世界上草原的面积广大，约占全球陆地表面积的40%。草原生态系统为人类提供了多种生态系统服务。草原上饲养的牲畜，包括牛、羊、马、驴和骆驼等，为人类提供肉类和皮毛等供给服务；草原为人类提供净化空气、调节气候、水土保持、水源涵养和防风固沙等调节服务；草原上丰富的动植物，悠久的历史文化和美丽景色吸引了大批游客，为游人提供休闲文化服务。牧民们世代生活在草原上，依靠草原的供给服务（畜牧业）获取收入，其生计和福祉与草原生态系统息息相关。在最近几十年中，全球近一半的草原出现不同程度的退化，对牧民的生计和福祉产生了极大的影响。同时，全球经济形势带来牲畜收购市场价格的波动，也对牧民的畜牧业收入产生影响。面对气候条件和经济形势的双重不确定性，牧民将如何应对未来的多重挑战？在维持草原生态系统服务的前提下，如何提高牧民福祉？ 锡林浩特市位于锡林郭勒大草原的中部，是内蒙古自治区锡林郭勒盟的政府所在地，可利用草场面积占比93%以上，是以草地为主的生态系统。2015年牧区人口为2.56万，占总人口的比例约为14%。在过去的几十年，气候条件的波动、一系列草地政策的实施、以及国内外羊肉市场的冲击，极大地影响了牧民的生计和福祉。为此，本研究以锡林浩特市为案例研究地，定量化分析草原生态系统供给服务与牧民福祉之间的关系，探讨未来气候变化和市场波动可能产生的影响，并提出相应的对策建议。 本研究的主要工作包括以下两个部分： 一方面，首先运用统计资料和遥感影像数据分析气候、草原生态系统和牧民福祉在近三十年的变化，然后结合实地调查数据系统的理解草原生态系统服务与牧民福祉之间的关系，完成两者关系框架图，识别牧民福祉的关键影响因素； 另一方面，本研究通过建立的草原生态系统服务与牧民福祉关系框架图和专家调查结果，在Netica软件中构建了牧民福祉的贝叶斯信念网络（Bayesian Belief Network，BBN）模型。BBN模型是基于概率论和图论的多变量统计模型，能够结合定性和定量的数据，运用概率数据来表征可能的状态，输出在特定情景下各变量的状态变化并随时更新，进行不确定性分析。模型建立之后，针对气候和市场的不确定性进行情景设计，讨论牧民未来三十年可能发生的四种情景，并且提出了相应的对策和建议。 主要研究结果包括以下五点： （1）近三十年，研究区土地利用变化明显分为两个阶段，前期（1985-1995）变化缓慢，后期（1995-2010）变化较快；草原生态系统在2000年前后经历了退化和再恢复的过程；牧区人口则在2000年前后经历了增加和减少的过程；牧民人均纯收入整体呈上升的趋势，但在2001-2003年间有所下降；近三十年城市人口和建成区面积增加，尤其在2000年后，城市化和产业化发展迅速。其中，采矿业和旅游业均在2006年之后发展较快，但2012年后采矿业开始下行，旅游业则大幅增长。旅游业被视为最有发展前景的产业。 （2）降水与归一化植被指数（Normalized Difference Vegetation Index，NDVI）、羊肉零售价格与牧民人均纯收入之间均存在显著的正相关关系。影响牧民福祉的最关键因素是牧民收入，而牧民收入则主要来自于畜牧业生产，受气候、羊价和政策的影响。 （3）BBN模型结果发现，1985-2015年研究区牧民收入和牧民福祉增加的概率最大，分别为56.4%和37%。随着草原生态系统供给服务的增加，牧民收入和福祉也增加。对牧民人均纯收入的影响程度由高到低的关键因子为：牧业净收入、日常生活消费、政府补贴和羊价，所对应的方差缩减值分别为0.0110、0.0021、0.00053和0.00052.。对牧民福祉的影响程度由高到低的关键因子为：牧民人均纯收入、自然灾害、牧业净收入和日常生活消费，所对应的方差缩减值分别为0.0198、0.0034、0.001和0.0003。 （4）在模拟的四种情景中，总体趋势都为牧民福祉增加的可能性最大，这符合研究区政策和经济发展的大背景。牧民人均纯收入增加的概率最多的为情景四“收入和福祉的权衡”，降低的概率最多的为情景一“维持贫穷的趋势”。情景二“美好生活”中的气候条件较好，羊价较高，牧民收入较高，自然灾害较少，牧民福祉最高。情景三“依赖政府的补贴勉强度日”中的气候条件较差，羊价较低，牧民收入较低，自然灾害较多，牧民福祉最低。 （5）以1985-2015年为基准（牧业净收入：51%，牧民人均纯收入：56.4%），情景一中的牧业净收入增加的概率（37.7%）降低，远远小于情景二中牧业净收入增加的概率（65.9%），但牧民人均纯收入增加概率之间的差距缩小（情景一：53.4%；情景二：59.5%），这主要是由于情景一中的牧民多从事多样型生计来弥补畜牧业损失；以1985-2015年为基准，情景三中牧业净收入增加概率（36.1%）降低的程度大于情景一中牧业净收入增加概率（37.7%），但人均纯收入增加概率（54.2%）降低的程度却小于情景一（53.4%），这主要是由于情景三中政府补贴的增加提高了牧民收入。因此，多样型生计和政府补贴对于牧民福祉至关重要。 根据研究结果，本研究提出了相应的政策建议：由单一的政府补贴过渡到商业保险，完善社会保障机制；注重生态旅游，发展多样型生计；进一步打造锡盟品牌，实现草原生态系统供给服务与文化服务的协同提升。

Grasslands are some of the largest ecosystems in the world and accounts for about 40% of the earth’s land surface. Grasslands provide humans with many ecosystem services. Livestock raised on the grassland, including cattle, sheep, horses, donkeys and camels, can provide human with meat and fur and other provisioning services. Grasslands also provide human with air purification, climate regulation, water and soil conservation, wind protection and sand fixation and other regulation services. The rich grassland plants and animals, the long history and the grassland culture and the beautiful scenery attract a large number of tourists, providing leisure and cultural services for people. Herdsmen have lived on the grassland for generations, mainly engaged in animal husbandry, and obtained income through the grassland provisioning service. Their livelihoods and human well-being are closely linked to the grasslands. In the past few decades, almost half of the grasslands showed degradation in different degrees, which greatly affected the herdsmen’s livelihood and well-being. At the same time, global economic situation has led to fluctuations in the livestock prices and also has an impact on herdsmen’s income of livestock. Faced with the uncertainties of climate conditions and economic situation, how will the herdsmen cope with the challenges in the future? How to improve the herdsmen’s well-being while maintaining grassland ecosystem services? Xilinhot City is located in the central part of Xilingol grassland and is the seat of the government of Xilingol League in the Inner Mongolia autonomous region. It is a grassland city with more than 93% of available grassland area. The population of rural pastoral areas was 25,600 in 2015, accounting for about 14% of the total population. In the past few decades, climate fluctuations, the implementation of a series of grassland policies, and the impact of domestic and foreign mutton markets have greatly affected the herdsmen’s livelihoods and well-being. Therefore, we selected Xilinhot City as the study area, quantitatively analyzed the relationship between the grassland provisioning services and herdsmen’s well-being, discussed the possible influence of future climate change and market fluctuations, and put forward some suggestions. The main work of this study includes the following two parts: On the one hand, we used statistical data and remote sensing data to analyze the changes in climate, grassland ecosystem and herdsmen’s well-being in the past three decades. Then we combined field survey data to understand the relationship between grassland ecosystem services and herdsmen’s wellbeing, complete the causal diagram and identified the main factors for herdsmen’s wellbeing. On the other hand, we established a Bayesian belief network (BBN) model of herdsmen’s wellbeing based on the diagram and expert survey data. BBN model is a multivariate statistical model based on probability theory and graph theory. It can combine qualitative and quantitative data and use probability data to represent the possible states, output the state beliefs of each variable in a specific situation and update it at any time for uncertainty analysis. After the model establishment, based on the uncertainty of future climate and market changes, we analyzed four plausible scenarios for herdsmen in the next 30 years, and put forward some suggestions. The main conclusions are represented as follows: (1) In the past 30 years, the land use change in the study area can be divided into two stages, slowly in the early stage (1985-1995) and faster in the later stage (1995-2010). Around 2000, a drought broke out in Xilinhot, and the grassland ecosystem experienced a process of degradation and then recovered. The population of pastoral areas also experienced a process of increase and decrease around 2000. The per capita net income of herdsmen increases on the whole, but it also decreased in 2001-2003. In the past three decades, the urban population and the area of built-up areas have increased. Especially after 2000, Urbanization and industrialization developed rapidly. Both the mining industry and tourism industry developed rapidly after 2006. However, the difference is that the mining industry began to decline after 2012, and the tourism industry began to see a significant increase in revenue since 2012. Tourism is regard as the most promising industry. (2) There was a significant positive correlation between precipitation and Normalized Difference Vegetation Index (NDVI), sheep retail price and herdsmen's per capita net income. Combined with the field survey results, it is found that the most critical factor for herdsmen’s wellbeing in the study area is the herdsmen’s income, which mainly comes from sheep husbandry and can be affected by climate, sheep price and policies. (3) BBN model results show that, from 1985 to 2015, the probability of increase states of herdsmen’s per capita net income and herdsmen’s wellbeing is the highest, 56.4% and 37% respectively. With the increase of grassland ecosystem provisioning services, the herdsmen’s per capita net income and herdsmen’s wellbeing increase. The key factors affecting herdsmen’s per capita net income ranking: net income from animal husbandry, expenditure in daily life, government subsidy and sheep price and the corresponding Variance of Beliefs values are 0.0110, 0.0021, 0.00053 and 0.00052 respectively. The key factors affecting the herdsmen well-being ranking: herdsmen’s per capita net income, natural disasters, net income from livestock and expenditure in daily life and the corresponding Variance of Beliefs values are 0.0198, 0.0034, 0.001 and 0.0003 respectively (4) In the four scenarios, the overall trend of herdsmen’s well-beingis most likely to increase, which is consistant with the policy and economic development background in the study area. The ratio of the highest increase of herdsmen's per capita net income is the scenario 4 “Trade-offs between herdsmen’s income and well-being”, and the ratio of the highest decrease is the scenario 1 “Maintain the poverty state”. Scenario 2 "The beautiful life" has better climatic conditions, higher sheep price, higher income, fewer natural disasters, therefore has the highest herdsmen’s well-being. Scenario 3 “Living on government subsidies” has poor climate conditions, lower sheep prices, lower income, more natural disasters, and therefore has the lowest herdsmen’s well-being. (5) The increase-beliefs of net income from livestock in the “Beautiful life” scenario (65.9%) is much bigger than the “Maintain the poverty state” scenario (37.7%). However, the gap of the increase beliefs of herdsmen’s per capita net income in these two scenarios shrinks significantly (59.5% and 53.4%), suggesting that diversified livelihood can increase herdsmen’s income. The increase-beliefs of net income from livestock in the “Maintain the poverty state” scenario is higher (37.7%) than in the “Living on government subsidies” scenario (36.1%), but the increase-beliefs of herdsmen’s per capita net income in the “Living on government subsidies” scenario is higher (54.2%) than the “Maintain the poverty state” scenario (53.4%), suggesting that the government subsidy is an important income source for herdsmen. Therefore, herdsmen’s decisions about their livelihoods and government subsidies are vital for herdsmen’s per capita net income and well-being. We put forward three suggestions: transition from a single government subsidy to commercial insurance; develop diversified livelihoods with ecotourism; and develop an effective brand to realize the win-wins improvement between grassland provisioning services and cultural services.