气候影响评价中统计降尺度若干问题的探讨

A discussion of some aspects of statistical downscaling in climate impacts assessment

全球气候模式是目前研究未来气候变化的重要工具,然而其较低的空间分辨率使其难以被直接用于区域尺度的气候影响评价中,统计降尺度常常被用于弥补这一不足。对统计降尺度的3种主要方法:转化函数法,天气分型法和天气发生器法的最新研究进展进行了归纳;论述了统计降尺度中的各种不确定性;总结了统计降尺度在中国的发展和应用。统计降尺度与动力降尺度的比较和结合、极端事件的降尺度以及统计降尺度的不确定性将成为未来的主要发展方向。

Global Climate Models （GCMs） are the primary tools for understanding how the global climate might change in the future. However, the relatively low spatial resolution of GCMs outputs is unsatisfactory for the localscale climate impact assessments. Compared to dynamic downscaling, the statistical downscaling approach is widely used to bridge this gap. In this review paper, recent advances in three fundamental statistical downscaling approaches （regression methods, weather type approaches and stochastic weather generators） were presented firstly. Furthermore, uncertainties in statistical downscaling were discussed. The developments and applications of statistical downscaling in China were then summarized. The review study concludes that the comparisons and combinations of statistical downscaling and dynamic downscaling approaches, downscaling of extreme events and uncertainty analysis in statistical downsealing will become the mainstream of future related studies.