data size error in run regional simulation

[xiaoxiaokuishu]

Hi，all

I run a regional simulation with clm5.0 at 0.5deg,

./create_newcase --case /work/n02/n02/ruxu/cesm/CESM/runs/test-region.20250108-1520 --compset 2000_DATM%GSWP3v1_CLM50%BGC-CROP_SICE_SOCN_MOSART_SGLC_SWAV --res CLM_USRDAT -run-unsupported


The regional surfdata is create with the CTSM5.1

./subset_data region --create-surface --create-mesh --create-domain --create-user-mods --lat1 -16 --lat2 16 --lon1 100 --lon2 130 --overwrite --verbose --inputdata-dir /work/n02/n02/ruxu/cesm/CESM/cesm_inputdata --outdir /work/n02/n02/ruxu/cesm/CESM/cesm_inputdata/ my_data_ctsm5.1



The setting for the experiment is as follow,


MYCSMDATA=/work/n02/n02/ruxu/cesm/CESM/cesm_inputdata/my_data_ctsm5.1

./xmlchange ATM_DOMAIN_PATH=$MYCSMDATA,LND_DOMAIN_PATH=$MYCSMDATA

./xmlchange ATM_DOMAIN_FILE=domain.lnd.fv0.9x1.25_gx1v7_100.0-130.0_-16.0-16.0_c250101.nc

./xmlchange LND_DOMAIN_FILE=domain.lnd.fv0.9x1.25_gx1v7_100.0-130.0_-16.0-16.0_c250101.nc

./xmlchange LND_NX=34

./xmlchange LND_NY=25

./xmlchange ATM_NX=34

./xmlchange ATM_NY=25



Then I mofify the path under the below files: (all the files are attached):

user_datm.streams.txt.CLMGSWP3v1.Precip

user_datm.streams.txt.CLMGSWP3v1.Solar

user_datm.streams.txt.CLMGSWP3v1.TPQW



When I start the simulation, the atm.log reported the below error:

(shr_dmodel_readstrm) file lb: /work/n02/n02/ruxu/cesm/CESM/cesm_inputdata//atm/datm7/atm_forcing.datm7.GSWP3.0.5d.v1.c170516/Solar/clmforc.GSWP3.c2011.0.5x0.5.Solr.1993-01.nc 1

(shr_dmodel_readstrm) ERROR in data sizes 720 360 248 850

ERROR: (shr_dmodel_readstrm) ERROR in data sizes


It seems that the forcing data is global, but I run regional simulation, according to our previous discussion, the regional simulation should use the global forcing? I am wonder if I miss some settings specificlly for regional simulation. When I set --res CLM_USRDAT, the same error existed.


Best

 

[oleson]

I think the problem is that you've replaced the GSWP3 domain file in your user datm streams with this file: domain.lnd.fv0.9x1.25_gx1v7_100.0-130.0_-16.0-16.0_c250101.nc. You should use the default GSWP3 global domain file. And if you are using the default GSWP3 forcing you shouldn't need to create user datm streams at all.

 

[xiaoxiaokuishu]

I think the problem is that you've replaced the GSWP3 domain file in your user datm streams with this file: domain.lnd.fv0.9x1.25_gx1v7_100.0-130.0_-16.0-16.0_c250101.nc. You should use the default GSWP3 global domain file. And if you are using the default GSWP3 forcing you shouldn't need to create user datm streams at all.

Hi, Oleson,


I am very confused about regional simulation, now my understand is that regional simulation
only set the surfdata as regional, the domain file and forcing should be global?
Now i change the domain files to the default one (it is a global dataset), but the lnd.log reported the below error:

sing /work/n02/n02/ruxu/cesm/CESM/cesm_inputdata/my_data_ctsm5.1/surfdata_100.0-130.0_-16.0-16.0_hist_16pfts_Irrig_CMIP6_simyr2000_c250101.nc
ncd_getiodesc ERROR in vsize 850 259200 0
ERROR: ERROR in /mnt/lustre/a2fs-work2/work/n02/n02/ruxu/cesm/CESM/clm5.0/src/main/ncdio_pio.F90.in at line 2388

So, i am not sure the key point of making a regional simulation...

Best
Ru

 

[oleson]

You should be able to run a regional simulation using the global forcing data.
The error indicates that the size of the surface dataset (number of lats X number of lons = 25X34 = 850) doesn't equal the size of the grid you are specifying with the domain file. The 259200 is 360 X 720 which is the size of the GSWP3 grid. I see that you are specifying the domain file correctly I think:

fatmlndfrc = '/work/n02/n02/ruxu/cesm/CESM/cesm_inputdata/my_data_ctsm5.1/domain.lnd.fv0.9x1.25_gx1v7_100.0-130.0_-16.0-16.0_c250101.nc'

What is the size of the domain file?
So, I'm not sure what is going on. Can you attach your lnd, atm, and cesm logs from your most recent simulation?

 

[xiaoxiaokuishu]

You should be able to run a regional simulation using the global forcing data.
The error indicates that the size of the surface dataset (number of lats X number of lons = 25X34 = 850) doesn't equal the size of the grid you are specifying with the domain file. The 259200 is 360 X 720 which is the size of the GSWP3 grid. I see that you are specifying the domain file correctly I think:

fatmlndfrc = '/work/n02/n02/ruxu/cesm/CESM/cesm_inputdata/my_data_ctsm5.1/domain.lnd.fv0.9x1.25_gx1v7_100.0-130.0_-16.0-16.0_c250101.nc'

What is the size of the domain file?
So, I'm not sure what is going on. Can you attach your lnd, atm, and cesm logs from your most recent simulation?

Hi, Oleson,

I attached all the possible files for your reference.

 

[oleson]

The lnd_in file from the newest attachment you sent indicates you are using this:

fatmlndfrc = '/work/n02/n02/ruxu/cesm/CESM/cesm_inputdata//share/domains/domain.clm/domain.lnd.360x720_cruncep.100429.nc'

which would not work for the surface dataset you are using:

fsurdat = "/work/n02/n02/ruxu/cesm/CESM/cesm_inputdata/my_data_ctsm5.1/surfdata_100.0-130.0_-16.0-16.0_hist_16pfts_Irrig_CMIP6_simyr2000_c250101.nc"

 

[xiaoxiaokuishu]

The lnd_in file from the newest attachment you sent indicates you are using this:

fatmlndfrc = '/work/n02/n02/ruxu/cesm/CESM/cesm_inputdata//share/domains/domain.clm/domain.lnd.360x720_cruncep.100429.nc'

which would not work for the surface dataset you are using:

fsurdat = "/work/n02/n02/ruxu/cesm/CESM/cesm_inputdata/my_data_ctsm5.1/surfdata_100.0-130.0_-16.0-16.0_hist_16pfts_Irrig_CMIP6_simyr2000_c250101.nc"

Hi, Oleson,

my first post is like this:

./xmlchange ATM_DOMAIN_PATH=$MYCSMDATA,LND_DOMAIN_PATH=$MYCSMDATA

./xmlchange ATM_DOMAIN_FILE=domain.lnd.fv0.9x1.25_gx1v7_100.0-130.0_-16.0-16.0_c250101.nc

./xmlchange LND_DOMAIN_FILE=domain.lnd.fv0.9x1.25_gx1v7_100.0-130.0_-16.0-16.0_c250101.nc

./xmlchange LND_NX=34

./xmlchange LND_NY=25

./xmlchange ATM_NX=34

./xmlchange ATM_NY=25

I already set domain file as regional, but there is still mismatch issues..
See my first post....

Best
Ru

 

[oleson]

Look at the lnd log you sent (lnd.log.8475782.250110-102702):

surface data = /work/n02/n02/ruxu/cesm/CESM/cesm_inputdata/my_data_ctsm5.1/surfdata_100.0-130.0_-16.0-16.0_hist_16pfts_Irrig_CMIP6_simyr2000_c250101.nc
land frac data = /work/n02/n02/ruxu/cesm/CESM/cesm_inputdata//share/domains/domain.clm/domain.lnd.360x720_cruncep.100429.nc

lnd.log.8475782.250110-102702

That's where your mismatch is.
Maybe your xmlchange commands are not having the desired effect.
You can do an xmlquery to see what they are actually set to.

 

[xiaoxiaokuishu]

Look at the lnd log you sent (lnd.log.8475782.250110-102702):

surface data = /work/n02/n02/ruxu/cesm/CESM/cesm_inputdata/my_data_ctsm5.1/surfdata_100.0-130.0_-16.0-16.0_hist_16pfts_Irrig_CMIP6_simyr2000_c250101.nc
land frac data = /work/n02/n02/ruxu/cesm/CESM/cesm_inputdata//share/domains/domain.clm/domain.lnd.360x720_cruncep.100429.nc

lnd.log.8475782.250110-102702

That's where your mismatch is.
Maybe your xmlchange commands are not having the desired effect.
You can do an xmlquery to see what they are actually set to.

Hi, Oleson,

Thanks for the support!!! I re-run the simulation, now the regional-simuation works.. even I did not notice which item i changed..

Besides, I change compset from
2000_DATM%GSWP3v1_CLM50%BGC-CROP_SICE_SOCN_MOSART_SGLC_SWAV
to
2000_DATM%GSWP3v1_CLM50%BGC-CROP_SICE_SOCN_SROF_SGLC_SWAV

the simulation works smoothly, if i use the 2000_DATM%GSWP3v1_CLM50%BGC-CROP_SICE_SOCN_MOSART_SGLC_SWAV,
there is also some tricky errors....

Best
Ru

 

[oleson]

Glad you got it working. Yes, for regional simulations, it is recommended to use the stub runoff model (SROF), not MOSART, because it would require a lot of effort/datasets to get MOSART to run regionally.