Welcome to the new DiscussCESM forum!
We are still working on the website migration, so you may experience downtime during this process.

Existing users, please reset your password before logging in here: https://xenforo.cgd.ucar.edu/cesm/index.php?lost-password/

iCESM land unit mismatch

Dear all, i was using iCESM to do a startup run using my old experiments restart file and i followed the instructions here. NCAR/iCESM1.2
However, when I use the new initial file, I got the following error:


application called MPI_Abort(comm=0xC4000006, 1) - process 13
check_dim ERROR: mismatch of input dimension 2494 with expected value
3098 for variable landunit


What does this error imply ?

Thanks very much

Xiao
 

nusbaume

Jesse Nusbaumer
CSEG and Liaisons
Staff member
Hi Xiao,

That error implies that your CLM restart file somehow does not match up with what your iCESM1 case expects. This could be due to a difference in resolution, time period (e.g 1850 vs 2000), land/sea mask, etc. between the iCESM1 case you created and the restart file you are trying to use. Without knowing the model configurations you have, one possible solution is to simply interpolate your input CLM restart file to match what iCESM1 wants. This can be done by using the "interpinic" program, which is found in the iCESM1.2 source code here:

iCESM1.2/models/lnd/clm/tools/clm4_0/interpinic

There is a README file in that directory that describes how to use the program. You can also find instructions online here:


Finally, if you aren't sure what kind of file the case needs, then simply try a startup run instead of a branch or hybrid run, and see what file is used for "finidat" in the iCESM case's lnd_in file (which will be in the run directory).

Hope that helps, and good luck with the interpolation!

Jesse
 
Thanks very much for the reply, Jesse,

I have tried the interpinic program, but the error is still there. I also tried a startup run.
However, another error showed up:

incompatible domain grid coordinates

After I printed out the lat and long values, the deviations were quite small as the following shown:


n= 1 data1= 38.9666104694540 data2=
38.9666103431570 diff= 1.262970528159713E-007 eps=
1.000000000000000E-012
n= 2 data1= 38.9666104694540 data2=
38.9666103431570 diff= 1.262970528159713E-007 eps=
1.000000000000000E-012


I am using the same domain file for CAM and CLM and I searched for the forum, it looks like a wrap around error.
Should I change some compiling options or should I generate new domain file?


Thanks very much

Xiao





Hi Xiao,

That error implies that your CLM restart file somehow does not match up with what your iCESM1 case expects. This could be due to a difference in resolution, time period (e.g 1850 vs 2000), land/sea mask, etc. between the iCESM1 case you created and the restart file you are trying to use. Without knowing the model configurations you have, one possible solution is to simply interpolate your input CLM restart file to match what iCESM1 wants. This can be done by using the "interpinic" program, which is found in the iCESM1.2 source code here:

iCESM1.2/models/lnd/clm/tools/clm4_0/interpinic

There is a README file in that directory that describes how to use the program. You can also find instructions online here:


Finally, if you aren't sure what kind of file the case needs, then simply try a startup run instead of a branch or hybrid run, and see what file is used for "finidat" in the iCESM case's lnd_in file (which will be in the run directory).

Hope that helps, and good luck with the interpolation!

Jesse
 
So, I checked the domain file, indeed, it is a wrap around error problem. After i updated to correct domain file. The model still showed problem.
The following is what I have tried.

Using the new generated CLM file by iclm4_IC_create.ncl to start 1) a hybrid run or 2) a startup run and 3) using iCESM PI CLM restart file interpolated using interpinic. In the first case, the model crashed due to 'floating devided by zero' error and the the other two cases, the model crashed due to land unit mismatch error. Any suggestions where I should heading ?

Thanks very much

Xiao
 

nusbaume

Jesse Nusbaumer
CSEG and Liaisons
Staff member
Hi Xiao,

Apologies for not seeing this post earlier! Can you send me your cesm.log.XXX and lnd.log.XXX files from the failed startup run (both should be present in your run directory), along with the lnd_in file from the iCESM run and, if relevant, the run that was used to generate the original restart files? My guess is that there is a mis-match in the input files somewhere, but until I can look at the logs I won't be able to say for sure.

Thanks, and have a great day!

Jesse
 
Top