Evapotranspiration partitioning of a winter wheat and summer maize double-cropping system using isotopic labeling

The oxygen and hydrogen isotope compositions of ecosystem water pools and fluxes represent important tracers in the water cycle. Combining with eddy covariance technique, it is possible to partition evapotranspiration (ET) into evaporation (E) and transpiration (T) components based on δ18O and δD measurements in the liquid and gas phases. The key challenge is to precisely determine the δ18O and δD of ET (δET), E (δE) and T (δT). In this study, we present high frequency 18O and D measurements of water pools (water vapor, precipitation, dew, groundwater, soil water, xylem water and leaf water) and water flux (evapotranspiration). We characterize, in conjunction with intensive field campaigns, the temporal dynamics of the δ18O and δD signals during the growing season of a in a winter wheat and summer maize double-cropping system in North China in 2008. The three flux end member isotopic signals were determined as follows: continuous δET measurement was made with the flux-gradient method using a tunable diode laser analyzer, δE was estimated with the Craig-Gordon model with a moisture-dependent soil resistance, and δT was approximated by the δ18O and δD of water in xylem (δx) in midday hours assuming δT being equal to δx under isotopic steady state, and by a non-steady state Craig-Gordon model for other times. We found that transpiration comprised ~90% of ET during peak wheat and corn growth periods. In a one-month period from wheat harvest to the establishment of a full corn canopy, transpiration contributed ~70% of ET. These results were consistent with soil lysimeter and eddy covariance measurements made during the same time period. Isotopic partitioning at sub-day time scales was however very scattered, due to large measurement noises in δET, uncertainties in the Craig-Gordon model predictions and spatial variability in δx.