Inverse modeling framework for estimating regional ecosystem carbon fluxes using satellite CO₂ observations

Objective estimates of CO₂ fluxes between the atmosphere and terrestrial ecosystems are needed to ensure reliable climate projections, as well as to address a number of practical problems, in particular, those related to the implementation of natural climate solutions. One of the main approaches that are widely used to develop such estimates is based on inverse modeling, which involves optimizing carbon fluxes by matching observations and simulations of atmospheric CO₂ mixing ratios using probabilistic methods.

Within the framework of this approach, the regional inversion modeling system RIGGO (Regional inversion of greenhouse gases observations) designed to assess ecosystem carbon fluxes in Russian regions is being developed in IAP RAS. The system includes, in particular, the chemistry transport model (CTM) CHIMERE, the meteorological model WRF, and the ecosystem exchange model VPRM. It is based on an original method for inverse modeling of ecosystem carbon fluxes using satellite observations of the CO₂ mixing ratio in the atmosphere [1]. Important features of the method are the estimation of spatially distributed ecosystem carbon fluxes through the Bayesian optimization of the VPRM parameters, as well as a reduction in the sensitivity of the flux estimates to the boundary conditions of the CTM due to a preliminary transformation of the input (measured and simulated) CO₂ data using calculations of the residence time for air masses within the region.

The capabilities of RIGGO were investigated by applying it to estimate the carbon balance in the European region covering the EU countries and the UK, as well as in Central Siberia during the warm season using OCO-2 satellite data. It was found that the assimilation of OCO-2 data by the system allows for a radical (up to two or three times) reduction of the a priori uncertainty of the estimated ecosystem fluxes. At the same time, consideration of several test cases using both synthetic and real CO₂ data demonstrated that the a posteriori estimates of carbon fluxes depend weakly on both the a priori estimates of the VPRM parameters and the boundary conditions of the CTM. It was also found that the inferred estimates of ecosystem fluxes are consistent with independent (and credible) estimates for both regions, including carbon flux estimates provided by the CTE-HR system (for the European region) and derived from the GFED database (for the Siberian region). In addition, optimization of the VPRM parameters improved the agreement of the simulated CO₂ mixing ratios over Western Europe with the data of the TCCON ground-based measurement network, while the calculations of net ecosystem exchange in Siberia were brought closer to the data of the corresponding measurements by the eddy covariance method at the Zotino tall tower observatory (ZOTTO).

1. Konovalov I.B., Golovushkin N.A., Mareev E.A. Using OCO-2 observations to constrain regional CO₂ fluxes estimated with Vegetation, Photosynthesis and Respiration Model // Remote Sens. 2025. V. 17, P. 177.