Page 16 - 2023年第54卷第7期
P. 16

Multimodelpredictionanditsensembleevaluationofrunoffand
                                     sedimentchangetrendintheYellowRiverBasin
                                                                                  3
                                     1
                                                       1
                                                                                               1
                                                                   2
                         HUChunhong,ZHANGXiaoming,YUKunxia,XUMengzhen,ZHAOYang
                           (1.ChinaInstituteofWaterResourcesandHydropowerResearch,StateKeyLaboratoryofSimulationand
                                       RegulationofWaterCycleinRiverBasin,Beijing 100038,China;
                  2.StateKeyLaboratoryBaseofEco - HydraulicEngineeringinAridArea,Xi’anUniversityofTechnology,Xi’an 710048,China;
                           3.StateKeyLaboratoryofHydroscienceandEngineering,TsinghuaUniversity,Beijing 100084,China)
                  Abstract:ThevariationofrunoffandsedimentintheYellowRiverprofoundlyaffectsthedecision - makingofmajor
                  issuesintheYellowRiverprotectionanddevelopment ,suchasthelayoutofrunoffandsedimentregulationpro
                  jects ,theallocationofwaterresourcesinsideandoutsidethebasin,andtheconstructionofcross - basinwater
                  transferprojects.However,theattributionanalysisandpredictionresultsofrunoffandsedimentvariationinthe
                  YellowRivervarygreatlyfromresearchtoresearch ,sincetheyareaffectedbymanyfactorssuchasresearchperi
                  od ,datumperiod,andresearchmethods.Afteranalyzingtheresultsofrunoffandsedimentchangeattributionand
                  prediction ,thisstudyconstructsamulti - indexandsystematicmodelaccuracyevaluationsystem andproposesa
                  technicalframeworkofensembleassessmenttopredictrunoffandsedimentchangesintheYellowRiverinthenext
                  50yearsbasedonthe9typesofmodelpredictionresultsset.Theresultsshowthatthelargedifferenceintheattri
                  butionanalysisresultsofrunoffandsedimentchangeinthetypicalYellowRiverbasiniscausedbydifferentdatare
                  quirements ,constituentvariables,andthesimulationaccuracyoftheevaluationmethods.Thus,themulti - index
                  andsystematicmodelaccuracyevaluationsystemincludestheinput - structure - outputassessment.Theapplicability
                  oftheexistingrunoffandsedimentchangepredictionmodelsisevaluatedfrom fiveaspectsincludingdatarequire
                  ments ,physicalbasis,applicationefficiency,outputscale,andpredictionaccuracybasedonthestandardized
                  datainput.The technicalframework ofensemble assessmentisfinally constructed based on multimodel
                  applicabilityevaluation - multiscenarioset - resultsensembleprediction, andtheaveragerunoffandsediment
                  amountatTongguanStationoftheYellowRiverarepredictedtobeabout24billionand300milliontonsperyear
                  respectivelyinthenext50yearsbasedonensembleassessmentof9typesofmodels.
                  Keywords:runoffandsedimentchangeoftheYellowRiver;attributionanalysis;uncertainty;runoffandsedi
                  mentmodel;ensembleevaluation

                                                                                    (责任编辑:李福田)

































                —  7 7  —
                     4
   11   12   13   14   15   16   17   18   19   20   21