北方草地是我国重要的畜牧业生产基地和华北地区重要的生态屏障，是国民经济发展的有机组成部分和敏感的生态保育前沿阵地。长期以来，在多种因素驱动下，北方草地的植被退化不仅影响到当地的自然-经济-社会复合体的安全与可持续发展，而且危及华北等地乃至全国的生态安全，也使该区成为我国少数几个主要的生态脆弱区之一。研究该地区的生态状况及其与人类活动的关系，对于开展生态系统的适应性管理、保障区域生态安全、实现社会经济可持续发展等，理论与实践意义重大。本研究选择北方草地典型地区–锡林郭勒盟为研究区，通过野外实地考察、数据搜集、文献查阅以及农牧民访谈等途径，运用地理信息系统、遥感技术和数理统计等方法，以NDVI 为主要数据源，分析了1986-2006年间植被覆盖的变化趋势，并利用空间相关分析方法辨析了植被覆盖空间分布与变化的影响因素；在此基础上，结合社会经济发展状况调研，构筑起以暴露性、敏感性和适应性能力为基础的生态脆弱性评价体系，借助空间主成分分析(SPCA)方法构建了暴露敏感性指数，采用主成分分析(PCA)方法构建了适应性能力评价指数，评价了锡林郭勒盟的生态脆弱性及其空间格局。研究分析了1995 年到2000 年的土地利用变化，并采用典范对应分析(CCA)方法考察了气候、地形、区位和人类活动的空间分布与土地利用变化空间分布的关系；在此基础上，采用多层感知人工神经网络模型(MLP-ANN)，分析区别出土地利用变化的主要类型及其转换潜力，并利用马尔可夫链法(Markov Chain)预测了未来的土地利用状况，指明了土地利用变化的趋势。以居民生计为切入点，利用持续生计方法(SLA)框架评估了现有草原管理政策对居民生产生活的影响。最后，根据各生态脆弱性分区的土地利用变化趋势，针对性地提出了未来区域生态与社会发展的策略。研究的主要结论如下：(1) 1998 年至2006 年，锡林郭勒盟植被覆盖变化趋势系数平均值为1.39，大部分地区呈现出植被改善的趋势，尤以南部和中西部地区为显著，东北部地区则呈现出植被衰退趋势。2000 年，锡林郭勒盟97%的植被发生不同程度的退化，并以轻度和中度为主，退化最严重的区域为中部典型草原区域，其次为东北部的草甸草原区域和东南部的农牧交错带(灌木草原区域)。气候指标与植被覆盖变化的相关性最高，其次是人类活动因子和地形因子。从气候因子与植被覆盖空间格局的关系看出，温度越低、降水越少，干燥度越高的区域植被退化越严重。海拔和地形起伏度是与草地植被变化格局关系最紧密的地形因素，反映出草地覆盖变化主要受到大地形特征的影响。到最近河流、道路、居民点的距离及放牧强度等指标反映了区位条件和人类活动对草地植被的影响。(2) 锡林郭勒盟的暴露敏感性指数以西北部为中心，向四周辐射状递增，最低值出现在东乌珠穆沁旗北部，并与南部阿巴嘎旗略呈马鞍形分布，最高值出现在苏尼特右旗东南部。轻度、中度和高度暴露敏感性区域的面积比重分别为51%，34%和15%。高适应性能力区域包括锡林浩特市、东乌珠穆沁旗和西乌珠穆沁旗；中适应性能力区域包括阿巴嘎旗、二连浩特、苏尼特左旗和苏尼特右旗；低适应性能力区域包括镶黄旗、正蓝旗、正镶白旗、太仆寺旗和多伦县。锡林郭勒盟适应性能力与暴露敏感性整体呈负相关关系。依据暴露敏感性和适应性能力的不同组合，将锡林郭勒盟划分为低敏感性弱适应性区、低敏感性中适应性区，低敏感性强适应性区、中敏感性弱适应性区、中敏感性中适应性区、中敏感性强适应性区、高敏感性弱适应性区、高敏感性中适应性区等八个生态脆弱性类型区。(3) 1995 年到2000 年，锡林郭勒盟的土地利用变化共区分出70 种类型，面积为59129.77km2，占全盟面积的29.69%。面积变化最大的土地利用类型为高覆盖度草地、中覆盖度草地和低覆盖度草地。其中，高度盖度草地转出最多，占全部变化土地面积的44%，主要流向中、低覆盖度草地；其次为中覆盖度草地，流向高覆盖草地和低覆盖草地。草地覆盖度提高的区域占土地利用变化面积的21%，显示草地退化与改善并存，但以退化草地面积净增为主。CCA 排序显示地形起伏度、海拔高度、坡向指数、年均降雨量、最近居民点距离等指标与土地利用变化空间格局的关系最为密切。预测显示，未来高覆盖度草地的减少幅度最大，中覆盖度草地的减少相对和缓，尤其是前者增加的幅度最大。高覆盖度草地减少的区域集中在西北部，主要转变为中低覆盖度草地，中覆盖度草地的减少集中在西南部，多流向未利用地(4) 草原“围封、禁牧、移民”政策解析为牧区休牧、牧区禁牧移民和农区禁牧三种形式。休禁牧政策的执行极大影响了农牧民的生计，使其生计资本和生活水平发生了重要变化，居民固定资本、自然资本、财务资本均大幅减少，肉类消费明显下降。其中，牧区禁牧移民对牧民生计的影响最大，其次是农区禁牧和牧区休牧。居民通过进城务工、调整种养和饮食结构等手段适应政策变化。居民对单一主体收入来源依赖性较高，生计多样性总体较低，其持续性受到气候和市场波动的威胁，生计风险较大。家庭收入的弹性恢复能力受到教育水平低下的限制。休禁牧政策在一定程度上提高了草地生产力和覆盖状况，增强了环境持续性。加强基础教育和职业教育，提高劳动者素质，加速城镇化发展，拓宽就业门路，增强城市的人口吸纳能力，完善决策过程中的公众参与机制，是减轻政策实施对居民生计影响的主要途径。(5) 据各生态脆弱性区未来土地利用变化模拟情况，发生变化面积最大的区域是低敏感-强适应性区域，其次为低敏感-中适应性区域和中敏感-中适应性区域。预期土地利用变化比重最大的区域是中敏感-弱适应区域。最后，针对不同区域的差异，概括了土地利用变化的方向，指出了加强人地系统生态管理的策略。

The grassland in Northern China is an elementary area not only for its sizable animalhusbandry but also for its crucial location as a buffer zone in sheltering desertification for NorthChina, playing an indispensable role in economic development and environmental conservationfor the whole China. The grassland degradation driven by a variety of factors in the past decadeshas affected the local ecological security and socialeconomic sustainability and is endangeringthe ecological security of the downstream area, turning this area one of the major ecologicallyfragile zones in China. Study on the relationship between ecological change and human activitiesis of great necessity for improving local adaptive ecosystem management, securing regionalecological balance, and maintaining the long-lasting socialeconomic sustainability.As a typical steppe region in Inner Mongolia, Northern China, Xilingol League was takenas the case study area. Through field investigation, socialeconomic data collection, literaturereview and household interviews, the relevant data were obtained and further analysis to thosedata was conducted based on the GIS and RS technology and quantitative statistics. Referring tothe NDVI data from 1986 to 2006 as the main source, the trend of vegetation cover change wasanalyzed and the driving forces leading to the change were distinguished with spatial correlationanalysis, which formed the base to further construct the exposure-sensitivity index (ES). Theindicators highly correlated to vegetation cover change were chosen to compose the ES withSpatial Principal Components Analysis (SPCA), and the socialeconomic data were used to buildup the adaptive capacity index (AC) with Principal Components Analysis (PCA), and thus avulnerability assessment system with ES and AC as main index components was established.Based on this assessment system the vulnerability of Xilingol was spatially evaluated byintegrating the ES index and the AC index. Meanwhile, the types, area and spatial distribution of land use and cover change (LUCC) were extracted with the land use data of 1995 and 2000, andthe driving forces of LUCC from perspectives of climate, topography, location and humanactivity were screened using Canonical Correspondence Analysis (CCA). The selected land usechange drivers in combination with the original land-use types were used to reveal the conversion potential of the main LUCC types with Multi-Layer Perception Artificial NeuralNetwork model (MLP-ANN), and the future LUCC tendency was predicted using the method ofMarkov Chain to point out region’s land use changing trend. To reveal the effect of grasslandmanagement on herdsman’s livelihood, the current grassland management policy of “FencingGrassland, Forbidding Grazing and Moving User” was analyzed by calibrating the policy intothree livelihood forms, rotational grazing in grassland region, forbidden grazing and user movingfrom grassland region, and barn feeding livestock raising with grazing control infarming-pastoral zone. Livelihood was used as a proxy for human living status and analyzed withthe framework of Sustainable Livelihoods Approach (SLA), and for each policy the livelihoodchange of residents involved was assessed through scenario analysis, to reflect the effects ofpolicy implementation so as to benefit the policy optimization. Based on those above-mentionedresearches, the land use changing trend of each vulnerability zone in Xilingol was finallyachieved and some pertinent strategies to mitigate their vulnerability are proposed for futureregional ecological conservation and socialeconomic development. The main conclusions of thestudy are as follow:(1) The average coefficient of vegetation change trend in Xilingol is 1.39 from 1998 to 2006,which is a reflection of both the vegetation improved in most areas of Xilingol, particularly inthe south and the mid-west, and the only degraded in the northeast. In 2000 in Xilingol, 97% ofthe vegetation degraded to varying degrees, with most areas belonging to light to moderate,whereas the others including the central typical steppe area to the most severe, and the meadowsteppe in northeast and the farming-pastoral zone in the southeast to the severe. Climaticindicators are more significantly correlated with the vegetation change than human activityfactors and topographic factors. From the perspective of climate, the less favorable rainfall andtemperature regime results in more severe degradation. Elevation and terrain undulation are thetwo topographic factors most significantly associated with land cover change in Xilingol,showing that the distribution of land cover change is closely controlled by the macro topographicfeatures. The distance to the nearest river, road, or settlement, and grazing intensity are alsoclosely associated with land cover change, which reflects the strong influence of locationadvantages and human activities on grassland vegetation.(2) The ES index of Xilingol increases in a radial pattern with the valley in the northwestand the peak value in south Right Sonid. The low ES area, moderate ES area and high ES areacover 51%, 34% and 15% of Xilingol, respectively. The counties with high AC scores includeXilinhot, West Ujumchin and East Ujumchin; whereas Abaga, Erenhot, Left Sonid and RightSonid share the moderate AC; and Xianghuang, Zhenglan, Zhengxiangbai, Taipusi and Duolunwith the low scores. The ES index is negatively correlated with the AC index. According to the combination of ES and AC, Xilingol can be compratively divided into eight subregions,including the low ES-low AC zone, the low ES-moderate AC zone, the low ES-high AC zone,the moderate ES-low AC zone, the moderate ES-moderate AC zone, the moderate ES-high ACzone, the high ES-low AC zone and the high ES-moderate AC zone. Each of the subregionalzones had its unique features both in ES and in AC.(3) From 1995 to 2000, the area where the land use has changed from type to type occupies30% (5 912 9.77 km2) of Xilingol, among which 70 land use change types in total wereidentified. 44% of the changed area converted from high-coverage grassland to moderate- and/orlow-coverage grassland, whereas only 21% of the changed area shifted from low-coveragegrassland to moderate-/high- coverage grassland or from moderate- to high-coverage. The CCAordination has shown that terrain undulation, altitude, TRASP, average annual precipitation, andthe distance to the nearest settlement are the main drivers of the land use change in Xilingol. Themain predicted flows of land use change in the future are from high/moderate coverage grasslandto vacant land and low-coverage grassland. During the process, high-coverage grassland willbroadly convert to moderate-coverage grassland in the northwest, while the vacant land willdramatically increase in the southwest from moderate-coverage grassland.(4) The implementation of the new grassland management policy has essentially influencedthe livelihood of involved herdsman residents by drastically changing their asset compositionand diet structure. The fixed capital, natural capital and financial capital declined to varyingextent according to the scenarios with different policy implementation. In the meantime, the dietstructure also changed, tending to much decreased red meat consumption but increasedvegetables and cereal ingestion. In general, the livelihood of the ecological migrants involved by“Grazing forbidden and user moving” policy have been nearly thoroughly changed, followed by“livestock rearing control” policy in the Farming-Pastoral Zone and “Grazing Rotation” policy inthe typical grassland region. Even the residents tried to adapt to the new policy environment byexpanding their livelihoods and life style, such as getting employed in urban or township area oradjusting the production structure of crop and livestock, however, their livelihood diversity wasstill limited to a low level, due to regional harsh natural environment, ethnic traditionalproduction customs and limited economic capacity. The local households with their livelihoodsmainly depending on one to two major income sources will set their livelihoods at much risk offrequently occurring natural disaster and market fluctuation. In addition, the livelihood elasticitywas constrained by the low education level. However, the grassland management policiesenhanced the sustainability of livelihood environment by improving the quality of grasslandvegetation. Fostering the elementary education, improving vocational education level, andpartially migrating the pastoral population to urban and township areas by guiding them withnecessary skills to expand their employment opportunities there, are some of the effectiveapproaches to improving household livelihoods in the region. To do so, some amendments of thegrassland management policy need to be presented and implemented thereafter, combined withmore participation of local residents in the decision-making process.(5) The low ES-high AC zone will be the area with the largest share of land-use change inthe future, followed by the low ES-moderate AC zone. However, the highest expected proportionof land use change may occur in the moderate ES-low AC zone with a percentage up to 25%.Based on the regional characteristics and predicted trends of future land use change, theresearch also put forward a few suggestions for each subregion for the decision makers at alllevels to improve the grassland human-environment system management.