Wavelet-genetic programming conjunction model for flood forecasting in rivers

摘要: In this study we have developed a conjunction model, WGP, of discrete wavelet transform (DWT) and genetic programming (GP) for forecasting river floods when the only data available is historical daily flows. DWT used denoising smoothening observed flow time series on which GP implemented to get next-day flood. The new model compared with autoregressive (AR) stand-alone models. All models are calibrated tested Kosi River one most devastating rivers world high spiky monsoon flows, modeling poses great challenge. With different inputs, twelve models, four in each class AR devised. best performing WGP WGP4, previous rates as input, forecasts an accuracy 87.9%, root mean square error 123.9 m 3 /s Nash–Sutcliffe coefficient 0.993, performance indices among all extreme also better simulated by than