Background/Question/Methods Across manipulative field studies elevated CO2 (ECO2) reduces stomatal conductance, and can reduce canopy transpiration if canopy surface area remains unchanged. Reductions in water use by trees can increase seasonal water availability, which could prolong carbon and nutrient uptake during drying conditions. In field studies, plant response to ECO2 can be confounded with response to prevailing climatic conditions. Thus, long-term ECO2 field studies that encompass inter-annual climatic variations can provide invaluable insight into stand-level impacts of changes in site water availability attributable to ECO2. Inter-annual patterns and implications of ECO2-induced changes in site water use were assessed in a long-term free-air carbon enrichment (FACE) study in a Liquidambar styraciflua (sweetgum) plantation, with ambient (ACO2) and ECO2 treatments now in their 12th year. Stem sap flux density was measured at various spatial (radial, circumferential) and temporal (diurnal, seasonal, inter-annual) resolutions and scaled to canopy transpiration. Data collected in 1999, 2004, 2007 and 2008 were correlated to inter-annual patterns of canopy development, precipitation inputs, N availability and biomass allocation patterns. Soil water content was monitored in upper soil and at multiple depths to 1 m (2008) and related to canopy transpiration. Results/Conclusions Canopy transpiration was consistently 11-16% lower in plots exposed to ECO2 during non-drought years. In a drought year (2007), annual transpiration was 25% lower in ECO2 plots than in ACO2 plots. In 2007, water use (and thus carbon uptake) was greatly reduced by drought resulting in slow canopy development during 2008, when maximum leaf area index (LAI) was just 75% (ACO2) or 66% (ECO2) that of 2006. Despite reduced LAI, site water use in 2008 was similar to other years, suggesting seasonal patterns of stomatal conductance were higher than in previous years. Upper soil water content was not correlated to reduced water use in ECO2 treatments (0-20 cm); however, there was enhanced soil water extraction in ECO2 treatments deeper in the profile, which is consistent with increased root production to at least to 60 cm in ECO2 treatments. This pattern is not inconsistent with reduced canopy transpiration for ECO2 trees, as a site water budget suggests heavier dependence on soil water deeper than 1 m for ACO2 trees. We conclude that ECO2 increases the sensitivity of stomatal response to changing edaphic and climatic conditions, a response that may be influenced by root-derived signaling as root distribution and water extraction patterns shift within the vertical soil profile.