Converting a NetCDF with multiple variables to GeoTIFF calling GDAL_TRANSLATE function from Python

Question

gdal_translate -a_srs EPSG:4326 NETCDF:"file_name.nc":variable_name output_file_name.tif  -a_ullr -9.21200 53.28058 -8.88045 53.11338

I am working on some NetCDF files that include multiple variables. The goal is to transform them into GeoTIFF files and georeference them given known x,y-coordinates and if possible apply this process for all the variables included in the file at once. The abovementioned command has successfully transformed the NetCDF file variable into a GeoTIFF file, but I want to execute this inside a Python script.

gdalinfo :

The code I used is here:

transform = gdal.Translate(outname,ds_read,format = 'Gtiff', outputSRS = 'EPSG:4326',outputBounds = [-9.21200, 53.28058, -8.88045, 53.11338] )

The problems I am facing are the following:

When running the code, without specifying a band I get the following Traceback :

Warning 1: Recode from UTF-8 to CP_ACP failed with the error: "Invalid argument".
ERROR 1: Band 1 requested, but only bands 1 to 0 available.

I tried to change the code a bit to test if the variables are interpreted as bands and I added the argument : bandList = 1 inside the gdal_translate fuction. After that, I got the same traceback call as before.

When using the gdal_translate function with the arguments mentioned here : https://gdal.org/python/osgeo.gdal-module.html#Translate , I do not know how to specify a variable (or all variables) of the NetCDF file inside the function. I also tried (unsuccessfully) to parse all the arguments using :

input_format = '-if NETCDF' 
output_format = ' -of GTiff'
projection = '-a_srs EPSG:4326 '
translateOptionText  = projection + input_format + output_format + ' -a_ullr' + str(-9.21200) + '' + str(53.28058) + '' + str(-8.88045) + '' + str(53.11338) 
translateOptions = gdal.TranslateOptions(gdal.ParseCommandLine(translateOptionText))
gdal.Translate(outname,new_name, options = translateOptions)

Any advice or idea?

Answer 1

Using vrt is one solution for sure. Apparently it worked like that for the votemper layer only:

#old_name = file name with path etc. , var_name = subdataset prefix name
#the prints are just for quick check

read_nc = gdal.Open(old_name,gdal.GA_ReadOnly)
sd = read_nc.GetMetadata('SUBDATASETS')
print(sd)
var_name = 'votemper'
sublayername =  gdal.Open('NETCDF:"{0}":{1}'.format(old_name, var_name ),gdal.GA_ReadOnly)
print('\n The sublayer name is: ',sublayername)
outname = old_name.replace('.nc','') + '_' + var_name + '.tif'
print('\n The output name is: ',outname)
projection = '-a_srs EPSG:4326 '
output_format = ' -of GTiff'
translateOptionText = projection + output_format + ' -a_ullr '  + str(28.60704) + ' ' + str(44.395184) + ' ' + str(28.87371) + ' ' + str(44.00740)
print('\n The command is: ', translateOptionText) 
translateOptions = gdal.TranslateOptions(gdal.ParseCommandLine(translateOptionText))
gdal.Translate(outname,sublayername,options = translateOptions)

Giving the subdataset name as an input in the gdal_translate function, together with the correct format and the options text is producing the correct results.

It was double checked in QGIS and everything was as it should be.