Confusion about LANDFRAC_PFT in the surface data generated by mksurfdata_map

[Song]

Dear all,

I have some difficulty in understanding the variable LANDFRAC_PFT in the surface data generated by mksurfdata_map (surfdata_0.9x1.25_hist_16pfts_Irrig_CMIP6_simyr1850_c24****.nc).

I originally thought the variable PCT_NATVEG means the percentage of grid cell, so the area of a PFT should be = AREA* PCT_NATVEG / 100 * PCT_NAT_PFT / 100.

However, from the previous discussion #5 in Calculating change in area of PFTs in CLM5 , it is said that the area of a PFT should be = LANDFRAC_PFT*AREA* PCT_NATVEG / 100 * PCT_NAT_PFT / 100. Since the LANDFRAC_PFT is multiplied here, does this mean PCT_NATVEG is not the percentage of natural vegetation in grid cell but the percentage of natural vegetation in vegetated land (including natural vegetation and crop land)? Meanwhile, does this also mean the LANDFRAC_PFT is the land fraction of vegetated land (including natural vegetation and crop land) in grid cell?
In this case, should the area of crop land be = LANDFRAC_PFT*AREA* PCT_CROP / 100 ?

In addition, I find the sum of PCT_NATVEG, PCT_CROP, PCT_URBAN, PCT_GLACIER, PCT_LAKE, PCT_WETLAND in a randomly selected grid is actually equal to 100. This seems to prove that PCT_NATVEG should be the percentage of the grid cell.

Any suggestions would be greatly appreciated!

Thanks,
Song

 

[slevis]

Peter Lawrence's response:
With the CLM5 data the correct way to calculate area of a pft or cft was
PFTArea = LANDFRAC_PFT * AREA * PCT_NATVEG / 100.0 * PCT_NAT_PFT / 100.0
CFTArea = LANDFRAC_PFT * AREA * PCT_CROP / 100.0 * PCT_CFT / 100.0
Due to some issues with this method not being consistent with satellite data numbers for land area and land cover types, I don't use this method in my analysis anymore. I use the same formula but replace the LANDFRAC_PFT * AREA with the landmask, landfrac and area fields that come from a CLM5 history file. This did a much better job of calculating global totals that were in agreement with my satellite and other global data.

 

[Song]

Peter Lawrence's response:
With the CLM5 data the correct way to calculate area of a pft or cft was
PFTArea = LANDFRAC_PFT * AREA * PCT_NATVEG / 100.0 * PCT_NAT_PFT / 100.0
CFTArea = LANDFRAC_PFT * AREA * PCT_CROP / 100.0 * PCT_CFT / 100.0
Due to some issues with this method not being consistent with satellite data numbers for land area and land cover types, I don't use this method in my analysis anymore. I use the same formula but replace the LANDFRAC_PFT * AREA with the landmask, landfrac and area fields that come from a CLM5 history file. This did a much better job of calculating global totals that were in agreement with my satellite and other global data.

OK, I see. I will change to use the fields from a history file.
Thank you so much for the explanation!