气候模拟日降水量的统计误差订正分析——以上海为例

ANALYSIS ON STATISTICAL BIAS CORRECTION OF DAILY PRECIPITATION SIMULATED BY REGIONAL CLIMATE MODEL

对区域气候模式系统PRECIS在SRES A1B情景下模拟的上海日降水输出按季节进行了统计误差订正。该方法首先对降水日数进行比率订正,以消除模式产生的微小值降水。然后利用Γ分布拟合日降水量的累计概率分布,采用整体和分段拟合两种方法构建传递函数TF(Transfer Function)进行订正。选取1962年12月—1992年11月作为控制时段,构建TF并将其应用于验证时段(1992年12月—2002年11月)。该订正方案消除了模式产生的微小值降水,解决了模拟的小降水值偏多的问题,频率误差保持在1%以下,分段拟合订正相比整体拟合订正具有更强的对极端降水的订正能力;对冬、春季的订正效果比夏、秋季更显著。该方案不仅有效消除了平均值的漂移,而且显著订正了变率,同时提高了极端降水事件的再现能力,是一种相对完善的订正方案。

A statistical bias correction is applied to the daily precipitation in Shanghai simulated by a regional climate model PRECIS under the SRES-AlB emission scenario. The correction is derived with seasonal distinction. Ratio correction is done to correct the precipitation days and Г distribution is used to fit the cumulative probability distribution of daily precipitation, with ensemble distribution fitting and piecewise distribution fitting applied. The transfer function is derived for the control period （December 1962-November 1992） and applied for the validation period （December 1992-November 2002）. Results showed that the spurious drizzle generated by the PRECIS model is well removed, small precipitation is decreased and the frequency bias turns to below 1 %. Piecewise correction does a better job on extreme precipitation than ensemble correction does. The daily precipitation is better corrected in winter/spring than in summer/autumn. The technique can not only correct the average daily precipitation, but also significantly correct the variability and increase the capability of reproducing extreme precipitation events. So it is a relatively satisfying correction method.