MM5与MM5-PAFOG模式区域雾预报效果评估比较  

Validation and Comparison of Regional Fog Forecast by MM5 and MM5-PAFOG Models

作　　者：石春娥[1] 吴照宪 邓学良[1] 张浩[1] 张苏[1] 

机构地区：[1]安徽省气象科学研究所/安徽省大气科学与卫星遥感重点实验室,安徽合肥230031 [2]安徽省池州市气象局,安徽池州247000

出　　处：《高原气象》2013年第5期1349-1359,共11页Plateau Meteorology

基　　金：国家自然科学基金项目(40775010);中国气象局新技术推广项目(CMATG2010M16);科技基础性工作专项重点项目(2007FY110700)

摘　　要：利用MM5和PAFOG模式对2010年10-11月中国东部地区的雾进行逐日预报,包括基于MM5模式结果的多要素诊断预报和PAFOG模式预报能见度的定量预报,应用TS/ETS评分方法和区域内所有测站每3h(或6h)一次的常规观测资料,根据该两方法的预报结果分别对中国东部地区和安徽进行了效果评估。结果表明:(1)两种预报方法对区域雾都有一定的预报能力,但都存在空报率较高的明显不足。基于MM5模式的多要素诊断预报,得到中国东部地区的TS/ETS评分为0.075/0.065,命中率(HR)为0.296;安徽的TS/ETS评分为0.094/0.081,HR为0.378。(2)PAFOG模式对中国东部地区的TS/ETS评分为0.038/0.027,HR为0.388;安徽的TS/ETS评分为0.053/0.038,HR为0.52。与MM5模式的诊断结果相比,TS/ETS评分较低,HR较高。(3)按日统计结果表明,MM5模式在整个中国东部地区的预报效果优于PAFOG模式,但在内陆省份(安徽),PAFOG模式的预报效果略优于MM5模式。(4)效果评估时若考虑过去1h内的天气现象记录,两种方法的预报效果都有明显改善。MM5 and PAFOG models were run to forecast daily regional fogs in eastern China from October to November of 2010. PAFOG model was run at each observational station with the initial and boundary conditions （IC/BC） from MM5 model. Fog was diagnosed based on the multi-variables of MM5 model out- put and forecasted by the visibility of PAFOG model. The results were evaluated against 3 h （or 6 h） rou- tine observations at all stations of China Meteorology Agency （CMA） observational net works by calculat- ing a set of scores for eastern China and Anhui Province, respectively. The evaluated results show that： （1） Both methods have a certain capability of forecasting the regional fog, with the common shortcoming of high false alarm rate. The values of Ts/ETs from based on MM5 model are 0. 075/0. 065 for eastern Chi- na and 0. 094：/0. 081 for Anhui Province, and the hit rate is 0. 296 for eastern China and 0. 378 for Anhui Province. （2）The Ts/Ers values from based on PAFOG model for eastern China and Anhui Province is 0. 038/0. 027 and 0. 053/0. 038, respectively, lower than those of MM5 model; however, the hit rates is 0. 387 for eastern China and 0. 517 for Anhui, higher than those of MM5 model. （3） Based on the daily statistic, MM5 model outperforms PAFOG model in eastern China, while PAFOG model outperforms MM5 model slightly in the inland province （Anhui Province） ; and PAFOG model tends to over-predict fog occurrence more than MM5 model. （4） If the weather records of the past 1 h were used for model evalua- tions, the validation results for both methods would be improved.

关 键 词：区域雾 诊断预报 MM5模式 PAFOG模式 效果评估 

分 类 号：P456.7[天文地球—大气科学及气象学]