Emergent constraints on tropical atmospheric aridity - Carbon feedbacks and the future of carbon sequestration

Carbon-climate feedbacks, which amplifies or attenuates atmospheric CO2 from fossil fuel emissions, are one of the largest sources of uncertainty in climate projections. However, these feedbacks depend both on temperature and its coupling to water and energy cycles, especially in the tropics. We show that atmospheric aridity - quantified as vapor pressure deficit (VPD) - is a good proxy for this coupling. Tropical VPD is strongly correlated to the global CO2 growth rate (CGR) with observed present-day sensitivities of -2.5 ± 0.4 GtC mb-1 yr-1. The sensitivity of CGR to tropical VPD interannual variability has increased by a factor of 1.7 ± 0.3 in the 21st century. A combination of causality and statistical analysis point to mechanistic moisture drivers of the VPD-CGR sensitivities, independent of temperature. Observational records provide evidence that tropical atmospheric aridity is linked to both water deficit and spatially correlated with evaporative fraction suggesting that CGR variability is indirectly driven by land - atmosphere coupling (compound soil and atmospheric drought). This coupling is manifest as a kind of carbon-climate feedback in CMIP6 Earth System Models where long-term increases in tropical VPD reduce tropical carbon storage but with a substantial inter-model range [-1.4 to -59.4 GtC mb-1]. However, by employing a hierarchical emergent constraint, the best estimate of atmospheric aridity - carbon cycle feedback (φ TL ) is -19 ± 10 GtC mb-1, which is 28% lower than model estimates with an uncertainty reduction of 50%. Our results bridge the role of moisture and land-atmosphere coupling on net carbon variability to the vulnerability of carbon storage in a changing climate