How can I create Pre-industry surface data for CLM?

[Dong Peng]

Hello everyone, I met some issues about creating pre-industry surface data for clm. I tried to type

-perl mksurfdata.pl -r 0.9x1.25 -years 1750
-** Invalid simulation year: 1750 # here reported an error to stop running.

- perl mksurfdata.pl -h

OPTIONS

-years [or -y] "years" Simulation year(s) to run over (by default 1750) # where use $opts{'years'} to obtain this 1750 value in mksurfdata.pl.
(can also be a simulation year range: i.e. 1850-2000)

Here I found the default simulation year was 1750. But, when I try to command -perl mksurfdata.pl -r 0.9x1.25 . It was actually run in 1850.

Has anybody tried to run for pre-industry surfdata?

 

[erik]

There's actually a simple fix to this. We limited the years you can run mksurfdata.pl for with the thinking that it would help eliminate typo's or mistakes in the files being made.

What you need to do is edit the file namelist_definition_ctsm.xml under bld/namelist_files for the sim_year entry which looks like...

<entry id="sim_year" type="char*4" category="default_settings"
       group="default_settings" valid_values=
"PtVg,1000,850,1100,1350,1600,1850,1855,1865,1875,1885,1895,1905,1915,1925,1935,1945,1955,1965,1975,1979,1980,1982,1985,1995,2000,2005,2010,2013,2015,2018,2025,2035,2045,2055,2065,2075,2085,2095,2105">
Year to simulate and to provide datasets for (such as surface datasets, initial conditions, aerosol-deposition, Nitrogen deposition rates etc.)
A sim_year of 1000 corresponds to data used for testing only, NOT corresponding to any real datasets.
A sim_year greater than 2015 corresponds to ssp_rcp scenario data
A sim_year of PtVg refers to the Potential Vegetation dataset, that doesn't include human influences
Most years are only used for clm_tools and there aren't CLM datasets that correspond to them.
CLM datasets exist for years: 1000 (for testing), 1850, and 2000

</entry>

As you can see 1750 isn't in the list, so you just need to add it to the above list.

We decided to go away from these types of restrictions in the newest version of the mksurfdata tool which will come out with CTSM5.2.

We also did make the above kind of change for the TRENDY datasets which start in 1700.

 

[Dong Peng]

There's actually a simple fix to this. We limited the years you can run mksurfdata.pl for with the thinking that it would help eliminate typo's or mistakes in the files being made.

What you need to do is edit the file namelist_definition_ctsm.xml under bld/namelist_files for the sim_year entry which looks like...

<entry id="sim_year" type="char*4" category="default_settings"
       group="default_settings" valid_values=
"PtVg,1000,850,1100,1350,1600,1850,1855,1865,1875,1885,1895,1905,1915,1925,1935,1945,1955,1965,1975,1979,1980,1982,1985,1995,2000,2005,2010,2013,2015,2018,2025,2035,2045,2055,2065,2075,2085,2095,2105">
Year to simulate and to provide datasets for (such as surface datasets, initial conditions, aerosol-deposition, Nitrogen deposition rates etc.)
A sim_year of 1000 corresponds to data used for testing only, NOT corresponding to any real datasets.
A sim_year greater than 2015 corresponds to ssp_rcp scenario data
A sim_year of PtVg refers to the Potential Vegetation dataset, that doesn't include human influences
Most years are only used for clm_tools and there aren't CLM datasets that correspond to them.
CLM datasets exist for years: 1000 (for testing), 1850, and 2000

</entry>

As you can see 1750 isn't in the list, so you just need to add it to the above list.

We decided to go away from these types of restrictions in the newest version of the mksurfdata tool which will come out with CTSM5.2.

We also did make the above kind of change for the TRENDY datasets which start in 1700.

Thank you for your help.

 

[liliyao]

There's actually a simple fix to this. We limited the years you can run mksurfdata.pl for with the thinking that it would help eliminate typo's or mistakes in the files being made.

What you need to do is edit the file namelist_definition_ctsm.xml under bld/namelist_files for the sim_year entry which looks like...

<entry id="sim_year" type="char*4" category="default_settings"
       group="default_settings" valid_values=
"PtVg,1000,850,1100,1350,1600,1850,1855,1865,1875,1885,1895,1905,1915,1925,1935,1945,1955,1965,1975,1979,1980,1982,1985,1995,2000,2005,2010,2013,2015,2018,2025,2035,2045,2055,2065,2075,2085,2095,2105">
Year to simulate and to provide datasets for (such as surface datasets, initial conditions, aerosol-deposition, Nitrogen deposition rates etc.)
A sim_year of 1000 corresponds to data used for testing only, NOT corresponding to any real datasets.
A sim_year greater than 2015 corresponds to ssp_rcp scenario data
A sim_year of PtVg refers to the Potential Vegetation dataset, that doesn't include human influences
Most years are only used for clm_tools and there aren't CLM datasets that correspond to them.
CLM datasets exist for years: 1000 (for testing), 1850, and 2000

</entry>

As you can see 1750 isn't in the list, so you just need to add it to the above list.

We decided to go away from these types of restrictions in the newest version of the mksurfdata tool which will come out with CTSM5.2.

We also did make the above kind of change for the TRENDY datasets which start in 1700.

Hi Erik, I have a related question on this topic. I would like to make 2020 as the first year of my future scenario (e.g., SSP5-8.5), and generate surface and land use timeseries files for the year range 2020-2100 by using this command: /mksurfdata.pl -res 0.125nldas2 -years 2020-2100 -ssp_rcp SSP5-8.5

I have modified the sim_year and sim_year_range entry in my_cesm_sandbox_2_2_0/components/clm/bld/namelist_files/namelist_definition_ctsm.xml

but the errors I get is:

ERROR: data file for this resolution does NOT exist /inputdata/cesm_inputdata/lnd/clm2/rawdata/pftcftdynharv.0.25x0.25.SSP1-2.6.simyr2016-2100.c181217/mksrf_landuse_SSP1RCP26_clm5_2020.c181217.nc

I am not sure why it is looking for SSP1-2.6 file when I have specified that I need SSP5-8.5 scenario. Could you please let me know how to fix this error? Thank you very much.

 

[oleson]

@erik

 

[erik]

So this is a problem with the cesm2.2.0 release version?

I tried the above in cesm2.2.0 release for CTSM and didn't see the problem you are referring to.

the SSP1-2.6 problem has to be this issue:

2100 surface datasets created using mksurfdata.pl are only SSP1-2.6 · Issue #955 · ESCOMP/CTSM

Brief summary of bug The 2100 fsurdat files created with release-clm5.0.31 are only for SSP1-2.6. General bug information CTSM version you are using: release-clm5.0.31 Does this bug cause significa...

github.com

As such I wonder if you shouldn't use a more recent version of mksurfdata_map to get what you want to do. I don't expect mksurfdata to work in each model version of CTSM, but we have made surface datasets in the 5.0 release versions (release-clm5.0.18, release-clm5.0.24, release-clm5.0.30, and release-clm5.0.34) and CTSM versions (ctsm5.1.dev030, ctsm5.1.dev038, ctsm5.1.dev052, ctsm5.1.dev116). So those versions would be ones that I'd recommend and in general I suggest using the latest one from that set.

 

[liliyao]

So this is a problem with the cesm2.2.0 release version?

I tried the above in cesm2.2.0 release for CTSM and didn't see the problem you are referring to.

the SSP1-2.6 problem has to be this issue:

2100 surface datasets created using mksurfdata.pl are only SSP1-2.6 · Issue #955 · ESCOMP/CTSM

Brief summary of bug The 2100 fsurdat files created with release-clm5.0.31 are only for SSP1-2.6. General bug information CTSM version you are using: release-clm5.0.31 Does this bug cause significa...

github.com

As such I wonder if you shouldn't use a more recent version of mksurfdata_map to get what you want to do. I don't expect mksurfdata to work in each model version of CTSM, but we have made surface datasets in the 5.0 release versions (release-clm5.0.18, release-clm5.0.24, release-clm5.0.30, and release-clm5.0.34) and CTSM versions (ctsm5.1.dev030, ctsm5.1.dev038, ctsm5.1.dev052, ctsm5.1.dev116). So those versions would be ones that I'd recommend and in general I suggest using the latest one from that set.

Thank you, Erik. Yes, I was using release-cesm2.2.0. unfortunately, I got the same error when I tried in ctsm5.1.dev116.

 

[oleson]

I can't replicate this problem either with release-cesm2.2.0. I added 2020-2100 to sim_year_range in namelist_definition_ctsm.xml and then ran ./mksurfdata.pl -res 0.125nldas2 -years 2020-2100 -ssp_rcp SSP5-8.5. That seems to be using SSP5-8.5 datasets, e.g.,

/glade/p/cesm/cseg/inputdata/lnd/clm2/rawdata/pftcftdynharv.0.25x0.25.SSP5-8.5.simyr2016-2100.c171005/mksrf_landuse_SSP5RCP85_clm5_2047.c171005.nc

 

[liliyao]

It's interesting. Both my colleague and I have the same problem. I have checked the /bld/CLMBuildNamelist.pm and tools/mksurfdata_map/mksurfdata.pl. Those modifications aiming to solve the SSP1-2.6 problem have been included. I am not sure why we still have the problem.

 

[liliyao]

I also have the SSP1-2.6 problem when I tried in release-clm5.0.34.

 

[oleson]

Can you attach your namelist_definition_ctsm.xml ?

 

[liliyao]

Hi Dr. Oleson, it is my namelist_definition_ctsm.xml from the release-cesm2.2.0 Thanks!

 

[oleson]

I get the same correct results when I use your namelist_definition_ctsm.xml in release-cesm2.2.0.
I guess I would look to see if you've inadvertently made any other changes to the model namelist files, e.g., with git status perhaps.

 

[liliyao]

from my_cesm_sandbox_2_2_0/components/clm/bld/namelist_files, I ran:
git config core.filemode false
git status > git_status.txt
It seems only the namelist_definition_ctsm.xml is the namelist file related to our current problem. I ran:
git diff namelist_definition_ctsm.xml > namelist_definition_ctsm.xml.diff

Both the git_status.txt and namelist_definition_ctsm.xml.diff are attached. Thank you very much.

 

[oleson]

What does git diff mksurfdata.pl show?

 

[liliyao]

It is the output of git diff mksurfdata.pl

 

[oleson]

Ok, thanks, well I guess I'm out of ideas at the moment since I can't replicate the problem even when using your xml file.

 

[liliyao]

Okay, thank you. I really appreciate your kindness and patience with me.

 

[oleson]

One more shot in the dark. Does this directory exist?

inputdata/lnd/clm2/rawdata/pftcftdynharv.0.25x0.25.SSP5-8.5.simyr2016-2100.c171005

Is it a regular directory or is it linked to something else maybe?

 

[liliyao]

This directory exits. It is a regular directory: /global/project/projectdirs/m2702/liliyao/inputdata/cesm_inputdata/lnd/clm2/rawdata/pftcftdynharv.0.25x0.25.SSP5-8.5.simyr2016-2100.c171005