河流系统实时洪水预报误差多点联合校正方法研究

实时校正是改善洪水实时预报精度的重要手段。河流系统中多个站点之间具有高度的水力联系, 各个站点之 间的误差也具有空间关联性。依据河道水流演进的基本方程和动态自适应的自回归方法, 建立了考虑误差空间演 化的河流系统实时洪水预报误差多点联合校正方法。利用洪峰段洪量误差、洪峰流量误差、纳什效率系数( NSE) 和 峰滞时间等指标分别对多点联合校正和不考虑误差空间关联性的单点校正开展校正效果评估。以淮河王家坝断面 以上为背景开展实证研究, 结果表明: 考虑河流系统误差空间关联性的多点联合校正效果优于单点校正, 洪峰段效 果更为显著, 能够更有效地提高河流系统洪水预报的精度。 Real-time correction is an im po rtant w ay to improv e the accur acy o f r eal2time flood fo recasting. T her e are hig h2level hy dr aulic connect ions among the multiple st ations in a river sy st em, and the err ors betw een the mult iple stat ions ar e spat ially in2 ter connected as w ell. Based on the basic equatio ns of riv er floo d routing and the dynamic adaptive aut oreg ressive method, t his study pro posed a mult i2point jo int co rr ect ion metho d fo r real2time flo od fo recasting erro rs wit h consideratio n to the spat ial evo2 lut ion of erro rs. T he correctio n effect was evaluated in terms o f the flood vo lume erro r during flood peak period, f loo d peak dis2 charg e err or, Nash2Sutclif fe efficiency co efficient, and lag t ime of peak. This paper conducted a case study o n t he reg ion up the Wangjiaba section o f H ua ihe Riv er to demonstr ate the pro po sed met ho do log y . The results indicated that the effect of multi2point joint co rr ection was better than that of single2po int correctio n, and the ef fect w as even mo re prominent fo r the floo d peak peri2 o d. The pr oposed method can effectiv ely impro ve the accuracy o f flo od for ecasting of the river sy st em. 国家重点研发计划( 2017YFC0405606) ; 国家自然科学基金( 51579068) ; 中央高校基本科研业务费专项资金( 2018B10514)