Rooting for water: Bridging Plant Hydraulics and Ecohydrology to Predict Drought Stress

Abstract

Forests regulate climate and sustain biodiversity, but their resilience is increasingly threatened by drought-induced hydraulic failure, when xylem water transport is impaired by embolism. A critical but poorly constrained determinant of this risk largely due to limited observations of deep root water uptake is the amount of root-accessible (sub)soil water storage (SR), which governs land-atmosphere exchanges during prolonged dry periods. While ecohydrological theory has long suggested a balance between soil water storage in the rooting zone, vegetation growth and drought tolerance, this concept has not yet been integrated into predictive models of hydraulic failure. Here we propose a novel, process-based inversion framework that integrates ecohydrological optimality theory with principles of plant hydraulic to infer SR. Using the SurEau plant hydraulic model, we estimate the SR value that balances the costs of soil exploration with the avoidance of drought-induced hydraulic damage. Applied first to a well-instrumented Mediterranean Quercus ilex forest, the inferred SRclosely matched independent neutron probe and eddy covariance measurements and accurately reproduced observed drought responses, including leaf water potential and sap flow dynamics. We then scaled the approach across European forests remotely-sensed data, to assess spatially explicit SR estimates and associated hydraulic failure risk. Across more than 20 species and sites, inferred SR and drought stress metrics aligned with field observations and outperformed estimates based on conventional soil databases or remote-sensing-only approaches. By linking plant physiology and ecohydrological theory within a scalable inversion framework, this approach improves predictions of forest drought risk under climate change. Significance Statement Forests rely on subsurface water to maintain functions during drought, but estimating the amount of soil water available to plants (SR) remains a major challenge. Current field-based methods cannot capture deep rooting water extraction or scale to landscape level, limiting predictions of drought impact. We present a novel, process-based approach that infers SR by combining ecohydrological optimum theory with plant hydraulic principles. We hypothesize that plants optimize rooting to balance exploration costs with the risk of hydraulic failure. Our approach predicts SR at European level, outperforming the predictions made using global soil databases or remotely-sensed estimates. This scalable framework advances understanding of forest water use and vulnerability, with broad implications for ecosystem modeling under climate change.