The Control of Plant and Soil Hydraulics on the Interannual Variability of Plant Carbon Uptake Over the Central US

The interannual variability (IAV) of gross primary productivity (GPP) reflects the sensitivity of GPP to climate variations and contributes substantially to the variations and long‐term trend of the atmospheric CO2 growth rate. Analyses of three observation‐based GPP products indicate that their IAVs are consistently correlated to terrestrial water storage anomaly over the central US, where episodic droughts occur. A land surface model explicitly representing plant hydraulics and groundwater capillary rise with an adequate soil hydraulics well captures the observed GPP IAV. Our sensitivity experiments indicate that, without representations of plant hydraulics and groundwater capillary rise or using an alternative soil hydraulics, the land model substantially overestimates the GPP IAV and the GPP sensitivity to water in the central US. This study strongly suggests the use of the van Genuchten water retention model to replace the most commonly used Brooks–Corey model, which generally produces too strong matric suction of soil water especially in dry conditions, in land surface modeling. This study highlights the importance of plant and soil hydraulics and surface–groundwater interactions to Earth system modeling for projections of future climates that may experience more intense and frequent droughts. Plain Language Summary The interannual variability (IAV) of land carbon uptake contributes substantially to the fluctuations of atmospheric CO2 growth rate. Consistent with previous studies, our data analyses of various observation‐based gross primary productivity (GPP) products and land water storage change data reveal a positive GPP–water relationship. This relationship also has been used to evaluate and constrain climate model projections. Our model sensitivity experiments suggest that current land surface models may overestimate the GPP–water sensitivity and potentially degrade the credibility of future climate projections, due to a lack of appropriate plant and soil hydraulics and surface–groundwater interactions. Our results highlight the importance of key ecohydrological processes on IAV of GPP as well as CO2 projections. Key Points Observation‐based estimates of annual gross primary productivity (GPP) show a strong dependence on water availability over the central US A land surface model with adequate representations of plant and soil hydraulics can capture the observed interannual variability of GPP A model with a low plant drought resilience substanti