Finding neighboring grid cells for flux transfer in CTSM

[akhtert]

Hello,

I am trying to simulate cell-to-cell groundwater flow in CTSM by activating the water table lower boundary condition. My goal is to simulate groundwater pumping for irrigation, adjust the water table based on the amount of water pumped, and allow lateral flow between neighboring cells to achieve a stabilized water table. This task is very similar to this previous work done in CLM5, where groundwater pumping as well as lateral groundwater flow was included. I am attempting to transfer the relevant code from this work into the latest CTSM (ctsm5.2.023).

However, I understand that the neighboring grid cell functionality has been discontinued in CTSM (finding neighboring grid | DiscussCESM Forums (ucar.edu)). For this specific purpose, I would like to ask if there is any way to reintroduce or adapt this feature to enable lateral groundwater flow between cells. I would greatly appreciate your thoughts on this matter and any suggestions you may have for achieving this.

Thank you.

 

[oleson]

Some comments from various TSS software/scientists:

it was never in clm, so I wouldn't say it was discontinued. There was a student (Farshid Felfelani) who put it in an older version, but it was never included in the main model. They could check with him to see if there is code they could try to adapt. As erik points out in the post that was referenced, it requires significant changes.

We should note that CTSM is designed such that gridcell to gridcell transfer is going to perform horribly. Especially if you are doing it every time step. If this was an important thing to do you'd want to redesign CTSM which would obviously be a huge effort. Or we could look at having a subcomponent of CTSM that we couple to for doing this efficiently.

The other thing of note is that FATES added seed dispersal which does this. And using what they are doing there would be the quick and dirty thing to do. Since seed dispersal is only done yearly it's inefficiency doesn't matter that much. In this case for every time step for water it will matter and will really affect performance and performance for increasing MPI tasks.

 

[amans]

The other thing of note is that FATES added seed dispersal which does this. And using what they are doing there would be the quick and dirty thing to do.

Hi @oleson. I am collaborating with @akhtert in this work. Could you point me to the FATES seed dispersal code where they figure out the neighboring grid cells?

 

[slevis]

I went to the FATES site on github and searched for seed dispersal and found the following:

Issues · NGEET/fates

repository for the Functionally Assembled Terrestrial Ecosystem Simulator (FATES) - Issues · NGEET/fates

github.com