Simulation Comparison#
Note
This page and its static assets are auto-generated by python -m tools.doc_gallery. The Sphinx build only reads committed PNG and JSON artifacts.
These cases compare multiple modelling methods on the same saved support. The figures stay lightweight enough for the docs while still exposing map-wide errors.
See also
Read How to read gallery, comparison, and validation pages if you want a guided entry point before opening the case pages below.
Current Coverage#
Study areas: Example12 / Naizin, Naizin catchment, Natural N1 10 km2 testbed.
Compared simulations: Boussinesq PETSc TS/SNESVI candidate, Boussinesq reusing the same mesh, MODFLOW 6 on committed Gmsh mesh, MODFLOW 6 reference.
Compared observables: Head Dry Late, Head First Computed, Head Last, Head Timeseries, Head Wet Year1, Watertable Depth, Watertable Elevation.
Comparison families: Same Support, Different Solvers (2), Different Supports, Same Regime (2), Multi-Simulation Suites (2), Natural-Geology MF6/Boussinesq Testbed (2).
Extend this section by adding committed comparison TOMLs to an existing family, or by declaring a new family when the comparison axis really changes.
Comparison Families#
The section is grouped by comparison intent first, so similar pages stay together even when the inventory grows.
Same Support, Different Solvers#
These cases keep the spatial support fixed so the main signal comes from solver-family differences rather than from a support change.
Shared-Mesh Solver Comparison on Naizin
Shared-mesh comparison of MODFLOW 6 and Boussinesq on Naizin, combining map snapshots and three head chronicle probes.
MF6 vs NWT on the Same 60x60 Grid
Annual moderate comparison on one shared structured grid, isolating MODFLOW-family differences from mesh effects.
Different Supports, Same Regime#
These cases keep the forcing regime fixed but intentionally change the mesh family, so the page captures both solver differences and support-transfer effects.
MF6 Triangular vs NWT Structured on Moderate Forcing
Annual moderate comparison where both solver family and mesh family change, with a common fine raster used for map alignment.
MF6 Triangular vs NWT Structured on Demonstrative Forcing
Different-support comparison on the more expressive demonstrative annual setup, including flux and timing diagnostics.
Multi-Simulation Suites#
These cases keep more than two simulations on one page so the reader can separate solver-family, support-family, and runtime-family effects without opening several independent comparisons.
Four-Simulation Moderate Suite on Example12
One annual moderate suite spanning MF6 and NWT on structured support plus MF6 and Boussinesq on committed triangles.
Four-Simulation Moderate Suite with Surface-Excess Diagnostics
Diagnostic extension of the four-simulation moderate suite, adding surface-excess observables and Boussinesq budget diagnostics.
Natural-Geology MF6/Boussinesq Testbed#
These pages publish reviewed Boussinesq/MODFLOW 6 comparison outputs from the testbed workflow without running the solvers during doc generation.
Site 03 Natural N1 10Km2 Mf6 Bouss
Published MODFLOW 6 and Boussinesq comparison artifacts, reused by the documentation without rerunning the solvers.
Site 08 Natural N1 10Km2 Mf6 Bouss
Published MODFLOW 6 and Boussinesq comparison artifacts, reused by the documentation without rerunning the solvers.