hydrological modeling across spatial and temporal scales

The traditional challenge charge to watershed hydrologic modelers has been to predict river discharge, given precipitation and other near-surface atmospheric variables that control evaporative demand. With the advent of the digital computer in the 1950s came spatially lumped hydrological models like the Stanford Watershed Model, and various more recent derivatives. These models attempted to represent runoff production using effective or spa-tially lumped parameters. These parameters represented the storage of water in the subsur-face, and the rates of exchange among the storage compartments, including infiltration and surface runoff production, and subsurface drainage or baseflow. Evapotranspiration was typically represented as a fixed (seasonally varying) potential, with actual evapotranspira-tion determined by the potential modulated by resistance controlled by the relative storage in subsurface zones. These models represented a considerable advance over pre-computer methods that typically were event-based, rather than temporally continuous. The move to time-continuous modeling avoided the difficulty of specifying initial conditions (primarily moisture storage) at the onset of a storm, which has always plagued event models in hydrol-ogy