Regional Inverse Modeling in North and South America for the NASA Carbon Monitoring System - Follow-On

With geostationary carbon satellites (such as GeoCARB) being launched, satellite footprint generation will become nearly impossible due to the expensive computational cost. A new inverse flux estimation strategy is needed to take advantage of satellite measurements in the coming years. Under previous efforts, we have added footprints (surface influence functions) for NASA remote sensing datasets such as OCO-2, and we have developed strategies to investigate consistency among in situ and remote sensing datasets and for combining in situ and remote sensing data for flux estimation. In this presentation, we will describe our objectives and initial progress developing a prototype regional Eulerian carbon flux estimation system using a hybrid 4DVar framework to leverage constraints from an increasingly large volume of satellite (i.e. OCO-2, OCO-3, GeoCARB) and in situ CO2 data over North America. An ensemble approach is used to quantify the uncertainty introduced by model transport, prior fluxes, and CO2 boundary inflow. The ensemble, calibrated by observations, will provide an objective estimate of uncertainty in these components of the inverse system. The estimated uncertainty components will be implemented in the flux estimation system. We plan to conduct a set of continental-scale Observation System Simulation Experiments in preparation for analysis of data from the newly announced GeoCARB mission. This project makes extensive use of NASA satellite and airborne missions, including OCO-2, ACT-America, and NSF and DOE field campaign measurements as well as NASA model products (e.g., MERRA transport fields and CMS flux products), thus supporting the development of an integrated Carbon Monitoring System. The work will develop strategies for incorporating diverse CO2 observations into CMS flux products and for quantifying fluxes and their uncertainties at policy-relevant scales and for Monitoring, Reporting and Verification (MRV).