For now, stemflow lacks an environmental data annotation tutorial/pipeline. Xarray could easily handle environmental data layers with different resolution and fast indexing for annotation. Could also be used for prediction. And it has a good feature of lazy loading.
Considering linking this package with the developing package EnvArray. Using ERA5 as an example.