ABSTRACT
Physically based models for catchment flow and solute transport use process descriptions that are solved in a spatially distributed manner to capture spatial structure in parameters, driving variables and flow processes. The numerical methods used to solve these process descriptions require a spatial discretization that serves the dual purposes of representing spatial structures and determining the mathematical accuracy of the process solutions. In many applications the computational burden associated with the spatial discretization has acted as a constraint to its resolution, leading to a coarse discretization and bringing into question the physical relevance of the process solutions (Beven, 1989).
