No Data Inbounds while clipping timeseries enabled netcdf

[SarvaPulla]

Hello,

I am trying to clip GRACE netcdf files using a shapefile. The following workflow works for certain regions/shapefiles: https://gis.stackexchange.com/questions/354782/how-to-mask-netcdf-time-series-data-from-a-shapefile-in-python/354798#354798. But it's intermittent. For example, it didn't work with a missouri shapefile. It keeps throwing a no data found inbounds error for variable 'lwe_thickness'. The data could be NaNs but there definitely is data all over the globe. Here is a sample dataset that I was trying against: https://tethys-staging.byu.edu/thredds/catalog/data/grace/grace/catalog.html?dataset=testDatasetScan/grace/grace/GRC_jpl_gw.nc. There are other datasets here as well: https://tethys-staging.byu.edu/thredds/catalog/data/grace/grace/catalog.html

[snowman2]

Are you passing in the CRS of your geometry when clipping?

Here is an example: https://corteva.github.io/rioxarray/stable/examples/clip_geom.html#Clip-using-a-GeoDataFrame

[SarvaPulla]

The slight deviation is that I use xarray.open_dataset to read the file instead of xarray.open_dataarray. Was having some other issues when I used open_dataarray. It still fails with open_dataset or open_dataarray.

[snowman2]

I would recommend using rio.clip_box and see what data is there based on the bounding box for the area of interest.

[SarvaPulla]

Clip Box worked on all the shapes that weren't working before. Noticed that the longitude values go from 0 to 360 instead of -180 to 180. Curios about why clip_box worked though.

[snowman2]

See #58

[SarvaPulla]

Thanks. That helped. Adding the write_crs syntax for reference. Was having issues setting the crs from the example in the docs. The following fixed it.

    ds = ds.assign({"lon": (((ds.lon + 180) % 360) - 180)}).sortby('lon')
    ds['lwe_thickness'] = ds['lwe_thickness'].rio.write_crs("epsg:4326")