Soil nutrient controls on water use efficiency in temperate forests: a regional and methodological comparison

Abstract:

While rising atmospheric CO₂ has been proposed as the primary driver of increases in forest water use efficiency, recent evidence suggests that acid deposition-driven changes in soil nutrients may be overlooked as a control on forest carbon and water relations. Here, we compare trends in water use efficiency between forests in the mid-Atlantic and northeastern United States, two regions exhibiting divergent trends in soil recovery as acid deposition has declined in recent decades. Water use efficiency was estimated using two different methods at the Fernow Experimental Forest in West Virginia and Bear Brook Watersheds in Maine. Each site consists of one watershed that has been experimentally acidified via aerial applications of ammonium sulfate since 1989, and an adjacent control watershed. Intrinsic water use efficiency (WUE_i) was derived from tree ring stable C isotope ratios (δ¹³C) for four codominant species in each watershed. From 1980-2015, WUE_i of red maple, the only co-occurring species between both regions, exhibited a strong positive trend at Bear Brook (0.88 mmol CO₂ mol^-1 H₂O yr^-1), but decreased at Fernow (-0.48 mmol CO₂ mol^-1 H₂O yr^-1). When mean WUE_i was aggregated for each site, there was a significant positive WUE_i trend at Bear Brook (0.62 mmol CO₂ mol^-1 H₂O yr^-1), but no trend at Fernow. We also present a novel watershed-based method of estimating water use efficiency (WUE_ws), defined as the ratio of aboveground tree biomass increment (derived from repeat forest inventory and tree ring data) to evapotranspiration (ET) at the catchment scale. The WUE_ws metric enables estimation of ecosystem WUE in complex terrain, where eddy covariance cannot accurately capture CO₂ and water fluxes, and can also be used to constrain the interpretation of WUE_i results. Between 1990 and 2015, WUE_ws declined at Fernow, driven by both a decline in growth increments and a modest increase in ET, but increased at Bear Brook, driven primarily by increases in biomass increment. Together, the results do not support a ubiquitous CO₂-driven increase in WUE_i in temperate forests, but instead point towards soil nutrient and stand-specific (i.e. species composition, age) controls on forest water use efficiency in the eastern US.