Groundwater modeling using flopy and Modflow 6 (MF6)
Modflow 6 changed the world for groundwater modelers. It promises a way into the future by being completely modular and extensible with new packages, and the ability to combine with more and diverse models with which it can exchange data.
Modflow 6 integrates most of the Modflow developments made in the past, which often led to separate versions of Modflow. One of the advances of MF6 is to allow for unstructured grids. But also internally MF6 has been modernized and improved. For instance, the use of Newton solutions was used to solve instabilities with non-linear problems. The new mf6 also soles the problem with rewetting of dry cells that have so long plagued geneartions of Modflow users because it often prevented to reach convergence.
Furthermore, md6 contains both the flow model and the transport model (previously MT3DMS was used for this) and can handle variable density and viscosity (replacing the previous seawat model).
Modflow can now be conveniently interacted with by python cripting through flopy, which was also newly developed by a team of people at the USGS together with scholars from over the world.
Mf6lab builds on floy and MF6 by standardizing setup and analysis of groundwater models with flopy and m6, thus streamlining its use in an overlying framework. mf6lab is applied in many projects. More will follow as I will use mf6lab for all new groundwater modelling projects in the future.
Examples can be found under mf6lab/projects//cases/
The idea central to mf6lab is to have a backbone file mf_setup.py that never changes which creates a dictionary with the data for all packages to be used in the simulation a standard file run.py which invokes it and the files mf_adapt.py, mf_analyze.py and <case_name>.xlsx which are specfiic for each model. Note however, that for most models the visualization of the simulation results, which is wat mf_analyze.py does is essentially the same for most models. mf_adapt is used to set the parameters for the packages to be used. It can conveniently import a dedicated file settings.py in which certain parameters are stored together and read from the Excel workbook file <case_name>.xlsx which has some sheets to guide the model. The first sheet is named NAM. This is where you tell wich flow and transport packages to use for the current simulation. You do this by setting the value of appropriate cells to 1 or zero. The transport model will be automatically invoked when one or more transport packages are switched to on. The second sheet is SIM6 it holds the parameters and their default values of the simulation packages. The third sheet is named GWF6. It holds the parameters of all availabe packages with their default values for the flow model. As long as your are satisfied with them, you don't have to change them, but you can do it here. Alternatively you can overwrite any parameter in mf_adapt.py, whatever is more natural to you. In any case, the sheet gives a complete overview of all parameters and their default values. mf_adapt.py will swallow this entire sheet to generate a dictionary with the parameters for each of the available packages. The next fourth sheet is named GWT6. It has the same structure as the previous sheets and holding all parametes and their default values pertaining to the tranport model. This too will seldom be changed. The fifth sheet is called PER, it specifies the stress periods. It is a table with at least the following header IPER PERLEN NSTEPS TSMULT needed to tell mf6 about the stress periods and subdivision for this simulatioin. One can add as many columns as desired to this table, which may then be utilized in mf_adapt.py. For instnace, in which stress period some action takes place. The sheet will be read in by mf_adapt.py and the resulting pandas DataFrame is yours to use at your liking. Only the mentioned columns are requied for mf6. A further sheet ofthen useful is LAY to specify layers and their properties. This sheet may be use if deemed useful. For instance, one may specify for each layer its conductivities, storativities, dispersivities as well as some code and geologic description. The sheet can be read in by mf_adapt.py and use to provide data to prepare the input, but data in the resulting DataFrame can also be used to set parameters for visualization such as itmes like formation names to be used in the legend. This is completely up to you.
Often it is convenient to create an exta file like settings.py to store divers simulation parametes on one place, and to set up, test and visualize the network. This file can be imported by 'mf_adapt.py' and by mf_analyze.py to gain access to the parameters in it from a central place. Wether to use such an extra file settings.py is up to the user, but it turns out to be convenient in many circumstances as a central place to find the parameters to fine-tune a given siulation-run.
Thus, having prepared mf_adapt.py run run.py to let flopy generate the mf6 files for this simulation. The smiulation is automatically started. It's progress can be followed on the screen. When successfully finished run mf_analyze.py to read, analyze, visualize, print and store the results. That's it.
For a new run change a few parameters which may be convenientlu stored in settings.py, launch run.py again and afterwards launch mf_analyze.py again to get the new results.
Some actions of done is:
switch from steady state to transient --> just swich on the Gwfsto pacakge in the NAM sheet of the Excel file and run again.
to switch on the Gwfchd package ---> just switch this package on in the NAM sheet of the Excel file.
to use the transport model ---> switch on the approprate Gwt-packages in the NAM sheet. Of course, in mf_adapt.py you set the appropriate parameters for these packages.
to change the stress periods, time steps and tsmult--> change them in the PER sheet of the Excdel file.
to change layer properties ---> change them in the LAY sheet of the Excel file.
to use density flow --> switch on the Gwfbuy package in the NAM shet of the Excel file.
The actual input tha MF6 is genearted in a number of files by flopy. They can be inspected in the folder *GWF of the current case. The files for the input generated by flopy for the transport simulation can be found in the folder GWT of the current case. The files for the input of the current simulaiton under which the flow and the tranport model resied can be found in the filder SIM of the current case. This folder also receives the output files that MF6 generated i.e. the heads (<case>.hds) and budget (<case>.cbc) files and possibly the concentration (<case>.unc) files.
To know which packages are needed in a specific model and whether or not the groundwar transport model is to be run, mf6lab reads the sheet NAM in the Excel file <case>.xlsx. Of course one can add as many special purpose sheets to this Excel file as one desires for specific actions or data in this simulation.
See the mf6lab/projects fiolders for the different projects. Under each project see the cases folder for the different individual cases. Note that every case has the same folders (more can be added of desired or removed if not needed). Gernally the visualization outcome is stored in the folder Images (i.e. fm6lab/Projects//cases//Images).
@TOlsthoorn 20231226, 20240202
Theo Olsthoorn (prof. dr.ir. TN), emeritus prof TUDelft, eritus hydrologist at Waternet Amsterdam. Heemstede 2024/02/02
TO 20231225