I am working with Flopy (3.3.5) and MF6 (6.2.2) transport to simulate a plume movement towards a creek from an industrial site. I am using a structured regular grid with 6 layers flow and transport are run simultaneously, and the creek is represented in the flow model using a RIV package with Drains along the bank. With the new ATS package the simulation works Great! and produces reasonable concentration results. The difficulty I run into is trying to extract the mass flux into the boundary conditions representing the creek, from the various budget and flux output options. I have set up an observation package to record the mass flux across every “JA-FACE” that the boundary condition cells share with the rest of the model, and this seems to be working, but I have been trying to compare this to budget outputs and it has highlighted some inconsistencies and problems with the current output.
• I noticed that the model simulates very small numbers for mass flux rather than zero away from the plume after a few iterations. When the numbers with three digit exponents are written to the list file and csv output from the obs package the ‘E’ in the number gets dropped(eg.” -2.90933775-241”, while 2 digit exponent numbers are written fine as “-2.90933775E-41”), I have tried changing the number of digits on output but the behaviour seems to happen consistently. I have scripted a fix for the output files of the observation package, but I imagine this is likely why I get an object with all “none” entries when I try to read the list file with the flopy listBudget utility.
• I have Set the SinkSourceMixing (SSM) package to both print and save budget information. When I look at the output written to the list file (in a text editor rather than in python) the cellid’s (l,r,c tuple) written to the list file with the SSM information do not corelate to the original cellids in the stress period data of the original flow model packages (inconsistencies across well, River, and Drain packages)
• I have loaded the binary budget file and can select the SSM data using “cbb.get_data(totem=t,text=”SOURCE-SINK MIX”) to get an appropriate recarray. I have been calculating the node number based on the formula in the “get_node” method of the “discretization/structuredgrid” package “node=Lnrowncol+R*ncol+C” which is different to the one used in the flopy/mf6 complex model example notebook. Using this node number I can select the rows of the recarray that has it in the “node” column. When I do this I noticed that all of the associated entries in the “node2” column don’t seem to be adjacent cells from any method of trying to back out the LRC from node number that I can find. When I look at the “FLOW-JA-FACE” data in the budget I get a consistent value for the flow across a face from the budget file using my calculated node number that I get from the observation package (defined with a L,R,C tuple) so that seems alright, but the summation of all flux into the cell from the “FLOW-JA-FACE” data does not seem to quite add up to the values in the SSM data in the budget.
I think for my purposes the flux into the boundary nodes I am getting out of the obs package is sufficient I just wanted to post a question documenting these observations in the hope that I can:
- understand if its due to something in my model setup, and
- make a bit of a contribution to the flopy/mf6 project from my experiences if they are due to issues in the codes that need to be addressed.
Thanks to all the contributors to this software package Kevin