摘要
利用AREM、MM5和WRF模式为试验模式,由对短期天气预报结果影响颇大的积云参数化方案和边界层方案构成15个集合预报成员,开展有限区域多模式短期集合预报在我国汛期时段的应用与研究。分别研究了单个模式集合预报和多模式集合预报在2003年汛期(7月)预报中的应用,预报对象主要包括降水、500 hPa位势高度和700 hPa相对湿度。试验结果表明:(1)由AREM、MM5和WRF模式构成的多模式集合对以上要素的集合预报总体效果比其任一单个模式的集合预报效果好;(2)对于降水的集合预报,单个模式的集合平均结果对多模式集合预报效果有影响。且对于不同的降水临界值影响不同;当降水临界值较小时,单模式集合平均结果对多模式集合效果影响较小;当降水临界值较大时,影响较大,甚至可以影响多模式集合的集合平均预报成败;(3)对于降水、500 hPa位势高度和700 hPa相对湿度,其单个模式以及多模式的48 h集合预报对确定性预报的改善度都比24 h的显著。(4)对于形势预报和相对湿度预报,多模式集合预报效果明显比同期T213模式的预报水平高。
An ensemble forecasting system with 15 members consisting of different cumulus parameterization and PBL schemes by changing AREM, MM5 and WRF model configuration is constructed to improve the application and research of multi-model short-range ensemble forecasting in rainy seasons of China. In this paper, the applications of single model and multi-model ensemble forecasting are respectively applied to study the rainy season （July） in 2003, with the objects of forecast mainly being the precipitation, 500 hPa geopotential height and 700 hPa relative humidity. The experiment has the following results. （1） Multi-model ensemble forecasting consisting of AREM, MM5 and WRF has a better presentation of the above elements than any single model ensemble forecasting. （2） For the ensemble forecast of precipitation, the mean results of the single model are better than those of the multi-model. With different accumulated precipitation threshold values, however, the impact is different. When the threshold value is small, the impact is weak whereas when it is large, the impact is strong. （3） For precipitation, 500 hPa geopotential height and 700 hPa relative humidity, the 48h ensemble forecast by either single models or multiple models improves deterministic predictions much more than the 24 h one. （4） For the forecast of weather situations and relative humidity, results produced by multi-model ensemble forecasting are much better than those by T213 for simultaneous periods.
出处
《热带气象学报》
CSCD
北大核心
2009年第4期449-457,共9页
Journal of Tropical Meteorology
基金
国家科学自然基金(No.40405010)资助
关键词
数值模拟
集合预报
多模式集合
江淮汛期
汛期降水
numerical simulation
ensemble-prediction
multi-model ensemble
Yangtze-Huaihe river basin rainy season
rainy season precipitation
