Was There a Tropical Land Carbon Sink During 2015-2023? Results from an Ensemble of Global Inversion Models Constrained by OCO-2 Version 11 and the Global In Situ Network.

Was There a Tropical Land Carbon Sink During 2015-2023? Results from an Ensemble of Global
Inversion Models Constrained by OCO-2 Version 11 and the Global In Situ Network.

The lack of sufficient in situ observations over the Tropics makes the carbon balance and seasonality for
critical terrestrial regions such as the Amazon, the tropical rain forests in central Africa, and Tropical Asia
difficult to constrain with confidence. In some years the carbon budget is dominated by biomass burning,
while in other years, the dominant factors seem to be driven by climate forcing due to the El Niño Southern
Oscillation (ENSO). Previous studies using OCO-2 and the global in situ network (Liu et al, 2017; Palmer et
al, 2017; Crowell et al, 2019; Peiro et al, 2022; Byrne et al, 2024) have shown a more dynamic tropical
carbon cycle than what was understood previously. The duration and quality of the satellite XCO2 record is
now sufficient to begin investigating interannual variability in regional terrestrial carbon fluxes in the Tropics
as well as how the carbon fluxes respond to environmental and climatic drivers like temperature,
precipitation, vapor pressure deficit, and fire.

In this work, we will present preliminary results from the OCO-2 model intercomparison project (OCO-2
MIP) that estimates fluxes from numerous independent models using different data assimilation schemes
using surface and airborne in situ data as well as OCO-2 Version 11 retrievals. As noted in the past, the
strong El Niño of 2015-2016 resulted in strong tropical efflux in all tropical regions. The results also continue
to suggest that in the last decade, the Tropics have been net-zero or small source of carbon rather than being a
carbon sink. In this presentation, we will also discuss the causes of uncertainty associated with these
estimates, including the significant uncertainty due to modeled transport errors.