基于TIGGE资料的西太平洋热带气旋多模式集成预报方法比较

Comparison of Multimodel Ensemble Methods for Western Pacific Tropical Cyclone Forecast Based on TIGGE Data

基于TIGGE资料中CMA、ECMWF、JMA和NCEP四中心2010、2011和2012年3年的资料,采用集合平均(EMN)、加权集合平均(WEMN)、消除偏差集合平均(BREM)和加权消除偏差集合平均(即超级集合,SUP)四种方法,对西太平洋地区热带气旋路径与中心气压进行时效为24、48、72、96和120 h的多模式集成预报,评估了不同单中心预报结果,并分析了不同多模式集成预报方法的特点,对比了不同多模式集成方法的效果。结果表明,对于热带气旋路径和中心气压预报,各中心预报技巧不同,其中3年的CMA预报效果均较差,2010、2011年的ECMWF预报和2012年的NCEP预报效果最优;总体上几种多模式集成方法在120 h预报时效内均优于单模式预报,其中EMN作为一种最简单的集成预报方法,具有一定的局限性,而WEMN由于给不同单模式预报赋予了权重,因此相对于EMN能够得到更好的多模式集成预报结果;BREM方案由于消除了模式预报中的系统性偏差,因此集成预报效果也优于EMN;由于去除了模式预报偏差,同时给不同模式赋予了权重,SUP方案得到的集成预报效果最优。

In this paper, four multimodel ensemble methods, namely ensemble mean （EMN）, weighted ensemble mean （WEMN）, bias-removed ensemble mean （BREM） and weighted bias-removed ensemble mean （superensemble, SUP） have been constructed based on TIGGE datasets, from which four single model forecasts of recent three years from CMA, ECMWF, JMA and NCEP are selected to conduct multi- model ensemble combination and comparison. Single and multimodel ensemble forecasts on the tracks and central pressure of tropical cyclone over Western Pacific with forecast lead time 24 h, 48 h, 72 h, 96 h and 120 h are evaluated and compared. The results show that different single models have different forecast skills. The CMA forecast has the worst performance during the three years, and the best single model forecast in 2010 and 2011 is ECMWF while NCEP is the best one in 2012. Generally, multimodels have significant superiority over the single models within 120 h forecast lead times. As the simplest method within all the four multimodels, EMN shows limited superiority. WEMN has better skill than EMN be- cause it can give different weights to the member models, while the BREM exhibits superiority for the sys- tematic bias is removed. Because of considering both the weights of member models and bias elimination, the SUP has a better skill in general.