Dynamics of transpiration and evaporation following a moisture pulse in semiarid grassland: A chamber-based isotope method for partitioning flux components

We describe a novel method for partitioning evapotranspiration (ET) from isotopic measurements of water vapor within large (4.86 m3) plot-scale gas exchange chambers. Using this approach, the short-term (15-day) dynamics of transpiration (T) and evaporation (E) in experimental replicated stands of the invasive grass Eragrostis lehmanniana and the native Heteropogon contortus were assessed following a 39-mm irrigation event in semiarid grassland in southeastern Arizona, USA. Water vapor samples (20-40 μL each) were collected sequentially during a 6-min transient increase of vapor concentration inside the chambers and used to produce Keeling plots (isotope mixing relationships) for identification of the isotopic composition of ET and partitioning of component fluxes. The method worked well in plots free of grass cover and in the sparsely covered plots of E. lehmanniana. Keeling plot estimates of the isotopic composition of soil evaporation (δE) in bare plots closely matched modeled values, lending strong support for the validity of the chamber approach. T/ET increased in stands of E. lehmanniana from 0.35 ± 0.07 on day 1 to 0.43 ± 0.08 on day 3 after the irrigation pulse, but decreased to 0.22 ± 0.05 by day 7 as the soil surface dried. Estimates of stand transpiration from the Keeling plot chamber method were positively correlated (Pearson's r = 0.76, p = 0.0004, n = 17) with independent estimates based on leaf-to-canopy scaling of stomatal conductance. We were unable to calculate T/ET on days 1 and 3 in plots of H. contortus because Keeling plot intercepts did not fall within the range of soil and canopy end-member isotope values. This likely occurred due to unaccounted effects of a wet litter layer on the estimation of δE. Our approach is useful for partitioning ET over a dynamic wetting event in semi-arid grassland at a scale relevant for experimental ecosystem studies, but requires further validation under a wide range of vegetation structures and environmental conditions.