GRAPES＿Meso V3.0模式预报效果检验

王雨; 李莉

GRAPES＿Meso V3.0模式预报效果检验

王雨,
李莉

中国气象局数值预报中心, 北京 100081

资助项目: 中国气象局“GRAPES数值预报系统"发展专项资助

计量
- 摘要浏览量: 3948
- HTML全文浏览量: 974
- PDF下载量: 1154
- 被引次数: 0
出版历程
- 收稿日期: 2010-02-10
- 修回日期: 2010-07-05
- 刊出日期: 2010-10-31

Verification of GRAPES_Meso V3.0 Model Forecast Results

Wang Yu,
Li Li.

Numerical Weather Prediction Center, CMA, Beijing 100081

摘要

摘要: 应用国家气象中心模式检验方法对GRAPES＿Meso V3.0模式2008年2月—2009年3月的试验预报产品，如降水、中低层高度、温度和风场预报进行统计检验。检验结果表明:V3.0模式降水预报性能得到明显改善，年及四季平均的各级降水TS评分显著提高，除了秋、冬季的48 h中雨和暴雨预报外，TS评分明显高于V2.5模式，但V3.0模式的预报偏差偏大，中雨以上偏大更明显。从预报的季节平均降水率分布来看，对秋、冬季我国东部24 h降水预报偏小改进明显，对春、夏季强降水中心位置及强度预报也好于V2.5模式，但是48 h降水预报明显偏大，逐日降水率演变图也印证了这一点。此外，V3.0模式对500 hPa高度和风场及48 h预报的850 hPa风场和温度场改进显著，对于850 hPa高度和温度的24 h预报，除夏季外，其他季节预报效果优于业务模式。
- GRAPES＿Meso V3.0模式;
- 预报;
- 效果检验
Abstract: The rainfall, temperature, height and wind forecast products of GRAPES_Meso model (V3.0), which is mesoscale model of the new generation multiple time scale numerical weather prediction system, are verified by operational statistical verification methods from March 2008 to February 2009. Dichotomous and continuous forecast verification methods for deterministic forecast model recommended by World Meteorological Organization or Chinese Meteorological Administration are employed. The verification results show the rainfall forecast of V3.0 model is better than V2.5 model. The annual mean TS of rainfall forecast is improved for all five grades and is higher than that of V2.5 model clearly for four seasons mean, all months mean of a year, except for 48-hour moderate rain and storm rain of autumn and winter. The forecasted rainfall of V3.0 model is mainly greater than observed rainfall, especially for moderate to heavy rain. The simulated distribution of the season mean raining rate is much more similar to the observed patterns. For the 24-hour forecast of eastern China in autumn and winter, rainfall is almost no longer underestimated. The forecast performance of location and intensity of strong rainfall center for spring and summer has also been improved, but the amount of 48-hour rainfall is obviously overestimated comparing to observation, shown clearly through the figures of daily mean rain rate time series. Meanwhile, the developing trend and intensity forecast of strong rainfall processes is better than that of V2.5 model for most time of the year. The height, temperature and wind forecasts of the upper prognostic forecast have made remarkable progresses especially for 500 hPa height forecast, 500 hPa wind forecast, and 48 hour 850 hPa wind forecast. However, the forecast of 24-hour low level temperature and height in summer are hardly improved, so the model still need more progresses especially in rainy season. For other seasons, the forecast of temperature and wind in low level atmosphere is fairly good.
- GRAPES_Meso V3.0 model;
- forecast;
- verification results

HTML全文

图 1 2008年3月—2009年2月GRAPES_MesoV2.5与V3.0模式降水预报全国加密累加检验TS评分

Fig. 1 The threat score of rainfall forecast on GRAPES_MesoV2.5 and V3.0 models for Chinese dense observation from March 2008 to February 2009

下载: 全尺寸图片幻灯片

图 2 2008年3月—2009年2月GRAPES_MesoV2.5与V3.0模式降水预报全国加密累加检验预报偏差

Fig. 2 The Bias of rainfall forecast on GRAPES_Meso V2.5 and V3.0 models for Chinese dense observation from March 2008 to February 2009

下载: 全尺寸图片幻灯片

图 3 2008年3月—2009年2月GRAPES_Meso V2.5和V3.0模式与其他模式及预报员降水预报全国加密累加检验

Fig. 3 The mean verification results of rainfall forecast on GRAPES_Meso V2.5 and V3.0, forecaster and other models for Chinese dense observation from March 2008 to February 2009

下载: 全尺寸图片幻灯片

图 4 2008—2009年不同季节降水率实况与GRAPES_Meso V3.0 24h和48 h预报平均降水率预报分布 (单位:mm·d^-1)

Fig. 4 Distributions of observational and 24-hour, 48-hour forecast rainfall rate of GRAPES_Meso V3.0 for Chinese dense observation in spring, summer, autumn and winter of 2008—2009 (unit:mm·d^-1)

下载: 全尺寸图片幻灯片

图 5 2008—2009年不同季节GRAPES_Meso V2.5与V3.0模式24 h预报全国平均降水率逐日变化

Fig. 5 The time variation of observational and 24-hour forecast rainfall rate of GRAPES_Meso V2.5 and V3.0 models for Chinese dense observation from March 2008 to February 2009

下载: 全尺寸图片幻灯片

图 6 利用观测资料对2008年4月—2009年2月GRAPES_Meso V2.5及V3.0模式850 hPa高度场24 h和48 h预报检验

Fig. 6 The monthly mean 24-hour and 48-hour forecast verification results of 850 hPa height on GRAPES_Meso V2.5 and V3.0 models for China against observation from April 2008 to February 2009

下载: 全尺寸图片幻灯片

图 7 利用观测资料对2008年4月—2009年2月GRAPES_Meso V2.5及V3.0模式24 h和48h 500 hPa高度场预报的检验

Fig. 7 The monthly mean 24-hour and 48-hour forecast verification results of 500 hPa height on GRAPES_Meso V2.5 and V3.0 models for China against observation from April 2008 to February 2009

下载: 全尺寸图片幻灯片

图 8 利用观测资料对2008年4月—2009年2月GRAPES_Meso V2.5及V3.0模式850 hPa温度场24h和48h预报检验

Fig. 8 The monthly mean 24-hour and 48-hour forecast verification results of 850 hPa temperature on GRAPES_Meso V2.5 and V3.0 models for China against observation from April 2008t oFebruary 2009

下载: 全尺寸图片幻灯片

图 9 利用观测资料对2008年4月—2009年2月GRAPES V2.5及3.0模式850 hPa风场24 h和48 h预报检验

Fig. 9 The monthly mean 24-hour and 48-hour forecast verification results of 850 hPa wind on GRAPES_Meso V2.5 and V3.0 models for China against observation from April 2008 to February 2009

下载: 全尺寸图片幻灯片

表 1 GRAPES_Meso V 2.5与V3.0模式主要差别

Table 1 The differences between GRAPES_Meso V2.5 and V3.0 models

参考文献(15)

[1]	张人禾,沈学顺.中国国家级新一代业务数值预报系统GRAPES的发展.科学通报,2008,53(20):2393-2395. http://www.cnki.com.cn/Article/CJFDTOTAL-KXTB200820001.htm
[2]	陈德辉,沈学顺.新一带数值预报系统GRPAES研究进展.应用气象学报,2006,17(6):773-777. http://www.cnki.com.cn/Article/CJFDTOTAL-YYQX200606013.htm
[3]	陈德辉,杨学胜,胡江林,等.多尺度通用动力模式框架的设计策略.应用气象学报,2003,14(4):452-461. http://www.cnki.com.cn/Article/CJFDTOTAL-YYQX200304008.htm
[4]	陈德辉,胡志晋.CAMS大气数值预报模式系统研究.北京:气象出版社,2004.
[5]	陈德辉,薛纪善,杨学胜,等.GRAPES新一代全球/区域多尺度统一数值预报模式总体设计研究.科学通报,2008,53(20):2396-2407. http://www.cnki.com.cn/Article/CJFDTOTAL-KXTB200820002.htm
[6]	叶成志,欧阳里程,李象玉,等.GRAPES中尺度模式对2005年长江流域重大灾害性天气过程预报性能的检验分析.热带气象学报,2006,26(4):393-399. http://www.cnki.com.cn/Article/CJFDTOTAL-RDQX200604011.htm
[7]	徐双柱,张兵,谌伟.GRAPES模式对长江流域天气预报的检验分析.气象,2007,33(11):65-71. http://www.cnki.com.cn/Article/CJFDTOTAL-QXXX200711011.htm
[8]	朱振铎,端义宏,陈德辉.GRAPES-TCM业务试验结果分析.气象.2007,33(7):44-54. http://www.cnki.com.cn/Article/CJFDTOTAL-QXXX200707006.htm
[9]	李勇,王雨.2007年夏季GRAPESMES015及30 km模式对比检验.气象,2008,34(10):81-89.
[10]	尤凤春,魏东,王雨.北京奥运期间多模式降水检验及集成试验.气象,2009,35(11):3-8. http://www.cnki.com.cn/Article/CJFDTOTAL-QXXX200911002.htm
[11]	徐国强,陈德辉,薛纪善,等.GRAPES物理过程的优化试验及程序结构设计.科学通报,2008,53(20):2428-2434. http://www.cnki.com.cn/Article/CJFDTOTAL-KXTB200820006.htm
[12]	彭新东,李兴良.多尺度大气数值预报的技术进展.应用气象学报,2010,21(2):129-138. http://qikan.camscma.cn/jams/ch/reader/view_abstract.aspx?file_no=20100201&flag=1
[13]	王雨,闫之辉.降水检验方案变化对降水检验评估效果的影响分析.气象,2007,33(12):53-62. http://www.cnki.com.cn/Article/CJFDTOTAL-QXXX200712009.htm
[14]	王雨.2004年主汛期各数值预报模式定量降水预报评估.应用气象学报,2006,17(3):316-323. http://qikan.camscma.cn/jams/ch/reader/view_abstract.aspx?file_no=20060357&flag=1
[15]	刘环珠,张绍晴.第四讲中期数值预报的统计检验分析.气象,1992,18(9):50-54. http://www.cnki.com.cn/Article/CJFDTOTAL-QXXX199209012.htm