transform xy to latlon from geostationary projection

[fedecutraro]

I processed data from GLM sensor with the glmtools library and got the next netCDF file OR_GLM-L2-GLMM1-M3_G16_s20222770457400_e20222770502400_c20223091710250.zip

I tried to convert the x,y coordinates to latitude and longitude with the next code

import xarray as xr
import metpy
ds = xr.open_dataset('OR_GLM-L2-GLMM1-M3_G16_s20222770457400_e20222770502400_c20223091710250.nc')
ds_crs = ds.metpy.assign_crs(ds['goes_imager_projection'].attrs)
ds_latlon = ds_crs.metpy.assign_latitude_longitude()

but the latitudes and longitudes I got are very next to the subsatellite point and not that I would expect.
I'm doing something wrong or maybe are some attributtes related to the projection that are missing in the file?

[dopplershift]

The problem here is that the data are encoded with x/y coordinates in radians, but the underlying projection code from PROJ expects values in meters. This can be fixed by multiplying the x/y coordinates by ds['goes_imager_projection'].attrs['perspective_point_height']. MetPy will also do this for you if you call our parse_cf() method (this is NOT done when manually assigning the crs as you do above). So things should be fixed if you adjust your original code to:

import xarray as xr
import metpy
ds = xr.open_dataset('OR_GLM-L2-GLMM1-M3_G16_s20222770457400_e20222770502400_c20223091710250.nc')
ds_crs = ds.metpy.parse_cf()
ds_latlon = ds_crs.metpy.assign_latitude_longitude()

Using that I can get something somewhat reasonable (not knowing what's in the file exactly):

[fedecutraro]

Thank you Ryan, your image is exactly what I expected to get.