洪水实时校正是洪水预报的重要组成内容, 也是水文预报研究的热点与难点。为进一步提高洪水预报精度, 提出了一种全过程联合校正的洪水预报修正方法。该方法的主旨是在雨量站数目较多的流域, 构建雨量站网密度 与洪水预报误差分配比例之间的定量关系; 对雨量站数目较少的流域, 借用该比例关系, 将洪水预报总误差按比例 划分为面雨量输入误差和模型误差两部分; 再基于系统响应理论, 对研究流域的这两部分误差进行修正, 从而实现 输入误差与模型误差的全过程联合校正。在淮河三个典型流域的应用结果表明, 相较于目前单变量的系统响应曲 线修正方法, 全过程联合校正方法可进一步提升误差修正效果, 提高洪水预报精度。

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.

国家重点研发计划( 2016YFC0402709) ; 国家自然科学基金( 41730750)