A Random Walk Particle Tracking Code for Solute Transport in Heterogeneous Aquifers
MODPATH-RW is a Random Walk Particle Tracking (RWPT) model, implemented by extension of the particle tracking program MODPATH-v7. The code employs discrete solute particles to solve the Advection Dispersion Equation (ADE), and inherits the parallel processing of particles from MODPATH-OMP.
The program is naturally integrated with groundwater flow models obtained with MODFLOW-2005, MODFLOW-USG and MODFLOW 6, for structured and rectangular unstructured grids.
Clone the repository
git clone https://github.com/upc-ghs/modpath-rw
Some external dependencies are stored in the src/lib
folder. The necessary files are all explicitly included in the repository for convenience.
-
gpkde
: module for smoothed reconstruction of concentrations. -
finterp
: module for linear interpolation. -
rng_par_zig
: module for random number generation in parallel with the Ziggurat method.
Makefiles are available at the folder make
:
makefile-gfortran-pc
: for thegfortran
compiler.makefile-ifort-pc
: for theifort
compiler, verified for[email protected]
.makefile-gfortran-mac
: for thegfortran
compiler.
By default, the compiled program is called mpathrw
. Compilation process will create a folder with temporary objects (make/objtemp
). When integrating program updates and recompiling, it is advised to remove this folder to avoid inconsistencies.
The program can be built with the meson build system:
- Setup the configuration:
meson setup builddir -Ddebug=false --prefix=$(pwd) --libdir=bin
- Compile:
meson install -C builddir
- Test:
meson test --verbose --no-rebuild -C builddir
The folder msvs
contains the project and solution files generated with Visual Studio 2022 on a Windows system.
An executable file and complementary dll
's providing the OpenMP library are available at the folder wbin/
. Users can make easy use of these files at a system level by extending the PATH
environment variable, adding their specific address to the wbin/
folder (as here or here).
The necessary dll
's can also be installed from the Intel and Visual Studio redistributables (x64
), respectively:
Details about the configuration of input files are provided in the program Documentation of Input-Output. It is recommended to follow this document closely with the Description of Model Input and Output of MODPATH-v7.
Users familiarized with the python
interface FloPy are encouraged to write input files for MODPATH-RW with the extension flopyrw. This package provides classes based on the interface for mp7, which have been adapted to the specific requirements of MODPATH-RW, plus new classes for the new model packages. The extension is consistent with the program documentation and example configuration routines and test cases can be found on the project page.
A command line interface provides some simple instructions and shortcuts for managing program execution. Asking for help (mpathrw -h
) displays the following message
MODPATH-RW version *.*.*
Program compiled MMM DD YYYY HH:MM:SS with ******** compiler (ver. *******)
A Random Walk Particle Tracking code for solute transport in heterogeneous aquifers
usage:
mpathrw [options] simfile
options:
-h --help Show this message
-i --init Initialize simulation without running
-l <str> --logname <str> Write program logs to <str>
-nl --nolog Do not write log file
-np <int> --nprocs <int> Run with <int> processes
-p --parallel Run in parallel
-ts <int> --tsoutput <int> Selects timeseries output <int>
-s --shortlog Simplified logs
-v --version Show program version
For bug reports and updates, follow:
https://github.com/upc-ghs/modpath-rw
Note: if the OpenMP library is not found by compilers, the program will be generated without parallel compatibility, without the input parameters nprocs
, parallel
and tsoutput
. If these are listed on the program help message (as above), the compilation with parallel compatibility was successful.
--init
: initialize the program packages and stop execution before displacing particles. Useful to monitor the initialization workflow and reports written to the mplst
file.
--nprocs
: run in parallel with the given number of processes (omp_set_num_threads(nprocs)
).
--parallel
: shorthand for running in parallel with the maximum number of processors (omp_get_num_procs()
).
Note: if neither --parallel
nor --nprocs
are given, the program will by default employ omp_get_max_threads()
, whose value can be controlled with the environment variable OMP_NUM_THREADS
.
--tsoutput
: controls the output protocol for timeseries runs while running in parallel
- 1: is the default format, output is performed into a single output unit with the OpenMP thread exclusive clause (critical). Only difference versus a serial run is that the output file contains non-sorted particle indexes.
- 2: timeseries records are written into thread specific binary units and then consolidated into a single file after each timeseries output time. Timeseries file generated with this format does not contains a file header.
- 3: timeseries records are written into thread specific output units. Timeseries file header is written only to the output unit of the first thread
1_example.timeseries
. Initial particle positions are also written to the file of the first thread.
Follow the contribution guidelines for this project.
MIT License