Current implementation of hydrograph assumes pandas objects, but with 2.0 we have (silently?) moved to having xarray as the default data object. Suggested implementation: check object type and convert internally to pandas is the easiest. Rewrite entirely to xarray might be better.