error in I1850Clm50Sp compset

[lee]

Good afternoon,

I want to get the dataset for the 1980-2000 period. I used compset is "I1850Clm50Sp (--run-unsupported)",

and I set the run_type = startup; run_startdate = 1980-01-01; stop_option = nmonths; stop_N = 12; run_stopdate = 20010101;resubmit = 19; datm_clmncep_yr_start = 1980; datm_clmncep_yr_end = 2000; clm_accelenrated_spinup = off; clm_force_coldstart = on; job_walkclock_time = 20:00:00; I didn't change any value except these.

The system shows that successfully case. build and case. submit, but there's no output except "case_1850.cism.initial_hist.1980-01-01-00000.nc".

Can you help me tell what the problem is?

Thanks in advance.

 

[oleson]

If you want model output for 1980-2000, then I would suggest either an I2000 compset or an IHISt compset, not I1850.
Look in your run directory for log files, e.g., for the lnd and atm. You can also look at the CaseStatus file in your case directory. These should provide an indication of model run success or failure.
If there are no log files then perhaps the model ran successfully. The output should then be in your archive directory.

 

[lee]

If you want model output for 1980-2000, then I would suggest either an I2000 compset or an IHISt compset, not I1850.
Look in your run directory for log files, e.g., for the lnd and atm. You can also look at the CaseStatus file in your case directory. These should provide an indication of model run success or failure.
If there are no log files then perhaps the model ran successfully. The output should then be in your archive directory.

Thank you so much for your reply!

Now, I ran the I2000Clm50SpGs compset at AU-Cum fluxnet site in Australia (lat/lon: -33.6152/150.7236), and got the daily net radiation during 2013-2014.
But, the simulation data from CLM5 (FSDSND) is different from the observation data from Fluxnet (NETRAD), the correlation is only 0.1, r^2 is 0.01.

I want to ask:
1. why did the two data so different?
2. How I can check the local time and run time and how to adjust time? Is it must the same?

Thank you so much!!! I really hope for the reply!

 

[lee]

If you want model output for 1980-2000, then I would suggest either an I2000 compset or an IHISt compset, not I1850.
Look in your run directory for log files, e.g., for the lnd and atm. You can also look at the CaseStatus file in your case directory. These should provide an indication of model run success or failure.
If there are no log files then perhaps the model ran successfully. The output should then be in your archive directory.

I also changed the ccsm_co2_ppmv value from 367 to 390;
coldstart to on;
GSWP3v1, -bgc sp

Thank you !

 

[oleson]

If NETRAD is indeed the observed net radiation, then FSDSND wouldn't be the right CLM variable to compare to. As the long_name attribute for that variable indicates, FSDSND is the "direct nir incident solar radiation". You should be able to calculate net radiation in CLM as FSA-FIRA (the absorbed (net) solar radiation minus the net longwave radiation).

 

[lee]

If NETRAD is indeed the observed net radiation, then FSDSND wouldn't be the right CLM variable to compare to. As the long_name attribute for that variable indicates, FSDSND is the "direct nir incident solar radiation". You should be able to calculate net radiation in CLM as FSA-FIRA (the absorbed (net) solar radiation minus the net longwave radiation).

Thank you so much for your reply! I will try this calculation.

For this site (AU-Cum), I also compared the sensible heat flux in CLM5 (FSH_G) and FLUXNET (H_F_MDS), the correlation is 0.58; the latent heat flux in CLM5 (EFLX_LH_TOT) and FLUXNET (LE_F_MDS), the correlation is 0.52 during 2013-2014; and the correlation of latent heat flux of AU-Rob site in Australia is 0.38 for 2014.

I want to ask:
1. Are the low correlations normal? If so, what is the reason? Does it mean the simulation of model performance bad?
2. I also want to ask if it is right to change the ccsm_co2_ppmv? Is the default value (367) suitable for any run time such as 1990, 2000, 2010?

Thank you !!

FLUXNET

 

[oleson]

For sensible heat flux, use FSH. FSH_G is the sensible heat flux from the ground only.
The correlations seem low. Are you forcing the model with the observed meteorology at the site or with GSWP3 data (the default forcing data for the I2000Clm50SpGs compset)? Correlations will likely be lower with GSWP3 data.
Regarding CO2, you can use an appropriate value for whatever year you are simulating. 367 is probably a reasonable value for year 2000.

 

[lee]

For sensible heat flux, use FSH. FSH_G is the sensible heat flux from the ground only.
The correlations seem low. Are you forcing the model with the observed meteorology at the site or with GSWP3 data (the default forcing data for the I2000Clm50SpGs compset)? Correlations will likely be lower with GSWP3 data.
Regarding CO2, you can use an appropriate value for whatever year you are simulating. 367 is probably a reasonable value for year 2000.

Thank you for your reply!
Exactly, I used the default forcing data for the CLM5 compset with GSWP3 data. I can run the model and get the output as for the default model.

Now, I want to run my own forcing data with GLDAS or meteorological observed data, but I cannot find the relative tutorial.
I want to learn in detail from creating new case to submit case with observed forcing data.

I'm looking forward to your reply,
Thank you!

 

[oleson]

You could start with the NEON single-point tutorial here:

Welcome to NEON Tutorial Jupyter Book — NEON Workshop Tutorial Book

as mentioned in this CTSM discussion item:

Tutorial opportunities for CTSM · ESCOMP CTSM · Discussion #1855

We've received a number of requests to launch the cloud environment from the CTSM mini-tutorial. This cloud environment has been taken down, but there are still ways to learn about CTSM. Here are s...

github.com

This discussion item may also be helpful:

How do I create a single-point run with the NUOPC (default) coupler? · ESCOMP CTSM · Discussion #1833

We get this question a lot: I am trying to set up a case for a specific latitude and longitude. I want to use GSWP3 climate data, how do I set this up?

github.com