TheMM5anditsfourdimensionalvariationaldataassimilation(4D-Var)systemareusedinthispaper.BasedontheNationalCentersforEnvironmentalPrediction/NationalCenterforAtmosphericResearch(NCEP/NCAR)reanalysisdata,theauthorsgenerateanoptimalinitialconditionforatyphoonbyusingthebogusdataassimilation(BDA)scheme.BDAisabletorecovermanyofthestructuralfeaturesoftyphoonsincludingawarm-corevertex,thecorrectcenterposition,andthestrongcirculation.AsaresultofBDAusingabogussurfacelow,dramaticimprovementisachievedinthe72hpredictionoftyphoonHerb.Throughseveralcases,theinitializationbyBDAeffectivelygeneratestheharmoniousinnerstructureofthetyphoon,butwhichislackingintheoriginalanalysisfield.Thereforetheintensityforecastisimprovedgreatly.Someimprovementsaremadeinthetrackforecast,butmoreworkstillneedstobedone.