[A-CG35] グローバル炭素循環の観測と解析

[ACG35-P02] Regional anthropogenic CO₂ emission estimates from a global inverse model

*Lorna Raja Nayagam¹、Shamil Maksyutov¹、Tomohiro Oda^2,3,4、Rajesh Janardanan¹、Pamela Trisolino⁵、Jiye Zeng¹、Tsuneo Matsunaga¹ (1.National Institute for Environmental Studies, Tsukuba, Japan、2.Earth from Space Institute, Universities Space Research Association (USRA), Washington, D.C., USA、3.Department of Atmospheric and Oceanic Science, University of Maryland, College Park, MD, USA、4.Graduate School of Engineering, Osaka University, Osaka, Japan、5.National Research Council of Italy-Institute of Atmospheric Science and Climate (CNR-ISAC), Bologna, Italy)

キーワード：Anthropogenic emissions, Greenhouse gas, Carbon dioxide, Inverse modeling

CO₂ emissions from fossil fuel combustion are the main source of anthropogenic greenhouse gases (GHG) emissions. A large fraction (~70%) of the emissions attributable to urban areas suggests the need for enhanced emission estimation/monitoring skills at smaller subnational scales beyond typical large aggregated regional/country scales for conventional atmospheric inversions. With GHG data obtained from the spatially dense ground-based networks and space borne measurements, such a requirement can be materialized in high-resolution inverse models. While the use of regional and local models has been more common for addressing anthropogenic emissions at small scales, here we use a global inverse model to comprehend the CO₂ emissions estimates at subnational scales. Our global Eulerian–Lagrangian coupled inverse model (NIES-TM-FLEXPART-variational) separately optimizes the terrestrial biosphere exchanges, ocean and fossil fuel fluxes for a period of 2015-2019. The inverse model uses the ground-based CO₂ observational data from the urban areas in the US and Europe as well as data from background sites, along with high-resolution (0.025°×0.025°) prior flux datasets. The result represents the large-scale land-ocean fluxes well and the flux estimates are comparable to NOAA’s CarbonTracker posterior fluxes. The RMSE and bias of posterior concentrations at individual sites are reduced compared to prior concentrations. The model’s capability in optimizing the fluxes and to evaluate emissions at subnational scales shows that the estimates are found to have reasonable agreement with self-reported inventories. Compared to data sparse regions, improved results are obtained in regions with denser observation sites. The fossil emissions estimate over US states of Indiana, Massachusetts, Connecticut and New York are comparable to the estimates from the U.S. Environment Protection Agency (EPA) inventory and the estimates over the countries of European Union and UK are compared to the TNO-CAMS inventory.