Groundwater access modulates plant water stress and evapotranspiration in water-limited ecosystems

Accurate representation of evapotranspiration (ET) and plant water stress is essential for understanding ecosystem resilience to hydroclimatic variability. However, most land surface and Earth System Models infer vegetation stress primarily from shallow soil moisture, implicitly assuming that declining surface water availability directly translates into physiological limitation. This assumption fails in ecosystems where vegetation can access deeper stored water, such as groundwater, leading to systematic mischaracterization of water stress and ecosystem functioning during dry periods.

Here we present a physics-based ecohydrology model that represents root water uptake as an emergent process governed by soil–plant–atmosphere water potential gradients, allowing plants to dynamically shift water sources between shallow soil and deeper reservoirs. This framework captures how access to groundwater modifies plant hydraulic status and regulates water stress across seasonal, interannual, and long-term drying. Applied to oak savanna ecosystems, the model reveals distinct uptake regimes in which groundwater increasingly contributes to transpiration as surface soils dry, buffering ET during dry seasons when shallow soil moisture alone would predict strong limitation.

Our results show that groundwater access fundamentally alters ecosystem stress trajectories, delaying the onset of hydraulic limitation and reducing ET sensitivity to surface drying. Water table depth emerges as a key control on the degree of buffering, highlighting feedbacks between rooting strategies, subsurface water availability, and ecosystem resilience. We further demonstrate that stress metrics based solely on shallow soil moisture substantially overestimate drought impacts in systems sustained by deeper water sources.

By providing a reduced-order representation of ET that explicitly accounts for both soil moisture and groundwater availability, this work offers a pathway to improve model representations of plant water stress and evapotranspiration in ecosystems sustained by deep water storage under a changing climate.