Documenting progress on urban-scale ML downscaling evaluation in CSET

This ticket documents progress on evaluating urban-scale ML downscaling predictions in CSET.

Aim: Develop the capability to conduct an integrated evaluation comprising urban-scale ML predictions, reference / "truth" data (i.e., 300 m model), predictor data for the reference variables (i.e., 2.2 km model), and observational data.