Output details
15 - General Engineering
University of Bristol
Runoff Prediction Using an Integrated Hybrid Modelling Scheme
Hydrological systems are complex and difficult to model. Paper describes sophisticated method using combination of wavelet decomposition and neural networks to integrate outputs from different sub-models in a cascading fashion. Scheme is tested in a catchment of Southwest England and demonstrates superior performance to all other models tested. New framework has wide implications in hydrological modelling since its general structure could be applied to various model combinations. Work led to Han's invitations to be Visiting Professor at Key National Water Lab (IWHR) in China and panelist in EarthCube Programme of National Science Foundation in the USA.