Towards a more accurate land model in a changing climate: Unwinding below-ground controls on vegetation carbon uptake in water-limited environments

With the projected increase in atmospheric CO2 concentration, making science-informed policies with minimized socioeconomic impacts requires today, more than ever, reliable Earth System Models (ESMs). This atmospheric carbon increase is characterized by large inter-annual variability (IAV) shown to be strongly associated with vegetation carbon uptake in semi-arid environments. ESMs used for projection of the future climate and biogeochemical cycle; however, suffer from difficulties in representing this biospheric-atmospheric carbon feedback mainly due to under-representation of below ground vegetation dynamics. The carbon and water uptake of plants are intrinsically linked through their leaves. The carbon sink is therefore controlled by precipitation (and air aridity) in water-limited environments, but we are witnessing important changes in precipitation with reduced frequency and higher intensity. This will have major implications for plant water stress on short to longer time scales. Plant mechanisms to withstand water deficit include stomatal and hydraulics regulation, compensated water uptake, biomass allocation - above and below ground, which in turn regulate the water and carbon cycles. How this is going to affect future carbon cycle and its IAV in particular is unclear. Therefore, we create a mixture of proven and new techniques to infer model parameters from satellite data together with computational methods to form a full quantitative understanding of vegetation water and carbon uptake as well as vegetation above and, in particular, below ground growth mechanisms in semi-arid environments. We propose to improve the representation of plants' role using the novel Climate Modelling Alliance (CliMa) model as a template, and inch towards enabling the inference of below ground processes from above ground measurements by combining computational tools with satellite data. We use Vegetation Optical Depth (VOD) from diurnal to seasonal time scales to constrain physiology (stomatal and hydraulics regulation) and above ground biomass and use optimality principle with CliMa model to infer unknown below ground processes. We validate our findings with measurements in FLUXNET sites. This project will be a milestone with respect to (i) increasing our basic understanding of vegetation carbon uptake mechanisms in semi-arid regions (ii) better representation of below ground vegetation characteristics in climate models and predictions. This new quantitative understanding will serve as a basis to assess the impact of future climate trajectories on the global carbon cycle and its IAV.