分布式水文模型与自回归误差校正相结合的低枯流量预报研究  

Research on low flow forecast based on a distributed hydrological model and autoregressive error correction

作　　者：何玉芬 杨汉波[1,2] 董宁澎 李昶明 HE Yufen;YANG Hanbo;DONG Ningpeng;LI Changming(Department of Hydraulic Engineering,Tsinghua University,Beijing 100084,China;State Key Laboratory of Hydroscience and Engineering,Tsinghua University,Beijing 100084,China;State Key Laboratory of Simulation and Regulation of Water Cycle in River Basin,China Institute of Water Resources and Hydropower Research,Beijing 100038,China)

机构地区：[1]清华大学水利水电工程系,北京100084 [2]清华大学水沙科学与水利水电工程国家重点实验室,北京100084 [3]中国水利水电科学研究院流域水循环模拟与调控国家重点实验室,北京100038

出　　处：《水利学报》2024年第12期1539-1547,共9页Journal of Hydraulic Engineering

基　　金：国家重点研发计划项目(2021YFC3000202)。

摘　　要：随着全球气候变化和人类活动影响加剧,干旱事件频发,枯水期的水资源供需矛盾日益突出,准确预报低枯流量愈受重视。本文采用分布式水文模型GBEHM,结合自回归(AR)误差校正方法修正径流模拟结果,进而结合气象预报降水信息,建立了低枯流量预报方法,并将其应用于长江石鼓水文站以上流域,开展了2000—2012年的径流模拟和候、旬、月尺度的预报研究。模拟结果表明:GBEHM模型能较好地重现逐日径流过程,率定期和验证期的纳什效率系数分别为0.94和0.91,相对水量平衡误差分别为1.0%和3.9%;枯水期的模拟径流较观测值偏低,经AR误差校正后率定期和验证期合格率提升至81%~96%。分析预报结果表明,枯水期和严重干旱期的径流预报合格率分别接近80%和85%,经AR误差校正后,预报合格率最大可提升至91%和97%。本研究结合分布式水文模型和误差实时校正技术,实现了候、旬、月尺度的高精度低枯流量预报,提高了枯水期和严重干旱期的径流预报精度,具有工程应用前景。With the increasing impact of climate change and human activities,drought events occur frequently,and water supply-demand conflicts during dry seasons become more prominent.Therefore,accurate low flow forecasting becomes increasingly important.In this paper,the distributed hydrological model(GBEHM)and autoregressive(AR)error correction method were used to correct the simulated runoff,and then,combined with predicted precipitation,a low flow forecast method was established and applied to the watershed above the Shigu hydrological station of the Yangtze River,and the runoff simulation and prediction research were carried out at five-day,ten-day,and monthly scales from 2000 to 2012.The results show that the GBEHM model has good simulation performance on daily runoff with Nash-Sutcliffe efficiency coefficient(NSE)of 0.94 and 0.91,and the relative water balance error(WBE)of 0.98%and 3.9%in the calibration and validation periods,respectively.However,the simulated runoff during dry seasons is lower than the observed.After the AR error correction,the simulation pass rate has increased to 81%to 96%in the calibration and validation periods,respectively.The forecasting pass rate during dry seasons and severe droughts are less than 80%and 85%,respectively.After AR error correction,the forecasting pass rates have been improved into 91%and 97%,respectively.This study has achieved high precision forecast of low flow at five-day,ten-day and monthly scales,significantly improving the forecasting accuracy during droughts and dry seasons.These results have promising applications in engineering.

关 键 词：径流预报 低枯流量 分布式水文模型 实时校正 长江上游 

分 类 号：P338[天文地球—水文科学]