1st International Congress on Environmental Modelling and Software - Lugano, Switzerland - June 2002
Keywords
black-box model, conceptual model, simulation, river flow forecasting system, neural network
Start Date
1-7-2002 12:00 AM
Abstract
Four black-box-type rainfall-runoff models, namely, the Simple Linear Model, the seasonallybasedLinear Perturbation Model, the wetness-index-based Linearly Varying Gain Factor Model, and theArtificial 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 structuralcomplexity and associated parameter parsimony, with corresponding degrees of complication in objectivefunction evaluation. Hence they represent a family of models suitable for application in both developed anddeveloping countries. Operating in continuous river-flow simulation mode, these models and techniqueswere 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 tocomparatively assess model efficiency. The ‘Galway Real-Time River Flow Forecasting System’ softwarepackage, 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.
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 seasonallybasedLinear Perturbation Model, the wetness-index-based Linearly Varying Gain Factor Model, and theArtificial 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 structuralcomplexity and associated parameter parsimony, with corresponding degrees of complication in objectivefunction evaluation. Hence they represent a family of models suitable for application in both developed anddeveloping countries. Operating in continuous river-flow simulation mode, these models and techniqueswere 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 tocomparatively assess model efficiency. The ‘Galway Real-Time River Flow Forecasting System’ softwarepackage, 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.
