【作者】 李明涛；

【导师】 王晓燕；

【作者基本信息】 首都师范大学 ， 自然地理学， 2014， 博士

【摘要】 全球气候变化和人类活动的影响,使得水资源危机和水环境污染成为当前世界各国亟需解决的难题。全球气候变化背景下,降雨、气温等气候因子的时空分布不均,必然会引起流域水循环过程及其生态环境的改变。同时,强烈的人类活动影响,尤其是土地利用变化,使得流域产汇流机制随之改变,进而引起流域水环境的变化。非点源污染作为典型的地表生态过程,受到水文气象等自然因子和土地利用变化等人为因子的综合影响。密云水库作为北京唯一的地表水水源地,具有重要的战略意义。相关研究表明,非点源污染目前已成为影响密云水库水环境质量的重要因素。因此,探讨土地利用和气候变化对流域非点源污染的影响,对认知流域非点源污染过程及其控制管理具有重要的实践意义。鉴于此,本研究以密云水库上游流域为研究区,从土地利用演变和气候变化与污染过程相互作用的角度出发,开展流域非点源污染过程对土地利用和气候变化的响应研究。基于流域1995、2000、2005年三期的土地利用数据,通过构建大尺度流域非点源模型SWAT,模拟评价流域非点源污染负荷分布特征,并应用景观格局指数、典范对应分析(CCA)和通径分析等方法,从全流域和三级保护区等多空间尺度,量化分析流域土地利用及其格局时空变化对非点源污染负荷的影响。基于流域长时间序列的气象观测资料,通过全球气候模式GCMs、SDSM统计降尺度模型和NCC/GU-WG天气发生器预测研究区未来的气候变化情景,并结合SWAT模型进行未来非点源污染负荷的趋势预估,进而探讨非点源污染对气候变化的响应。研究主要内容及结论如下：(1)本研究基于密云水库流域SWAT模型的数据库构建、参数率定和模型验证,模拟评价了流域非点源污染的分布特征。模型校准验证结果表明SWAT模型在密云水库流域具有较好的适用性。1995-2005年的污染负荷结果表明,TN、TP负荷随降雨量的变化基本呈减少的趋势,非点源污染的时间分布受降雨影响显著,其中降雨对TN负荷输出的影响要明显大于TP负荷。污染负荷的空间分布受降雨和土地利用的共同作用影响明显,高风险区主要分布白河流域的赤城县和崇礼县,潮河流域丰宁县,这些区县的乡镇是非点源污染产生的关键源区,需对这些区域影响污染发生的因素进行重点控制。(2)基于研究区1995、2000、2005年三期的土地利用数据,应用空间统计分析和景观格局指数方法,量化分析了密云水库流域1995-2005年十年间的土地利用变化的时空特征。分析表明,10年间密云水库流域的土地利用发生了较大地变化,1995-2000年间林地面积持续减少,草地和耕地面积相应有所增加,土地利用的变化主要体现在林地、草地和耕地之间的转换；2000-2005年间流域的土地利用的变化趋于稳定,各用地类型间略微波动,基本保持不变。10年间土地利用的空间格局则表现出破碎化程度加剧,形状趋于复杂化,斑块类型变得多样等特征,且空间格局的变化具有明显的尺度效应,随着空间尺度的增大,景观多样性和破碎度减小,斑块形状变复杂,而团聚程度增强。(3)利用典范对应分析(CCA)和通径分析等统计学方法,从全流域和三级保护区等多尺度,探讨了土地利用及其格局对非点源污染负荷的影响。结果表明,流域的非点源负荷与土地利用格局间存在着密切的联系。格局指数能累积解释流域非点源污染负荷变化的56.3%。污染负荷在景观水平上受土地利用格局的破碎度和形状的影响较大,而对于不同的用地类型,影响负荷的格局因子的作用表现不同。通径分析的结果表明,耕地、林地面积比例、形状指数和斑块密度是影响研究区非点源污染负荷输出的主要因子,其中形状指数和耕地面积比例对TP负荷的解释能力要明显高于其他指标。从空间尺度上看,各格局因子与非点源污染负荷的关系具有尺度效应,随着空间尺度的递增,格局对负荷的解释程度降低,在较小的尺度范围内,尤其是一级保护区的解释能力最高,达到62.9%,表明离水库越近的区域应是非点源防治高度重视的区域。(4)基于流域长时间序列的气象观测资料,应用SDSM统计降尺度模型和NCC/GU-WG天气发生器2种统计降尺度方法模拟比较了研究区的气候变化。结果表明,2种方法对降水量和气温的模拟均取得了比较理想的结果,其中对气温的模拟效果要优于降水的模拟；NCC/GU-WG在降水量数值模拟方面具有一定的优势,但SDSM对降水和气温变化过程的模拟效果较好。(5)采用全球气候模式GCMsA2、B2和NCC/GU-WG三种情景模拟的区域未来30年的气候变化,并结合SWAT模型预测气候变化条件下的非点源污染负荷响应。3种气候情景模拟的未来降水量和气温的变化趋势基本一致,均呈增加趋势,其中未来降水量的变化相对更为显著。3种气候情景中,B2情景模拟的降水量和气温变化幅度明显高于其他情景。在气候变化的影响下,密云水库流域未来30年的径流量、TN负荷、TP负荷变化也均表现出增加的趋势,其中径流量较基准期的平均增幅最大,增幅为15%,TN、TP负荷的平均增幅相对较小,但年际波动幅度较大。更多还原

【Abstract】 With the effects of global climate change and human activities, the crisis of water resources and water environment pollution has become an urgent issue to be solved in the world. Under the background of climate change, the spatial and temporal heterogeneity of precipitation, temperature, and other climate factors, will inevitably lead changes in water cycle process and the ecological environment. Meanwhile, due to the strong influence of human activities, especially land use/cover change, making the mechanism of the runoff yield in watershed changed, thereby causing changes in basin water environment. Non-point source pollution (NPS), as a typical ecological processes, was affected by the combined effects of natural factor and human factors. Miyun Reservoir, as the only surface water sources, has important strategic significance for water surply of Beijing city. Researches showed that NPS pollution has become an important source affecting the quality of the Miyun Reservoir water environment. Therefore, exploring the impact of climate and land use change on NPS pollution, has important practical implications for cognitive non-point source pollution process and its management.In view of this background and condition, this dissertation took the Miyun Reservoir watershed as the study area, to carry out research in the response of non-point source pollution to land use and climate change. This study used three period land use data of1995,2000,2005and adopted Soil and Water Assessment Tool (SWAT), Canonical Correspondence Analysis (CCA) and Path Analysis to explore the relationship between land use change and non-point source pollution loads on a series of spatial seale. Based on the long time series of meteorological data, Global Climate Models (GCMs), Statistical Downscaling Model (SDSM) and Weather Generator (NCC/GU-WG) were used to predict the future climate change scenarios. Combined with SWAT model, then, the trends of future NPS pollution loads was estimated. The main contents and conclusions are listed as follows:(1) The SWAT model was applied for NPS pollution simulation in the Miyun Reservoir watershed. The model were calibrated and validated with monthly observed data from1995to2004. The results showed that Nash-Suttcliffe coefficient (Ens) was greater than0.8for runoff simiulation, and greater than0.5for TN and TP, which suggested a good agreement between observed and simulated runoff and water quality. The results suggested that SWAT model was suitable for the study area. The pollution loads simulation showed that, TN and TP loads decreased with the rainfall varies from1995to2005, indicating that rainfall has important impact on the temporal variability of pollution loads, and its impact on TN loads was greater than TP loads. The spatial distribution of TN and TP loads were significantly affected by the comprehensive influence of rainfall and land use. The high-risk areas were mainly located in Chicheng, Chongli, and Fengning County.(2) Analysis and compare land use change of Miyun Reservoir watershed between1995,2000, and2005. By applying landscape metrics method, quantitative analysis of the spatial and temporal variation of land use pattern was carried out. It showed that land use occurred significantly change between the10years. From1995to2000, the forestland area reduced remarkably, and grassland, arable land area correspondingly increased. The changes mainly occured in the transitions of forest land, grassland and arable land. Land use change from2000to2005maintain stability, and fluctuate slightly btween different land use types. The characteristic of spatial pattern of land use in the10years tended to be more fragmentation, diversity, and more complicated shape. Land use patterns was closely associated with the spatial scale. With the increase of scale, landscape diversity and fragmentation were reduced, patch shape became complicated, and the agglomeration level enhancement. (3) Canonical correspondence analysis (CCA) and path analysis method were used to explore the relationship between land use change and NPS pollution loads at different scales. The results showed that, there was a significant correlation between NPS pollution loads and landscape characteristics in the study area. Landscape variables can explain56%variation of nutrient loads. At the landscape level, the fragmentation metrics, and shape metrics were the main pattern indices effectively affecting the variation of nutrient losses. At the patch-class level, landscape metrics affecting the spatial variation of pollution process varied with land use types. The result of path analysis suggested that the proportion of arable land and forest land area, patch density, and shape index were the main pattern indices effectively affecting the nutrient pollution process. The interpretation capacity of proportion of arable land and shape index to TN and TP loads was greater than the other indicators. The relationship was closely associated with the spatial scale. With the spatial scales increased, the interpretation capacity of landscape variables reduced. The first grade protection zones had the highest interpretation capacity, which reached62.9%, indicating that the area around the reservoir area should be highly valued for the prevention and treatment of NPS pollution.(4) Based on the long time series of meteorological data, Statistical Downscaling Model (SDSM) and Weather Generator (NCC/GU-WG) were simulated and compared. The results showed that the two statistical downscaling methods both had achieved satisfactory results, and the simulation of temperature was better than precipitation. NCC/GU-WG has certain advantages at simulating amount of precipitation, while SDSM was better at simulating variability of precipitation and temperature.(5) Under GCMs A2, B2and NCC/GU-WG climatic scenarios, the change of precipitation and temperature in the Miyun Reservoir watershed during2021to2050were predicted. The results showed that there were significant increasing trends for the precipitation and temperature in this area. Future climate change simulated by B2 scenario was significantly higher than that of other scenarios. Impacts of climate changes on pollution loads were simulated by using SWAT model with input downscaled climate variables. Runoff, TN and loads showed increasing trends in the next30years. Runoff would increase significantly in the period of2021to2050compared to the period of1961to2000, which had an increase of15%. The inerease of TN and TP loads were relatively smaller than runoff, but had larger interannual fluctuations.更多还原