I use gdalwarp on Windows (GDAL 3.0.2, released 2019/10/28) to perform this projection change.
gdalwarp -t_srs EPSG:4326 -te -180 -90 180 90 -of NETCDF in.nc out.nc
The result file out.nc seems to have the correct EPSG:4326 projection but there are no data inside -32767 (no data value from the in file)
I tried to : - Extract the chla subdataset specifying the input file as NETCDF:"chla":in.nc but it does not change anything - Change the -te values : i always have the issue
Specify directly the source EPSG with -s_srs "EPSG:3411" and with -s_srs "+proj=stere +lat_0=90 +lat_ts=70 +lon_0=-45 +k=1 +x_0=0 +y_0=0 +a=6378273 +b=6356889.449 +units=m +no_defs" : i always have the issue
Specify the interpolation method, bilinear for instance : i always have the issue
What is strange is that in the in.nc when i check in Panoply the No Datavalue are Nan (and the internal code is -32767 and the scale factor and offset are specified in the metadata) whereas in the out.nc the grid displays Nan values directly as -32767 and the scale factor and addoffset are not specified.
I also tried the advises specified here : Re-projecting from Polar projection to EPSG:4326?
Below find the log of the projection :
After additional investigations it seems that the NETCDF original file may contain bad information for the the extent compared to unit in m defined in the crs. Indeed it seems that extent is defined in 100 of km.
Moreover when i compare the PROJCRS between the failing netcdf file and a grib file with EPSG:3411 that i manage to convert into EPSG:4326, i notice that the EPSG Reference is 4326 for the failing one whereas it is different for the good grib file even if the netecdf file specifies the Polar Stereographic for the Conversion. Any ideas ?