subsetting ndep, lightng, and popdens data at half degree using getregional_datasets.pl

[xgao304]

Dear Scientist,

I successfully ran the cesm 2.1.3. with the compset I2000Clm50BgcCrop at a half degree on the global domain. Here are the relevant input files used:

fatmlndfrc = 'domain.lnd.360x720_cruncep.100429.nc'
fsurdat = 'surfdata_360x720cru_hist_78pfts_CMIP6_simyr2000_c210423.nc' (created using mksurfdata.pl)
stream_fldfilename_ndep = '/ndepdata/fndep_clm_hist_b.e21.BWHIST.f09_g17.CMIP6-historical-WACCM.ensmean_1849-2015_monthly_0.9x1.25_c180926.nc'
stream_fldfilename_lightng = '/datm7/NASA_LIS/clmforc.Li_2012_climo1995-2011.T62.lnfm_Total_c140423.nc'
stream_fldfilename_popdens = '/firedata/clmforc.Li_2017_HYDEv3.2_CMIP6_hdm_0.5x0.5_AVHRR_simyr1850-2016_c180202.nc'

To run the same compset at a half degree over the USA region, I tried to use getregional_datasets.pl to subset these global datasets.
I am able to create the domain and surfdata over the USA domain, but have some issues with the other three data sets (ndep, lightng, and popdens).

Here are my questions:

1. I assume the subsetted ndep, lightng, and popdens data at a half degree are required to run the model over the USA. In other words, the interpolation and subsetting are
not taken care of during the model run. Is that correct? (I tried to run the model with the subsetted domain and surfdata as well as the other three global data sets, but the model run failed with some error message).

2. getregional_datasets.pl requires "fatmlndfrc" is on the same grid as all of the other data files.

for popdens, both global files are at a half degree. However, I found that the longitude mapping is not consistent between two files:

for fatmlndfrc: longitude spans from 0.25 to 359.75; while for popdens: longitude spans from -179.75 to 179.75. If I provide the "-ne 54.75,297.75 -sw 20.75,231.25" for the "getregional_datasets.pl", the resulting longitude domain for popdens is not correct.

for "ndep" and "lightng" files, since both global files are not at the same grid as the "fatmlndfrc", the "getregional_datasets.pl" simply won't generate any corresponding files over the USA domain.

How could I proceed?

Thanks.

 

[oleson]

It is not necessary to subset those datasets for a regional run.
What error are you getting?

 

[xgao304]

The attached is the cesm log file. I don't know much about how to diagnose the error message. The error seems related to
inconsistent vector length and row count. Thanks.

 

[oleson]

What is the create_newcase command you are using for the regional simulation?

 

[xgao304]

create_newcase --case /net/fs05/d1/xgao/cesm2.1.3/cases/BgcCrop2000_USA --compset I2000Clm50BgcCrop --res hcru_hcru --user-mods-dir /net/fs05/d1/xgao/cesm2.1.3/USAdata --machine svante --compiler intel --run-unsupported

-user-modis-dir contains

domain.lnd.360x720_cruncep_USA_c210428.nc
surfdata_360x720cru_hist_78pfts_CMIP6_simyr2000_USA_c210428.nc

user_nl_clm:
fsurdat = '/net/fs05/d1/xgao/cesm2.1.3/USAdata/surfdata_360x720cru_hist_78pfts_CMIP6_simyr2000_USA_c210428.nc'

xmlchange_cmnds:

./xmlchange ATM_DOMAIN_PATH=/net/fs05/d1/xgao/cesm2.1.3/USAdata
./xmlchange LND_DOMAIN_PATH=/net/fs05/d1/xgao/cesm2.1.3/USAdata
./xmlchange ATM_DOMAIN_FILE=domain.lnd.360x720_cruncep_USA_c210428.nc
./xmlchange LND_DOMAIN_FILE=domain.lnd.360x720_cruncep_USA_c210428.nc

 

[oleson]

The longname for your compset is:

2000_DATM%GSWP3v1_CLM50%BGC-CROP_SICE_SOCN_MOSART_CISM2%NOEVOLVE_SWAV

I think you need to use the stub glacier and runoff models for a regional case. Try:

2000_DATM%GSWP3v1_CLM50%BGC-CROP_SICE_SOCN_SROF_SGLC_SWAV

 

[xgao304]

Thanks. It works and the model is running. One more question:

When I use "./query_config --compset clm" to list the available compsets, the choice of the runoff model only
has "MOSART" versus "RTM". Is the "SROF" same as "RTM"? Or now all the corresponding compsets
with RTM should be replaced with "SROF". If so, what is the shortname to use, for example:

I2000Clm50BgcCropRtm : 2000_DATM%GSWP3v1_CLM50%BGC-CROP_SICE_SOCN_RTM_CISM2%NOEVOLVE_SWAV

Thanks.

 

[oleson]

SROF is not the same as RTM. RTM is the older runoff model used with CLM4.5 and old versions. SROF is the "stub" runoff model, meaning runoff will not be routed.

Actually, you might be able to run MOSART for your regional run, you could try it.

 

[xgao304]

I see, about SROF. I tried with MOSART and stub glacier, but it does not work. I got the similar error as before.

create_newcase --case /net/fs05/d1/xgao/cesm2.1.3/cases/BgcCrop2000_Gs_USA --compset 2000_DATM%GSWP3v1_CLM50%BGC-CROP_SICE_SOCN_MOSART_SGLC_SWAV --res hcru_hcru --user-mods-dir /net/fs05/d1/xgao/cesm2.1.3/USAdata --machine svante --compiler intel --run-unsupported

So I guess for my case, "CLM5, BGC with crop on, and over USA domain", I have to use stub runoff and glacier? Certainly, it would be preferable to have runoff model on.

 

[oleson]

I should have said runoff routing model. You will get runoff from the land model, it just won't be routed to the ocean by MOSART. Since you don't have an active ocean, it probably doesn't really matter.

 

[xgao304]

Oh, I see what you meant now. Thanks for the help.