【作者】 李洋；

【导师】 魏晓妹；

【作者基本信息】 西北农林科技大学 ， 水文学及水资源， 2008， 硕士

【摘要】 水资源是国民经济的命脉和生态环境的支柱,是实现可持续发展的重要物质基础。近年来,全球气候变化以及频繁的人类活动对水循环产生了重要影响,内陆干旱区因其独特的地理气候条件,影响更为显著。因此,研究气候变化和人类活动背景下的水循环要素变化特征,对深入认识内陆干旱区水资源形成和演变规律、合理开发利用水资源和保护生态环境具有重要的科学价值。本文以典型的内陆干旱区——石羊河流域作为研究区域,依据流域内民勤、永昌、武威、阿拉善右旗、古浪、天祝、门源、景泰8个气象站点1960～2005年的降水量和蒸发量资料和流域8条主要河流出山口水文测站及红崖山水库等9个水文站1950～2000径流资料,运用Kendall秩次相关检验法定量的研究了流域水循环要素年内、年际变化特征,利用EOF正交分解法以及EOF-SVD综合分析法分析了流域降水量和蒸发量的空间分布特征,并在此基础上建立主成份回归模型和人工神经网络模型对红崖山水库入库径流量进行模拟,对其主要影响因素(自然因素和人类活动)及影响程度进行了分析。主要研究结果如下:(1)流域降水量年内分配不均,受夏季季风影响,夏季降水量大,占全年降水量56%,冬季降水量最小;流域多年降水量呈增加趋势,但变化不显著。20世纪70年代～90年代流域降水量比较稳定,60年代～70年代以及90年代初降水量年际间波动幅度较大。流域多年降水量包括三种主要空间分布模态:南北递减型,东西反相分布型以及南北反相分布型。降水量变化受到相对湿度和水汽压的影响,民勤和阿拉善右旗是影响比较显著的区域。(2)流域蒸发量主要集中于春季和夏季,分别占全年蒸发量的32%和41%,冬季蒸发量小;流域多年平均蒸发量呈显著下降趋势,且夏季蒸发量下降趋势显著于其它季节。20世纪60年代末期～21世纪初期流域蒸发量较为稳定,60年代前期及21世纪初(2000～2005)蒸发量波动较大。流域多年蒸发量包括两种主要的空间分布模态:周边-中心分布型以及东西反相分布型。流域蒸发量受到风速的影响,民勤,阿拉善右旗、门源和天祝是影响比较显著的区域。(3)径流量集中集中于6～9月份,占全年径流量58%。各站径流量在20世纪50～70年代均大幅下降,70～90年代各站径流量呈小幅波动状态,但下降趋势不明显;90年代之后又开始出现明显减少趋势,但幅度小于50～70年代。各站多年径流量均呈下降趋势,除西大河和大靖河外,其余各站下降趋势显著,尤其以红崖山水库站下降最为明显。八条出山河流的径流量主要受祁连山区气候因素的影响,而红崖山水库入库径流量既受气候影响,也受流域中游人类活动的作用。(4)基于流域径流量影响因素的分析,建立了主成份回归模型和人工神经网络模型,对红崖山水库入库径流量进行了模拟研究,研究结果表明:人类活动因子对红崖山水库入库径流量的变化起主导作用,占主影响力的59%,自然因素影响力占41%;两模型的拟合效果均比较理想,可用以红崖山水库入库径流量的模拟研究,相对而言,人工神经网络模型的拟合精度略高于主成份回归模型的拟合精度。更多还原

【Abstract】 Water resources are the support of national economy and ecological environment, and are also the material base of sustainable development. These years, the Global Warming and the more and more human activities make important influence, the inland arid region which has special weather and environment has been influenced deeply. Hence, the analysis for the characteristics variety of water cycle factors in the background of climate change and human activities are essential for deeply knowing the water resources formation and evolution rules, rational exploitation and utilization and environment protection.This thesis is based on the typical inland arid region ----Shiyang River Basin. The data is from 8 weather stations (Minqin, Yongchang, Wuwei, Alashanyouqi, Gulang, Tianzhu, Menyuan and Jingtai) and 9 hydrologic stations. Data for rainfall and evaporation is from1960 to 2005 and data for runoff is from 1950 to 2000. The main methods for analysis are Kendall rank test, by which we can know the degree of the change of Water Cycle factors, EOF (empirical orthogonal function) and EOF-SVD method, though which we will know the rainfall and evaporation spatial distribution in this basin. After that, the principal component regression model and ANN model have been built to simulate the runoff of Hongyashan Reservoir and analysis the main influence factors (nature and human activities) of it. The main results are as follows:(1) Rainfall is not even in the year, it is highest in summer (56% of the year) and lowest in winter. The multi-year average precipitation shows the uptrend, but trend is not significant. From 1970s to 1990s the rainfall remains a stable level, 1960s to 1970s and early 1990s it fluctuates obviously. The multi-year average precipitation has three spatial distribution types: south-north decrease, east-west opposite and south-north opposite. The relative humidity and vapor pressure influence the rainfall significant and Minqin and Alashanyouqi are the areas where rainfall is influenced significantly.(2) Evaporation is highest in summer and spring, (32% and 41% of the year respectively) and lowest in winter; the multi-year average evaporation shows the significant downtrend, and in summer it changes the most. From late 1960s to late 20th century the evaporation remains a stable level; before 1960s and early 21st century (2000-2005) it fluctuates obviously. The multi-year average evaporation has two spatial distribution types: around-center type and east-west opposite. The wind speed influences the evaporation significant and Minqin, Alashanyouqi, Menyuan and Tianzhu are the areas where evaporation is influenced significantly.(3) Runoff is highest in the period from June to September which is 58%of the year. From 1950s to 1970s runoff shows significant downtrend; from 1970s to 1990s a little downtrend; after 1990s down obviously again, but less than the first period (1950-1970). Except West River and Dajing River, all the rivers show significant downtrend and the Hongyashan Reservoir down the most. For the eight river, the nature factors in Qilian Mountain are the main reason for reduce, but for the runoff in Hongyashan Reservoir, human activities are another important reason.(4) Base on the foregoing analysis, the principal component regression model and ANN model have been built to simulate the runoff of Hongyashan Reservoir. And the analysis shows human activities have 59% of the power and nature factors have 41% of the power. That is to say the human activities are the main influence factor for Hongyashan Reservoir runoff change. Both the two model has good effect in the simulation, although comparatively speaking the ANN model has better precision in simulation.更多还原