[M-IS01] Environmental, Socio-Economic and Climatic Changes in Northern Eurasia

[MIS01-P04] Inverse modeling of the carbon sink in the northern boreal regions and comparison with bottom-up data.

*Shamil Maksyutov¹, Tomohiro Oda², Jiye Zeng¹, Motoki Sasakawa¹, Toshinobu Machida¹, Mikhail Arshinov³, Tsuneo Matsunaga¹ (1.National Institute for Environmental Studies, 2.Universities Space Research Association/University of Maryland, 3.V.V. Zuev Institute of Atmospheric Optics)

Keywords:carbon sink, inverse modeling, Siberia

Recent estimates of the carbon sink in the northern boreal regions based on atmospheric CO₂ observations and modelling persistently show large sink that is larger than those reported by forest inventory-based estimates for Russia and Canada. We attempted to revisit the problem using our recently updated inverse modelling system and a set of different observations collected from both ground-based sites and a satellite. We employed a global inverse model NTFVAR (NIES-TM-FLEXPART-variational) to estimate the regional carbon dioxide (CO₂) fluxes using surface/ground-based observation data included in the Obspack dataset, the Japan-Russia Siberian tall tower inland observation network (JR-STATION) data, and v.02.95 XCO2 retrievals from GOSAT satellite. Fossil emissions were taken from the ODIAC inventory, fire emissions from GFAS, and biospheric and oceanic fluxes from the upscaling products based on surface observations. All the fluxes over land were provided at a 0.1-degree resolution. The prior respiration and oceanic fluxes were scaled to match the global CO₂ growth rate. The corrections to the prior fluxes by the inverse model were estimated on a bi-weekly time step separately for land biosphere and ocean regions. With the set of prior fluxes, the forward model simulation replicated the observed seasonal cycle at most global monitoring sites reasonably well. The estimated fluxes for the period of 2009-2019 were summed over the 22 TransCom regions as well as large countries including Russia, Canada, and USA. We found that the flux estimates for Siberia and the Boreal North America are robust and insensitive to the changes in the observing system configurations for different datasets: all the surface observation sites, the background-only (oceanic) sites, all the surface sites plus GOSAT data, the background plus the Siberian data. In all four configurations, the carbon sink in Siberia was stronger than in the Boreal North America, in agreement with forest inventory-based estimates. The magnitude of the sink estimated for Russia appeared to be larger than the national inventory estimates, but close to the estimates based on the remote sensing of the forest cover.