Estimating Regional Greenhouse Gas Sources and Sinks through Ensemble-Based Simultaneous State and Parameter Estimation

The increased data density of greenhouse gas (GHG) measurements opens up doors for estimating GHG sources and sinks at finer spatio-temporal scales through top-down approaches, but simultaneously places higher demands on efficient data-model integration and accurate uncertainty quantification. Building on advanced data assimilation techniques, we present a generalized data assimilation framework on Ensemble-based Simultaneous State and Parameter Estimation (ESSPE) for GHG source and sink estimation. In this talk we focus on the carbon cycle and demonstrate a data assimilation system for estimating CO₂ sources and sinks at regional scales over North America. By coupling the carbon and atmospheric states, we show that the ESSPE system can optimize CO₂ flux parameters while explicitly accounting for atmospheric transport uncertainties and other error sources. The data assimilation system is based on the ensemble Kalman Filter and is able to both assimilate high-resolution observations and optimize flux parameters at the native model resolution. Beyond source and sink estimating and error quantification, we show that the ESSPE framework is useful for designing better observation strategies, exploring the effects of heterogeneity and equifinality, and investigating the impacts of different error sources such as transport uncertainties.