Typecasting catchments: Universal donors and acceptors

Abstract:

Classifying catchments represents a significant challenge to hydrology and hydrologic modeling, restricting widespread transfer of knowledge from well-studied sites. Traditionally, catchment classification has focused on identifying important physical, climatological, or hydrologic attributes (to varying degrees depending on application/data availability). Such classification approaches are regularly assessed with regard to their ability to provide suitable hydrologic predictions – commonly via the transfer of hydrologic model parameters from the catchment at which they were fitted to a catchment the classification deemed similar. While intuitive, such an approach makes the implicit assumption that physical similarity begets functional (model parameterization) similarity and ignores the most uncertain aspect of the process – the model itself. Here, we explore catchment classification and parameter transferability of the internally verifiable Catchment Connectivity Model across seven catchments located within the Tenderfoot Creek Experimental Forest. Our analysis investigates the roles of model consistency, parameter sensitivity, and catchment structure with regard to catchment classification and introduces the concept of universal donor and universal acceptor catchments.