Identifying signals of tree water stress with data from a new stress monitoring network

Water stress in forest ecosystems limits carbon assimilation and growth, and can increase the risk of exogenous disturbance events like wildfire and insect outbreaks. Climate models predict warmer and drier conditions across much of the Southeastern U.S. in the future, creating a need for decision-support tools that permit land managers to assess stress levels at the spatial and temporal scales relevant to the size and scope of their management capabilities. Drought indices derived from data collected within the current network of climate stations and remotely-sensed platforms tend to have coarse spatial resolutions that are insufficient to capture ecosystem-scale variations in vegetation cover, topography, or edaphic conditions. The Remote Assessment of Forest Ecosystem Stress (RAFES) network is a new stress monitoring data network and decision support tool which provides a framework for assessing stress conditions directly at the ecosystem scale and in real-time. As part of RAFES, a range of meteorological and physiological variables, including sap flux density, air temperature and humidity, soil moisture, temperature, and matric potential, and woody fuel moisture, have been collected in nine at Eastern U.S. forest ecosystems since the summer of 2010. Here, we introduce the network and identify drought stress signals in the raw data records that may be used in the future to develop a generic stress index for RAFES sites. We use a Jarvis-model framework with a particular focus on the relationship between canopy stomatal conductance (determined from sap flux data) and vapor pressure deficit, noting that recent studies in the literature suggest that this relationship is governed by marginal water use efficiency, which in turn should vary in response to water stress.