Prediction of extreme precipitation using a neural network: application to summer flood occurence in Moravia

摘要: Abstract Dramatic floods occurred in Central Europe summer 1997, and Czech Republic has been seriously affected its eastern part—Moravia. A predictive approach based on modelling flood recurrence may be helpful management. Summer are typically characterized by saturated catchment due to long-lasting heavy precipitation followed a sudden extreme rainfall. In present work an artificial neural network (ANN) model was evaluated for forecasting. Back propagation networks were trained with actual monthly data from two Moravian meteorological stations time period of 38 years. Predicted amounts next-month-precipitation the next year. The ANN models provided good fit data, have shown high feasibility prediction precipitation.