ABSTRACT
There are many underlying or implicit assumptions involved in the development of deterministic, physically based hydrologic models and their application to real catchments. These range from the model structure to the constituent algorithms and their parameterization. Physically based models usually consist of linked process equations derived from physics principles (conservation of mass, momentum and energy) at a point, laboratory or plot scale. These models are developed with the belief (or hope) that the model parameters are measurable. In contrast, empirical or functional models can be thought of as simple input-output models whose parameters can rarely be interpreted in terms of physical processes. Another term used with physically based hydrologic models is “distributed-parameter” which refers to the ability to represent the spatial variability of parameters and/or processes within catchments. The application of distributed-parameter, physically based models to real catchments presupposes: (a) that physical processes can be represented in a deterministic way and that the overall catchment response can be represented by the combined action of the constituent process algorithms; (b) that the
spatial variability of a catchment can be represented by distributed values of the model parameters; and (c) that model parameters can be meaningfully measured or estimated at the element scale of the model.
