: Same matrix BP file: nrows, ncols, width; fixed-width nelement,eindex,value or CSR ridx,cidx,v,nnz.

Source: XGC_core/cpp/utils/matrix.cpp:394.

matrix .bp files: width (MatrixWidth)

Native units: Positive integer.

Maximum number of stored nonzero entries per row in the stored matrix.

  • plot_label: Matrix row capacity

Source: XGC_core/cpp/utils/matrix.cpp:352.

moments_check.bp: density (MomentsCheckDensity)

Native units: m^-3.

Reference density moments on [toroidal plane, mesh node, species] for the moments check.

  • plot_label: $n_{\mathrm{ref}}$

  • axes: toroidal_plane,mesh_node,species

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • species_context: Last axis indexes only species with dynamic_f0=true, in simulation order (compact f0_idx order). Species without a dynamic f0 are omitted; electrons come first when included. Resolve species mass/charge with xgc.units.bp. This is an internal initialization/consistency snapshot without physical time coordinates; do not treat ADIOS position as time.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/electrons/density(:)/values; plasma_profiles/ggd(:)/ion(:)/density(:)/values

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/plasma/moments_check.cpp:311.

moments_check.bp: nnodes (MomentsCheckNnodes)

Native units: Positive integer.

Number of 2D mesh nodes per toroidal plane in the moments-check reference data.

  • plot_label: $N_n$

Source: XGC_core/cpp/plasma/moments_check.cpp:301.

moments_check.bp: nplanes (MomentsCheckNplanes)

Native units: Positive integer.

Number of toroidal planes in the moments-check reference data.

  • plot_label: $N_\varphi$

Source: XGC_core/cpp/plasma/moments_check.cpp:296.

moments_check.bp: nspecies (MomentsCheckNspecies)

Native units: Positive integer.

Number of species with a dynamic f0 in the moments-check reference data.

  • plot_label: $N_{\mathrm{sp}}$

Source: XGC_core/cpp/plasma/moments_check.cpp:291.

moments_check.bp: ptl_resolution (MomentsCheckPtlResolution)

Native units: Non-negative integer.

Total marker-particle count used to generate the moments-check reference data.

  • plot_label: Reference marker count

Source: XGC_core/cpp/plasma/moments_check.cpp:306.

moments_check.bp: temp (MomentsCheckTemp)

Native units: eV.

Reference temperature moments on [toroidal plane, mesh node, species] for the moments check.

  • plot_label: $T_{\mathrm{ref}}$

  • axes: toroidal_plane,mesh_node,species

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • species_context: Last axis indexes only species with dynamic_f0=true, in simulation order (compact f0_idx order). Species without a dynamic f0 are omitted; electrons come first when included. Resolve species mass/charge with xgc.units.bp. This is an internal initialization/consistency snapshot without physical time coordinates; do not treat ADIOS position as time.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/electrons/temperature(:)/values; plasma_profiles/ggd(:)/ion(:)/temperature(:)/values

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/plasma/moments_check.cpp:327.

moments_check.bp: temp_para (MomentsCheckTempPara)

Native units: eV.

Reference parallel-temperature moments on [toroidal plane, mesh node, species] for the moments check.

  • plot_label: $T_{\parallel,\mathrm{ref}}$

  • axes: toroidal_plane,mesh_node,species

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • species_context: Last axis indexes only species with dynamic_f0=true, in simulation order (compact f0_idx order). Species without a dynamic f0 are omitted; electrons come first when included. Resolve species mass/charge with xgc.units.bp. This is an internal initialization/consistency snapshot without physical time coordinates; do not treat ADIOS position as time.

  • centering: node

Source: XGC_core/cpp/plasma/moments_check.cpp:344.

moments_check.bp: temp_perp (MomentsCheckTempPerp)

Native units: eV.

Reference perpendicular-temperature moments on [toroidal plane, mesh node, species] for the moments check.

  • plot_label: $T_{\perp,\mathrm{ref}}$

  • axes: toroidal_plane,mesh_node,species

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • species_context: Last axis indexes only species with dynamic_f0=true, in simulation order (compact f0_idx order). Species without a dynamic f0 are omitted; electrons come first when included. Resolve species mass/charge with xgc.units.bp. This is an internal initialization/consistency snapshot without physical time coordinates; do not treat ADIOS position as time.

  • centering: node

Source: XGC_core/cpp/plasma/moments_check.cpp:335.

moments_check.bp: upara (MomentsCheckUpara)

Native units: m s^-1.

Reference parallel-flow moments on [toroidal plane, mesh node, species] for the moments check.

  • plot_label: $u_{\parallel,\mathrm{ref}}$

  • axes: toroidal_plane,mesh_node,species

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • species_context: Last axis indexes only species with dynamic_f0=true, in simulation order (compact f0_idx order). Species without a dynamic f0 are omitted; electrons come first when included. Resolve species mass/charge with xgc.units.bp. This is an internal initialization/consistency snapshot without physical time coordinates; do not treat ADIOS position as time.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/electrons/velocity(:)/parallel; plasma_profiles/ggd(:)/ion(:)/velocity(:)/parallel

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/plasma/moments_check.cpp:319.

particle_check*.bp: refgid (GlobalParticleCheckRefgid)

Native units: Integer index.

Reference globally unique particle IDs used by the global particle consistency check.

  • plot_label: Reference marker ID

  • axes: particle

  • mesh_context: Particle sample, not a mesh field. The same species’ phase array contains row-wise cylindrical R (m), Z (m), phi (rad) in columns 0,1,2; use gid to identify rows/particles across output slices. MPI repartitioning can change row order. Visualize as points or perform a stated statistical/binning procedure, not triangle interpolation. Other phase columns and units are specified in description.

  • centering: particle

Source: XGC_core/cpp/plasma/global_particle_check.cpp:95.

particle_check*.bp: refphase (GlobalParticleCheckRefphase)

Native units: [m, m, rad, m, dimensionless, dimensionless, J/T, dimensionless, m^-3 eV^(-3/2)].

Reference particle phase array with columns [R, Z, phi, rho, w1, w2, mu, w0, g], where \(R\), \(Z\), and \(\\phi\) are the cylindrical coordinates of the particle, \(\\rho\) is the parallel gyroradius (proxy for \(v_{\\parallel}\)), \(w_1\) is the delta-f weight, \(w_2\) is a weight used in the weight evolution equation, \(\\mu\) is the magnetic moment, \(w_0\) is the full-f weight (from importance sampling), and \(g\) is the phase-space marker density. Used by the global particle consistency check.

  • plot_label: Reference particle phase space

  • axes: particle,phase_component

  • mesh_context: Particle sample, not a mesh field. The same species’ phase array contains row-wise cylindrical R (m), Z (m), phi (rad) in columns 0,1,2; use gid to identify rows/particles across output slices. MPI repartitioning can change row order. Visualize as points or perform a stated statistical/binning procedure, not triangle interpolation. Other phase columns and units are specified in description.

  • centering: particle

  • normalization: The g column is marker density in XGC velocity-space normalization, with units m^-3 eV^(-3/2); f=w0*g. For a Maxwellian this normalization is proportional to n*T_eV^(-3/2)*exp(-E_eV/T_eV), with the SI velocity normalization factor (mass/(2*pi*EV_2_J))^(3/2) removed. This is distinct from the spatial triangle marker_den in xgc.loading.*.bp.

Source: XGC_core/cpp/plasma/global_particle_check.cpp:102.

restart_dir/xgc.restart*.bp: 3db_rampup_factor (PerturbedBField3dbRampupFactor)

Native units: dimensionless.

Restart value of the 3D perturbed magnetic-field ramp-up factor.

  • plot_label: 3D magnetic perturbation ramp-up factor

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

Source: XGC_core/cpp/solvers/perturbed_B_field.cpp:835.

restart_dir/xgc.restart*.bp: apars (ElectricFieldApars)

Native units: T m.

Restart value of perturbed parallel vector potential \(A_{\\parallel}\) on toroidal planes and mesh nodes.

  • plot_label: $A_{\parallel}^{(s)}$

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz, nd_connect_list for axisymmetric geometry; rplane_rz, rplane_nd_connect_list for stellarator right-plane geometry. Coordinates are R,Z in m and triangle connectivity is zero-based. Values are node-centered on this geometry.

  • coordinate_context: Right-plane symplectic parallel vector potential, sampled at toroidal angle (plane_index+1)*delta_phi. The written array is [toroidal_plane,mesh_node].

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

  • centering: node

Source: XGC_core/cpp/solvers/electric_field.cpp:288.

restart_dir/xgc.restart*.bp: cden0_rmp (PerturbedBFieldCden0Rmp)

Native units: m^-3.

Checkpoint of the per-toroidal-plane Poisson-RHS charge-number-density accumulation, indexed by Runge-Kutta stage and solver iflag during the RMP EM/ES cycle.

  • plot_label: RMP Poisson-RHS accumulator

  • axes: toroidal_plane,runge_kutta_stage,solver_iflag,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: The first size-two state axis is ipc-1 and the second is the zero-based Poisson iflag; neither is a field-following side. Values are plane-specific accumulators.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

  • centering: node

  • normalization: Charge-number density (charge density divided by elementary charge), accumulated with 1/(mstep_em+mstep_es) and consumed/reset at electromagnetic steps.

Source: XGC_core/cpp/solvers/perturbed_B_field.cpp:854.

restart_dir/xgc.restart*.bp: checkpoint_version (WriteCheckpointFilesCheckpointVersion)

Native units: Positive integer.

Version index indicating the format of this restart file.

  • plot_label: Checkpoint format version

Source: XGC_core/cpp/main/write_checkpoint_files.cpp:16.

restart_dir/xgc.restart*.bp: dpotsave (ElectricFieldDpotsave)

Native units: V.

Restart history of perturbed electrostatic potential on toroidal planes and mesh nodes.

  • plot_label: Potential perturbation history

  • axes: runge_kutta_stage,field_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz, nd_connect_list for axisymmetric geometry; lplane_rz/lplane_nd_connect_list and rplane_rz/rplane_nd_connect_list for the corresponding stellarator field planes. Coordinates are R,Z in m and connectivity is zero-based.

  • coordinate_context: First axis is the two ipc-1 Runge-Kutta history slots. Second axis interleaves the left and right interior field planes for each toroidal partition, at offsets 2*plane_index and 2*plane_index+1. Saved values are used to difference potentials between updates of the same RK slot. Per-slot physical times are not stored; this axis is not a diagnostic timeline.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

  • centering: node

Source: XGC_core/cpp/solvers/electric_field.cpp:245.

restart_dir/xgc.restart*.bp: inode1 (DomainDecompositionInode1)

Native units: Integer index.

One-based global index of the first mesh node assigned to each MPI rank in the checkpoint domain decomposition.

  • plot_label: First global mesh-node index by rank

  • axes: mpi_rank

  • coordinate_context: One-based node-start index for each rank in the checkpoint decomposition; subtract one to recover the zero-based node_offset.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

Source: XGC_core/cpp/domain_decomposition/domain_decomposition.cpp:295.

restart_dir/xgc.restart*.bp: jpar0_rmp (PerturbedBFieldJpar0Rmp)

Native units: m^-2 s^-1.

Checkpoint of the per-toroidal-plane Ampere-RHS accumulation for each Runge-Kutta stage during the RMP EM/ES cycle; elementary charge is not included.

  • plot_label: RMP Ampere-RHS accumulator

  • axes: toroidal_plane,runge_kutta_stage,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Second axis is ipc-1 (two Runge-Kutta stage slots), not a left/right field-following side. Values belong to each plane rather than a toroidal average.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

  • centering: node

  • normalization: Native current-density state is divided by the elementary charge. Multiply by UNIT_CHARGE to obtain the corresponding A m^-2 contribution. Accumulation uses 1/(mstep_em+mstep_es) and is consumed/reset during an electromagnetic step.

Source: XGC_core/cpp/solvers/perturbed_B_field.cpp:845.

restart_dir/xgc.restart*.bp: jpara_diff_integral (CurrentDriveJparaDiffIntegral)

Native units: A m^-2.

Cumulative controller sum of bt_sign*(j_parallel,equilibrium-j_parallel,plasma) over current-drive applications; no physical timestep factor is included.

  • plot_label: Cumulative current-drive error sum

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

  • centering: node

  • normalization: Define delta_j=bt_sign*(j_parallel,equilibrium-j_parallel,plasma), where bt_sign is the equilibrium toroidal-field sign. The controller updates integral += delta_j on each application, so this is a discrete sum in A m^-2, not a physical-time integral.

Source: XGC_core/cpp/sources/current_drive.cpp:389.

restart_dir/xgc.restart*.bp: jpara_diff_previous (CurrentDriveJparaDiffPrevious)

Native units: A m^-2.

Exponentially smoothed controller error bt_sign*(j_parallel,equilibrium-j_parallel,plasma), updated as previous=0.8*previous+0.2*current on each current-drive application.

  • plot_label: Smoothed current-drive error

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

  • centering: node

  • normalization: The array contains the smoothed error history, not the derivative or current-minus-previous difference. It is updated before diagnostic writing and retained verbatim in the checkpoint.

Source: XGC_core/cpp/sources/current_drive.cpp:397.

restart_dir/xgc.restart*.bp: loop_voltage (ElectricFieldLoopVoltage)

Native units: V.

Restart value of the loop-voltage profile on mesh nodes.

  • plot_label: $V_{\mathrm{loop}}$ (checkpoint)

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

  • centering: node

Source: XGC_core/cpp/solvers/electric_field.cpp:202.

restart_dir/xgc.restart*.bp: neu_node_dens (NeutralsNeuNodeDens)

Native units: m^-3.

Checkpointed neutral background density on mesh nodes.

  • plot_label: $n_{\mathrm{neutral}}$ (checkpoint)

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

  • centering: node

Source: XGC_core/cpp/sources/neutrals.cpp:923.

restart_dir/xgc.restart*.bp: neu_node_temp (NeutralsNeuNodeTemp)

Native units: eV.

Checkpointed neutral background temperature on mesh nodes.

  • plot_label: $T_{\mathrm{neutral}}$ (checkpoint)

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

  • centering: node

Source: XGC_core/cpp/sources/neutrals.cpp:928.

restart_dir/xgc.restart*.bp: neu_weight_lost_ebin (NeutralsNeuWeightLostEbin)

Native units: simulation-particle weight.

Checkpointed sum of lost main-ion particle weights binned by wall node and incident-ion energy, used to determine the neutral recycling source.

  • plot_label: Energy-binned main-ion wall-loss weight

  • axes: energy_bin,wall_node

  • mesh_context: xgc.mesh.bp: grid_wall_nodes contains zero-based indices into rz (R,Z in m) in the ordered mesh-boundary polygon sequence produced by order_wall. Wall-node values follow this sequence; successive entries and the final-to-first pair define boundary segments for a valid closed wall. The wall-node list is not the triangle connectivity array nd_connect_list.

  • coordinate_context: First axis is incident main-ion energy bin for the wall-loss tally; endpoints come from xgc.used_inputs.bp:neu_param/neu_ebin_min,neu_ebin_max (eV), neu_ebin_num and neu_ebin_log. Linear edges are Emin+j*(Emax-Emin)/N; logarithmic edges are exp(log(Emin)+j*(log(Emax)-log(Emin))/N). Under/overflow contributions are clamped to first/last bins. If neu_degas2 is false, there is one all-energy bin. Wall entries follow grid_wall_nodes. These are cumulative loss tallies, not a thermal distribution.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

  • centering: wall_node

Source: XGC_core/cpp/sources/neutrals.cpp:917.

restart_dir/xgc.restart*.bp: neu_weight_sum_lost (NeutralsNeuWeightSumLost)

Native units: simulation-particle weight.

Checkpointed sum of lost main-ion particle weights, used to normalize the recycled-neutral source.

  • plot_label: Accumulated main-ion wall-loss weight

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

Source: XGC_core/cpp/sources/neutrals.cpp:906.

restart_dir/xgc.restart*.bp: neu_weight_wall_lost (NeutralsNeuWeightWallLost)

Native units: simulation-particle weight.

Checkpointed sum of lost main-ion particle weights at each wall node, used to determine the neutral recycling source.

  • plot_label: Main-ion wall-loss weight per wall node

  • mesh_context: xgc.mesh.bp: grid_wall_nodes contains zero-based indices into rz (R,Z in m) in the ordered mesh-boundary polygon sequence produced by order_wall. Wall-node values follow this sequence; successive entries and the final-to-first pair define boundary segments for a valid closed wall. The wall-node list is not the triangle connectivity array nd_connect_list.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

  • centering: wall_node

Source: XGC_core/cpp/sources/neutrals.cpp:911.

restart_dir/xgc.restart*.bp: nmup1 (PlasmaNmup1)

Native units: Positive integer.

Number of perpendicular velocity grid points in f0 checkpoint arrays.

  • plot_label: $N_{v,\perp}$

Source: XGC_core/cpp/plasma/plasma.cpp:213.

restart_dir/xgc.restart*.bp: nnode (PlasmaNnode)

Native units: Positive integer.

Number of 2D mesh nodes.

  • plot_label: $N_n$

Source: XGC_core/cpp/plasma/plasma.cpp:208.

restart_dir/xgc.restart*.bp: nphi (PlasmaNphi)

Native units: Positive integer.

Number of toroidal mesh planes \(N_{\\mathrm{planes}}\).

  • plot_label: $N_\varphi$

Source: XGC_core/cpp/plasma/plasma.cpp:218.

restart_dir/xgc.restart*.bp: predictor_ne00_tau (PotentialPredictorPredictorNe00Tau)

Native units: main-ion toroidal transit times.

Averaging time used to accumulate predictor_ne00_time_avg.

  • plot_label: Predictor density averaging time

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

Source: XGC_core/cpp/solvers/potential_predictor.cpp:628.

restart_dir/xgc.restart*.bp: predictor_ne00_time_avg (PotentialPredictorPredictorNe00TimeAvg)

Native units: m^-3.

Restart value of the exponentially time-averaged predictor electron-density correction.

  • plot_label: Time-averaged predictor electron-density correction

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz, nd_connect_list for axisymmetric geometry; rplane_rz, rplane_nd_connect_list for stellarator right-plane geometry. Connectivity is zero-based. Node-centered field with continuous linear barycentric basis functions on each triangle.

  • coordinate_context: Axisymmetric predictor state on the right-plane nodal geometry used by the Poisson solve; a flux-surface average/median remains represented at mesh nodes.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

  • centering: node

Source: XGC_core/cpp/solvers/potential_predictor.cpp:623.

restart_dir/xgc.restart*.bp: prefix + “max_used_gid” (SpeciesPrefixMaxUsedGid)

Native units: Non-negative integer.

Largest globally unique marker-particle ID already assigned for this species.

  • plot_label: Largest assigned marker ID ({name})

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

Source: XGC_core/cpp/plasma/species.hpp:114.

restart_dir/xgc.restart*.bp: rhs_offset0 (ChargeRhsOffset0)

Native units: m^-3.

Axisymmetric right-hand-side density offset used when restarting the Poisson/charge solve.

  • plot_label: Axisymmetric charge-solve density offset

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

  • centering: node

Source: XGC_core/cpp/solvers/charge.hpp:132.

restart_dir/xgc.restart*.bp: sheath_ilost (SheathSheathIlost)

Native units: C.

Checkpointed ion charge-loss accumulator on toroidal planes and wall nodes.

  • plot_label: $Q_{\mathrm{lost,i}}$

  • mesh_context: xgc.mesh.bp: grid_wall_nodes contains zero-based indices into rz (R,Z in m) in the ordered mesh-boundary polygon sequence produced by order_wall. Wall-node values follow this sequence; successive entries and the final-to-first pair define boundary segments for a valid closed wall. The wall-node list is not the triangle connectivity array nd_connect_list.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

  • centering: wall_node

Source: XGC_core/cpp/sheath/sheath.cpp:771.

restart_dir/xgc.restart*.bp: sheath_lost (SheathSheathLost)

Native units: C.

Checkpointed signed net charge-loss accumulator on toroidal planes and wall nodes.

  • plot_label: $Q_{\mathrm{lost}}$

  • mesh_context: xgc.mesh.bp: grid_wall_nodes contains zero-based indices into rz (R,Z in m) in the ordered mesh-boundary polygon sequence produced by order_wall. Wall-node values follow this sequence; successive entries and the final-to-first pair define boundary segments for a valid closed wall. The wall-node list is not the triangle connectivity array nd_connect_list.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

  • centering: wall_node

Source: XGC_core/cpp/sheath/sheath.cpp:766.

restart_dir/xgc.restart*.bp: sheath_nphi (SheathSheathNphi)

Native units: Positive integer.

Number of toroidal planes in sheath checkpoint arrays.

  • plot_label: $N_{\varphi,\mathrm{sh}}$

Source: XGC_core/cpp/sheath/sheath.cpp:756.

restart_dir/xgc.restart*.bp: sheath_nwall (SheathSheathNwall)

Native units: Positive integer.

Number of wall elements in sheath checkpoint arrays.

  • plot_label: $N_{\mathrm{wall,sh}}$

Source: XGC_core/cpp/sheath/sheath.cpp:751.

restart_dir/xgc.restart*.bp: sheath_pot (SheathSheathPot)

Native units: V.

Checkpointed sheath potential on toroidal planes and wall elements.

  • plot_label: $\phi_{\mathrm{sh}}$ (checkpoint)

  • mesh_context: xgc.mesh.bp: grid_wall_nodes contains zero-based indices into rz (R,Z in m) in the ordered mesh-boundary polygon sequence produced by order_wall. Wall-node values follow this sequence; successive entries and the final-to-first pair define boundary segments for a valid closed wall. The wall-node list is not the triangle connectivity array nd_connect_list.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

  • centering: wall_node

Source: XGC_core/cpp/sheath/sheath.cpp:761.

restart_dir/xgc.restart*.bp: sp_name + “_f0_delta_T” (SpeciesSpNameF0DeltaT)

Native units: eV.

Restart value of species temperature perturbation added to the analytic f0 background on mesh nodes.

  • plot_label: $\delta T_0$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

  • centering: node

Source: XGC_core/cpp/plasma/species.cpp:1011.

restart_dir/xgc.restart*.bp: sp_name + “_f0_delta_n” (SpeciesSpNameF0DeltaN)

Native units: m^-3.

Restart value of species density perturbation added to the analytic f0 background on mesh nodes.

  • plot_label: $\delta n_0$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

  • centering: node

Source: XGC_core/cpp/plasma/species.cpp:1001.

restart_dir/xgc.restart*.bp: sp_name + “_f0_delta_u” (SpeciesSpNameF0DeltaU)

Native units: m s^-1.

Restart value of species flow perturbation added to the analytic f0 background on mesh nodes.

  • plot_label: $\delta u_0$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

  • centering: node

Source: XGC_core/cpp/plasma/species.cpp:1006.

restart_dir/xgc.restart*.bp: sp_name + “_f0g” (SpeciesSpNameF0g)

Native units: m^-3 eV^-1.

Species residual/grid contribution to the f0 distribution function on [toroidal plane, perpendicular velocity, mesh node, parallel velocity].

  • plot_label: Residual f0 distribution ({name})

  • axes: toroidal_plane,velocity_mu,mesh_node,velocity_parallel

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh. Velocity-grid coordinates and quadrature are defined by xgc.f0.mesh.bp: f0_nmu, f0_nvp, f0_dsmu, f0_dvp, f0_vp_max, f0_fg_T_ev, f0_grid_vol_vonly; see normalization.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

  • centering: node

  • normalization: Velocity coordinates use xgc.f0.mesh.bp: f0_nmu/f0_nvp are interval counts; there are f0_nmu+1 perpendicular and 2*f0_nvp+1 parallel vertices. At zero-based indices j,k, vperp_norm=j*f0_dsmu except vperp_norm[0]=f0_dsmu/3, and vpara_norm=k*f0_dvp-f0_vp_max. Physical speed is normalized speed times sqrt(EV_2_J*f0_fg_T_ev[species,node]/mass_kg), EV_2_J=1.6022e-19 J/eV; mass_kg=xgc.units.bp:<species>_ptl_mass_au*1.6720e-27 kg. These are the native XGC constants EV_2_J and PROTON_MASS, not the exact SI elementary charge or unified atomic mass constant. This is XGC thermal_velocity; the analysis reader uses abs(charge_C)*T/m and agrees for singly charged species only. Stored f/f0g already absorbs the perpendicular Jacobian. Density is f0_grid_vol_vonly[species,node]*sum(mu_vol[j]*vp_vol[k]*f[j,node,k]); endpoint factors are 0.5, interior factors 1. Do not add another vperp, dsmu*dvp or 2*pi factor to that formula. f0n and accumulation counters use these coordinates but are not distributions to integrate.

Source: XGC_core/cpp/plasma/species.cpp:989.

restart_dir/xgc.restart*.bp: sp_name + “flag” (SpeciesSpNameFlag)

Native units: integer flag.

Packed 64-bit ESC_PTL particle state for this checkpoint species, including low status bits, the stored escape-step payload, and the signed fixed-point weight-change payload; this is not the decoded low-bit flag array in xgc.escaped_ptls.*.bp.

  • plot_label: Packed particle state ({name})

  • axes: particle

  • mesh_context: Particle sample, not a mesh field. The same species’ phase array contains row-wise cylindrical R (m), Z (m), phi (rad) in columns 0,1,2; use gid to identify rows/particles across output slices. MPI repartitioning can change row order. Visualize as points or perform a stated statistical/binning procedure, not triangle interpolation. Other phase columns and units are specified in description.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

  • centering: particle

  • normalization: Decode using the native ESC_PTL packing constants and helpers: ps_flag_bits=8, ps_step_bits=17, ps_dw_factor_bits=28. The escaped-particle diagnostic writes the decoded flag, step, and dw separately.

  • related_variables: Same checkpoint species phase, gid, num_arr, off_arr; time and timestep.

Source: XGC_core/cpp/plasma/species.cpp:876.

restart_dir/xgc.restart*.bp: sp_name + “gid” (SpeciesSpNameGid)

Native units: Integer index.

Globally unique marker-particle IDs for this species.

  • plot_label: Marker ID ({name})

  • axes: particle

  • mesh_context: Particle sample, not a mesh field. The same species’ phase array contains row-wise cylindrical R (m), Z (m), phi (rad) in columns 0,1,2; use gid to identify rows/particles across output slices. MPI repartitioning can change row order. Visualize as points or perform a stated statistical/binning procedure, not triangle interpolation. Other phase columns and units are specified in description.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

  • centering: particle

  • related_variables: Same checkpoint species phase, gid, num_arr, off_arr; time and timestep.

Source: XGC_core/cpp/plasma/species.cpp:868.

restart_dir/xgc.restart*.bp: sp_name + “nphase” (SpeciesSpNameNphase)

Native units: Positive integer.

Number of columns in the species particle phase checkpoint array.

  • plot_label: Checkpoint phase-component count ({name})

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

Source: XGC_core/cpp/plasma/species.cpp:844.

restart_dir/xgc.restart*.bp: sp_name + “num_arr” (SpeciesSpNameNumArr)

Native units: Non-negative integer.

Number of checkpointed marker particles for this species written by each MPI rank.

  • plot_label: Rank-local checkpoint marker count ({name})

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

Source: XGC_core/cpp/plasma/species.cpp:854.

restart_dir/xgc.restart*.bp: sp_name + “num_total” (SpeciesSpNameNumTotal)

Native units: Non-negative integer.

Total number of checkpointed marker particles for this species across all MPI ranks.

  • plot_label: Checkpoint marker count ({name})

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

Source: XGC_core/cpp/plasma/species.cpp:849.

restart_dir/xgc.restart*.bp: sp_name + “off_arr” (SpeciesSpNameOffArr)

Native units: Non-negative integer.

Global particle-array offset for this species for each MPI rank in the checkpoint.

  • plot_label: Rank-local checkpoint particle offset ({name})

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

Source: XGC_core/cpp/plasma/species.cpp:859.

restart_dir/xgc.restart*.bp: sp_name + “phase” (SpeciesSpNamePhase)

Native units: [m, m, rad, m, dimensionless, dimensionless, J/T, dimensionless, m^-3 eV^(-3/2)].

Species particle phase checkpoint array with columns [R, Z, phi, rho, w1, w2, mu, w0, g], where \(R\), \(Z\), and \(\\phi\) are the cylindrical coordinates of the particle, \(\\rho\) is the parallel gyroradius (proxy for \(v_{\\parallel}\)), \(w_1\) is the delta-f weight, \(w_2\) is a weight used in the weight evolution equation, \(\\mu\) is the magnetic moment, \(w_0\) is the full-f weight (from importance sampling), and \(g\) is the phase-space marker density.

  • plot_label: Checkpoint particle phase space ({name})

  • axes: particle,phase_component

  • mesh_context: Particle sample, not a mesh field. The same species’ phase array contains row-wise cylindrical R (m), Z (m), phi (rad) in columns 0,1,2; use gid to identify rows/particles across output slices. MPI repartitioning can change row order. Visualize as points or perform a stated statistical/binning procedure, not triangle interpolation. Other phase columns and units are specified in description.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

  • centering: particle

  • normalization: The g column is marker density in XGC velocity-space normalization, with units m^-3 eV^(-3/2); f=w0*g. For a Maxwellian this normalization is proportional to n*T_eV^(-3/2)*exp(-E_eV/T_eV), with the SI velocity normalization factor (mass/(2*pi*EV_2_J))^(3/2) removed. This is distinct from the spatial triangle marker_den in xgc.loading.*.bp.

  • related_variables: Same checkpoint species phase, gid, num_arr, off_arr; time and timestep.

Source: XGC_core/cpp/plasma/species.cpp:886.

restart_dir/xgc.restart*.bp: species.short_label.str() + “_f0_df0g_persistent” (PlasmaSpeciesShortLabelStrF0Df0gPersistent)

Native units: m^-3 eV^-1.

Persistent residual df0g distribution retained for later particle transfer on [toroidal plane, perpendicular velocity, mesh node, parallel velocity].

  • plot_label: Persistent residual f0 distribution ({name})

  • axes: toroidal_plane,velocity_mu,mesh_node,velocity_parallel

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh. Velocity-grid coordinates and quadrature are defined by xgc.f0.mesh.bp: f0_nmu, f0_nvp, f0_dsmu, f0_dvp, f0_vp_max, f0_fg_T_ev, f0_grid_vol_vonly; see normalization.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

  • centering: node

  • normalization: Velocity coordinates use xgc.f0.mesh.bp: f0_nmu/f0_nvp are interval counts; there are f0_nmu+1 perpendicular and 2*f0_nvp+1 parallel vertices. At zero-based indices j,k, vperp_norm=j*f0_dsmu except vperp_norm[0]=f0_dsmu/3, and vpara_norm=k*f0_dvp-f0_vp_max. Physical speed is normalized speed times sqrt(EV_2_J*f0_fg_T_ev[species,node]/mass_kg), EV_2_J=1.6022e-19 J/eV; mass_kg=xgc.units.bp:<species>_ptl_mass_au*1.6720e-27 kg. These are the native XGC constants EV_2_J and PROTON_MASS, not the exact SI elementary charge or unified atomic mass constant. This is XGC thermal_velocity; the analysis reader uses abs(charge_C)*T/m and agrees for singly charged species only. Stored f/f0g already absorbs the perpendicular Jacobian. Density is f0_grid_vol_vonly[species,node]*sum(mu_vol[j]*vp_vol[k]*f[j,node,k]); endpoint factors are 0.5, interior factors 1. Do not add another vperp, dsmu*dvp or 2*pi factor to that formula. f0n and accumulation counters use these coordinates but are not distributions to integrate.

Source: XGC_core/cpp/plasma/plasma.cpp:250.

restart_dir/xgc.restart*.bp: species.short_label.str() + “_f0_df0g_persistent_counter” (PlasmaSpeciesShortLabelStrF0Df0gPersistentCounter)

Native units: source-application count.

Age counter for persistent residual df0g entries on [toroidal plane, perpendicular velocity, mesh node, parallel velocity].

  • plot_label: Persistent f0 residual age ({name})

  • axes: toroidal_plane,velocity_mu,mesh_node,velocity_parallel

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh. Velocity-grid coordinates and quadrature are defined by xgc.f0.mesh.bp: f0_nmu, f0_nvp, f0_dsmu, f0_dvp, f0_vp_max, f0_fg_T_ev, f0_grid_vol_vonly; see normalization.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

  • centering: node

  • normalization: Count of source applications/age at each phase-space grid entry, not a distribution function. No density integral or physical elapsed time is implied.

Source: XGC_core/cpp/plasma/plasma.cpp:257.

restart_dir/xgc.restart*.bp: time (SmlTime)

Native units: s.

Simulation time of step in seconds.

  • plot_label: $t$

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times. Clock convention: in the regular main loop gstep is incremented before diagnostic calls, but sml.time is incremented afterward. A checkpoint written later at the same gstep therefore records time larger by one physical timestep. Use the written time, global step and product/integration stage together; equal gstep alone does not establish simultaneous diagnostic and checkpoint states.

Source: XGC_core/cpp/main/sml.cpp:17.

restart_dir/xgc.restart*.bp: timestep (SmlTimestep)

Native units: Non-negative integer.

Simulation time step stored in the checkpoint.

  • plot_label: Checkpoint time step

  • time_context: Restart snapshot at restart_dir/xgc.restart*.bp: time is physical simulation time in s and timestep is the global integration-step index. These are checkpoint states, not diagnostic-cadence averages. Some variables hold histories/cumulative controller state as described; do not interpret a history axis as successive diagnostic times.

Source: XGC_core/cpp/main/sml.cpp:12.

restart_dir/xgc.restart*.bp: totalpe (DomainDecompositionTotalpe)

Native units: Positive integer.

Number of MPI ranks used when the checkpoint was written.

  • plot_label: MPI rank count

Source: XGC_core/cpp/domain_decomposition/domain_decomposition.cpp:290.

restart_dir/xgc.restart*.bp: vpdata (PlasmaVpdata)

Native units: Positive integer.

Number of parallel velocity grid points in f0 checkpoint arrays.

  • plot_label: $N_{v,\parallel}$

Source: XGC_core/cpp/plasma/plasma.cpp:203.

triangle_color_scheme.bp: avg_set32_density (TriangleColorSchemeAvgSet32Density)

Native units: dimensionless.

Triangle-lane utilization over contiguous blocks of at most 32 triangles: total triangle count divided by the sum of block_size times the number of distinct colors in each block.

  • plot_label: Triangle color-pass lane utilization

Source: XGC_core/cpp/grid/triangle_color_scheme.cpp:177.

triangle_color_scheme.bp: color_offsets (TriangleColorSchemeColorOffsets)

Native units: Non-negative integer.

Zero-based start offsets of triangle lists in triangle_ids_by_color; final entry is the total number of triangles.

  • plot_label: Color-group offset

  • coordinate_context: Same triangle_color_scheme.bp: zero-based color c selects triangle_ids_by_color[color_offsets[c]:color_offsets[c+1]], with zero-based triangle IDs into nodeid. These are disjoint categorical groups, not mesh coordinates.

Source: XGC_core/cpp/grid/triangle_color_scheme.cpp:151.

triangle_color_scheme.bp: coord (TriangleColorSchemeCoord)

Native units: m.

Cylindrical R and Z coordinates of mesh vertices in this standalone debug product.

  • plot_label: $(R,Z)$

  • axes: mesh_node,rz_component

  • coordinate_context: Standalone triangle_color_scheme.bp:coord[:,0] is R in m and coord[:,1] is Z in m.

Source: XGC_core/cpp/grid/triangle_color_scheme.cpp:134.

triangle_color_scheme.bp: max_color_bound (TriangleColorSchemeMaxColorBound)

Native units: Positive integer.

Upper bound on the number of colors used by the coloring algorithm.

  • plot_label: Triangle-color upper bound

Source: XGC_core/cpp/grid/triangle_color_scheme.cpp:172.

triangle_color_scheme.bp: max_num_t_node (TriangleColorSchemeMaxNumTNode)

Native units: Non-negative integer.

Maximum number of triangles adjacent to one mesh node.

  • plot_label: Maximum node-adjacent triangle count

Source: XGC_core/cpp/grid/triangle_color_scheme.cpp:167.

triangle_color_scheme.bp: ncolor (TriangleColorSchemeNcolor)

Native units: Positive integer.

Number of triangle colors.

  • plot_label: Triangle-color count

Source: XGC_core/cpp/grid/triangle_color_scheme.cpp:162.

triangle_color_scheme.bp: nodeid (TriangleColorSchemeNodeid)

Native units: Integer index.

Triangle vertices as zero-based mesh-node indices into coord.

  • plot_label: Triangle vertex mesh-node index

  • axes: triangle,triangle_vertex

  • coordinate_context: Standalone triangle_color_scheme.bp:nodeid has three zero-based indices into coord per triangle.

Source: XGC_core/cpp/grid/triangle_color_scheme.cpp:139.

triangle_color_scheme.bp: triangle_color (TriangleColorSchemeTriangleColor)

Native units: Integer index.

Zero-based color assigned to each triangle; triangles sharing a vertex have different colors.

  • plot_label: Triangle color ID

  • axes: triangle

  • mesh_context: Same triangle_color_scheme.bp:coord gives R,Z vertices in m; nodeid gives zero-based triangle connectivity. Color is a cell-centered categorical label, not a continuous nodal field.

  • centering: triangle_category

Source: XGC_core/cpp/grid/triangle_color_scheme.cpp:146.

triangle_color_scheme.bp: triangle_ids_by_color (TriangleColorSchemeTriangleIdsByColor)

Native units: Integer index.

Zero-based triangle IDs packed by color; use color_offsets to slice each color group.

  • plot_label: Triangle ID grouped by color

  • coordinate_context: Same triangle_color_scheme.bp: zero-based color c selects triangle_ids_by_color[color_offsets[c]:color_offsets[c+1]], with zero-based triangle IDs into nodeid. These are disjoint categorical groups, not mesh coordinates.

Source: XGC_core/cpp/grid/triangle_color_scheme.cpp:157.

xgc.2d/3d.*.bp: As_rho_ff (Diag3dAsRhoFf)

Native units: T m.

Field-following As components at gyrocenter

  • plot_label: $A_\parallel^{(s)}$ (field-following)

  • axes: toroidal_plane,mesh_node,field_following_side

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: At zero gyro-radius (rho index 0), two scalar As contributions in field-following coordinates: side 0 is the left plane (phi_left), side 1 the right plane (phi_right), traced to the local mesh. The side axis is distinct from toroidal_plane. Interpolate the two contributions with field-following plane weights.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:655.

xgc.2d/3d.*.bp: aparh (Diag3dAparh)

Native units: T m.

Perturbed parallel vector potential A_h

  • plot_label: $A_{\parallel}^{(h)}$

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz, nd_connect_list for axisymmetric geometry; rplane_rz, rplane_nd_connect_list for stellarator right-plane geometry. Connectivity is zero-based. Node-centered field with continuous linear barycentric basis functions on each triangle.

  • coordinate_context: Cylindrical R,Z,phi. Field planes are at the same-product phi/phi_right; density and deposited moments are on midplanes halfway between phi_left and phi_right. Toroidal periodicity is xgc.mesh.bp: wedge_angle, delta_phi (wedge repeats 2*pi/wedge_angle times). A toroidal_plane axis precedes mesh_node where present. For toroidal interpolation, follow magnetic field lines using xgc.bfield.bp and the field-following projection matrices xgc.ff_1dp_fwd.bp/xgc.ff_1dp_rev.bp when available; straight interpolation between equal node IDs is not a field-line interpolation.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/a_field(:)/parallel

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:619.

xgc.2d/3d.*.bp: apars (Diag3dApars)

Native units: T m.

Symplectic part As of the perturbed parallel magnetic vector potential.

  • plot_label: $A_{\parallel}^{(s)}$

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz, nd_connect_list for axisymmetric geometry; rplane_rz, rplane_nd_connect_list for stellarator right-plane geometry. Connectivity is zero-based. Node-centered field with continuous linear barycentric basis functions on each triangle.

  • coordinate_context: Cylindrical R,Z,phi. Field planes are at the same-product phi/phi_right; density and deposited moments are on midplanes halfway between phi_left and phi_right. Toroidal periodicity is xgc.mesh.bp: wedge_angle, delta_phi (wedge repeats 2*pi/wedge_angle times). A toroidal_plane axis precedes mesh_node where present. For toroidal interpolation, follow magnetic field lines using xgc.bfield.bp and the field-following projection matrices xgc.ff_1dp_fwd.bp/xgc.ff_1dp_rev.bp when available; straight interpolation between equal node IDs is not a field-line interpolation.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/a_field(:)/parallel

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:647.

xgc.2d/3d.*.bp: dAs_rho_ff (Diag3dDAsRhoFf)

Native units: T.

Negative spatial derivatives of the symplectic parallel vector potential As at zero gyro-radius, in field-following coordinates.

  • plot_label: $-\nabla A_\parallel^{(s)}$ (field-following)

  • axes: toroidal_plane,mesh_node,derivative_component,field_following_side

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: At zero gyro-radius (rho index 0), negative spatial derivatives of As in field-following coordinates. Side 0 is the left plane (phi_left), side 1 the right plane (phi_right), traced to the local mesh. Derivative slots 0,1,2 are the two poloidal basis components and the parallel component. The poloidal basis depends on xgc.grad_rz.bp:basis (0=psi/theta unit vectors, 1=R/Z). For basis=0, e_psi=(psi_R,psi_Z)/|grad psi| and e_theta=(-psi_Z,psi_R)/|grad psi|; rotate the first two components into R/Z before treating them as cylindrical vectors. The parallel slot is not the cylindrical toroidal component. xgc.grad_rz.bp is omitted by MINIMAL_OUTPUT and contains only the historical rank-zero plane; it does not provide every stellarator plane basis. Without the matching node-wise basis this output alone does not identify the poloidal component frame.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Spatial derivatives of As with respect to physical length; no conversion is applied to the stored values.

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:660.

xgc.2d/3d.*.bp: dBphi (Diag3dDBphi)

Native units: T.

Toroidal magnetic field perturbation in Tesla

  • plot_label: $\delta B_\varphi$

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: XGC1 right field plane at same-product phi_right (also historical phi), after conversion from field-following coordinates to real edge-plane coordinates. Density/current deposits include averaging of adjacent contributions at that edge plane. Use the corresponding right-plane geometry; axisymmetric geometry reuses xgc.mesh.bp:rz,nd_connect_list, while stellarator right-plane geometry is rplane_rz,rplane_nd_connect_list. Axisymmetric XGCa has one poloidal mesh.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/b_field(:)/phi

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:788.

xgc.2d/3d.*.bp: dBpsi (Diag3dDBpsi)

Native units: T.

Radial magnetic field perturbation in Tesla

  • plot_label: $\delta B_\psi$

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: XGC1 right field plane at same-product phi_right (also historical phi), after conversion from field-following coordinates to real edge-plane coordinates. Density/current deposits include averaging of adjacent contributions at that edge plane. Use the corresponding right-plane geometry; axisymmetric geometry reuses xgc.mesh.bp:rz,nd_connect_list, while stellarator right-plane geometry is rplane_rz,rplane_nd_connect_list. Axisymmetric XGCa has one poloidal mesh.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/b_field(:)/radial

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:777.

xgc.2d/3d.*.bp: dBtheta (Diag3dDBtheta)

Native units: T.

Poloidal magnetic field perturbation in Tesla

  • plot_label: $\delta B_\theta$

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: XGC1 right field plane at same-product phi_right (also historical phi), after conversion from field-following coordinates to real edge-plane coordinates. Density/current deposits include averaging of adjacent contributions at that edge plane. Use the corresponding right-plane geometry; axisymmetric geometry reuses xgc.mesh.bp:rz,nd_connect_list, while stellarator right-plane geometry is rplane_rz,rplane_nd_connect_list. Axisymmetric XGCa has one poloidal mesh.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/b_field(:)/poloidal

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:782.

xgc.2d/3d.*.bp: ddpotdt (Diag3dDdpotdt)

Native units: V s^-1.

Time derivative of perturbed potential

  • plot_label: $\partial\delta\phi/\partial t$

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Field-following right-side (side 1) values indexed by midplane nodes. The right field-plane finite-difference potential derivative is projected along magnetic field lines onto xgc.mesh.bp:rz,nd_connect_list; the stored value is not the average of the left and right sides. Midplane angle=(phi_left+phi_right)/2, with periodic unwrapping at the wedge boundary; the historical phi array denotes the right-plane angle in XGC1.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:562.

xgc.2d/3d.*.bp: depsidt (Diag3dDepsidt)

Native units: V m^-1 s^-1.

Time derivative of electric field across flux surfaces

  • plot_label: $\partial E_\psi/\partial t$

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry xgc.mesh.bp:rz,nd_connect_list; toroidal angle=(phi_left+phi_right)/2 on each plane, with periodic unwrapping at the wedge boundary. The historical phi array is the right-plane angle in XGC1; use phi_left/phi_right to locate midplane data.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:815.

xgc.2d/3d.*.bp: dethetadt (Diag3dDethetadt)

Native units: V m^-1 s^-1.

Time derivative of electric field along flux surfaces

  • plot_label: $\partial E_\theta/\partial t$

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry xgc.mesh.bp:rz,nd_connect_list; toroidal angle=(phi_left+phi_right)/2 on each plane, with periodic unwrapping at the wedge boundary. The historical phi array is the right-plane angle in XGC1; use phi_left/phi_right to locate midplane data.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:821.

xgc.2d/3d.*.bp: dpot (Diag3dDpot)

Native units: V.

Perturbed potential in volt.

  • plot_label: $\delta\phi$

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz, nd_connect_list for axisymmetric geometry; rplane_rz, rplane_nd_connect_list for stellarator right-plane geometry. Connectivity is zero-based. Node-centered field with continuous linear barycentric basis functions on each triangle.

  • coordinate_context: Cylindrical R,Z,phi. Field planes are at the same-product phi/phi_right; density and deposited moments are on midplanes halfway between phi_left and phi_right. Toroidal periodicity is xgc.mesh.bp: wedge_angle, delta_phi (wedge repeats 2*pi/wedge_angle times). A toroidal_plane axis precedes mesh_node where present. For toroidal interpolation, follow magnetic field lines using xgc.bfield.bp and the field-following projection matrices xgc.ff_1dp_fwd.bp/xgc.ff_1dp_rev.bp when available; straight interpolation between equal node IDs is not a field-line interpolation.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/phi_potential(:)/values

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:531.

xgc.2d/3d.*.bp: dpot_es (Diag3dDpotEs)

Native units: V.

Electrostatic potential

  • plot_label: $\phi$

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz, nd_connect_list for axisymmetric geometry; rplane_rz, rplane_nd_connect_list for stellarator right-plane geometry. Connectivity is zero-based. Node-centered field with continuous linear barycentric basis functions on each triangle.

  • coordinate_context: Electrostatic solver potential on the right-plane mesh, at phi_right in this product; for stellarators use xgc.mesh.bp:rplane_rz,rplane_nd_connect_list.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/phi_potential(:)/values

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:584.

xgc.2d/3d.*.bp: dpot_n0 (Diag3dDpotN0)

Native units: V.

n=0 component of perturbed potential

  • plot_label: $\delta\phi_{n=0}$

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz, nd_connect_list for axisymmetric geometry; rplane_rz, rplane_nd_connect_list for stellarator right-plane geometry. Connectivity is zero-based. Node-centered field with continuous linear barycentric basis functions on each triangle.

  • coordinate_context: Axisymmetric potential on the right-plane nodal geometry. For stellarators use xgc.mesh.bp:rplane_rz,rplane_nd_connect_list. No toroidal_plane axis is stored.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/phi_potential(:)/values

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:610.

xgc.2d/3d.*.bp: eden (Diag3dEden)

Native units: m^-3.

Electron field-solve number-density contribution at mesh nodes.

  • plot_label: $\delta n_e$

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: XGC1 right field plane at same-product phi_right (also historical phi), after conversion from field-following coordinates to real edge-plane coordinates. Density/current deposits include averaging of adjacent contributions at that edge plane. Use the corresponding right-plane geometry; axisymmetric geometry reuses xgc.mesh.bp:rz,nd_connect_list, while stellarator right-plane geometry is rplane_rz,rplane_nd_connect_list. Axisymmetric XGCa has one poloidal mesh.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Electron deposition uses charge factor +1 rather than the electron electric charge. The stored value is a number-density contribution in m^-3; multiply by minus the positive elementary charge for its charge-density contribution in C/m^3. Its weights and optional f0-grid contribution follow the field solve; do not assume it is the complete physical electron density.

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:546.

xgc.2d/3d.*.bp: ejpar (Diag3dEjpar)

Native units: m^-2 s^-1.

Electron parallel number-flux moment deposited for the field solve, stored with positive electron number weight. The electron charge sign and elementary-charge factor are not applied.

  • plot_label: $\Gamma_{\parallel,e}$

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: XGC1 right field plane at same-product phi_right (also historical phi), after conversion from field-following coordinates to real edge-plane coordinates. Density/current deposits include averaging of adjacent contributions at that edge plane. Use the corresponding right-plane geometry; axisymmetric geometry reuses xgc.mesh.bp:rz,nd_connect_list, while stellarator right-plane geometry is rplane_rz,rplane_nd_connect_list. Axisymmetric XGCa has one poloidal mesh.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Multiply the stored number-flux moment by -UNIT_CHARGE (-1.6022e-19 C) to obtain this deposited contribution to parallel current density in A m^-2. This field-solve contribution can use delta-f/control-variate weights and does not include all analytic-background current terms.

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:594.

xgc.2d/3d.*.bp: epara (Diag3dEpara)

Native units: V/m.

Parallel (to the magnetic field) electric field

  • plot_label: $E_\parallel$

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: XGC1 right field plane at same-product phi_right (also historical phi), after conversion from field-following coordinates to real edge-plane coordinates. Density/current deposits include averaging of adjacent contributions at that edge plane. Use the corresponding right-plane geometry; axisymmetric geometry reuses xgc.mesh.bp:rz,nd_connect_list, while stellarator right-plane geometry is rplane_rz,rplane_nd_connect_list. Axisymmetric XGCa has one poloidal mesh.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/e_field(:)/parallel

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:830.

xgc.2d/3d.*.bp: epara2 (Diag3dEpara2)

Native units: V/m.

Right-plane reconstruction of dAdt_guess used to advance the symplectic parallel vector potential As in the IdealMHD pullback method; zero for other pullback methods.

  • plot_label: $\partial A_\parallel^{(s)}/\partial t$ (IdealMHD predictor)

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: XGC1 right field plane at same-product phi_right (also historical phi), after conversion from field-following coordinates to real edge-plane coordinates. Density/current deposits include averaging of adjacent contributions at that edge plane. Use the corresponding right-plane geometry; axisymmetric geometry reuses xgc.mesh.bp:rz,nd_connect_list, while stellarator right-plane geometry is rplane_rz,rplane_nd_connect_list. Axisymmetric XGCa has one poloidal mesh.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:840.

xgc.2d/3d.*.bp: epsi (Diag3dEpsi)

Native units: V/m.

Radial electric field in Volt per meter

  • plot_label: $E_\psi$

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: XGC1 right field plane at same-product phi_right (also historical phi), after conversion from field-following coordinates to real edge-plane coordinates. Density/current deposits include averaging of adjacent contributions at that edge plane. Use the corresponding right-plane geometry; axisymmetric geometry reuses xgc.mesh.bp:rz,nd_connect_list, while stellarator right-plane geometry is rplane_rz,rplane_nd_connect_list. Axisymmetric XGCa has one poloidal mesh.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/e_field(:)/radial

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:801.

xgc.2d/3d.*.bp: etheta (Diag3dEtheta)

Native units: V/m.

Poloidal electric field in Volt per meter

  • plot_label: $E_\theta$

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: XGC1 right field plane at same-product phi_right (also historical phi), after conversion from field-following coordinates to real edge-plane coordinates. Density/current deposits include averaging of adjacent contributions at that edge plane. Use the corresponding right-plane geometry; axisymmetric geometry reuses xgc.mesh.bp:rz,nd_connect_list, while stellarator right-plane geometry is rplane_rz,rplane_nd_connect_list. Axisymmetric XGCa has one poloidal mesh.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/e_field(:)/poloidal

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:806.

xgc.2d/3d.*.bp: gstep (Diag3dGstep)

Native units: Non-negative integer.

XGC global simulation time step for this output slice.

  • plot_label: Global time step

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:390.

xgc.2d/3d.*.bp: iden (Diag3dIden)

Native units: m^-3.

Ion density at each mesh node interpolated from surrounding marker particles

  • plot_label: $\overline{\delta n_{i,gc}}$

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry xgc.mesh.bp:rz,nd_connect_list; toroidal angle=(phi_left+phi_right)/2 on each plane, with periodic unwrapping at the wedge boundary. The historical phi array is the right-plane angle in XGC1; use phi_left/phi_right to locate midplane data.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/ion(:)/density(:)/values

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:537.

xgc.2d/3d.*.bp: ijpar (Diag3dIjpar)

Native units: m^-2 s^-1.

Ion charge-number-weighted parallel number-flux moment deposited for the field solve. Sum over ion species, including charge_eu weights; not yet multiplied by elementary charge.

  • plot_label: $\sum_i Z_i\Gamma_{\parallel,i}$

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: XGC1 right field plane at same-product phi_right (also historical phi), after conversion from field-following coordinates to real edge-plane coordinates. Density/current deposits include averaging of adjacent contributions at that edge plane. Use the corresponding right-plane geometry; axisymmetric geometry reuses xgc.mesh.bp:rz,nd_connect_list, while stellarator right-plane geometry is rplane_rz,rplane_nd_connect_list. Axisymmetric XGCa has one poloidal mesh.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Multiply the stored number-flux moment by +UNIT_CHARGE (+1.6022e-19 C) to obtain this deposited contribution to parallel current density in A m^-2. This field-solve contribution can use delta-f/control-variate weights and does not include all analytic-background current terms.

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:575.

xgc.2d/3d.*.bp: iphi (Diag3dIphi)

Native units: Non-negative integer.

Index of local plane (0-based)

  • plot_label: $i_\varphi$

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:438.

xgc.2d/3d.*.bp: ne00_deposited_fsa (PotentialPredictorNe00DepositedFsa)

Native units: m^-3.

Instantaneous unprofiled electron-density FSA using sml_field_solver_fsa_method after electron charge deposition, excluding private and wall regions

  • plot_label: Deposited flux-averaged electron density

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz, nd_connect_list for axisymmetric geometry; rplane_rz, rplane_nd_connect_list for stellarator right-plane geometry. Connectivity is zero-based. Node-centered field with continuous linear barycentric basis functions on each triangle.

  • coordinate_context: Axisymmetric predictor state on the right-plane nodal geometry used by the Poisson solve; a flux-surface average/median remains represented at mesh nodes.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

Source: XGC_core/cpp/solvers/potential_predictor.cpp:663.

xgc.2d/3d.*.bp: ne00_deposited_median (PotentialPredictorNe00DepositedMedian)

Native units: m^-3.

Instantaneous node-volume-weighted median of aligned surface-node electron density, mapped to nodes with sml_field_solver_fsa_method after electron charge deposition, before diffusion or radial continuation and excluding the private and wall regions

  • plot_label: Deposited median electron density

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz, nd_connect_list for axisymmetric geometry; rplane_rz, rplane_nd_connect_list for stellarator right-plane geometry. Connectivity is zero-based. Node-centered field with continuous linear barycentric basis functions on each triangle.

  • coordinate_context: Axisymmetric predictor state on the right-plane nodal geometry used by the Poisson solve; a flux-surface average/median remains represented at mesh nodes.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

Source: XGC_core/cpp/solvers/potential_predictor.cpp:672.

xgc.2d/3d.*.bp: ne00_predictor (PotentialPredictorNe00Predictor)

Native units: m^-3.

Electron-density correction used by the axisymmetric predictor after FSA or median selection, configured radial handling, and exponential time averaging; FSA reductions and surface-to-node mappings follow sml_field_solver_fsa_method, excluding private and wall regions

  • plot_label: $\delta n_{e,00}$ (predictor)

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz, nd_connect_list for axisymmetric geometry; rplane_rz, rplane_nd_connect_list for stellarator right-plane geometry. Connectivity is zero-based. Node-centered field with continuous linear barycentric basis functions on each triangle.

  • coordinate_context: Axisymmetric predictor state on the right-plane nodal geometry used by the Poisson solve; a flux-surface average/median remains represented at mesh nodes.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

Source: XGC_core/cpp/solvers/potential_predictor.cpp:656.

xgc.2d/3d.*.bp: nnode (Diag3dNnode)

Native units: Positive integer.

Number of mesh nodes per poloidal plane.

  • plot_label: $N_n$

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:383.

xgc.2d/3d.*.bp: node_to_wall (Diag3dNodeToWall)

Native units: Integer index.

Mapping from mesh node to wall index

  • plot_label: Node-to-wall index

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Integer lookup from zero-based mesh node to zero-based wall index; -1 denotes a non-wall node. Use non-negative entries to index wall-node arrays. Do not interpolate integer labels.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node_index

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:424.

xgc.2d/3d.*.bp: nphi (Diag3dNphi)

Native units: Positive integer.

Number of toroidal mesh planes \(N_{\\mathrm{planes}}\).

  • plot_label: $N_\varphi$

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:433.

xgc.2d/3d.*.bp: nwall (Diag3dNwall)

Native units: Positive integer.

Number of wall vertices

  • plot_label: $N_w$

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:409.

xgc.2d/3d.*.bp: phi (Diag3dPhi)

Native units: rad.

Toroidal angle of this plane

  • plot_label: $\varphi$

  • axes: toroidal_plane

  • coordinate_context: Toroidal angle coordinate in rad; same-product phi_left/phi_right identify bracketing planes. Geometry/periodicity: xgc.mesh.bp wedge_angle and delta_phi.

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:513.

xgc.2d/3d.*.bp: phi_left (Diag3dPhiLeft)

Native units: rad.

Toroidal angle of left plane

  • plot_label: $\varphi_{\mathrm{left}}$

  • axes: toroidal_plane

  • coordinate_context: Toroidal angle coordinate in rad; same-product phi_left/phi_right identify bracketing planes. Geometry/periodicity: xgc.mesh.bp wedge_angle and delta_phi.

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:518.

xgc.2d/3d.*.bp: phi_right (Diag3dPhiRight)

Native units: rad.

Toroidal angle of right plane

  • plot_label: $\varphi_{\mathrm{right}}$

  • axes: toroidal_plane

  • coordinate_context: Toroidal angle coordinate in rad; same-product phi_left/phi_right identify bracketing planes. Geometry/periodicity: xgc.mesh.bp wedge_angle and delta_phi.

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:523.

xgc.2d/3d.*.bp: poisson_axisym_converged_reason (Diag3dPoissonAxisymConvergedReason)

Native units: PETSc KSPConvergedReason integer.

PETSc KSP termination-reason code; index 0 is adiabatic and index 1 is full.

  • plot_label: Poisson termination code

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:486.

xgc.2d/3d.*.bp: poisson_axisym_iteration_count (Diag3dPoissonAxisymIterationCount)

Native units: Non-negative integer.

Number of completed outer KSP iterations; index 0 is adiabatic and index 1 is full.

  • plot_label: Axisymmetric Poisson iteration count

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:480.

xgc.2d/3d.*.bp: poisson_axisym_residual_count (Diag3dPoissonAxisymResidualCount)

Native units: Non-negative integer.

Number of valid entries in each axisymmetric residual-history row; index 0 is adiabatic and index 1 is full.

  • plot_label: Poisson residual-history length

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:473.

xgc.2d/3d.*.bp: poisson_axisym_residual_history (Diag3dPoissonAxisymResidualHistory)

Native units: PETSc residual norm.

Latest outer-KSP residual history; row 0 is the adiabatic solve and row 1 is the full solve. Use poisson_axisym_residual_count to identify valid entries.

  • plot_label: Axisymmetric Poisson residual history

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:466.

xgc.2d/3d.*.bp: pot0 (Diag3dPot0)

Native units: V.

Axisymmetric, flux-surface averaged (n=0, m=0) electric potential

  • plot_label: $\langle \phi \rangle$

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz, nd_connect_list for axisymmetric geometry; rplane_rz, rplane_nd_connect_list for stellarator right-plane geometry. Connectivity is zero-based. Node-centered field with continuous linear barycentric basis functions on each triangle.

  • coordinate_context: Axisymmetric potential on the right-plane nodal geometry. For stellarators use xgc.mesh.bp:rplane_rz,rplane_nd_connect_list. No toroidal_plane axis is stored.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/phi_potential(:)/values

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:449.

xgc.2d/3d.*.bp: pot0m (Diag3dPot0m)

Native units: V.

Full axisymmetric (n=0) electrostatic potential, including its flux-surface average and poloidally varying parts.

  • plot_label: $\phi_{n=0}$

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz, nd_connect_list for axisymmetric geometry; rplane_rz, rplane_nd_connect_list for stellarator right-plane geometry. Connectivity is zero-based. Node-centered field with continuous linear barycentric basis functions on each triangle.

  • coordinate_context: Axisymmetric potential on the right-plane nodal geometry. For stellarators use xgc.mesh.bp:rplane_rz,rplane_nd_connect_list. No toroidal_plane axis is stored.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/phi_potential(:)/values

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:454.

xgc.2d/3d.*.bp: pot0m_predictor (PotentialPredictorPot0mPredictor)

Native units: V.

Axisymmetric electric potential obtained from the adiabatic predictor before electron charge deposition, available with or without the ne00 correction

  • plot_label: $\phi_{0m}$ (predictor)

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz, nd_connect_list for axisymmetric geometry; rplane_rz, rplane_nd_connect_list for stellarator right-plane geometry. Connectivity is zero-based. Node-centered field with continuous linear barycentric basis functions on each triangle.

  • coordinate_context: Axisymmetric predictor state on the right-plane nodal geometry used by the Poisson solve; a flux-surface average/median remains represented at mesh nodes.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/phi_potential(:)/values

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/solvers/potential_predictor.cpp:647.

xgc.2d/3d.*.bp: sheath_nphi (Diag3dSheathNphi)

Native units: Positive integer.

Number of unique planes of sheath potential

  • plot_label: $N_{\varphi,sh}$

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:414.

xgc.2d/3d.*.bp: shpot (Diag3dShpot)

Native units: V.

Sheath potential

  • plot_label: $\phi_{\mathrm{sh}}$

  • axes: toroidal_plane,wall_node

  • mesh_context: xgc.mesh.bp: grid_wall_nodes gives zero-based indices into rz (R,Z in m); this list is explicitly unordered. Wall-node data follow grid_wall_nodes order. Do not connect successive entries into a wall polygon without reconstructing boundary adjacency from nd_connect_list; grid_wall_nodes is not triangle connectivity.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: wall_node

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:419.

xgc.2d/3d.*.bp: sp_name + “marker_den” (Diag3dSpNameMarkerDen)

Native units: m^-3.

Toroidally averaged marker number density at mesh nodes, obtained from deposited marker counts divided by node volume.

  • plot_label: $n_{\mathrm{marker}}$ ({name})

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry xgc.mesh.bp:rz,nd_connect_list; toroidal angle=(phi_left+phi_right)/2 on each plane, with periodic unwrapping at the wedge boundary. The historical phi array is the right-plane angle in XGC1; use phi_left/phi_right to locate midplane data.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Computed from marker deposits on midplane mesh nodes. XGC1 averages each stored statistic over toroidal planes using equal node IDs; the resulting toroidal average is repeated along the output toroidal_plane axis. marker_den is marker count divided by node volume; mean_weight and weight_variance are respectively the per-marker mean and variance of the deposition weight.

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:679.

xgc.2d/3d.*.bp: sp_name + “mean_weight” (Diag3dSpNameMeanWeight)

Native units: #real / #marker.

Mean particle weight (real particles / marker particles) at each mesh node

  • plot_label: $\langle w\rangle$ ({name})

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry xgc.mesh.bp:rz,nd_connect_list; toroidal angle=(phi_left+phi_right)/2 on each plane, with periodic unwrapping at the wedge boundary. The historical phi array is the right-plane angle in XGC1; use phi_left/phi_right to locate midplane data.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Computed from marker deposits on midplane mesh nodes. XGC1 averages each stored statistic over toroidal planes using equal node IDs; the resulting toroidal average is repeated along the output toroidal_plane axis. marker_den is marker count divided by node volume; mean_weight and weight_variance are respectively the per-marker mean and variance of the deposition weight.

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:688.

xgc.2d/3d.*.bp: sp_name + “temp” (Diag3dSpNameTemp)

Native units: J.

Latest optionally collected main-ion particle-weighted total kinetic energy per particle at each mesh node, including parallel-flow energy. No 2/3 temperature conversion or eV conversion is applied; non-main-ion channels remain zero.

  • plot_label: $\langle K\rangle$ ({name})

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry xgc.mesh.bp:rz,nd_connect_list; toroidal angle=(phi_left+phi_right)/2 on each plane, with periodic unwrapping at the wedge boundary. The historical phi array is the right-plane angle in XGC1; use phi_left/phi_right to locate midplane data.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: These optional additional moments are collected on the 1D push-diagnostic trigger (diag_1d_period, predictor stage 1) and written on the 3D field diagnostic trigger. The file time/gstep identify the field write; an additional-moment array may still contain an earlier collected sample when the periods differ. Only the main-ion collector runs; other species arrays remain zero.

  • centering: node

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:748.

xgc.2d/3d.*.bp: sp_name + “vp” (Diag3dSpNameVp)

Native units: m/s.

Latest optionally collected main-ion particle-weighted parallel velocity at each mesh node. Non-main-ion channels are not computed and remain zero.

  • plot_label: $\langle v_\parallel\rangle$ ({name})

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry xgc.mesh.bp:rz,nd_connect_list; toroidal angle=(phi_left+phi_right)/2 on each plane, with periodic unwrapping at the wedge boundary. The historical phi array is the right-plane angle in XGC1; use phi_left/phi_right to locate midplane data.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: These optional additional moments are collected on the 1D push-diagnostic trigger (diag_1d_period, predictor stage 1) and written on the 3D field diagnostic trigger. The file time/gstep identify the field write; an additional-moment array may still contain an earlier collected sample when the periods differ. Only the main-ion collector runs; other species arrays remain zero.

  • centering: node

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:739.

xgc.2d/3d.*.bp: sp_name + “weight_variance” (Diag3dSpNameWeightVariance)

Native units: (#real / #marker)^2.

Particle weight variance at each mesh node

  • plot_label: $\mathrm{Var}(w)$ ({name})

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry xgc.mesh.bp:rz,nd_connect_list; toroidal angle=(phi_left+phi_right)/2 on each plane, with periodic unwrapping at the wedge boundary. The historical phi array is the right-plane angle in XGC1; use phi_left/phi_right to locate midplane data.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Computed from marker deposits on midplane mesh nodes. XGC1 averages each stored statistic over toroidal planes using equal node IDs; the resulting toroidal average is repeated along the output toroidal_plane axis. marker_den is marker count divided by node volume; mean_weight and weight_variance are respectively the per-marker mean and variance of the deposition weight.

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:694.

xgc.2d/3d.*.bp: sp_name + “wsum” (Diag3dSpNameWsum)

Native units: Number of real particles.

Latest optionally collected main-ion deposited particle-weight sum at each mesh node. It has not been divided by node volume; non-main-ion channels are not computed and remain zero.

  • plot_label: Deposited particle-weight sum ({name})

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry xgc.mesh.bp:rz,nd_connect_list; toroidal angle=(phi_left+phi_right)/2 on each plane, with periodic unwrapping at the wedge boundary. The historical phi array is the right-plane angle in XGC1; use phi_left/phi_right to locate midplane data.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: These optional additional moments are collected on the 1D push-diagnostic trigger (diag_1d_period, predictor stage 1) and written on the 3D field diagnostic trigger. The file time/gstep identify the field write; an additional-moment array may still contain an earlier collected sample when the periods differ. Only the main-ion collector runs; other species arrays remain zero.

  • centering: node

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:730.

xgc.2d/3d.*.bp: time (Diag3dTime)

Native units: s.

Simulation time of step in seconds

  • plot_label: $t$

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions. Clock convention: in the regular main loop gstep is incremented before diagnostic calls, but sml.time is incremented afterward. A checkpoint written later at the same gstep therefore records time larger by one physical timestep. Use the written time, global step and product/integration stage together; equal gstep alone does not establish simultaneous diagnostic and checkpoint states.

  • imas_candidates: plasma_profiles/time

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_high

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:400.

xgc.2d/3d.*.bp: tindex (Diag3dTindex)

Native units: Positive integer.

Index of this field diagnostic output time slice. Values: 1, 2, …, # outputs.

  • plot_label: Output slice index

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/diagnostics/diag_3d.cpp:395.

xgc.bfield.bp: bfield (DumpBfieldBfield)

Native units: T.

Magnetic field vector

  • plot_label: $\mathbf{B}$

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: bfield is [vector_component,mesh_node], with R,Z,phi component order; stellarator data add a leading toroidal_plane axis. The field is equilibrium/static, not a magnetic perturbation.

  • centering: node

Source: XGC_core/cpp/grid/dump_bfield.cpp:47.

xgc.bfield.bp: bfield (DumpBfieldBfield2)

Native units: T.

Magnetic field vector

  • plot_label: $\mathbf{B}$

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: bfield is [vector_component,mesh_node], with R,Z,phi component order; stellarator data add a leading toroidal_plane axis. The field is equilibrium/static, not a magnetic perturbation.

  • centering: node

Source: XGC_core/cpp/grid/dump_bfield.cpp:54.

xgc.bfield.bp: n_n (DumpBfieldNN)

Native units: Positive integer.

Number of grid vertices

  • plot_label: $N_n$

Source: XGC_core/cpp/grid/dump_bfield.cpp:30.

xgc.bfield.bp: psi (DumpBfieldPsi)

Native units: T m^2.

Equilibrium magnetic-flux label evaluated at mesh vertices; axisymmetric runs use poloidal flux divided by 2*pi.

  • plot_label: $\psi/(2\pi)$

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Scalar flux label on the mesh midplane geometry, not a magnetic-vector component. This is static equilibrium data; the flux convention follows the supplied equilibrium.

  • centering: node

Source: XGC_core/cpp/grid/dump_bfield.cpp:37.

xgc.bfieldm.bp: N (MagneticFieldN)

Native units: Positive integer.

Number of midplane points (1000).

  • plot_label: Midplane point count

Source: XGC_core/cpp/magnetic_field/magnetic_field.cpp:65.

xgc.bfieldm.bp: abs_bphi (MagneticFieldAbsBphi)

Native units: T.

Absolute value of toroidal magnetic field.

  • plot_label: $|B_\varphi|$

  • axes: major_radius

  • mesh_context: Same xgc.bfieldm.bp: rmajor (m) is a one-dimensional major-radius grid at constant z (m), the magnetic-axis height. Interpolate in rmajor; this is not the triangle mesh.

Source: XGC_core/cpp/magnetic_field/magnetic_field.cpp:96.

xgc.bfieldm.bp: dpdr (MagneticFieldDpdr)

Native units: T m.

Nonnegative R*sqrt(B_R^2+B_Z^2) on Z=Z_axis, phi=0; for an axisymmetric equilibrium this is the magnitude of grad(poloidal flux/(2*pi)), not its signed R derivative.

  • plot_label: $R\sqrt{B_R^2+B_Z^2}$

  • axes: major_radius

  • mesh_context: Same xgc.bfieldm.bp: rmajor (m) is a one-dimensional major-radius grid at constant z (m), the magnetic-axis height. Interpolate in rmajor; this is not the triangle mesh.

Source: XGC_core/cpp/magnetic_field/magnetic_field.cpp:88.

xgc.bfieldm.bp: psi_n (MagneticFieldPsiN)

Native units: dimensionless.

Equilibrium flux label divided by eq_psi_norm, sampled on Z=Z_axis, phi=0; the denominator is the X-point flux for axisymmetric equilibria and the configured reference-point flux for stellarators.

  • plot_label: $\psi_N$

  • axes: major_radius

  • mesh_context: Same xgc.bfieldm.bp: rmajor (m) is a one-dimensional major-radius grid at constant z (m), the magnetic-axis height. Interpolate in rmajor; this is not the triangle mesh.

  • normalization: psi_n=psi/eq_psi_norm. Use xgc.equil.bp:eq_psi_norm; the stellarator reference flux need not be an X-point value.

Source: XGC_core/cpp/magnetic_field/magnetic_field.cpp:80.

xgc.bfieldm.bp: rmajor (MagneticFieldRmajor)

Native units: m.

Major radius.

  • plot_label: $R(Z=Z_{\mathrm{axis}},\varphi=0)$

Source: XGC_core/cpp/magnetic_field/magnetic_field.cpp:75.

xgc.bfieldm.bp: z (MagneticFieldZ)

Native units: m.

Z(height) at the magnetic axis.

  • plot_label: $Z_{\mathrm{axis}}$

Source: XGC_core/cpp/magnetic_field/magnetic_field.cpp:70.

xgc.col_conv_status.bp: convergence_status (CollisionsConvergenceStatus)

Native units: integer flag.

Collision solver convergence status at each toroidal plane and mesh node. Status codes are -5: NaN or Inf found, -4: maximum Picard iterations reached, -3: too many negative distribution-function values after the solve, -2: positivity correction failed, -1: conservation check failed before the first Picard solve, 0: not attempted, 1: converged successfully, 2: still in progress/internal transient state.

  • plot_label: Collision convergence code

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Categorical convergence flags/subcycle counts. Plot nearest-node or categorical colors; linear interpolation of status codes has no solver meaning.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

Source: XGC_core/cpp/sources/collisions.cpp:134.

xgc.col_conv_status.bp: gstep (CollisionsGstep)

Native units: Non-negative integer.

XGC global simulation time step for this collision convergence output slice.

  • plot_label: Global time step

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/sources/collisions.cpp:119.

xgc.col_conv_status.bp: n_subcycles (CollisionsNSubcycles)

Native units: Non-negative integer.

Configured collision subcycle count at each toroidal plane and mesh node after convergence-based adjustment for the next collision application. On a failed node this may exceed the count used in the just-completed solve.

  • plot_label: Next collision subcycle count

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Categorical convergence flags/subcycle counts. Plot nearest-node or categorical colors; linear interpolation of status codes has no solver meaning.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

Source: XGC_core/cpp/sources/collisions.cpp:148.

xgc.col_conv_status.bp: time (CollisionsTime)

Native units: s.

Simulation time of this collision convergence output slice.

  • plot_label: $t$

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions. Clock convention: in the regular main loop gstep is incremented before diagnostic calls, but sml.time is incremented afterward. A checkpoint written later at the same gstep therefore records time larger by one physical timestep. Use the written time, global step and product/integration stage together; equal gstep alone does not establish simultaneous diagnostic and checkpoint states.

  • imas_candidates: plasma_profiles/time

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_high

Source: XGC_core/cpp/sources/collisions.cpp:129.

xgc.col_conv_status.bp: tindex (CollisionsTindex)

Native units: Positive integer.

Index of this collision convergence output slice. Values: 1, 2, …, # outputs.

  • plot_label: Output slice index

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/sources/collisions.cpp:124.

xgc.current_drive.bp: jpar_bg (CurrentDriveJparBg)

Native units: A m^-2.

Equilibrium parallel current density \(j_{\\parallel}=\\hat{b}\\cdot(\\nabla\\times B)/\\mu_0\) on mesh nodes.

  • plot_label: $j_{\parallel,\mathrm{eq}}$

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/j_total(:)/parallel

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/sources/current_drive.cpp:93.

xgc.current_drive.bp: jpar_bg_fs_avg (CurrentDriveJparBgFsAvg)

Native units: A m^-2.

Flux-surface-averaged equilibrium parallel current density on mesh nodes.

  • plot_label: $\langle j_{\parallel,\mathrm{eq}}\rangle$

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/j_total(:)/parallel

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/sources/current_drive.cpp:99.

xgc.current_drive.bp: n_n (CurrentDriveNN)

Native units: Positive integer.

Number of mesh nodes per plane.

  • plot_label: $N_n$

Source: XGC_core/cpp/sources/current_drive.cpp:88.

xgc.diffusion_profiles.bp: density (DiagDiffusionProfilesDensity)

Native units: m^-3.

Flux-surface-averaged density profiles on [species, sample, flux surface].

  • plot_label: $\langle n\rangle$

  • axes: species,time_sample,flux_surface

  • mesh_context: Same xgc.diffusion_profiles.bp step: psi is the radial poloidal-flux/(2*pi) coordinate. time[] and steps[] label the middle time_sample axis. Spatial data are flux-surface averages; interpolate radially in psi.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Each ADIOS step is one output buffer, not one simulation sample. time[j] (s) and steps[j] (global integration step) index the time_sample axis of density/flow/temp[:,j,:]. Scalar gstep labels the buffer write and tindex=gstep/output_rate labels the buffer output. n_samples is the number of buffered samples. Join by time[j]/steps[j], not by outer ADIOS position. These are sampled flux-surface averages, not a time average over the buffer.

  • centering: flux_surface

  • imas_candidates: core_profiles/profiles_1d(:)/electrons/density; core_profiles/profiles_1d(:)/ion(:)/density

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_diffusion_profiles.cpp:258.

xgc.diffusion_profiles.bp: flow (DiagDiffusionProfilesFlow)

Native units: m s^-1.

Flux-surface-averaged parallel-flow profiles on [species, sample, flux surface].

  • plot_label: $\langle u_\parallel\rangle$

  • axes: species,time_sample,flux_surface

  • mesh_context: Same xgc.diffusion_profiles.bp step: psi is the radial poloidal-flux/(2*pi) coordinate. time[] and steps[] label the middle time_sample axis. Spatial data are flux-surface averages; interpolate radially in psi.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Each ADIOS step is one output buffer, not one simulation sample. time[j] (s) and steps[j] (global integration step) index the time_sample axis of density/flow/temp[:,j,:]. Scalar gstep labels the buffer write and tindex=gstep/output_rate labels the buffer output. n_samples is the number of buffered samples. Join by time[j]/steps[j], not by outer ADIOS position. These are sampled flux-surface averages, not a time average over the buffer.

  • centering: flux_surface

  • imas_candidates: core_profiles/profiles_1d(:)/ion(:)/velocity/parallel

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_diffusion_profiles.cpp:263.

xgc.diffusion_profiles.bp: gstep (DiagDiffusionProfilesGstep)

Native units: Non-negative integer.

XGC global simulation time step for this diffusion-profile diagnostic output.

  • plot_label: Global time step

  • time_context: Each ADIOS step is one output buffer, not one simulation sample. time[j] (s) and steps[j] (global integration step) index the time_sample axis of density/flow/temp[:,j,:]. Scalar gstep labels the buffer write and tindex=gstep/output_rate labels the buffer output. n_samples is the number of buffered samples. Join by time[j]/steps[j], not by outer ADIOS position. These are sampled flux-surface averages, not a time average over the buffer.

Source: XGC_core/cpp/diagnostics/diag_diffusion_profiles.cpp:237.

xgc.diffusion_profiles.bp: n_samples (DiagDiffusionProfilesNSamples)

Native units: Positive integer.

Number of buffered time samples in each diffusion-profile diagnostic output.

  • plot_label: $N_t$

Source: XGC_core/cpp/diagnostics/diag_diffusion_profiles.cpp:222.

xgc.diffusion_profiles.bp: n_species (DiagDiffusionProfilesNSpecies)

Native units: Positive integer.

Number of non-adiabatic species in the diffusion-profile diagnostic.

  • plot_label: $n_{sp}$

Source: XGC_core/cpp/diagnostics/diag_diffusion_profiles.cpp:217.

xgc.diffusion_profiles.bp: n_surf (DiagDiffusionProfilesNSurf)

Native units: Positive integer.

Number of flux-surface grid points in the diffusion-profile diagnostic.

  • plot_label: $N_{surf}$

Source: XGC_core/cpp/diagnostics/diag_diffusion_profiles.cpp:227.

xgc.diffusion_profiles.bp: psi (DiagDiffusionProfilesPsi)

Native units: T m^2.

Poloidal magnetic flux \(\\psi/(2\\pi)\).

  • plot_label: $\psi/(2\pi)$

Source: XGC_core/cpp/diagnostics/diag_diffusion_profiles.cpp:232.

xgc.diffusion_profiles.bp: steps (DiagDiffusionProfilesSteps)

Native units: Non-negative integer.

Simulation time steps corresponding to buffered diffusion-profile samples.

  • plot_label: Sample global time step

  • axes: time_sample

  • time_context: Each ADIOS step is one output buffer, not one simulation sample. time[j] (s) and steps[j] (global integration step) index the time_sample axis of density/flow/temp[:,j,:]. Scalar gstep labels the buffer write and tindex=gstep/output_rate labels the buffer output. n_samples is the number of buffered samples. Join by time[j]/steps[j], not by outer ADIOS position. These are sampled flux-surface averages, not a time average over the buffer.

Source: XGC_core/cpp/diagnostics/diag_diffusion_profiles.cpp:253.

xgc.diffusion_profiles.bp: temp (DiagDiffusionProfilesTemp)

Native units: eV.

Flux-surface-averaged temperature profiles on [species, sample, flux surface].

  • plot_label: $\langle T\rangle$

  • axes: species,time_sample,flux_surface

  • mesh_context: Same xgc.diffusion_profiles.bp step: psi is the radial poloidal-flux/(2*pi) coordinate. time[] and steps[] label the middle time_sample axis. Spatial data are flux-surface averages; interpolate radially in psi.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Each ADIOS step is one output buffer, not one simulation sample. time[j] (s) and steps[j] (global integration step) index the time_sample axis of density/flow/temp[:,j,:]. Scalar gstep labels the buffer write and tindex=gstep/output_rate labels the buffer output. n_samples is the number of buffered samples. Join by time[j]/steps[j], not by outer ADIOS position. These are sampled flux-surface averages, not a time average over the buffer.

  • centering: flux_surface

  • imas_candidates: core_profiles/profiles_1d(:)/electrons/temperature; core_profiles/profiles_1d(:)/ion(:)/temperature

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_diffusion_profiles.cpp:268.

xgc.diffusion_profiles.bp: time (DiagDiffusionProfilesTime)

Native units: s.

Physical time in seconds for each buffered sample, not one scalar time per ADIOS step.

  • plot_label: $t$

  • axes: time_sample

  • time_context: Each ADIOS step is one output buffer, not one simulation sample. time[j] (s) and steps[j] (global integration step) index the time_sample axis of density/flow/temp[:,j,:]. Scalar gstep labels the buffer write and tindex=gstep/output_rate labels the buffer output. n_samples is the number of buffered samples. Join by time[j]/steps[j], not by outer ADIOS position. These are sampled flux-surface averages, not a time average over the buffer. Clock convention: in the regular main loop gstep is incremented before diagnostic calls, but sml.time is incremented afterward. A checkpoint written later at the same gstep therefore records time larger by one physical timestep. Use the written time, global step and product/integration stage together; equal gstep alone does not establish simultaneous diagnostic and checkpoint states.

  • imas_candidates: core_profiles/time

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_diffusion_profiles.cpp:248.

xgc.diffusion_profiles.bp: tindex (DiagDiffusionProfilesTindex)

Native units: Non-negative integer.

Index of the buffered diffusion-profile output; individual sample steps are in steps and sample times in time.

  • plot_label: Output slice index

  • time_context: Each ADIOS step is one output buffer, not one simulation sample. time[j] (s) and steps[j] (global integration step) index the time_sample axis of density/flow/temp[:,j,:]. Scalar gstep labels the buffer write and tindex=gstep/output_rate labels the buffer output. n_samples is the number of buffered samples. Join by time[j]/steps[j], not by outer ADIOS position. These are sampled flux-surface averages, not a time average over the buffer.

Source: XGC_core/cpp/diagnostics/diag_diffusion_profiles.cpp:242.

xgc.diffusion_solver_moments.bp: dt (DiffusionDt)

Native units: s.

Physical time interval advanced by this diffusion solve.

  • plot_label: $\Delta t$

  • time_context: One optional before/after snapshot of a single diffusion solve, overwritten when this debug output is written again. dt is the physical solve interval in seconds; no absolute gstep/time is written. Do not interpret separate ADIOS positions as simulation times.

Source: XGC_core/cpp/sources/diffusion.cpp:117.

xgc.diffusion_solver_moments.bp: moments_after (DiffusionMomentsAfter)

Native units: [m^-3, m s^-1, eV, eV, m s^-1, eV, eV].

Packed transport moments immediately after diffusion_solve; dimensions are [moment_block, node].

  • plot_label: Transport moments after diffusion

  • axes: moment_block,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • time_context: One optional before/after snapshot of a single diffusion solve, overwritten when this debug output is written again. dt is the physical solve interval in seconds; no absolute gstep/time is written. Do not interpret separate ADIOS positions as simulation times.

  • centering: node

  • normalization: The moment_block order is n_e,u_parallel_i,T_parallel_i,T_perp_i,u_parallel_e,T_parallel_e,T_perp_e, with units m^-3,m s^-1,eV,eV,m s^-1,eV,eV. These are physical caller moments before/after the solve; internal PETSc field scaling is removed before this output.

Source: XGC_core/cpp/sources/diffusion.cpp:131.

xgc.diffusion_solver_moments.bp: moments_before (DiffusionMomentsBefore)

Native units: [m^-3, m s^-1, eV, eV, m s^-1, eV, eV].

Packed transport moments immediately before diffusion_solve; dimensions are [moment_block, node].

  • plot_label: Transport moments before diffusion

  • axes: moment_block,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • time_context: One optional before/after snapshot of a single diffusion solve, overwritten when this debug output is written again. dt is the physical solve interval in seconds; no absolute gstep/time is written. Do not interpret separate ADIOS positions as simulation times.

  • centering: node

  • normalization: The moment_block order is n_e,u_parallel_i,T_parallel_i,T_perp_i,u_parallel_e,T_parallel_e,T_perp_e, with units m^-3,m s^-1,eV,eV,m s^-1,eV,eV. These are physical caller moments before/after the solve; internal PETSc field scaling is removed before this output.

Source: XGC_core/cpp/sources/diffusion.cpp:124.

xgc.diffusion_solver_moments.bp: n_moments (DiffusionNMoments)

Native units: Positive integer.

Number of packed anomalous-transport moment fields.

  • plot_label: Packed transport-moment count

Source: XGC_core/cpp/sources/diffusion.cpp:112.

xgc.diffusion_solver_moments.bp: nnode (DiffusionNnode)

Native units: Positive integer.

Number of nodes in each 2D unstructured-mesh moment field.

  • plot_label: $N_n$

Source: XGC_core/cpp/sources/diffusion.cpp:107.

xgc.equil.bp: bp_sign (EquilBpSign)

Native units: dimensionless.

Direction of poloidal field (1 or -1)

  • plot_label: Poloidal magnetic-field sign

  • mesh_context: B_P=bp_sign*grad(psi/2pi) x grad(phi); psi=0 on axis and increasing outward is an XGC convention. The coordinates in XGC are right handed (R,phi,Z). The positive toroidal direction (given by grad(phi)) is counter-clockwise when looking from the top (down the Z-axis)

Source: XGC_core/cpp/magnetic_field/equil.cpp:143.

xgc.equil.bp: bt_sign (EquilBtSign)

Native units: dimensionless.

Direction of toroidal field (1 or -1)

  • plot_label: Toroidal magnetic-field sign

  • mesh_context: B_T=bt_sign*eq_I/R; eq_I>0 is an XGC convention. The coordinates in XGC are right handed (R,phi,Z). The positive toroidal direction is counter-clockwise when looking from the top (down the Z-axis). bt_sign is relative to this definition.

Source: XGC_core/cpp/magnetic_field/equil.cpp:136.

xgc.equil.bp: eq_B_R_rz (EquilEqBRRz)

Native units: T.

Radial magnetic field component on rectangular grid points

  • plot_label: $B_R$ (equilibrium)

  • axes: toroidal_plane,vertical_coordinate,major_radius

  • mesh_context: xgc.equil.bp: eq_mr, eq_mz and eq_min_r/eq_max_r, eq_min_z/eq_max_z define a uniformly spaced rectangular RZ grid with endpoints included. R[i]=eq_min_r+i*(eq_max_r-eq_min_r)/(eq_mr-1), Z[j]=eq_min_z+j*(eq_max_z-eq_min_z)/(eq_mz-1). Stored order is [Z,R] (R contiguous), with a leading toroidal_plane axis for stellarators. Use interpolation on this rectangular grid, not xgc.mesh.bp triangulation.

  • centering: grid_point

  • imas_candidates: equilibrium/time_slice(:)/profiles_2d(:)/b_field_r

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/magnetic_field/equil.cpp:209.

xgc.equil.bp: eq_B_Z_rz (EquilEqBZRz)

Native units: T.

Z magnetic field component on rectangular grid points

  • plot_label: $B_Z$ (equilibrium)

  • axes: toroidal_plane,vertical_coordinate,major_radius

  • mesh_context: xgc.equil.bp: eq_mr, eq_mz and eq_min_r/eq_max_r, eq_min_z/eq_max_z define a uniformly spaced rectangular RZ grid with endpoints included. R[i]=eq_min_r+i*(eq_max_r-eq_min_r)/(eq_mr-1), Z[j]=eq_min_z+j*(eq_max_z-eq_min_z)/(eq_mz-1). Stored order is [Z,R] (R contiguous), with a leading toroidal_plane axis for stellarators. Use interpolation on this rectangular grid, not xgc.mesh.bp triangulation.

  • centering: grid_point

  • imas_candidates: equilibrium/time_slice(:)/profiles_2d(:)/b_field_z

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/magnetic_field/equil.cpp:215.

xgc.equil.bp: eq_B_phi_rz (EquilEqBPhiRz)

Native units: T.

Toroidal magnetic field component on rectangular grid points

  • plot_label: $B_\varphi$ (equilibrium)

  • axes: toroidal_plane,vertical_coordinate,major_radius

  • mesh_context: xgc.equil.bp: eq_mr, eq_mz and eq_min_r/eq_max_r, eq_min_z/eq_max_z define a uniformly spaced rectangular RZ grid with endpoints included. R[i]=eq_min_r+i*(eq_max_r-eq_min_r)/(eq_mr-1), Z[j]=eq_min_z+j*(eq_max_z-eq_min_z)/(eq_mz-1). Stored order is [Z,R] (R contiguous), with a leading toroidal_plane axis for stellarators. Use interpolation on this rectangular grid, not xgc.mesh.bp triangulation.

  • centering: grid_point

  • imas_candidates: equilibrium/time_slice(:)/profiles_2d(:)/b_field_phi

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/magnetic_field/equil.cpp:221.

xgc.equil.bp: eq_I (EquilEqI)

Native units: T m.

Positive poloidal-current function magnitude as a function of the equilibrium flux label; the signed toroidal-field function is bt_sign*eq_I.

  • plot_label: $|R B_\varphi|$

  • axes: flux_surface

  • mesh_context: Same xgc.equil.bp: eq_psi_grid is the radial poloidal-flux/(2*pi) coordinate; interpolate the poloidal-current function in this coordinate.

  • normalization: eq_I>0 in XGC by convention. The signed toroidal-field function R*B_phi, including the IDS profiles_1d/f candidate, is bt_sign*eq_I; retain the equilibrium COCOS/sign convention.

  • imas_candidates: equilibrium/time_slice(:)/profiles_1d/f

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/magnetic_field/equil.cpp:274.

xgc.equil.bp: eq_PEST_angle_rz (EquilEqPESTAngleRz)

Native units: rad.

PEST coordinate on rectangular grid points

  • plot_label: $\theta^\ast_{\mathrm{PEST}}$

  • axes: toroidal_plane,vertical_coordinate,major_radius

  • mesh_context: xgc.equil.bp: eq_mr, eq_mz and eq_min_r/eq_max_r, eq_min_z/eq_max_z define a uniformly spaced rectangular RZ grid with endpoints included. R[i]=eq_min_r+i*(eq_max_r-eq_min_r)/(eq_mr-1), Z[j]=eq_min_z+j*(eq_max_z-eq_min_z)/(eq_mz-1). Stored order is [Z,R] (R contiguous), with a leading toroidal_plane axis for stellarators. Use interpolation on this rectangular grid, not xgc.mesh.bp triangulation.

  • centering: grid_point

Source: XGC_core/cpp/magnetic_field/equil.cpp:233.

xgc.equil.bp: eq_VMEC_angle_rz (EquilEqVMECAngleRz)

Native units: rad.

VMEC coordinate on rectangular grid points

  • plot_label: $\theta^\ast_{\mathrm{VMEC}}$

  • axes: toroidal_plane,vertical_coordinate,major_radius

  • mesh_context: xgc.equil.bp: eq_mr, eq_mz and eq_min_r/eq_max_r, eq_min_z/eq_max_z define a uniformly spaced rectangular RZ grid with endpoints included. R[i]=eq_min_r+i*(eq_max_r-eq_min_r)/(eq_mr-1), Z[j]=eq_min_z+j*(eq_max_z-eq_min_z)/(eq_mz-1). Stored order is [Z,R] (R contiguous), with a leading toroidal_plane axis for stellarators. Use interpolation on this rectangular grid, not xgc.mesh.bp triangulation.

  • centering: grid_point

Source: XGC_core/cpp/magnetic_field/equil.cpp:227.

xgc.equil.bp: eq_axis_b (EquilEqAxisB)

Native units: T.

Magnetic field at the magnetic axis

  • plot_label: $|\mathbf{B}|_{\mathrm{axis}}$

Source: XGC_core/cpp/magnetic_field/equil.cpp:170.

xgc.equil.bp: eq_axis_b (EquilEqAxisB2)

Native units: T.

Magnetic field at the magnetic axis

  • plot_label: $|\mathbf{B}|_{\mathrm{axis}}$

Source: XGC_core/cpp/magnetic_field/equil.cpp:244.

xgc.equil.bp: eq_axis_r (EquilEqAxisR)

Native units: m.

Major radius at the magnetic axis

  • plot_label: $R_{\mathrm{axis}}$

  • imas_candidates: equilibrium/time_slice(:)/global_quantities/magnetic_axis/r

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/magnetic_field/equil.cpp:175.

xgc.equil.bp: eq_axis_r (EquilEqAxisR2)

Native units: m.

Major radius at the magnetic axis

  • plot_label: $R_{\mathrm{axis}}$

  • imas_candidates: equilibrium/time_slice(:)/global_quantities/magnetic_axis/r

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/magnetic_field/equil.cpp:249.

xgc.equil.bp: eq_axis_z (EquilEqAxisZ)

Native units: m.

Z (height) at the magnetic axis

  • plot_label: $Z_{\mathrm{axis}}$

  • imas_candidates: equilibrium/time_slice(:)/global_quantities/magnetic_axis/z

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/magnetic_field/equil.cpp:180.

xgc.equil.bp: eq_axis_z (EquilEqAxisZ2)

Native units: m.

Z (height) at the magnetic axis

  • plot_label: $Z_{\mathrm{axis}}$

  • imas_candidates: equilibrium/time_slice(:)/global_quantities/magnetic_axis/z

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/magnetic_field/equil.cpp:254.

xgc.equil.bp: eq_max_r (EquilEqMaxR)

Native units: m.

Maximum major radius of interpolation boundary

  • plot_label: $\mathrm{max}(R)$ (equilibrium)

Source: XGC_core/cpp/magnetic_field/equil.cpp:104.

xgc.equil.bp: eq_max_z (EquilEqMaxZ)

Native units: m.

Maximum height of interpolation boundary

  • plot_label: $\mathrm{max}(Z)$ (equilibrium)

Source: XGC_core/cpp/magnetic_field/equil.cpp:114.

xgc.equil.bp: eq_min_r (EquilEqMinR)

Native units: m.

Minimum major radius of interpolation boundary

  • plot_label: $\mathrm{min}(R)$ (equilibrium)

Source: XGC_core/cpp/magnetic_field/equil.cpp:99.

xgc.equil.bp: eq_min_z (EquilEqMinZ)

Native units: m.

Minimum height of interpolation boundary

  • plot_label: $\mathrm{min}(Z)$ (equilibrium)

Source: XGC_core/cpp/magnetic_field/equil.cpp:109.

xgc.equil.bp: eq_mp (EquilEqMp)

Native units: Positive integer.

Number of planes in input grid (stellarator only)

  • plot_label: Equilibrium plane count

Source: XGC_core/cpp/magnetic_field/equil.cpp:165.

xgc.equil.bp: eq_mpsi (EquilEqMpsi)

Native units: Positive integer.

Number of psi (poloidal flux) grid points

  • plot_label: Equilibrium $\psi$-grid size

Source: XGC_core/cpp/magnetic_field/equil.cpp:94.

xgc.equil.bp: eq_mr (EquilEqMr)

Native units: Positive integer.

Number of horizontal R grid points

  • plot_label: Equilibrium radial grid size

Source: XGC_core/cpp/magnetic_field/equil.cpp:82.

xgc.equil.bp: eq_mz (EquilEqMz)

Native units: Positive integer.

Number of vertical Z grid points

  • plot_label: Equilibrium vertical grid size

Source: XGC_core/cpp/magnetic_field/equil.cpp:87.

xgc.equil.bp: eq_psi_grid (EquilEqPsiGrid)

Native units: T m^2 (axisymmetric); dimensionless (stellarator).

Equilibrium flux-grid coordinate: dimensional poloidal flux divided by 2*pi for axisymmetric equilibria, or the supplied stellarator flux label divided by eq_psi_norm.

  • plot_label: Equilibrium flux-grid coordinate

  • normalization: Axisymmetric values are dimensional poloidal flux/(2*pi). Stellarator values are the input equilibrium flux label divided by eq_psi_norm; multiply by eq_psi_norm to restore that input label. A poloidal-flux mapping to IDS Wb requires 2*pi and the appropriate COCOS/sign conversion, after verifying the supplied stellarator flux definition.

  • imas_candidates: equilibrium/time_slice(:)/profiles_1d/psi

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/magnetic_field/equil.cpp:121.

xgc.equil.bp: eq_psi_norm (EquilEqPsiNorm)

Native units: T m^2.

Poloidal flux used for normalization

  • plot_label: $\psi_{\mathrm{norm}}$

Source: XGC_core/cpp/magnetic_field/equil.cpp:129.

xgc.equil.bp: eq_psi_rz (EquilEqPsiRz)

Native units: dimensionless.

Normalized equilibrium flux label on the stellarator rectangular RZ grid, equal to the supplied flux label divided by eq_psi_norm.

  • plot_label: $\psi/\psi_{\mathrm{norm}}$ (equilibrium)

  • axes: toroidal_plane,vertical_coordinate,major_radius

  • mesh_context: xgc.equil.bp: eq_mr, eq_mz and eq_min_r/eq_max_r, eq_min_z/eq_max_z define a uniformly spaced rectangular RZ grid with endpoints included. R[i]=eq_min_r+i*(eq_max_r-eq_min_r)/(eq_mr-1), Z[j]=eq_min_z+j*(eq_max_z-eq_min_z)/(eq_mz-1). Stored order is [Z,R] (R contiguous), with a leading toroidal_plane axis for stellarators. Use interpolation on this rectangular grid, not xgc.mesh.bp triangulation.

  • centering: grid_point

  • normalization: Written values are the supplied stellarator equilibrium flux label divided by eq_psi_norm. Multiply by eq_psi_norm to recover the dimensional input label before any IDS conversion; verify whether that input is poloidal flux and establish the COCOS/sign convention.

  • imas_candidates: equilibrium/time_slice(:)/profiles_2d(:)/psi

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/magnetic_field/equil.cpp:200.

xgc.equil.bp: eq_psi_rz (EquilEqPsiRz2)

Native units: T m^2.

Poloidal flux on the rectangular grid points

  • plot_label: $\psi/(2\pi)$ (equilibrium)

  • axes: vertical_coordinate,major_radius

  • mesh_context: xgc.equil.bp: eq_mr, eq_mz and eq_min_r/eq_max_r, eq_min_z/eq_max_z define a uniformly spaced rectangular RZ grid with endpoints included. R[i]=eq_min_r+i*(eq_max_r-eq_min_r)/(eq_mr-1), Z[j]=eq_min_z+j*(eq_max_z-eq_min_z)/(eq_mz-1). Stored order is [Z,R] (R contiguous), with a leading toroidal_plane axis for stellarators. Use interpolation on this rectangular grid, not xgc.mesh.bp triangulation.

  • centering: grid_point

  • normalization: Axisymmetric output uses dimensional poloidal flux/(2*pi), in T m^2. Stellarator output divides eq_psi_grid and eq_psi_rz by eq_psi_norm before writing, despite their historical T m^2 unit string. Restore dimensional flux with eq_psi_norm first; converting dimensional XGC flux to IDS Wb then requires 2*pi and a COCOS/sign check. Psi=0 on axis, increasing outwards. B-field orientation in tokamaks defined by bt_sign and bp_sign.

  • imas_candidates: equilibrium/time_slice(:)/profiles_2d(:)/psi

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/magnetic_field/equil.cpp:282.

xgc.equil.bp: eq_x_psi (EquilEqXPsi)

Native units: T m^2.

Poloidal flux at the first (lower) X-point

  • plot_label: $\psi_{X1}/(2\pi)$

Source: XGC_core/cpp/magnetic_field/equil.cpp:185.

xgc.equil.bp: eq_x_psi (EquilEqXPsi2)

Native units: T m^2.

Poloidal flux at the first (lower) X-point

  • plot_label: $\psi_{X1}/(2\pi)$

Source: XGC_core/cpp/magnetic_field/equil.cpp:259.

xgc.equil.bp: eq_x_r (EquilEqXR)

Native units: m.

Major radius at the first (lower) X-point

  • plot_label: $R_{X1}$

Source: XGC_core/cpp/magnetic_field/equil.cpp:190.

xgc.equil.bp: eq_x_r (EquilEqXR2)

Native units: m.

Major radius at the first (lower) X-point

  • plot_label: $R_{X1}$

Source: XGC_core/cpp/magnetic_field/equil.cpp:264.

xgc.equil.bp: eq_x_z (EquilEqXZ)

Native units: m.

Height at the first (lower) X-point

  • plot_label: $Z_{X1}$

Source: XGC_core/cpp/magnetic_field/equil.cpp:195.

xgc.equil.bp: eq_x_z (EquilEqXZ2)

Native units: m.

Height at the first (lower) X-point

  • plot_label: $Z_{X1}$

Source: XGC_core/cpp/magnetic_field/equil.cpp:269.

xgc.equil.bp: nphi (EquilNphi)

Native units: Positive integer.

Number of toroidal mesh planes \(N_{\\mathrm{planes}}\)

  • plot_label: $N_\varphi$

Source: XGC_core/cpp/magnetic_field/equil.cpp:155.

xgc.equil.bp: periodicity (EquilPeriodicity)

Native units: Positive integer.

Toroidal periodicity from VMEC or torus wedge number

  • plot_label: Toroidal periodicity

Source: XGC_core/cpp/magnetic_field/equil.cpp:160.

xgc.escaped_ptls.*.bp: gstep (DiagEscapedPtlsGstep)

Native units: Non-negative integer.

XGC global simulation time step for this escaped-particle diagnostic output slice.

  • plot_label: Global time step

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/diagnostics/diag_escaped_ptls.cpp:247.

xgc.escaped_ptls.*.bp: prefix + “dw” (DiagEscapedPtlsPrefixDw)

Native units: dimensionless.

Decoded escaped-particle weight-change payload from the packed particle flag.

  • plot_label: Escaped particle weight ({name})

  • axes: particle

  • mesh_context: Particle sample, not a mesh field. The same species’ phase array contains row-wise cylindrical R (m), Z (m), phi (rad) in columns 0,1,2; use gid to identify rows/particles across output slices. MPI repartitioning can change row order. Visualize as points or perform a stated statistical/binning procedure, not triangle interpolation. Other phase columns and units are specified in description.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Escape events buffered until scalar gstep/time (s) at output. The same-species step[particle] gives each particle’s escape/marking step, which differs from buffer-flush gstep. Event physical times are not stored; do not assign every particle the flush time. With constant physical dt only, differences of event steps times dt give elapsed time.

  • centering: particle

Source: XGC_core/cpp/diagnostics/diag_escaped_ptls.cpp:192.

xgc.escaped_ptls.*.bp: prefix + “flag” (DiagEscapedPtlsPrefixFlag)

Native units: integer flag.

Decoded escaped-particle low-bit status flags for this species.

  • plot_label: Escape status flag ({name})

  • axes: particle

  • mesh_context: Particle sample, not a mesh field. The same species’ phase array contains row-wise cylindrical R (m), Z (m), phi (rad) in columns 0,1,2; use gid to identify rows/particles across output slices. MPI repartitioning can change row order. Visualize as points or perform a stated statistical/binning procedure, not triangle interpolation. Other phase columns and units are specified in description.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Escape events buffered until scalar gstep/time (s) at output. The same-species step[particle] gives each particle’s escape/marking step, which differs from buffer-flush gstep. Event physical times are not stored; do not assign every particle the flush time. With constant physical dt only, differences of event steps times dt give elapsed time.

  • centering: particle

Source: XGC_core/cpp/diagnostics/diag_escaped_ptls.cpp:178.

xgc.escaped_ptls.*.bp: prefix + “gid” (DiagEscapedPtlsPrefixGid)

Native units: Integer index.

Globally unique marker-particle IDs for escaped particles of this species.

  • plot_label: Escaped marker ID ({name})

  • axes: particle

  • mesh_context: Particle sample, not a mesh field. The same species’ phase array contains row-wise cylindrical R (m), Z (m), phi (rad) in columns 0,1,2; use gid to identify rows/particles across output slices. MPI repartitioning can change row order. Visualize as points or perform a stated statistical/binning procedure, not triangle interpolation. Other phase columns and units are specified in description.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Escape events buffered until scalar gstep/time (s) at output. The same-species step[particle] gives each particle’s escape/marking step, which differs from buffer-flush gstep. Event physical times are not stored; do not assign every particle the flush time. With constant physical dt only, differences of event steps times dt give elapsed time.

  • centering: particle

Source: XGC_core/cpp/diagnostics/diag_escaped_ptls.cpp:171.

xgc.escaped_ptls.*.bp: prefix + “maxgid” (DiagEscapedPtlsPrefixMaxgid)

Native units: Non-negative integer.

Largest globally unique marker-particle ID already assigned for this species when escaped particles are written.

  • plot_label: Largest assigned marker ID ({name})

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/diagnostics/diag_escaped_ptls.cpp:155.

xgc.escaped_ptls.*.bp: prefix + “nphase” (DiagEscapedPtlsPrefixNphase)

Native units: Positive integer.

Number of columns in the escaped-particle phase array.

  • plot_label: Escape phase-component count ({name})

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Per-particle escape state buffered until the output at scalar gstep/time; use the same-species step array for each particle’s event step. Event physical time is not explicitly stored.

Source: XGC_core/cpp/diagnostics/diag_escaped_ptls.cpp:166.

xgc.escaped_ptls.*.bp: prefix + “num_total” (DiagEscapedPtlsPrefixNumTotal)

Native units: Non-negative integer.

Total number of escaped marker particles written for this species in this diagnostic step.

  • plot_label: Escaped marker count ({name})

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/diagnostics/diag_escaped_ptls.cpp:161.

xgc.escaped_ptls.*.bp: prefix + “phase” (DiagEscapedPtlsPrefixPhase)

Native units: [m, m, rad, m, dimensionless, dimensionless, J/T, dimensionless, m^-3 eV^(-3/2), optional: m^-3 eV^(-3/2), T m^2, T].

Escaped-particle phase array with columns [R, Z, phi, rho, w1, w2, mu, w0, g, optional g2, psi, |B|], where \(R\), \(Z\), and \(\\phi\) are the cylindrical coordinates of the particle, \(\\rho\) is the parallel gyroradius (proxy for \(v_{\\parallel}\)), \(w_1\) is the delta-f weight, \(w_2\) is a weight used in the weight evolution equation, \(\\mu\) is the magnetic moment, \(g\) is the phase-space marker density, \(\\psi\) is the poloidal magnetic flux, and \(\|B\|\) is the magnetic-field magnitude. With PTL_G2, g2 is appended after g and before psi and |B|.

  • plot_label: Escaped marker phase space coord. ({name})

  • axes: particle,phase_component

  • mesh_context: Particle sample, not a mesh field. The same species’ phase array contains row-wise cylindrical R (m), Z (m), phi (rad) in columns 0,1,2; use gid to identify rows/particles across output slices. MPI repartitioning can change row order. Visualize as points or perform a stated statistical/binning procedure, not triangle interpolation. Other phase columns and units are specified in description.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Rows are selected by the buffered ps_is_to_write flag and written at scalar gstep/time (s). The phase/constants and psi/|B| are sampled from the current particle at buffer flush, whereas step[particle] is the decoded stored escape-step payload and can predate the sample. The array is not a saved phase-space snapshot at each escape event; physical event times are not written.

  • centering: particle

  • normalization: The g column is marker density in XGC velocity-space normalization, with units m^-3 eV^(-3/2); f=w0*g. For a Maxwellian this normalization is proportional to n*T_eV^(-3/2)*exp(-E_eV/T_eV), with the SI velocity normalization factor (mass/(2*pi*EV_2_J))^(3/2) removed. This is distinct from the spatial triangle marker_den in xgc.loading.*.bp.

Source: XGC_core/cpp/diagnostics/diag_escaped_ptls.cpp:197.

xgc.escaped_ptls.*.bp: prefix + “step” (DiagEscapedPtlsPrefixStep)

Native units: Non-negative integer.

Stored escape-step payload decoded from the packed particle flag. This is set on a separatrix escape; a divertor-only output mark does not necessarily update it, so it is not a guaranteed timestamp of the latest output-marking event.

  • plot_label: Escape event step ({name})

  • axes: particle

  • mesh_context: Particle sample, not a mesh field. The same species’ phase array contains row-wise cylindrical R (m), Z (m), phi (rad) in columns 0,1,2; use gid to identify rows/particles across output slices. MPI repartitioning can change row order. Visualize as points or perform a stated statistical/binning procedure, not triangle interpolation. Other phase columns and units are specified in description.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Escape events buffered until scalar gstep/time (s) at output. The same-species step[particle] gives each particle’s escape/marking step, which differs from buffer-flush gstep. Event physical times are not stored; do not assign every particle the flush time. With constant physical dt only, differences of event steps times dt give elapsed time.

  • centering: particle

Source: XGC_core/cpp/diagnostics/diag_escaped_ptls.cpp:183.

xgc.escaped_ptls.*.bp: time (DiagEscapedPtlsTime)

Native units: s.

Simulation time of step in seconds.

  • plot_label: $t$

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions. Clock convention: in the regular main loop gstep is incremented before diagnostic calls, but sml.time is incremented afterward. A checkpoint written later at the same gstep therefore records time larger by one physical timestep. Use the written time, global step and product/integration stage together; equal gstep alone does not establish simultaneous diagnostic and checkpoint states.

  • imas_candidates: plasma_profiles/time

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_high

Source: XGC_core/cpp/diagnostics/diag_escaped_ptls.cpp:260.

xgc.escaped_ptls.*.bp: tindex (DiagEscapedPtlsTindex)

Native units: Positive integer.

Diagnostic cadence index computed as integer gstep divided by the configured diagnostic period. This index is not the zero-based ADIOS step and need not start at 1 in a restarted or selected product.

  • plot_label: Output slice index

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/diagnostics/diag_escaped_ptls.cpp:252.

xgc.f0.*.bp: dpot (DiagF0Dpot)

Native units: V.

Perturbed electrostatic potential on each toroidal plane and 2D mesh node.

  • plot_label: $\delta\phi$

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz, nd_connect_list for axisymmetric geometry; rplane_rz, rplane_nd_connect_list for stellarator right-plane geometry. Connectivity is zero-based. Node-centered field with continuous linear barycentric basis functions on each triangle.

  • coordinate_context: Right field plane in XGC1: phi=(plane_index+1)*xgc.mesh.bp:delta_phi; use rplane_rz/rplane_nd_connect_list for stellarators. Axisymmetric XGCa has one poloidal mesh.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/phi_potential(:)/values

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_f0.cpp:164.

xgc.f0.*.bp: dpot (DiagF0Dpot2)

Native units: V.

Axisymmetric perturbed electrostatic potential on the 2D mesh.

  • plot_label: $\delta\phi_{n=0}$

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz, nd_connect_list for axisymmetric geometry; rplane_rz, rplane_nd_connect_list for stellarator right-plane geometry. Connectivity is zero-based. Node-centered field with continuous linear barycentric basis functions on each triangle.

  • coordinate_context: Right field plane in XGC1: phi=(plane_index+1)*xgc.mesh.bp:delta_phi; use rplane_rz/rplane_nd_connect_list for stellarators. Axisymmetric XGCa has one poloidal mesh.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/phi_potential(:)/values

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_f0.cpp:170.

xgc.f0.*.bp: eden_f0 (DiagF0EdenF0)

Native units: m^-3.

Electron density contribution from the residual/grid distribution f0g on the midplane mesh; the analytic background and marker contributions are not included.

  • plot_label: $\delta n_{e,f0}$

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

Source: XGC_core/cpp/diagnostics/diag_f0.cpp:143.

xgc.f0.*.bp: gstep (DiagF0Gstep)

Native units: Non-negative integer.

XGC global simulation time step for this f0 diagnostic output slice.

  • plot_label: Global time step

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/diagnostics/diag_f0.cpp:67.

xgc.f0.*.bp: iden_f0_approx (DiagF0IdenF0Approx)

Native units: m^-3.

Approximate ion charge-number-weighted density contribution from residual/grid f0g, summed over ion species and gyro-radius bins on the midplane mesh; no ion gyroaverage is applied to this sum.

  • plot_label: $\delta n_{i,f0}$

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

Source: XGC_core/cpp/diagnostics/diag_f0.cpp:134.

xgc.f0.*.bp: imu1m1 (DiagF0Imu1m1)

Native units: Non-negative integer.

Legacy decomposition metadata: zero-based perpendicular-velocity-grid offset of this write.

  • plot_label: Perpendicular-velocity offset

Source: XGC_core/cpp/diagnostics/diag_f0.cpp:99.

xgc.f0.*.bp: inode1m1 (DiagF0Inode1m1)

Native units: Non-negative integer.

Legacy decomposition metadata: zero-based global node offset of the first local mesh node in this write.

  • plot_label: Global mesh-node offset

Source: XGC_core/cpp/diagnostics/diag_f0.cpp:87.

xgc.f0.*.bp: iphi (DiagF0Iphi)

Native units: Non-negative integer.

Legacy decomposition metadata: zero-based toroidal plane index written by this rank.

  • plot_label: $i_\varphi$

Source: XGC_core/cpp/diagnostics/diag_f0.cpp:82.

xgc.f0.*.bp: mudata (DiagF0Mudata)

Native units: Positive integer.

Number of perpendicular velocity grid points.

  • plot_label: $N_{v,\perp}$

Source: XGC_core/cpp/diagnostics/diag_f0.cpp:109.

xgc.f0.*.bp: ndata (DiagF0Ndata)

Native units: Positive integer.

Legacy decomposition metadata: number of local mesh nodes written by this rank.

  • plot_label: Local mesh-node count

Source: XGC_core/cpp/diagnostics/diag_f0.cpp:94.

xgc.f0.*.bp: nmup1 (DiagF0Nmup1)

Native units: Positive integer.

Number of perpendicular velocity grid points.

  • plot_label: $N_{v,\perp}$

Source: XGC_core/cpp/diagnostics/diag_f0.cpp:104.

xgc.f0.*.bp: nnode (DiagF0Nnode)

Native units: Positive integer.

Number of 2D mesh nodes per plane.

  • plot_label: $N_n$

Source: XGC_core/cpp/diagnostics/diag_f0.cpp:57.

xgc.f0.*.bp: nphi (DiagF0Nphi)

Native units: Positive integer.

Number of toroidal mesh planes \(N_{\\mathrm{planes}}\).

  • plot_label: $N_\varphi$

Source: XGC_core/cpp/diagnostics/diag_f0.cpp:77.

xgc.f0.*.bp: sp_name + “f” (DiagF0SpNameF)

Native units: m^-3 eV^-1.

Full distribution function multiplied by the normalized perpendicular velocity factor, i.e., \(v_{\\perp} f(v_{\\perp}, v_{\\parallel})\), on [toroidal plane, perpendicular velocity, mesh node, parallel velocity].

  • plot_label: $f(\boldsymbol{x}, v_\perp, v_\parallel)$ ({name})

  • axes: toroidal_plane,velocity_mu,mesh_node,velocity_parallel

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh. Velocity-grid coordinates and quadrature are defined by xgc.f0.mesh.bp: f0_nmu, f0_nvp, f0_dsmu, f0_dvp, f0_vp_max, f0_fg_T_ev, f0_grid_vol_vonly; see normalization.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Velocity coordinates use xgc.f0.mesh.bp: f0_nmu/f0_nvp are interval counts; there are f0_nmu+1 perpendicular and 2*f0_nvp+1 parallel vertices. At zero-based indices j,k, vperp_norm=j*f0_dsmu except vperp_norm[0]=f0_dsmu/3, and vpara_norm=k*f0_dvp-f0_vp_max. Physical speed is normalized speed times sqrt(EV_2_J*f0_fg_T_ev[species,node]/mass_kg), EV_2_J=1.6022e-19 J/eV; mass_kg=xgc.units.bp:<species>_ptl_mass_au*1.6720e-27 kg. These are the native XGC constants EV_2_J and PROTON_MASS, not the exact SI elementary charge or unified atomic mass constant. This is XGC thermal_velocity; the analysis reader uses abs(charge_C)*T/m and agrees for singly charged species only. Stored f/f0g already absorbs the perpendicular Jacobian. Density is f0_grid_vol_vonly[species,node]*sum(mu_vol[j]*vp_vol[k]*f[j,node,k]); endpoint factors are 0.5, interior factors 1. Do not add another vperp, dsmu*dvp or 2*pi factor to that formula. f0n and accumulation counters use these coordinates but are not distributions to integrate.

Source: XGC_core/cpp/diagnostics/diag_f0.cpp:185.

xgc.f0.*.bp: sp_name + “f0g” (DiagF0SpNameF0g)

Native units: m^-3 eV^-1.

Residual/grid contribution to the species f0 distribution function on [toroidal plane, perpendicular velocity, mesh node, parallel velocity].

  • plot_label: $f_g(\boldsymbol{x}, v_\perp, v_\parallel)$ ({name})

  • axes: toroidal_plane,velocity_mu,mesh_node,velocity_parallel

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh. Velocity-grid coordinates and quadrature are defined by xgc.f0.mesh.bp: f0_nmu, f0_nvp, f0_dsmu, f0_dvp, f0_vp_max, f0_fg_T_ev, f0_grid_vol_vonly; see normalization.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Velocity coordinates use xgc.f0.mesh.bp: f0_nmu/f0_nvp are interval counts; there are f0_nmu+1 perpendicular and 2*f0_nvp+1 parallel vertices. At zero-based indices j,k, vperp_norm=j*f0_dsmu except vperp_norm[0]=f0_dsmu/3, and vpara_norm=k*f0_dvp-f0_vp_max. Physical speed is normalized speed times sqrt(EV_2_J*f0_fg_T_ev[species,node]/mass_kg), EV_2_J=1.6022e-19 J/eV; mass_kg=xgc.units.bp:<species>_ptl_mass_au*1.6720e-27 kg. These are the native XGC constants EV_2_J and PROTON_MASS, not the exact SI elementary charge or unified atomic mass constant. This is XGC thermal_velocity; the analysis reader uses abs(charge_C)*T/m and agrees for singly charged species only. Stored f/f0g already absorbs the perpendicular Jacobian. Density is f0_grid_vol_vonly[species,node]*sum(mu_vol[j]*vp_vol[k]*f[j,node,k]); endpoint factors are 0.5, interior factors 1. Do not add another vperp, dsmu*dvp or 2*pi factor to that formula. f0n and accumulation counters use these coordinates but are not distributions to integrate.

Source: XGC_core/cpp/diagnostics/diag_f0.cpp:199.

xgc.f0.*.bp: sp_name + “f0n” (DiagF0SpNameF0n)

Native units: dimensionless.

Particle-to-grid interpolation normalization for the species f0 distribution on [toroidal plane, perpendicular velocity, mesh node, parallel velocity].

  • plot_label: $g(\boldsymbol{x}, v_\perp, v_\parallel)$ ({name})

  • axes: toroidal_plane,velocity_mu,mesh_node,velocity_parallel

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh. Velocity-grid coordinates and quadrature are defined by xgc.f0.mesh.bp: f0_nmu, f0_nvp, f0_dsmu, f0_dvp, f0_vp_max, f0_fg_T_ev, f0_grid_vol_vonly; see normalization.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Dimensionless particle-to-grid interpolation normalization on the same f0 velocity grid. Not a distribution function; do not apply a density moment formula to this channel.

Source: XGC_core/cpp/diagnostics/diag_f0.cpp:213.

xgc.f0.*.bp: time (DiagF0Time)

Native units: s.

Simulation time in seconds.

  • plot_label: $t$

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions. Clock convention: in the regular main loop gstep is incremented before diagnostic calls, but sml.time is incremented afterward. A checkpoint written later at the same gstep therefore records time larger by one physical timestep. Use the written time, global step and product/integration stage together; equal gstep alone does not establish simultaneous diagnostic and checkpoint states.

  • imas_candidates: distributions/time

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_high

Source: XGC_core/cpp/diagnostics/diag_f0.cpp:62.

xgc.f0.*.bp: tindex (DiagF0Tindex)

Native units: Positive integer.

Index of this f0 diagnostic output time slice. Values: 1, 2, …, # outputs.

  • plot_label: Output slice index

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/diagnostics/diag_f0.cpp:72.

xgc.f0.*.bp: vpdata (DiagF0Vpdata)

Native units: Positive integer.

Number of parallel velocity grid points.

  • plot_label: $N_{v,\parallel}$

Source: XGC_core/cpp/diagnostics/diag_f0.cpp:114.

xgc.f0.mesh.bp: f0_T_ev (DumpF0GridF0TEv)

Native units: eV.

Initial analytic-background temperature for each species with a dynamic f0 and mesh node.

  • plot_label: $T_0$

  • axes: species,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • species_context: Species axis contains only species with dynamic_f0=true, in simulation order (compact species.f0_idx order), with nsp rows. Species without a dynamic f0 are omitted; electrons come first when included. Mass/charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu; composition is in xgc.used_inputs.bp.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/electrons/temperature(:)/values; plasma_profiles/ggd(:)/ion(:)/temperature(:)/values

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/plasma/dump_f0_grid.cpp:113.

xgc.f0.mesh.bp: f0_den (DumpF0GridF0Den)

Native units: m^-3.

Initial analytic-background density for each species with a dynamic f0 and mesh node.

  • plot_label: $n_0$

  • axes: species,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • species_context: Species axis contains only species with dynamic_f0=true, in simulation order (compact species.f0_idx order), with nsp rows. Species without a dynamic f0 are omitted; electrons come first when included. Mass/charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu; composition is in xgc.used_inputs.bp.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/electrons/density(:)/values; plasma_profiles/ggd(:)/ion(:)/density(:)/values

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/plasma/dump_f0_grid.cpp:106.

xgc.f0.mesh.bp: f0_dsmu (DumpF0GridF0Dsmu)

Native units: dimensionless.

\(\\Delta v_{\\perp}\) in the 2D velocity grid of species \(s\).

  • plot_label: $\Delta v_{\perp,\mathrm{norm}}$

  • normalization: Velocity coordinates use xgc.f0.mesh.bp: f0_nmu/f0_nvp are interval counts; there are f0_nmu+1 perpendicular and 2*f0_nvp+1 parallel vertices. At zero-based indices j,k, vperp_norm=j*f0_dsmu except vperp_norm[0]=f0_dsmu/3, and vpara_norm=k*f0_dvp-f0_vp_max. Physical speed is normalized speed times sqrt(EV_2_J*f0_fg_T_ev[species,node]/mass_kg), EV_2_J=1.6022e-19 J/eV; mass_kg=xgc.units.bp:<species>_ptl_mass_au*1.6720e-27 kg. These are the native XGC constants EV_2_J and PROTON_MASS, not the exact SI elementary charge or unified atomic mass constant. This is XGC thermal_velocity; the analysis reader uses abs(charge_C)*T/m and agrees for singly charged species only. Stored f/f0g already absorbs the perpendicular Jacobian. Density is f0_grid_vol_vonly[species,node]*sum(mu_vol[j]*vp_vol[k]*f[j,node,k]); endpoint factors are 0.5, interior factors 1. Do not add another vperp, dsmu*dvp or 2*pi factor to that formula. f0n and accumulation counters use these coordinates but are not distributions to integrate.

Source: XGC_core/cpp/plasma/dump_f0_grid.cpp:92.

xgc.f0.mesh.bp: f0_dvp (DumpF0GridF0Dvp)

Native units: dimensionless.

\(\\Delta v_{\\parallel}\) in the 2D velocity grid of species \(s\).

  • plot_label: $\Delta v_{\parallel,\mathrm{norm}}$

  • normalization: Velocity coordinates use xgc.f0.mesh.bp: f0_nmu/f0_nvp are interval counts; there are f0_nmu+1 perpendicular and 2*f0_nvp+1 parallel vertices. At zero-based indices j,k, vperp_norm=j*f0_dsmu except vperp_norm[0]=f0_dsmu/3, and vpara_norm=k*f0_dvp-f0_vp_max. Physical speed is normalized speed times sqrt(EV_2_J*f0_fg_T_ev[species,node]/mass_kg), EV_2_J=1.6022e-19 J/eV; mass_kg=xgc.units.bp:<species>_ptl_mass_au*1.6720e-27 kg. These are the native XGC constants EV_2_J and PROTON_MASS, not the exact SI elementary charge or unified atomic mass constant. This is XGC thermal_velocity; the analysis reader uses abs(charge_C)*T/m and agrees for singly charged species only. Stored f/f0g already absorbs the perpendicular Jacobian. Density is f0_grid_vol_vonly[species,node]*sum(mu_vol[j]*vp_vol[k]*f[j,node,k]); endpoint factors are 0.5, interior factors 1. Do not add another vperp, dsmu*dvp or 2*pi factor to that formula. f0n and accumulation counters use these coordinates but are not distributions to integrate.

Source: XGC_core/cpp/plasma/dump_f0_grid.cpp:99.

xgc.f0.mesh.bp: f0_fg_T_ev (DumpF0GridF0FgTEv)

Native units: eV.

Temperature used for calculating the thermal speed used to normalize velocities.

  • plot_label: $T_{\mathrm{grid}}$

  • axes: species,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • species_context: Species axis contains only species with dynamic_f0=true, in simulation order (compact species.f0_idx order), with nsp rows. Species without a dynamic f0 are omitted; electrons come first when included. Mass/charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu; composition is in xgc.used_inputs.bp.

  • centering: node

  • normalization: Velocity coordinates use xgc.f0.mesh.bp: f0_nmu/f0_nvp are interval counts; there are f0_nmu+1 perpendicular and 2*f0_nvp+1 parallel vertices. At zero-based indices j,k, vperp_norm=j*f0_dsmu except vperp_norm[0]=f0_dsmu/3, and vpara_norm=k*f0_dvp-f0_vp_max. Physical speed is normalized speed times sqrt(EV_2_J*f0_fg_T_ev[species,node]/mass_kg), EV_2_J=1.6022e-19 J/eV; mass_kg=xgc.units.bp:<species>_ptl_mass_au*1.6720e-27 kg. These are the native XGC constants EV_2_J and PROTON_MASS, not the exact SI elementary charge or unified atomic mass constant. This is XGC thermal_velocity; the analysis reader uses abs(charge_C)*T/m and agrees for singly charged species only. Stored f/f0g already absorbs the perpendicular Jacobian. Density is f0_grid_vol_vonly[species,node]*sum(mu_vol[j]*vp_vol[k]*f[j,node,k]); endpoint factors are 0.5, interior factors 1. Do not add another vperp, dsmu*dvp or 2*pi factor to that formula. f0n and accumulation counters use these coordinates but are not distributions to integrate.

Source: XGC_core/cpp/plasma/dump_f0_grid.cpp:120.

xgc.f0.mesh.bp: f0_flow (DumpF0GridF0Flow)

Native units: m/s.

Initial parallel bulk flow of the plasma. Positive flow implies motion in the direction of the magnetic field.

  • plot_label: $u_{\parallel,0}$

  • axes: species,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • species_context: Species axis contains only species with dynamic_f0=true, in simulation order (compact species.f0_idx order), with nsp rows. Species without a dynamic f0 are omitted; electrons come first when included. Mass/charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu; composition is in xgc.used_inputs.bp.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/electrons/velocity(:)/parallel; plasma_profiles/ggd(:)/ion(:)/velocity(:)/parallel

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/plasma/dump_f0_grid.cpp:125.

xgc.f0.mesh.bp: f0_grid_vol_vonly (DumpF0GridF0GridVolVonly)

Native units: eV.

Prefactor for calculating velocity-space moments: \(T_s/\\sqrt{2\\pi} \\, dsmu_n \\, dvp_n\). This assumes that the distribution function \(f\) has absorbed the \(v_{\\perp}\) Jacobian factor and that the velocity-boundary coefficients mu_vol and vp_vol (0.5 on the boundary, 1 otherwise) are not included. For example, the density moment is \(n_s(k)=f0\\_grid\\_vol\\_vonly(k) \\sum_{i,j} mu\\_vol(i) vp\\_vol(j) f(i,j,k)\).

  • plot_label: Velocity-space quadrature prefactor

  • axes: species,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • species_context: Species axis contains only species with dynamic_f0=true, in simulation order (compact species.f0_idx order), with nsp rows. Species without a dynamic f0 are omitted; electrons come first when included. Mass/charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu; composition is in xgc.used_inputs.bp.

  • centering: node

  • normalization: Velocity coordinates use xgc.f0.mesh.bp: f0_nmu/f0_nvp are interval counts; there are f0_nmu+1 perpendicular and 2*f0_nvp+1 parallel vertices. At zero-based indices j,k, vperp_norm=j*f0_dsmu except vperp_norm[0]=f0_dsmu/3, and vpara_norm=k*f0_dvp-f0_vp_max. Physical speed is normalized speed times sqrt(EV_2_J*f0_fg_T_ev[species,node]/mass_kg), EV_2_J=1.6022e-19 J/eV; mass_kg=xgc.units.bp:<species>_ptl_mass_au*1.6720e-27 kg. These are the native XGC constants EV_2_J and PROTON_MASS, not the exact SI elementary charge or unified atomic mass constant. This is XGC thermal_velocity; the analysis reader uses abs(charge_C)*T/m and agrees for singly charged species only. Stored f/f0g already absorbs the perpendicular Jacobian. Density is f0_grid_vol_vonly[species,node]*sum(mu_vol[j]*vp_vol[k]*f[j,node,k]); endpoint factors are 0.5, interior factors 1. Do not add another vperp, dsmu*dvp or 2*pi factor to that formula. f0n and accumulation counters use these coordinates but are not distributions to integrate.

Source: XGC_core/cpp/plasma/dump_f0_grid.cpp:131.

xgc.f0.mesh.bp: f0_nmu (DumpF0GridF0Nmu)

Native units: Positive integer.

Number of intervals from zero to f0_smu_max in the normalized perpendicular-velocity grid. Equal to input f0_nmu; the stored distribution has f0_nmu + 1 perpendicular grid vertices.

  • plot_label: $N_{v,\perp}-1$

  • normalization: Velocity coordinates use xgc.f0.mesh.bp: f0_nmu/f0_nvp are interval counts; there are f0_nmu+1 perpendicular and 2*f0_nvp+1 parallel vertices. At zero-based indices j,k, vperp_norm=j*f0_dsmu except vperp_norm[0]=f0_dsmu/3, and vpara_norm=k*f0_dvp-f0_vp_max. Physical speed is normalized speed times sqrt(EV_2_J*f0_fg_T_ev[species,node]/mass_kg), EV_2_J=1.6022e-19 J/eV; mass_kg=xgc.units.bp:<species>_ptl_mass_au*1.6720e-27 kg. These are the native XGC constants EV_2_J and PROTON_MASS, not the exact SI elementary charge or unified atomic mass constant. This is XGC thermal_velocity; the analysis reader uses abs(charge_C)*T/m and agrees for singly charged species only. Stored f/f0g already absorbs the perpendicular Jacobian. Density is f0_grid_vol_vonly[species,node]*sum(mu_vol[j]*vp_vol[k]*f[j,node,k]); endpoint factors are 0.5, interior factors 1. Do not add another vperp, dsmu*dvp or 2*pi factor to that formula. f0n and accumulation counters use these coordinates but are not distributions to integrate.

Source: XGC_core/cpp/plasma/dump_f0_grid.cpp:61.

xgc.f0.mesh.bp: f0_nvp (DumpF0GridF0Nvp)

Native units: Positive integer.

Number of intervals from zero to f0_vp_max on each half of the normalized parallel-velocity grid. Equal to input f0_nvp; the stored distribution has 2*f0_nvp + 1 parallel grid vertices.

  • plot_label: $(N_{v,\parallel}-1)/2$

  • normalization: Velocity coordinates use xgc.f0.mesh.bp: f0_nmu/f0_nvp are interval counts; there are f0_nmu+1 perpendicular and 2*f0_nvp+1 parallel vertices. At zero-based indices j,k, vperp_norm=j*f0_dsmu except vperp_norm[0]=f0_dsmu/3, and vpara_norm=k*f0_dvp-f0_vp_max. Physical speed is normalized speed times sqrt(EV_2_J*f0_fg_T_ev[species,node]/mass_kg), EV_2_J=1.6022e-19 J/eV; mass_kg=xgc.units.bp:<species>_ptl_mass_au*1.6720e-27 kg. These are the native XGC constants EV_2_J and PROTON_MASS, not the exact SI elementary charge or unified atomic mass constant. This is XGC thermal_velocity; the analysis reader uses abs(charge_C)*T/m and agrees for singly charged species only. Stored f/f0g already absorbs the perpendicular Jacobian. Density is f0_grid_vol_vonly[species,node]*sum(mu_vol[j]*vp_vol[k]*f[j,node,k]); endpoint factors are 0.5, interior factors 1. Do not add another vperp, dsmu*dvp or 2*pi factor to that formula. f0n and accumulation counters use these coordinates but are not distributions to integrate.

Source: XGC_core/cpp/plasma/dump_f0_grid.cpp:69.

xgc.f0.mesh.bp: f0_smu_max (DumpF0GridF0SmuMax)

Native units: dimensionless.

Maximum \(v_{\\perp}\) in the 2D velocity grid (minimum is 0). Same as input parameter f0_smu_max.

  • plot_label: $\mathrm{max}(v_{\perp,\mathrm{norm}})$

  • normalization: Velocity coordinates use xgc.f0.mesh.bp: f0_nmu/f0_nvp are interval counts; there are f0_nmu+1 perpendicular and 2*f0_nvp+1 parallel vertices. At zero-based indices j,k, vperp_norm=j*f0_dsmu except vperp_norm[0]=f0_dsmu/3, and vpara_norm=k*f0_dvp-f0_vp_max. Physical speed is normalized speed times sqrt(EV_2_J*f0_fg_T_ev[species,node]/mass_kg), EV_2_J=1.6022e-19 J/eV; mass_kg=xgc.units.bp:<species>_ptl_mass_au*1.6720e-27 kg. These are the native XGC constants EV_2_J and PROTON_MASS, not the exact SI elementary charge or unified atomic mass constant. This is XGC thermal_velocity; the analysis reader uses abs(charge_C)*T/m and agrees for singly charged species only. Stored f/f0g already absorbs the perpendicular Jacobian. Density is f0_grid_vol_vonly[species,node]*sum(mu_vol[j]*vp_vol[k]*f[j,node,k]); endpoint factors are 0.5, interior factors 1. Do not add another vperp, dsmu*dvp or 2*pi factor to that formula. f0n and accumulation counters use these coordinates but are not distributions to integrate.

Source: XGC_core/cpp/plasma/dump_f0_grid.cpp:77.

xgc.f0.mesh.bp: f0_vp_max (DumpF0GridF0VpMax)

Native units: dimensionless.

Maximum absolute \(v_{\\parallel}\) in the 2D velocity grid; the minimum is the negative of this value. Same as input parameter f0_vp_max.

  • plot_label: $\mathrm{max}(v_{\parallel,\mathrm{norm}})$

  • normalization: Velocity coordinates use xgc.f0.mesh.bp: f0_nmu/f0_nvp are interval counts; there are f0_nmu+1 perpendicular and 2*f0_nvp+1 parallel vertices. At zero-based indices j,k, vperp_norm=j*f0_dsmu except vperp_norm[0]=f0_dsmu/3, and vpara_norm=k*f0_dvp-f0_vp_max. Physical speed is normalized speed times sqrt(EV_2_J*f0_fg_T_ev[species,node]/mass_kg), EV_2_J=1.6022e-19 J/eV; mass_kg=xgc.units.bp:<species>_ptl_mass_au*1.6720e-27 kg. These are the native XGC constants EV_2_J and PROTON_MASS, not the exact SI elementary charge or unified atomic mass constant. This is XGC thermal_velocity; the analysis reader uses abs(charge_C)*T/m and agrees for singly charged species only. Stored f/f0g already absorbs the perpendicular Jacobian. Density is f0_grid_vol_vonly[species,node]*sum(mu_vol[j]*vp_vol[k]*f[j,node,k]); endpoint factors are 0.5, interior factors 1. Do not add another vperp, dsmu*dvp or 2*pi factor to that formula. f0n and accumulation counters use these coordinates but are not distributions to integrate.

Source: XGC_core/cpp/plasma/dump_f0_grid.cpp:84.

xgc.f0.mesh.bp: gradpsi (DumpF0GridGradpsi)

Native units: T m.

Poloidal magnetic flux gradient at each mesh node.

  • plot_label: $\nabla\psi$

  • axes: mesh_node,vector_component

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Gradient components in cylindrical R,Z order.

  • centering: node

Source: XGC_core/cpp/plasma/dump_f0_grid.cpp:175.

xgc.f0.mesh.bp: gvid0_pid (DumpF0GridGvid0Pid)

Native units: Positive integer.

Poloidal mesh partition table. It has length pe_per_plane + 1 and stores the first global vertex ID owned by each rank within a toroidal plane, using XGC’s legacy 1-based convention.

  • plot_label: Mesh partition node offset

Source: XGC_core/cpp/plasma/dump_f0_grid.cpp:180.

xgc.f0.mesh.bp: inode1m1 (DumpF0GridInode1m1)

Native units: Non-negative integer.

Legacy decomposition metadata: offset of the first vertex on the last MPI rank that wrote to this file.

  • plot_label: Global mesh-node offset

Source: XGC_core/cpp/plasma/dump_f0_grid.cpp:54.

xgc.f0.mesh.bp: n_n (DumpF0GridNN)

Native units: Positive integer.

Number of mesh nodes per plane.

  • plot_label: $N_n$

Source: XGC_core/cpp/plasma/dump_f0_grid.cpp:38.

xgc.f0.mesh.bp: nb_curl_nb (DumpF0GridNbCurlNb)

Native units: m^-1.

Part of the phase space Jacobian (D in XGC’s equations), \hat{\boldsymbol{b}}\cdot\nabla\times\hat{\boldsymbol{b}}.

  • plot_label: $\hat{\boldsymbol{b}} \cdot \nabla\times \hat{\boldsymbol{b}}$

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • centering: node

Source: XGC_core/cpp/plasma/dump_f0_grid.cpp:169.

xgc.f0.mesh.bp: ndata (DumpF0GridNdata)

Native units: Positive integer.

Legacy decomposition metadata: number of mesh vertices assigned to the last MPI rank that wrote to this file.

  • plot_label: Local mesh-node count

Source: XGC_core/cpp/plasma/dump_f0_grid.cpp:48.

xgc.f0.mesh.bp: nsp (DumpF0GridNsp)

Native units: Positive integer.

Number of species with a dynamic f0 in the output.

  • plot_label: $N_{\mathrm{sp}}$

Source: XGC_core/cpp/plasma/dump_f0_grid.cpp:43.

xgc.f0.mesh.bp: pe_per_plane (DumpF0GridPePerPlane)

Native units: Positive integer.

Number of MPI ranks per plane.

  • plot_label: MPI ranks per plane

Source: XGC_core/cpp/plasma/dump_f0_grid.cpp:152.

xgc.f0.mesh.bp: v_curv (DumpF0GridVCurv)

Native units: radial: T^2, poloidal and toroidal: T m^-1.

Factors for curvature drift, \(\\nabla\\times B\), in radial, poloidal, and toroidal components. The radial component includes a factor \(\|\\nabla\\psi\|=R B_P\), hence the different units.

  • plot_label: $F_{v,\kappa}$

  • axes: mesh_node,vector_component

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Components ordered radial (psi), poloidal, toroidal; radial components include |grad psi| and have different units from the others (see unit).

  • centering: node

Source: XGC_core/cpp/plasma/dump_f0_grid.cpp:163.

xgc.f0.mesh.bp: v_gradb (DumpF0GridVGradb)

Native units: radial: T^3, poloidal and toroidal: T^2 m^-1.

Factors for grad-B drift, \(B\\times\\nabla \|B\|\), in radial, poloidal, and toroidal components. The radial component includes a factor \(\|\\nabla\\psi\|=R B_P\), hence the different units.

  • plot_label: $F_{v,\nabla(B)}$

  • axes: mesh_node,vector_component

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Components ordered radial (psi), poloidal, toroidal; radial components include |grad psi| and have different units from the others (see unit).

  • centering: node

Source: XGC_core/cpp/plasma/dump_f0_grid.cpp:157.

xgc.f2d/f3d.*.bp: [species]_par_en_flux_[n0/turb]_[f0/df] (Diag3dF0FParEnFlux)

Native units: J m^-2 s^-1.

Parallel kinetic-energy flux density: velocity integral of f E v_parallel, toroidally averaged; density included and convective energy transport retained.

  • plot_label: $Q_\parallel$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Density-weighted velocity integrals; no division by density, so do not multiply by density. Parallel kernels use physical v_parallel in m/s; no poloidal-flux-gradient conversion is needed. par_mom_flux transports toroidal angular momentum L_phi=m v_parallel R B_phi/|B|. n0 and turb are toroidally averaged mean-product and correlated fluctuation contributions; f0 and df distinguish adiabatic and nonadiabatic distribution contributions. XGCa omits n0/turb in names. These channels have no energy-bin or toroidal-plane axis.

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:133.

xgc.f2d/f3d.*.bp: [species]_par_mom_flux_[n0/turb]_[f0/df] (Diag3dF0FParMomFlux)

Native units: kg s^-2.

Parallel flux density of toroidal angular momentum: velocity integral of f (m v_parallel R B_phi/|B|) v_parallel, toroidally averaged; mass and density included.

  • plot_label: $\Pi_{\varphi,\parallel}$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Density-weighted velocity integrals; no division by density, so do not multiply by density. Parallel kernels use physical v_parallel in m/s; no poloidal-flux-gradient conversion is needed. par_mom_flux transports toroidal angular momentum L_phi=m v_parallel R B_phi/|B|. n0 and turb are toroidally averaged mean-product and correlated fluctuation contributions; f0 and df distinguish adiabatic and nonadiabatic distribution contributions. XGCa omits n0/turb in names. These channels have no energy-bin or toroidal-plane axis.

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:127.

xgc.f2d/f3d.*.bp: [species]_par_pot_en_flux_[n0/turb]_[f0/df] (Diag3dF0FParPotEnFlux)

Native units: J m^-2 s^-1.

Parallel potential-energy flux density: velocity integral of f charge phi v_parallel, toroidally averaged; density included.

  • plot_label: $Q_{\mathrm{pot},\parallel}$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Density-weighted velocity integrals; no division by density, so do not multiply by density. Parallel kernels use physical v_parallel in m/s; no poloidal-flux-gradient conversion is needed. par_mom_flux transports toroidal angular momentum L_phi=m v_parallel R B_phi/|B|. n0 and turb are toroidally averaged mean-product and correlated fluctuation contributions; f0 and df distinguish adiabatic and nonadiabatic distribution contributions. XGCa omits n0/turb in names. These channels have no energy-bin or toroidal-plane axis.

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:139.

xgc.f2d/f3d.*.bp: [species]_parallel_flow_[n0/turb]_[f0/df] (Diag3dF0FParallelFlow)

Native units: m/s.

Toroidally averaged, density-normalized parallel-flow contribution.

  • plot_label: $u_\parallel$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Only the toroidal, poloidal and parallel flow-velocity channels are divided by the local total gyrocenter density before toroidal averaging. Their f0/df and n0/turb contributions use that same denominator. n0 and turb are toroidally averaged mean-product and correlated fluctuation contributions; f0 and df distinguish adiabatic and nonadiabatic distribution contributions. XGCa omits n0/turb in names. These channels have no energy-bin or toroidal-plane axis.

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:30.

xgc.f2d/f3d.*.bp: [species]_poloidal_flow_[n0/turb]_[f0/df] (Diag3dF0FPoloidalFlow)

Native units: m/s.

Toroidally averaged, density-normalized gyrocenter poloidal-flow contribution. Differs from the physical poloidal flow by the diamagnetic flow.

  • plot_label: $u_{\theta,\mathrm{gc}}$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Only the toroidal, poloidal and parallel flow-velocity channels are divided by the local total gyrocenter density before toroidal averaging. Their f0/df and n0/turb contributions use that same denominator. n0 and turb are toroidally averaged mean-product and correlated fluctuation contributions; f0 and df distinguish adiabatic and nonadiabatic distribution contributions. XGCa omits n0/turb in names. These channels have no energy-bin or toroidal-plane axis.

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:24.

xgc.f2d/f3d.*.bp: [species]_rad_mom_flux_3db_[n0/turb]_[f0/df] (Diag3dF0FRadMomFlux3db)

Native units: kg T m s^-2.

Radial flux density of toroidal angular momentum in psi space from parallel motion across flux surfaces in the 3D magnetic field, toroidally averaged; mass and density included.

  • plot_label: $\Pi_{\varphi,\psi}^{\delta B}$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Density-weighted velocity integrals; no division by density, so do not multiply by density. Radial kernels use v dot grad(psi), with psi the native poloidal flux/(2*pi) in T m^2. Divide pointwise by |grad psi| from xgc.f0.mesh.bp:gradpsi to obtain flux per physical area. For surface profiles, the XGC-Analysis neoclassical convention is FSA(F_psi)/sqrt(FSA(|grad psi|^2)); use line-integral FSA for magnetic-drift cancellation. n0 and turb are toroidally averaged mean-product and correlated fluctuation contributions; f0 and df distinguish adiabatic and nonadiabatic distribution contributions. XGCa omits n0/turb in names. These channels have no energy-bin or toroidal-plane axis.

  • related_variables: xgc.f0.mesh.bp: gradpsi (R,Z components, T m)

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:71.

xgc.f2d/f3d.*.bp: [species]_rad_mom_flux_ExB_[n0/turb]_[f0/df] (Diag3dF0FRadMomFluxExB)

Native units: kg T m s^-2.

Radial flux density of toroidal angular momentum in psi space from ExB drift, toroidally averaged; mass and density included.

  • plot_label: $\Pi_{\varphi,\psi}^{E\times B}$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Density-weighted velocity integrals; no division by density, so do not multiply by density. Radial kernels use v dot grad(psi), with psi the native poloidal flux/(2*pi) in T m^2. Divide pointwise by |grad psi| from xgc.f0.mesh.bp:gradpsi to obtain flux per physical area. For surface profiles, the XGC-Analysis neoclassical convention is FSA(F_psi)/sqrt(FSA(|grad psi|^2)); use line-integral FSA for magnetic-drift cancellation. n0 and turb are toroidally averaged mean-product and correlated fluctuation contributions; f0 and df distinguish adiabatic and nonadiabatic distribution contributions. XGCa omits n0/turb in names. These channels have no energy-bin or toroidal-plane axis.

  • related_variables: xgc.f0.mesh.bp: gradpsi (R,Z components, T m)

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:59.

xgc.f2d/f3d.*.bp: [species]_rad_mom_flux_mag_[n0/turb]_[f0/df] (Diag3dF0FRadMomFluxMag)

Native units: kg T m s^-2.

Radial flux density of toroidal angular momentum in psi space from magnetic drifts, toroidally averaged; mass and density included. Use line-integral flux-surface averaging.

  • plot_label: $\Pi_{\varphi,\psi}^{\mathrm{mag}}$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Density-weighted velocity integrals; no division by density, so do not multiply by density. Radial kernels use v dot grad(psi), with psi the native poloidal flux/(2*pi) in T m^2. Divide pointwise by |grad psi| from xgc.f0.mesh.bp:gradpsi to obtain flux per physical area. For surface profiles, the XGC-Analysis neoclassical convention is FSA(F_psi)/sqrt(FSA(|grad psi|^2)); use line-integral FSA for magnetic-drift cancellation. n0 and turb are toroidally averaged mean-product and correlated fluctuation contributions; f0 and df distinguish adiabatic and nonadiabatic distribution contributions. XGCa omits n0/turb in names. These channels have no energy-bin or toroidal-plane axis.

  • related_variables: xgc.f0.mesh.bp: gradpsi (R,Z components, T m)

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:65.

xgc.f2d/f3d.*.bp: [species]_radial_en_flux_3db_[n0/turb]_[f0/df] (Diag3dF0FRadialEnFlux3db)

Native units: J T m^-1 s^-1.

Radial kinetic-energy flux density in psi space from parallel motion across flux surfaces in the 3D magnetic field, toroidally averaged; density included and convective energy transport retained.

  • plot_label: $Q_\psi^{\delta B}$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Density-weighted velocity integrals; no division by density, so do not multiply by density. Radial kernels use v dot grad(psi), with psi the native poloidal flux/(2*pi) in T m^2. Divide pointwise by |grad psi| from xgc.f0.mesh.bp:gradpsi to obtain flux per physical area. For surface profiles, the XGC-Analysis neoclassical convention is FSA(F_psi)/sqrt(FSA(|grad psi|^2)); use line-integral FSA for magnetic-drift cancellation. n0 and turb are toroidally averaged mean-product and correlated fluctuation contributions; f0 and df distinguish adiabatic and nonadiabatic distribution contributions. XGCa omits n0/turb in names. These channels have no energy-bin or toroidal-plane axis.

  • related_variables: xgc.f0.mesh.bp: gradpsi (R,Z components, T m)

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:89.

xgc.f2d/f3d.*.bp: [species]_radial_en_flux_ExB_[n0/turb]_[f0/df] (Diag3dF0FRadialEnFluxExB)

Native units: J T m^-1 s^-1.

Radial kinetic-energy flux density in psi space from ExB drift, toroidally averaged; density included and convective energy transport retained.

  • plot_label: $Q_\psi^{E\times B}$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Density-weighted velocity integrals; no division by density, so do not multiply by density. Radial kernels use v dot grad(psi), with psi the native poloidal flux/(2*pi) in T m^2. Divide pointwise by |grad psi| from xgc.f0.mesh.bp:gradpsi to obtain flux per physical area. For surface profiles, the XGC-Analysis neoclassical convention is FSA(F_psi)/sqrt(FSA(|grad psi|^2)); use line-integral FSA for magnetic-drift cancellation. n0 and turb are toroidally averaged mean-product and correlated fluctuation contributions; f0 and df distinguish adiabatic and nonadiabatic distribution contributions. XGCa omits n0/turb in names. These channels have no energy-bin or toroidal-plane axis.

  • related_variables: xgc.f0.mesh.bp: gradpsi (R,Z components, T m)

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:77.

xgc.f2d/f3d.*.bp: [species]_radial_en_flux_mag_[n0/turb]_[f0/df] (Diag3dF0FRadialEnFluxMag)

Native units: J T m^-1 s^-1.

Radial kinetic-energy flux density in psi space from magnetic drifts, toroidally averaged; density included and convective energy transport retained. Use line-integral flux-surface averaging.

  • plot_label: $Q_\psi^{\mathrm{mag}}$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Density-weighted velocity integrals; no division by density, so do not multiply by density. Radial kernels use v dot grad(psi), with psi the native poloidal flux/(2*pi) in T m^2. Divide pointwise by |grad psi| from xgc.f0.mesh.bp:gradpsi to obtain flux per physical area. For surface profiles, the XGC-Analysis neoclassical convention is FSA(F_psi)/sqrt(FSA(|grad psi|^2)); use line-integral FSA for magnetic-drift cancellation. n0 and turb are toroidally averaged mean-product and correlated fluctuation contributions; f0 and df distinguish adiabatic and nonadiabatic distribution contributions. XGCa omits n0/turb in names. These channels have no energy-bin or toroidal-plane axis.

  • related_variables: xgc.f0.mesh.bp: gradpsi (R,Z components, T m)

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:83.

xgc.f2d/f3d.*.bp: [species]_radial_flux_3db_[n0/turb]_[f0/df] (Diag3dF0FRadialFlux3db)

Native units: T m^-1 s^-1.

Radial particle flux density in psi space from parallel motion across flux surfaces in the 3D magnetic field, toroidally averaged; not divided by density.

  • plot_label: $\Gamma_\psi^{\delta B}$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Density-weighted velocity integrals; no division by density, so do not multiply by density. Radial kernels use v dot grad(psi), with psi the native poloidal flux/(2*pi) in T m^2. Divide pointwise by |grad psi| from xgc.f0.mesh.bp:gradpsi to obtain flux per physical area. For surface profiles, the XGC-Analysis neoclassical convention is FSA(F_psi)/sqrt(FSA(|grad psi|^2)); use line-integral FSA for magnetic-drift cancellation. n0 and turb are toroidally averaged mean-product and correlated fluctuation contributions; f0 and df distinguish adiabatic and nonadiabatic distribution contributions. XGCa omits n0/turb in names. These channels have no energy-bin or toroidal-plane axis.

  • related_variables: xgc.f0.mesh.bp: gradpsi (R,Z components, T m)

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:47.

xgc.f2d/f3d.*.bp: [species]_radial_flux_ExB_[n0/turb]_[f0/df] (Diag3dF0FRadialFluxExB)

Native units: T m^-1 s^-1.

Radial particle flux density in psi space from ExB drift: velocity integral of f v_ExB dot grad(psi), toroidally averaged; not divided by density.

  • plot_label: $\Gamma_\psi^{E\times B}$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Density-weighted velocity integrals; no division by density, so do not multiply by density. Radial kernels use v dot grad(psi), with psi the native poloidal flux/(2*pi) in T m^2. Divide pointwise by |grad psi| from xgc.f0.mesh.bp:gradpsi to obtain flux per physical area. For surface profiles, the XGC-Analysis neoclassical convention is FSA(F_psi)/sqrt(FSA(|grad psi|^2)); use line-integral FSA for magnetic-drift cancellation. n0 and turb are toroidally averaged mean-product and correlated fluctuation contributions; f0 and df distinguish adiabatic and nonadiabatic distribution contributions. XGCa omits n0/turb in names. These channels have no energy-bin or toroidal-plane axis.

  • related_variables: xgc.f0.mesh.bp: gradpsi (R,Z components, T m)

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:35.

xgc.f2d/f3d.*.bp: [species]_radial_flux_mag_[n0/turb]_[f0/df] (Diag3dF0FRadialFluxMag)

Native units: T m^-1 s^-1.

Radial particle flux density in psi space from magnetic drifts, toroidally averaged; not divided by density. Use line-integral flux-surface averaging to avoid magnetic-drift cancellation errors.

  • plot_label: $\Gamma_\psi^{\mathrm{mag}}$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Density-weighted velocity integrals; no division by density, so do not multiply by density. Radial kernels use v dot grad(psi), with psi the native poloidal flux/(2*pi) in T m^2. Divide pointwise by |grad psi| from xgc.f0.mesh.bp:gradpsi to obtain flux per physical area. For surface profiles, the XGC-Analysis neoclassical convention is FSA(F_psi)/sqrt(FSA(|grad psi|^2)); use line-integral FSA for magnetic-drift cancellation. n0 and turb are toroidally averaged mean-product and correlated fluctuation contributions; f0 and df distinguish adiabatic and nonadiabatic distribution contributions. XGCa omits n0/turb in names. These channels have no energy-bin or toroidal-plane axis.

  • related_variables: xgc.f0.mesh.bp: gradpsi (R,Z components, T m)

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:41.

xgc.f2d/f3d.*.bp: [species]_radial_pot_en_flux_3db_[n0/turb]_[f0/df] (Diag3dF0FRadialPotEnFlux3db)

Native units: J T m^-1 s^-1.

Radial potential-energy flux density in psi space from parallel motion across flux surfaces in the 3D magnetic field, toroidally averaged; density included.

  • plot_label: $Q_{\mathrm{pot},\psi}^{\delta B}$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Density-weighted velocity integrals; no division by density, so do not multiply by density. Radial kernels use v dot grad(psi), with psi the native poloidal flux/(2*pi) in T m^2. Divide pointwise by |grad psi| from xgc.f0.mesh.bp:gradpsi to obtain flux per physical area. For surface profiles, the XGC-Analysis neoclassical convention is FSA(F_psi)/sqrt(FSA(|grad psi|^2)); use line-integral FSA for magnetic-drift cancellation. n0 and turb are toroidally averaged mean-product and correlated fluctuation contributions; f0 and df distinguish adiabatic and nonadiabatic distribution contributions. XGCa omits n0/turb in names. These channels have no energy-bin or toroidal-plane axis.

  • related_variables: xgc.f0.mesh.bp: gradpsi (R,Z components, T m)

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:121.

xgc.f2d/f3d.*.bp: [species]_radial_pot_en_flux_ExB0_[n0/turb]_[f0/df] (Diag3dF0FRadialPotEnFluxExB0)

Native units: J T m^-1 s^-1.

Radial potential-energy flux density in psi space, XGC1 ExB0 channel using the axisymmetric ExB velocity; toroidally averaged, density included.

  • plot_label: $Q_{\mathrm{pot},\psi}^{E\times B,0}$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Density-weighted velocity integrals; no division by density, so do not multiply by density. Radial kernels use v dot grad(psi), with psi the native poloidal flux/(2*pi) in T m^2. Divide pointwise by |grad psi| from xgc.f0.mesh.bp:gradpsi to obtain flux per physical area. For surface profiles, the XGC-Analysis neoclassical convention is FSA(F_psi)/sqrt(FSA(|grad psi|^2)); use line-integral FSA for magnetic-drift cancellation. n0 and turb are toroidally averaged mean-product and correlated fluctuation contributions; f0 and df distinguish adiabatic and nonadiabatic distribution contributions. XGCa omits n0/turb in names. These channels have no energy-bin or toroidal-plane axis.

  • related_variables: xgc.f0.mesh.bp: gradpsi (R,Z components, T m)

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:96.

xgc.f2d/f3d.*.bp: [species]_radial_pot_en_flux_ExB_[n0/turb]_[f0/df] (Diag3dF0FRadialPotEnFluxExB)

Native units: J T m^-1 s^-1.

Radial potential-energy flux density in psi space from ExB drift, toroidally averaged; density included.

  • plot_label: $Q_{\mathrm{pot},\psi}^{E\times B}$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Density-weighted velocity integrals; no division by density, so do not multiply by density. Radial kernels use v dot grad(psi), with psi the native poloidal flux/(2*pi) in T m^2. Divide pointwise by |grad psi| from xgc.f0.mesh.bp:gradpsi to obtain flux per physical area. For surface profiles, the XGC-Analysis neoclassical convention is FSA(F_psi)/sqrt(FSA(|grad psi|^2)); use line-integral FSA for magnetic-drift cancellation. n0 and turb are toroidally averaged mean-product and correlated fluctuation contributions; f0 and df distinguish adiabatic and nonadiabatic distribution contributions. XGCa omits n0/turb in names. These channels have no energy-bin or toroidal-plane axis.

  • related_variables: xgc.f0.mesh.bp: gradpsi (R,Z components, T m)

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:109.

xgc.f2d/f3d.*.bp: [species]_radial_pot_en_flux_ExBt_[n0/turb]_[f0/df] (Diag3dF0FRadialPotEnFluxExBt)

Native units: J T m^-1 s^-1.

Radial potential-energy flux density in psi space, XGC1 ExBt channel using the fluctuating ExB velocity; toroidally averaged, density included.

  • plot_label: $Q_{\mathrm{pot},\psi}^{E\times B,\mathrm{t}}$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Density-weighted velocity integrals; no division by density, so do not multiply by density. Radial kernels use v dot grad(psi), with psi the native poloidal flux/(2*pi) in T m^2. Divide pointwise by |grad psi| from xgc.f0.mesh.bp:gradpsi to obtain flux per physical area. For surface profiles, the XGC-Analysis neoclassical convention is FSA(F_psi)/sqrt(FSA(|grad psi|^2)); use line-integral FSA for magnetic-drift cancellation. n0 and turb are toroidally averaged mean-product and correlated fluctuation contributions; f0 and df distinguish adiabatic and nonadiabatic distribution contributions. XGCa omits n0/turb in names. These channels have no energy-bin or toroidal-plane axis.

  • related_variables: xgc.f0.mesh.bp: gradpsi (R,Z components, T m)

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:102.

xgc.f2d/f3d.*.bp: [species]_radial_pot_en_flux_mag_[n0/turb]_[f0/df] (Diag3dF0FRadialPotEnFluxMag)

Native units: J T m^-1 s^-1.

Radial potential-energy flux density in psi space from magnetic drifts, toroidally averaged; density included. Use line-integral flux-surface averaging.

  • plot_label: $Q_{\mathrm{pot},\psi}^{\mathrm{mag}}$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Density-weighted velocity integrals; no division by density, so do not multiply by density. Radial kernels use v dot grad(psi), with psi the native poloidal flux/(2*pi) in T m^2. Divide pointwise by |grad psi| from xgc.f0.mesh.bp:gradpsi to obtain flux per physical area. For surface profiles, the XGC-Analysis neoclassical convention is FSA(F_psi)/sqrt(FSA(|grad psi|^2)); use line-integral FSA for magnetic-drift cancellation. n0 and turb are toroidally averaged mean-product and correlated fluctuation contributions; f0 and df distinguish adiabatic and nonadiabatic distribution contributions. XGCa omits n0/turb in names. These channels have no energy-bin or toroidal-plane axis.

  • related_variables: xgc.f0.mesh.bp: gradpsi (R,Z components, T m)

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:115.

xgc.f2d/f3d.*.bp: [species]_tor_ang_mom_[n0/turb]_[f0/df] (Diag3dF0FTorAngMom)

Native units: kg m^-1 s^-1.

Toroidal angular-momentum density: velocity integral of f m v_parallel R B_phi/|B|, toroidally averaged; not divided by density.

  • plot_label: Toroidal angular-momentum density ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Angular-momentum density, not a per-particle mean: velocity integral of f L_phi, L_phi=m v_parallel R B_phi/|B|. Mass and density are included; no division by density. n0 and turb are toroidally averaged mean-product and correlated fluctuation contributions; f0 and df distinguish adiabatic and nonadiabatic distribution contributions. XGCa omits n0/turb in names. These channels have no energy-bin or toroidal-plane axis.

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:53.

xgc.f2d/f3d.*.bp: [species]_toroidal_flow_[n0/turb]_[f0/df] (Diag3dF0FToroidalFlow)

Native units: m/s.

Toroidally averaged, density-normalized gyrocenter toroidal-flow contribution. Differs from the physical toroidal flow by the diamagnetic flow.

  • plot_label: $u_{\varphi,\mathrm{gc}}$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Only the toroidal, poloidal and parallel flow-velocity channels are divided by the local total gyrocenter density before toroidal averaging. Their f0/df and n0/turb contributions use that same denominator. n0 and turb are toroidally averaged mean-product and correlated fluctuation contributions; f0 and df distinguish adiabatic and nonadiabatic distribution contributions. XGCa omits n0/turb in names. These channels have no energy-bin or toroidal-plane axis.

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:18.

xgc.f2d/f3d.*.bp: base_name + “ExB_enflux_en” (Diag3dF0FBaseNameExBEnfluxEn)

Native units: J T m^-1 s^-1.

Energy-binned radial ExB kinetic-energy flux density in psi space from the nonadiabatic turbulent contribution; density included, summed within each bin, convective energy transport retained.

  • plot_label: $Q_\psi^{E\times B}$ (energy binned) ({name})

  • axes: mesh_node,energy_bin

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Bins are assigned by rounding 8*sqrt((vperp_norm^2+vpar_norm^2)/(f0_smu_max^2+f0_vp_max^2)) to the nearest integer, then clamping to 0…8. They are uniform in normalized speed, not in energy. energy_en stores the velocity integral of f*E/m within a bin, not an energy coordinate. Where den_en is nonzero, a sampled mean energy in eV is mass_kg*energy_en/(den_en*EV_2_J). Preserve the bin axis; empty bins have no sampled mean energy.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Density-weighted velocity integrals; no division by density, so do not multiply by density. Radial kernels use v dot grad(psi), with psi the native poloidal flux/(2*pi) in T m^2. Divide pointwise by |grad psi| from xgc.f0.mesh.bp:gradpsi to obtain flux per physical area. For surface profiles, the XGC-Analysis neoclassical convention is FSA(F_psi)/sqrt(FSA(|grad psi|^2)); use line-integral FSA for magnetic-drift cancellation. Each energy bin stores an integral over its bin, not a density per unit energy. The kernel is the nonadiabatic turbulent ExB particle/kinetic-energy flux from f_nonadia_turb times v_ExB,turb dot grad(psi). Unlike the channel-family moments, the energy-binned array is not toroidally averaged before writing: the writer outputs toroidal plane rank 0’s local values rather than a toroidal average.

  • related_variables: xgc.f0.mesh.bp: gradpsi (R,Z components, T m)

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:955.

xgc.f2d/f3d.*.bp: base_name + “ExB_flux_en” (Diag3dF0FBaseNameExBFluxEn)

Native units: T m^-1 s^-1.

Energy-binned radial ExB particle flux density in psi space from the nonadiabatic turbulent contribution; density included, summed within each bin.

  • plot_label: $\Gamma_\psi^{E\times B}$ (energy binned) ({name})

  • axes: mesh_node,energy_bin

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Bins are assigned by rounding 8*sqrt((vperp_norm^2+vpar_norm^2)/(f0_smu_max^2+f0_vp_max^2)) to the nearest integer, then clamping to 0…8. They are uniform in normalized speed, not in energy. energy_en stores the velocity integral of f*E/m within a bin, not an energy coordinate. Where den_en is nonzero, a sampled mean energy in eV is mass_kg*energy_en/(den_en*EV_2_J). Preserve the bin axis; empty bins have no sampled mean energy.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Density-weighted velocity integrals; no division by density, so do not multiply by density. Radial kernels use v dot grad(psi), with psi the native poloidal flux/(2*pi) in T m^2. Divide pointwise by |grad psi| from xgc.f0.mesh.bp:gradpsi to obtain flux per physical area. For surface profiles, the XGC-Analysis neoclassical convention is FSA(F_psi)/sqrt(FSA(|grad psi|^2)); use line-integral FSA for magnetic-drift cancellation. Each energy bin stores an integral over its bin, not a density per unit energy. The kernel is the nonadiabatic turbulent ExB particle/kinetic-energy flux from f_nonadia_turb times v_ExB,turb dot grad(psi). Unlike the channel-family moments, the energy-binned array is not toroidally averaged before writing: the writer outputs toroidal plane rank 0’s local values rather than a toroidal average.

  • related_variables: xgc.f0.mesh.bp: gradpsi (R,Z components, T m)

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:948.

xgc.f2d/f3d.*.bp: base_name + “_f0_delta_T” (Diag3dF0FBaseNameF0DeltaT)

Native units: eV.

f0 perturbation: delta temperature

  • plot_label: $\delta T_0$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:983.

xgc.f2d/f3d.*.bp: base_name + “_f0_delta_n” (Diag3dF0FBaseNameF0DeltaN)

Native units: m^-3.

f0 perturbation: delta density

  • plot_label: $\delta n_0$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:971.

xgc.f2d/f3d.*.bp: base_name + “_f0_delta_u” (Diag3dF0FBaseNameF0DeltaU)

Native units: m/s.

f0 perturbation: delta flow

  • plot_label: $\delta u_0$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:977.

xgc.f2d/f3d.*.bp: base_name + “den_en” (Diag3dF0FBaseNameDenEn)

Native units: m^-3.

Full-distribution density integrated within each normalized-speed bin, written from toroidal plane rank 0 without toroidal averaging.

  • plot_label: $n$ (energy binned) ({name})

  • axes: mesh_node,energy_bin

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Bins are assigned by rounding 8*sqrt((vperp_norm^2+vpar_norm^2)/(f0_smu_max^2+f0_vp_max^2)) to the nearest integer, then clamping to 0…8. They are uniform in normalized speed, not in energy. energy_en stores the velocity integral of f*E/m within a bin, not an energy coordinate. Where den_en is nonzero, a sampled mean energy in eV is mass_kg*energy_en/(den_en*EV_2_J). Preserve the bin axis; empty bins have no sampled mean energy.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Integral of the full distribution within each bin, including velocity quadrature. No division by bin width or density is applied. This array uses plane rank 0 values and is not toroidally averaged.

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:932.

xgc.f2d/f3d.*.bp: base_name + “energy_en” (Diag3dF0FBaseNameEnergyEn)

Native units: m^-1 s^-2.

Energy-binned kinetic-energy density per species mass: velocity integral of f E/m within each bin; not divided by bin density.

  • plot_label: $K$ (energy binned) ({name})

  • axes: mesh_node,energy_bin

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Bins are assigned by rounding 8*sqrt((vperp_norm^2+vpar_norm^2)/(f0_smu_max^2+f0_vp_max^2)) to the nearest integer, then clamping to 0…8. They are uniform in normalized speed, not in energy. energy_en stores the velocity integral of f*E/m within a bin, not an energy coordinate. Where den_en is nonzero, a sampled mean energy in eV is mass_kg*energy_en/(den_en*EV_2_J). Preserve the bin axis; empty bins have no sampled mean energy.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Stored sum is f_total*d^3v*en*vth^2, with en=(vperp_norm^2+vpar_norm^2)/2 and vth^2=EV_2_J*f0_fg_T_ev/mass_kg. Thus it is kinetic-energy density divided by species mass, integrated within each bin. Multiply by species mass in kg (xgc.units.bp:<species>_ptl_mass_au times XGC PROTON_MASS=1.6720e-27 kg) to obtain J m^-3; divide that by the matching den_en bin density to obtain J per particle, and by XGC EV_2_J=1.6022e-19 J/eV for eV per particle. No division by bin width or density has been applied. The writer outputs toroidal plane rank 0 values without toroidal averaging.

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:941.

xgc.f2d/f3d.*.bp: dpot (Diag3dF0FDpot)

Native units: V.

Perturbed potential at midplane

  • plot_label: $\delta\phi$ (midplane)

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/phi_potential(:)/values

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:998.

xgc.f2d/f3d.*.bp: gstep (Diag3dF0FGstep)

Native units: Non-negative integer.

XGC global simulation time step for this f2d/f3d diagnostic output slice.

  • plot_label: Global time step

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:780.

xgc.f2d/f3d.*.bp: inode1m1 (Diag3dF0FInode1m1)

Native units: Non-negative integer.

Node offset for this rank

  • plot_label: Global mesh-node offset

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:802.

xgc.f2d/f3d.*.bp: iphim1 (Diag3dF0FIphim1)

Native units: Non-negative integer.

Local plane index

  • plot_label: $i_\varphi$

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:795.

xgc.f2d/f3d.*.bp: n_energy (Diag3dF0FNEnergy)

Native units: Positive integer.

Number of energy bins

  • plot_label: $N_{en}$

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:814.

xgc.f2d/f3d.*.bp: ndata (Diag3dF0FNdata)

Native units: Positive integer.

Number of nodes on this rank

  • plot_label: Local mesh-node count

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:809.

xgc.f2d/f3d.*.bp: nnode (Diag3dF0FNnode)

Native units: Positive integer.

Number of mesh nodes per poloidal plane.

  • plot_label: $N_n$

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:768.

xgc.f2d/f3d.*.bp: nphi (Diag3dF0FNphi)

Native units: Positive integer.

Number of planes

  • plot_label: $N_\varphi$

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:790.

xgc.f2d/f3d.*.bp: phi (Diag3dF0FPhi)

Native units: rad.

Toroidal angle of this plane

  • plot_label: $\varphi$

  • axes: toroidal_plane

  • coordinate_context: Toroidal angle in rad of the midplane used by f2d/f3d. Geometry/periodicity: xgc.mesh.bp:rz,nd_connect_list,wedge_angle,delta_phi. No phi_left/phi_right arrays are written in this product.

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:835.

xgc.f2d/f3d.*.bp: pot0 (Diag3dF0FPot0)

Native units: V.

Flux-surface-averaged component of the axisymmetric electrostatic potential, stored on mesh nodes; zero when the flux-surface-averaged electric-field component is disabled.

  • plot_label: $\langle \phi \rangle$

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz, nd_connect_list for axisymmetric geometry; rplane_rz, rplane_nd_connect_list for stellarator right-plane geometry. Connectivity is zero-based. Node-centered field with continuous linear barycentric basis functions on each triangle.

  • coordinate_context: The same electric_field.pot0_h field is written as pot0 in xgc.2d/3d.*.bp. It uses the right-plane nodal geometry, including xgc.mesh.bp:rplane_rz,rplane_nd_connect_list for stellarators. No toroidal_plane axis is stored. This field is flux-surface averaged, rather than the full poloidally varying n=0 potential.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/phi_potential(:)/values

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:1010.

xgc.f2d/f3d.*.bp: radial_flux_fsa_method (Diag3dF0FRadialFluxFsaMethod)

Native units: dimensionless.

Recommended flux-surface average (FSA) method for radial transport fluxes, particularly the neoclassical ones: 0=volume integral, 1=line integral weighted with 1/B_\theta. 0 or the absence of this parameter imply that the FSA of the radial components of the grad-B and curvature drifts have been subtracted from the magnetic drifts used to evaluate the fluxes, and a consistent FSA must be used for the FSA of the fluxes. The line-integral FSA automatically satisfies <v_{D,r}>=0.

  • plot_label: Radial-flux-surface averaging method

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:819.

xgc.f2d/f3d.*.bp: sp_name + “T_para” (Diag3dF0FSpNameTPara)

Native units: eV.

Parallel temperature

  • plot_label: $T_\parallel$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:855.

xgc.f2d/f3d.*.bp: sp_name + “T_para” (Diag3dF0FSpNameTPara2)

Native units: eV.

Parallel temperature

  • plot_label: $T_\parallel$ ({name})

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:879.

xgc.f2d/f3d.*.bp: sp_name + “T_perp” (Diag3dF0FSpNameTPerp)

Native units: eV.

Perpendicular temperature

  • plot_label: $T_\perp$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:850.

xgc.f2d/f3d.*.bp: sp_name + “T_perp” (Diag3dF0FSpNameTPerp2)

Native units: eV.

Perpendicular temperature

  • plot_label: $T_\perp$ ({name})

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:873.

xgc.f2d/f3d.*.bp: sp_name + “den” (Diag3dF0FSpNameDen)

Native units: m^-3.

Density at each mesh node

  • plot_label: $n$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/electrons/density(:)/values; plasma_profiles/ggd(:)/ion(:)/density(:)/values

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:865.

xgc.f2d/f3d.*.bp: sp_name + “den” (Diag3dF0FSpNameDen2)

Native units: m^-3.

Density at each mesh node

  • plot_label: $n$ ({name})

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/electrons/density(:)/values; plasma_profiles/ggd(:)/ion(:)/density(:)/values

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:891.

xgc.f2d/f3d.*.bp: sp_name + “u_para” (Diag3dF0FSpNameUPara)

Native units: m/s.

Parallel bulk flow

  • plot_label: $u_\parallel$ ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/electrons/velocity(:)/parallel; plasma_profiles/ggd(:)/ion(:)/velocity(:)/parallel

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:860.

xgc.f2d/f3d.*.bp: sp_name + “u_para” (Diag3dF0FSpNameUPara2)

Native units: m/s.

Parallel bulk flow

  • plot_label: $u_\parallel$ ({name})

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Midplane geometry and the same-product phi coordinate (rad). Use xgc.mesh.bp: wedge_angle and delta_phi for the periodic wedge. Variables with _n0 are axisymmetric; _turb are toroidally averaged turbulent correlations, both still spatial node arrays. They do not carry a toroidal_plane axis. Toroidal interpolation of plane-resolved arrays needs magnetic-field-line projection rather than equal-index interpolation.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/electrons/velocity(:)/parallel; plasma_profiles/ggd(:)/ion(:)/velocity(:)/parallel

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:885.

xgc.f2d/f3d.*.bp: time (Diag3dF0FTime)

Native units: s.

Simulation time of step in seconds

  • plot_label: $t$

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions. Clock convention: in the regular main loop gstep is incremented before diagnostic calls, but sml.time is incremented afterward. A checkpoint written later at the same gstep therefore records time larger by one physical timestep. Use the written time, global step and product/integration stage together; equal gstep alone does not establish simultaneous diagnostic and checkpoint states.

  • imas_candidates: plasma_profiles/time

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_high

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:775.

xgc.f2d/f3d.*.bp: tindex (Diag3dF0FTindex)

Native units: Positive integer.

Index of this f2d/f3d diagnostic output time slice. Values: 1, 2, …, # outputs.

  • plot_label: Output slice index

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/diagnostics/diag_3d_f0_f.cpp:785.

xgc.fsourcediag.*.bp: fullname (DiagF0DfFullname)

Native units: [m^3, 1/m^3, dimensionless, J, kg m/s].

Per-node f0_df diagnostic [species]_[channel]_[source]; channel is volume, density, density_change, energy_change, or momentum_change for the named source. Change channels are normalized by the source density/weight sum, so the actual density, momentum, and energy changes are obtained by multiplying by the corresponding density variable. The source [total] is a special case whose meaning depends on f0_df0g_transfer_algorithm. For PersistentRetainedDf0g, [total] is the conservation error from dropping distribution-function components that could not be interpolated to particles. For LegacyAddToF0gH, [total] is the remainder of the mesh-to-particle transfer and is added to the background distribution function; in that case, the sum of [total] and [total2] should equal the sum of the other sources.

  • plot_label: Source diagnostic ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One source-diagnostic slice per ADIOS step or numbered BP file. time is the written physical simulation time in s and gstep the XGC global integration-step index. tindex is the one-based output-slice index, gstep/lcm(sml_f_source_period,diag_f0_df_period), continued across restarts with unchanged periods. Align using written time/gstep. The slice reports operations of the current source call, not an accumulation since the previous diagnostic output.

  • centering: node

  • normalization: Source change channels are divided by the source density/weight sum. Multiply each change by its matching [species]_density_[source] before comparing absolute density, energy or parallel-momentum changes or summing sources. For LegacyAddToF0gH, compare absolute total plus total2 with the other sources. PersistentRetainedDf0g total is newly discarded retention-expiry material; retained changes can originate in earlier calls. Outputs describe the current source call and are not automatically rates per second or interval-integrated changes since the previous output.

Source: XGC_core/cpp/diagnostics/diag_f0_df.cpp:154.

xgc.fsourcediag.*.bp: fullname (DiagF0DfFullname2)

Native units: [dimensionless, dimensionless, dimensionless, J, kg m/s].

Per-node particle source diagnostic [species]_[channel]_total2. density_change, energy_change and momentum_change are the net deposited particle changes across the current source call, divided by the corresponding [species]_density_total mesh weight sum. The volume and density total2 channels are not accumulated and remain zero. Multiply change channels by [species]_density_total to obtain changes per physical volume; divide by the applicable source-update duration for rates.

  • plot_label: Particle-based source diagnostic ({name})

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One source-diagnostic slice per ADIOS step or numbered BP file. time is the written physical simulation time in s and gstep the XGC global integration-step index. tindex is the one-based output-slice index, gstep/lcm(sml_f_source_period,diag_f0_df_period), continued across restarts with unchanged periods. Align using written time/gstep. The slice reports operations of the current source call, not an accumulation since the previous diagnostic output.

  • centering: node

  • normalization: Change channels are normalized by the mesh weight sum underlying [species]_density_total, not by [species]_density_total2 (which is zero). Multiply by [species]_density_total for absolute changes per volume. volume_total2 and density_total2 are zero placeholders. These changes span the current source call; they are not automatically rates or changes accumulated since the previous diagnostic output.

Source: XGC_core/cpp/diagnostics/diag_f0_df.cpp:193.

xgc.fsourcediag.*.bp: gstep (DiagF0DfGstep)

Native units: Non-negative integer.

XGC global simulation time step for this f-source diagnostic output slice.

  • plot_label: Global time step

  • time_context: One source-diagnostic slice per ADIOS step or numbered BP file. time is the written physical simulation time in s and gstep the XGC global integration-step index. tindex is the one-based output-slice index, gstep/lcm(sml_f_source_period,diag_f0_df_period), continued across restarts with unchanged periods. Align using written time/gstep. The slice reports operations of the current source call, not an accumulation since the previous diagnostic output.

Source: XGC_core/cpp/diagnostics/diag_f0_df.cpp:88.

xgc.fsourcediag.*.bp: psi (DiagF0DfPsi)

Native units: dimensionless.

Normalized poloidal magnetic flux at each mesh node: grid.psi/magnetic_field.psi_norm().

  • plot_label: $\psi_N$

  • axes: mesh_node

  • normalization: Dimensionless poloidal flux normalized by magnetic_field.psi_norm(). The matching psi_mks array stores the dimensional poloidal flux divided by 2*pi.

Source: XGC_core/cpp/diagnostics/diag_f0_df.cpp:105.

xgc.fsourcediag.*.bp: psi_mks (DiagF0DfPsiMks)

Native units: Wb/rad.

Poloidal magnetic flux \(\\psi/(2\\pi)\) in MKS units.

  • plot_label: $\psi/(2\pi)$

  • axes: mesh_node

Source: XGC_core/cpp/diagnostics/diag_f0_df.cpp:100.

xgc.fsourcediag.*.bp: samples (DiagF0DfSamples)

Native units: Positive integer.

Output size (number of mesh vertices).

  • plot_label: $N_n$

  • time_context: One source-diagnostic slice per ADIOS step or numbered BP file. time is the written physical simulation time in s and gstep the XGC global integration-step index. tindex is the one-based output-slice index, gstep/lcm(sml_f_source_period,diag_f0_df_period), continued across restarts with unchanged periods. Align using written time/gstep. The slice reports operations of the current source call, not an accumulation since the previous diagnostic output.

Source: XGC_core/cpp/diagnostics/diag_f0_df.cpp:76.

xgc.fsourcediag.*.bp: time (DiagF0DfTime)

Native units: s.

Simulation time of step in seconds.

  • plot_label: $t$

  • time_context: One source-diagnostic slice per ADIOS step or numbered BP file. time is the written physical simulation time in s and gstep the XGC global integration-step index. tindex is the one-based output-slice index, gstep/lcm(sml_f_source_period,diag_f0_df_period), continued across restarts with unchanged periods. Align using written time/gstep. The slice reports operations of the current source call, not an accumulation since the previous diagnostic output. Clock convention: in the regular main loop gstep is incremented before diagnostic calls, but sml.time is incremented afterward. A checkpoint written later at the same gstep therefore records time larger by one physical timestep. Use the written time, global step and product/integration stage together; equal gstep alone does not establish simultaneous diagnostic and checkpoint states.

  • imas_candidates: plasma_profiles/time

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_high

Source: XGC_core/cpp/diagnostics/diag_f0_df.cpp:83.

xgc.fsourcediag.*.bp: tindex (DiagF0DfTindex)

Native units: Positive integer.

Index of this f-source diagnostic output time slice. Values: 1, 2, …, # outputs.

  • plot_label: Output slice index

  • time_context: One source-diagnostic slice per ADIOS step or numbered BP file. time is the written physical simulation time in s and gstep the XGC global integration-step index. tindex is the one-based output-slice index, gstep/lcm(sml_f_source_period,diag_f0_df_period), continued across restarts with unchanged periods. Align using written time/gstep. The slice reports operations of the current source call, not an accumulation since the previous diagnostic output.

Source: XGC_core/cpp/diagnostics/diag_f0_df.cpp:93.

xgc.grad_rz.bp: basis (GradientMatBasis)

Native units: dimensionless.

Per-node poloidal derivative basis: 0 selects unit grad(psi) and its counter-clockwise perpendicular; 1 selects cylindrical R and Z.

  • plot_label: Poloidal derivative basis code

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz and nd_connect_list for axisymmetric geometry; rplane_rz[0,:,:] and rplane_nd_connect_list[0,:,:] for the rank-zero right plane of a non-axisymmetric/stellarator mesh. Coordinates are (R,Z) in m and triangle connectivity is zero-based. This file preserves the historical single rank-zero plane, not an operator for every stellarator plane. Matrix rows are output mesh nodes and columns are input mesh nodes in that geometry; do not interpolate matrix storage slots as a spatial field. The resulting derivative is nodal and can be visualized with linear barycentric interpolation on the matching triangles.

  • coordinate_context: Same-file basis[i] selects the poloidal output frame per node: 1 means (R,Z); 0 means (e_psi,e_theta), with e_psi=(psi_R,psi_Z)/sqrt(psi_R^2+psi_Z^2) and e_theta=(-psi_Z,psi_R)/sqrt(psi_R^2+psi_Z^2), counter-clockwise in the R/Z plane. These are derivatives per physical metre along unit directions, not derivatives with respect to flux or dimensionless angle. Given the equilibrium gradient at that node, rotate D_r,D_z to cylindrical components as D_R=e_psi_R*D_r+e_theta_R*D_z, D_Z=e_psi_Z*D_r+e_theta_Z*D_z for basis=0. The file does not store the unit vectors themselves. For basis=1, D_R=D_r and D_Z=D_z.

  • centering: node

  • normalization: For suffix s=r or z and a scalar nodal vector f, D_s[i]=sum(value_s[i,j]*f[eindex_s[i,j]-1], j=0..nelement_s[i]-1). In Python/C readers arrays are [mesh_node,matrix_slot]; ignore unused padding (including negative index sentinels). These are positive gradient operators; E=-grad(phi) requires an additional minus sign. Output units are units(f)/m. Coefficients include XGC discretization/boundary/fallback choices; do not substitute generic triangle derivatives if reproducing XGC results.

  • related_variables: xgc.grad_rz.bp: n_r,m_r,w_r,nelement_r,eindex_r,value_r,n_z,m_z,w_z,nelement_z,eindex_z,value_z,basis; xgc.mesh.bp: rz,nd_connect_list,rplane_rz,rplane_nd_connect_list

Source: XGC_core/cpp/grid/write_gradient_mat.cpp:97.

xgc.grad_rz.bp: eindex_r (GradientMatEindexR)

Native units: Integer index.

One-based input mesh-node column indices for the first poloidal directional-derivative matrix; only nelement_r entries per row are valid.

  • plot_label: First-direction gradient column index

  • axes: mesh_node,matrix_slot

  • mesh_context: xgc.mesh.bp: rz and nd_connect_list for axisymmetric geometry; rplane_rz[0,:,:] and rplane_nd_connect_list[0,:,:] for the rank-zero right plane of a non-axisymmetric/stellarator mesh. Coordinates are (R,Z) in m and triangle connectivity is zero-based. This file preserves the historical single rank-zero plane, not an operator for every stellarator plane. Matrix rows are output mesh nodes and columns are input mesh nodes in that geometry; do not interpolate matrix storage slots as a spatial field. The resulting derivative is nodal and can be visualized with linear barycentric interpolation on the matching triangles.

  • coordinate_context: Same-file basis[i] selects the poloidal output frame per node: 1 means (R,Z); 0 means (e_psi,e_theta), with e_psi=(psi_R,psi_Z)/sqrt(psi_R^2+psi_Z^2) and e_theta=(-psi_Z,psi_R)/sqrt(psi_R^2+psi_Z^2), counter-clockwise in the R/Z plane. These are derivatives per physical metre along unit directions, not derivatives with respect to flux or dimensionless angle. Given the equilibrium gradient at that node, rotate D_r,D_z to cylindrical components as D_R=e_psi_R*D_r+e_theta_R*D_z, D_Z=e_psi_Z*D_r+e_theta_Z*D_z for basis=0. The file does not store the unit vectors themselves. For basis=1, D_R=D_r and D_Z=D_z.

  • normalization: For suffix s=r or z and a scalar nodal vector f, D_s[i]=sum(value_s[i,j]*f[eindex_s[i,j]-1], j=0..nelement_s[i]-1). In Python/C readers arrays are [mesh_node,matrix_slot]; ignore unused padding (including negative index sentinels). These are positive gradient operators; E=-grad(phi) requires an additional minus sign. Output units are units(f)/m. Coefficients include XGC discretization/boundary/fallback choices; do not substitute generic triangle derivatives if reproducing XGC results.

  • related_variables: xgc.grad_rz.bp: n_r,m_r,w_r,nelement_r,eindex_r,value_r,n_z,m_z,w_z,nelement_z,eindex_z,value_z,basis; xgc.mesh.bp: rz,nd_connect_list,rplane_rz,rplane_nd_connect_list

Source: XGC_core/cpp/grid/write_gradient_mat.cpp:52.

xgc.grad_rz.bp: eindex_z (GradientMatEindexZ)

Native units: Integer index.

One-based input mesh-node column indices for the second poloidal directional-derivative matrix; only nelement_z entries per row are valid.

  • plot_label: Second-direction gradient column index

  • axes: mesh_node,matrix_slot

  • mesh_context: xgc.mesh.bp: rz and nd_connect_list for axisymmetric geometry; rplane_rz[0,:,:] and rplane_nd_connect_list[0,:,:] for the rank-zero right plane of a non-axisymmetric/stellarator mesh. Coordinates are (R,Z) in m and triangle connectivity is zero-based. This file preserves the historical single rank-zero plane, not an operator for every stellarator plane. Matrix rows are output mesh nodes and columns are input mesh nodes in that geometry; do not interpolate matrix storage slots as a spatial field. The resulting derivative is nodal and can be visualized with linear barycentric interpolation on the matching triangles.

  • coordinate_context: Same-file basis[i] selects the poloidal output frame per node: 1 means (R,Z); 0 means (e_psi,e_theta), with e_psi=(psi_R,psi_Z)/sqrt(psi_R^2+psi_Z^2) and e_theta=(-psi_Z,psi_R)/sqrt(psi_R^2+psi_Z^2), counter-clockwise in the R/Z plane. These are derivatives per physical metre along unit directions, not derivatives with respect to flux or dimensionless angle. Given the equilibrium gradient at that node, rotate D_r,D_z to cylindrical components as D_R=e_psi_R*D_r+e_theta_R*D_z, D_Z=e_psi_Z*D_r+e_theta_Z*D_z for basis=0. The file does not store the unit vectors themselves. For basis=1, D_R=D_r and D_Z=D_z.

  • normalization: For suffix s=r or z and a scalar nodal vector f, D_s[i]=sum(value_s[i,j]*f[eindex_s[i,j]-1], j=0..nelement_s[i]-1). In Python/C readers arrays are [mesh_node,matrix_slot]; ignore unused padding (including negative index sentinels). These are positive gradient operators; E=-grad(phi) requires an additional minus sign. Output units are units(f)/m. Coefficients include XGC discretization/boundary/fallback choices; do not substitute generic triangle derivatives if reproducing XGC results.

  • related_variables: xgc.grad_rz.bp: n_r,m_r,w_r,nelement_r,eindex_r,value_r,n_z,m_z,w_z,nelement_z,eindex_z,value_z,basis; xgc.mesh.bp: rz,nd_connect_list,rplane_rz,rplane_nd_connect_list

Source: XGC_core/cpp/grid/write_gradient_mat.cpp:85.

xgc.grad_rz.bp: m_r (GradientMatMR)

Native units: Positive integer.

Number of output mesh-node rows in the first poloidal directional-derivative matrix.

  • plot_label: Gradient output-node count

Source: XGC_core/cpp/grid/write_gradient_mat.cpp:36.

xgc.grad_rz.bp: m_z (GradientMatMZ)

Native units: Positive integer.

Number of output mesh-node rows in the second poloidal directional-derivative matrix.

  • plot_label: Gradient output-node count

Source: XGC_core/cpp/grid/write_gradient_mat.cpp:69.

xgc.grad_rz.bp: n_r (GradientMatNR)

Native units: Positive integer.

Number of input mesh-node columns in the first poloidal directional-derivative matrix.

  • plot_label: Gradient input-node count

Source: XGC_core/cpp/grid/write_gradient_mat.cpp:31.

xgc.grad_rz.bp: n_z (GradientMatNZ)

Native units: Positive integer.

Number of input mesh-node columns in the second poloidal directional-derivative matrix.

  • plot_label: Gradient input-node count

Source: XGC_core/cpp/grid/write_gradient_mat.cpp:64.

xgc.grad_rz.bp: nelement_r (GradientMatNelementR)

Native units: Non-negative integer.

Valid entry count for each row of the first poloidal directional-derivative matrix; ignore slots at or beyond this count.

  • plot_label: First-direction gradient row entry count

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz and nd_connect_list for axisymmetric geometry; rplane_rz[0,:,:] and rplane_nd_connect_list[0,:,:] for the rank-zero right plane of a non-axisymmetric/stellarator mesh. Coordinates are (R,Z) in m and triangle connectivity is zero-based. This file preserves the historical single rank-zero plane, not an operator for every stellarator plane. Matrix rows are output mesh nodes and columns are input mesh nodes in that geometry; do not interpolate matrix storage slots as a spatial field. The resulting derivative is nodal and can be visualized with linear barycentric interpolation on the matching triangles.

  • coordinate_context: Same-file basis[i] selects the poloidal output frame per node: 1 means (R,Z); 0 means (e_psi,e_theta), with e_psi=(psi_R,psi_Z)/sqrt(psi_R^2+psi_Z^2) and e_theta=(-psi_Z,psi_R)/sqrt(psi_R^2+psi_Z^2), counter-clockwise in the R/Z plane. These are derivatives per physical metre along unit directions, not derivatives with respect to flux or dimensionless angle. Given the equilibrium gradient at that node, rotate D_r,D_z to cylindrical components as D_R=e_psi_R*D_r+e_theta_R*D_z, D_Z=e_psi_Z*D_r+e_theta_Z*D_z for basis=0. The file does not store the unit vectors themselves. For basis=1, D_R=D_r and D_Z=D_z.

  • normalization: For suffix s=r or z and a scalar nodal vector f, D_s[i]=sum(value_s[i,j]*f[eindex_s[i,j]-1], j=0..nelement_s[i]-1). In Python/C readers arrays are [mesh_node,matrix_slot]; ignore unused padding (including negative index sentinels). These are positive gradient operators; E=-grad(phi) requires an additional minus sign. Output units are units(f)/m. Coefficients include XGC discretization/boundary/fallback choices; do not substitute generic triangle derivatives if reproducing XGC results.

  • related_variables: xgc.grad_rz.bp: n_r,m_r,w_r,nelement_r,eindex_r,value_r,n_z,m_z,w_z,nelement_z,eindex_z,value_z,basis; xgc.mesh.bp: rz,nd_connect_list,rplane_rz,rplane_nd_connect_list

Source: XGC_core/cpp/grid/write_gradient_mat.cpp:46.

xgc.grad_rz.bp: nelement_z (GradientMatNelementZ)

Native units: Non-negative integer.

Valid entry count for each row of the second poloidal directional-derivative matrix; ignore slots at or beyond this count.

  • plot_label: Second-direction gradient row entry count

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz and nd_connect_list for axisymmetric geometry; rplane_rz[0,:,:] and rplane_nd_connect_list[0,:,:] for the rank-zero right plane of a non-axisymmetric/stellarator mesh. Coordinates are (R,Z) in m and triangle connectivity is zero-based. This file preserves the historical single rank-zero plane, not an operator for every stellarator plane. Matrix rows are output mesh nodes and columns are input mesh nodes in that geometry; do not interpolate matrix storage slots as a spatial field. The resulting derivative is nodal and can be visualized with linear barycentric interpolation on the matching triangles.

  • coordinate_context: Same-file basis[i] selects the poloidal output frame per node: 1 means (R,Z); 0 means (e_psi,e_theta), with e_psi=(psi_R,psi_Z)/sqrt(psi_R^2+psi_Z^2) and e_theta=(-psi_Z,psi_R)/sqrt(psi_R^2+psi_Z^2), counter-clockwise in the R/Z plane. These are derivatives per physical metre along unit directions, not derivatives with respect to flux or dimensionless angle. Given the equilibrium gradient at that node, rotate D_r,D_z to cylindrical components as D_R=e_psi_R*D_r+e_theta_R*D_z, D_Z=e_psi_Z*D_r+e_theta_Z*D_z for basis=0. The file does not store the unit vectors themselves. For basis=1, D_R=D_r and D_Z=D_z.

  • normalization: For suffix s=r or z and a scalar nodal vector f, D_s[i]=sum(value_s[i,j]*f[eindex_s[i,j]-1], j=0..nelement_s[i]-1). In Python/C readers arrays are [mesh_node,matrix_slot]; ignore unused padding (including negative index sentinels). These are positive gradient operators; E=-grad(phi) requires an additional minus sign. Output units are units(f)/m. Coefficients include XGC discretization/boundary/fallback choices; do not substitute generic triangle derivatives if reproducing XGC results.

  • related_variables: xgc.grad_rz.bp: n_r,m_r,w_r,nelement_r,eindex_r,value_r,n_z,m_z,w_z,nelement_z,eindex_z,value_z,basis; xgc.mesh.bp: rz,nd_connect_list,rplane_rz,rplane_nd_connect_list

Source: XGC_core/cpp/grid/write_gradient_mat.cpp:79.

xgc.grad_rz.bp: value_r (GradientMatValueR)

Native units: m^-1.

First poloidal directional-derivative coefficients: R derivative where basis=1, derivative along the unit grad(psi) direction where basis=0.

  • plot_label: First-direction gradient coefficient

  • axes: mesh_node,matrix_slot

  • mesh_context: xgc.mesh.bp: rz and nd_connect_list for axisymmetric geometry; rplane_rz[0,:,:] and rplane_nd_connect_list[0,:,:] for the rank-zero right plane of a non-axisymmetric/stellarator mesh. Coordinates are (R,Z) in m and triangle connectivity is zero-based. This file preserves the historical single rank-zero plane, not an operator for every stellarator plane. Matrix rows are output mesh nodes and columns are input mesh nodes in that geometry; do not interpolate matrix storage slots as a spatial field. The resulting derivative is nodal and can be visualized with linear barycentric interpolation on the matching triangles.

  • coordinate_context: Same-file basis[i] selects the poloidal output frame per node: 1 means (R,Z); 0 means (e_psi,e_theta), with e_psi=(psi_R,psi_Z)/sqrt(psi_R^2+psi_Z^2) and e_theta=(-psi_Z,psi_R)/sqrt(psi_R^2+psi_Z^2), counter-clockwise in the R/Z plane. These are derivatives per physical metre along unit directions, not derivatives with respect to flux or dimensionless angle. Given the equilibrium gradient at that node, rotate D_r,D_z to cylindrical components as D_R=e_psi_R*D_r+e_theta_R*D_z, D_Z=e_psi_Z*D_r+e_theta_Z*D_z for basis=0. The file does not store the unit vectors themselves. For basis=1, D_R=D_r and D_Z=D_z.

  • normalization: For suffix s=r or z and a scalar nodal vector f, D_s[i]=sum(value_s[i,j]*f[eindex_s[i,j]-1], j=0..nelement_s[i]-1). In Python/C readers arrays are [mesh_node,matrix_slot]; ignore unused padding (including negative index sentinels). These are positive gradient operators; E=-grad(phi) requires an additional minus sign. Output units are units(f)/m. Coefficients include XGC discretization/boundary/fallback choices; do not substitute generic triangle derivatives if reproducing XGC results.

  • related_variables: xgc.grad_rz.bp: n_r,m_r,w_r,nelement_r,eindex_r,value_r,n_z,m_z,w_z,nelement_z,eindex_z,value_z,basis; xgc.mesh.bp: rz,nd_connect_list,rplane_rz,rplane_nd_connect_list

Source: XGC_core/cpp/grid/write_gradient_mat.cpp:58.

xgc.grad_rz.bp: value_z (GradientMatValueZ)

Native units: m^-1.

Second poloidal directional-derivative coefficients: Z derivative where basis=1, counter-clockwise poloidal unit-direction derivative where basis=0.

  • plot_label: Second-direction gradient coefficient

  • axes: mesh_node,matrix_slot

  • mesh_context: xgc.mesh.bp: rz and nd_connect_list for axisymmetric geometry; rplane_rz[0,:,:] and rplane_nd_connect_list[0,:,:] for the rank-zero right plane of a non-axisymmetric/stellarator mesh. Coordinates are (R,Z) in m and triangle connectivity is zero-based. This file preserves the historical single rank-zero plane, not an operator for every stellarator plane. Matrix rows are output mesh nodes and columns are input mesh nodes in that geometry; do not interpolate matrix storage slots as a spatial field. The resulting derivative is nodal and can be visualized with linear barycentric interpolation on the matching triangles.

  • coordinate_context: Same-file basis[i] selects the poloidal output frame per node: 1 means (R,Z); 0 means (e_psi,e_theta), with e_psi=(psi_R,psi_Z)/sqrt(psi_R^2+psi_Z^2) and e_theta=(-psi_Z,psi_R)/sqrt(psi_R^2+psi_Z^2), counter-clockwise in the R/Z plane. These are derivatives per physical metre along unit directions, not derivatives with respect to flux or dimensionless angle. Given the equilibrium gradient at that node, rotate D_r,D_z to cylindrical components as D_R=e_psi_R*D_r+e_theta_R*D_z, D_Z=e_psi_Z*D_r+e_theta_Z*D_z for basis=0. The file does not store the unit vectors themselves. For basis=1, D_R=D_r and D_Z=D_z.

  • normalization: For suffix s=r or z and a scalar nodal vector f, D_s[i]=sum(value_s[i,j]*f[eindex_s[i,j]-1], j=0..nelement_s[i]-1). In Python/C readers arrays are [mesh_node,matrix_slot]; ignore unused padding (including negative index sentinels). These are positive gradient operators; E=-grad(phi) requires an additional minus sign. Output units are units(f)/m. Coefficients include XGC discretization/boundary/fallback choices; do not substitute generic triangle derivatives if reproducing XGC results.

  • related_variables: xgc.grad_rz.bp: n_r,m_r,w_r,nelement_r,eindex_r,value_r,n_z,m_z,w_z,nelement_z,eindex_z,value_z,basis; xgc.mesh.bp: rz,nd_connect_list,rplane_rz,rplane_nd_connect_list

Source: XGC_core/cpp/grid/write_gradient_mat.cpp:91.

xgc.grad_rz.bp: w_r (GradientMatWR)

Native units: Positive integer.

Maximum stored entries per row of the first poloidal directional-derivative matrix.

  • plot_label: First-direction gradient row capacity

Source: XGC_core/cpp/grid/write_gradient_mat.cpp:41.

xgc.grad_rz.bp: w_z (GradientMatWZ)

Native units: Positive integer.

Maximum stored entries per row of the second poloidal directional-derivative matrix.

  • plot_label: Second-direction gradient row capacity

Source: XGC_core/cpp/grid/write_gradient_mat.cpp:74.

xgc.gyraze.bp: bfield_impact_angle (SheathBfieldImpactAngle)

Native units: rad.

Magnetic grazing angle relative to the local wall surface at wall nodes in xgc.mesh.bp grid_wall_nodes ordering

  • plot_label: $\theta_{\mathrm{B,wall}}$

  • axes: wall_node

  • mesh_context: xgc.mesh.bp: grid_wall_nodes contains zero-based indices into rz (R,Z in m) in the ordered mesh-boundary polygon sequence produced by order_wall. Wall-node values follow this sequence; successive entries and the final-to-first pair define boundary segments for a valid closed wall. The wall-node list is not the triangle connectivity array nd_connect_list.

  • coordinate_context: Wall-node ordering from grid_wall_nodes. Sheath toroidal plane spacing is xgc.mesh.bp wedge_angle / same-product sheath_nphi. Do not assume sheath_nphi equals the field-plane count.

  • centering: wall_node

Source: XGC_core/cpp/sheath/sheath.cpp:94.

xgc.gyraze.bp: grid_nwall (SheathGridNwall)

Native units: Positive integer.

Number of wall nodes in xgc.gyraze.bp, matching grid_wall_nodes ordering in xgc.mesh.bp.

  • plot_label: $N_{\mathrm{wall}}$

Source: XGC_core/cpp/sheath/sheath.cpp:89.

xgc.heatdiag*.bp: ds (HeatDiagnosticsDs)

Native units: m^2.

Surface area of each wall-curve segment swept over one heat-diagnostic toroidal bin of width wedge_angle/nphi.

  • plot_label: $A_{\mathrm{seg}}$

  • axes: wall_segment

  • mesh_context: Same heat diagnostic product: r,z are ordered wall-polygon vertices, psi is the corresponding flux coordinate, ds is collection area. Species data use [toroidal_bin,wall_segment_with_reserved_zero]; bin 0 is reserved and has no physical segment. Associate bins 1..nseg with successive polygon segments, including the closing segment. This polygon may include extra interpolated vertices relative to the ordered grid_wall_nodes mesh boundary. ds[j] and strike_angle[j] refer to segment from vertex j to (j+1)%nseg; unlike species data, these geometry arrays have no unassigned bin.

  • centering: wall_segment

  • normalization: ds[j]=R_mid[j]*segment_length[j]*(xgc.mesh.bp:wedge_angle/nphi). Divide accumulated energy/count in species bin j+1 by this area and elapsed physical time for flux density.

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:555.

xgc.heatdiag*.bp: gstep (HeatDiagnosticsGstep)

Native units: Non-negative integer.

XGC global simulation time step for this heat diagnostic output slice.

  • plot_label: Global time step

  • time_context: Accumulated wall-hit contributions since the previous heat write, reset after output. The scalar time is the physical sml.time value passed at the write; main_loop advances sml.time only later, so it is not an independently established interval-end timestamp. gstep identifies the current integration loop and tindex=gstep/period. Energy/count channels are interval sums; use differences of successive physical time values and the applicable collection area for rates/flux densities. Potential and displacement channels require their own number normalization.

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:314.

xgc.heatdiag*.bp: gstep (HeatDiagnosticsGstep2)

Native units: Non-negative integer.

XGC global simulation time step for this heat diagnostic output slice.

  • plot_label: Global time step

  • time_context: Accumulated wall-hit contributions since the previous heat write, reset after output. The scalar time is the physical sml.time value passed at the write; main_loop advances sml.time only later, so it is not an independently established interval-end timestamp. gstep identifies the current integration loop and tindex=gstep/period. Energy/count channels are interval sums; use differences of successive physical time values and the applicable collection area for rates/flux densities. Potential and displacement channels require their own number normalization.

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:527.

xgc.heatdiag*.bp: nphi (HeatDiagnosticsNphi)

Native units: Positive integer.

Number of toroidal heat-diagnostic bins within the simulated toroidal wedge, configured by diag_heat_nphi independently of the field mesh-plane count.

  • plot_label: Toroidal heat-bin count

  • coordinate_context: Toroidal bin width is xgc.mesh.bp:wedge_angle divided by same-product nphi; these bins need not coincide with field mesh planes.

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:514.

xgc.heatdiag*.bp: nseg (HeatDiagnosticsNseg)

Native units: Positive integer.

Number of wall segments on which the heat load is sampled.

  • plot_label: Wall-segment count

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:501.

xgc.heatdiag*.bp: nseg1 (HeatDiagnosticsNseg1)

Native units: Positive integer.

Number of heat wall-segment entries plus the reserved leading bin (nseg+1). Current deposition places matched and nearest-segment fallback crossings in bins 1..nseg, leaving bin 0 unused.

  • plot_label: Wall-bin count including reserved leading bin

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:506.

xgc.heatdiag*.bp: psamples (HeatDiagnosticsPsamples)

Native units: Positive integer.

Number of sampling points in physical poloidal flux divided by 2*pi (psi), in each rectangular heat-diagnostic section.

  • plot_label: Flux sample count

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:296.

xgc.heatdiag*.bp: psi (HeatDiagnosticsPsi)

Native units: T m^2.

Poloidal magnetic flux \(\\psi/(2\\pi)\) at the heat load sampling points.

  • plot_label: $\psi/(2\pi)$ (sampling grid)

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:337.

xgc.heatdiag*.bp: psi (HeatDiagnosticsPsi2)

Native units: T m^2.

Poloidal magnetic flux \(\\psi/(2\\pi)\).

  • plot_label: $\psi/(2\pi)$

  • axes: wall_segment

  • coordinate_context: psi[j] is evaluated at the midpoint of the polygon segment from (r[j],z[j]) to (r[(j+1)%nseg],z[(j+1)%nseg]); geometry arrays have no unassigned bin, whereas species bin 0 is unassigned.

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:550.

xgc.heatdiag*.bp: r (HeatDiagnosticsR)

Native units: m.

Radial coordinates of the sampling points.

  • plot_label: $R$ (sampling grid)

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:327.

xgc.heatdiag*.bp: r (HeatDiagnosticsR2)

Native units: m.

Radial coordinates of the sampling points.

  • plot_label: $R$

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:540.

xgc.heatdiag*.bp: rsamples (HeatDiagnosticsRsamples)

Native units: Positive integer.

Number of sampling points in R-direction.

  • plot_label: R sample count

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:286.

xgc.heatdiag*.bp: section (HeatDiagnosticsSection)

Native units: Positive integer.

Number of sections in which the divertor heat load is sampled.

  • plot_label: Divertor section count

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:304.

xgc.heatdiag*.bp: sp_name + “_distance” (HeatDiagnosticsSpNameDistance)

Native units: m.

Accumulated sum of each escaping marker RZ displacement sqrt((R_old-R_new)^2+(Z_old-Z_new)^2) times its deposited plasma-particle weight over the heat diagnostic interval. Divide by the same-species _number in nonempty bins for the weighted mean displacement.

  • plot_label: Accumulated weighted RZ escape displacement ({name})

  • axes: section,z_bin,r_bin

  • mesh_context: Same heat diagnostic BP product: r[section,R_bin] and z[section,Z_bin] are sampling coordinates for [section,Z_bin,R_bin] deposition arrays; these are bins on a rectangular RZ sampling grid, not the unstructured field mesh.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Accumulated wall-hit contributions since the previous heat write, reset after output. The scalar time is the physical sml.time value passed at the write; main_loop advances sml.time only later, so it is not an independently established interval-end timestamp. gstep identifies the current integration loop and tindex=gstep/period. Energy/count channels are interval sums; use differences of successive physical time values and the applicable collection area for rates/flux densities. Potential and displacement channels require their own number normalization.

  • centering: bin

  • normalization: Raw sum of RZ displacement*(w*w0) using the same deposition weights as the species number channel. The scalar displacement uses original old/new RZ coordinates; only placement into sampling bins uses the mode-specific field-following procedure.

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:400.

xgc.heatdiag*.bp: sp_name + “_distance_psi” (HeatDiagnosticsSpNameDistancePsi)

Native units: m.

Accumulated sum of each escaping marker RZ displacement sqrt((R_old-R_new)^2+(Z_old-Z_new)^2) times its deposited plasma-particle weight over the heat diagnostic interval. Divide by the same-species _number_psi in nonempty bins for the weighted mean displacement.

  • plot_label: Accumulated weighted RZ escape displacement ({name})

  • axes: section,flux_bin

  • mesh_context: Same heat diagnostic BP product: psi[section,flux_bin] gives the flux-bin sampling coordinate; values are histogram/deposition bins, not triangle vertices.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Accumulated wall-hit contributions since the previous heat write, reset after output. The scalar time is the physical sml.time value passed at the write; main_loop advances sml.time only later, so it is not an independently established interval-end timestamp. gstep identifies the current integration loop and tindex=gstep/period. Energy/count channels are interval sums; use differences of successive physical time values and the applicable collection area for rates/flux densities. Potential and displacement channels require their own number normalization.

  • centering: bin

  • normalization: Raw sum of RZ displacement*(w*w0) using the same deposition weights as the species number channel. The scalar displacement uses original old/new RZ coordinates; only placement into sampling bins uses the mode-specific field-following procedure.

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:442.

xgc.heatdiag*.bp: sp_name + “_number” (HeatDiagnosticsSpNameNumber)

Native units: particle count.

Number of plasma particles (not markers) hitting each sampling point (in RZ-coordinates). Data is accumulated over all time steps between two subsequent heat diagnostic outputs.

  • plot_label: Deposited particle count ({name})

  • axes: section,z_bin,r_bin

  • mesh_context: Same heat diagnostic BP product: r[section,R_bin] and z[section,Z_bin] are sampling coordinates for [section,Z_bin,R_bin] deposition arrays; these are bins on a rectangular RZ sampling grid, not the unstructured field mesh.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Accumulated wall-hit contributions since the previous heat write, reset after output. The scalar time is the physical sml.time value passed at the write; main_loop advances sml.time only later, so it is not an independently established interval-end timestamp. gstep identifies the current integration loop and tindex=gstep/period. Energy/count channels are interval sums; use differences of successive physical time values and the applicable collection area for rates/flux densities. Potential and displacement channels require their own number normalization.

  • centering: bin

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:369.

xgc.heatdiag*.bp: sp_name + “_number” (HeatDiagnosticsSpNameNumber2)

Native units: particle count.

Number of plasma particles (not markers) hitting each wall segment; nseg1 includes a reserved leading bin 0; current deposition assigns unmatched crossings to the nearest segment. Data is accumulated over all time steps between two subsequent heat diagnostic outputs.

  • plot_label: Wall particle count ({name})

  • axes: toroidal_bin,wall_segment

  • mesh_context: Same heat diagnostic product: r,z are ordered wall-polygon vertices, psi is the corresponding flux coordinate, ds is collection area. Species data use [toroidal_bin,wall_segment_with_reserved_zero]; bin 0 is reserved and has no physical segment; current nearest-segment fallback leaves it unused. Associate bins 1..nseg with successive polygon segments, including the closing segment. This polygon may include extra interpolated vertices relative to the ordered grid_wall_nodes mesh boundary. Data bin j+1 corresponds to geometry ds[j]/strike_angle[j]. ds is the segment area for one toroidal diagnostic bin delta_phi (R_mid*segment_length*delta_phi); it is not the full-torus area.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Accumulated wall-hit contributions since the previous heat write, reset after output. The scalar time is the physical sml.time value passed at the write; main_loop advances sml.time only later, so it is not an independently established interval-end timestamp. gstep identifies the current integration loop and tindex=gstep/period. Energy/count channels are interval sums; use differences of successive physical time values and the applicable collection area for rates/flux densities. Potential and displacement channels require their own number normalization.

  • centering: wall_segment

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:588.

xgc.heatdiag*.bp: sp_name + “_number_psi” (HeatDiagnosticsSpNameNumberPsi)

Native units: particle count.

Number of plasma particles (not markers) hitting each sampling point (in psi). Data is accumulated over all time steps between two subsequent heat diagnostic outputs.

  • plot_label: Deposited particle count in flux bins ({name})

  • axes: section,flux_bin

  • mesh_context: Same heat diagnostic BP product: psi[section,flux_bin] gives the flux-bin sampling coordinate; values are histogram/deposition bins, not triangle vertices.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Accumulated wall-hit contributions since the previous heat write, reset after output. The scalar time is the physical sml.time value passed at the write; main_loop advances sml.time only later, so it is not an independently established interval-end timestamp. gstep identifies the current integration loop and tindex=gstep/period. Energy/count channels are interval sums; use differences of successive physical time values and the applicable collection area for rates/flux densities. Potential and displacement channels require their own number normalization.

  • centering: bin

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:411.

xgc.heatdiag*.bp: sp_name + “_para_energy” (HeatDiagnosticsSpNameParaEnergy)

Native units: J.

Parallel kinetic energy hitting each sampling point (in RZ-coordinates). Data is accumulated over all time steps between two subsequent heat diagnostic outputs.

  • plot_label: $K_\parallel$ (deposited) ({name})

  • axes: section,z_bin,r_bin

  • mesh_context: Same heat diagnostic BP product: r[section,R_bin] and z[section,Z_bin] are sampling coordinates for [section,Z_bin,R_bin] deposition arrays; these are bins on a rectangular RZ sampling grid, not the unstructured field mesh.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Accumulated wall-hit contributions since the previous heat write, reset after output. The scalar time is the physical sml.time value passed at the write; main_loop advances sml.time only later, so it is not an independently established interval-end timestamp. gstep identifies the current integration loop and tindex=gstep/period. Energy/count channels are interval sums; use differences of successive physical time values and the applicable collection area for rates/flux densities. Potential and displacement channels require their own number normalization.

  • centering: bin

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:376.

xgc.heatdiag*.bp: sp_name + “_para_energy” (HeatDiagnosticsSpNameParaEnergy2)

Native units: J.

Parallel kinetic energy hitting each wall segment; nseg1 includes a reserved leading bin 0; current deposition assigns unmatched crossings to the nearest segment. Data is accumulated over all time steps between two subsequent heat diagnostic outputs.

  • plot_label: $K_\parallel$ (wall deposition) ({name})

  • axes: toroidal_bin,wall_segment

  • mesh_context: Same heat diagnostic product: r,z are ordered wall-polygon vertices, psi is the corresponding flux coordinate, ds is collection area. Species data use [toroidal_bin,wall_segment_with_reserved_zero]; bin 0 is reserved and has no physical segment; current nearest-segment fallback leaves it unused. Associate bins 1..nseg with successive polygon segments, including the closing segment. This polygon may include extra interpolated vertices relative to the ordered grid_wall_nodes mesh boundary. Data bin j+1 corresponds to geometry ds[j]/strike_angle[j]. ds is the segment area for one toroidal diagnostic bin delta_phi (R_mid*segment_length*delta_phi); it is not the full-torus area.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Accumulated wall-hit contributions since the previous heat write, reset after output. The scalar time is the physical sml.time value passed at the write; main_loop advances sml.time only later, so it is not an independently established interval-end timestamp. gstep identifies the current integration loop and tindex=gstep/period. Energy/count channels are interval sums; use differences of successive physical time values and the applicable collection area for rates/flux densities. Potential and displacement channels require their own number normalization.

  • centering: wall_segment

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:598.

xgc.heatdiag*.bp: sp_name + “_para_energy_psi” (HeatDiagnosticsSpNameParaEnergyPsi)

Native units: J.

Parallel kinetic energy hitting each sampling point (in psi). Data is accumulated over all time steps between two subsequent heat diagnostic outputs.

  • plot_label: $K_\parallel$ (deposited, flux bins) ({name})

  • axes: section,flux_bin

  • mesh_context: Same heat diagnostic BP product: psi[section,flux_bin] gives the flux-bin sampling coordinate; values are histogram/deposition bins, not triangle vertices.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Accumulated wall-hit contributions since the previous heat write, reset after output. The scalar time is the physical sml.time value passed at the write; main_loop advances sml.time only later, so it is not an independently established interval-end timestamp. gstep identifies the current integration loop and tindex=gstep/period. Energy/count channels are interval sums; use differences of successive physical time values and the applicable collection area for rates/flux densities. Potential and displacement channels require their own number normalization.

  • centering: bin

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:418.

xgc.heatdiag*.bp: sp_name + “_perp_energy” (HeatDiagnosticsSpNamePerpEnergy)

Native units: J.

Perpendicular kinetic energy hitting each sampling point (in RZ-coordinates). Data is accumulated over all time steps between two subsequent heat diagnostic outputs.

  • plot_label: $K_\perp$ (deposited) ({name})

  • axes: section,z_bin,r_bin

  • mesh_context: Same heat diagnostic BP product: r[section,R_bin] and z[section,Z_bin] are sampling coordinates for [section,Z_bin,R_bin] deposition arrays; these are bins on a rectangular RZ sampling grid, not the unstructured field mesh.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Accumulated wall-hit contributions since the previous heat write, reset after output. The scalar time is the physical sml.time value passed at the write; main_loop advances sml.time only later, so it is not an independently established interval-end timestamp. gstep identifies the current integration loop and tindex=gstep/period. Energy/count channels are interval sums; use differences of successive physical time values and the applicable collection area for rates/flux densities. Potential and displacement channels require their own number normalization.

  • centering: bin

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:383.

xgc.heatdiag*.bp: sp_name + “_perp_energy” (HeatDiagnosticsSpNamePerpEnergy2)

Native units: J.

Perpendicular kinetic energy hitting each wall segment; nseg1 includes a reserved leading bin 0; current deposition assigns unmatched crossings to the nearest segment. Data is accumulated over all time steps between two subsequent heat diagnostic outputs.

  • plot_label: $K_\perp$ (wall deposition) ({name})

  • axes: toroidal_bin,wall_segment

  • mesh_context: Same heat diagnostic product: r,z are ordered wall-polygon vertices, psi is the corresponding flux coordinate, ds is collection area. Species data use [toroidal_bin,wall_segment_with_reserved_zero]; bin 0 is reserved and has no physical segment; current nearest-segment fallback leaves it unused. Associate bins 1..nseg with successive polygon segments, including the closing segment. This polygon may include extra interpolated vertices relative to the ordered grid_wall_nodes mesh boundary. Data bin j+1 corresponds to geometry ds[j]/strike_angle[j]. ds is the segment area for one toroidal diagnostic bin delta_phi (R_mid*segment_length*delta_phi); it is not the full-torus area.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Accumulated wall-hit contributions since the previous heat write, reset after output. The scalar time is the physical sml.time value passed at the write; main_loop advances sml.time only later, so it is not an independently established interval-end timestamp. gstep identifies the current integration loop and tindex=gstep/period. Energy/count channels are interval sums; use differences of successive physical time values and the applicable collection area for rates/flux densities. Potential and displacement channels require their own number normalization.

  • centering: wall_segment

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:608.

xgc.heatdiag*.bp: sp_name + “_perp_energy_psi” (HeatDiagnosticsSpNamePerpEnergyPsi)

Native units: J.

Perpendicular kinetic energy hitting each sampling point (in psi). Data is accumulated over all time steps between two subsequent heat diagnostic outputs.

  • plot_label: $K_\perp$ (deposited, flux bins) ({name})

  • axes: section,flux_bin

  • mesh_context: Same heat diagnostic BP product: psi[section,flux_bin] gives the flux-bin sampling coordinate; values are histogram/deposition bins, not triangle vertices.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Accumulated wall-hit contributions since the previous heat write, reset after output. The scalar time is the physical sml.time value passed at the write; main_loop advances sml.time only later, so it is not an independently established interval-end timestamp. gstep identifies the current integration loop and tindex=gstep/period. Energy/count channels are interval sums; use differences of successive physical time values and the applicable collection area for rates/flux densities. Potential and displacement channels require their own number normalization.

  • centering: bin

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:425.

xgc.heatdiag*.bp: sp_name + “_potential” (HeatDiagnosticsSpNamePotential)

Native units: V.

Accumulated sum of sheath-potential samples times deposited plasma-particle weight over the heat diagnostic interval. Divide by the same species _number in each nonempty bin to obtain the deposited-particle-weighted mean sheath potential.

  • plot_label: Accumulated weighted sheath potential ({name})

  • axes: section,z_bin,r_bin

  • mesh_context: Same heat diagnostic BP product: r[section,R_bin] and z[section,Z_bin] are sampling coordinates for [section,Z_bin,R_bin] deposition arrays; these are bins on a rectangular RZ sampling grid, not the unstructured field mesh.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Accumulated wall-hit contributions since the previous heat write, reset after output. The scalar time is the physical sml.time value passed at the write; main_loop advances sml.time only later, so it is not an independently established interval-end timestamp. gstep identifies the current integration loop and tindex=gstep/period. Energy/count channels are interval sums; use differences of successive physical time values and the applicable collection area for rates/flux densities. Potential and displacement channels require their own number normalization.

  • centering: bin

  • normalization: Raw sum of pot*(w*w0) using the same deposition weights as the species number channel; no division by number, time, or area is applied. Dividing by the corresponding nonzero number channel yields a weighted mean potential in V.

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:390.

xgc.heatdiag*.bp: sp_name + “_potential” (HeatDiagnosticsSpNamePotential2)

Native units: V.

Accumulated sum of sheath-potential samples times deposited plasma-particle weight over the heat diagnostic interval. Divide by the same species _number in each nonempty bin to obtain the deposited-particle-weighted mean sheath potential. Current segment deposition leaves reserved bin 0 unused and assigns unmatched trajectories to the nearest wall segment.

  • plot_label: Accumulated weighted sheath potential ({name})

  • axes: toroidal_bin,wall_segment

  • mesh_context: Same heat diagnostic product: r,z are ordered wall-polygon vertices, psi is the corresponding flux coordinate, ds is collection area. Species data use [toroidal_bin,wall_segment_with_reserved_zero]; bin 0 is reserved and has no physical segment; current nearest-segment fallback leaves it unused. Associate bins 1..nseg with successive polygon segments, including the closing segment. This polygon may include extra interpolated vertices relative to the ordered grid_wall_nodes mesh boundary. Data bin j+1 corresponds to geometry ds[j]/strike_angle[j]. ds is the segment area for one toroidal diagnostic bin delta_phi (R_mid*segment_length*delta_phi); it is not the full-torus area.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Accumulated wall-hit contributions since the previous heat write, reset after output. The scalar time is the physical sml.time value passed at the write; main_loop advances sml.time only later, so it is not an independently established interval-end timestamp. gstep identifies the current integration loop and tindex=gstep/period. Energy/count channels are interval sums; use differences of successive physical time values and the applicable collection area for rates/flux densities. Potential and displacement channels require their own number normalization.

  • centering: wall_segment

  • normalization: Raw sum of pot*(w*w0) using the same deposition weights as the species number channel; no division by number, time, or area is applied. Dividing by the corresponding nonzero number channel yields a weighted mean potential in V.

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:618.

xgc.heatdiag*.bp: sp_name + “_potential_psi” (HeatDiagnosticsSpNamePotentialPsi)

Native units: V.

Accumulated sum of sheath-potential samples times deposited plasma-particle weight over the heat diagnostic interval. Divide by the same species _number_psi in each nonempty bin to obtain the deposited-particle-weighted mean sheath potential.

  • plot_label: Accumulated weighted sheath potential ({name})

  • axes: section,flux_bin

  • mesh_context: Same heat diagnostic BP product: psi[section,flux_bin] gives the flux-bin sampling coordinate; values are histogram/deposition bins, not triangle vertices.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: Accumulated wall-hit contributions since the previous heat write, reset after output. The scalar time is the physical sml.time value passed at the write; main_loop advances sml.time only later, so it is not an independently established interval-end timestamp. gstep identifies the current integration loop and tindex=gstep/period. Energy/count channels are interval sums; use differences of successive physical time values and the applicable collection area for rates/flux densities. Potential and displacement channels require their own number normalization.

  • centering: bin

  • normalization: Raw sum of pot*(w*w0) using the same deposition weights as the species number channel; no division by number, time, or area is applied. Dividing by the corresponding nonzero number channel yields a weighted mean potential in V.

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:432.

xgc.heatdiag*.bp: strike_angle (HeatDiagnosticsStrikeAngle)

Native units: rad.

Angle of the magnetic field relative to the wall surface normal.

  • plot_label: $\theta_{\mathrm{B,wall}}$

  • axes: wall_segment

  • mesh_context: Same heat diagnostic product: r,z are ordered wall-polygon vertices, psi is the corresponding flux coordinate, ds is collection area. Species data use [toroidal_bin,wall_segment_with_unassigned]; bin 0 is unassigned and has no physical segment. Associate bins 1..nseg with successive polygon segments, including the closing segment. This polygon may include extra interpolated vertices relative to the ordered grid_wall_nodes mesh boundary. ds[j] and strike_angle[j] refer to segment from vertex j to (j+1)%nseg; unlike species data, these geometry arrays have no unassigned bin.

  • centering: wall_segment

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:563.

xgc.heatdiag*.bp: time (HeatDiagnosticsTime)

Native units: s.

Simulation time of step in seconds.

  • plot_label: $t$

  • time_context: Accumulated wall-hit contributions since the previous heat write, reset after output. The scalar time is the physical sml.time value passed at the write; main_loop advances sml.time only later, so it is not an independently established interval-end timestamp. gstep identifies the current integration loop and tindex=gstep/period. Energy/count channels are interval sums; use differences of successive physical time values and the applicable collection area for rates/flux densities. Potential and displacement channels require their own number normalization. Clock convention: in the regular main loop gstep is incremented before diagnostic calls, but sml.time is incremented afterward. A checkpoint written later at the same gstep therefore records time larger by one physical timestep. Use the written time, global step and product/integration stage together; equal gstep alone does not establish simultaneous diagnostic and checkpoint states.

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:309.

xgc.heatdiag*.bp: time (HeatDiagnosticsTime2)

Native units: s.

Simulation time of step in seconds.

  • plot_label: $t$

  • time_context: Accumulated wall-hit contributions since the previous heat write, reset after output. The scalar time is the physical sml.time value passed at the write; main_loop advances sml.time only later, so it is not an independently established interval-end timestamp. gstep identifies the current integration loop and tindex=gstep/period. Energy/count channels are interval sums; use differences of successive physical time values and the applicable collection area for rates/flux densities. Potential and displacement channels require their own number normalization. Clock convention: in the regular main loop gstep is incremented before diagnostic calls, but sml.time is incremented afterward. A checkpoint written later at the same gstep therefore records time larger by one physical timestep. Use the written time, global step and product/integration stage together; equal gstep alone does not establish simultaneous diagnostic and checkpoint states.

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:522.

xgc.heatdiag*.bp: tindex (HeatDiagnosticsTindex)

Native units: Positive integer.

Diagnostic cadence index computed as integer gstep divided by the configured diagnostic period. This index is not the zero-based ADIOS step and need not start at 1 in a restarted or selected product.

  • plot_label: Output slice index

  • time_context: Accumulated wall-hit contributions since the previous heat write, reset after output. The scalar time is the physical sml.time value passed at the write; main_loop advances sml.time only later, so it is not an independently established interval-end timestamp. gstep identifies the current integration loop and tindex=gstep/period. Energy/count channels are interval sums; use differences of successive physical time values and the applicable collection area for rates/flux densities. Potential and displacement channels require their own number normalization.

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:319.

xgc.heatdiag*.bp: tindex (HeatDiagnosticsTindex2)

Native units: Positive integer.

Diagnostic cadence index computed as integer gstep divided by the configured diagnostic period. This index is not the zero-based ADIOS step and need not start at 1 in a restarted or selected product.

  • plot_label: Output slice index

  • time_context: Accumulated wall-hit contributions since the previous heat write, reset after output. The scalar time is the physical sml.time value passed at the write; main_loop advances sml.time only later, so it is not an independently established interval-end timestamp. gstep identifies the current integration loop and tindex=gstep/period. Energy/count channels are interval sums; use differences of successive physical time values and the applicable collection area for rates/flux densities. Potential and displacement channels require their own number normalization.

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:532.

xgc.heatdiag*.bp: z (HeatDiagnosticsZ)

Native units: m.

Z-coordinates of the sampling points.

  • plot_label: $Z$ (sampling grid)

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:332.

xgc.heatdiag*.bp: z (HeatDiagnosticsZ2)

Native units: m.

Z-coordinates of the sampling points.

  • plot_label: $Z$

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:545.

xgc.heatdiag*.bp: zsamples (HeatDiagnosticsZsamples)

Native units: Positive integer.

Number of sampling points in Z-direction.

  • plot_label: Z sample count

Source: XGC_core/cpp/diagnostics/heat_diagnostics.cpp:291.

xgc.loading.*.bp: marker_den (InitializeParticlesMarkerDen)

Native units: m^-3.

Marker particle density in each triangle of the planar mesh, toroidally averaged.

  • plot_label: $\langle n_{\mathrm{marker}}\rangle_\varphi$

  • axes: triangle

  • mesh_context: xgc.mesh.bp: rz, nd_connect_list (zero-based). One loading diagnostic value per triangle, toroidally averaged; display as cell values, not nodal linear-basis data.

  • centering: triangle

Source: XGC_core/cpp/plasma/initialize_particles.cpp:348.

xgc.loading.*.bp: ntriangle (InitializeParticlesNtriangle)

Native units: Positive integer.

Number of triangles in the planar mesh.

  • plot_label: $N_{\mathrm{tr}}$

Source: XGC_core/cpp/plasma/initialize_particles.cpp:343.

xgc.loop_vol.bp: gstep (LoopVoltageDiagGstep)

Native units: Non-negative integer.

XGC global simulation time step for this loop-voltage diagnostic output slice.

  • plot_label: Global time step

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/sources/loop_voltage_diag.cpp:45.

xgc.loop_vol.bp: jpara_diff_current (LoopVoltageDiagJparaDiffCurrent)

Native units: A m^-2.

Current controller error bt_sign*(j_parallel,equilibrium-j_parallel,plasma), with bt_sign the equilibrium toroidal-field sign.

  • plot_label: Current-drive error

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Positive values request stronger current drive in the positive toroidal direction. Current density is toroidally averaged and filtered; equilibrium/plasma currents are also flux-surface averaged when sml_loop_voltage_fsa is enabled.

Source: XGC_core/cpp/sources/loop_voltage_diag.cpp:85.

xgc.loop_vol.bp: jpara_diff_integral (LoopVoltageDiagJparaDiffIntegral)

Native units: A m^-2.

Cumulative controller sum of bt_sign*(j_parallel,equilibrium-j_parallel,plasma) over current-drive applications; no physical timestep factor is included.

  • plot_label: Cumulative current-drive error sum

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: Define delta_j=bt_sign*(j_parallel,equilibrium-j_parallel,plasma), where bt_sign is the equilibrium toroidal-field sign. The controller updates integral += delta_j on each application, so this is a discrete sum in A m^-2, not a physical-time integral.

Source: XGC_core/cpp/sources/loop_voltage_diag.cpp:69.

xgc.loop_vol.bp: jpara_diff_previous (LoopVoltageDiagJparaDiffPrevious)

Native units: A m^-2.

Exponentially smoothed controller error bt_sign*(j_parallel,equilibrium-j_parallel,plasma), updated as previous=0.8*previous+0.2*current on each current-drive application.

  • plot_label: Smoothed current-drive error

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • normalization: The array contains the smoothed error history, not the derivative or current-minus-previous difference. It is updated before diagnostic writing and retained verbatim in the checkpoint.

Source: XGC_core/cpp/sources/loop_voltage_diag.cpp:77.

xgc.loop_vol.bp: loop_voltage (LoopVoltageDiagLoopVoltage)

Native units: V.

Loop voltage. The corresponding toroidal electric field is \(E_{\\phi}=V_{\\mathrm{loop}}/(2\\pi R)\), so positive loop voltage implies an electric field in the positive toroidal direction.

  • plot_label: $V_{\mathrm{loop}}$

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • coordinate_context: Node-centered spatial loop-voltage profile, not a single global voltage. The corresponding imposed toroidal electric field is E_phi(R,Z)=V_loop(R,Z)/(2*pi*R). Any global IDS voltage hint requires an explicitly chosen representative value or physical reduction.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: equilibrium/time_slice(:)/global_quantities/v_external

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/sources/loop_voltage_diag.cpp:63.

xgc.loop_vol.bp: time (LoopVoltageDiagTime)

Native units: s.

Simulation time of step in seconds.

  • plot_label: $t$

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions. Clock convention: in the regular main loop gstep is incremented before diagnostic calls, but sml.time is incremented afterward. A checkpoint written later at the same gstep therefore records time larger by one physical timestep. Use the written time, global step and product/integration stage together; equal gstep alone does not establish simultaneous diagnostic and checkpoint states.

  • imas_candidates: plasma_profiles/time

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_high

Source: XGC_core/cpp/sources/loop_voltage_diag.cpp:58.

xgc.loop_vol.bp: tindex (LoopVoltageDiagTindex)

Native units: Positive integer.

Diagnostic cadence index computed as integer gstep divided by the configured diagnostic period. This index is not the zero-based ADIOS step and need not start at 1 in a restarted or selected product.

  • plot_label: Output slice index

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/sources/loop_voltage_diag.cpp:50.

xgc.mesh.bp: PEST_angle (GridPESTAngle)

Native units: rad.

PEST coordinate on the mesh nodes (only written for stellarator simulations).

  • plot_label: $\theta^\ast_{\mathrm{PEST}}$

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • centering: node

Source: XGC_core/cpp/grid/grid.cpp:430.

xgc.mesh.bp: VMEC_angle (GridVMECAngle)

Native units: rad.

VMEC coordinate on the mesh nodes (only written for stellarator simulations).

  • plot_label: $\theta^\ast_{\mathrm{VMEC}}$

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • centering: node

Source: XGC_core/cpp/grid/grid.cpp:425.

xgc.mesh.bp: epsilon (GridEpsilon)

Native units: dimensionless.

Inverse aspect ratio \(r_{\\min}/R_{\\maj}\) for the flux surface.

  • plot_label: $\epsilon$

  • axes: flux_surface

  • mesh_context: xgc.mesh.bp: psi_surf labels flux surfaces; surf_len gives each surface’s valid node count; surf_idx stores one-based node IDs with zero padding beyond surf_len. Subtract one only from valid surf_idx entries before looking up rz. theta gives the straight-field-line poloidal angle. Treat disconnected/private-flux regions separately when interpolating in flux.

  • centering: flux_surface

Source: XGC_core/cpp/grid/grid.cpp:533.

xgc.mesh.bp: ff_1dp_p (GridFf1dpP)

Native units: dimensionless.

Barycentric interpolation weights at the full-plane field-following projection of each mesh node.

  • plot_label: Full-plane barycentric weight

  • mesh_context: Same xgc.mesh.bp:rz,nd_connect_list define the source mesh. ff_hdp_p and ff_1dp_p are barycentric weights for field-line images, not field values. Corresponding xgc.ff_hdp_fwd/rev.bp or xgc.ff_1dp_fwd/rev.bp:eindex,value identify the target nodes for each row. Use one-based eindex only in valid nelement slots.

  • coordinate_context: Barycentric weights on half/full-step field-line image triangles, indexed by projection direction and mesh node. Weights alone do not identify target vertices; use matching field-following matrix eindex/value data for reconstruction.

  • centering: projection_weight

Source: XGC_core/cpp/grid/grid.cpp:508.

xgc.mesh.bp: ff_hdp_p (GridFfHdpP)

Native units: dimensionless.

Barycentric interpolation weights at the half-plane field-following projection of each mesh node.

  • plot_label: Half-plane barycentric weight

  • mesh_context: Same xgc.mesh.bp:rz,nd_connect_list define the source mesh. ff_hdp_p and ff_1dp_p are barycentric weights for field-line images, not field values. Corresponding xgc.ff_hdp_fwd/rev.bp or xgc.ff_1dp_fwd/rev.bp:eindex,value identify the target nodes for each row. Use one-based eindex only in valid nelement slots.

  • coordinate_context: Barycentric weights on half/full-step field-line image triangles, indexed by projection direction and mesh node. Weights alone do not identify target vertices; use matching field-following matrix eindex/value data for reconstruction.

  • centering: projection_weight

Source: XGC_core/cpp/grid/grid.cpp:503.

xgc.mesh.bp: grid_nwall (GridGridNwall)

Native units: Non-negative integer.

Number of nodes that make up the boundary of the domain (and represent the wall).

  • plot_label: Wall node count

Source: XGC_core/cpp/grid/grid.cpp:633.

xgc.mesh.bp: grid_wall_nodes (GridGridWallNodes)

Native units: Integer index.

Zero-based mesh-node indices ordered along the domain boundary by boundary-edge adjacency. The setup orients the boundary clockwise and, when the magnetic-axis-height crossing exists, starts near its inboard crossing.

  • plot_label: Wall node index

  • coordinate_context: Ordered zero-based boundary-node IDs into the corresponding mesh geometry. Boundary order applies to this complete grid_wall_nodes list; a separately selected wall-node subset need not preserve neighboring segments.

Source: XGC_core/cpp/grid/grid.cpp:643.

xgc.mesh.bp: m_max_surf (GridMMaxSurf)

Native units: Non-negative integer.

Poloidal mode cutoff obtained by integer truncation of 2*pi/(sml_mode_select_cutoff*selected delta-theta). The selected straight-field-line angle spacing is the maximum adjacent-node spacing by default, or the spacing at the maximum-B node for sml_mode_select_cutoff_mode=2. Unprocessed surfaces retain zero. The primary-separatrix value is replaced using the smaller neighboring cutoff scaled by the ratio of q values.

  • plot_label: Maximum resolved poloidal mode number

  • axes: flux_surface

  • mesh_context: xgc.mesh.bp: psi_surf labels flux surfaces; surf_len gives each surface’s valid node count; surf_idx stores one-based node IDs with zero padding beyond surf_len. Subtract one only from valid surf_idx entries before looking up rz. theta gives the straight-field-line poloidal angle. Treat disconnected/private-flux regions separately when interpolating in flux.

  • centering: flux_surface

Source: XGC_core/cpp/grid/grid.cpp:618.

xgc.mesh.bp: n_geo (GridNGeo)

Native units: Positive integer.

Number of flux surfaces; same value as nsurf.

  • plot_label: $N_{\mathrm{surf}}$

Source: XGC_core/cpp/grid/grid.cpp:372.

xgc.mesh.bp: n_n (GridNN)

Native units: Positive integer.

Number of mesh vertices (per plane).

  • plot_label: $N_n$

Source: XGC_core/cpp/grid/grid.cpp:362.

xgc.mesh.bp: n_t (GridNT)

Native units: Positive integer.

Number of mesh triangles (per plane).

  • plot_label: $N_{\mathrm{tr}}$

Source: XGC_core/cpp/grid/grid.cpp:367.

xgc.mesh.bp: nd_connect_list (PlaneNdConnectList)

Native units: index.

Triangle connectivity list for the poloidal plane mesh; each triplet contains zero-based mesh-node indices.

  • plot_label: Triangle vertex index

  • coordinate_context: Triangle vertex indices are zero-based, last axis selects the three triangle vertices. rz coordinates use R,Z (m). Stellarator variables with lplane_/rplane_ prefixes refer to those planes.

Source: XGC_core/cpp/grid/plane.cpp:479.

xgc.mesh.bp: nextnode (GridNextnode)

Native units: Integer index.

Zero-based destination mesh-node index with the largest barycentric weight in the forward full-plane field-line projection of each source node.

  • plot_label: Nearest field-line image node

  • coordinate_context: Choose the destination vertex having the largest barycentric interpolation weight in the forward projection; indices are zero-based. This selection need not be the Euclidean nearest destination vertex.

Source: XGC_core/cpp/grid/grid.cpp:490.

xgc.mesh.bp: node_vol (GridNodeVol)

Native units: m^3.

Nodal volume weight for linear barycentric interpolation and straight toroidal extrusion over one local toroidal slab of angular width delta_phi. For axisymmetric geometry the full-torus volume is wedge_n*nphi*sum(node_vol); for varying stellarator geometry sum the corresponding per-plane slab volumes before multiplying by wedge_n.

  • plot_label: $\Delta V_{\mathrm{lin}}$

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • centering: node

Source: XGC_core/cpp/grid/grid.cpp:385.

xgc.mesh.bp: node_vol_ff0 (GridNodeVolFf0)

Native units: m^3.

Field-following nodal volume weight on the midplane geometry for toroidal deposition side 0 (left). The weight integrates linear barycentric and toroidal deposition over one local slab.

  • plot_label: $\Delta V_{\mathrm{ff},0}$ (midplane)

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • centering: node

Source: XGC_core/cpp/grid/grid.cpp:440.

xgc.mesh.bp: node_vol_ff0_l (GridNodeVolFf0L)

Native units: m^3.

Field-following nodal volume weight on the left-plane geometry for toroidal deposition side 0 (left). The weight integrates linear barycentric and toroidal deposition over one local slab.

  • plot_label: $\Delta V_{\mathrm{ff},0}$ (left-plane)

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • centering: node

Source: XGC_core/cpp/grid/grid.cpp:457.

xgc.mesh.bp: node_vol_ff0_r (GridNodeVolFf0R)

Native units: m^3.

Field-following nodal volume weight on the right-plane geometry for toroidal deposition side 0 (left). The weight integrates linear barycentric and toroidal deposition over one local slab.

  • plot_label: $\Delta V_{\mathrm{ff},0}$ (right-plane)

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • centering: node

Source: XGC_core/cpp/grid/grid.cpp:473.

xgc.mesh.bp: node_vol_ff1 (GridNodeVolFf1)

Native units: m^3.

Field-following nodal volume weight on the midplane geometry for toroidal deposition side 1 (right). The weight integrates linear barycentric and toroidal deposition over one local slab.

  • plot_label: $\Delta V_{\mathrm{ff},1}$ (midplane)

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • centering: node

Source: XGC_core/cpp/grid/grid.cpp:448.

xgc.mesh.bp: node_vol_ff1_l (GridNodeVolFf1L)

Native units: m^3.

Field-following nodal volume weight on the left-plane geometry for toroidal deposition side 1 (right). The weight integrates linear barycentric and toroidal deposition over one local slab.

  • plot_label: $\Delta V_{\mathrm{ff},1}$ (left-plane)

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • centering: node

Source: XGC_core/cpp/grid/grid.cpp:465.

xgc.mesh.bp: node_vol_ff1_r (GridNodeVolFf1R)

Native units: m^3.

Field-following nodal volume weight on the right-plane geometry for toroidal deposition side 1 (right). The weight integrates linear barycentric and toroidal deposition over one local slab.

  • plot_label: $\Delta V_{\mathrm{ff},1}$ (right-plane)

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • centering: node

Source: XGC_core/cpp/grid/grid.cpp:481.

xgc.mesh.bp: node_vol_nearest (GridNodeVolNearest)

Native units: m^3.

Voronoi volume of the mesh vertices under nearest neighbor interpolation and field-following projection.

  • plot_label: $\Delta V_{\mathrm{nearest}}$

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • centering: node

Source: XGC_core/cpp/grid/grid.cpp:400.

xgc.mesh.bp: nphi (GridNphi)

Native units: Positive integer.

Number of toroidal mesh planes \(N_{\\mathrm{planes}}\) (only written for stellarator simulations).

  • plot_label: $N_\varphi$

Source: XGC_core/cpp/grid/grid.cpp:415.

xgc.mesh.bp: nsurf (GridNsurf)

Native units: Positive integer.

Number of flux surfaces.

  • plot_label: $N_{\mathrm{surf}}$

Source: XGC_core/cpp/grid/grid.cpp:572.

xgc.mesh.bp: one_per_dx (GridOnePerDx)

Native units: m.

Distance traveled along the magnetic field line to adjacent full toroidal mesh planes.

  • plot_label: Full-plane field-line distance

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • centering: node

Source: XGC_core/cpp/grid/grid.cpp:498.

xgc.mesh.bp: prefix+”delta_phi” (PlanePrefixDeltaPhi)

Native units: rad.

Toroidal angular spacing between neighboring mesh planes.

  • plot_label: $\Delta \varphi$ ({name})

Source: XGC_core/cpp/grid/plane.cpp:502.

xgc.mesh.bp: prefix+”mapping” (PlanePrefixMapping)

Native units: [1/m, 1/m, m].

Per-triangle affine mapping coefficients from physical (R,Z) coordinates to local barycentric triangle coordinates; entries 0 and 1 store the inverse edge matrix and entry 2 stores the reference vertex position.

  • plot_label: Triangle affine mapping coefficient ({name})

  • mesh_context: Same xgc.mesh.bp plane-prefix rz and nd_connect_list define the triangles. mapping coefficients are per triangle; component slots 0/1 contain inverse edges and slot 2 the reference RZ position.

  • centering: triangle

Source: XGC_core/cpp/grid/plane.cpp:485.

xgc.mesh.bp: prefix+”nd_connect_list” (PlanePrefixNdConnectList)

Native units: index.

Triangle connectivity list for each plane (multiple planes in case of stellarators); each triplet contains zero-based mesh-node indices.

  • plot_label: Triangle vertex index ({name})

  • coordinate_context: Triangle vertex indices are zero-based, last axis selects the three triangle vertices. rz coordinates use R,Z (m). Stellarator variables with lplane_/rplane_ prefixes refer to those planes.

Source: XGC_core/cpp/grid/plane.cpp:473.

xgc.mesh.bp: prefix+”region” (PlanePrefixRegion)

Native units: index.

Region identifier for each mesh vertex.

  • plot_label: Mesh region ID ({name})

  • mesh_context: Same xgc.mesh.bp prefix+rz defines the vertices; prefix+nd_connect_list supplies zero-based triangle connectivity. Prefix is a plane identifier (empty, lplane_, rplane_), not species.

  • coordinate_context: Categorical region IDs; show discrete colors, do not interpolate region numbers.

  • centering: node_category

Source: XGC_core/cpp/grid/plane.cpp:492.

xgc.mesh.bp: prefix+”rz” (PlanePrefixRz)

Native units: m.

Spatial location in (R,Z) coordinates of each mesh vertex in each stellarator plane.

  • plot_label: $(R,Z)$ ({name})

  • coordinate_context: Cylindrical coordinates R,Z in m; last axis component 0 is R and 1 is Z. For stellarators the leading axis selects toroidal plane; prefixes lplane_/rplane_ select bracketing geometry.

Source: XGC_core/cpp/grid/plane.cpp:463.

xgc.mesh.bp: prefix+”wedge_angle” (PlanePrefixWedgeAngle)

Native units: rad.

Toroidal wedge angle represented by this simulation.

  • plot_label: Toroidal wedge angle ({name})

Source: XGC_core/cpp/grid/plane.cpp:497.

xgc.mesh.bp: psi (GridPsi)

Native units: Wb/rad.

Equilibrium magnetic-flux label evaluated at mesh vertices; axisymmetric runs use poloidal flux divided by 2*pi.

  • plot_label: $\psi/(2\pi)$

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • centering: node

Source: XGC_core/cpp/grid/grid.cpp:377.

xgc.mesh.bp: psi_surf (GridPsiSurf)

Native units: Wb/rad.

Aligned-node volume-weighted equilibrium flux label for each mesh-file surface, toroidally combined for stellarators; axisymmetric runs use poloidal flux divided by 2*pi.

  • plot_label: $\psi/(2\pi)$ (flux surface)

  • axes: flux_surface

  • mesh_context: xgc.mesh.bp: psi_surf labels flux surfaces; surf_len gives each surface’s valid node count; surf_idx stores one-based node IDs with zero padding beyond surf_len. Subtract one only from valid surf_idx entries before looking up rz. theta gives the straight-field-line poloidal angle. Treat disconnected/private-flux regions separately when interpolating in flux.

  • centering: flux_surface

Source: XGC_core/cpp/grid/grid.cpp:597.

xgc.mesh.bp: qsafety (GridQsafety)

Native units: dimensionless.

Nonnegative winding measure abs(surf_arc_len)/(2*pi); interpreted as the safety-factor magnitude on processed closed surfaces and as a connection-integral measure on processed open surfaces. Unprocessed surfaces retain zero.

  • plot_label: $|q|$

  • axes: flux_surface

  • mesh_context: xgc.mesh.bp: psi_surf labels flux surfaces; surf_len gives each surface’s valid node count; surf_idx stores one-based node IDs with zero padding beyond surf_len. Subtract one only from valid surf_idx entries before looking up rz. theta gives the straight-field-line poloidal angle. Treat disconnected/private-flux regions separately when interpolating in flux.

  • centering: flux_surface

  • normalization: Native q is abs(surf_arc_len)/(2*pi), so its sign is discarded. The signed IDS safety factor requires the equilibrium current/field orientation and COCOS convention; do not interpret unprocessed zero entries as physical q=0.

  • imas_candidates: equilibrium/time_slice(:)/profiles_1d/q

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/grid/grid.cpp:538.

xgc.mesh.bp: qsafety (GridQsafety2)

Native units: dimensionless.

Nonnegative winding measure abs(surf_arc_len)/(2*pi); interpreted as the safety-factor magnitude on processed closed surfaces and as a connection-integral measure on processed open surfaces. Unprocessed surfaces retain zero.

  • plot_label: $|q|$

  • axes: flux_surface

  • mesh_context: xgc.mesh.bp: psi_surf labels flux surfaces; surf_len gives each surface’s valid node count; surf_idx stores one-based node IDs with zero padding beyond surf_len. Subtract one only from valid surf_idx entries before looking up rz. theta gives the straight-field-line poloidal angle. Treat disconnected/private-flux regions separately when interpolating in flux.

  • centering: flux_surface

  • normalization: Native q is abs(surf_arc_len)/(2*pi), so its sign is discarded. The signed IDS safety factor requires the equilibrium current/field orientation and COCOS convention; do not interpret unprocessed zero entries as physical q=0.

  • imas_candidates: equilibrium/time_slice(:)/profiles_1d/q

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/grid/grid.cpp:558.

xgc.mesh.bp: rmaj (GridRmaj)

Native units: m.

Radial midpoint of the sampled right-plane flux-surface nodes, (max(R)+min(R))/2.

  • plot_label: $R_{\mathrm{maj}}$ (flux surface)

  • axes: flux_surface

  • mesh_context: xgc.mesh.bp: psi_surf labels flux surfaces; surf_len gives each surface’s valid node count; surf_idx stores one-based node IDs with zero padding beyond surf_len. Subtract one only from valid surf_idx entries before looking up rz. theta gives the straight-field-line poloidal angle. Treat disconnected/private-flux regions separately when interpolating in flux.

  • centering: flux_surface

Source: XGC_core/cpp/grid/grid.cpp:526.

xgc.mesh.bp: rmin (GridRmin)

Native units: m.

Radial half-width of the sampled right-plane flux-surface nodes, (max(R)-min(R))/2.

  • plot_label: $r_{\mathrm{min}}$ (flux surface)

  • axes: flux_surface

  • mesh_context: xgc.mesh.bp: psi_surf labels flux surfaces; surf_len gives each surface’s valid node count; surf_idx stores one-based node IDs with zero padding beyond surf_len. Subtract one only from valid surf_idx entries before looking up rz. theta gives the straight-field-line poloidal angle. Treat disconnected/private-flux regions separately when interpolating in flux.

  • centering: flux_surface

Source: XGC_core/cpp/grid/grid.cpp:519.

xgc.mesh.bp: rz (PlaneRz)

Native units: m.

Spatial location of each mesh node in the poloidal plane; the first index selects the node and the second index selects the R or Z coordinate.

  • plot_label: $(R,Z)$

  • coordinate_context: Cylindrical coordinates R,Z in m; last axis component 0 is R and 1 is Z. For stellarators the leading axis selects toroidal plane; prefixes lplane_/rplane_ select bracketing geometry.

Source: XGC_core/cpp/grid/plane.cpp:467.

xgc.mesh.bp: surf_arc_len (GridSurfArcLen)

Native units: dimensionless.

Signed field-line toroidal advance integral over the processed poloidal surface segment: integral B_phi/(R*|B_p|) d l_p. Closed surfaces give 2*pi times the signed field-line winding measure; unprocessed surfaces retain zero.

  • plot_label: $\int B_\varphi/(R|B_p|)\,\mathrm{d}l_p$

  • axes: flux_surface

  • mesh_context: xgc.mesh.bp: psi_surf labels flux surfaces; surf_len gives each surface’s valid node count; surf_idx stores one-based node IDs with zero padding beyond surf_len. Subtract one only from valid surf_idx entries before looking up rz. theta gives the straight-field-line poloidal angle. Treat disconnected/private-flux regions separately when interpolating in flux.

  • centering: flux_surface

  • normalization: qsafety=abs(surf_arc_len)/(2*pi). This is an angular winding integral, not a metric length; cylindrical geometry uses the magnetic-axis radius in place of local R.

Source: XGC_core/cpp/grid/grid.cpp:588.

xgc.mesh.bp: surf_idx (GridSurfIdx)

Native units: Integer index.

Node index for each node on a flux surface. First index is flux surface label and second index is the node on the flux surface. Nodes counted from 1. Filled with zeros beyond surf_len.

  • plot_label: Surface node index

  • coordinate_context: One-based node IDs; use surf_len to discard zero padding before subtracting one. Never subtract one from nd_connect_list or grid_wall_nodes, which are already zero-based.

Source: XGC_core/cpp/grid/grid.cpp:610.

xgc.mesh.bp: surf_len (GridSurfLen)

Native units: Positive integer.

Number of nodes in each flux surface.

  • plot_label: Surface node count

  • axes: flux_surface

  • mesh_context: xgc.mesh.bp: psi_surf labels flux surfaces; surf_len gives each surface’s valid node count; surf_idx stores one-based node IDs with zero padding beyond surf_len. Subtract one only from valid surf_idx entries before looking up rz. theta gives the straight-field-line poloidal angle. Treat disconnected/private-flux regions separately when interpolating in flux.

  • centering: flux_surface

Source: XGC_core/cpp/grid/grid.cpp:605.

xgc.mesh.bp: surf_maxlen (GridSurfMaxlen)

Native units: Positive integer.

The maximum number of nodes on any flux surface. Used for indexing surf_idx.

  • plot_label: Maximum surface node count

Source: XGC_core/cpp/grid/grid.cpp:577.

xgc.mesh.bp: theta (GridTheta)

Native units: rad.

Straight-field-line poloidal angle. Core: \(\\theta=0\) on the outer midplane; SOL: zero on the outer target and \(2\\pi\) on the inner target.

  • plot_label: $\theta^\ast$

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • centering: node

  • normalization: Generalized poloidal angle is assigned on processed surface nodes; -10 is the unassigned sentinel on nodes not processed by calc_gen_theta_psi. Mask these values before angular interpolation or plotting.

Source: XGC_core/cpp/grid/grid.cpp:582.

xgc.mesh.bp: tr_area (GridTrArea)

Native units: m^2.

Surface area of the triangles.

  • plot_label: $\Delta A_{\mathrm{tr}}$

  • axes: triangle

  • mesh_context: Same xgc.mesh.bp: rz, nd_connect_list (zero-based). Triangle area is cell-centered, not a nodal field.

  • centering: triangle

Source: XGC_core/cpp/grid/grid.cpp:395.

xgc.mesh.bp: trapped (GridTrapped)

Native units: dimensionless.

Trapped-particle fraction calculated on processed closed flux surfaces; entries for open, excluded, or too-short surfaces remain zero and are not computed trapped-fraction estimates.

  • plot_label: $f_t$

  • axes: flux_surface

  • mesh_context: xgc.mesh.bp: psi_surf labels flux surfaces; surf_len gives each surface’s valid node count; surf_idx stores one-based node IDs with zero padding beyond surf_len. Subtract one only from valid surf_idx entries before looking up rz. theta gives the straight-field-line poloidal angle. Treat disconnected/private-flux regions separately when interpolating in flux.

  • centering: flux_surface

  • normalization: Only processed periodic surfaces have physical trapped-fraction estimates; initialized zero values on other surfaces indicate no evaluation.

Source: XGC_core/cpp/grid/grid.cpp:547.

xgc.mesh.bp: wedge_n (GridWedgeN)

Native units: Positive integer.

Torus wedge number or toroidal periodicity (only written for stellarator simulations).

  • plot_label: $N_{\mathrm{wedge}}$

Source: XGC_core/cpp/grid/grid.cpp:420.

xgc.neutrals.*.bp: den_neut (NeutralsDenNeut)

Native units: m^-3.

Neutrals density.

  • plot_label: $n_{\mathrm{neutral}}$

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/neutral(:)/density(:)/values

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/sources/neutrals.cpp:1071.

xgc.neutrals.*.bp: gstep (NeutralsGstep)

Native units: Non-negative integer.

XGC global simulation time step for this neutral output slice.

  • plot_label: Global time step

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/sources/neutrals.cpp:1051.

xgc.neutrals.*.bp: n_n (NeutralsNN)

Native units: Positive integer.

Number of mesh nodes per plane.

  • plot_label: $N_n$

Source: XGC_core/cpp/sources/neutrals.cpp:1066.

xgc.neutrals.*.bp: rel_std (NeutralsRelStd)

Native units: dimensionless.

Relative statistical uncertainty of the neutral background density on mesh nodes.

  • plot_label: Relative neutral-density uncertainty

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

Source: XGC_core/cpp/sources/neutrals.cpp:1081.

xgc.neutrals.*.bp: temp_neut (NeutralsTempNeut)

Native units: eV.

Neutrals temperature.

  • plot_label: $T_{\mathrm{neutral}}$

  • axes: mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: node

  • imas_candidates: plasma_profiles/ggd(:)/neutral(:)/temperature(:)/values

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/sources/neutrals.cpp:1076.

xgc.neutrals.*.bp: time (NeutralsTime)

Native units: s.

Simulation time of this neutral output slice.

  • plot_label: $t$

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions. Clock convention: in the regular main loop gstep is incremented before diagnostic calls, but sml.time is incremented afterward. A checkpoint written later at the same gstep therefore records time larger by one physical timestep. Use the written time, global step and product/integration stage together; equal gstep alone does not establish simultaneous diagnostic and checkpoint states.

  • imas_candidates: plasma_profiles/time

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_high

Source: XGC_core/cpp/sources/neutrals.cpp:1061.

xgc.neutrals.*.bp: tindex (NeutralsTindex)

Native units: Positive integer.

Index of this neutral diagnostic output time slice. Values: 1, 2, …, # outputs.

  • plot_label: Output slice index

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/sources/neutrals.cpp:1056.

xgc.oneddiag.bp: [species]_gc_density_1d or _df_1d (Diag1dGcDensity)

Native units: m^-3.

Gyro-center density. Note: gyro-center density is close to the actual plasma density for electrons; for ions add polarization density to get actual ion density.

  • plot_label: $n_{\mathrm{gc}}$ ({name})

  • axes: flux_surface

  • mesh_context: xgc.oneddiag.bp (same ADIOS step): psi_mks (poloidal flux divided by 2*pi, T m^2) and psi (normalized flux) are radial coordinates of flux-surface averages. Interpolate profiles in the chosen flux coordinate; do not triangulate radial profile arrays.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: flux_surface

  • normalization: Channel-specific units and definition apply. _1d is sampled full-f-marker contribution; _df_1d uses delta-f weights, with grid/background contributions for total-f species. gc_density and marker_density divide their sums by bin volume; all other particle-averaged channels divide by the appropriate gyrocenter particle/density weight; psi-space radial fluxes require division by <|grad psi|> or sqrt([species]_grad_psi_sqr_1d) for a per-particle physical radial velocity, energy transport, or momentum transport; multiply by matching gyrocenter density to obtain a flux density.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:388.

xgc.oneddiag.bp: [species]_gc_poloidal_flow_1d or _df_1d (Diag1dGcPoloidalFlow)

Native units: m/s.

Flux averaged poloidal flow of gyro-center. Differs from the poloidal flow by the diamagnetic flow.

  • plot_label: $u_{\theta,\mathrm{gc}}$ ({name})

  • axes: flux_surface

  • mesh_context: xgc.oneddiag.bp (same ADIOS step): psi_mks (poloidal flux divided by 2*pi, T m^2) and psi (normalized flux) are radial coordinates of flux-surface averages. Interpolate profiles in the chosen flux coordinate; do not triangulate radial profile arrays.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: flux_surface

  • normalization: Channel-specific units and definition apply. _1d is sampled full-f-marker contribution; _df_1d uses delta-f weights, with grid/background contributions for total-f species. gc_density and marker_density divide their sums by bin volume; all other particle-averaged channels divide by the appropriate gyrocenter particle/density weight; psi-space radial fluxes require division by <|grad psi|> or sqrt([species]_grad_psi_sqr_1d) for a per-particle physical radial velocity, energy transport, or momentum transport; multiply by matching gyrocenter density to obtain a flux density.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:399.

xgc.oneddiag.bp: [species]_gc_toroidal_flow_1d or _df_1d (Diag1dGcToroidalFlow)

Native units: m/s.

Flux averaged toroidal flow of gyro-center. Differs from the toroidal flow by the diamagnetic flow.

  • plot_label: $u_{\varphi,\mathrm{gc}}$ ({name})

  • axes: flux_surface

  • mesh_context: xgc.oneddiag.bp (same ADIOS step): psi_mks (poloidal flux divided by 2*pi, T m^2) and psi (normalized flux) are radial coordinates of flux-surface averages. Interpolate profiles in the chosen flux coordinate; do not triangulate radial profile arrays.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: flux_surface

  • normalization: Channel-specific units and definition apply. _1d is sampled full-f-marker contribution; _df_1d uses delta-f weights, with grid/background contributions for total-f species. gc_density and marker_density divide their sums by bin volume; all other particle-averaged channels divide by the appropriate gyrocenter particle/density weight; psi-space radial fluxes require division by <|grad psi|> or sqrt([species]_grad_psi_sqr_1d) for a per-particle physical radial velocity, energy transport, or momentum transport; multiply by matching gyrocenter density to obtain a flux density.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:394.

xgc.oneddiag.bp: [species]_grad_psi_sqr_1d or _df_1d (Diag1dGradPsiSqr)

Native units: (T m)^2.

Particle average of \({\\mathrm{\|\\nabla \\psi\|^2}}\) for each flux surface. Roughly flux average of \({\\mathrm{\|\\nabla \\psi\|^2}}\) This is usually used for converting \({\\mathrm{{\\partial \\psi}}\) to \({\\mathrm{\\partial r}}\)

  • plot_label: $\langle |\nabla\psi|^2\rangle$ ({name})

  • axes: flux_surface

  • mesh_context: xgc.oneddiag.bp (same ADIOS step): psi_mks (poloidal flux divided by 2*pi, T m^2) and psi (normalized flux) are radial coordinates of flux-surface averages. Interpolate profiles in the chosen flux coordinate; do not triangulate radial profile arrays.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: flux_surface

  • normalization: Channel-specific units and definition apply. _1d is sampled full-f-marker contribution; _df_1d uses delta-f weights, with grid/background contributions for total-f species. gc_density and marker_density divide their sums by bin volume; all other particle-averaged channels divide by the appropriate gyrocenter particle/density weight; psi-space radial fluxes require division by <|grad psi|> or sqrt([species]_grad_psi_sqr_1d) for a per-particle physical radial velocity, energy transport, or momentum transport; multiply by matching gyrocenter density to obtain a flux density.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:478.

xgc.oneddiag.bp: [species]_marker_density_1d or _df_1d (Diag1dMarkerDensity)

Native units: m^-3.

Marker count per bin volume for _1d; the corresponding _df_1d channel is the sum of marker w1 weights per bin volume and can be signed. Neither marker-density channel is normalized by the gyrocenter-density sum.

  • plot_label: $n_{\mathrm{marker}}$ ({name})

  • axes: flux_surface

  • mesh_context: xgc.oneddiag.bp (same ADIOS step): psi_mks (poloidal flux divided by 2*pi, T m^2) and psi (normalized flux) are radial coordinates of flux-surface averages. Interpolate profiles in the chosen flux coordinate; do not triangulate radial profile arrays.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: flux_surface

  • normalization: Channel-specific units and definition apply. _1d is sampled full-f-marker contribution; _df_1d uses delta-f weights, with grid/background contributions for total-f species. gc_density and marker_density divide their sums by bin volume; all other particle-averaged channels divide by the appropriate gyrocenter particle/density weight; psi-space radial fluxes require division by <|grad psi|> or sqrt([species]_grad_psi_sqr_1d) for a per-particle physical radial velocity, energy transport, or momentum transport; multiply by matching gyrocenter density to obtain a flux density.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:495.

xgc.oneddiag.bp: [species]_parallel_flow_1d or _df_1d (Diag1dParallelFlow)

Native units: m/s.

Parallel flow velocity.

  • plot_label: $u_\parallel$ ({name})

  • axes: flux_surface

  • mesh_context: xgc.oneddiag.bp (same ADIOS step): psi_mks (poloidal flux divided by 2*pi, T m^2) and psi (normalized flux) are radial coordinates of flux-surface averages. Interpolate profiles in the chosen flux coordinate; do not triangulate radial profile arrays.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: flux_surface

  • normalization: Channel-specific units and definition apply. _1d is sampled full-f-marker contribution; _df_1d uses delta-f weights, with grid/background contributions for total-f species. gc_density and marker_density divide their sums by bin volume; all other particle-averaged channels divide by the appropriate gyrocenter particle/density weight; psi-space radial fluxes require division by <|grad psi|> or sqrt([species]_grad_psi_sqr_1d) for a per-particle physical radial velocity, energy transport, or momentum transport; multiply by matching gyrocenter density to obtain a flux density.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:404.

xgc.oneddiag.bp: [species]_parallel_mean_en_1d or _df_1d (Diag1dParallelMeanEn)

Native units: eV.

Parallel kinetic energy: <0.5 m v^2>. For parallel temperature, subtract mean flow energy.

  • plot_label: $K_\parallel$ ({name})

  • axes: flux_surface

  • mesh_context: xgc.oneddiag.bp (same ADIOS step): psi_mks (poloidal flux divided by 2*pi, T m^2) and psi (normalized flux) are radial coordinates of flux-surface averages. Interpolate profiles in the chosen flux coordinate; do not triangulate radial profile arrays.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: flux_surface

  • normalization: Channel-specific units and definition apply. _1d is sampled full-f-marker contribution; _df_1d uses delta-f weights, with grid/background contributions for total-f species. gc_density and marker_density divide their sums by bin volume; all other particle-averaged channels divide by the appropriate gyrocenter particle/density weight; psi-space radial fluxes require division by <|grad psi|> or sqrt([species]_grad_psi_sqr_1d) for a per-particle physical radial velocity, energy transport, or momentum transport; multiply by matching gyrocenter density to obtain a flux density.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:420.

xgc.oneddiag.bp: [species]_perp_temperature_1d or _df_1d (Diag1dPerpTemperature)

Native units: eV.

Perpendicular temperature \(<\\mu B>\)

  • plot_label: $T_\perp$ ({name})

  • axes: flux_surface

  • mesh_context: xgc.oneddiag.bp (same ADIOS step): psi_mks (poloidal flux divided by 2*pi, T m^2) and psi (normalized flux) are radial coordinates of flux-surface averages. Interpolate profiles in the chosen flux coordinate; do not triangulate radial profile arrays.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: flux_surface

  • normalization: Channel-specific units and definition apply. _1d is sampled full-f-marker contribution; _df_1d uses delta-f weights, with grid/background contributions for total-f species. gc_density and marker_density divide their sums by bin volume; all other particle-averaged channels divide by the appropriate gyrocenter particle/density weight; psi-space radial fluxes require division by <|grad psi|> or sqrt([species]_grad_psi_sqr_1d) for a per-particle physical radial velocity, energy transport, or momentum transport; multiply by matching gyrocenter density to obtain a flux density.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:425.

xgc.oneddiag.bp: [species]_poloidal_ExB_flow_1d or _df_1d (Diag1dPoloidalExBFlow)

Native units: m/s.

Particle average of poloidal flow by ExB drift.

  • plot_label: $u_{\theta,E\times B}$ ({name})

  • axes: flux_surface

  • mesh_context: xgc.oneddiag.bp (same ADIOS step): psi_mks (poloidal flux divided by 2*pi, T m^2) and psi (normalized flux) are radial coordinates of flux-surface averages. Interpolate profiles in the chosen flux coordinate; do not triangulate radial profile arrays.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: flux_surface

  • normalization: Channel-specific units and definition apply. _1d is sampled full-f-marker contribution; _df_1d uses delta-f weights, with grid/background contributions for total-f species. gc_density and marker_density divide their sums by bin volume; all other particle-averaged channels divide by the appropriate gyrocenter particle/density weight; psi-space radial fluxes require division by <|grad psi|> or sqrt([species]_grad_psi_sqr_1d) for a per-particle physical radial velocity, energy transport, or momentum transport; multiply by matching gyrocenter density to obtain a flux density.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:466.

xgc.oneddiag.bp: [species]_radial_en_flux_1d or _df_1d (Diag1dRadialEnFlux)

Native units: J T m^2 s^-1.

Particle averaged radial energy flux in \(\\psi\) space. \(<E V \\cdot \\nabla \\psi>\) Averaged \(\\nabla \\psi\) is required to get radial flux in real space (J m/s)

  • plot_label: $Q \langle|\nabla \psi|\rangle/n_{gc}$ ({name})

  • axes: flux_surface

  • mesh_context: xgc.oneddiag.bp (same ADIOS step): psi_mks (poloidal flux divided by 2*pi, T m^2) and psi (normalized flux) are radial coordinates of flux-surface averages. Interpolate profiles in the chosen flux coordinate; do not triangulate radial profile arrays.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: flux_surface

  • normalization: Channel-specific units and definition apply. _1d is sampled full-f-marker contribution; _df_1d uses delta-f weights, with grid/background contributions for total-f species. gc_density and marker_density divide their sums by bin volume; all other particle-averaged channels divide by the appropriate gyrocenter particle/density weight; psi-space radial fluxes require division by <|grad psi|> or sqrt([species]_grad_psi_sqr_1d) for a per-particle physical radial velocity, energy transport, or momentum transport; multiply by matching gyrocenter density to obtain a flux density.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:430.

xgc.oneddiag.bp: [species]_radial_en_flux_3db_1d or _df_1d (Diag1dRadialEnFlux3db)

Native units: J T m^2 s^-1.

Particle averaged radial energy flux by dB in psi space.

  • plot_label: $Q_{\delta B_\perp} \langle|\nabla \psi|\rangle/n_{gc}$ ({name})

  • axes: flux_surface

  • mesh_context: xgc.oneddiag.bp (same ADIOS step): psi_mks (poloidal flux divided by 2*pi, T m^2) and psi (normalized flux) are radial coordinates of flux-surface averages. Interpolate profiles in the chosen flux coordinate; do not triangulate radial profile arrays.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: flux_surface

  • normalization: Channel-specific units and definition apply. _1d is sampled full-f-marker contribution; _df_1d uses delta-f weights, with grid/background contributions for total-f species. gc_density and marker_density divide their sums by bin volume; all other particle-averaged channels divide by the appropriate gyrocenter particle/density weight; psi-space radial fluxes require division by <|grad psi|> or sqrt([species]_grad_psi_sqr_1d) for a per-particle physical radial velocity, energy transport, or momentum transport; multiply by matching gyrocenter density to obtain a flux density.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:490.

xgc.oneddiag.bp: [species]_radial_en_flux_ExB_1d or _df_1d (Diag1dRadialEnFluxExB)

Native units: J T m^2 s^-1.

Particle averaged radial energy flux by ExB in psi space.

  • plot_label: $Q_{E\times B} \langle|\nabla \psi|\rangle/n_{gc}$  ({name})

  • axes: flux_surface

  • mesh_context: xgc.oneddiag.bp (same ADIOS step): psi_mks (poloidal flux divided by 2*pi, T m^2) and psi (normalized flux) are radial coordinates of flux-surface averages. Interpolate profiles in the chosen flux coordinate; do not triangulate radial profile arrays.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: flux_surface

  • normalization: Channel-specific units and definition apply. _1d is sampled full-f-marker contribution; _df_1d uses delta-f weights, with grid/background contributions for total-f species. gc_density and marker_density divide their sums by bin volume; all other particle-averaged channels divide by the appropriate gyrocenter particle/density weight; psi-space radial fluxes require division by <|grad psi|> or sqrt([species]_grad_psi_sqr_1d) for a per-particle physical radial velocity, energy transport, or momentum transport; multiply by matching gyrocenter density to obtain a flux density.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:461.

xgc.oneddiag.bp: [species]_radial_flux_1d or _df_1d (Diag1dRadialFlux)

Native units: T m^2 s^-1.

Particle averaged radial drift in \(\\psi\) space. \(<V \\cdot \\nabla \\psi>\) Averaged \(\\nabla \\psi\) is required to get radial flux in real space (m/s)

  • plot_label: $\Gamma \langle|\nabla\psi|\rangle/n_{gc}$ ({name})

  • axes: flux_surface

  • mesh_context: xgc.oneddiag.bp (same ADIOS step): psi_mks (poloidal flux divided by 2*pi, T m^2) and psi (normalized flux) are radial coordinates of flux-surface averages. Interpolate profiles in the chosen flux coordinate; do not triangulate radial profile arrays.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: flux_surface

  • normalization: Channel-specific units and definition apply. _1d is sampled full-f-marker contribution; _df_1d uses delta-f weights, with grid/background contributions for total-f species. gc_density and marker_density divide their sums by bin volume; all other particle-averaged channels divide by the appropriate gyrocenter particle/density weight; psi-space radial fluxes require division by <|grad psi|> or sqrt([species]_grad_psi_sqr_1d) for a per-particle physical radial velocity, energy transport, or momentum transport; multiply by matching gyrocenter density to obtain a flux density.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:414.

xgc.oneddiag.bp: [species]_radial_flux_3db_1d or _df_1d (Diag1dRadialFlux3db)

Native units: T m^2 s^-1.

Particle averaged radial drift by dB in psi space.

  • plot_label: $\Gamma_{\delta B_\perp} \langle|\nabla\psi|\rangle/n_{gc}$ ({name})

  • axes: flux_surface

  • mesh_context: xgc.oneddiag.bp (same ADIOS step): psi_mks (poloidal flux divided by 2*pi, T m^2) and psi (normalized flux) are radial coordinates of flux-surface averages. Interpolate profiles in the chosen flux coordinate; do not triangulate radial profile arrays.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: flux_surface

  • normalization: Channel-specific units and definition apply. _1d is sampled full-f-marker contribution; _df_1d uses delta-f weights, with grid/background contributions for total-f species. gc_density and marker_density divide their sums by bin volume; all other particle-averaged channels divide by the appropriate gyrocenter particle/density weight; psi-space radial fluxes require division by <|grad psi|> or sqrt([species]_grad_psi_sqr_1d) for a per-particle physical radial velocity, energy transport, or momentum transport; multiply by matching gyrocenter density to obtain a flux density.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:485.

xgc.oneddiag.bp: [species]_radial_flux_ExB_1d or _df_1d (Diag1dRadialFluxExB)

Native units: T m^2 s^-1.

Particle averaged radial drift by ExB in psi space.

  • plot_label: $\Gamma_{E\times B} \langle|\nabla\psi|\rangle/n_{gc}$ ({name})

  • axes: flux_surface

  • mesh_context: xgc.oneddiag.bp (same ADIOS step): psi_mks (poloidal flux divided by 2*pi, T m^2) and psi (normalized flux) are radial coordinates of flux-surface averages. Interpolate profiles in the chosen flux coordinate; do not triangulate radial profile arrays.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: flux_surface

  • normalization: Channel-specific units and definition apply. _1d is sampled full-f-marker contribution; _df_1d uses delta-f weights, with grid/background contributions for total-f species. gc_density and marker_density divide their sums by bin volume; all other particle-averaged channels divide by the appropriate gyrocenter particle/density weight; psi-space radial fluxes require division by <|grad psi|> or sqrt([species]_grad_psi_sqr_1d) for a per-particle physical radial velocity, energy transport, or momentum transport; multiply by matching gyrocenter density to obtain a flux density.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:456.

xgc.oneddiag.bp: [species]_radial_mom_flux_1d or _df_1d (Diag1dRadialMomFlux)

Native units: T m^3 s^-2.

Particle-weighted mean of (v_parallel B_phi/|B|) (v_gc dot grad(psi)), without species mass or a major-radius moment arm. Dividing by a surface gradient scale gives per-particle linear-momentum transport divided by mass, in m^2 s^-2; multiply by density and species mass for linear-momentum flux density.

  • plot_label: $\Pi_{\varphi} \langle|\nabla\psi|\rangle/n_{gc}$ ({name})

  • axes: flux_surface

  • mesh_context: xgc.oneddiag.bp (same ADIOS step): psi_mks (poloidal flux divided by 2*pi, T m^2) and psi (normalized flux) are radial coordinates of flux-surface averages. Interpolate profiles in the chosen flux coordinate; do not triangulate radial profile arrays.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: flux_surface

  • normalization: Channel-specific units and definition apply. _1d is sampled full-f-marker contribution; _df_1d uses delta-f weights, with grid/background contributions for total-f species. gc_density and marker_density divide their sums by bin volume; all other particle-averaged channels divide by the appropriate gyrocenter particle/density weight; psi-space radial fluxes require division by <|grad psi|> or sqrt([species]_grad_psi_sqr_1d) for a per-particle physical radial velocity, energy transport, or momentum transport; multiply by matching gyrocenter density to obtain a flux density.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:446.

xgc.oneddiag.bp: [species]_radial_mom_flux_ExB_1d or _df_1d (Diag1dRadialMomFluxExB)

Native units: T m^3 s^-2.

Particle-weighted mean of (v_parallel B_phi/|B|) (v_ExB dot grad(psi)), without species mass or a major-radius moment arm. Dividing by a surface gradient scale gives per-particle linear-momentum transport divided by mass, in m^2 s^-2; multiply by density and species mass for linear-momentum flux density.

  • plot_label: $\Pi_{\varphi,E\times B} \langle|\nabla\psi|\rangle/n_{gc}$ ({name})

  • axes: flux_surface

  • mesh_context: xgc.oneddiag.bp (same ADIOS step): psi_mks (poloidal flux divided by 2*pi, T m^2) and psi (normalized flux) are radial coordinates of flux-surface averages. Interpolate profiles in the chosen flux coordinate; do not triangulate radial profile arrays.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: flux_surface

  • normalization: Channel-specific units and definition apply. _1d is sampled full-f-marker contribution; _df_1d uses delta-f weights, with grid/background contributions for total-f species. gc_density and marker_density divide their sums by bin volume; all other particle-averaged channels divide by the appropriate gyrocenter particle/density weight; psi-space radial fluxes require division by <|grad psi|> or sqrt([species]_grad_psi_sqr_1d) for a per-particle physical radial velocity, energy transport, or momentum transport; multiply by matching gyrocenter density to obtain a flux density.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:436.

xgc.oneddiag.bp: [species]_tor_angular_momentum_1d or _df_1d (Diag1dTorAngularMomentum)

Native units: m^2 s^-1.

Toroidal angular momentum (without mass): <V_tor R>.

  • plot_label: $L_T/m$ ({name})

  • axes: flux_surface

  • mesh_context: xgc.oneddiag.bp (same ADIOS step): psi_mks (poloidal flux divided by 2*pi, T m^2) and psi (normalized flux) are radial coordinates of flux-surface averages. Interpolate profiles in the chosen flux coordinate; do not triangulate radial profile arrays.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: flux_surface

  • normalization: Channel-specific units and definition apply. _1d is sampled full-f-marker contribution; _df_1d uses delta-f weights, with grid/background contributions for total-f species. gc_density and marker_density divide their sums by bin volume; all other particle-averaged channels divide by the appropriate gyrocenter particle/density weight; psi-space radial fluxes require division by <|grad psi|> or sqrt([species]_grad_psi_sqr_1d) for a per-particle physical radial velocity, energy transport, or momentum transport; multiply by matching gyrocenter density to obtain a flux density.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:409.

xgc.oneddiag.bp: [species]_vpar_x_B_1d or _df_1d (Diag1dVparXB)

Native units: T m s^-1.

Particle-weighted average of the scalar product v_parallel |B|, used in bootstrap-current analysis.

  • plot_label: $\langle v_\parallel B\rangle$ ({name})

  • axes: flux_surface

  • mesh_context: xgc.oneddiag.bp (same ADIOS step): psi_mks (poloidal flux divided by 2*pi, T m^2) and psi (normalized flux) are radial coordinates of flux-surface averages. Interpolate profiles in the chosen flux coordinate; do not triangulate radial profile arrays.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: flux_surface

  • normalization: Channel-specific units and definition apply. _1d is sampled full-f-marker contribution; _df_1d uses delta-f weights, with grid/background contributions for total-f species. gc_density and marker_density divide their sums by bin volume; all other particle-averaged channels divide by the appropriate gyrocenter particle/density weight; psi-space radial fluxes require division by <|grad psi|> or sqrt([species]_grad_psi_sqr_1d) for a per-particle physical radial velocity, energy transport, or momentum transport; multiply by matching gyrocenter density to obtain a flux density.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:471.

xgc.oneddiag.bp: cden00_1d (Diag1dCden001d)

Native units: m^-3.

Flux-surface average of the processed axisymmetric Poisson right-hand-side density on psi00, in elementary-charge-normalized units. It includes the enabled boundary masks, offsets, analytic-background perturbations and charge corrections.

  • plot_label: Axisymmetric Poisson RHS density

  • axes: flux_surface

  • mesh_context: xgc.oneddiag.bp (same ADIOS step): psi00 is the uniformly spaced poloidal-flux grid for this quantity. It differs from the psi/psi_mks grid used by the particle-moment channels.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: flux_surface

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:682.

xgc.oneddiag.bp: eden00_1d (Diag1dEden001d)

Native units: m^-3.

Flux-surface averaged electron field-solve number-density channel.

  • plot_label: $\langle \delta n_{e,gc}\rangle$

  • axes: flux_surface

  • mesh_context: xgc.oneddiag.bp (same ADIOS step): psi00 is the uniformly spaced poloidal-flux grid for this quantity. It differs from the psi/psi_mks grid used by the particle-moment channels.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: flux_surface

  • normalization: For kinetic electrons, the channel uses charge factor +1, so multiplying by minus the positive elementary charge converts its density contribution from m^-3 to C/m^3. For adiabatic electrons it is initialized from the full-f ion reference profile rather than a kinetic-electron marker deposition.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:675.

xgc.oneddiag.bp: gsamples (Diag1dGsamples)

Native units: Positive integer.

Number of samples (for evenly spaced psi00)

  • plot_label: $N_{\psi,00}$

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:630.

xgc.oneddiag.bp: gstep (Diag1dGstep)

Native units: Non-negative integer.

XGC global simulation time step for this 1D diagnostic output slice.

  • plot_label: Global time step

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:640.

xgc.oneddiag.bp: iden00_1d (Diag1dIden001d)

Native units: m^-3.

Flux-surface average of the ion field-solve density contribution, summed over ion species with charge-number weighting.

  • plot_label: $\langle \overline{\delta n_{i,gc}}\rangle$

  • axes: flux_surface

  • mesh_context: xgc.oneddiag.bp (same ADIOS step): psi00 is the uniformly spaced poloidal-flux grid for this quantity. It differs from the psi/psi_mks grid used by the particle-moment channels.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: flux_surface

  • normalization: This is the ion charge-number-weighted density contribution in m^-3 after toroidal/flux-surface reduction and interpolation onto psi00. Multiply by the positive elementary charge for C/m^3; it is not a per-species ion number density.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:667.

xgc.oneddiag.bp: pot00_1d (Diag1dPot001d)

Native units: V.

Flux averaged electrostatic potential

  • plot_label: $\langle\phi\rangle$

  • axes: flux_surface

  • mesh_context: xgc.oneddiag.bp (same ADIOS step): psi00 is the uniformly spaced poloidal-flux grid for this quantity. It differs from the psi/psi_mks grid used by the particle-moment channels.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: flux_surface

  • imas_candidates: core_profiles/profiles_1d(:)/phi_potential

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_medium

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:691.

xgc.oneddiag.bp: psi (Diag1dPsi)

Native units: dimensionless.

Normalized poloidal magnetic flux at each output on each flux surface

  • plot_label: $\psi_N$

  • axes: flux_surface

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:655.

xgc.oneddiag.bp: psi00 (Diag1dPsi00)

Native units: T m^2.

Unnormalized poloidal magnetic flux at each output on each flux surface

  • plot_label: $\psi_{00}/(2\pi)$ (uniform grid)

  • axes: flux_surface

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:662.

xgc.oneddiag.bp: psi_mks (Diag1dPsiMks)

Native units: T m^2.

Unnormalized poloidal magnetic flux at each output on each flux surface

  • plot_label: $\psi/(2\pi)$

  • axes: flux_surface

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:650.

xgc.oneddiag.bp: samples (Diag1dSamples)

Native units: Positive integer.

Number of samples (i.e. number of flux surfaces)

  • plot_label: $N_\psi$

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:623.

xgc.oneddiag.bp: time (Diag1dTime)

Native units: s.

Simulation time at each output in seconds

  • plot_label: $t$

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions. Clock convention: in the regular main loop gstep is incremented before diagnostic calls, but sml.time is incremented afterward. A checkpoint written later at the same gstep therefore records time larger by one physical timestep. Use the written time, global step and product/integration stage together; equal gstep alone does not establish simultaneous diagnostic and checkpoint states.

  • imas_candidates: core_profiles/time

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_high

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:635.

xgc.oneddiag.bp: tindex (Diag1dTindex)

Native units: Positive integer.

Index of diagnostic output (gstep/diagnostic period). Values: 1, 2, …, # outputs

  • plot_label: Output slice index

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:645.

xgc.particle.*.bp: gstep (DiagParticleGstep)

Native units: Non-negative integer.

XGC global simulation time step for this particle diagnostic output slice.

  • plot_label: Global time step

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/diagnostics/diag_particle.cpp:58.

xgc.particle.*.bp: sp_name + “gid” (DiagParticleSpNameGid)

Native units: Positive integer.

Global ID of the markers in the sample.

  • plot_label: Sampled marker GID ({name})

  • axes: particle

  • mesh_context: Particle sample, not a mesh field. The same species’ phase array contains row-wise cylindrical R (m), Z (m), phi (rad) in columns 0,1,2; use gid to identify rows/particles across output slices. MPI repartitioning can change row order. Visualize as points or perform a stated statistical/binning procedure, not triangle interpolation. Other phase columns and units are specified in description.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: particle

Source: XGC_core/cpp/diagnostics/diag_particle.cpp:139.

xgc.particle.*.bp: sp_name + “nphase” (DiagParticleSpNameNphase)

Native units: Positive integer.

Number of particle phase variables.

  • plot_label: Particle phase-component count ({name})

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

Source: XGC_core/cpp/diagnostics/diag_particle.cpp:134.

xgc.particle.*.bp: sp_name + “num_total” (DiagParticleSpNameNumTotal)

Native units: Non-negative integer.

Number of markers in the sample.

  • plot_label: Sampled marker count ({name})

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/diagnostics/diag_particle.cpp:127.

xgc.particle.*.bp: sp_name + “phase” (DiagParticleSpNamePhase)

Native units: [m, m, rad, m, dimensionless, dimensionless, J/T, dimensionless, m^-3 eV^(-3/2), optional: m^-3 eV^(-3/2)].

Selected species particle phase array with columns [R, Z, phi, rho, w1, w2, mu, w0, g], where \(R\), \(Z\), and \(\\phi\) are the cylindrical coordinates of the particle, \(\\rho\) is the parallel gyroradius (proxy for \(v_{\\parallel}\)), \(w_1\) is the delta-f weight, \(w_2\) is a weight used in the weight evolution equation, \(\\mu\) is the magnetic moment, \(w_0\) is the full-f weight (from importance sampling), and \(g\) is the phase-space marker density. Particles are selected when their global particle ID is positive and divisible by diag_particle_mod. If PTL_G2 is enabled, an additional g2 marker-density column is appended.

  • plot_label: Sampled marker phase space coord. ({name})

  • axes: particle,phase_component

  • mesh_context: Particle sample, not a mesh field. The same species’ phase array contains row-wise cylindrical R (m), Z (m), phi (rad) in columns 0,1,2; use gid to identify rows/particles across output slices. MPI repartitioning can change row order. Visualize as points or perform a stated statistical/binning procedure, not triangle interpolation. Other phase columns and units are specified in description.

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: particle

  • normalization: The g column is marker density in XGC velocity-space normalization, with units m^-3 eV^(-3/2); f=w0*g. For a Maxwellian this normalization is proportional to n*T_eV^(-3/2)*exp(-E_eV/T_eV), with the SI velocity normalization factor (mass/(2*pi*EV_2_J))^(3/2) removed. This is distinct from the spatial triangle marker_den in xgc.loading.*.bp.

Source: XGC_core/cpp/diagnostics/diag_particle.cpp:144.

xgc.particle.*.bp: time (DiagParticleTime)

Native units: s.

Simulation time of step in seconds.

  • plot_label: $t$

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions. Clock convention: in the regular main loop gstep is incremented before diagnostic calls, but sml.time is incremented afterward. A checkpoint written later at the same gstep therefore records time larger by one physical timestep. Use the written time, global step and product/integration stage together; equal gstep alone does not establish simultaneous diagnostic and checkpoint states.

  • imas_candidates: plasma_profiles/time

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_high

Source: XGC_core/cpp/diagnostics/diag_particle.cpp:71.

xgc.particle.*.bp: tindex (DiagParticleTindex)

Native units: Positive integer.

Diagnostic cadence index computed as integer gstep divided by the configured diagnostic period. This index is not the zero-based ADIOS step and need not start at 1 in a restarted or selected product.

  • plot_label: Output slice index

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/diagnostics/diag_particle.cpp:63.

xgc.particle_stream.bp: e_E (TracerDiagEE)

Native units: J.

Kinetic energy 0.5*mass*vpar^2 + mu*|B| of the tracked particle.

  • plot_label: $K_{e}$

  • axes: particle

  • mesh_context: Same xgc.particle_stream.bp step: [i/e]_R,[i/e]_Z (m), [i/e]_phi (rad) give row-wise point coordinates; [i/e]_gid joins a marker across writes. Rows can move with MPI repartitioning. Point samples have no triangular basis; binning requires a stated statistical method.

  • species_context: Prefix i_ denotes the ion pool and e_ the electron pool. The legacy writer combines all non-electron species into i_; it does not preserve separate ion-species identity.

  • time_context: Instantaneous particle sample at same-product gstep. This legacy ESC_PTL writer has no time or tindex variables. A physical-time difference can be obtained from step differences times xgc.units.bp:sml_dt only for a constant timestep. The tracked-particle collection path is currently disabled by a fatal error for removed Fortran field calls; no valid field-derived particle values are produced on that path.

  • centering: particle

Source: XGC_core/cpp/diagnostics/tracer_diag.cpp:129.

xgc.particle_stream.bp: e_R (TracerDiagER)

Native units: m.

Cylindrical major radius of the tracked particle.

  • plot_label: $R_{e}$

  • axes: particle

  • mesh_context: Same xgc.particle_stream.bp step: [i/e]_R,[i/e]_Z (m), [i/e]_phi (rad) give row-wise point coordinates; [i/e]_gid joins a marker across writes. Rows can move with MPI repartitioning. Point samples have no triangular basis; binning requires a stated statistical method.

  • species_context: Prefix i_ denotes the ion pool and e_ the electron pool. The legacy writer combines all non-electron species into i_; it does not preserve separate ion-species identity.

  • time_context: Instantaneous particle sample at same-product gstep. This legacy ESC_PTL writer has no time or tindex variables. A physical-time difference can be obtained from step differences times xgc.units.bp:sml_dt only for a constant timestep. The tracked-particle collection path is currently disabled by a fatal error for removed Fortran field calls; no valid field-derived particle values are produced on that path.

  • centering: particle

Source: XGC_core/cpp/diagnostics/tracer_diag.cpp:137.

xgc.particle_stream.bp: e_Z (TracerDiagEZ)

Native units: m.

Vertical position of the tracked particle.

  • plot_label: $Z_{e}$

  • axes: particle

  • mesh_context: Same xgc.particle_stream.bp step: [i/e]_R,[i/e]_Z (m), [i/e]_phi (rad) give row-wise point coordinates; [i/e]_gid joins a marker across writes. Rows can move with MPI repartitioning. Point samples have no triangular basis; binning requires a stated statistical method.

  • species_context: Prefix i_ denotes the ion pool and e_ the electron pool. The legacy writer combines all non-electron species into i_; it does not preserve separate ion-species identity.

  • time_context: Instantaneous particle sample at same-product gstep. This legacy ESC_PTL writer has no time or tindex variables. A physical-time difference can be obtained from step differences times xgc.units.bp:sml_dt only for a constant timestep. The tracked-particle collection path is currently disabled by a fatal error for removed Fortran field calls; no valid field-derived particle values are produced on that path.

  • centering: particle

Source: XGC_core/cpp/diagnostics/tracer_diag.cpp:145.

xgc.particle_stream.bp: e_gid (TracerDiagEGid)

Native units: Integer index.

Global marker ID; join particle samples by ID rather than row position.

  • plot_label: Electron marker ID

  • axes: particle

  • mesh_context: Same xgc.particle_stream.bp step: [i/e]_R,[i/e]_Z (m), [i/e]_phi (rad) give row-wise point coordinates; [i/e]_gid joins a marker across writes. Rows can move with MPI repartitioning. Point samples have no triangular basis; binning requires a stated statistical method.

  • species_context: Prefix i_ denotes the ion pool and e_ the electron pool. The legacy writer combines all non-electron species into i_; it does not preserve separate ion-species identity.

  • time_context: Instantaneous particle sample at same-product gstep. This legacy ESC_PTL writer has no time or tindex variables. A physical-time difference can be obtained from step differences times xgc.units.bp:sml_dt only for a constant timestep. The tracked-particle collection path is currently disabled by a fatal error for removed Fortran field calls; no valid field-derived particle values are produced on that path.

  • centering: particle

Source: XGC_core/cpp/diagnostics/tracer_diag.cpp:105.

xgc.particle_stream.bp: e_phi (TracerDiagEPhi)

Native units: rad.

Cylindrical toroidal angle of the tracked particle.

  • plot_label: $\varphi_{e}$

  • axes: particle

  • mesh_context: Same xgc.particle_stream.bp step: [i/e]_R,[i/e]_Z (m), [i/e]_phi (rad) give row-wise point coordinates; [i/e]_gid joins a marker across writes. Rows can move with MPI repartitioning. Point samples have no triangular basis; binning requires a stated statistical method.

  • species_context: Prefix i_ denotes the ion pool and e_ the electron pool. The legacy writer combines all non-electron species into i_; it does not preserve separate ion-species identity.

  • time_context: Instantaneous particle sample at same-product gstep. This legacy ESC_PTL writer has no time or tindex variables. A physical-time difference can be obtained from step differences times xgc.units.bp:sml_dt only for a constant timestep. The tracked-particle collection path is currently disabled by a fatal error for removed Fortran field calls; no valid field-derived particle values are produced on that path.

  • centering: particle

Source: XGC_core/cpp/diagnostics/tracer_diag.cpp:153.

xgc.particle_stream.bp: e_psi (TracerDiagEPsi)

Native units: T m^2.

Poloidal magnetic flux divided by 2*pi at the tracked particle.

  • plot_label: $\psi_{e}/(2\pi)$

  • axes: particle

  • mesh_context: Same xgc.particle_stream.bp step: [i/e]_R,[i/e]_Z (m), [i/e]_phi (rad) give row-wise point coordinates; [i/e]_gid joins a marker across writes. Rows can move with MPI repartitioning. Point samples have no triangular basis; binning requires a stated statistical method.

  • species_context: Prefix i_ denotes the ion pool and e_ the electron pool. The legacy writer combines all non-electron species into i_; it does not preserve separate ion-species identity.

  • time_context: Instantaneous particle sample at same-product gstep. This legacy ESC_PTL writer has no time or tindex variables. A physical-time difference can be obtained from step differences times xgc.units.bp:sml_dt only for a constant timestep. The tracked-particle collection path is currently disabled by a fatal error for removed Fortran field calls; no valid field-derived particle values are produced on that path.

  • centering: particle

Source: XGC_core/cpp/diagnostics/tracer_diag.cpp:113.

xgc.particle_stream.bp: e_psi_c (TracerDiagEPsiC)

Native units: T m^2.

Canonical flux-like coordinate psi + sign(-Bt)*rho*Bphi*R for the tracked particle.

  • plot_label: Electron canonical flux coordinate

  • axes: particle

  • mesh_context: Same xgc.particle_stream.bp step: [i/e]_R,[i/e]_Z (m), [i/e]_phi (rad) give row-wise point coordinates; [i/e]_gid joins a marker across writes. Rows can move with MPI repartitioning. Point samples have no triangular basis; binning requires a stated statistical method.

  • species_context: Prefix i_ denotes the ion pool and e_ the electron pool. The legacy writer combines all non-electron species into i_; it does not preserve separate ion-species identity.

  • time_context: Instantaneous particle sample at same-product gstep. This legacy ESC_PTL writer has no time or tindex variables. A physical-time difference can be obtained from step differences times xgc.units.bp:sml_dt only for a constant timestep. The tracked-particle collection path is currently disabled by a fatal error for removed Fortran field calls; no valid field-derived particle values are produced on that path.

  • centering: particle

Source: XGC_core/cpp/diagnostics/tracer_diag.cpp:121.

xgc.particle_stream.bp: e_vpar (TracerDiagEVpar)

Native units: m s^-1.

Velocity parallel to the magnetic field of the tracked particle.

  • plot_label: $v_{\parallel,e}$

  • axes: particle

  • mesh_context: Same xgc.particle_stream.bp step: [i/e]_R,[i/e]_Z (m), [i/e]_phi (rad) give row-wise point coordinates; [i/e]_gid joins a marker across writes. Rows can move with MPI repartitioning. Point samples have no triangular basis; binning requires a stated statistical method.

  • species_context: Prefix i_ denotes the ion pool and e_ the electron pool. The legacy writer combines all non-electron species into i_; it does not preserve separate ion-species identity.

  • time_context: Instantaneous particle sample at same-product gstep. This legacy ESC_PTL writer has no time or tindex variables. A physical-time difference can be obtained from step differences times xgc.units.bp:sml_dt only for a constant timestep. The tracked-particle collection path is currently disabled by a fatal error for removed Fortran field calls; no valid field-derived particle values are produced on that path.

  • centering: particle

Source: XGC_core/cpp/diagnostics/tracer_diag.cpp:161.

xgc.particle_stream.bp: e_w0 (TracerDiagEW0)

Native units: dimensionless.

Constant particle weight; -1 is written as a divertor-hit sentinel.

  • plot_label: $w_{0,e}$ (constant)

  • axes: particle

  • mesh_context: Same xgc.particle_stream.bp step: [i/e]_R,[i/e]_Z (m), [i/e]_phi (rad) give row-wise point coordinates; [i/e]_gid joins a marker across writes. Rows can move with MPI repartitioning. Point samples have no triangular basis; binning requires a stated statistical method.

  • species_context: Prefix i_ denotes the ion pool and e_ the electron pool. The legacy writer combines all non-electron species into i_; it does not preserve separate ion-species identity.

  • time_context: Instantaneous particle sample at same-product gstep. This legacy ESC_PTL writer has no time or tindex variables. A physical-time difference can be obtained from step differences times xgc.units.bp:sml_dt only for a constant timestep. The tracked-particle collection path is currently disabled by a fatal error for removed Fortran field calls; no valid field-derived particle values are produced on that path.

  • centering: particle

Source: XGC_core/cpp/diagnostics/tracer_diag.cpp:169.

xgc.particle_stream.bp: e_w1 (TracerDiagEW1)

Native units: dimensionless.

Particle delta-f weight component w1.

  • plot_label: $w_{1,e}$ (delta-f)

  • axes: particle

  • mesh_context: Same xgc.particle_stream.bp step: [i/e]_R,[i/e]_Z (m), [i/e]_phi (rad) give row-wise point coordinates; [i/e]_gid joins a marker across writes. Rows can move with MPI repartitioning. Point samples have no triangular basis; binning requires a stated statistical method.

  • species_context: Prefix i_ denotes the ion pool and e_ the electron pool. The legacy writer combines all non-electron species into i_; it does not preserve separate ion-species identity.

  • time_context: Instantaneous particle sample at same-product gstep. This legacy ESC_PTL writer has no time or tindex variables. A physical-time difference can be obtained from step differences times xgc.units.bp:sml_dt only for a constant timestep. The tracked-particle collection path is currently disabled by a fatal error for removed Fortran field calls; no valid field-derived particle values are produced on that path.

  • centering: particle

Source: XGC_core/cpp/diagnostics/tracer_diag.cpp:177.

xgc.particle_stream.bp: e_w2 (TracerDiagEW2)

Native units: dimensionless.

Particle auxiliary weight w2=1-f0/(w0*g), copied from the particle phase array.

  • plot_label: $w_{2,e}$ (auxiliary)

  • axes: particle

  • mesh_context: Same xgc.particle_stream.bp step: [i/e]_R,[i/e]_Z (m), [i/e]_phi (rad) give row-wise point coordinates; [i/e]_gid joins a marker across writes. Rows can move with MPI repartitioning. Point samples have no triangular basis; binning requires a stated statistical method.

  • species_context: Prefix i_ denotes the ion pool and e_ the electron pool. The legacy writer combines all non-electron species into i_; it does not preserve separate ion-species identity.

  • time_context: Instantaneous particle sample at same-product gstep. This legacy ESC_PTL writer has no time or tindex variables. A physical-time difference can be obtained from step differences times xgc.units.bp:sml_dt only for a constant timestep. The tracked-particle collection path is currently disabled by a fatal error for removed Fortran field calls; no valid field-derived particle values are produced on that path.

  • centering: particle

Source: XGC_core/cpp/diagnostics/tracer_diag.cpp:185.

xgc.particle_stream.bp: gstep (TracerDiagGstep)

Native units: Non-negative integer.

Global integration step of this particle-stream sample.

  • plot_label: Global time step

  • time_context: Instantaneous particle sample at same-product gstep. This legacy ESC_PTL writer has no time or tindex variables. A physical-time difference can be obtained from step differences times xgc.units.bp:sml_dt only for a constant timestep. The tracked-particle collection path is currently disabled by a fatal error for removed Fortran field calls; no valid field-derived particle values are produced on that path.

Source: XGC_core/cpp/diagnostics/tracer_diag.cpp:607.

xgc.particle_stream.bp: i_E (TracerDiagIE)

Native units: J.

Kinetic energy 0.5*mass*vpar^2 + mu*|B| of the tracked particle.

  • plot_label: $K_{i}$

  • axes: particle

  • mesh_context: Same xgc.particle_stream.bp step: [i/e]_R,[i/e]_Z (m), [i/e]_phi (rad) give row-wise point coordinates; [i/e]_gid joins a marker across writes. Rows can move with MPI repartitioning. Point samples have no triangular basis; binning requires a stated statistical method.

  • species_context: Prefix i_ denotes the ion pool and e_ the electron pool. The legacy writer combines all non-electron species into i_; it does not preserve separate ion-species identity.

  • time_context: Instantaneous particle sample at same-product gstep. This legacy ESC_PTL writer has no time or tindex variables. A physical-time difference can be obtained from step differences times xgc.units.bp:sml_dt only for a constant timestep. The tracked-particle collection path is currently disabled by a fatal error for removed Fortran field calls; no valid field-derived particle values are produced on that path.

  • centering: particle

Source: XGC_core/cpp/diagnostics/tracer_diag.cpp:38.

xgc.particle_stream.bp: i_R (TracerDiagIR)

Native units: m.

Cylindrical major radius of the tracked particle.

  • plot_label: $R_{i}$

  • axes: particle

  • mesh_context: Same xgc.particle_stream.bp step: [i/e]_R,[i/e]_Z (m), [i/e]_phi (rad) give row-wise point coordinates; [i/e]_gid joins a marker across writes. Rows can move with MPI repartitioning. Point samples have no triangular basis; binning requires a stated statistical method.

  • species_context: Prefix i_ denotes the ion pool and e_ the electron pool. The legacy writer combines all non-electron species into i_; it does not preserve separate ion-species identity.

  • time_context: Instantaneous particle sample at same-product gstep. This legacy ESC_PTL writer has no time or tindex variables. A physical-time difference can be obtained from step differences times xgc.units.bp:sml_dt only for a constant timestep. The tracked-particle collection path is currently disabled by a fatal error for removed Fortran field calls; no valid field-derived particle values are produced on that path.

  • centering: particle

Source: XGC_core/cpp/diagnostics/tracer_diag.cpp:46.

xgc.particle_stream.bp: i_Z (TracerDiagIZ)

Native units: m.

Vertical position of the tracked particle.

  • plot_label: $Z_{i}$

  • axes: particle

  • mesh_context: Same xgc.particle_stream.bp step: [i/e]_R,[i/e]_Z (m), [i/e]_phi (rad) give row-wise point coordinates; [i/e]_gid joins a marker across writes. Rows can move with MPI repartitioning. Point samples have no triangular basis; binning requires a stated statistical method.

  • species_context: Prefix i_ denotes the ion pool and e_ the electron pool. The legacy writer combines all non-electron species into i_; it does not preserve separate ion-species identity.

  • time_context: Instantaneous particle sample at same-product gstep. This legacy ESC_PTL writer has no time or tindex variables. A physical-time difference can be obtained from step differences times xgc.units.bp:sml_dt only for a constant timestep. The tracked-particle collection path is currently disabled by a fatal error for removed Fortran field calls; no valid field-derived particle values are produced on that path.

  • centering: particle

Source: XGC_core/cpp/diagnostics/tracer_diag.cpp:54.

xgc.particle_stream.bp: i_gid (TracerDiagIGid)

Native units: Integer index.

Global marker ID; join particle samples by ID rather than row position.

  • plot_label: Ion marker ID

  • axes: particle

  • mesh_context: Same xgc.particle_stream.bp step: [i/e]_R,[i/e]_Z (m), [i/e]_phi (rad) give row-wise point coordinates; [i/e]_gid joins a marker across writes. Rows can move with MPI repartitioning. Point samples have no triangular basis; binning requires a stated statistical method.

  • species_context: Prefix i_ denotes the ion pool and e_ the electron pool. The legacy writer combines all non-electron species into i_; it does not preserve separate ion-species identity.

  • time_context: Instantaneous particle sample at same-product gstep. This legacy ESC_PTL writer has no time or tindex variables. A physical-time difference can be obtained from step differences times xgc.units.bp:sml_dt only for a constant timestep. The tracked-particle collection path is currently disabled by a fatal error for removed Fortran field calls; no valid field-derived particle values are produced on that path.

  • centering: particle

Source: XGC_core/cpp/diagnostics/tracer_diag.cpp:14.

xgc.particle_stream.bp: i_phi (TracerDiagIPhi)

Native units: rad.

Cylindrical toroidal angle of the tracked particle.

  • plot_label: $\varphi_{i}$

  • axes: particle

  • mesh_context: Same xgc.particle_stream.bp step: [i/e]_R,[i/e]_Z (m), [i/e]_phi (rad) give row-wise point coordinates; [i/e]_gid joins a marker across writes. Rows can move with MPI repartitioning. Point samples have no triangular basis; binning requires a stated statistical method.

  • species_context: Prefix i_ denotes the ion pool and e_ the electron pool. The legacy writer combines all non-electron species into i_; it does not preserve separate ion-species identity.

  • time_context: Instantaneous particle sample at same-product gstep. This legacy ESC_PTL writer has no time or tindex variables. A physical-time difference can be obtained from step differences times xgc.units.bp:sml_dt only for a constant timestep. The tracked-particle collection path is currently disabled by a fatal error for removed Fortran field calls; no valid field-derived particle values are produced on that path.

  • centering: particle

Source: XGC_core/cpp/diagnostics/tracer_diag.cpp:62.

xgc.particle_stream.bp: i_psi (TracerDiagIPsi)

Native units: T m^2.

Poloidal magnetic flux divided by 2*pi at the tracked particle.

  • plot_label: $\psi_{i}/(2\pi)$

  • axes: particle

  • mesh_context: Same xgc.particle_stream.bp step: [i/e]_R,[i/e]_Z (m), [i/e]_phi (rad) give row-wise point coordinates; [i/e]_gid joins a marker across writes. Rows can move with MPI repartitioning. Point samples have no triangular basis; binning requires a stated statistical method.

  • species_context: Prefix i_ denotes the ion pool and e_ the electron pool. The legacy writer combines all non-electron species into i_; it does not preserve separate ion-species identity.

  • time_context: Instantaneous particle sample at same-product gstep. This legacy ESC_PTL writer has no time or tindex variables. A physical-time difference can be obtained from step differences times xgc.units.bp:sml_dt only for a constant timestep. The tracked-particle collection path is currently disabled by a fatal error for removed Fortran field calls; no valid field-derived particle values are produced on that path.

  • centering: particle

Source: XGC_core/cpp/diagnostics/tracer_diag.cpp:22.

xgc.particle_stream.bp: i_psi_c (TracerDiagIPsiC)

Native units: T m^2.

Canonical flux-like coordinate psi + sign(-Bt)*rho*Bphi*R for the tracked particle.

  • plot_label: Ion canonical flux coordinate

  • axes: particle

  • mesh_context: Same xgc.particle_stream.bp step: [i/e]_R,[i/e]_Z (m), [i/e]_phi (rad) give row-wise point coordinates; [i/e]_gid joins a marker across writes. Rows can move with MPI repartitioning. Point samples have no triangular basis; binning requires a stated statistical method.

  • species_context: Prefix i_ denotes the ion pool and e_ the electron pool. The legacy writer combines all non-electron species into i_; it does not preserve separate ion-species identity.

  • time_context: Instantaneous particle sample at same-product gstep. This legacy ESC_PTL writer has no time or tindex variables. A physical-time difference can be obtained from step differences times xgc.units.bp:sml_dt only for a constant timestep. The tracked-particle collection path is currently disabled by a fatal error for removed Fortran field calls; no valid field-derived particle values are produced on that path.

  • centering: particle

Source: XGC_core/cpp/diagnostics/tracer_diag.cpp:30.

xgc.particle_stream.bp: i_vpar (TracerDiagIVpar)

Native units: m s^-1.

Velocity parallel to the magnetic field of the tracked particle.

  • plot_label: $v_{\parallel,i}$

  • axes: particle

  • mesh_context: Same xgc.particle_stream.bp step: [i/e]_R,[i/e]_Z (m), [i/e]_phi (rad) give row-wise point coordinates; [i/e]_gid joins a marker across writes. Rows can move with MPI repartitioning. Point samples have no triangular basis; binning requires a stated statistical method.

  • species_context: Prefix i_ denotes the ion pool and e_ the electron pool. The legacy writer combines all non-electron species into i_; it does not preserve separate ion-species identity.

  • time_context: Instantaneous particle sample at same-product gstep. This legacy ESC_PTL writer has no time or tindex variables. A physical-time difference can be obtained from step differences times xgc.units.bp:sml_dt only for a constant timestep. The tracked-particle collection path is currently disabled by a fatal error for removed Fortran field calls; no valid field-derived particle values are produced on that path.

  • centering: particle

Source: XGC_core/cpp/diagnostics/tracer_diag.cpp:70.

xgc.particle_stream.bp: i_w0 (TracerDiagIW0)

Native units: dimensionless.

Constant particle weight; -1 is written as a divertor-hit sentinel.

  • plot_label: $w_{0,i}$ (constant)

  • axes: particle

  • mesh_context: Same xgc.particle_stream.bp step: [i/e]_R,[i/e]_Z (m), [i/e]_phi (rad) give row-wise point coordinates; [i/e]_gid joins a marker across writes. Rows can move with MPI repartitioning. Point samples have no triangular basis; binning requires a stated statistical method.

  • species_context: Prefix i_ denotes the ion pool and e_ the electron pool. The legacy writer combines all non-electron species into i_; it does not preserve separate ion-species identity.

  • time_context: Instantaneous particle sample at same-product gstep. This legacy ESC_PTL writer has no time or tindex variables. A physical-time difference can be obtained from step differences times xgc.units.bp:sml_dt only for a constant timestep. The tracked-particle collection path is currently disabled by a fatal error for removed Fortran field calls; no valid field-derived particle values are produced on that path.

  • centering: particle

Source: XGC_core/cpp/diagnostics/tracer_diag.cpp:78.

xgc.particle_stream.bp: i_w1 (TracerDiagIW1)

Native units: dimensionless.

Particle delta-f weight component w1.

  • plot_label: $w_{1,i}$ (delta-f)

  • axes: particle

  • mesh_context: Same xgc.particle_stream.bp step: [i/e]_R,[i/e]_Z (m), [i/e]_phi (rad) give row-wise point coordinates; [i/e]_gid joins a marker across writes. Rows can move with MPI repartitioning. Point samples have no triangular basis; binning requires a stated statistical method.

  • species_context: Prefix i_ denotes the ion pool and e_ the electron pool. The legacy writer combines all non-electron species into i_; it does not preserve separate ion-species identity.

  • time_context: Instantaneous particle sample at same-product gstep. This legacy ESC_PTL writer has no time or tindex variables. A physical-time difference can be obtained from step differences times xgc.units.bp:sml_dt only for a constant timestep. The tracked-particle collection path is currently disabled by a fatal error for removed Fortran field calls; no valid field-derived particle values are produced on that path.

  • centering: particle

Source: XGC_core/cpp/diagnostics/tracer_diag.cpp:86.

xgc.particle_stream.bp: i_w2 (TracerDiagIW2)

Native units: dimensionless.

Particle auxiliary weight w2=1-f0/(w0*g), copied from the particle phase array.

  • plot_label: $w_{2,i}$ (auxiliary)

  • axes: particle

  • mesh_context: Same xgc.particle_stream.bp step: [i/e]_R,[i/e]_Z (m), [i/e]_phi (rad) give row-wise point coordinates; [i/e]_gid joins a marker across writes. Rows can move with MPI repartitioning. Point samples have no triangular basis; binning requires a stated statistical method.

  • species_context: Prefix i_ denotes the ion pool and e_ the electron pool. The legacy writer combines all non-electron species into i_; it does not preserve separate ion-species identity.

  • time_context: Instantaneous particle sample at same-product gstep. This legacy ESC_PTL writer has no time or tindex variables. A physical-time difference can be obtained from step differences times xgc.units.bp:sml_dt only for a constant timestep. The tracked-particle collection path is currently disabled by a fatal error for removed Fortran field calls; no valid field-derived particle values are produced on that path.

  • centering: particle

Source: XGC_core/cpp/diagnostics/tracer_diag.cpp:94.

xgc.rmp_deltaa.bp: deltaa (PerturbedBFieldDeltaa)

Native units: T m.

RMP perturbation of the parallel vector potential.

  • plot_label: RMP parallel vector potential perturbation

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz, nd_connect_list for axisymmetric geometry; rplane_rz, rplane_nd_connect_list for stellarator right-plane geometry. Connectivity is zero-based. Node-centered field with continuous linear barycentric basis functions on each triangle.

  • coordinate_context: Static vacuum RMP vector-potential-related field sampled on right-plane mesh vertices, [toroidal_plane,mesh_node]. Toroidal angle=(plane_index+1)*xgc.mesh.bp:delta_phi. Same-file ntor lists input harmonics, not an array dimension or real/imaginary components.

  • centering: node

  • normalization: Parallel projection A_parallel=A_vector dot B/|B|, in the same physical T m units as the electromagnetic As field. No additional length factor is applied by the writer.

Source: XGC_core/cpp/solvers/perturbed_B_field.cpp:358.

xgc.rmp_deltaa.bp: n_n (PerturbedBFieldNN)

Native units: Positive integer.

Number of mesh nodes per plane.

  • plot_label: $N_n$

Source: XGC_core/cpp/solvers/perturbed_B_field.cpp:332.

xgc.rmp_deltaa.bp: n_phi (PerturbedBFieldNPhi)

Native units: Positive integer.

Number of toroidal mesh planes \(N_{\\mathrm{planes}}\).

  • plot_label: $N_\varphi$

Source: XGC_core/cpp/solvers/perturbed_B_field.cpp:337.

xgc.rmp_deltaa.bp: ntor (PerturbedBFieldNtor)

Native units: integer mode number.

Toroidal mode numbers for RMP simulations.

  • plot_label: Toroidal mode number

Source: XGC_core/cpp/solvers/perturbed_B_field.cpp:349.

xgc.rmp_deltaa.bp: num_ntor (PerturbedBFieldNumNtor)

Native units: Positive integer.

Number of configured input toroidal harmonics in the static vacuum RMP field (ptb_3db_num_ntor).

  • plot_label: RMP toroidal mode count

Source: XGC_core/cpp/solvers/perturbed_B_field.cpp:342.

xgc.rmp_residual.bp: Apar_mn (PerturbedBFieldAparMn)

Native units: T m.

Parallel vector-potential Fourier amplitudes by toroidal mode, poloidal mode, and flux surface.

  • plot_label: $A_\parallel(m,n)$

  • axes: fourier_n,fourier_m,flux_surface

  • mesh_context: Same xgc.rmp_residual.bp: mpol labels the poloidal mode axis and psi_res gives unnormalized flux labels on the radial axis. Leading toroidal DFT slots have length nphi from the simulation, not generally len(ntor). xgc.mesh.bp:theta,psi_surf,surf_idx,surf_len describes the projection geometry.

  • coordinate_context: Leading axis stores toroidal DFT slots, with zero-based slot k corresponding to n=k for the sampled toroidal-period coordinate; the underlying DFT uses phi_i=2*pi*i/nphi. It spans nphi slots, whereas the same-file ntor array lists only configured input RMP harmonics and is not a general coordinate array for this axis. The second axis follows mpol. Apply the simulation wedge convention when translating to full-torus toroidal mode numbers. Stored Fourier amplitudes are nonnegative sqrt(cosine_coefficient^2+sine_coefficient^2); phase is discarded and cannot be reconstructed from these amplitudes alone.

  • time_context: Same step: time (s), gstep, tindex and ipc (Runge-Kutta stage). Multiple stages may have identical tindex/gstep; preserve ipc and ADIOS position. tindex counts only EM steps after the RMP start; it is not a uniform physical clock.

Source: XGC_core/cpp/solvers/perturbed_B_field.cpp:488.

xgc.rmp_residual.bp: dApar (PerturbedBFieldDApar)

Native units: T m.

Perturbed parallel vector-potential correction on mesh nodes and toroidal planes.

  • plot_label: $\Delta A_\parallel$

  • axes: toroidal_plane,mesh_node

  • mesh_context: xgc.mesh.bp: rz (R,Z in m), nd_connect_list (zero-based triangle vertices). Values are at mesh vertices; use continuous piecewise-linear barycentric basis functions within each unstructured triangle, not a rectangular reshape. For stellarators select the matching toroidal plane geometry; axisymmetric runs reuse one RZ mesh.

  • time_context: Same step: time (s), gstep, tindex and ipc (Runge-Kutta stage). Multiple stages may have identical tindex/gstep; preserve ipc and ADIOS position. tindex counts only EM steps after the RMP start; it is not a uniform physical clock.

  • centering: node

Source: XGC_core/cpp/solvers/perturbed_B_field.cpp:483.

xgc.rmp_residual.bp: dApar_mn (PerturbedBFieldDAparMn)

Native units: T m.

Parallel vector-potential correction Fourier amplitudes by toroidal mode, poloidal mode, and flux surface.

  • plot_label: $\Delta A_\parallel(m,n)$

  • axes: fourier_n,fourier_m,flux_surface

  • mesh_context: Same xgc.rmp_residual.bp: mpol labels the poloidal mode axis and psi_res gives unnormalized flux labels on the radial axis. Leading toroidal DFT slots have length nphi from the simulation, not generally len(ntor). xgc.mesh.bp:theta,psi_surf,surf_idx,surf_len describes the projection geometry.

  • coordinate_context: Leading axis stores toroidal DFT slots, with zero-based slot k corresponding to n=k for the sampled toroidal-period coordinate; the underlying DFT uses phi_i=2*pi*i/nphi. It spans nphi slots, whereas the same-file ntor array lists only configured input RMP harmonics and is not a general coordinate array for this axis. The second axis follows mpol. Apply the simulation wedge convention when translating to full-torus toroidal mode numbers. Stored Fourier amplitudes are nonnegative sqrt(cosine_coefficient^2+sine_coefficient^2); phase is discarded and cannot be reconstructed from these amplitudes alone.

  • time_context: Same step: time (s), gstep, tindex and ipc (Runge-Kutta stage). Multiple stages may have identical tindex/gstep; preserve ipc and ADIOS position. tindex counts only EM steps after the RMP start; it is not a uniform physical clock.

Source: XGC_core/cpp/solvers/perturbed_B_field.cpp:493.

xgc.rmp_residual.bp: damp_fac (PerturbedBFieldDampFac)

Native units: dimensionless.

Damping factor applied to each RMP Fourier mode.

  • plot_label: RMP mode damping factor

  • axes: fourier_n,fourier_m

  • mesh_context: Same xgc.rmp_residual.bp: mpol labels the poloidal mode axis. Leading toroidal DFT slots have length nphi from the simulation, not generally len(ntor); these arrays contain one scalar per DFT-slot/poloidal-mode pair.

  • coordinate_context: Leading axis stores toroidal DFT slots, with zero-based slot k corresponding to n=k for the sampled toroidal-period coordinate; the underlying DFT uses phi_i=2*pi*i/nphi. It spans nphi slots, whereas the same-file ntor array lists only configured input RMP harmonics and is not a general coordinate array for this axis. The second axis follows mpol. Apply the simulation wedge convention when translating to full-torus toroidal mode numbers.

  • time_context: Same step: time (s), gstep, tindex and ipc (Runge-Kutta stage). Multiple stages may have identical tindex/gstep; preserve ipc and ADIOS position. tindex counts only EM steps after the RMP start; it is not a uniform physical clock.

Source: XGC_core/cpp/solvers/perturbed_B_field.cpp:470.

xgc.rmp_residual.bp: gstep (PerturbedBFieldGstep)

Native units: Non-negative integer.

XGC global simulation time step for this RMP residual output slice.

  • plot_label: Global time step

  • time_context: Same step: time (s), gstep, tindex and ipc (Runge-Kutta stage). Multiple stages may have identical tindex/gstep; preserve ipc and ADIOS position. tindex counts only EM steps after the RMP start; it is not a uniform physical clock.

Source: XGC_core/cpp/solvers/perturbed_B_field.cpp:410.

xgc.rmp_residual.bp: ipc (PerturbedBFieldIpc)

Native units: Positive integer.

Runge-Kutta stage index for this RMP residual output slice.

  • plot_label: Runge-Kutta stage

  • time_context: Same step: time (s), gstep, tindex and ipc (Runge-Kutta stage). Multiple stages may have identical tindex/gstep; preserve ipc and ADIOS position. tindex counts only EM steps after the RMP start; it is not a uniform physical clock.

Source: XGC_core/cpp/solvers/perturbed_B_field.cpp:422.

xgc.rmp_residual.bp: mpol (PerturbedBFieldMpol)

Native units: Integer mode number.

Poloidal mode numbers used in the RMP residual diagnostic.

  • plot_label: Poloidal mode number

Source: XGC_core/cpp/solvers/perturbed_B_field.cpp:465.

xgc.rmp_residual.bp: nnode (PerturbedBFieldNnode)

Native units: Positive integer.

Number of mesh nodes in the RMP residual diagnostic.

  • plot_label: $N_n$

Source: XGC_core/cpp/solvers/perturbed_B_field.cpp:437.

xgc.rmp_residual.bp: npsi (PerturbedBFieldNpsi)

Native units: Positive integer.

Number of radial flux-surface samples for the RMP residual diagnostic.

  • plot_label: RMP residual radial sample count

Source: XGC_core/cpp/solvers/perturbed_B_field.cpp:432.

xgc.rmp_residual.bp: ntor (PerturbedBFieldNtor2)

Native units: Integer mode number.

Toroidal mode numbers used in the RMP residual diagnostic.

  • plot_label: Toroidal mode number

Source: XGC_core/cpp/solvers/perturbed_B_field.cpp:460.

xgc.rmp_residual.bp: num_mpol (PerturbedBFieldNumMpol)

Native units: Positive integer.

Number of poloidal mode numbers in the RMP residual diagnostic.

  • plot_label: RMP poloidal mode count

Source: XGC_core/cpp/solvers/perturbed_B_field.cpp:455.

xgc.rmp_residual.bp: num_ntor (PerturbedBFieldNumNtor2)

Native units: Positive integer.

Number of toroidal mode numbers in the RMP residual diagnostic.

  • plot_label: RMP toroidal mode count

Source: XGC_core/cpp/solvers/perturbed_B_field.cpp:450.

xgc.rmp_residual.bp: psi_res (PerturbedBFieldPsiRes)

Native units: T m^2.

Unnormalized equilibrium flux labels on the selected psi_surf2 grid used by the RMP residual diagnostic; axisymmetric values are poloidal flux divided by 2*pi.

  • plot_label: $\psi/(2\pi)$

  • axes: flux_surface

  • normalization: Stored values are not divided by eq_psi_norm. Divide by xgc.equil.bp:eq_psi_norm to form the normalized flux label.

Source: XGC_core/cpp/solvers/perturbed_B_field.cpp:442.

xgc.rmp_residual.bp: rampup_factor (PerturbedBFieldRampupFactor)

Native units: dimensionless.

RMP ramp-up factor applied at this diagnostic time.

  • plot_label: RMP ramp-up factor

  • time_context: Same step: time (s), gstep, tindex and ipc (Runge-Kutta stage). Multiple stages may have identical tindex/gstep; preserve ipc and ADIOS position. tindex counts only EM steps after the RMP start; it is not a uniform physical clock.

Source: XGC_core/cpp/solvers/perturbed_B_field.cpp:499.

xgc.rmp_residual.bp: rmp_residual (PerturbedBFieldRmpResidual)

Native units: dimensionless.

Relative RMP update residual for each toroidal DFT slot and poloidal mode: radial L2 norm of dApar_mn divided by max(tol0, radial L2 norm of Apar_mn). Ignored toroidal slots are zero.

  • plot_label: $\|\Delta A_\parallel\|_2/\|A_\parallel\|_2$

  • axes: fourier_n,fourier_m

  • mesh_context: Same xgc.rmp_residual.bp: mpol labels the poloidal mode axis. Leading toroidal DFT slots have length nphi from the simulation, not generally len(ntor); these arrays contain one scalar per DFT-slot/poloidal-mode pair.

  • coordinate_context: Leading axis stores toroidal DFT slots, with zero-based slot k corresponding to n=k for the sampled toroidal-period coordinate; the underlying DFT uses phi_i=2*pi*i/nphi. It spans nphi slots, whereas the same-file ntor array lists only configured input RMP harmonics and is not a general coordinate array for this axis. The second axis follows mpol. Apply the simulation wedge convention when translating to full-torus toroidal mode numbers.

  • time_context: Same step: time (s), gstep, tindex and ipc (Runge-Kutta stage). Multiple stages may have identical tindex/gstep; preserve ipc and ADIOS position. tindex counts only EM steps after the RMP start; it is not a uniform physical clock.

Source: XGC_core/cpp/solvers/perturbed_B_field.cpp:475.

xgc.rmp_residual.bp: time (PerturbedBFieldTime)

Native units: s.

Simulation time for the RMP damping-factor residual diagnostic.

  • plot_label: $t$

  • time_context: Same step: time (s), gstep, tindex and ipc (Runge-Kutta stage). Multiple stages may have identical tindex/gstep; preserve ipc and ADIOS position. tindex counts only EM steps after the RMP start; it is not a uniform physical clock. Clock convention: in the regular main loop gstep is incremented before diagnostic calls, but sml.time is incremented afterward. A checkpoint written later at the same gstep therefore records time larger by one physical timestep. Use the written time, global step and product/integration stage together; equal gstep alone does not establish simultaneous diagnostic and checkpoint states.

  • imas_candidates: plasma_profiles/time

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_high

Source: XGC_core/cpp/solvers/perturbed_B_field.cpp:427.

xgc.rmp_residual.bp: tindex (PerturbedBFieldTindex)

Native units: Positive integer.

Index of this RMP residual output time slice. Before ptb_3db_start_time this follows gstep; after ptb_3db_start_time it counts only electromagnetic steps in the RMP EM/ES cycle. Use ipc to distinguish Runge-Kutta stages written at the same tindex.

  • plot_label: Output slice index

  • time_context: Same step: time (s), gstep, tindex and ipc (Runge-Kutta stage). Multiple stages may have identical tindex/gstep; preserve ipc and ADIOS position. tindex counts only EM steps after the RMP start; it is not a uniform physical clock.

Source: XGC_core/cpp/solvers/perturbed_B_field.cpp:415.

xgc.sheathdiag.bp: gstep (SheathDiagGstep)

Native units: Non-negative integer.

XGC global simulation time step for this sheath output slice.

  • plot_label: Global time step

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/sheath/sheath_diag.cpp:47.

xgc.sheathdiag.bp: gyraze_converged_count (SheathDiagGyrazeConvergedCount)

Native units: Non-negative integer.

Number of Gyraze surrogate calls classified as converged at each wall node

  • plot_label: Gyraze converged-call count

  • axes: wall_node

  • mesh_context: xgc.mesh.bp: grid_wall_nodes contains zero-based indices into rz (R,Z in m) in the ordered mesh-boundary polygon sequence produced by order_wall. Wall-node values follow this sequence; successive entries and the final-to-first pair define boundary segments for a valid closed wall. The wall-node list is not the triangle connectivity array nd_connect_list.

  • coordinate_context: Wall-node ordering from grid_wall_nodes. Sheath toroidal plane spacing is xgc.mesh.bp wedge_angle / same-product sheath_nphi. Do not assume sheath_nphi equals the field-plane count.

  • time_context: Reduction over the elapsed sheath diagnostic interval ending at same-step time (s), gstep and tindex. *_count are counts over that interval and *_avg are hit-weighted averages, not instantaneous fields.

  • centering: wall_node

Source: XGC_core/cpp/sheath/sheath_diag.cpp:121.

xgc.sheathdiag.bp: gyraze_cutoff_factor_avg (SheathDiagGyrazeCutoffFactorAvg)

Native units: dimensionless.

Average Gyraze sheath cutoff scaling factor at each wall node over this diagnostic interval

  • plot_label: Mean Gyraze cutoff factor

  • axes: wall_node

  • mesh_context: xgc.mesh.bp: grid_wall_nodes contains zero-based indices into rz (R,Z in m) in the ordered mesh-boundary polygon sequence produced by order_wall. Wall-node values follow this sequence; successive entries and the final-to-first pair define boundary segments for a valid closed wall. The wall-node list is not the triangle connectivity array nd_connect_list.

  • coordinate_context: Wall-node ordering from grid_wall_nodes. Sheath toroidal plane spacing is xgc.mesh.bp wedge_angle / same-product sheath_nphi. Do not assume sheath_nphi equals the field-plane count.

  • time_context: Reduction over the elapsed sheath diagnostic interval ending at same-step time (s), gstep and tindex. *_count are counts over that interval and *_avg are hit-weighted averages, not instantaneous fields.

  • centering: wall_node

Source: XGC_core/cpp/sheath/sheath_diag.cpp:104.

xgc.sheathdiag.bp: gyraze_cutoff_factor_count (SheathDiagGyrazeCutoffFactorCount)

Native units: Non-negative integer.

Number of electron sheath hits included in gyraze_cutoff_factor_avg at each wall node

  • plot_label: Gyraze electron-hit count

  • axes: wall_node

  • mesh_context: xgc.mesh.bp: grid_wall_nodes contains zero-based indices into rz (R,Z in m) in the ordered mesh-boundary polygon sequence produced by order_wall. Wall-node values follow this sequence; successive entries and the final-to-first pair define boundary segments for a valid closed wall. The wall-node list is not the triangle connectivity array nd_connect_list.

  • coordinate_context: Wall-node ordering from grid_wall_nodes. Sheath toroidal plane spacing is xgc.mesh.bp wedge_angle / same-product sheath_nphi. Do not assume sheath_nphi equals the field-plane count.

  • time_context: Reduction over the elapsed sheath diagnostic interval ending at same-step time (s), gstep and tindex. *_count are counts over that interval and *_avg are hit-weighted averages, not instantaneous fields.

  • centering: wall_node

Source: XGC_core/cpp/sheath/sheath_diag.cpp:111.

xgc.sheathdiag.bp: gyraze_fallback_count (SheathDiagGyrazeFallbackCount)

Native units: Non-negative integer.

Number of Gyraze surrogate calls that fell back to the original cutoff factor at each wall node

  • plot_label: Gyraze fallback-call count

  • axes: wall_node

  • mesh_context: xgc.mesh.bp: grid_wall_nodes contains zero-based indices into rz (R,Z in m) in the ordered mesh-boundary polygon sequence produced by order_wall. Wall-node values follow this sequence; successive entries and the final-to-first pair define boundary segments for a valid closed wall. The wall-node list is not the triangle connectivity array nd_connect_list.

  • coordinate_context: Wall-node ordering from grid_wall_nodes. Sheath toroidal plane spacing is xgc.mesh.bp wedge_angle / same-product sheath_nphi. Do not assume sheath_nphi equals the field-plane count.

  • time_context: Reduction over the elapsed sheath diagnostic interval ending at same-step time (s), gstep and tindex. *_count are counts over that interval and *_avg are hit-weighted averages, not instantaneous fields.

  • centering: wall_node

Source: XGC_core/cpp/sheath/sheath_diag.cpp:136.

xgc.sheathdiag.bp: gyraze_potential_avg (SheathDiagGyrazePotentialAvg)

Native units: V.

Average local sheath potential passed to Gyraze at each wall node over this diagnostic interval

  • plot_label: $\langle\phi_{\mathrm{sh}}\rangle$ (Gyraze)

  • axes: wall_node

  • mesh_context: xgc.mesh.bp: grid_wall_nodes contains zero-based indices into rz (R,Z in m) in the ordered mesh-boundary polygon sequence produced by order_wall. Wall-node values follow this sequence; successive entries and the final-to-first pair define boundary segments for a valid closed wall. The wall-node list is not the triangle connectivity array nd_connect_list.

  • coordinate_context: Wall-node ordering from grid_wall_nodes. Sheath toroidal plane spacing is xgc.mesh.bp wedge_angle / same-product sheath_nphi. Do not assume sheath_nphi equals the field-plane count.

  • time_context: Reduction over the elapsed sheath diagnostic interval ending at same-step time (s), gstep and tindex. *_count are counts over that interval and *_avg are hit-weighted averages, not instantaneous fields.

  • centering: wall_node

Source: XGC_core/cpp/sheath/sheath_diag.cpp:116.

xgc.sheathdiag.bp: gyraze_unconverged_count (SheathDiagGyrazeUnconvergedCount)

Native units: Non-negative integer.

Number of Gyraze surrogate calls classified as unconverged and evaluated with the full MPE model at each wall node

  • plot_label: Gyraze unconverged-call count

  • axes: wall_node

  • mesh_context: xgc.mesh.bp: grid_wall_nodes contains zero-based indices into rz (R,Z in m) in the ordered mesh-boundary polygon sequence produced by order_wall. Wall-node values follow this sequence; successive entries and the final-to-first pair define boundary segments for a valid closed wall. The wall-node list is not the triangle connectivity array nd_connect_list.

  • coordinate_context: Wall-node ordering from grid_wall_nodes. Sheath toroidal plane spacing is xgc.mesh.bp wedge_angle / same-product sheath_nphi. Do not assume sheath_nphi equals the field-plane count.

  • time_context: Reduction over the elapsed sheath diagnostic interval ending at same-step time (s), gstep and tindex. *_count are counts over that interval and *_avg are hit-weighted averages, not instantaneous fields.

  • centering: wall_node

Source: XGC_core/cpp/sheath/sheath_diag.cpp:128.

xgc.sheathdiag.bp: nwall (SheathDiagNwall)

Native units: Positive integer.

Number of vertices in the wall polygon.

  • plot_label: $N_w$

Source: XGC_core/cpp/sheath/sheath_diag.cpp:65.

xgc.sheathdiag.bp: sheath_ilost (SheathDiagSheathIlost)

Native units: C.

Lost ion charge at each toroidal plane and wall node for the accumulator interval consumed by this sheath adjustment. Modes 1/3 write before the current ipc=2 push, while mode 2 writes after current-step sheath-hit processing; this is not a sum over the full diagnostic cadence.

  • plot_label: $Q_{\mathrm{lost,i}}$

  • axes: toroidal_plane,wall_node

  • mesh_context: xgc.mesh.bp: grid_wall_nodes contains zero-based indices into rz (R,Z in m) in the ordered mesh-boundary polygon sequence produced by order_wall. Wall-node values follow this sequence; successive entries and the final-to-first pair define boundary segments for a valid closed wall. The wall-node list is not the triangle connectivity array nd_connect_list.

  • coordinate_context: Wall-node ordering from grid_wall_nodes. Sheath toroidal plane spacing is xgc.mesh.bp wedge_angle / same-product sheath_nphi. Do not assume sheath_nphi equals the field-plane count.

  • time_context: Loss accumulator slice consumed by the current sheath adjustment, reset after adjustment rather than after each diagnostic cadence. Modes 1/3 output before the current ipc=2 push; mode 2 outputs after current-step hit processing. time is the passed physical sml.time state before its end-of-loop increment, and gstep is the current integration-loop index.

  • centering: wall_node

Source: XGC_core/cpp/sheath/sheath_diag.cpp:91.

xgc.sheathdiag.bp: sheath_lost (SheathDiagSheathLost)

Native units: C.

Lost signed net charge at each toroidal plane and wall node for the accumulator interval consumed by this sheath adjustment. Modes 1/3 write before the current ipc=2 push, while mode 2 writes after current-step sheath-hit processing; this is not a sum over the full diagnostic cadence.

  • plot_label: $Q_{\mathrm{lost}}$

  • axes: toroidal_plane,wall_node

  • mesh_context: xgc.mesh.bp: grid_wall_nodes contains zero-based indices into rz (R,Z in m) in the ordered mesh-boundary polygon sequence produced by order_wall. Wall-node values follow this sequence; successive entries and the final-to-first pair define boundary segments for a valid closed wall. The wall-node list is not the triangle connectivity array nd_connect_list.

  • coordinate_context: Wall-node ordering from grid_wall_nodes. Sheath toroidal plane spacing is xgc.mesh.bp wedge_angle / same-product sheath_nphi. Do not assume sheath_nphi equals the field-plane count.

  • time_context: Loss accumulator slice consumed by the current sheath adjustment, reset after adjustment rather than after each diagnostic cadence. Modes 1/3 output before the current ipc=2 push; mode 2 outputs after current-step hit processing. time is the passed physical sml.time state before its end-of-loop increment, and gstep is the current integration-loop index.

  • centering: wall_node

Source: XGC_core/cpp/sheath/sheath_diag.cpp:82.

xgc.sheathdiag.bp: sheath_nphi (SheathDiagSheathNphi)

Native units: Positive integer.

Number of toroidal mesh planes resolved by the sheath calculation; toroidal angle spacing depends on nphi and wedge_n in xgc.mesh.bp: \(\\Delta\\phi = 2\\pi/(N_{\\phi} N_{\\mathrm{wedge}})\).

  • plot_label: $N_{\varphi,\mathrm{sh}}$

Source: XGC_core/cpp/sheath/sheath_diag.cpp:70.

xgc.sheathdiag.bp: sheath_pot (SheathDiagSheathPot)

Native units: V.

Sheath potential drop (from magnetic pre-sheath to the grounded wall) evaluated by the logical sheath module.

  • plot_label: $\phi_{\mathrm{sh}}$

  • axes: toroidal_plane,wall_node

  • mesh_context: xgc.mesh.bp: grid_wall_nodes contains zero-based indices into rz (R,Z in m) in the ordered mesh-boundary polygon sequence produced by order_wall. Wall-node values follow this sequence; successive entries and the final-to-first pair define boundary segments for a valid closed wall. The wall-node list is not the triangle connectivity array nd_connect_list.

  • coordinate_context: Wall-node ordering from grid_wall_nodes. Sheath toroidal plane spacing is xgc.mesh.bp wedge_angle / same-product sheath_nphi. Do not assume sheath_nphi equals the field-plane count.

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

  • centering: wall_node

Source: XGC_core/cpp/sheath/sheath_diag.cpp:76.

xgc.sheathdiag.bp: time (SheathDiagTime)

Native units: s.

Simulation time of this sheath output slice.

  • plot_label: $t$

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions. Clock convention: in the regular main loop gstep is incremented before diagnostic calls, but sml.time is incremented afterward. A checkpoint written later at the same gstep therefore records time larger by one physical timestep. Use the written time, global step and product/integration stage together; equal gstep alone does not establish simultaneous diagnostic and checkpoint states.

  • imas_candidates: plasma_profiles/time

  • imas_dd_version: 4.1.1

  • imas_confidence: auto_high

Source: XGC_core/cpp/sheath/sheath_diag.cpp:60.

xgc.sheathdiag.bp: tindex (SheathDiagTindex)

Native units: Positive integer.

Diagnostic cadence index computed as integer gstep divided by the configured diagnostic period. This index is not the zero-based ADIOS step and need not start at 1 in a restarted or selected product.

  • plot_label: Output slice index

  • time_context: One diagnostic time slice per ADIOS step (or numbered BP file). In this product, time is physical simulation time in s, gstep is the XGC global integration-step index, and tindex labels this product’s output slice; ADIOS step numbers are zero-based storage positions, not gstep or seconds. Align products using time/gstep rather than ADIOS positions.

Source: XGC_core/cpp/sheath/sheath_diag.cpp:52.

xgc.units.bp: diag_1d_period (WriteUnitsDiag1dPeriod)

Native units: Positive integer.

Input parameter diag_1d_period, i.e. output frequency of 1D diagnostics.

  • plot_label: 1D diagnostic output period

Source: XGC_core/cpp/diagnostics/write_units.cpp:129.

xgc.units.bp: eq_axis_b (WriteUnitsEqAxisB)

Native units: T.

Magnetic-field magnitude at midplane mesh node 0 (the magnetic axis when that node is located there).

  • plot_label: $|\mathbf{B}|$ at mesh node 0

Source: XGC_core/cpp/diagnostics/write_units.cpp:116.

xgc.units.bp: eq_axis_r (WriteUnitsEqAxisR)

Native units: m.

Major radius at the magnetic axis

  • plot_label: $R_{\mathrm{axis}}$

Source: XGC_core/cpp/diagnostics/write_units.cpp:106.

xgc.units.bp: eq_axis_z (WriteUnitsEqAxisZ)

Native units: m.

Height at the magnetic axis

  • plot_label: $Z_{\mathrm{axis}}$

Source: XGC_core/cpp/diagnostics/write_units.cpp:111.

xgc.units.bp: eq_den_v1 (WriteUnitsEqDenV1)

Native units: \(\\mathrm{m^{-3}}\).

Plasma density of eq_den_v1. This value gives a characteristic density value ONLY when an analytic profile is used.

  • plot_label: $n_{\mathrm{ref}}$

Source: XGC_core/cpp/diagnostics/write_units.cpp:149.

xgc.units.bp: eq_tempi_v1 (WriteUnitsEqTempiV1)

Native units: eV.

Ion temperature of eq_tempi_v1. This value gives a characteristic temperature value ONLY when an analytic profile is used.

  • plot_label: $T_{\mathrm{i,ref}}$

Source: XGC_core/cpp/diagnostics/write_units.cpp:123.

xgc.units.bp: eq_x_psi (WriteUnitsEqXPsi)

Native units: \(\\mathrm{ T m^2 }\).

Poloidal flux (over \(2\\pi\)) at the (first) X-point

  • plot_label: $\psi_{X1}/(2\pi)$

Source: XGC_core/cpp/diagnostics/write_units.cpp:134.

xgc.units.bp: eq_x_r (WriteUnitsEqXR)

Native units: m.

Major radius at the (first) X-point

  • plot_label: $R_{X1}$

Source: XGC_core/cpp/diagnostics/write_units.cpp:144.

xgc.units.bp: eq_x_z (WriteUnitsEqXZ)

Native units: m.

Height at the (first) X-point

  • plot_label: $Z_{X1}$

Source: XGC_core/cpp/diagnostics/write_units.cpp:139.

xgc.units.bp: sml_dt (WriteUnitsSmlDt)

Native units: \(\\mathrm{s}\).

Time step \(\\Delta t\) in seconds. Note that the unit is different from the input parameter sml_dt which is given in units of toroidal transit time.

  • plot_label: $\Delta t$

Source: XGC_core/cpp/diagnostics/write_units.cpp:95.

xgc.units.bp: sml_totalpe (WriteUnitsSmlTotalpe)

Native units: Positive integer.

Number of MPI processors

  • plot_label: MPI rank count

Source: XGC_core/cpp/diagnostics/write_units.cpp:90.

xgc.units.bp: sml_tran (WriteUnitsSmlTran)

Native units: \(\\mathrm{s}\).

Ion transit time of [sml_en_order_kev] main ion. \(\\frac{2 \\pi R}{\\sqrt{2 E/m}}\)

  • plot_label: $\tau_{\mathrm{tr,i}}$

Source: XGC_core/cpp/diagnostics/write_units.cpp:101.

xgc.units.bp: sml_wedge_n (WriteUnitsSmlWedgeN)

Native units: Positive integer.

Number of wedges of simulation domain to cover \(2\\pi\).

  • plot_label: $N_{\mathrm{wedge}}$

Source: XGC_core/cpp/diagnostics/write_units.cpp:155.

xgc.units.bp: sp_name + “_ptl_charge_eu” (WriteUnitsSpNamePtlChargeEu)

Native units: dimensionless.

Input parameter ptl_charge_eu, i.e., the species charge number (charge in units of elementary charge \(e\)).

  • plot_label: $q_s/e$ ({name})

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • normalization: Recover signed charge in C by multiplying this charge number by native XGC UNIT_CHARGE=1.6022e-19 C. EV_2_J uses the same numeric constant in J/eV.

Source: XGC_core/cpp/diagnostics/write_units.cpp:172.

xgc.units.bp: sp_name + “_ptl_mass_au” (WriteUnitsSpNamePtlMassAu)

Native units: dimensionless.

Species mass divided by the native XGC PROTON_MASS constant (1.6720e-27 kg); this is a dimensionless proton-mass ratio.

  • plot_label: $m_s/m_p$ ({name})

  • species_context: Native prefix e/e_ denotes electrons, i/i_ the main ion, i2 etc additional ion species. Mass and charge: xgc.units.bp:<species>_ptl_mass_au and <species>_ptl_charge_eu. Tables using nonadiabatic_idx enumerate non-adiabatic species in simulation order (electrons first if kinetic). Velocity-grid distributions and xgc.f0.mesh.bp include only dynamic_f0=true species in compact f0_idx order. Read a product-specific species_context override where present; composition/inputs are recorded in xgc.used_inputs.bp.

  • normalization: The historical ptl_mass_au name/unit uses mass_kg/PROTON_MASS, with native XGC PROTON_MASS=1.6720e-27 kg. Recover SI mass by multiplying this value by 1.6720e-27 kg; do not use the unified atomic mass constant.

Source: XGC_core/cpp/diagnostics/write_units.cpp:164.

xgc.used_inputs.bp: var_name (NamelistUsageWriterVarName)

Native units: dimensionless.

Boolean namelist value returned by a get call, encoded as 0 or 1.

  • plot_label: Boolean input value ({name})

Source: XGC_core/cpp/io/namelist_usage_writer.cpp:92.

xgc.used_inputs.bp: var_name (NamelistUsageWriterVarName2)

Native units: Namelist parameter units.

Namelist value returned by a get call.

  • plot_label: Input value ({name})

Source: XGC_core/cpp/io/namelist_usage_writer.cpp:98.

xgc.volumes.bp: diag_1d_npsi (Diag1dDiag1dNpsi)

Native units: Positive integer.

Number of psi surfaces

  • plot_label: $N_\psi$

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:336.

xgc.volumes.bp: diag_1d_vol (Diag1dDiag1dVol)

Native units: m^3.

Volume associated with each flux surface

  • plot_label: $\Delta V(\psi)$

  • axes: flux_surface

  • mesh_context: xgc.oneddiag.bp (same ADIOS step): psi_mks (poloidal flux divided by 2*pi, T m^2) and psi (normalized flux) are radial coordinates of flux-surface averages. Interpolate profiles in the chosen flux coordinate; do not triangulate radial profile arrays.

  • centering: flux_surface

  • normalization: Per-bin integration volume, not cumulative enclosed volume. The FullDomain Monte-Carlo normalization combines nplanes local wedges, so in toroidal geometry this spans the simulated toroidal wedge; multiply by 2*pi/wedge_angle only when a full-torus volume is wanted. It is not a single-plane node volume.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:341.

xgc.volumes.bp: vol00 (Diag1dVol00)

Native units: m^3.

Volume associated with each psi00 (evenly spaced in psi) surface

  • plot_label: $\Delta V(\psi)$

  • axes: flux_surface

  • mesh_context: xgc.oneddiag.bp: psi00 is the matching uniform poloidal-flux/(2*pi) coordinate. This static array contains per-bin integration volumes, not cumulative enclosed volumes.

  • centering: flux_surface

  • normalization: Per-bin integration volume, not cumulative enclosed volume. The FullDomain Monte-Carlo normalization combines nplanes local wedges, so in toroidal geometry this spans the simulated toroidal wedge; multiply by 2*pi/wedge_angle only when a full-torus volume is wanted. It is not a single-plane node volume.

Source: XGC_core/cpp/diagnostics/diag_1d.cpp:346.