设为首页
加入收藏
Application of a Bias Correction Method to Meteorological Forecast for the Pyeongchang Winter Olympic Games
-
摘要: 为了提高GRAPES_3 km（Global/Regional Assimilation and Prediction System）模式在2018年平昌冬奥会气象服务中的预报能力，采用一阶自适应的卡尔曼滤波方法对GRAPES_3 km模式的2 m气温、2 m相对湿度和10 m风开展偏差订正。结果表明：偏差订正方法明显提高了地面要素的预报效果，其中2 m气温的均方根误差整体减小到2℃左右，站点订正改善率为10%~60%；10 m风速的均方根误差减小到2 m·s-1左右，站点订正改善率为10%~45%；2 m相对湿度减小到20%以下，站点订正改善率为0~20%。与韩国气象厅LDAPS（Local Data Assimilation and Prediction System）及美国宇航局NU-WRF（NASA-Unified WRF）模式相比，GRAPES_3 km模式的风速预报表现更为优异，各站点整体预报效果明显优于LDAPS和NU-WRF模式。偏差订正方法可有效改善模式在复杂地形条件下的预报能力，是提高精细化预报准确率的重要手段。
-
关键词:
- GRAPES_3 km模式;
- 平昌冬奥会;
- ICE-POP 2018;
- 偏差订正
Abstract: The 23rd Winter Olympics Games and the 13th Winter Paralympic Games are held in Pyeongchang, South Korea during 9-25 February 2018 and 8-18 March 2018. Supported by the WMO(World Meteorological Organization) WWRP(World Weather Research Project Group), ICE-POP 2018(International Collaborative Experiments for Pyeongchang 2018 Olympic and Paralympic winter games) is organized by KMA (Korea Meteorological Administration), which aims to improve the forecasting ability of convective scale numerical models for high-impact weather system under complex terrain and to support weather forecasting and meteorological services during the Winter Olympics through international cooperation. GRAPES_3 km, a high-resolution model independently developed by CMA (China Meteorological Administration), participated in this project and provided real-time forecasting products at the specific sites during the Winter Olympics Games. In order to improve the forecasting ability of GRAPES_3 km, a bias correction method named one-order adaptive Kalman filtering is applied, which compensate for the fact that the resolution of GRAPES_3 km is not sufficient to simulate the complex terrain of the Winter Olympic Games and GRAPES_3 km doesn't assimilate Korea's observations due to some objective reasons such as data transmission. 1-24-hour calibration products are provided twice a day, including 2 m temperature, 2 m relative humidity, 10 m wind speed, and 10 m wind direction on sixteen sites. The verification and evaluation are carried out in two aspects. First, it is to check whether the bias correction improves the accuracy of GRAPES_3 km, examined from the model's diurnal cycle, daily variation and the forecast ability of complex terrain. Second, to the performance of calibrated GRAPES_3 km during the Winter Olympic Games is and examined compared to LDAPS model from KMA and NU-WRF from NASA, which shows that the high-resolution GRAPES_3 km model has abilities to simulate the near-ground elements in this service and the bias correction technology makes model products better and effective. After bias correction, the root mean square error of model products is reduced to about 2℃ for the temperature, about 2 m·s-1 for the wind speed, and less than 20% for the relative humidity is reduced to. Compared with, NU-WRF and LDAPS models, the corrected GRAPES_3 km has the best wind speed forecast capability and root mean square error of the wind speed is significantly reduced. For the wind direction forecast of complex terrain, although all numerical models have limited capabilities, and their wind direction forecasting of all the stations are almost dominated by westerly during the Winter Games, the corrected GRAPES_3 km perform relatively well in some non-westerly dominated stations. However, GRAPES_3 km has no significant advantage in temperature and humidity compared with the other two models although they are improved more obviously than wind after the calibration. In addition, the bias correction can attenuate the diurnal variation characteristics of GRPAES_3 km, and to some extent, improve the simulation capability for complex terrain. In short, the application of the bias correction in the Pyeongchang Winter Olympic Games is an effective way for GRAPES_3 km. -
图 2 2018年2月9日—3月18日平均的误差订正前后均方根误差(折线)和改善率(柱状)日变化特征 (a)2 m气温, (b)10 m风速, (c)2 m相对湿度
Fig. 2 Diurnal variation of root mean square error (the line) before and after calibration with improvement rate (the bar) averaged from 9 Feb to 18 Mar in 2018 (a)2 m temperature, (b)10 m wind speed, (c)2 m relative humidity
图 3 2018年2月9日—3月18日24个预报时效平均的误差订正前后均方根误差(折线)和改善率(柱状)逐日变化特征 (a)2 m气温, (b)10 m风速, (c)2 m相对湿度
Fig. 3 Daily variation of root mean square error (the line) before and after calibration with improvement rate (the bar) of 1-24 h forecast mean from 9 Feb to 18 Mar in 2018 (a)2 m temperature, (b)10 m wind speed, (c)2 m relative humidity
图 4 2018年2月9日—3月18日平均的误差订正前后均方根误差(折线)和改善率(柱状)变化特征 (a)2 m气温, (b)10 m风速, (c)2 m相对湿度
Fig. 4 Variation of root mean square error (the line) before and after calibration and improvement rate (the bar) averaged from 9 Feb to 18 Mar in 2018 (a)2 m temperature, (b)10 m wind speed, (c)2 m relative humidity
图 5 不同高度站点订正前后的均方根误差(折线)的对比(柱状, 颜色表示观测站点的所在的场馆) (a)2 m气温, (b)10 m风速, (c)2 m相对湿度
Fig. 5 Comparison of root mean square error (the line) before and after calibration at stations with different height (the bar, color of bar denotes venues) (a)2 m temperature, (b)10 m wind speed, (c)2 m relative humidity
图 6 2018年2月12日订正前后2 m气温(折线)及不同站点高度(柱状、颜色表示不同场馆)对比 (a)6 h时效预报, (b)12 h时效预报, (c)24 h时效预报
Fig. 6 Comparison of 2 m temperature (the line) before and after calibration with altitudes of different sites (the bar, the color denotes venue) on 12 Feb 2018 (a)6 h forecast, (b)12 h forecast, (c)24 h forecast
图 7 不同模式预报的2 m气温均方根误差的对比 (a)日变化, (b)逐日变化, (c)逐站点变化
Fig. 7 Comparison of 2 m temperature's root mean square error of different models (a)diurnal variation, (b)daily variation, (c)station variation
图 8 不同模式预报的2 m相对湿度均方根误差对比 (a)日变化, (b)逐日变化, (c)逐站点变化
Fig. 8 Comparison of 2 m relative humidity's root mean square error of different models (a)diurnal variation, (b)daily variation, (c)station variation
图 9 不同模式预报的10 m风速均方根误差的对比 (a)日变化, (b)逐日变化, (c)逐站点变化
Fig. 9 Comparison of 10 m wind speed's root mean square error of different models (a)diurnal variation, (b)daily variation, (c)station variation
图 10 2018年2月9日—3月18日2561站10 m风玫瑰图 (a)观测, (b)GRAPES_3 km, (c)NU-WRF, (d)LDAPS
Fig. 10 Wind rose diagram of Station 2561 from 9 Feb to 18 Mar in 2018 (a)observation, (b)GRAPES_3 km, (c)NU-WRF, (d)LDAPS
表 1 站点信息表
Table 1 Information of sites
场馆位置 站号 地形高度/m 站点描述 阿尔卑西亚 2575 785.0 跳台滑雪 2557 760.0 冬季两项赛道起点 2577 764.0 越野滑雪赛道起点 2554 812.0 滑行运动赛道终点 龙坪 2560 1416.0 高山滑雪大回转赛道起点 2579 1180.0 高山滑雪大回转赛道中点 2561 975.0 高山滑雪大回转赛道终点 旌善 2584 1370.0 高山滑雪速降赛道 2586 919.0 高山滑雪速降赛道中点 2587 639.0 高山滑雪速降赛道终点 宝光 2580 856.0 越野竞速赛道起点 2581 664.0 越野竞速赛道终点 2588 874.0 坡道障碍赛道起点 2583 709.0 坡道障碍赛道终点 大关岭 47100 772.6 大关岭站 江陵 47105 26.0 江陵站 
下载: 导出CSV
表 2 模式对比
Table 2 Comparison of models
模式配置 GRAPES_3 km LDAPS NU-WRF 水平分辨率 3 km×3 km 1.5 km×1.5 km 1 km×1 km 更新频率 每日2次 每日4次 每日4次 预报时效 0~24 h 0~36 h 0~24 h 初始场和侧边界条件 T639模式预报场驱动 GDAPS(UM 17 km)预报场驱动 NCEP/EMC GFS预报场驱动 辐射方案 RRTM LW&Dudhia SW 通用二流方案[26] NASA/GSFC 微物理方案 WSM-6 混合相降水方案[27] NASA/GSFC 4ICE 边界层方案 MRF 一阶非局地方案[28] MYJ 同化方案 三维云分析(GRAPES-GCAS) 三维变分同化(FGAT, IAU) Goddard卫星数据模拟器单元, 利用了土地信息系统LSM和同化框架
下载: 导出CSV
-
[1] Joe P.Weather services, science advances, and the Vancouver 2010 Olympic and Paralympic Winter Games.Bull Amer Meteor Soc, 2010, 91(1):31-36. doi: 10.1175/2009BAMS2998.1 [2] Kiktev D, Joe P, Isaac G A, et al.FROST-2014: The Sochi Winter Olympics International Project.Bull Amer Meteor Soc, 2017, 98:1908-1929. doi: 10.1175/BAMS-D-15-00307.1 [3] 陈超君, 王东海, 李国平, 等.冬季高海拔复杂地形下GRAPES-MESO要素预报的检验评估.气象, 2012, 38(6):657-668. http://d.old.wanfangdata.com.cn/Conference/8188248 [4] 黄丽萍, 陈德辉, 邓莲堂, 等.GRAPES_Meso V4.0主要技术改进和预报效果检验.应用气象学报, 2017, 28(1):25-37. doi: 10.11898/1001-7313.20170103 [5] 朱立娟, 龚建东, 黄丽萍, 等.GRAPES三维云初始场形成及在短临预报中的应用.应用气象学报, 2017, 28(1):38-51. doi: 10.11898/1001-7313.20170104 [6] 郝民, 龚建东, 田伟红, 等.L波段探空仪湿度资料偏差订正及同化试验.应用气象学报, 2018, 29(5):559-570. doi: 10.11898/1001-7313.20180505 [7] 唐文苑, 郑永光, 张小雯.基于FSS的高分辨率模式华北对流预报能力评估.应用气象学报, 2018, 29(5):513-523. doi: 10.11898/1001-7313.20180501 [8] 许晨璐, 王建捷, 黄丽萍.千米尺度分辨率下GRAPES-Meso 4.0模式定量降水预报性能评估.气象学报, 2017, 75(6):851-876. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=qxxb201706001 [9] Cui B, Toth Z, Zhu Y, et al.Bias correction for global ensemble forecast.Wea Forecasting, 2012, 27:396-410. doi: 10.1175/WAF-D-11-00011.1 [10] Toth Z, Zhu Y and Marchok T.The use of ensembles to identify forecasts with small and large uncertainty.Wea Forecasting, 16(4): 463-477. [11] Persson A.Kalman Filtering-A New Approach to Adaptive Statistical Interpretation of Numerical Forecasts.WMO Training Workshop on the Interpretation of NWP Products, WMO, Wageningen, the Netherlands, 1991: XX-27-XX-32. [12] Persson A, Strauss B.On the skill and consistency in mediumrange weather forecasts.ECMWF Newsletter, 1995, 70:12-15. [13] Klein W H, Glahn H R.Forecasting local weather by means of model output statistics.Bull Amer Meteor Soc, 1974, 55:1217-1227. doi: 10.1175/1520-0477(1974)055<1217:FLWBMO>2.0.CO;2 [14] 曾晓青, 赵声蓉, 段云霞.基于MOS方法的风向预测方案对比研究.气象与环境学报, 2013, 29(6):140-144. http://d.old.wanfangdata.com.cn/Periodical/lnqx201306022 [15] 吴启树, 韩美, 郭弘, 等.MOS温度预报中最优训练期方案.应用气象学报, 2016, 27(4):426-434. doi: 10.11898/1001-7313.20160405 [16] 刘还珠, 赵声蓉, 陆志善, 等.国家气象中心气象要素的客观预报-MOS系统.应用气象学报, 2004, 15(2):181-191. doi: 10.3969/j.issn.1001-7313.2004.02.006 [17] 车钦, 赵声蓉, 范广洲.华北地区极端温度MOS预报的季节划分.应用气象学报, 2011, 22(4):429-436. doi: 10.3969/j.issn.1001-7313.2011.04.005 [18] 陈豫英, 刘还珠, 陈楠, 等.基于聚类天气分型的KNN方法在风预报中的应用.应用气象学报, 2008, 19(5):564-572. doi: 10.3969/j.issn.1001-7313.2008.05.007 [19] 王焕毅, 谭政华, 杨萌, 等.三种数值模式气温预报产品的检验及误差订正方法研究.气象与环境学报, 2018, 34(1):22-29. doi: 10.3969/j.issn.1673-503X.2018.01.003 [20] 佟华, 郭品文, 朱跃建, 等.基于大尺度模式产品的误差订正与统计降尺度气象要素预报技术.气象, 2014, 40(1):66-75. doi: 10.3969/j.issn.1007-6190.2014.01.015 [21] 李莉, 李应林, 田华, 等.T213全球集合预报系统性误差订正研究.气象, 2011, 37(1):31-38. http://d.old.wanfangdata.com.cn/Periodical/qx201101004 [22] 邱学兴, 王东勇, 陈宝峰.T639模式预报系统误差统计和订正方法研究.气象, 2012, 38(5):526-532. doi: 10.3969/2012jms.0058 [23] Wang J, Chen J, Du J, et al. Sensitivity of ensemble forecast verification to model bias.Mon Wea Rev, 2018, 146:781-796. doi: 10.1175/MWR-D-17-0223.1 [24] Prasanna V.High-resolution wind simulation over Incheon International Airport with the Unified Model's Rose Nesting Suite from KMA Operational forecasts.Asia-Pac J Atmos Sci, 2018, 54(2):187-203. doi: 10.1007/s13143-018-0003-5 [25] Peters-Lidard C D, Kemp E M, Matsui T, et al. Integrated modeling of aerosol, cloud, precipitation and land processes at satellite-resolved scales.Environ Modell Softw, 2015, 67:149-159. doi: 10.1016/j.envsoft.2015.01.007 [26] Edwards J M, Slingo A.Studies with a flexible new radiation code.I:Choosing a configuration for a large-scale model.Quart J Roy Meteor Soc, 1996, 122:689-719. doi: 10.1002/qj.49712253107 [27] Wilson D R, Ballard S P.A microphysically based precipitation scheme for the UK meteorological office unified model.Quart J Roy Meteor Soc, 1999, 125:1607-1636. doi: 10.1002/qj.49712555707 [28] Lock A P.A new boundary layer mixing scheme.Part Ⅰ:Scheme description and single-column model tests.Mon Wea Rev, 2000, 128:3187-3199. doi: 10.1175/1520-0493(2000)128<3187:ANBLMS>2.0.CO;2 [29] Kalman R F.A new approach to linear filtering and prediction problems.Trans ASME, J Basic Eng(Series D), 1960, 82:35-45. doi: 10.1115/1.3662552 -
计量
- 摘要浏览量: 5950
- HTML全文浏览量: 2820
- PDF下载量: 214
- 被引次数: 0
