GC23L-1262
Deuterium excess in the water cycle of Cona Lake, central Tibetan Plateau

Tuesday, 15 December 2015
Poster Hall (Moscone South)
Jiangpeng Cui1, Lide Tian1, Trent W Biggs2 and Rong Wen1, (1)ITP Institute of Tibetan Plateau Research, Chinese Academy of Sciences, Beijing, China, (2)San Diego State University, San Diego, CA, United States
Abstract:
Large numbers of lakes on the Tibetan Plateau (TP) play an important role in the regional hydrological cycle, and the river systems they occur in serve as important water resources for more than a billion people, but systematic observations of the lake water balance is scarce on the TP. Stable isotopes of water (δ18O and δD) can be used to understand the lake water cycle and its impact on water resources. Here we present a detailed study on the water cycle of Cona Lake, at the headwater of Nujiang-Salween River, based on three years of observations of δ18O and δD, including samples from precipitation, upper lake water (<1 m), outlet surface water, and atmospheric water vapor. Theδ18O-δD relationship in lake water (δD = 6.67δ18O – 20.37) differed from that of precipitation (δD = 8.29δ18O + 12.50). The deuterium excess (d-excess = δD - 8δ18O) was lower in the Cona Lake water (-7.5‰) than in local precipitation (10.7‰), indicating evaporation of lake water. The ratio of evaporation to inflow (E/I) of the lake water was estimated using the Craig-Gordon model. The E/I ratios of Cona lake ranged from 0.24 to 0.27 in the three years, indicating that 73-76% of lake water flowed downstream. Compared to the Craig-Gordon model, the Rayleigh distillation model had limited success in estimating lake evaporation, especially when lake water experienced intensive evaporation. In-situ observation of atmospheric water vapor isotope improved the estimate of E/I by 0.04 for δ18O and 0.08 for δD on average compared to water vapor isotope derived from the precipitation-vapor equilibrium method. Continuous sampling of lake water was necessary to estimate the E/I ratio accurately for lakes impacted by seasonal precipitation or meltwater, with sampling in the late summer or fall recommended for regional surveys. The results from this work imply the feasibility of d-excess in the study of lake water cycle for other lakes on the TP. The result is also helpful for the catchment-scale hydrological simulation work, as usually the role of lakes in the water cycle might be ignored in hydrological model as in this basin.