spin up IHIST case

[Fuhow]

Hi, I am using cesm 2.2.0 and I want to create an IHIST case, so that I can define time series landuse and monthly LAI from 1980 to 2020 manually.

I followed the way to create I2000Clm50BgcCropGs case, then I create HIST_DATM%GSWP3v1_CLM50%BGC-CROP_SICE_SOCN_MOSART_SGLC_SWAV, using ESMF to create regional surfdata, setting spin-up, but I have some questions about spin-up IHIST case:

1. In user_nl_clm file, before I set 'landuse_timeseries', it shows that missing this file during build, after I download this file directly from https://svn-ccsm-inputdata.cgd.ucar...ftdyn_0.9x1.25_hist_simyr1850-2005_c110114.nc, it shows an error code in lnd log, whether I should make a regional landuse_timeseries file using ESMF?

2. I want to run CLM using my own land use and MONTHLY_LAI by modifying transient land file, instead of the dynamic Bgc like case, is it correct for me to use an IHIST case?

Thanks for any advice.

 

[slevis]

1) I do not know what error you got, so I do not have suggestions.
2) IHist is a valid choice if you are interested in reading a landuse file to update pft distributions over time.

 

[Fuhow]

1) I do not know what error you got, so I do not have suggestions.
2) IHist is a valid choice if you are interested in reading a landuse file to update pft distributions over time.

Hi, slevis, thanks for your sugestion.

For Q1, sorry I didn't label the error, as shown emphasised below which is the end of my lnd.run.log. It is interrupted after opening existing landuse_timeseries file with a code of 524288, and cesm.run.log is in initialize lnd stage, the submit quited without output. So, I think this might be the reson of landuse_timeseries file, should I transfer this file for regional or single point? (I am testing a single point now)

For Q2, I think your point fits well with mine.

Thanks a lot.

 

[slevis]

@Fuhow this is not an error. It just shows where the lnd.log stopped. When that happens, usually you will find an error in one of the other .log. files, for example cesm.log, etc.

 

[slevis]

...but you say that you do not see an explicit error in any of the log files.That's difficult to troubleshoot.

 

[Fuhow]

...but you say that you do not see an explicit error in any of the log files.That's difficult to troubleshoot.

Hi, slevis @slevis, I attached my log file for I am using a lnd only case. As shown in the cesm log file, the process was aborted. I found there is a same aborted code(CESM2 stop running after reading regrid_vegetation.nc), and jedwards said it might be the reason of stack size setting (CESM2 stop running after reading regrid_vegetation.nc).

So, my questions are:

1. As mentioned before, whether should I transfer the raw global landuse_timeseries file to a regional or single point file using ESMF, or the raw landuse timeseries file is suitable for a single point case?

2. My stacksize setting in config_machine.xml is only 256M, am I point to the right place? I will try this setting to unlimited soon.

Many thanks for your replying.

(cesm log file aborted)


(the same aborted code before)


(my stack size setting in config_machine.xml)

 

[Fuhow]

Hi, slevis @slevis, I attached my log file for I am using a lnd only case. As shown in the cesm log file, the process was aborted. I found there is a same aborted code(CESM2 stop running after reading regrid_vegetation.nc), and jedwards said it might be the reason of stack size setting (CESM2 stop running after reading regrid_vegetation.nc).

So, my questions are:

1. As mentioned before, whether should I transfer the raw global landuse_timeseries file to a regional or single point file using ESMF, or the raw landuse timeseries file is suitable for a single point case?

2. My stacksize setting in config_machine.xml is only 256M, am I point to the right place? I will try this setting to unlimited soon.

Many thanks for your replying.

(cesm log file aborted)
View attachment 6298

(the same aborted code before)
View attachment 6299

(my stack size setting in config_machine.xml)
View attachment 6300

Dear @slevis, I have tried to set OMP_STACKSIZE to 512M, but the same error remained, so I think this landuse_timeseries file might should be modified to single point before submitted to case run.

 

[Fuhow]

Hi, @slevis, another update, I found that it is possible to make regional land data using mksurfdata_map to create surfdata and history landuse data (surface_map and landuse_timeseries datasets required for the stand-alone CLM5 run), or using site_and_regional in ctsm tools (Error during regional simulation startup).

Now, I am testing creating regional (single point) landuse time_series file using command like this:
./mksurfdata.pl -r usrspec -usr_gname $GRIDNAME -usr_gdate $CDATE -l $CSMDATA -no_surfdata -y 2003-2006 -ssp_rcp "SSP2-4.5".

However, it looks like only the full time line can be used as parameter, for example: 1850 to 2000, or 2000 to 2100, the command should be:
./mksurfdata.pl -r usrspec -usr_gname $GRIDNAME -usr_gdate $CDATE -l $CSMDATA -no_surfdata -y 2000-2100 -ssp_rcp "SSP2-4.5".

So, by the way, my question is, can't this parameter choice be the target period? I think it might be a lot of time and memory for download this full datasets.