中文摘要：

水分条件不仅影响半干旱区群落的组成,而且在一定程度上决定了群落的功能。处于不同水分条件生境下群落的优势物种在水分利用和同化物利用效率方面的功能特征会存在差异,这些差异将导致群落对于气候变化产生不同的响应,进而影响到景观和区域尺度上对于全球变化下碳动态和格局的分析。该文选取了锡林河流域典型草原区沿水分梯度的4个代表群落,在野外实验测定并结合长期定位研究成果基础上,利用BIOME-BGC模型对代表群落的长期净初级生产力（Net primary productivity,NPP）动态进行了模拟和模型验证。通过分析该地区1953~2005年气候变化趋势,推测了未来可能的气候变化情景,进而模拟了气候变化下4个群落长期NPP动态的响应。结果表明,当前气候条件下,羊草（Leymus chinensis）群落NPP平均值为197.76gC·m^-2（SE=7.11）,大针茅（Stipa grandis）群落NPP平均值为198.95gC·m^-2（SE=6.41）,贝加尔针茅（Stipa baicalensis）群落NPP平均值为210.41gC·m^-2（SE=7.87）,克氏针茅（Stipa krylovii）群落NPP平均值为144.92gC·m-2（SE=4.64）,4个群落NPP平均值为188.01gC·m^-2（SE=3.72）;气候变化情景下,温度增加下（P0T1）,NPP平均下降14.2%,降水增加下（P1T0）,NPP平均增加13.2%,温度与降水都增加情景下（P1T1）,NPP平均下降2.7%,但由于生境水分条件差别和优势物种功能特征差异,4个群落表现出了增减幅度不同的趋势。对气候因子的敏感性分析及回归分析表明,降水是该地区NPP最主要的决定因子,而温度决定作用相对较小,主要通过影响植物的呼吸和水分蒸散等过程影响NPP。在最有可能代表未来气候变化的温度增加的两种情景下（P0T1、P1T1）,NPP均呈下降趋势。群落NPP对气候变化的响应趋势与水分胁迫系数（Water stress index,WSI）、碳胁迫系数（Carbon stress index,CSI）变化密切相关。克氏针茅群落由于所处生?

英文摘要：

Aims Water availability influences community composition and function in semi-arid areas. Habitat water availability may differentiate functional traits on water and carbon assimilation use efficiencies of dominant species, which may induce different response patterns of communities to climatic change. Our objectives are to 1） quantitatively analyze long-term net primary productivity （NPP） dynamics of typical grassland communities along a water gradient based on field measurements, long-time monitoring results and a processbased model; 2） analyze trends of temperature and precipitation based on local long-term meteorological data and predict future climate change scenarios and 3） predict NPP responses of the communities along the water gradient to climate change and interpret the results by means of water stress index （WSI） and carbon stress index （CSI）. Methods We selected four grassland communities along a water gradient in the Xilingel River Basin of China as our study sites： Stipa baicalensis, Leymus chinensis, Stipa grandis, and Stipa krylovii. We conducted field surveys and measurements of physiological parameters and soil parameters in summer 2005 and used local meteorological data （1953 to 2005） to analyze trends of precipitation and temperature and prescribe future scenarios. A process-based model, BIOME-BGC, was parameterized and validated using field data and long-term monitoring data and mn on daily steps to simulate NPP dynamics of the four communities under the current climate and future scenarios. We calculated water and carbon stress indices for each community under each scenario to interpret possible mechanisms. Important findings Minimum and maximum temperatures increased from 1953 to 2005, while the precipitation varied. Average NPP of the four communities decreased 14.2% under the scenario of increasing temperature, increased 13.2% under the scenario of increasing precipitation, and declined 2.7 % under the scenario of increasing both. Precipitation is the predomina