A question about running CESM 2.1.3

[USER_M]

I am a user of CESM2.1.3, I plan to study the land NPP from 1979 to 2015. But I'm confused about the spin up process of CESM2.1.3. The Fig.4(attach) in the article named The Community Earth System Model Version 2 (CESM2) is not easy to understand. If I want to run CESM from 1979 to 2015 at B1850, are the following steps correct?
1.spin up the model based on 0134-01-01 data run from 0001-01-01 to 0200-01-01.
2.directly run the model from 1979 to 2015 based on the r file of 0200-01-01.

 

[mlevy]

I don't think you want to spin up the model yourself, it's a long and expensive process. You can use the BHIST compset to get transient forcings for the historical period (B1850 repeats the year-1850 forcings, which is not what you want if you are looking for modern-day comparisons). I've asked @hannay for more guidance on starting from a specific year -- you will need to download restart files, and I'm unsure of the exact process. There is a little more information in https://bb.cgd.ucar.edu/cesm/threads/suggestions-for-modifying-the-start-date-cesm2.4921/#post-35058 and the thread that comment was posted in.

 

[hannay]

For your case, I would use the restart coming from one of the historical runs. I don't have restart from year 1979 but I have restart from 1979.
I am putting them on the ftp server.
ftp://ftp.cgd.ucar.edu/pub/hannay/
the file is: b.e21.BHIST.f09_g17.CMIP6-historical.011.rest.1976-01-01-00000.tar

 

[ucas_qs]

For your case, I would use the restart coming from one of the historical runs. I don't have restart from year 1979 but I have restart from 1979.
I am putting them on the ftp server.
ftp://ftp.cgd.ucar.edu/pub/hannay/
the file is: b.e21.BHIST.f09_g17.CMIP6-historical.011.rest.1976-01-01-00000.tar

Hello, Hannay. I'm also confused about the spin up. I'm running the case BHISTcmip6 with a resolution of f19_ g17, but I found it seems that this is running from the initial state. This is "RUN_TYPE" is "startup" and "RUN_REFDATE" is "0001-01-01". As far as I know, the case BHISTcmip6 with a resolution of f09_ g17 has an initial field with an integral of 500 years, so there's no spin up. But my simulation doesn't need such a high resolution. I need to change the physical process and run it for more than 100 years, so high resolution will cost more.
So I might need to run hundreds of model years to spin up for the case BHISTcmip6 with f19_g17. Do you have the spin up files that I can use as initial data to start running from 1850 without needing spin up? Thank you very much.