Reproject WRF data

[francobarrionuevoenv21]

Hi, guys!

I'm trying to work with some WRF model data from the national weather service. After retrieving the dataset for an specific forecast from AWS, I get the following dataset

As you can see in metadata, it's projected in lambert conformal. The thing is that I need to reproject it into epsg 4326. I've tried using write_crs (from rioxarray) and reproject methods but with no luck. I don't understand if it's something I don't get about CRS in this dataset or just a dataset format issue.

Any help is welcome. Thank you!!

[scottyhq]

Rioxarray (ds.rio.reproject()) is expecting a regular grid, but the WRF lat/lon spacing is irregular. So you'd first need to interpolate onto a regular grid - see relevant discussion corteva/rioxarray#426. The recommended solution in that discussion for WRF specifically was to use the more specialized salem library built on top of Xarray that has routines for working with WRF, and grid interpolation: https://salem.readthedocs.io/en/v0.3.10/wrf.html. Did you eventually find a solution @francobarrionuevoenv21 , if you're able to share the code you ended up using that would be great!