The entire-process correction approach for flood forecasting

Real-t ime co rr ection is an important compo nent of floo d fo recast ing , and also a hot and diff icult to pic in hydrolog ical forecasting resea rch. Fo r this purpose, an ent ire2pro cess correctio n approach was pro po sed to improv e the accuracy of r eal2time floo d fo recasting. Firstly, w e to ok a basin w ith plentiful r ain g auges. Based on rainfall data of that basin, the relationship be2 tw een the density o f r ain g aug es and the rat io of flo od for ecast ing erro r was established. T hen, fo r the basin w ith a small number rain gaug es, ratio of the real2t ime floo d fo recast ing err or int o areal rainfall err or and mo del erro r w as used to calculate the err or of the flo od fo recast ing . Finally, based o n system r esponse theo ry, the tw o for ecasting erro rs w ere simultaneo usly corrected as to realize the entir e2process correctio n of model input and model structur e. The application results in thr ee ty pical basins o f H ua ihe River showed that the entire2process co rr ect ion appr oach co uld improv e the effect of erro r correctio n and the accur acy of flo od forecasting compar ed to the curr ent sing le2v ariable dy namic sy stem r esponse cur ve method.

Project Supported:

Clc Number: