Want to write netCDF to tiff but cannot identify the crs?

[qAp]

Hi
I'm trying to convert this netCDF file:
http://thredds-gfnl.usc.es/thredds/fileServer/GLOBALWTDFTP/annualmeans/NAMERICA_WTD_annualmean.nc
to a tiff file using rioxarray and to eventually upload the tiff onto the Google Earth Engine, which requires the tiff file to have a projection.

I know from Converting NetCDF dataset array to GeoTiff using rasterio Python, that the projection can be set using something like xds.rio.set_crs("epsg:4326")

I have also found, following Search order for the CRS (DataArray and Dataset), that for my dataset ds, ds['spatial_ref'].attrs['GeoTransform'] is:

-180.0000010006781 0.008333300184369964 0.0 83.99968819563394 0.0 -0.008333300447549506

But how do I use this info to set a projection before writing out to tiff?

[snowman2]

I recommend using:

xds.rio.write_crs("epsg:4326", inplace=True)

It stays around better.

[qAp]

Thanks. I was able to set the projection successfully this way. However, when writing it out to a tiff, it seems to hang indefinitely. Was wondering if this is the right way to write out tiffs, from a array?

ds = rioxarray.open_rasterio(f'../data/{url.name}')
ds.rio.write_crs('epsg:4326', inplace=True)
print(ds.rio.crs)
ds = ds.squeeze()
ds.WTD.rio.to_raster(f'../data/NAMERICA_WTD_annualmean.tiff')

[qAp]

@snowman2 Apologies, it did write, just had to wait ~ 25 minutes. If this is unusual, please let me know. Thanks

[snowman2]

@qAp, there are a couple of things to think about:

• Sometimes I have has better luck using open_dataset with netCDF files when reading in the data.
• You are reading it over the network, how long does it take to download?
• Looking into parallel reading/writing with rioxarray 0.2 dask example & read parallel.

[RichardScottOZ]

Yes, also how big it is - if something is gigabytes plus it will take a bit of time.