极坐标投影到规则的纬度/经度网格

Question

This is a follow on from my previous question, in that this is much more specific with examples and data. I have a 304 by 448 array of ice data (in polar stereographic projection), of which the netcdf can be obtained here . The corresponding lat/lon points of each grid box can be found in this file . I would like to transform this data onto a regular lat/lon grid (say 180x360) where each grid box is a degree. Is there any python way of doing this? Etc using Basemap, or alternatively using CDO (climate data operators)?

Many thanks in advance. James

Answer 1

If you first add the grid to the data file with, eg,

ncks -A -v lat,lon NSIDC_LatLon.nc sample.nc

Then you can use NCO's ncremap to regrid that with, eg,

ncremap -i sample.nc -d 1x1.nc -o out.nc

Unfortunately this will require the current snapshot of NCO (rather than a released version) because the input files contained dimensions in an unusual order that I just patched NCO to handle. A sample of the regridded output is here .

At the OP's request I regridded his file then uploaded it ( here ). I didn't have a 1x1 dataset sitting around so I first created a 1x1 grid file as shown in the manual, then regridded with that (with ncremap -g option):

ncremap -i ~/NSIDC.nc -g ${DATA}/grids/180x360_SCRIP.20150901.nc -o ~/NSIDC_1x1.nc

Answer 2

Usually the command to do this in CDO would be

cdo remapcon,r360x180 in.nc out.nc 

instead of con=conservative remapping, you can use bil=bilinear, nn=nearest neighbour, con2=2nd order conservative remapping.

According to this posting, https://code.mpimet.mpg.de/boards/1/topics/8302?r=8326 the latest versions of cdo should be able to handle polar stereographic projections.