CMIP6全球气候模式对中国气温模拟的BMA方法评估

Evaluation of BMA for temperature simulation in China by CMIP6 global climate models

选用CMIP6中13种全球气候模式数据,以CN05.1数据作为实测资料,对1961—2014年中国气温进行模拟及模式能力评估。采用BMA、泰勒图评估模式排名,并将BMA与算术平均(AVG)集合结果进行比较。结果表明,泰勒图评分和BMA权重在最优和最劣模式评价中基本一致,模拟效果最好的两种模式为ACCESS-ESM1-5、INM-CM5-0。BMA集合模拟结果优于AVG方法,CN05.1、BMA、AVG方法得到的中国多年平均气温分别为6.18、5.95和4.92℃,BMA方法通过权重调节使整体系统误差最小。BMA和AVG方法集合的CMIP6气候模式在对中国气温模拟的空间分布形式上与实测差距不大,而局部地域分布情况有所区别。BMA方法不仅可以对CMIP6模式进行有效评估,并且其集合模拟结果的时间及空间变化情况都与实测值更接近。

The 13 global climate model in CMIP6 were selected,and CN05.1 data was used as the measured data to simulate China s temperature from 1961 to 2014.The model capabilities were ranked using BMA and Taylor graph,and BMA was compared with the Arithmetic Average(AVG)ensemble results.Results show that the Taylor graph score and BMA weight are basically consistent in the evaluation of the best and worst models,and the two models with the best simulation effect are ACCESS-ESM1-5 and INM-CM5-0.The BMA ensemble results are better than that of AVG.The annual average temperatures in China by CN05.1,BMA and AVG methods are 6.18,5.95 and 4.92℃,respectively.The BMA method minimizes the overall systematic error through weight adjustment.The CMIP6 climate model ensemble of BMA and AVG has little difference between the spatial distribution of China s temperature simulation and the actual measurement,but the regional distributions are different.The BMA method can not only effectively evaluate the CMIP6 models,but also the temporal and spatial changes of the ensemble results are closer to the measured values.