摘要
国家气候中心气候系统模式BCC_CSM1.1参与了第五次耦合模式比较计划(CMIP5)的年代际试验,研究中采用了一种系统误差的订正方法对该模式年代际试验的回报数据进行订正,检验了该方法能否提高模式对热带SST的回报效果。对1960-1990年每5年开展一组的年代际回报试验分析表明,未订正前,年代际试验尽管使用了观测海温资料进行初始化,但对随后海温的实际演变预测能力很低。不同组起报时间的回报试验对逐月海温的预测与对应时段的观测资料,仅在西太平洋及热带北大西洋海域存在一致的正相关关系。经订正后,考虑了模式回报与观测之间误差的统计信息,对全球海温的回报技巧明显提高,尤其是在热带太平洋和南半球印度洋。在热带太平洋海域,订正的模式结果与对应观测的空间相关系数在起报后120个月基本保持在0.8以上,经订正的模式结果对太平洋海温的模态分布更接近观测事实。表明这一误差订正的方法有助于减小模式误差,对预测热带SST有重要的科学参考。
The decadal prediction component of CMIP5 experiment aims at improving our understanding of physical climate system and ability to predict its evolution in the near term. These decadal simulations make it possible and meaningful to assess model skill in forecasting climate under the observed ocean initial conditions. In this study, the decadal hindcast and prediction experiment carried out by BCC_CSM1. 1 climate model for CMIP5 was used to evaluate the model＇ s prediction capability of SST in tropical area. Using a model bias cor- recting method, we corrected the raw model simulation results and examined whether this method can improve the model＇ s hindcast performance in tropical SST. By analyzing the decadal experiment which is conducted ev- ery 5 years from 1960 to 1990, results show as following： in spite of initiation with observation SST was taken into account in the decadal hindcast experiments, the model still underperforms in simulating the realistic SST evolution. The relationship between different groups of the experiments and the observation only show consistent positive correlation in western Pacific and the tropical North Atlantic Ocean area, while hindcast skills of correct- ed model simulation are improved significantly in the global ocean area, particularly in the southern Indian and tropical Pacific Ocean area. In the tropical western Pacific, the space correlation efficient between corrected mod- el simulation and observation is above 0. 8 in all the 120 months after the predicted time. The corrected model simulations can reproduce the mode of the Pacific SST in the observation. Therefore, the bias correcting method in the present study is important and useful for reducing the systematic bias and conducting the climate prediction of tropical SST on time-scales.
出处
《高原气象》
CSCD
北大核心
2016年第5期1364-1375,共12页
Plateau Meteorology
基金
国家重点研发计划项目(2016YFA0602100)
