近57年我国气温格点数据集的建立和质量评估

在普通克里金插值方法的基础上, 引入高程因子并充分考虑插值的边界效应, 对1951:2007年我国气温站点资料进行空间结构分析和插值, 得到我国地面气温日、月、年平均值1°×1°格点数据集。数据集的质量评估结果表明:高程在我国区域气温空间结构分析和插值中起着重要作用, 高程资料的引入有效提高了大部分高山地区的插值效果; 相比站点资料, 所建立的格点数据集在描述我国年平均气温以及季节平均气温分布时更为合理, 突出了温度场的大尺度特征; 数据集反映出了我国年平均气温变化趋势主要的空间差异; 数据集较好地反映了我国年平均气温变化状况, 1951- 2007年气温变暖幅度约为1.6 ℃, 增温速率0.28 ℃/10 a, 比全球或半球同期平均增温速率明显偏高, 且气温增暖主要发生在最近的20余年之内。另外, 格点数据集显示, 1998—2007年是1951年以来最暖的10年, 其中2006年全国平均气温距平接近2000年之前的历史最高年份1998年, 而2007年全国平均气温距平值超过1998年, 达到1951年以来的历史最高值1.3 ℃, 为最暖的一年。

Establishment and Assessment of the Grid Air Temperature Data Sets in China for the Past 57 Years

Temperature data from meteorological stations in China for the past 57 years are interpolated by introducing digital elevation model(DEM) and taking the edge effect of interpolations into consideration on the basis of the ordinary Kriging method. Using this method, the daily, monthly and annual mean data sets of the grid-based temperature in China with the resolution of 1°latitude by 1°longitude are obtained. The results of the grid data sets quality assessment show that DEM has an apparent effect on the spatial structure of temperatures and plays an important role on interpolations. Great improvements on the spatial interpolation in alpine spots are achieved after exploiting DEM. Compared with station data, grid data sets are more plausible to depict the annual and seasonal mean temperature distribution. The grid data sets can well simulate the spatial difference of the annual mean temperature trends in China. The grid data sets can also well demonstrate the change of annual mean temperature in China which as a whole rises by about 1.6 ℃ for 1951—2007, with a warming rate of about 0.28 ℃/10 a. The warming in the later half 20th century is more rapid than the average values of the world and the Northern Hemisphere. The most evident warming occurs in the past over 20 years. The grid data sets also indicate that: Period from 1998 to 2007 is a decade when the annual mean temperature is the highest since 1951, amongst which the annual mean temperature anomaly in 2006 is close to that in 1998 with the highest value before 2000, and the annual mean temperature anomaly in 2007 with a value of 1.3 ℃ beyond that in 1998, is the highest since 1951.