Characterizing biospheric carbon balance using CO2 observations from the OCO-2 satellite: the impact of improved satellite retrievals

NASA's Orbiting Carbon Observatory 2 (OCO-2) satellite could allow scientists to constrain CO₂ fluxes across regions or continents that were previously difficult to monitor. However, one continuing challenge is the development of a robust satellite retrieval: an estimate atmospheric CO₂ from satellite observations of near infrared radiation. The NASA Atmospheric CO₂ Observations from Space (ACOS) retrievals have undergone multiple improvements and updates since the satellite's launch, and the retrieval algorithm is now on version 9. For example, early work indicated biases between land nadir and ocean glint observations and errors in some terrestrial regions that correlated with topography. Recent improvements to the algorithm in versions 8 and 9 have remedied many of these discrepancies.

In the current study, we evaluate the extent to which current OCO-2 observations can constrain monthly CO₂ sources and sinks from the biosphere, and we particularly focus on how this constraint has evolved with improvements to the OCO-2 satellite retrieval. Our goal is to guide top-down, inverse modeling studies and identify areas for future improvement. Specifically, we develop a series of top-down experiments aimed at identifying the smallest regions for which OCO-2 retrievals can be used to robustly detect variations in CO₂ sources and sinks, and we apply these experiments to several recent retrieval versions. We find that an earlier version of the OCO-2 retrieval (version 7) is best-equipped to constrain the biospheric carbon balance across only continental or hemispheric regions. By contrast, newer versions of the retrieval algorithm yield a far more detailed constraint; we are able to constrain CO₂ budgets for individual global biomes (e.g., temperate forests or tropical grasslands), particularly during the Northern Hemisphere summer when biospheric CO₂ uptake is greatest. These results indicate that improvements in CO₂ retrieval algorithms are having a potentially transformative effect on the ability of satellite-based observations to constrain biospheric carbon balance.