寒旱区流域冰雪径流的动态过程分析

A variety of research actions such as (i) studying on ice-snow runoff processes for watershed in cold-and-dry region, (ii) in-deepth understanding special hydrological processes and mechanisms of high-mountain region, and (iii) disclosing impacts of uncertainties existing in many systems' components and their interactions on hydrological processes of inland river, would be of significance for improving watershed rainfall-runoff simulation and prediction precision. A case study will be carried out in the Kaiduhe watershed, which is one of typical watersheds of Tarim River headstream and possesses a variety of frost-circle elements such as snow, glacier and tundra. The mean elevation of Kaiduhe ranges from 2,400 to 2,600m above sea level in the basin, and from 4,000 to 5,500m in the mountain. The spatial and temporal distribution of precipitation is strongly heterogeneous. Based on data collection and multi-year on-site monitoring, this research will focus on analyzing water and energy exchange in ice-snow runoff hydrological processes. It will study variations of energy parameter of ice-snow melting processes along with altitude grads, gradient, and slope-direction, vertical-direction movement mechanism of water melted from glacier and snow, as well as surface accordant junction of glacier-tundra layer. Consequently, it can better disclose watershed-scale hydrological rule of cold-and-dry region. In addition, techniques of fuzzy mathematics programming, stochastic analysis, scenario analysis, and factorial analysis will be employed to study effects of multiple uncertain factors (e.g., climate, hydrology and land use) as well as their interactions on ice-snow runoff processes. These methods can effectively handle uncertainties associated with multiple parameters with multi-layer interactions that exist in watershed systems. Furthermore, a number of methods such as Monte Carlo simulation, multivariate analysis, and fuzzy mathematics, will be advanced to quantitatively evaluate uncertainties in ice-snow runoff processes, such that more effective simulation and prediction to watershed runoff would be achieved.

研究寒旱区流域冰雪径流过程，深入认识高山区特殊水文过程与机理，揭示大量系统成份及其间关系的不确定性对内陆河水文过程的影响，对提高流域降水径流模拟和预测精度有重要研究意义。本研究选取位于塔里木河源区南部天山山区的开都河流域作为研究区。该流域是塔里木河源区具典型代表性的高寒山区流域，集积雪、冰川、冻土等冰冻圈要素于一身。在数据收集、现场观测及特征识别研究的基础上，项目将致力冰雪径流水文过程的水分与能量交换分析。研究冰雪消融过程能量参数随海拔梯度、坡度、坡向的变化，冰川积雪消融水的垂向运动机理，以及冰川冻土层地表汇流，从而更好地揭示流域尺度的寒旱区水文规律。此外，运用随机分析、情景分析、析因子分析等手段研究气候、水文和下垫面多因子不确定性及其互动对冰雪径流过程的影响；并应用蒙特卡洛模拟、多元分析、模糊数学等方法，定量表征冰雪径流过程中的不确定性，实现对流域冰雪径流的有效模拟与预测。

研究寒旱区流域冰雪径流过程，揭示大量系统成份及其间关系的不确定性复杂流域水文过程的影响，对提高流域降水径流模拟和预测精度有重要研究意义。本研究以开都河、泾河流域为依托，对流域系统的复杂性进行了辨识，并剖析了各系统组分的特征以及它们之间的复杂互动关系。具体如下：（1）评估不同的DEM分辨率和子流域划分水平的组合方式与水文模型模拟效率之间的关系，研究发现：不同空间分辨率组合对径流的影响具有一定差异性,提高DEM精度和增加流域划分水平有助于提高模型表现，但不能显著改善模型的径流预测能力；（2）耦合SLURP分布式水文模型、二型模糊集、析因子分析，处理模型参数不确定性，研究发现：模型参数不确定性对模拟结果影响显著，评估模型参数的敏感性，以及参数间相互作用对模型输出的影响非常必要；（3）集合马尔科夫蒙特卡罗、析因分析与SWAT水文模型，研究参数不确定性以及参数对水文响应的影响及交互效应，结果表明：径流曲线数（CN2）对开都河流域径流夏季峰值主效应最高，融雪过程参数（SNO50COV）对融雪期峰值非线性效应明显，土壤蒸发补偿系数（ESCO）和融雪过程参数（SNO50COV）对年均值效应明显；（4）耦合多GCM、随机天气发生器、逐步聚类分析，揭示了开都河流域的冰川与融雪水直接与气温变化相关，降水量直接导致径流的变化；（5）基于贝叶斯分析的DRAM方法，分析水文模型参数和输入数据（温度、降雨）的随机性特征，研究发现：输入数据不确定性不仅影响单个参数的随机性特征，还影响参数间的交互作用，综合考虑温度和降雨输入数据的不确定性以及模型参数的不确定性有助于提高复杂流域（混合产流）水文模型的模拟预报能力。项目在Journal of Hydrometeorology, Climate Dynamics, Journal of Hydrology, International Journal of Climatology, Hydrology Research等国际SCI杂志上发表论文28篇(已标注)，培养博士和硕士研究生8名。