Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Potential issue with GSFLOW runoff calculations #19

Open
etrianaGIT opened this issue Sep 28, 2022 · 11 comments
Open

Potential issue with GSFLOW runoff calculations #19

etrianaGIT opened this issue Sep 28, 2022 · 11 comments

Comments

@etrianaGIT
Copy link
Collaborator

I experimented with two runs with the same GSFLOW model and compared the results showing different runoff estimates resulting in the upper basin, which doesn't have any lakes or diversions. The two runs are:
(1) using a recent GSFLOW.exe compiled to generate the segments output. The segment output (net depletion) is simulated in MODSIM as static. (Mode: GSFLOW in the control file)
(2) using the coupled MODSIM-GSFLOW code with a MODSIM network that uses iterations and convergence between the two models. (Mode: MODSIM-GSFLOW in the control file)
The comparison of the result of the routed runoff from the upper basin in MODSIM shows the following behavior

image
and the difference is more notorious in the peak flows.
image

  • the blue line being the GSFLOW mode run.
@etrianaGIT
Copy link
Collaborator Author

I tired running the same model with the coupled code but on GSFLOW5 mode and the model seem to struggle with convergence
image

@rsregan
Copy link
Collaborator

rsregan commented Sep 29, 2022 via email

@etrianaGIT
Copy link
Collaborator Author

etrianaGIT commented Sep 29, 2022 •
edited
Loading

Ahh, okay -with just GSFLOW mode - the start of the simulation has similar messages.
image

@etrianaGIT
Copy link
Collaborator Author

etrianaGIT commented Sep 30, 2022 •
edited
Loading

[Update]

The input files of the test model were in GSFLOW 4 version, so they needed to be updated to GSFLOW5, which uses fractions and offset parameters for precipitation and temperature. The new parameters file was generated using the convert function of GSFLOW.
....\bin\gsflow_v2.exe .\ag-gsp.control -set model_mode CONVERT

The parameter file created was placed in the input\prms folder and the new parameter file was added to the control file. This parameter files allow simulation with GSFLOW5 and MODSIM-GSFLOW modes.

This run uses the C# code but in GSFLOW5 mode -
image

Inflows at the control point are the same to the GSFLOW stand-alone results.

Need to do a test with MODSIM-GSFLOW mode to see if the initial discrepancies improve.

@etrianaGIT
Copy link
Collaborator Author

etrianaGIT commented Sep 30, 2022 •
edited
Loading

[Update 2]

These results correspond to a MODSIM-GSFLOW mode run with the C# code. At the reservoir inflow - control point - inflows are similar to the previous GSFLOW5 run; however, the peak flows are lower in the MODSIM-GSFLOW mode than in GSFLOW5 mode (running with the coupling code).
image

The MODSIM-GSFLOW mode peaks are lower than the stand-alone GSFLOW run.

Are these results worrisome? Any ideas how to start debugging these differences?

@emorway-usgs
Copy link
Collaborator

@etrianaGIT It might be more illustrative if you can show a plot of the differences. It'll be easier to get a feel for just how different things are with that type of plot.

@etrianaGIT
Copy link
Collaborator Author

etrianaGIT commented Sep 30, 2022 •
edited
Loading

Here is a plot that compares each flow value between the two modes (not sure how meaningful it is)
image
The differences are
image

@emorway-usgs
Copy link
Collaborator

There are the obvious spikes, but there also looks to be the large differences early in time on the order of 20 acft/day, which would equate to 900,000ft^3/day, which is a lot of water in today water scarce world.

@rsregan
Copy link
Collaborator

rsregan commented Oct 3, 2022 via email

@rsregan
Copy link
Collaborator

rsregan commented Oct 3, 2022 via email

@emorway-usgs
Copy link
Collaborator

@rsregan The files you mentioned, namely dynamic_soil_param_read.f90 and strmflow_character.f90 are not yet in the develop_modsim branch, as far as I can tell. Perhaps when you jump on the code sprint call, we will be able to get this sorted out.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
3 participants
@emorway-usgs @rsregan @etrianaGIT