ABSTRACT
Analysis of large-scale watershed processes and the development of efficient and integrated modeling platforms which can be used for this purpose has been an important research focus for hydrologists over the past several decades. In particular, in this research area, our activities have focused on the development of modeling tools that can be used in the simulation of the overall response of a watershed based on a localized or distributed hydrologic event which may be identified as input. As is well known, this objective can be achieved through the use of empirically based lumped-parameter models or physics-based distributed-parameter models. The main difference between these two approaches lies in the physical representation of the system (Gunduz, 2004). In the lumped-parameter representation, a watershed is considered to be a single unit behaving in accordance to a completely empirical or quasi-empirical response function with little or no dependence to the analytic description of physical processes and spatial heterogeneity. On the other hand, the distributedparameter representation is based on the idea of treating the system as a discretized set of small homogeneous units that address the spatial heterogeneity with full reference to the analytic representation of physical processes that act on each unit. These differences in turn reflect not only on the simplicity of the model formulation but also on the model’s implementation. Typically, lumpedparameter models provide a general understanding of the system but do not give a comprehensive coverage of the details, whereas physics-based distributed models supply the much-needed detail pertaining to the watershed system. This is most probably the main reason for the preference of the distributed-parameter models over the lumped-parameter models in the hydrological modeling community. This is despite all limitations associated with data availability and the computationally intensive nature of these platforms.
