How limited belowground water storage can shield forests from drought

Abstract:

Understanding how annual swings in precipitation impact belowground water storage is essential for predicting ecosystem response to change. In seasonally dry climates in particular, where energy and water availability peak at different times of the year, dry season plant water use is supplied almost exclusively from wet season storage. Here, we hypothesize that dry season water availability is determined by the ratio winter rainfall to belowground water storage capacity. High storage capacity relative to winter rainfall results in year-to-year differences in storage that reflect variations in rainfall totals and consequently, lead to different year-to-year evapotranspiration. In contrast, where storage capacity is low relative to winter rainfall, storage fills in both wet and dry years, and dry season water use does not vary with winter rainfall totals. We test this hypothesis with a probabilistic hydrologic model, detailed field studies at three northern California hillslopes, and water balances conducted over 26 catchments across California. At our three field sites, hillslope storage dynamics are directly quantified via neutron probe monitoring in deep wells, and at the catchment scale, we estimate storage via water balance using publicly available hydrologic data. Our modeling and datasets lead to three key findings. First, water held in weathered bedrock beneath soils (“rock moisture”) is a significant, and sometimes dominant, fraction of storage and is therefore a key determinant of an ecosystem’s sensitivity to wet season rainfall delivery. Second, for sites with low storage capacity relative to winter rainfall, differences in canopy cover are explained by differences in storage capacity. Finally, sensitivity is not determined by the magnitude of storage alone, but rather by the non-dimensional ratio of storage capacity to mean winter rainfall. This contradicts the notion that large storage volumes lead to forest resilience. Contrary to predictions based primarily on tree density and rainfall deficits, sites across a range of storage capacities that exhibited low storage capacity relative to winter rainfall did not experience widespread mortality in the 2011‐2016 extreme drought. Likewise, some sites with high storage capacity relative to winter rainfall did experience mortality.