Title: Do plants respond to variations in atmospheric humidity?
Abstract: A growing literature argues that ecosystem-scale evapotranspiration is more sensitive to drying of the atmosphere because of stomatal regulation by plants than to reductions in surface soil moisture. Past studies analysed observations, for which it is difficult to conclusively control for potential relations among plant physiology, measurable state variables such as vapor pressure deficit (VPD) or soil moisture, and ecosystem-scale water flux. Here we analyse natural mechanism-denial experiments at non-vegetated but hydrologically active salt flats. At these sites, any apparent sensitivity of the ecosystem-scale surface conductance (gs, a bulk measure of how the land surface influences evapotranspiration) to VPD cannot be due to stomatal closure. Over the salt flats we find a VPD–gs relation similar to that commonly attributed to stomatal closure, and reproduce similar relations using a parsimonious boundary layer model that excludes plants. We conclude that observational studies probably overstate the sensitivity of ecosystem-scale surface conductance to atmospheric drying and understate the importance of variations in surface soil moisture. This finding has broad implications for future ecosystems, because anthropogenic trends in soil moisture are uncertain and spatially heterogeneous whereas ubiquitous atmospheric drying is expected due to global warming.