简介：Thesurface-relatedmultipleelimination(SRME)methodisbasedonfeedbackformulationandhasbecomeoneofthemostpreferredmultiplesuppressionmethodsused.However,somedifferencesareapparentbetweenthepredictedmultiplesandthoseinthesourceseismicrecords,whichmayresultinconventionaladaptivemultiplesubtractionmethodsbeingbarelyabletoeffectivelysuppressmultiplesinactualproduction.ThispaperintroducesacombinedadaptivemultipleattenuationmethodbasedontheoptimizedeventtracingtechniqueandextendedWienerfiltering.ThemethodfirstlyusesmultiplerecordspredictedbySRMEtogenerateamultiplevelocityspectrum,thenseparatestheoriginalrecordtoanapproximateprimaryrecordandanapproximatemultiplerecordbyapplyingtheoptimizedeventtracingmethodandshort-timewindowFKfilteringmethod.AfterapplyingtheextendedWienerfilteringmethod,residualmultiplesintheapproximateprimaryrecordcanthenbeeliminatedandthedamagedprimarycanberestoredfromtheapproximatemultiplerecord.ThismethodcombinestheadvantagesofmultipleeliminationbasedontheoptimizedeventtracingmethodandtheextendedWienerfilteringtechnique.Itisanidealmethodforsuppressingtypicalhyperbolicandothertypesofmultiples,withtheadvantageofminimizingdamageoftheprimary.Syntheticandfielddatatestsshowthatthismethodproducesbettermultipleeliminationresultsthanthetraditionalmulti-channelWienerfiltermethodandismoresuitableformultipleeliminationincomplicatedgeologicalareas.