A PRELIMINARY STUDY ON THE MEDIUM-RANGE FORECAST BUSTS OF GRAPES_GFS IN EAST ASIA
-
摘要: 利用2016年4月1日—2018年12月31日GRAPES_GFS模式的业务预报数据,将GRAPES_GFS模式在东亚地区144小时预报500 hPa高度场的距平相关系数小于0.4,均方根误差大于60 gpm的事件定义为模式在东亚地区的中期预报技巧极端下降事件,据此标准筛选出29个个例进行了研究。结果表明,GRAPES_GFS在东亚地区中期预报技巧极端下降事件的高发期主要在春秋季,春季和秋季分别占比31.03%、41.38%,预报技巧极端下降事件通常表现在对于东亚大槽、西伯利亚槽脊、副热带高压的预报失败,四个季节合成的模式偏差均与该季节影响东亚地区主要天气系统的预报偏差大有关系。进一步通过典型个例的研究表明,东亚地区中期预报技巧极端下降事件的误差来源在不同季节表现出不同特征。冬夏季的预报偏差来源于高纬极区,多与模式在极区存在较大预报误差关联;春秋季的预报偏差主要来源于上游地区,与模式在东亚上游预报误差向下游的传播有关,未能合理预报台风活动也是预报偏差来源之一。
-
关键词:
- 中期预报技巧极端下降 /
- GRAPES_GFS /
- 距平相关系数 /
- 东亚预报误差
Abstract: The"forecast bust"of GRAPES_GFS is investigated over the East Asian area by using the operational forecast data of GRAPES_GFS model during the period from April 1, 2016 to December 31, 2018. The"forecast bust"is defined as the cases when the area-averaged root mean square error (RMSE) of the day-6 forecast of geopotential height at 500hPa exceeds 60 gpm, as well as the its anomaly correlation being less than 0.4. According to this definition, the 29 cases have been identified. Of which, the high incidence of GRAPES_GFS medium-range forecast busts in East Asia is mainly in spring and autumn, with spring and autumn accounting for 31.03% and 41.38%, respectively. The medium-range forecast busts of GRAPES_GFS model in East Asia is usually manifested in the failure of the forecast of the East Asian trough, the Siberian trough, the Siberian ridge and the subtropical high. The composite of forecast errors in the four seasons are closely related to the forecast errors of the main weather systems that affect East Asia during that season. Further case studies have shown that the sources of errors in the forecast busts in East Asia exhibit different characteristics in different seasons. The forecast errors in winter and summer come from high latitude polar region, which are mostly related to the large forecast errors of the model in the polar region; the forecast errors in spring and autumn mainly come from the upstream areas, which are related to the propagation of the model's upstream forecast errors to the downstream. In addition, the failure of the model to properly forecast typhoon activity is also one of the sources of forecast errors.-
Key words:
- medium-range forecast bust /
- GRAPES_GFS /
- ACC /
- forecast errors in East Asia
-
图 4 同图 3,但为2017年4月13日0000 UTC起报
图 6 同图 3,但为2017年7月8日0000 UTC起报
图 7 同图 3,但为2017年10月24日0000 UTC起报
图 8 同图 3,但为2018年10月25日0000 UTC起报
红色台风符号表示实际台风所在的位置。
表 1 GRAPES_GFS模式在东亚地区的中期预报技巧极端下降现象
起报日期(0000 UTC) ACC RMSE/gpm 初始时刻相关的高影响天气 144 h预报相关的高影响天气 2016/5/14 0.361 2 86.721 1 / / 2016/7/9 0.266 7 61.392 4 台风“尼伯特” / 2017/1/19 0.145 4 74.145 0 / / 2017/2/28 0.282 6 76.083 5 / / 2017/3/2 0.122 6 80.417 6 / / 2017/3/14 0.356 9 96.820 5 / / 2017/4/13 0.286 0 92.857 9 / / 2017/4/20 0.307 9 72.335 5 / 热带风暴“梅花” 2017/7/8 0.328 6 61.238 5 / / 2017/8/1 -0.086 1 67.310 7 强台风“奥鹿” 台风“奥鹿” 2017/8/4 0.094 5 63.942 2 台风“奥鹿”、热带风暴“尼格” / 2017/9/2 0.027 2 65.323 3 台风“珊瑚”、热带风暴“玛娃” / 2017/9/12 0.114 1 70.503 9 台风“泰利”、热带低压“杜苏芮” 强热带风暴“泰利” 2017/9/16 0.189 5 61.901 9 台风“泰利”、热带低压“杜苏芮” / 2017/9/22 0.232 7 70.457 3 / / 2017/10/11 0.261 6 74.102 1 / 热带风暴“兰恩” 2017/10/20 0.225 0 62.535 1 台风“兰恩” 热带风暴“苏拉” 2017/10/21 0.221 0 66.197 0 超强台风“兰恩” 热带风暴“苏拉” 2017/10/23 0.309 8 62.155 4 强热带风暴“兰恩” 强热带风暴“苏拉” 2017/10/24 0.257 7 62.251 4 / / 2017/10/29 -0.090 3 89.998 8 强热带风暴“苏拉” 台风“达维” 2017/12/21 0.282 5 85.432 7 热带风暴“启德”、热带风暴“天秤” / 2018/3/7 0.225 6 94.071 7 / / 2018/5/2 0.308 2 89.573 5 / / 2018/5/4 0.369 8 89.386 6 / / 2018/5/29 0.260 9 67.239 5 / / 2018/6/3 0.096 3 82.783 8 热带低压“艾云尼” 热带风暴“马力斯” 2018/10/25 -0.178 7 86.935 3 超强台风“玉兔” 强热带风暴“玉兔” 2018/11/20 0.201 7 83.058 3 / 强热带风暴“万宜” -
[1] RODWELL M J, MAGNUSSON L, BAUER P et al. Characteristics of occasional poor medium-range weather forecasts for Europe[J]. Bull Amer Meteor Soc, 2013, 94(9): 1 393-1 405. [2] MAGNUSSON L. Diagnostic methods for understanding the origin of forecast errors[J]. Quart J Roy Meteor Soc, 2017, 143(706): 2 129-2 142. [3] FERRANTI L, CORTI S, JANOUSEK M. Flow-dependent verification of the ECMWF ensemble over the Euro-Atlantic sector[J]. Quart J Roy Meteor Soc, 2015, 141(688): 916-924. [4] LILLO S P, PARSONS D B. Investigating the dynamics of error growth in ECMWF medium range forecast busts[J]. Quart J Roy Meteor Soc, 2017, 143(704): 1 211-1 226. [5] GRAZZINI F, ISAKSEN L. North America Increments[R]. UK: ECMWF Operation Department, 2002. [6] NAMIAS J. Interactions of circulation and weather between hemispheres[J]. Mon Wea Rev, 2009, 91(10): 482-486. [7] HARR P A, DEA J M. Downstream development associated with the extratropical transition of tropical cyclones over the Western North Pacific[J]. Mon Wea Rev, 2009, 137(4): 1 295-1 319. [8] STENSRUD D J, ANDERSON J L. Is Midlatitude Convection an Active or a Passive Player in Producing Global Circulation Patterns? [J]. J Climate, 2001, 14(10): 2 222-2 237. [9] STENSRUD D J. Effects of persistent, midlatitude mesoscale regions of convection on the large-scale environment during the warm season[J]. J Atmos Sci, 1996, 53(23): 3 503-3 527. [10] 沈学顺, 苏勇, 胡江林, 等. GRAPES_GFS全球中期预报系统的研发和业务化[J]. 应用气象学报, 2017, 28(1): 1-10. [11] 沈学顺, 王建捷, 李泽椿, 等. 中国数值天气预报的自主创新发展[J]. 气象学报, 2020, 78(3): 451-476. [12] NITTA T. Convective activities in the tropical Western Pacific and their impact on the northern hemisphere summer circulation[J]. Journal of the Meteorological Society of Japan. Ser. Ⅱ, 1987, 65(3): 373-390. [13] HUANG R H, LI W J. Influence of the heat source anomaly over the tropical western Pacific on the subtropical high over East Asia [C]// Proceedings of International Conference on the General Circulation of East Asia. Chengdu, 10-15 April 1987, 40-51. [14] XU P Q, WANG L, CHEN W, et al. Structural changes in the pacific-japan pattern in the late 1990s[J]. J Climate, 2019, 63(3): 607-621.