Evapotranspiration partitioning through water stable isotopic measurements in a subtropical coniferous forest

Abstract

AbstractEvapotranspiration (ET) partitioning distinguishes the soil evaporation (E) and plant transpiration (T) components and is crucial for understanding the land‐atmosphere interactions and ecosystem water budget. However, the mechanism and controls of ET partitioning for subtropical forests in heterogeneous environments remain poorly understood. Here, we present δ18O and δ2H of about 1,527 isotope samples including atmospheric water, soil and plant water during different seasons in 2 years of 2020–2021 from a coniferous forest across Southeast China. We used the isotopic mass balance of ecosystem water pools, the Craig‐Gordon model and the Keeling‐Plot method to partition T from ET (T/ET) and quantify the controls on T/ET. Results indicated that the uncertainty in the T/ET was principally from the soil water evaporation (δE) value, about 20–30 cm was found to be a reasonable evaporating front depth for estimating δE in this coniferous forest. T/ET presented a “U” shape diurnal pattern and varied from 66.7% to 89.9%. Isotope‐based T/ET in autumn with high temperatures and little rain was higher than those in the summer and winter seasons. Relative humidity (or vapour pressure deficit) dominated the diurnal T/ET variations (relative contributions of > 40%) in summer and autumn, while air temperature and soil water content were the main controls in winter. Our study also showed that δ18O‐derived T/ET was consistent with that of δ2H, although δ2H was found to be more stable in ET partitioning, the dual stable isotope approach should be employed in future studies for the uncertainties brought by samplings or measurements. The agreement between the isotope‐based T/ET and ET partitioning approach that uses eddy covariance and sap flux data was stronger at midday. These isotope‐inferred ET partitioning can inform land surface models and provide more insights into water management in subtropical forests.