CLM initial field and version issues

[Joanna Liu]

Hi,
I've recently been exploring land-atmosphere interactions and would like to run the 2019 simulation with a more accurate land surface initial field. In the input data, the land surface initial field for the most recent year is clmi.BHIST.2010-01-01.0.9x1.25_gx1v7_simyr2010_c181015.nc. I'd like to confirm if I can use this initial field in combination with the IHistClm50Bgc component to simulate the land surface from 2010-2019 and then use the January 2019 clm restart file as my land surface initial field for the 2019 test. Or do I need a longer spin-up?

Also, I have two questions about land initial fields and CLM versions
1）clmi.BHIST.2010-01-01.0.9x1.25_gx1v7_simyr2010_c181015.nc This initial field should be generated from BHIST(HIST_CAM60_CLM50%BGC-CROP_CICE_POP2%ECO_MOSART_CISM2% NOEVOLVE_WW3_BGC%BDRD) compset, is that available for other compsets to use as an initial field? (e.g., different CLM versions, _CLM60%BGC)
2）What are the main changes in the parameters of CLM60 compared to CLM50 (I haven't found any references yet)?

 

[slevis]

You can use initial files from I-cases to start B or F-cases and vice versa in any combination or grid that you wish. However, keep in mind that the model will interpolate the initial data when the grids or sub-grids (i.e. patches, columns, landunits) do not line up. In that case you will get an error suggesting that you set use_init_interp = .true. and then the simulation will most likely work.

Clm6 is in development, so you will not find references for a little while still. Eventually clm6 will be released and we hope to provide some initial files for that version. Meanwhile it's worth trying what I suggested above to see if your clm5 initial file will start a clm6 case. If not, you could perform your own spin-up.

 

[Joanna Liu]

You can use initial files from I-cases to start B or F-cases and vice versa in any combination or grid that you wish. However, keep in mind that the model will interpolate the initial data when the grids or sub-grids (i.e. patches, columns, landunits) do not line up. In that case you will get an error suggesting that you set use_init_interp = .true. and then the simulation will most likely work.

Clm6 is in development, so you will not find references for a little while still. Eventually clm6 will be released and we hope to provide some initial files for that version. Meanwhile it's worth trying what I suggested above to see if your clm5 initial file will start a clm6 case. If not, you could perform your own spin-up.

I have tried that, and it runs successfully. However, the simulated overestimation of LAI for the CLM60 using this initial field is more pronounced than for the CLM50. So I don't know if it's only related to the initial field or the CLM version parameter.

 

[slevis]

The answer may depend on the model version that you're using. There's a long discussion regarding the model's simulated LAI here and the current version overestimates the LAI:

Dead Arctic Vegetation · NCAR LMWG_dev · Discussion #3

Summary: We have persistent dead Arctic vegetation in CLM. This was an issue in CESM2 because with dead vegetation soil C pools are also zero, leading to lower than expected soil C stocks in permaf...

github.com

 

[slevis]

...so with clm6 I think you may get overestimated LAI regardless of the initial conditions.

 

[Joanna Liu]

The answer may depend on the model version that you're using. There's a long discussion regarding the model's simulated LAI here and the current version overestimates the LAI:

Dead Arctic Vegetation · NCAR LMWG_dev · Discussion #3

Summary: We have persistent dead Arctic vegetation in CLM. This was an issue in CESM2 because with dead vegetation soil C pools are also zero, leading to lower than expected soil C stocks in permaf...

github.com

I'll look into it. Many thanks!