MODFLOW 6 PRT Particle Tracking#

Use transport/modflow6prt when the question is advective travel paths or travel time in a MODFLOW 6 flow field. This is particle tracking, not a concentration-transport solve. For concentration, use transport/modflow6gwt.

Execution Model#

HydroModPy runs PRT as a downstream transport process:

  1. solve a flow/modflow6 GWF model;

  2. attach a MODFLOW 6 PRT model to that flow model through a GWF6-PRT6 exchange;

  3. release particles from the configured cells;

  4. let MODFLOW 6 PRT integrate their pathlines through the flow field;

  5. read the PRT track output into the HydroModPy results store.

Particles are independent tracers. They do not change the flow solution and do not interact with each other. The current HydroModPy integration supports forward tracking on MODFLOW 6 DISV-style meshes.

File Handoff and Direct Reads#

With the current FloPy/external-executable workflow, MODFLOW 6 writes PRT track outputs to disk. The PRT Output Control package exposes binary TRACK and CSV TRACKCSV files, and user-selected output times are supplied through TRACKTIMES. HydroModPy therefore cannot receive pathlines directly as Python objects from the running mf6 process.

The implemented data path is:

  1. MODFLOW 6 writes *.trk.csv and optionally *.trk in the solver scratch directory;

  2. Modflow6PrtOutputAdapter parses the CSV track file;

  3. HydroModPy stores vectorized arrays under pathlines/ in the run Zarr store:

    • pathlines/x;

    • pathlines/y;

    • pathlines/z;

    • pathlines/time;

    • optional status and termination-reason arrays.

A more direct plotting path is also available for reports that do not need the catalogued Zarr arrays: the shared CSV parser can feed an HTML report directly from *.trk.csv. That avoids a second Zarr read for a one-off page, but it still consumes a MODFLOW 6 track file because the solver is an external executable. Avoiding the file boundary entirely would require a different execution architecture, such as an in-process MODFLOW 6 library binding or a dedicated streaming callback; that is not how HydroModPy currently runs MODFLOW 6.

Minimal TOML Pattern#

Declare the flow process first, then the PRT process:

[[simulation.process]]
id = "flow_main"
type = "flow"
solvers = ["modflow6"]

[[simulation.process]]
id = "prt_main"
type = "transport"
solvers = ["modflow6_prt"]

[simulation.results]
keep_solver_files = true

[transport.modflow6prt.parameters]
release_zone = "domain_nonriver"
upstream_top_quantile = 0.88
track_dir = "forward"
porosity = 0.01
local_z = 0.5
max_particles = 300
release_times_days = [0.0]
track_time_step_days = 2.0
stop_time_days = 3650.0
stop_travel_time_days = 3650.0
extend_tracking = true
dry_tracking_method = "drop"
exit_solve_tolerance = 1.0e-10
write_track_csv = true
write_track_binary = true

The demonstrator in examples/projects/14_transport_nancon_gwt_visual_guard/run_nancon_steady_mf6_prt_pathlines.toml uses this pattern on the Nancon DISV mesh. keep_solver_files = true is not required for catalogued Zarr extraction, but it keeps the raw *.trk.csv file available for direct inspection and report generation.

Release Zones#

The first implementation supports a small set of deterministic release-zone selectors:

Zone

Meaning

domain

All active cells.

domain_nonriver

Active cells excluding river-support cells, spread over the modeled domain when max_particles is smaller than the selected zone.

upstream

Active cells above the configured top-elevation quantile.

upstream_nonriver

Upstream cells excluding river-support cells. This is useful for testing hillslope-to-river travel paths without releasing directly on the river.

river

River-support cells.

outlet

Low-elevation river cells when possible, otherwise low-elevation active cells.

custom

Explicit zero-based DISV cell ids from particle_cell_ids.

Tracking-Time Discretization#

Two mechanisms control stored pathline density:

track_times_days

Explicit list of user tracking output times.

track_time_step_days

Convenience spacing used when track_times_days is omitted. HydroModPy expands it into regular tracking output times from zero to stop_time_days.

Increasing tracking-time density does not change the underlying GWF flow solve. It asks PRT to report more intermediate pathline positions, so the stored and plotted pathlines look smoother. It can increase CSV/Zarr size.

Velocity and Porosity#

PRT advective speed depends on the flow field and the porosity used by the PRT MIP package. In HydroModPy, porosity can be set explicitly as a uniform value. When it is omitted, positive specific yield values from the flow model are used where available. Lower porosity increases pore velocity for the same Darcy flux, so particles travel farther in the same simulation time.

Version Requirements#

Use a recent MODFLOW 6 executable with PRT support. The Nancon PRT demonstrator has been verified with MODFLOW 6.7.0. If HydroModPy resolves an older cached mf6 binary, set an explicit path:

[modflow6.runtime]
mf6_executable_name = "mf6"

References#

On this page