`Hazard.from_xarray_raster` should read CRS information, if available

[peanutfun]

I tested the Hazard.from_xarray_raster function using daily humidity data (hurs) from DWD EPISODES, which uses the Lambert Conformal Conic projection (EPSG:3034). The data was successfully loaded as a Hazard object with no issues related to the CRS format.

However, by default, the coordinate reference system (CRS) recorded in the object was EPSG:4326, as this is coded in the function’s properties.

Like :
haz.centroids.crs
<Geographic 2D CRS: EPSG:4326>
Name: WGS 84
Axis Info [ellipsoidal]:

• Lat[north]: Geodetic latitude (degree)
• Lon[east]: Geodetic longitude (degree)
  Area of Use:
• name: World.
• bounds: (-180.0, -90.0, 180.0, 90.0)
  Datum: World Geodetic System 1984 ensemble
• Ellipsoid: WGS 84
• Prime Meridian: Greenwich_

However, my data properties are: ETRS_1989_LCC, Datum: D_ETRS_1989, Spheroid: GRS_1980, Authority: EPSG:3034, Prime Meridian: Greenwich (0.0), Projection: Lambert Conformal Conic.

Perhaps the function could trigger a warning indicating that the CRS should be manually defined, or (more ambitiously ) try to extract it automatically from the data. However, in my case, this is not straightforward.

For now, I did:

ds = xr.open_dataset("/Users/daraya/Downloads/hurs_day_GCFS21--DWD-EPISODES2022--DE-5_sfc20240601_r1i1p1_20240601-20241130.nc")
intensity_variable = "hurs"
hazard_type = "HU"
intensity_unit = "%"

Create hazard from xarray

haz = Hazard.from_xarray_raster(
data=ds,
hazard_type=hazard_type,
intensity_unit=intensity_unit,
intensity=intensity_variable,
crs="EPSG:3034",
coordinate_vars={
"event": "time",
"longitude": "lon",
"latitude": "lat",

}

) Obviously, with my data I can't use functions like dist_to_coast, which are based on EPSG:4326. I think it's a great idea to have the flexibility to incorporate data with different coordinate systems.

Originally posted by @DahyannAraya in #1020 (comment)

[peanutfun]

@DahyannAraya Thanks for reporting this.

The problem is that xarray itself does not read CRS information. The crs parameter in from_xarray_raster must be filled by the user to convey the correct CRS information of the data. The coordinates are directly read from file and no transformation is applied. If you do not supply the crs parameter, the method assumes the coordinates are in WGS 84. Note that this does not fall into the two categories of functions targeted by #1020 :

• [...] functions that must be in epsg:4326 but do take only value as input to catch when input coordinates are not geographic (e.g. in meters)
• [...] functions that assume epsg:4326 without reason to be able to use other crs

I just noticed that we can actually use rioxarray to load CRS data from NetCDF files if they follow the CF conventions: https://corteva.github.io/rioxarray/stable/getting_started/crs_management.html So far, we used rioxarray only to load GeoTIFF files. We might change that in the future 🤔