Computational & Technology Resources
an online resource for computational,
engineering & technology publications
DEVELOPMENTS IN NEURAL NETWORKS AND EVOLUTIONARY COMPUTING FOR CIVIL AND STRUCTURAL ENGINEERING
Edited by: B.H.V. Topping
Artificial Neural Networks in Hydro-Meteorological Modelling
A.W. Jayawardena and D.A.K. Fernando
Department of Civil and Structural Engineering, University of Hong Kong, Hong Kong
A.W. Jayawardena, D.A.K. Fernando, "Artificial Neural Networks in Hydro-Meteorological Modelling", in B.H.V. Topping, (Editor), "Developments in Neural Networks and Evolutionary Computing for Civil and Structural Engineering", Civil-Comp Press, Edinburgh, UK, pp 115-120, 1995. doi:10.4203/ccp.34.6.1
Artificial Neural Networks which mimic the brain and the neuron system are useful computational methods that have attracted the attention of researchers from many fields in the recent past. They are particularly helpful in situations where the input-output relationship of the system under study is not explicitly known. In this paper, an application of the approach to predict streamflows in two small catchments in Hong Kong using corresponding rainfall data as input parameters is described.
Daily discharge predictions are compared with actual observations and with those from a traditional mathematical model. The ANN approach was found to give better agreements. For short term event simulations, the ANN approach is equally applicable, but the accuracy of prediction seems to depend heavily on the use of the past values of streamflow as input parameters. For the catchment considered, it was found that past streamflow values at least 45 minutes prior to the present time level are necessary for accurate predictions.
purchase the full-text of this paper (price £20)