Page 135 - 2024年第55卷第12期
P. 135
[30] 王泰华.青藏高原冻土变化的水碳效应研究[D].北京:清华大学,2021.
[31] CUOL,ZHANGYX,BOHNTJ,etal.Frozensoildegradationanditseffectsonsurfacehydrologyinthenorth
ernTibetanPlateau[J].JournalofGeophysicalResearch,2015,120(16):8276 - 8298.
[32] XIONGL,O’CONNORKM.Comparisonoffourupdatingmodelsforreal - timeriverflowforecasting[J].Hydro
logicalSciencesJournal ,2002,47(4):621 - 639.
[33] 熊怡,周建中,孙 娜,等.基 于 自 适 应 变 分 模 态 分 解 和 长 短 期 记 忆 网 络 的 月 径 流 预 报 [J].水 利 学 报,
2023,54(2):172 - 183,198.
[34] 杨姗姗,包为民,杨小强,等.微分响应在降雨误差修正中的应用[J].中 国 农 村 水 利 水 电,2015(1):
75 - 79.
[35] 蒋语繤,瞿思敏,蒋思军,等.系统微分响应误差修正方法在 VIC模型中的应用与研究[J].中国农村水
利水电,2023(10):96 - 103.
[36] 王慧敏,宋松柏,张更喜.基于季 节 长 记 忆 双 自 回 归 模 型 的 日 径 流 模 拟 [J].水 利 学 报,2023,54(6):
686 - 695.
[37] 薛洋,杨光,许雷.基于 Kalman - ARIMA模型的大坝变形预测[J].中国农村水利水电,2016(12):117 -
119 ,123.
Researchonlowflowforecastbasedonadistributedhydrologicalmodel
andautoregressiveerrorcorrection
1,2
1,2
1,2
3
HEYufen ,YANGHanbo ,DONGNingpeng,LIChangming
(1.DepartmentofHydraulicEngineering,TsinghuaUniversity,Beijing 100084,China;
2.StateKeyLaboratoryofHydroscienceandEngineering,TsinghuaUniversity,Beijing 100084,China;
3.StateKeyLaboratoryofSimulationandRegulationofWaterCycleinRiverBasin,
ChinaInstituteofWaterResourcesandHydropowerResearch,Beijing 100038,China)
Abstract:Withtheincreasingimpactofclimatechangeandhumanactivities,droughteventsoccurfrequently,
andwatersupply - demandconflictsduringdryseasonsbecomemoreprominent.Therefore,accuratelowflowfore
castingbecomesincreasinglyimportant.Inthispaper ,thedistributedhydrologicalmodel(GBEHM)andautore
gressive(AR)errorcorrectionmethodwereusedtocorrectthesimulatedrunoff,andthen,combinedwithpredic
tedprecipitation ,alowflowforecastmethodwasestablishedandappliedtothewatershedabovetheShiguhydro
logicalstationoftheYangtzeRiver,andtherunoffsimulationandpredictionresearchwerecarriedoutatfive - day,
ten - day,andmonthlyscalesfrom2000to2012.TheresultsshowthattheGBEHMmodelhasgoodsimulationper
formanceondailyrunoffwithNash - Sutcliffeefficiencycoefficient (NSE)of0.94and0.91,andtherelativewater
balanceerror(WBE)of0.98% and3.9% inthecalibrationandvalidationperiods,respectively.However,the
simulatedrunoffduringdryseasonsislowerthantheobserved.AftertheARerrorcorrection ,thesimulationpass
ratehasincreasedto81% to96% inthecalibrationandvalidationperiods,respectively.Theforecastingpassrate
duringdryseasonsandseveredroughtsarelessthan80% and85%,respectively.AfterARerrorcorrection,the
forecastingpassrateshavebeenimprovedinto91% and97%,respectively.Thisstudyhasachievedhighprecision
forecastoflowflowatfive - day,ten - dayandmonthlyscales,significantlyimprovingtheforecastingaccuracydur
ingdroughtsanddryseasons.Theseresultshavepromisingapplicationsinengineering.
Keywords:runoffforecasting;lowflow;distributedhydrologicalmodel;real - timecorrection;theUpperYangtze
River
(责任编辑:韩 昆)
5
— 1 4 7 —