水稳定同位素示踪的冰川流域水文模拟及不确定性研究

水文模拟不确定性长期以来是制约寒区水文发展的瓶颈问题。水稳定同位素示踪为认识冰川流域径流过程提供了重要“指纹”信息，但仍缺乏有效模型将该信息与冰川水文模型耦合，而且同位素信息对冰川流域水文模型不确定性的约束效果也有待检验。将水稳定同位素信息（δ¹⁸O）与冰川流域水文模型FLEX^G相耦合，实现对冰川流域水稳定同位素和径流过程的耦合建模（FLEX^G-iso），并在乌鲁木齐河源1号冰川流域进行模拟检验和径流分割。结果表明：模型不仅对2013—2016年径流过程有良好的模拟效果，还可以重现水稳定同位素、冰川物质平衡等重要过程。利用水稳定同位素这一辅助数据，提高了模型参数的识别能力，减少了模拟过程中各水源的相互妥协效应和不确定性范围。2013—2016年1号冰川断面径流32%~34%来自融雪，48%~51%来自融冰，0%~7%来自地下水，12%~15%来自降雨径流。水稳定同位素对雪和冰川相关中间过程有明显的约束能力，原有模型对融冰贡献的模拟偏高约7%。FLEX^G-iso模型的建立有助于推动寒区水文学理论和方法的发展，以及寒区水资源、生态环境保护等相关决策制定。

Hydrological simulation traced by water stable isotope and uncertainty analysis in a glacier catchment:taking Glacier No.1 in headwater of Urumqi River as an example

The hydrological model of glacier basin is an important method to study glacier water resources， but its uncertainty affects its wide application and the ability to support decision-making. Water stable isotope tracing provides important “fingerprint” information for understanding the runoff process of glacier basins， but there is still a lack of effective models to couple this information with glacier hydrological models， and the effect of stable isotope information on the uncertainty of hydrological models of glacier basins at different time and space scales also needs to be tested. In this study， the water stable isotope information （δ¹⁸O） is coupled with the glacier-hydrological model FLEX^G to realize the coupled modeling of the water stable isotope and runoff process （FLEX^G-iso） of the glacier basin. We take the catchment of Glacier No. 1 in headwater of Urumqi River as the research area. The data used are the temperature， precipitation and runoff data of the Glacier No. 1 hydrological station from 2013 to 2016， the stable isotope data （δ¹⁸O） of rainfall and runoff measured during the same period， and the Glacier No. 1 glacial material balancing data， terrain data and the Second Chinese Glacier Inventory. To compare the effects of stable water isotope information on the quantitative simulation and uncertainty range of runoff process， four schemes with different calibration data design by using the glacier-hydrological model （FLEX^G-iso） coupled with water stable isotope information （δ¹⁸O）. Finally， based on the simulation results， the runoff composition of the Glacier No.1 in headwater of Urumqi River section is analyzed. The results show that： under the four schemes， the model not only has a good simulation of the runoff process during 2013—2016， but also can reproduce important processes such as stable water isotope and glacial mass balance. The use of stable water isotope data for the model calibration test improves the reliability of the simulation results during calibration period and evaluation period， resulting in a lower uncertainty range. During calibration period and evaluation period， Scheme 4 has the smallest uncertainty range and good simulation results （the mean KGE of runoff depth is greater than 0.8， the mean square error of runoff δ¹⁸O is less than 0.61‰， and the mean volume deviation efficiency of the glacier mass balance is less than 0.37）. Comparing the four schemes， most of the parameters of Scheme 4 produce the smallest uncertainty range， which shows that the water isotope data has the ability to further restrict the parameters. The use of water stable isotopes as auxiliary data improves the identification of model parameters， especially the parameters related to snow and glacier runoff processes （T_t， F_dd， C_wh and K_f，g）， and reduces the mutual compromise effect of various water sources in the simulation process and the uncertainty range. Calculated from Scheme 4， from 2013 to 2016， about 32%~34% of the runoff from Glacier No.1 in headwater of Urumqi River came from snowmelt， 48%~51% from melting ice， 0%~7% from groundwater， and 12%~15% from rainfall runoff. In the ablation season， melting ice is the largest runoff component， while snowmelt is the second largest runoff component， followed by rainfall runoff， and finally groundwater. In the non- ablation season， snowmelt and groundwater dominate， and ice melting and rainfall runoff contribute little. Stable water isotopes have obvious constraints on the intermediate processes related to snow and glaciers， and the original model’s simulation of the contribution to ice melting is about 7% higher. The establishment of the FLEX^G-iso model establishes a bridge for communication between experiment and model scientists， and helps to promote the development of hydrological theories and methods in cold regions， as well as decision-making related to water resources and ecological environment protection in cold regions.