Weak Carbon and Water Flux Coupling of Amazonian Rainforests

Tropical carbon and water fluxes are often assumed to be tightly coupled as plants exchange water for carbon via stomata. However, long-term, observation-based evapotranspiration (ET; representing water loss) and gross primary productivity (GPP; representing carbon uptake) estimates show extensive weak or anti-correlations in the central Amazon Basin, implying carbon-water flux decoupling. We investigate this phenomenon across climate and land use gradients in the Amazon Basin spatially and temporally.  We map wet versus dry (using a 2000 mm/year precipitation threshold) and water deficit gradients to examine water availability conditions. Additionally, we map intact versus fragmented forest to examine the anthropogenic influences on carbon-water coupling. Finally, we map these correlations during the dry and wet seasons. We find: (1) Intact and wet forests display weak correlations mostly during the dry season, while non-intact and dry forests remain strongly coupled throughout the year. (2) Intact and wet forests that do not experience water deficits exhibit the weakest carbon-water relationship, especially during the dry season - in such areas, GPP begins increasing one month prior to the dry season, while ET begins increasing two months later. We hypothesize that on an ecosystem level, forested areas with limited human disturbance and sufficient water availability, i.e., undergoing limited stress, reduce carbon and water coupling via the stomata such that GPP and ET instead respond to different conditions such as seasonal changes in radiation and atmospheric water demand, at different times. Conversely, areas that undergo stress from vegetation structure or water availability changes strongly couple GPP and ET. This could be due to: (1) pioneer or secondary vegetation not being structured to decouple their carbon and water fluxes; and/or (2) water stress-induced stomatal responses. This work highlights the spatial and temporal variability in, and large-scale impact of vegetation changes on tropical rainforest carbon and water cycling .