CLM with additional PFTs

[abadger@gmu_edu]

I am having issues running CLM offline.   I was trying to add a new PFT, I got that added to the physiology file and put the new PFT in the fsurdat file in the namelist.  I then changed some parts in the code so it could read in my new PFT.  The code compiled. The model did not run though, I found this error in the cesm.log file:1: pio_support::pio_die:: myrank=          -1 : ERROR: pionfget_mod.F90:1:         380 : NetCDF: Start+count exceeds dimension bound1:abort:1:INFO: 0031-306  pm_atexit: pm_exit_value is 134.ERROR: 0031-250  task 1: Aborted I have been trying to track down what the issue is for the past week or so, but have not had any success.  I think it might have something to do with calcdecomp.F90 in pio, but not 100% sure on that.  In the email I received when the job was exited, it said "CPU time : 5417.37 sec."  This leads  me to believe that the changes I made in the code did not stop the model from running, but is more of an I/O issue.

 

[jedwards]

You didn't say which version of cesm / clm you are using?   What machine you are on?   What IO options are you using?  What grid?   What pe layout?

 

[slevis]

Andrew,Given that the model stopped working after you made changes to code and datasets, I would suggest going back to code that worked and making one change at a time. Considering that the error is about netcdf, then your problem most likely has to do with the files that you changed and/or how you modified the way that they are read. It will be helpful for you to figure out whether the problem is with your changes to the pft-physiology file or your changes to the fsurdat file.Sam

 

[abadger@gmu_edu]

I am using CESM1.1.1 (CLM4) on Yellowstone with a regular lon/lat grid (f09_g16). I am using the default IO options and am using the default PE layout that the scripts setup.

 

[slevis]

We got this request:[...] my project involves creating our own PFTs for specific crop types. Would it be
possible to have access to release code or guidance on creating custom
PFTs to run with CLM4?Response:CLM release code is accessible on the CESM web site, though this may not help with creating your own pfts. I will give you basic guidance and the rest will be up to you.

The clm reads most of its pft info from the surface data (fsurdat in the clm namelist) and the pft physiology file (fpftcon in the clm namelist). A few parameters may be hardwired directly in the clm code.

You will find it easier to replace existing pfts with your own if that's an option for your application. In that case replace existing specific crop types with your specific crop types. In particular, modify parameter values in the pft-physiology file and other pft info in the fsurdat file.

If you need to add your new pfts without replacing existing ones, then add them to the aforementioned files and also modify the clm code so that it may expect the increased number of pfts. In this case, look at how the code works now and follow it as a template when you modify it. You may find this process challenging even if you are a seasoned programmer.Sam Levis