Problem with creating surface data for single-point run on userdefined machine

[schulzm6@staff_uni-marburg_de]

 Dear all, I am trying to create a surface data set for a single-point in CESM 1.2.2 run on a userdefined machine. Therefore, I am following Example 5-5 under this link: http://www.cesm.ucar.edu/models/cesm1.2/clm/models/lnd/clm/doc/UsersGuide/x12938.html mkmapdata.sh creates most of the necessary files, but before the last file "map_1km-merge-10min_HYDRO1K-merge-nomask_to_test1x1_nomask_aave_da_c170216.nc" is created the script stops with the  message "No support for --netcdf4 or UGRID on machines other than yellowstone/caldera/geyser". Other people seem to encounter the same problem: https://bb.cgd.ucar.edu/clm45using-esmfregridweightgen-version-62-tool-create-map-fileformat-netcdf-4-occurs-errors Is there any solution for this problem or is it simply impossible to run a custom single-point run on a userdefined machine? Thanks for any help, Martin

 

[h_duarte@utah_edu]

Hi Martin,I've run into the same problem while creating surface data for a regional run. Did you find a solution/workaround?Best,Henrique

 

[schulzm6@staff_uni-marburg_de]

Hi Henrique,I did not find a real solution. mkmapdata.sh checks wether you are running the script on a NCAR machine or not. If it runs on a NCAR machine, it will go to a hardcoded directory on these machines and run a script there (Search for "MY_ESMF_REGRID" in mkmapdata.sh to find more info).Anyway, this is only the case for CLM 4.5. Therefore I decided to switch to CLM 4.0 by running mkmapdata with the "-p clm4_0" option.BTW: If you are running a single point run, your next error will most probably be "(shr_sys_abort) ERROR: (shr_tInterp_getCosz)  ERROR: lon lat cosz sizes disagree". This is because (on my machine at least) single point runs do not work. Therefore, I switched to a 2x2 run.Cheers,Martin 

 

[felix_wiss@kit_edu]

Hi,I am currently running into the exact same problem...
Right now, I will start trying to change the mkmapdata.sh so that it fits to our machine, but we will see.

Changing to CLM 4.0 is no option for me, as I want to evaluate my source code changes within the models of CLM 4.5.

Felix