Structures and Performances of Five Rainfall-Runoff Models for Continuous River-Flow Simulation

Four black-box-type rainfall-runoff models, namely, the Simple Linear Model, the seasonallybased Linear Perturbation Model, the wetness-index-based Linearly Varying Gain Factor Model, and the Artificial Neural Network Model, along with the conceptual Soil Moisture Accounting and Routing Model, were used for this study. The models exhibit a considerable range of variation in degree of structural complexity and associated parameter parsimony, with corresponding degrees of complication in objective function evaluation. Hence they represent a family of models suitable for application in both developed and developing countries. Operating in continuous river-flow simulation mode, these models and techniques were applied to six test catchments representing wide variability in geographic location, climatic condition, areal extent and physiographical characteristics. A number of performance evaluation criteria were used to comparatively assess model efficiency. The ‘Galway Real-Time River Flow Forecasting System’ software package, developed at the Department of Engineering Hydrology, of the National University of Ireland, Galway, was used to produce all the numerical results presented in the study.