Obtaining column-level QIRRIG variables for irrigated CFTs

[amans]

I saved pft-level outputs for QIRRIG_FROM_SURFACE, QIRRIG_FROM_GW_UNCONFINED, and QIRRIG_FROM_GW_CONFINED for BgcCrop compset. I am using alpha-ctsm5.2.mksrf.23_ctsm5.1.dev171 version of CTSM. I am looking to obtain irrigation withdrawal values for individual irrigated crops. However, as per this thread and the open github issue, pft-level irrigation withdrawals are not usable currently.
Is it possible to obtain CFT-level irrigation outputs then? The CLM5 description paper mentions, "On the crop land unit, each CFT (irrigated and unirrigated) resides on its own soil column...".
My pft-level hist file has the following dimensions:

• Dimensions:
  • levgrnd: 25
  • levsoi: 20
  • levlak: 10
  • levdcmp: 25
  • time: 365
  • hist_interval: 2
  • lon: 770
  • lat: 480
  • gridcell: 205895
  • landunit: 563040
  • column: 5812693
  • pft: 8695223

I have been accessing pft-level GRAINC_TO_FOOD by reshaping the pft dimension to (79, lat, lon). For the irrigation variables, would reshaping the column dimension to (64, lat, lon) work similarly, or is there something else I should be aware of?

 

[oleson]

Right, as you noted, these are all column-level quantities, and so requesting them at pft-level can results in unanticipated problems. But yes, since crops are on their own soil column, you can get CFT-level irrigation outputs. I think you should be able to identify crop types in column-level output by looking at cols1d_itype_col. Output from the global attributes in a history file:

ctype_crop_noncompete = "2*100+m, m=cft_lb,cft_ub"
:cft_c3_crop = 1 ;
:cft_c3_irrigated = 2 ;
:cft_temperate_corn = 3 ;
:cft_irrigated_temperate_corn = 4 ;
:cft_spring_wheat = 5 ;
....