Differences in watershed evaporation indicated by hydrogen and oxygen single and dual isotopes: Evidence from controlled simulation tests under different land uses

Abstract

Partitioning watershed evapotranspiration into evaporation and transpiration is essential for studying the water–carbon cycle, and the key step in this partitioning is to calculate watershed evaporation based on isotopic indexes. The δ18O and δ2H single isotopes and d-excess or lc-excess dual isotopes derived from δ18O and δ2H are widely used for this purpose. Although δ18O, δ2H, and d-excess have different characteristics, few studies have assessed how these differences influence the usefulness of these isotopes as indicators of watershed evaporation. Here, we simulated five watersheds with different land uses and degrees of evaporation in a region of Southwest China with a subtropical monsoon climate. We then investigated differences in the ability of the isotopic indexes to indicate watershed evaporation over two hydrological years. We found that d-excess and lc-excess dual isotopes were better indicators of watershed evaporation than δ18O and δ2H single isotopes. This is because the δ18O and δ2H in precipitation are easily influenced by rainout during the rainy season, when groundwater is primarily recharged. They are more variable than d-excess and lc-excess, which leads to δ18O and δ2H in groundwater being more sensitive to changes in the isotopic signal input in precipitation. This probably interferes with, or even masks watershed evaporation signals carried by δ18O and δ2H. The annual mean isotopic compositions were more depleted in groundwaters (δ18O = − 8.7 ‰ to − 8.5 ‰, δ2H = − 61.9 ‰ to − 58.7 ‰) than in precipitation (δ18O = − 8.3 ‰, δ2H = − 54.3 ‰) from November 2015 to October 2016; these compositions fail to support the isotopic enrichment of groundwater compared with that of precipitation. Evaporation rates over two years derived from d-excess dual isotopes agreed well with those from the water balance in bare-rock and bare-soil lands, which had no data from plants available for verification. Compared with hydrogen and oxygen single isotopes, their dual isotopes reflected watershed evaporation more accurately, especially in Asian monsoon regions.