A dynamic, non-steady state approach for paritioning of soil evaporation and plant water use at landscape scales

Abstract:

Seperate characterization of plant water use and soil evaporation are critical to understanding ecohydrological dynamics of dryland ecosystems and for efficiently managing water in dryland agriculture. The application of stable isotopes as a tracer of these individual fluxes has been constrained by obtaining robust measurements of the isotopic composition of plant water use (δT) that may be scaled up to the ecosystem level. Of particular concern is the fact that the isotopic composition of plant transpiration is usually assumed to be equal to the isotopic composition of xylem water; the so-called steady-state assumption. However, our results and the findings of other published studies strongly suggest that steady state conditions are unrealistic for vegetation in dynamic natural environments. This talk focuses on the development of a simple framework for using relationships between plant transpiration and δT to partition ET at the landscape level. Our method uses a newly-derived empirical relationship between leaf conductance and isotopic fractionation during transpiration to solve a system of equations that can provide solutions to the fraction of total ET composed of bare soil evaporation and transpiration. We apply our method to a time series of evapotranspiration fluxes and near-surface water vapor isotopic composition at a field station in central Kenya and compare the results with partitioning obtained from both steady-state approaches and non-isotopic approaches for partitioning.