摘要

ANN model can not provide a good simulation of big storm events because of the lack of big storm events in its training data. To overcome this limitation, this paper establishes a rainfall-runoff model that is a combination of a conceptual model and an ANN model. In this rainfall-runoff model, big synthetic storm events are generated by using Xinanjiang model, and then these events are used as the training data of the ANN model so that the number of big storm events is increased. This new model gathers the merits of the conceptual model and the ANN model, and a case study shows that this model can produce good simulation for storm events.

  • 出版日期2010

全文