全球平均气温未来情景的降尺度分析

Downscaling of Global Mean Annual Temperature under Different Scenairos

如何提高全球气候模拟数据的分辨率,以满足全球、区域乃至局地陆地生态系统全球变化响应的定量分析,是当今全球气候变化研究的核心内容之一。在全球尺度上,本文利用全球气象观测站点的气候数据和DEM数据,对全球年平均气温与纬度和海拔高程之间相关性进行回归分析,建立全球气候降尺度空间模拟的统计转移函数,并与高精度曲面建模(HASM)方法进行集成,从而实现IPCC GCM HadCM3的模拟数据从3.75°×2.5°到0.125°×0.125°的降尺度处理。研究结果表明,在3种气候情景的T1-T4时段内,格陵兰岛平均气温在0℃以下的区域和南极洲平均气温在-35℃以下的区域均呈逐渐缩减趋势,赤道至南北回归线之间的平均气温大于40℃以上的区域呈逐渐增加趋势。其中,A1Fi情景的平均气温上升速度最快,A2情景次之,B2情景的平均气温上升速度最慢。构建降尺度方法有效地将IPCC GCMs的粗分辨率的气候情景数据降尺度转换成高分辨率的气候数据,并克服和弥补了目前IPCC GCMs的模拟数据因分辨率低而不能对区域乃至局地气候变化的细节及趋势进行刻画的缺陷。

One of the key issues of global change research is how to improve the simulated data resolution of Global Climate Models （GCMs） for the quantitative analysis of terrestrial ecosystems in response to the climate change at global, regional and local levels. In this paper, the statistcial transfer funcitons are developed by estab- lishing the regression analysis of relation between mean annual temperature and latitude and elevation with the digital elevation models and climate data from global meteorological stations aton global level. The High Accu- racy Surface Modelling （HASM） method integrated the statistical transfer functions, is used to downscale the simulated data of HadCM3 from a spatial resolution of 3.75~ x 2.5~ to 0.125~ ~ 0.125~. The simualted results of A 1Fi, A2 and B2 scenarios show that the mean annual temperature would be increasing in the 21 st century, the areas in Greenland where the mean annual temperature is below 0~C and in Antarctica below-35~C would shrink, and the areas between north and south tropics where the mean annual temperature is higher than 40~C would expand. The increase rate under scenario A1Fi is the highest and that under scenario B2 is slowest among three scenarios during the period from T1 to T4. The results also show that the coarse resolution data of IPCC GCMs can be availably downscaled to high resolution data by integrating the statistcial transfer funcitons and HASM methods, which could overcome the limitation that the current simulated data resolution of IPCC GCMs can not be used to explain and describe the details of climate change at regional level, especially at local level.