Aregionalsurfacecarbondioxide(CO_2)fluxinversionsystem,theTan-Tracker-Region,wasdevelopedbyincorporatinganassimilationschemeintotheCommunityMultiscaleAirQuality(CMAQ)regionalchemicaltransportmodeltoresolvefine-scaleCO_2variabilityoverEastAsia.Theproperorthogonaldecomposition-basedensemblefour-dimensionalvariationaldataassimilationapproach(POD-4DVar)isthecorealgorithmforthejointassimilationframework,andsimultaneousassimilationsofCO_2concentrationsandsurfaceCO_2fluxesareappliedtohelpreducetheuncertaintyininitialCO_2concentrations.ApersistencedynamicalmodelwasdevelopedtodescribetheevolutionofthesurfaceCO_2fluxesandhelpavoidthe'signal-to-noise'problem;thus,CO_2fluxescouldbeestimatedasawholeatthemodelgridscale,withbetteruseofobservationinformation.Theperformanceoftheregionalinversionsystemwasevaluatedthroughagroupofsingle-observation-basedobservingsystemsimulationexperiments(OSSEs).TheresultsoftheexperimentssuggestthatareliableperformanceofTan-Tracker-Regionisdependentoncertainassimilationparameterchoices,forexample,anoptimizedwindowlengthofapproximately3h,anensemblesizeofapproximately100,andacovariancelocalizationradiusofapproximately320km.Thisisprobablyduetothestrongdiurnalvariationandspatialheterogeneityinthefine-scaleCMAQsimulation,whichcouldaffecttheperformanceoftheregionalinversionsystem.Inaddition,becauseallobservationscanbeartificiallyobtainedinOSSEs,theperformanceofTan-Tracker-RegionwasfurtherevaluatedthroughdifferentdensitiesoftheartificialobservationnetworkindifferentCO_2fluxsituations.TheresultsindicatethatmoreobservationsiteswouldbeusefultosystematicallyimprovetheestimationofCO_2concentrationandfluxinlargeareasoverthemodeldomain.TheworkpresentedhereformsafoundationforfutureresearchinwhichathoroughestimationofCO_2fluxvariability
