Prediction of extreme precipitation using a neural network: application tosummer flood occurrence in Moravia

Dramatic floods occurred in Central Europe in summer 1997, and Czech Republ ic has been seriously affected in its eastern part-Moravia. A predictive ap proach based on modelling flood recurrence may be helpful in flood manageme nt. Summer floods are typically characterized by saturated catchment due to long-lasting heavy precipitation followed by a sudden extreme rainfall. In present work an artificial neural network (ANN) model was evaluated for pr ecipitation forecasting. Back propagation neural networks were trained with actual monthly precipitation data from two Moravian meteorological station s for a time period of 38 years. Predicted amounts are of next-month-precip itation and summer precipitation in the next year. The ANN models provided a good fit with the actual data, and have shown a high feasibility in predi ction of extreme precipitation. (C) 2000 Elsevier Science Ltd. All rights r eserved.