Running CLM 4.5 out of CESM

[jlechuga@uni-bonn_de]

From the University of Bonn (Germany), we have been using CLM, version 3.5, into our meteorological package TerrSysMP (http://tr32new.uni-koeln.de/index.php/component/comprofiler/userslist/Project%20Z4).

Nowadays, our aim consists in the update to CLM 4.5, which we would like to include into TerrSysMP, however, we have noticed that, as-is, it is difficult to run CLM 4.5 outside the framework of CESM 1.2.2.

Is it possible to obtain a standalone CLM 4.5 version which can be used outside the CESM framework in order to include it into our TerrSysMP package?

Best regards,
Dr. Javier Lechuga

 

[yudong_tian@nasa_gov]

I have the same question. We want to put CLM 4.5 inside our own Land Information System framework at NASA.

 

[jlechuga@uni-bonn_de]

I keep on working on it. It is possible to achieve a coupling through oasis-mct library. At the moment, I am having difficulties to run the system into a Blue gene/Q machine we have here. Which is exactly your question? Maybe I have solve it already.

 

[yudong_tian@nasa_gov]

Thanks for your response. I think our situation is similar to yours. We used to have an earlier version of CLM (2.x?) imbedded in our own modeling framework. But now CLM4.5 has to run within CESM, and this makes it unclear how we can imbed it in our framework without CESM.  --Yudong

 

[jlechuga@uni-bonn_de]

Hello yudong,Using CESM 1.2.2, I found the main program file into  ~/models/drv/driver/ccsm_driver.F90 (from there and ccsm_comp_mod.F90, in the same directory), it is possible to start a coupling interface. CLM 4.5 main loop is found at: ~/models/lnd/clm/src/lnd_comp_mct.F90 You might start from there. To get a cesm.exe with CLM 4.5 alone, from ~/script directory, you may do something like: ./create_newcase -case $CASEROOT -mach $MACH -compset I_2000_CLM45 -res f45_g37 Then, you can get your files into the run directory *_in files, to generate your particular input files case.Regards,Javier

 

[dlawren]

Hi Yudong and Javier,We are aware of the need from the community to be able to run CLM outside of CESM coupled with either another atmosphere model or another data atmosphere model.  We are working on providing this capability, with some progress, but the code is not yet ready to be provided to collaborators. Unfortunately, we don't have a date when it will become available.  About all I can say at this point is that we will likely be able to do this at some point in the not-too-distant future.  Please check back on the bulletin board or send an email to me at dlawren@ucar.edu if you are curious about progress.  The timeline is at least order of months.Dave Lawrence, LMWG co-chair

 

[yudong_tian@nasa_gov]

Dave,   Thanks for the update. This is helpful. I am wondering, if you guys have done some standalone tests of CLM 4.5 before integrating it into CESM? If so, the standalone test cases will be helpful for my purpose. Yudong

 

[dlawren]

CLM can be run offline from within the CESM framework (and I compset in CESM language).  We have run many offline simulations and have posted a few of them on the CLM4.5 webpage. http://www.cesm.ucar.edu/models/cesm1.2/clm/

 

[yudong_tian@nasa_gov]

Thanks. I am trying to reproduce some of the runs listed in http://www.cesm.ucar.edu/experiments/cesm1.2/diagnostics/clm_diag.html, for example:1° CLM45SP
clm45sp_1deg4502_hist
CRUNCEP but I can not find the 1-deg grid definition. I tried the following command but it did not work:scripts/create_newcase -case $CASEROOT/CLM4.5SP -mach discover -compset I1850CLM45 -res a%rx1_l%rx1_oi%rx1_r%rx1_m%rx1_g%null_w%null

Can I have a sample case script and user_nl_clm file so I can compare my output with the plots you posted? Thanks,Yudong Tian    

 

[dlawren]

The resolution that you want to use if f09_g16 , e.g. in your create_newcase command use  -res f09_g16 (this is the nominal 1deg resolution).fsurdat = '/glade/p/cesmdata/cseg/inputdata/lnd/clm2/surfdata_map/surfdata_0.9x1.25_simyr1850_c130927.nc'

 

[jlechuga@uni-bonn_de]

Just arrived from a meeting in Garmisch-Partenkirchen, south Germany, Gabrielle De Lannoy and Chris Snyder came to visit.
Nice to see you, guys

 

[yudong_tian@nasa_gov]

Dave,   Thanks for your help, I can now succesffully run CLM4.5SP, created by the following command: create_newcase -case $CASEDIR/CLM4.5SP -mach discover -compset I1850CLM45 -res f09_g16I've run it for over 40 years, and here is the plot of the global mean TSA:http://sigma.umd.edu/ytian/Projects/NLDAS/CESM/CLM4.5SP/tsa-annual.pngbut it is quite different from what was posted on CESM's website:http://www.cgd.ucar.edu/tss/clm/diagnostics/clm4.5_dev/clm45sp_1deg4502_hista-clm40sp_1deg4502_hista/set1/set1_TSA.gifWhat do you think caused the big difference? Maybe my configuration is still different? How can I reproduce what was posted on the website? Or can my local run output be trusted as benchmark data?BTW, I saw the following surfdata was used in my run:CESM_INPUT_DATA/lnd/clm2/surfdata_map/surfdata_0.9x1.25_simyr1850_c130415.nc Thanks,Yudong 

 

[yudong_tian@nasa_gov]

Dave and all,   Would you share your thoughts and insights on what I described above? Thanks,Yudong 

 

[dlawren]

I expect that this is simply reflecting a different forcing dataset.  I think you are using the "QIAN" forcing and the runs you are showing in the standard CLM plots are with CRUNCEP forcing.See the difference with these two forcing datasets here.
http://www.cgd.ucar.edu/tss/clm/diagnostics/clm4.5_dev/clm45sp_1deg4502_hista-clm45sp_1deg4502_histqiana/set1/set1_TSA.gif

 

[yudong_tian@nasa_gov]

Javier -- how's progress of transplanting CLM on your end? I've been able to write a light weight driver of my own to make CLM read the parameter files and do the initialization. Now I am trying to figure out the plumbing for the atmospheric forcing data to reach CLM.   
--Yudong

 

[zp_xie@foxmail_com]

Hi, Yudong, do you have any suggestion if I want to add a new variable into the atmospheric forcing data. I am comfused about the processes of the atmospheric forcing data to CLM. Thanksvery much