HR: 16:15h
AN: H54B-02 INVITED     [Abstracts]
TI: Utility of Artificial Neural Networks for Addressing Different Hydrologic Modeling Problems
AU: * Simonovic, S
EM: simonovic@uwo.ca
AF: University of Western Ontario, Department of Civil and Environmental Engineering Institute for Catastrophic Loss Reduction , London, ON N6A5B9 Canada
AB: This presentation investigates the utility of Artificial Neural Networks (ANNs) in addressing different hydrologic problems. The experiences with (a) short term forecasting of streamflows, (b) spring runoff prediction, and (c) prediction of the peak flow, timing and shape of runoff hydrograph are used to explore and assess the capabilities of ANNs and compare the performance of this tool to conventional hydrologic approaches. Several issues associated with the use of an ANN are examined including the type of input data and the number, and the size of hidden layer(s) to be included in the network. Perceived strengths of ANNs are their capability of representing complex, non-linear relationships as well as being able to model interaction effects. Two case study problems are used to illustrate these capabilities of ANNs. The ANN approach for streamflow forecasting is applied to a specific reach of the Winnipeg River system in Northwest Ontario, Canada. The results from this study were most promising. A very close fit was obtained during the calibration (training) phase and the ANNs developed consistently outperformed a conventional model during the verification (testing) phase. The artificial neural network model was applied for spring runoff prediction in the Red River Valley, southern Manitoba, Canada, and compared to linear and nonlinear regression techniques. In this study, according to the accuracy of results, the ANN models show superiority in most of the cases. Data from the Red River in Manitoba, Canada is used to develop an ANN model for predicting the peak flow, timing and shape of runoff hydrograph, based on causal meteorological parameters. High correlation between observed and simulated values of peak flow and time of peak suggests the potential benefits of using ANN for developing runoff hydrographs. Artificial neural networks can be an efficient way of modeling the runoff process in situations where explicit knowledge of the internal hydrologic processes is not required. The presentation will also discuss the advantages and limitations of ANNs based on results obtained from earlier studies that were related to streamflow forecasting.
DE: 1800 HYDROLOGY
DE: 1821 Floods
DE: 1829 Groundwater hydrology
DE: 1857 Reservoirs (surface)
DE: 1860 Streamflow
SC: Hydrology [H]
MN: Fall Meeting 2005