Solver Capability Matrix#

Purpose#

This page is meant to become the compact comparison sheet that answers:

“Which backend should I use for which kind of HydroModPy problem, and how well is that path currently documented and validated?”

The first selection axis is now the process type. Backend-family comparisons remain useful, but they should not hide the distinction between flow, transport, postprocess, and display.

Axis-level capability snapshot#

Compact matrix scanned in seconds. yes means a public path exists, partial means the code surface exists but the docs or validation coverage is still incomplete, no means the feature is not exposed.

Backend	Steady	Transient	Unstructured mesh	Transport	Calibration	1D	2D	3D
`modflownwt`	yes	yes	no	via MT3DMS	yes	yes	yes	yes
`modflow6`	yes	yes	yes (DISV)	yes (GWT)	yes	yes	yes	yes
`boussinesq`	yes	yes	yes (triangular)	no	partial	yes	yes	no
`gr4j`	yes	yes	n/a (lumped)	no	yes	n/a	n/a	n/a

The columns above are the recommended decision axes when scoping a new study. For richer context, read each backend row in Process-Level Matrix below.

Visual Anchors#

The matrix below is easier to read if each solver family is connected to a real result artifact. These figures are representative anchors, not exhaustive capability claims.

MODFLOW 6 basin run

MODFLOW 6 transient basin flow-state triptych — A modern MODFLOW 6 run can publish basin-scale state fields such as topography, hydraulic head, and water-table depth.

MODFLOW-NWT basin run

Water budget from the Nancon transient MODFLOW-NWT run — The NWT path remains useful for structured legacy-style simulations where budgets, heads, and hydrographs can be read with established MODFLOW conventions.

Boussinesq validation run

Boussinesq hillslope recharge-step interception validation result — The in-house Boussinesq backend is best read through controlled validation figures before broad production claims are made.

Process-Level Matrix#

Backend	Mesh support	Regimes	Main scientific role	Validation anchor	Current main limits
`modflownwt`	Structured `sgrid`	Steady and transient	Legacy MODFLOW-family flow path, still important for comparison, MODPATH, and MT3DMS workflows	`validation_cases/README.md` plus MODFLOW-NWT Architecture	No runtime Gmsh mesh path, scientific method note still missing, sunset strategy not yet fully reflected in user-facing docs
`modflow6`	Structured plus runtime DISV-style unstructured meshes	Steady and transient	Modern MODFLOW-family backend for flow, with irregular-mesh support and growing method-choice documentation	XT3D method-choice assets plus MODFLOW 6 Architecture	Scientific note still incomplete on package choices, vertical assumptions, and when to prefer it over NWT
`boussinesq`	Triangular runtime meshes	Steady and transient	In-house finite-volume shallow-groundwater backend with explicit groundwater/surface-interaction formulations	Boussinesq Theory plus `validation_cases/README.md`	Still under active validation, not yet documented as a fully mature production envelope for all workflows

Documentation Maturity Matrix#

Topic	Current maturity	Best current source	Main gap
Project-level scientific scope	Low	System Scope And Assumptions	Still a scaffold, not yet a complete physical framing page
Solver-agnostic groundwater problem	Low to medium	`hydromodpy.physics.flow` plus Groundwater Flow Problem Definition	Scientific contract still implicit in code
Hydrological forcing chain	Low	`hydromodpy.physics.forcing` and `pyhelp`	No canonical scientific page on recharge-generation and forcing semantics
MODFLOW-family scientific choices	Medium	MODFLOW Governing Equation And CVFD Formulation, MODFLOW 6 Versus MODFLOW-NWT: Scientific Comparison, XT3D note, architecture pages	Core rationale now exists publicly, but package-level and forcing-level synthesis still remains dispersed
Boussinesq mathematical formulation	Medium to high	Boussinesq Theory	Good core note, but backend envelope and comparison guidance still need consolidation

How This Page Should Be Used#

This matrix should later serve three different audiences:

contributors choosing where to add documentation next,
users choosing a backend for a new study,
reviewers checking whether a modelling path is both implemented and justified.

Immediate Next Additions Recommended#

The next high-value additions to this page would be:

one row-level capability breakdown for boundary conditions and forcing families,
one explicit column for mesh topology and vertical representation,
one column for transport compatibility,
one column for documentation confidence versus validation confidence.