Page 70 - 《水资源与水工程学报》2023年第4期
P. 70
!34 " ! 4 # & ' ( ) & * + , - Vol.34No.4
2023 $ 8% JournalofWaterResources&WaterEngineering Aug.,2023
DOI:10.11705/j.issn.1672-643X.2023.04.08
'a EnKS s SWAT CZdxþ/0h/z{2
1,2 1,2 3 1,2
i j , <kl , mno , pq/
(1. JÑI& ùa¢!;<P( JL), JL JÑ 350108;2. JÑI& ãù×íîWF»Ü6°Í&ѾÛå,
JL JÑ 350108;3. JL@pFò¾Û, JL JÑ 350001)
H I: Áï@AîO¢UZoSBUZ^b¬âPôl, ?uXdù×>V^ÁZ, ÎÝ·de?
;<à, 'x,°àÁ4I(EnKS) W SWAT qr, ¶Ldù×>V=ª, ÄK,°àÁÞ(EnKF) =p
}¶oº, #ê>>Vqr^ÁZ, ·1ù×>Voê>d·Z^ab。¯ï:EnKS ±qâßÈ8Z
ê>@A\N、 dekj; ü)@Aqr^âPôlµÎÿ.uXqr^>VÁZ, oº EnKF =p,En
KS =p^Sà. Wù (NSE)áãâ、 M、 ãä 3ÑL·kuào 0.03、0.12、0.03, 5=5¢k
(RMSE)·kO(o 7.43%、26.81%、4.25%; ù×>V=poê>d·Z^a}OZãj¡lWâ
j¡l, Xx{ p|Wo@àe EnKS =pQc|¢d9¥k_`^a}, @N EnKS =poÁïd^a
}û&@Nk_。
JKL: d; ù×>V;EnKS p;SWAT qr; Pôl; Ý·de
+
MNOPQ:P338;TV121 .4 7RSTU:A 7VWQ:1672643X(2023)04006610
StremflowdataassimilationoftheMinjiangRiverBasinbased
onEnKSapproachandSWATmodel
1,2
3
1,2
XIANGYong ,CHENYunzhi ,TANGLifang,WANGXiaoqin 1,2
(1.AcademyofDigitalChina(Fujian),FuzhouUniversity,Fuzhou350108,China;2.KeyLaboratoryofSpatialData
MiningandInformationSharing,MinistryofEducation,FuzhouUniversity,Fuzhou350108,China;
3.FujianSoilandWaterConservationExperimentalStation,Fuzhou350001,China)
Abstract:Atimelagexistsintheresponseofobservedvariablestostatevariablesinsurfacehydrological
process.Toimprovetheaccuracyofstreamflowdataassimilation ,astreamflowdataassimilationscheme
wasconstructedbasedontheensembleKalmansmoother (EnKS)andSWATmodelfortheMinjiangRiv
erBasin ,whichwasthencomparedwiththeensembleKalmanfilter(EnKF)method,soastoevaluate
theaccuracyofdifferentassimilationmethodsandtoanalyzetheeffectofdataassimilationondifferent
streamflowcomponents.TheresultsindicatethattheoptimaltimewindowlengthofEnKSdiffersineach
hydrologicalperiodandbasin ,andtheconsiderationofthetimelagofthehydrologicalmodelcaneffec
tivelyimprovetheaccuracyofmodelassimilation.ComparedwiththeEnKFmethod,theNSEoftheEn
KSmethodincreasedby0.03 ,0.12,0.03attheQilijiestation,ShaxianstationandZhuqistation,re
spectively;whereastheRMSEofthethreestationsdecreasedby7.43%,26.81% and4.25%,respec
tively.Thereisspatialandtemporalheterogeneityintheimprovementeffectofdataassimilationmethod
ondifferentstreamflowcomponents.TheEnKSmethodcansignificantlyimprovetheaccuracyofthelater
alflowinregionswithhighpermeabilitysoilandsteepslopes ,andithasabetterperformanceinwetsea
sonthandryseasonfortheimprovementofsurfaceflowprediction.
Keywords:streamflow;dataassimilation;ensembleKalmansmoother(EnKS)method;SWATmodel;
timelag;theMinjiangRiverBasin
!"#$:20221010; %$:20230223
'()*: ¢!%&Pÿ!l9%H)Û(XDA23100503); JL@J%H)*(MSK202210、MSK202214)
+,-.: ) ¾(1998—), ., Û·Ñ3, ß5;<8, ;<=>?@Aù×>V。
¥¦+,: ®"¹(1982—), ã, JLJÑ3, 45, ª;<ó, ;<=>?CDK8B2。