Deuterium excess reveals diurnal sources of water vapor in forest air

An understanding of atmospheric water vapor content and its isotopic composition is important if we are to be able to model future water vapor dynamics and their potential feedback on future climate change. Here we present diurnal and vertical patterns of water isotope ratios in forest air (delta H-2(v) and delta O-18(v)) not observed previously. Water vapor observed at three heights over 3 consecutive days in a coniferous forest in the Pacific Northwest of the United States, shows a stratified nocturnal structure of delta H-2(v) and delta O-18(v), with the most positive values consistently observed above the canopy (60 m). Differences between 0.5 m and 60 m range between 2-6aEuro degrees for delta O-18 and 20-40aEuro degrees for delta H-2 at night. Using a box model, we simulated H2O isotope fluxes and showed that the low to high delta H-2(v) and delta O-18(v) profiles can be explained by the vapor flux associated with evaporation from the forest floor and canopy transpiration. We used d-excess as a diagnostic tracer to identify processes that contribute to the diurnal variation in atmospheric moisture. Values of d-excess derived from water vapor measurements showed a repeated diel pattern, with the lowest values occurring in the early morning and the highest values occurring at midday. The isotopic composition of rain water, collected during a light rain event in the first morning of our experiment, suggested that considerable below-cloud secondary evaporation occurred during the descent of raindrops. We conclude that atmospheric entrainment appears to drive the isotopic variation of water vapor in the early morning when the convective boundary layer rapidly develops, while evapotranspiration becomes more important in the mid-afternoon as a primary moisture source of water vapor in this forest. Our results demonstrate the interplay between the effects of vegetation and boundary layer mixing under the influence of rain evaporation, which has implications for larger-scale predictions of precipitation across the terrestrial landscape.