设为首页
加入收藏
An Objective Hailstorm Labeling Algorithm Based on Ground Observation
-
摘要: 标识工作是建立深度学习数据集的关键基础，对观测数据匮乏的冰雹等灾害性天气智能预报尤为重要。选取2008—2019年重庆地区灾情报告中有准确时间的13次降雹过程（分为参照集和验证集），利用模糊逻辑算法，建立基于地面实况的降雹风暴体客观标识方法。为获取冰雹与风暴体的合理匹配，选取风暴质心与降雹地点间距离、风暴最大反射率因子、45 dBZ反射率因子最大高度、最大垂直积分液态水含量、最大回波顶高作为判别因子。对于参照集，客观标识方法正确标识7次，其中有5次标识时间与灾情报告记录时间相差在6 min以内。对于验证集，算法标识正确率为100%。为了扩大检验范围，将算法用于无准确时间的22次降雹过程，并将结果与预报员人工标识结果进行比较后发现，二者往往标识的是同一风暴体。上述结果表明：该方法在时间信息模糊的情况下可进行标识。同时发现该方法不依赖于冰雹尺寸、发生时间及风暴体生命史长度，但对初始猜测位置、风暴体识别算法较为敏感。Abstract: Data labeling makes the key foundation of building data sets for deep learning, especially in the intelligent forecasts of severe weather, such as hail, the observations of which are lacking. Disaster report is a kind of information that describes the details of meteorological disasters which is collected by meteorological information officer. Due to the high coverage rate of informants throughout villages and communities, disaster report is considered to have good consistency and high spatial resolution. However, the vague description of disaster occurrence time in disaster report limits its application. To solve this problem, 13 hail cases(divided into reference set and verification set) with accurate occurrence time in hail reports in Chongqing during 2008-2019 are selected, and an objective hailstorm labeling algorithm based on actual hail observations is developed using fuzzy logic algorithm. In order to obtain a reasonable match between hail occurrence location and convective storm, the distance between the centroid of the storm and initial guess location of hail occurrence, the maximum values of reflectivity, height of 45 dBZ reflectivity, vertical integral liquid water content and echo top are selected as discriminant factors, and the storm is identified by the storm cell identification and tracking (SCIT) algorithm. In reference set, 7 hail cases can be labeled correctly and only 1 case is failed to identify storms. The time bias between the labeling time and ground disaster report is less than 6 minutes during 5 cases. Inspected by verification set (5 cases in 2019), the algorithm labeling accuracy is 100% and the matching degree ranges from 0.887 to 1.000. Furthermore, the algorithm is applied to 22 hailstorm labeling cases lacking accurate time, and the results are compared with the manual labeling results by forecasters. Subjective and objective methods tend to identify the same storm cell and have little impact on data set construction. Forecasters tend to label the same storm cell 6-12 minutes ahead. Further analysis shows that the size of hail has no significant effects on the labeling result. The algorithm is not sensitive to the occurrence time of hail disaster, and it can give reliable labeling results for both long time living storm and local hail disaster. However, when the identification algorithm fails to figure out storms, or the initial guess location deviation is large, it will have a significant negative impact on the labeling results.
-
Key words:
- hail;
- hailstorm labeling;
- fuzzy logic;
- disaster report
-
图 2 降雹风暴体客观标识方法中各指标的隶属函数
Fig. 2 Membership functions of each index of object-labeling algorithm for hailstorm
图 4 各冰雹过程中算法标识时刻的垂直积分液态水含量(填色)、组合反射率因子(实线, 单位：dBZ)及SCIT识别的风暴体路径(三角，红色为标识时刻)
Fig. 4 Vertical integrated liquid water (the shaded), composite reflectivity (the line, unit:dBZ) and storm path from SCIT at labeling time for each hail case (triangles, the red one means labeling time)
图 5 2008年6月5日南川、丰都冰雹过程组合反射率因子演变及风暴路径信息
Fig. 5 Composite reflectivity evolution with storm path information in the hail case at Nanchuan and Fengdu on 5 Jun 2008
图 6 2008年4月10日南川冰雹组合反射率因子演变及风暴路径信息
Fig. 6 Composite reflectivity evolution with storm path information in the hail case at Nanchuan on 10 Apr 2008
图 7 初猜位置扰动对标识结果的影响
Fig. 7 Influence of initial guess position disturbance on labeling results
图 8 2018年4月30日20：12雷达组合反射率因子(a)及其沿图 8a中AB所作剖面(b)
Fig. 8 Composite reflectivity at 2012 BT 30 Apr 2018(a) and profile along AB in Fig. 8a(b)
表 1 2008—2019年重庆地区有相对准确时间和雷达观测数据的冰雹过程
Table 1 Hail cases with exactly time and radar observation in Chongqing during 2008-2019
过程日期 地面记录降雹时刻 受灾区县 冰雹大小 2008-04-10 21:30 南川 小冰雹 2008-06-05 15:20 南川 不明 15:28—15:32 丰都 小冰雹 2010-05-06 01:12 梁平 大冰雹 2011-07-23 17:30 渝北 小冰雹 2014-04-17 约23:48 沙坪坝 大冰雹 2015-07-25 约13:31 丰都 不明 2018-03-17 约22:20 彭水 大冰雹 2019-02-19 22:50—23:00 酉阳 小冰雹 2019-03-19 23:20—23:30 黔江 大冰雹 2019-04-08 18:30—18:40 彭水 不明 2019-04-24 20:30—20:40 巫溪 小冰雹 2019-08-01 16:10—16:30 巫山 小冰雹 
下载: 导出CSV
表 2 2014—2018年3—9月重庆沙坪坝站平均0℃和-20℃层高度
Table 2 Average heights of 0℃ and -20℃ at Shapingba sounding station from Mar to Sep during 2014-2018
月份 0℃层高度/km -20℃层高度/km 3 3.2783 6.6654 4 3.9190 7.2017 5 4.6056 7.9129 6 5.2451 8.6644 7 5.4405 8.8348 8 5.3239 8.7461 9 5.0934 8.5597
下载: 导出CSV
表 3 雹暴标识客观算法在参照集的标识效果
Table 3 Results of objective labeling algorithm for hailstorm based on reference set
过程日期 冰雹大小 冰雹发生时间 输入数据时间范围 客观算法标识时间 匹配程度 2008-04-10 小冰雹 21:30 17:00—次日00:00 21:42 1.000 2008-06-05 不明 15:20 11:20—19:20 15:06 0.899 小冰雹 15:28—15:32 11:20—19:20 2010-05-06 大冰雹 次日01:12 20:00—次日04:00 01:12 0.981 2011-07-23 小冰雹 17:30 13:30—21:30 17:36 1.000 2014-04-17 大冰雹 23:48 21:00—次日05:00 23:48 1.000 2015-07-25 不明 13:31 11:00—16:30 13:30 0.785 2018-03-17 大冰雹 22:20 19:00—次日01:00 22:18 0.805
下载: 导出CSV
表 4 雹暴标识客观算法在验证集的标识效果
Table 4 Performance of objective labeling algorithm for hailstorm based on validation set
过程日期 冰雹大小 冰雹发生时间 输入数据时间范围 客观算法标识时间 匹配程度 2019-02-19 小冰雹 22:50—23:00 22:00—次日00:00 23:00 0.893 2019-03-19 大冰雹 23:20—23:30 22:00—次日02:00 23:30 0.903 2019-04-08 不明 18:30—18:40 17:30—19:30 18:36 1.000 2019-04-24 小冰雹 20:30—20:40 19:00—21:00 20:18 0.984 2019-08-01 小冰雹 16:10—16:30 15:00—18:00 16:18 0.887
下载: 导出CSV
-
[1] Mcgovern A, Elmore K L, Gagne D J, et al.Using artificial intelligence to improve real-time decision making for high-impact weather.Bull Amer Meteor Soc, 2017, 98(10): 2073-2090. doi: 10.1175/BAMS-D-16-0123.1 [2] Gange D J, McGovern A, Haupt S E, et al.Storm-based probabilistic hail forecasting with machine learning applied to convection-allowing ensembles.Wea Forecasting, 2017, 32(5): 1819-1840. doi: 10.1175/WAF-D-17-0010.1 [3] 韩丰, 龙明盛, 李月安, 等.循环神经网络在雷达临近预报中的应用.应用气象学报, 2019, 30(1): 61-69. doi: 10.11898/1001-7313.20190106Han F, Long M S, Li Y A, et al.The application of recurrent neural network to nowcasting.J Appl Meteor Sci, 2019, 30(1): 61-69. doi: 10.11898/1001-7313.20190106 [4] Witt A, Eilts M D, Stumpf G J, et al.An enhanced hail detection algorithm for the WSR-88D.Wea Forecasting, 1998, 13: 286-303. doi: 10.1175/1520-0434(1998)013<0286:AEHDAF>2.0.CO;2 [5] Cintineo J L, Smith T M, Lakshmanan V, et al.An objective high-resolution hail climatology of the contiguous United States.Wea Forecasting, 2012, 27(5): 1235-1248. doi: 10.1175/WAF-D-11-00151.1 [6] 张秉祥, 李国翠, 刘黎平, 等.基于模糊逻辑的冰雹天气雷达识别算法.应用气象学报, 2014, 25(4): 415-426. doi: 10.3969/j.issn.1001-7313.2014.04.004Zhang B X, Li G C, Liu L P, et al.Identification method of hail weather based on fuzzy-logical principle.J Appl Meteor Sci, 2014, 25(4): 415-426. doi: 10.3969/j.issn.1001-7313.2014.04.004 [7] 王瑾, 刘黎平.WSR-88D冰雹探测算法在贵州地区的评估检验.应用气象学报, 2011, 22(1): 96-106. doi: 10.3969/j.issn.1001-7313.2011.01.010Wang J, Liu L P.The evaluation of WSR-88D hail detection algorithm over Guizhou region.J Appl Meteor Sci, 2011, 22(1): 96-106. doi: 10.3969/j.issn.1001-7313.2011.01.010 [8] 肖艳姣, 刘黎平.新一代天气雷达组网资料的三维格点化及拼图方法研究.气象学报, 2006, 64(5): 647-656. doi: 10.3321/j.issn:0577-6619.2006.05.011Xiao Y J, Liu L P.Study of methods for interpolating from weather radar network to 3-D grid and mosaics.Acta Meteor Sinica, 2006, 64(5): 647-656. doi: 10.3321/j.issn:0577-6619.2006.05.011 [9] 胡胜, 顾松山, 庄旭东, 等.风暴的多普勒雷达自动识别.气象学报, 2006, 64(6): 796-808.Hu S, Gu S S, Zhuang X D, et al.Automatic identification of storm cells using Doppler radars.Acta Meteor Sinica, 2006, 64(6): 796-808. [10] 虎雅琼, 边宇轩, 黄梦宇, 等.基于灾情信息的1981-2017年北京地区降雹特征.应用气象学报, 2019, 30(6): 710-721. doi: 10.11898/1001-7313.20190607Hu Y Q, Bian Y X, Huang M Y, et al.Characteristics of hailstone distribution based on disaster information in Beijing from 1981 to 2017.J Appl Meteor Sci, 2019, 30(6): 710-721. doi: 10.11898/1001-7313.20190607 [11] Tian F Y, Zheng Y G, Zhang T, et al.Statistical characteristics of environmental parameters for warm season short-duration heavy rainfall over central and eastern China.J Meteor Res, 2015, 29(3): 370-384. doi: 10.1007/s13351-014-4119-y [12] 曹艳察, 田付友, 郑永光, 等.中国两级阶梯地势区域冰雹天气的环境物理量统计特征.高原气象, 2018, 31(1): 185-196.Cao Y C, Tian F Y, Zheng Y G, et al.Statistical characteristics of environmental parameters for hail over the two-step terrains of China.Plateau Meteor, 2018, 31(1): 185-196. [13] 郑永光, 周康辉, 盛杰, 等.强对流天气监测预报预警技术进展.应用气象学报, 2015, 26(6): 641-657. doi: 10.11898/1001-7313.20150601Zheng Y G, Zhou K H, Sheng J, et al.Advances in techniques of monitoring, forecasting and warning of severe convection weather.J Appl Meteor Sci, 2015, 26(6): 641-657. doi: 10.11898/1001-7313.20150601 [14] Zadeh L A.Fuzzy sets.Information & Control, 1965, 8(3): 338-353. [15] Cho Y H, Lee G, Kim K E, et al.Identification and removal of ground echoes and anomalous propagation using the characteristics of radar echoes.J Atmos Oceanic Tech, 2006, 23(9): 1206-1222. doi: 10.1175/JTECH1913.1 [16] Gourley J J, Tabary P, Jacques P D C.A fuzzy logic algorithm for the separation of precipitating from nonprecipitating echoes using polarimetric radar observations.J Atmos Oceanic Tech, 2006, 24(8): 1439-1451. [17] 陈明轩, 高峰, 孔荣, 等.自动临近预报系统及其在北京奥运期间的应用.应用气象学报, 2010, 21(4): 395-404. doi: 10.3969/j.issn.1001-7313.2010.04.002Chen M X, Gao F, Kong R, et al.Introduction of auto-nowcasting system for convective storm and its performance in Beijing Olympics meteorological service.J Appl Meteor Sci, 2010, 21(4): 395-404. doi: 10.3969/j.issn.1001-7313.2010.04.002 [18] 李丰, 刘黎平, 王红艳, 等.S波段多普勒天气雷达非降水气象回波识别.应用气象学报, 2012, 23(2): 147-158. doi: 10.3969/j.issn.1001-7313.2012.02.003Li F, Liu L P, Wang H Y, et al.Identification of non-precipitation meteorological echoes with Doppler weather radar.J Appl Meteor Sci, 2012, 23(2): 147-158. doi: 10.3969/j.issn.1001-7313.2012.02.003 [19] 周康辉, 郑永光, 王婷波, 等.基于模糊逻辑的雷暴大风和非雷暴大风区分方法.气象, 2017, 43(7): 781-791.Zhou K H, Zheng Y G, Wang T B, et al.Fuzzy logic algorithm of thunderstorm gale identification using multisource data.Meteor Mon, 2017, 43(7): 781-791. [20] 齐晨, 金晨曦, 郭文利, 等.基于模糊逻辑的飞机积冰预测指数.应用气象学报, 2019, 30(5): 619-628. doi: 10.11898/1001-7313.20190510Qi C, Jin C X, Guo W L, et al.Icing potential index of aircraft icing based on fuzzy logic.J Appl Meteor Sci, 2019, 30(5): 619-628. doi: 10.11898/1001-7313.20190510 [21] 徐舒扬, 吴翀, 刘黎平.双偏振雷达水凝物相态识别算法的参数改进.应用气象学报, 2020, 31(3): 350-360. doi: 10.11898/1001-7313.20200309Xu S Y, Wu C, Liu L P.Parameter improvements of hydrometeor classification algorithm for the dual-polarimetric radar.J Appl Meteor Sci, 2020, 31(3): 350-360. doi: 10.11898/1001-7313.20200309 [22] 俞小鼎, 姚秀萍, 熊廷南, 等.多普勒天气雷达原理与业务应用.北京:气象出版社, 2006.Yu X D, Yao X P, Xiong T N, et al.Principle and Operational Application of Doppler Weather Radar.Beijing:China Meteorological Press, 2006. [23] 吴剑坤, 俞小鼎.强冰雹天气的多普勒天气雷达探测与预警技术综述.干旱气象, 2009, 27(3): 197-206. doi: 10.3969/j.issn.1006-7639.2009.03.001Wu J K, Yu X D.Review of detection and warning methods for severe hail events by Doppler weather radars.J Arid Meteor, 2009, 27(3): 197-206. doi: 10.3969/j.issn.1006-7639.2009.03.001 [24] 李云川, 王福侠, 裴宇杰, 等.用CINRAD-SA雷达产品识别冰雹、大风和强降水.气象, 2006, 32(10): 64-69.Li Y C, Wang F X, Pei Y J, et al.Products of CINRAD-SA Doppler radar applied to different typical weather.Meteor Mon, 2006, 32(10): 64-69. [25] 刘德, 张亚萍, 陈贵川, 等.重庆市天气预报技术手册.北京:气象出版社, 2012.Liu D, Zhang Y P, Chen G C, et al.Chongqing Weather Forecast Technical Manual.Beijing:China Meteorological Press, 2012. [26] Orlanski I.A rational subdivision of scales for atmospheric processes.Bull Amer Meteor Soc, 1975, 56: 527-530. doi: 10.1175/1520-0477-56.5.527 [27] 方翀, 俞小鼎, 朱文剑, 等.2013年3月20日湖南和广东雷暴大风过程的特征分析.气象, 2015, 41(11): 1305-1314. doi: 10.7519/j.issn.1000-0526.2015.11.001Fang C, Yu X D, Zhu W J, et al.Characteristics of the thunderstorm gale process in Hunan and Guangdong on 20 March 2013.Meteor Mon, 2015, 41(11): 1305-1314. doi: 10.7519/j.issn.1000-0526.2015.11.001 [28] Darrah R P.On the relationship of severe weather to radar tops.Mon Wea Rev, 1978, 106(9): 1332-1339. doi: 10.1175/1520-0493(1978)106<1332:OTROSW>2.0.CO;2 [29] Smith P L, Myers C G, Orville H D.Radar reflectivity factor calculations in numerical cloud models using bulk parameterization of precipitation.J Appl Meteor, 1975, 14(9): 1156-1165. -
点击查看大图
计量
- 摘要浏览量: 1271
- HTML全文浏览量: 398
- PDF下载量: 118
- 被引次数: 0
