Bayesian network model for flood forecasting based on atmospheric ensemble forecasts

摘要

The purpose of this study is to propose the Bayesiannetwork (BN) model to estimate flood peaks from atmospheric ensembleforecasts (AEFs). The Weather Research and Forecasting (WRF) model was used tosimulate historic storms using five cumulus parameterization schemes. The BNmodel was trained to compute flood peak forecasts from AEFs and hydrologicalpre-conditions. The mean absolute relative error was calculated as 0.076 forvalidation data. An artificial neural network(ANN) was applied for the same problem but showed inferior performance with a mean absolute relative error of 0.39. It seems that BN is less sensitive to smalldata sets, thus it is more suited for flood peak forecasting than ANN.