Model Partner Software
GMDSI sees strategic use of model partner software as essential to decision-support modelling. Two specifications are fundamental to model partner software. They are:
- a non-intrusive interface with the model
- an ability to parallelize model runs.
GMDSI recognizes that there are two major obstacles to widespread use of model partner software by the groundwater industry at large. These include:
- lack of understanding of the tasks performed by this software, and of the theory behind these tasks
- difficulties in constructing input datasets for this software, especially in contexts where parameters are many and observation numbers are large
- the high cost, in terms of model runs, incurred by this software when undertaking parameter estimation and uncertainty analysis.
- improved use of pilot points in parameterization of spatially variable hydraulic properties in two and three dimensions, and parameterization of polylinear boundary conditions. These improvements are implemented in the PEST PLPROC pre-processor
- development of a new utility named TS6PROC which allows parameter-based manipulation of time series contained in MODFLOW 6 time series files
- development of a new automatic setup utility named OLPROC. OLPROC facilitates construction of PEST and PEST++ input datasets based on complex, multi-component objective functions that involve observations and observation differences
- improvements to regularization and randomized Jacobian functionality provided by PEST_HP.
All of this software will be downloadable for free from the PEST web pages.
In addition, enhancements and additions will be made to PEST, PEST_HP and other members of the PEST suite on an as-needed basis to support GMDSI research and worked example work. These will be made available to the industry at large as they are developed.
Independently from GMDSI, the PEST++ software suite is under continuous development. So too is the PyEMU suite. The PyEMU suite complements the PEST++ suite by enabling easy setup of PEST and PEST++ datasets, and by performing ancillary and complementary tasks such as linear analysis, plotting and visualization. See the PEST++ and PyEMU web pages.
Other software and resources
A simple, lumped parameter recharge model named LUMPREM, with linkages to MODFLOW, is available from the PEST web pages.
GMDSI is exploring the possibility of collaborating with the USGS personnel to improve MODFLOW 6 performance in areas where near-surface hydraulic properties vary vertically. This will be accomplished without the need to add extra layers to a model. This is expected to be an attractive option in simulating the regional impact of mining. Faster run times and increased numerical stability will facilitate data assimilation and uncertainty analysis.
Python-based software to facilitate MODFLOW setup and pre- and post-processing, is available though flopy. Other software to assist in Python-based MODFLOW setup is available from Andy Leaf’s GitHub site.
Training and workflow resources for PyEMU can be downloaded from here.
Banner photo: BHP