在我国半干旱地区，地表水和地下水缺乏，降水是生产和生活的主要水源，但时空分布不均，利用效率低。雨水集流是解决半干旱区水资源短缺的有效途径之一。本论文通过对半干旱区降水特性的系统分析，估算了半干旱区的集雨潜力。通过供需水量平衡分析，提取反映集雨系统缺水状况的风险因子，构建了雨水集流风险指数模型。在此基础上，以半干旱区雨水集流典型区域－甘肃省定西县和内蒙古准格尔旗为例，从经济、社会与生态三个层面剖析雨水集流效益，建立综合效益评价指标体系和评价模型，为雨水集流效益的量化提供参考。主要结论如下：1、半干旱区降水时空变率较大，以小雨事件为主，但降水总量主要取决于大于10 mm的降水。小于10 mm降水占总降水事件的85%，占总降水量的39.8%；其中，小于5 mm的降雨总量年际变化较小，在61-99.4 mm之间，占年降水量的14.1-24.6%。大于10 mm降水多年平均次数介于6.8-17.4，占年降水量的60.2%，呈东南至西北逐渐减少趋势。连续无降水事件主要发生于9月至次年5月间。最长连续无降水天数介于15-64.2，呈中部至东北、西南方向递减趋势，半干旱区东北与西南边缘为15-30天。2、针对半干旱区混凝土人工庭院和自然黄土坡面两种集雨类型，利用日降雨量-累积年雨量曲线模型估算其集雨潜力。结果表明，混凝土人工庭院集雨类型比自然黄土坡面集雨类型集雨潜力高60%。在平水年、干旱年和极端干旱年，庭院集雨类型集雨潜力的平均值分别为369 mm、290 mm、261 mm；黄土坡面集雨类型的集雨潜力分别为245 mm、175 mm、161 mm。总体上，集雨潜力具有由东南向西北逐渐减少的空间分布特征。3、针对半干旱区现有的“人畜饮水”和“人畜饮水+集雨补灌”两种用水模式进行缺水风险综合分析。结果表明，“人畜饮水+集雨补灌”模式下，半干旱区各地均存在缺水风险，且风险概率超过0.5的区域占半干旱区的40%。“人畜饮水”模式下，缺水风险性较小，缺水风险区域仅占半干旱区的62.5%，且风险概率均小于0.5。雨水集流高风险区位于半干旱区西北部；低风险区位于半干旱区东南边缘；内蒙古中东部，河北、山西、陕西、青海北部为高低风险过渡地带。根据经济效益最大化原则，雨水集流高风险区应以发展“人畜饮水”模式为主，中风险区适宜发展“人畜饮水+集雨养殖”模式。4、以黄土高原半干旱区甘肃省定西县和鄂尔多斯高原半干旱区内蒙古准格尔旗为例，进行集雨效益综合评价。结果表明，雨水集流能够显著提高社会、经济和生态效益，体现在能够提高作物水分利用效率，保障粮食安全，增加农民收入，促进产业结构调整，改变消费结构，提高人口素质，拦蓄降雨径流，减轻土壤冲刷侵蚀，改善区域生态环境等方面。通过雨水集流，黄土高原半干旱区和内蒙古鄂尔多斯高原半干旱区的综合效益指数分别由0.147、0.204提高至0.614、0.843，综合效益均提高三倍。

In semiarid region of China, rainfall is the major water resource, and poorly distributed in time and space. Rainfall harvesting offers a new approach to mitigate the recurrent drought and water shortage in the region.The paper, with semiarid region as an interesting study area, set up mathematical model and evaluated rainwater harvesting potential, based on the analysis of the precipitation data and the study of typical rainwater harvesting (RWH) modes. Then a risk assessment model was put forward for analyzing the water shortage risk of rainfall harvesting system. In addition, the paper assessed the integrated benefit of rainfall harvesting in the loess plateau and the Erdos plateau, by using weighted comprehensive analysis and analytic hierarchy process. Results are listed as follows:1. Temporal and spatial variation of precipitation in semiarid region was quite unstable. Precipitation events, which were 10 mm or less, accounted for 85% of precipitation events and 39.8% of total precipitation. It indicated that light precipitation contributed to most of the precipitation events, but total precipitation was determined by the events of 10 mm or more. The amount of rainfall received in the form of 5 mm or less varied slightly among years. Mean effective precipitation frequency per year was 6.8-17.4 days in semiarid region. Extreme dry spell was defined as the longest continuous non-precipitation days, and it usually occurred from September to May. The shortest extreme dry spell located in northeast and southwest semiarid region.2. Based on established a function that described the relationship between daily rainfall above given intensities and annual rainfall in excess of the given daily amount, the paper evaluated for water harvesting potential. The result indicated that under guarantee of 50%, 75%, 90%, RWH potential of concrete catchments and natural loess slope were 369 mm, 290 mm, 261 mm and 245 mm, 175 mm, 161 mm respectively. Generally, the former catchments’ potential was 1.6 times of the latter.3. Choosing the primary rainwater harvesting types, one was RWH for domestic and livestock, the other was RWH for domestic and supplemental irrigation, analyzed the risk of water shortage under these two typical models. The result indicated that, the former had lower risk than the latter. The high risk areas of water shortage located in northwest of semiarid region, the low risk areas located in southeast of the region, the middle belt of the semiarid region was the medium risk zone. According to principle of economic optimization, high risk zone was suitable for developing the courtyard rainfall collection model, and rainwater harvesting for cultivation purposes was fit for medium risk zone.4. Taking Dingxi of Gansu Province, Zhungeer of Inner Mongolia as research area, set up the mathematical model and indicators system for integrated benefit evaluation of rainwater harvesting. The research indicated that, water harvesting had the benefit of solving drinking water problems, increasing water use efficiency, getting rid of poverty, reducing soil erosion, improving eco-environmental status. Through harvested rainwater utilization, the integrated benefit of the loess plateau and the Erdos plateau had increased 3 times either.