Technical note: A high-resolution inverse modelling technique for estimating surface CO<sub>2</sub> fluxes based on the NIES-TM–FLEXPART coupled transport model and its adjoint

Abstract. We developed a high-resolution surface flux inversion system based on the global Eulerian–Lagrangian coupled tracer transport model composed of National Institute for Environmental Studies (NIES) (TM; collectively NIES-TM) and FLEXible PARTicle dispersion (FLEXPART). The is named NTFVAR (NIES-TM–FLEXPART-variational) as it applies variational optimization to estimate fluxes. tested by estimating optimized corrections natural CO2 fluxes achieve best fit atmospheric data collected in situ network necessary step towards capability anthropogenic emissions. employed Lagrangian particle (LPDM) FLEXPART calculate footprints observations at spatial resolution 0.1?×0.1?. LPDM with (NIES-TM). Our technique uses an adjoint iterative procedure. error covariance operator was implemented via implicit diffusion. Biweekly prior fields were estimated years 2010–2012 from included Observation Package (ObsPack) set. High-resolution prepared using Open-Data Inventory Anthropogenic Carbon dioxide (ODIAC) fossil fuel combustion, Global Fire Assimilation System (GFAS) biomass burning, Vegetation Integrative SImulator Trace gases (VISIT) terrestrial biosphere exchange, Ocean Tracer Transport Model (OTTM) oceanic exchange. biospheric field constructed vegetation mosaic map separate simulation daily time VISIT each type. uncertainty scaled proportionally monthly mean gross primary production (GPP) Moderate Resolution Imaging Spectroradiometer (MODIS) MOD17 product. inverse calculates form relatively smooth multiplied patterns uncertainties land ocean, following coastlines fine-scale productivity gradients. resulting estimates improved taken continuous observation sites, reproducing both seasonal short-term concentration variabilities including high events associated use inversions has advantage better resolving transported mixed signals sources densely populated continental regions. Thus, potential separation between ecosystems strong localized sources, such emissions forest fires. Further improvements modelling are needed our posterior than that Oceanic Atmospheric Administration (NOAA)'s CarbonTracker only fraction monitoring i.e. mostly coastal island locations where background local mixed.