中文摘要：

由于GCM模拟的气温数据分辨率不高，很难用于区域尺度上各种生态系统的模拟。本文基于长时间序列（1964～2007年）的全国气温观测数据，结合经纬度数据、以及DEM、坡向、坡度等系列地形特征数据，利用空间统计方法，在构建年平均气温降尺度模型的基础上，运用高精度曲面建模（HASM）方法对HadCM3的A1Fi、A2a和B2a三种情景1961～1990、2010～2039、2040～2069、2070～2099年时段的全国未来平均气温进行高精度曲面模拟。模拟结果显示，在四个时段内，A1Fi、A2a和B2a三种情景的全国未来平均气温均呈持续上升趋势。其中，平均气温在A1Fi情景中上升速度最快，B2a情景中增速最慢；A1Fi和A2a两种情景的平均气温均呈加速上升和增加趋势，而B2a情景的平均气温则呈减速上升和增加趋势。本文构建的降尺度模拟方法可以有效地实现IPCC GCM的低分辨率的气温未来情景数据降尺度转换成高分辨率气温数据。

英文摘要：

Scenario simulations of ecosystems and their services require the climate data mostly from the Global Climate Models （GCMs）. Because of most GCM simulated data with the coarse resolution （about 200-500 kin）, it is difficult to use these data to assess impacts of temperature change on various ecosystems on regional and local scales, although the data can be used to effectively predict the future temperature change on a global scale. To address the above issue, the downscaling model of mean temperature is developed with the spatial statistical method in this paper, combined with series data of DEM, latitude and longitude. For validating the downscaling model, the mean annual temperature is simulated under the three scenarios of HadCM3 A1Fi, A2a and B2a during the periods T1 （1961～1990）, T2 （2010～2039）, T3 （2040～2069）, and T4 （2070～2099）. In the simulation process, the data resolution is downscaled from 3.75°×0. 125° to 1 km × 1 km by High Accuracy Modeling （HASM）. Simulation results show that mean temperature would continue to increase in the future 100 years under the three scenarios of HadCM3 （A1Fi, A2a and B2a）. During the periods from T1 to T4, the rising rates of mean annual temperature are the greatest under scenario A1Fi, the average levels under scenario A2a, and the lowest under scenario B2a, which might increase per decade by 0.29℃, 0. 23℃, and 0.17℃, respectively. Furthermore, mean annual temperature would have a changing trend of accelerated increase under scenarios AIFi and A2a, but the increasing trend would be slow under scenario B2a in the next 100 years. The results show that the temperature data with IPCC GCM＇s coarse resolution can be effectively downscaled to high-resolution temperature data that could be used to assess the ecosystems and their services on the regional and local scales.