Help getting SP LAI to match LAI from another simulation - issue with crop weighting?

[czarakas@uw_edu]

I'm having a hard time getting SP mode to run such that the LAI climatology exactly matches the LAI of another run for all PFTs. Ultimately I’m trying to run these different model configurations:

• BGC_standard is the typical BGC model set up (compset 1850_DATM%GSWP3v1_CLM50%BGC_SICE_SOCN_MOSART_CISM2%NOEVOLVE_SWAV_SIAC_SESP)
• SP_standard is the standard SP model set up with the default surface dataset (compset 1850_DATM%GSWP3v1_CLM50%SP_SICE_SOCN_MOSART_CISM2%NOEVOLVE_SWAV_SIAC_SESP)
• SP_newlai is the same as SP_standard but I change the surface dataset so that the MONTHLY_LAI climatology is the same as BGC_standard

The default surface dataset I’m using is /glade/p/cesmdata/cseg/inputdata/lnd/clm2/surfdata_map/release-clm5.0.18/surfdata_1.9x2.5_hist_16pfts_Irrig_CMIP6_simyr2000_c190304.nc

I wrote a script to calculate MONTHLY_LAI (dimensions: 12 months, 17 PFTs, lat, lon) from PFT-level TLAI output. I thought this worked correctly because when I tested the script on PFT-level output from SP_standard, I generate MONTHLY_LAI which matches the default surface dataset pretty closely. However, when I tried running the SP_newlai case with a new surface dataset based on BGC_standard, TLAI in SP_newlai doesn’t match the original TLAI from BGC_standard (which I had used to generate the new surface dataset). Do I also need to change the area weightings (e.g. PCT_CROP, PCT_CFT) in the surface dataset to get this to work? When comparing the weights for PFT_level output (pfts1d_wtgcell) in the BGC and SP runs, I noticed that crops seem to occupy more gridcell area in SP than in BGC, but I don’t understand why this would be the case.

Thanks in advance for your help!

 

[oleson]

From czarakas@uw_edu for the benefit of other users:

I wanted to follow up on this since I think I resolved this issue.

The inconsistency was that my BGC case was getting the land cover fractions from the land use time series, while the SP case was getting the land cover fractions from the surface dataset. I think this difference exists in the default BGC and SP model configurations, but it's certainly possible I made some error in setting up these cases. Either way, there is a pretty big difference in the 1850 percent crop between the land use time series (e.g. landuse.timeseries_1.9x2.5_hist_16pfts_Irrig_CMIP6_simyr1850-2015_c170824.nc) and default surface dataset (surfdata_1.9x2.5_hist_16pfts_Irrig_CMIP6_simyr2000_c190304.nc).

I changed PCT_NATVEG, PCT_CROP, PCT_NAT_PFT, and PCT_CFT in the surface dataset to match the percentages in the land use time series, and that change made the TLAI in SP_newlai match the TLAI from the default BGC case.