Methods for variables where paleo reconstructions are not available

[fernandadialzira]

Goal: adapt bash script to python application

For variables where reconstructions are not available, the following have presented a good result applying distance weighted average interpolation, therefore being good candidates for nearest neighbour interpolation in python or creepfill.

interp_vars =[al, cnhk,i stl1, istl2, istl3, istl4, licd, lict, lmld, lmlt, lshf, ltlt, alnid, alnip, ci, src, skt, asn, rsn, slt1, slt2, slt3, slt4, tsn, aluvd, aluvp, lblt] 

variables lsrh and sr have minimum and maximum as 0, therefore no change should applied.

the bash script where cdo commands were implemented to create masks for the interpolation:

make_masks.sh

#!/bin/sh

lp_path=/Documents/phd/data/pliomip3/LP/
input_mask_path=${lp_path}/input_masks

#making sure that missing value in the PRISM4D data is not 0
cdo -setmissval,-99999 ../LP_LSM_v1.0.nc lp.lsm.nc
cdo -setmissval,-99999 ../LP_icemask_v1.0.nc lp.ice.nc


#interpolating PRISM4 lsm to OpenIFS grid
cdo griddes ../ICMGGaackINIT > ../grid_icmgg.txt
cdo -f grb -copy -int -setgrid,../ICMGGaackINIT -remapdis,../grid_icmgg.txt lp.lsm.nc lp.lsm.rmp.grb
cdo -f grb -copy -int -setgrid,../ICMGGaackINIT -remapdis,../grid_icmgg.txt lp.ice.nc lp.ice.rmp.grb

#creating the mask where ocean is 1 and land is nan
cdo -setctomiss,2 -addc,1 lp.lsm.rmp.grb oce1.mask.grb

#creating the mask where land is 1 and ocean is nan
cdo -setctomiss,0 lp.lsm.rmp.grb land1.mask.grb

#saving openIFS land-sea and ice sheet masks
cdo -setmissval,-99999 -selcode,172 ../ICMGGaackINIT oifs.lsm.grb
cdo -setmissval,-99999 -selcode,141 ../ICMGGaackINIT oifs.ice.grb

#creating anomaly of the land sea mask
cdo -setrtoc,0.5,1,1 -setrtoc,0,0.5,0 oifs.lsm.grb oifs.lsm.mask.grb
cdo sub lp.lsm.rmp.grb oifs.lsm.mask.grb lsm.anom.grb

#creating added, drowned, gis and wais masks
cdo -setctomiss,2 -addc,1 -setrtoc,-1,0,0 lsm.anom.grb lsm.added.miss.grb
cdo -setctomiss,2 -addc,1 -setrtoc,-1,0,0 -mulc,-1 lsm.anom.grb lsm.drowned.miss.grb

#creating mask where the icesheet is reduced
cdo sub lp.ice.rmp.grb oifs.ice.grb icesheet.anom.grb
cdo setrtomiss,1.1,3 -addc,1 -lec,-9 -setmisstoc,0 icesheet.anom.grb lsm.reduced.icesheet.grb

and to change each variable according to the method:

change_icmgg_vars.sh

#!/bin/bash


# Extract all unique codes from the ICMGG files
codes=$(cdo showcode ../ICMGGaackINIT)

# Loop through each unique code
for code in $codes; do

    # Saving openIFS land-sea and ice sheet masks
    cdo -setmissval,-99999 -selcode,${code} ../ICMGGaackINIT oifs.var${code}.grb

    # Interpolating drowned points
    cdo -ifthen oce1.mask.grb oifs.var${code}.grb var${code}.oce.grb
    cdo -setmisstodis,50 -mul var${code}.oce.grb lsm.drowned.miss.grb var${code}.oce.drowned.dwa.grb
    cdo -setmisstoc,0 -ifthen oce1.mask.grb var${code}.oce.drowned.dwa.grb var${code}.oce.drowned.dwa.oce.grb

    # Interpolating added and ice sheet points
    cdo -ifthen land1.mask.grb oifs.var${code}.grb var${code}.land.grb
    cdo -setmisstodis,50 -mul var${code}.land.grb lsm.added.miss.grb var${code}.land.added.grb
    cdo -setmisstodis,50 -mul var${code}.land.added.grb lsm.reduced.icesheet.grb var${code}.land.added.is.grb
    cdo -setmisstoc,0 -ifthen land1.mask.grb var${code}.land.added.is.grb var${code}.land.added.is.dwa.land.grb

    # Adding both fields
    cdo add var${code}.oce.drowned.dwa.oce.grb var${code}.land.added.is.dwa.land.grb var${code}.plio.grb
done

cdo merge var*.plio.grb oifs.plio.init.grb
cdo -addc,1 -mulc,0 ../ICMGGaackINIT icmgg_ones
cdo replace ../ICMGGaackINIT oifs.plio.init.grb ICMGGaackINIT_plio_variables
cdo mul ICMGGackINIT_plio_variables icmgg_ones ICMGGaackINIT_plio_variables_corrected

rm *var*.grb

DWA_NN_remap.pdf

[JanStreffing]

Hello @fernandadialzira, would you still have time to update this if necessary, or point me to where I can find up-to-date info?