Applicability assessment of ROSE2.1 short-duration heavy precipitation automatic alarm product in northern Ningxia

doi:10.11755/j.issn.1006-7639-2025-03-0469

Abstract

Abstract:

To enhance the radar monitoring and early warning efficacy for severe convective weather in nowcasting operations, this study systematically analyzed the spatiotemporal characteristics of short-duration heavy precipitation in northern Ningxia from 2018 to 2022, by using Yinchuan C-band radar data and precipitation data from meteorological automatic stations. Based on the ROSE2.1 radar system, the parameter threshold R of the automatic alarm product for short-duration heavy precipitation (UAM product) was revised and optimized. By evaluating key indicators such as the accuracy rate of the UAM product under different parameters, a locally applicable parameter framework was ultimately established. The results show that short-duration heavy precipitation in northern Ningxia occurs most frequently on the eastern foothills of the main peak of the Helan Mountain, followed by the secondary peak. The area with the strongest short-duration heavy precipitation intensity is concentrated around the secondary peak. In Yinchuan City, short-duration heavy precipitation frequently occurs in Xixia District, Helan County, and Lingwu City, while in Shizuishan City, it is more common in Pingluo County and Huinong District, though with relatively lower extremity compared to Yinchuan. Parameter evaluation shows that in scenarios where the radar distance is less than 50 km and the hourly rainfall is greater than or equal to 20 mm and less than 50 mm in Yinchuan City, and where the radar distance is greater than or equal to 50 km and less than or equal to100 km and the hourly rainfall is greater than or equal to 20 mm and less than 50 mm in Shizuishan City, the hit rate of parameter R11 exceeded 70%, with the best comprehensive score. The UAM product optimized with these parameters shows improvements in alarm location, lead time, and overall accuracy rate, but for widespread and intense short-duration heavy precipitation events, it exhibits issues of overestimation of alarm intensity and high false alarm rate, which require further improvement.

Key words: short-duration heavy precipitation, ROSE2.1 system, automatic alarm product, applicability

摘要： 为提升雷达在短临天气预报业务中对强对流天气的监测预警效能，本文利用银川C波段雷达数据及气象自动站降水资料，系统分析了2018—2022年宁夏北部地区短时强降水时空分布特征，并基于ROSE2.1雷达系统，对短时强降水自动报警产品（User Alarm Information，简称“UAM产品”）参数阈值（R）进行修正优化，通过检验评估不同参数下UAM产品准确率等关键指标，最终构建适用于本地的参数体系。结果表明：宁夏北部短时强降水以贺兰山主峰东麓最多，次峰次之，短时强降水强度最强的区域集中在次峰。银川市短时强降水多发于西夏区、贺兰县和灵武市，石嘴山市多发于平罗县、惠农区，但极端性相对银川市较差。参数评估表明，银川市雷达距离<50 km、小时雨量为20~<50 mm及石嘴山市雷达距离50~100 km、小时雨量为20~<50 mm的场景下，R11参数命中率超70%，且综合评分最优。基于该参数优化的UAM产品在报警落区、时间提前量和准确率等方面均有提升，但对于范围较广、强度较强的短时强降水过程，存在报警强度虚高、虚警率偏高等问题，需进一步改进完善。

关键词: 短时强降水, ROSE2.1系统, 自动报警产品, 适用性

CLC Number:

CHENG Yao, ZHAO Teng, YANG Yuanyuan. Applicability assessment of ROSE2.1 short-duration heavy precipitation automatic alarm product in northern Ningxia[J]. Journal of Arid Meteorology, 2025, 43(3): 469-478.

程瑶, 赵腾, 杨苑媛. ROSE2.1短时强降水自动报警产品在宁夏北部的适用性评估[J]. 干旱气象, 2025, 43(3): 469-478.

Figures/Tables 9

Tab.1 Frequency of short-duration heavy precipitation with different levels in administrative regions of Yinchuan and Shizuishan City from 2018 to 2022 单位：次

小时雨量/mm	银川市							石嘴山市
小时雨量/mm	兴庆区	金凤区	西夏区	贺兰县	永宁县	灵武市	总计	大武口区	惠农区	平罗县	总计
20~<50	60	31	122	84	46	68	411	33	72	106	211
50~<70	2	1	9	2	0	4	18	2	0	3	5
≥70	1	0	2	1	0	3	7	0	1	0	1

Fig.1 The spatial distribution of intensity (a, Unit: mm·h-1) and occurrence frequency (b, Unit: times) of short-duration heavy precipitation in northern Ningxia from 2018 to 2022

Fig.2 Evaluation results of UAM products in Yinchuan City based on different combinations of rainfall intensity and radar distance from 2018 to 2022

Fig.3 Evaluation results of UAM products in Shizuishan City based on different combinations of rainfall intensity and radar distance from 2018 to 2022

Tab.2 Recommended threshold of UAM parameters of different rainfall intensity and different radar distance combinations for short-duration heavy precipitation in northern Ningxia

区域	雨强20~<50 mm·h^-1			雨强≥50 mm·h^-1
区域	雷达距离<50 km	雷达距离50~100 km	雷达距离>100 km	雷达距离<50 km	雷达距离50~100 km
银川市	R11*	R10~R20		R20	R10~R20
石嘴山市	R20	R11*	R10~R20		R19~R20
宁夏北部	R11	R11	R10~R20	R20	R20

Tab.3 Evaluation results of UAM products for short-duration heavy precipitation in northern Ningxia in 2023 under different radar range classification

雷达距离分级	参数	累计命中/个	累计漏报/个	累计虚警/个	平均命中率/%	平均虚警率/%	平均漏报率/%	平均临界成功指数
所有范围（50次个例）	R11	40	9	39	42.86	20.00	49.54	0.36
	R15	30	18	20	29.03	51.61	30.28	0.25
	R20	17	27	6	14.53	73.93	19.23	0.15
<50 km （37次个例）	R11	29	8	10	57.14	42.86	20.61	0.51
	R15	22	13	7	31.25	68.75	17.00	0.27
	R20	15	16	1	12.70	87.30	7.14	0.13
50~100 km （13次个例）	R11	10	1	16	50.00	7.14	53.23	0.40
	R15	8	5	6	33.33	41.67	30.56	0.28
	R20	2	12	2	18.18	72.73	18.18	0.18

Tab.4 Assessment of time lead and initial intensity of short-duration heavy precipitation UAM products cases in northern Ningxia in 2023

统计量	时间提前量/min			强度较实况变率/%
统计量	R11	R15	R20	R11	R15	R20
范围	-11~73	-12~68	-8~56	-41~398	0~398	46~501
均值	34	30	27	58.6	84.9	206.5

Fig.4 Spaial distribution of observed short-duration heavy precipitation and corresponding Yinchuan radar products in northern Ningxia on 10 August 2023 (Green-circled areas indicate short-duration heavy precipitation occurrences, the red shaded regions represent system-activated alert zones, the same as below)

Fig.5 Spaial distribution of observed short-duration heavy precipitation and corresponding Yinchuan radar products in northern Ningxia on 14 September 2023

References 34

[1]	曾广宇, 郭泽勇, 周钦强, 等, 2021. 基于激光雨滴谱的双偏振雷达定量降水估测算法应用评估[J]. 气象研究与应用, 42(4): 95-100.
[2]	陈晓娟, 王咏青, 毛璐, 等, 2020. 贺兰山区两次极端暴雨动力作用数值模拟分析[J]. 干旱区研究, 37(3): 680-688.
[3]	程鹏, 罗汉, 陈佩璇, 等, 2020. 庆阳黄土高原短时强降水统计特征及预警指标[J]. 干旱气象, 38(2): 319-328.
[4]	程瑛, 黄武斌, 沙宏娥, 2018. 甘肃岷县两次强降水致山洪泥石流灾害特征对比分析[J]. 干旱区地理, 41(3):443-448.
[5]	樊江伟, 董冰, 2024. 新一代天气雷达在防灾减灾工作中的应用[J]. 气候变化研究快报, 13(4): 780-785.
[6]	郭飞燕, 刁秀广, 褚颖佳, 等, 2023. 两次极端强降水风暴双偏振参量特征对比分析[J]. 干旱气象, 41(1): 103-113. DOI
[7]	郭佳, 吴艳锋, 罗丽, 等, 2020. CINRAD-SA偏振雷达定量降水估测算法改进及应用评估[J]. 气候与环境研究, 25(3): 305-319.
[8]	兰明才, 周莉, 蒋帅, 等, 2022. 西太平洋副热带高压控制下湖南一次短时强降水成因分析[J]. 干旱气象, 40(4): 656-666. DOI
[9]	李悦, 2022. 新一代天气雷达业务软件正式业务运行[N]. 中国气象报, 06-14(1).
[10]	刘倩, 安涛, 刘世玺, 2023. 中国气象局新一代天气雷达业务软件ROSE2.1正式业务运行[N]. 中国气象报, 12- 05(1).
[11]	穆建华, 纪晓玲, 贾乐, 等, 2023. 基于CLDAS的贺兰山区5—9月降水时空分布特征及其与地形的关系分析[J]. 气象科技, 51(6): 824-834.
[12]	潘赫拉, 许东蓓, 陈明轩, 等, 2020. 天气雷达气候学研究新进展[J]. 干旱气象, 38(6): 887-894.
[13]	苏华英, 曾莉萍, 罗乃兴, 等, 2020. 基于雷达资料的定量降水预报外推算法试验[J]. 沙漠与绿洲气象, 14(5): 27-35.
[14]	王俊超, 王志斌, 赖安伟, 等, 2022. 基于数值预报与雷达外推预报融合方法的降水回波短临预报试验研究[J]. 干旱气象, 40(3): 485-499. DOI
[15]	王萍, 窦冰杰, 2018. 基于多普勒雷达数据的强辐合场识别方法[J]. 天津大学学报:自然科学与工程技术版, 51(8): 797-809.
[16]	韦凯华, 张录青, 黄晓莹, 2023. ROSE2.0定量降水估测产品在广东地区的适用性检验[J]. 广东气象, 45(2): 44-47.
[17]	肖贻青, 马永永, 陈小婷, 等, 2023. 陕南汉江盆地一次冷锋触发的短时强降水过程及预报检验[J]. 干旱气象, 41(6): 972-983.
[18]	辛玮琦, 马中元, 谌云, 等, 2021. 宜丰短时强降水雷达回波特征分析[J]. 沙漠与绿洲气象, 15(2): 70-80.
[19]	荀爱萍, 赵玉春, 李菲, 等, 2023. X波段双偏振相控阵雷达定量降水估测研究[J]. 暴雨灾害, 42(4): 437-445.
[20]	杨侃, 纪晓玲, 毛璐, 等, 2020. 贺兰山两次特大致洪暴雨的数值模拟与地形影响对比[J]. 干旱气象, 38(4): 581-590.
[21]	杨涛, 杨莲梅, 张云惠, 等, 2021. 新疆短时强降水天气系统环流配置及雷达回波特征[J]. 干旱气象, 39(4): 631-640.
[22]	袁凯, 庞晶, 李武阶, 等, 2023. 深度学习模型在2021年汛期武汉市雷达回波临近预报中的应用评估[J]. 干旱气象, 41(1): 173-185. DOI
[23]	赵畅, 魏鸣, 刘红亚, 等, 2019. 雷达定量估测降水量中双偏振参量的算法[J]. 科学技术与工程, 19(18): 40-46.
[24]	CHEN H N, LIM S, CHANDRASEKAR V, et al, 2017. Urban hydrological applications of dual-polarization X-band radar: Case study in Korea[J]. Journal of Hydrologic Engineering, 22(5): E5016001. DOI:10.1061/(ASCE)HE. 1943-5584.00014.
[25]	HASSAN D, ISAAC G A, TAYLOR P A, et al, 2022. Optimizing radar-based rainfall estimation using machine learning models[J]. Remote Sensing, 14(20): 5188. DOI: 10.3390/rs14205188.
[26]	GHAEMI E, GABELLA M, FOELSCHE U, et al, 2023. The effect of altitude on the uncertainty of radar-based precipitation estimates over Switzerland[J]. International Journal of Remote Sensing, 44(8):2495-2 517.
[27]	JI C X, XU Y, 2024. trajPredRNN+: A new approach for precipitation nowcasting with weather radar echo images based on deep learning[J]. Heliyon, 10(18): e36134. DOI:10.1016/j.heliyon.2024.e36134.
[28]	NIKAHD A, SHABANI M, 2024. New calibration formula for radar-rainfall relationships analysis[J]. Russian Meteorology and Hydrology, 49: 504-512.
[29]	SLEZIAK P, JANČO M, DANKO M, et al, 2023. Accuracy of radar-estimated precipitation in a mountain catchment in Slovakia[J]. Journal of Hydrology and Hydromechanics, 71(1): 111-122.
[30]	SAMA B, UMA K N, DAS S K, 2024. A comprehensive assessment of the temporal and spatial variation of hail producing convective storms over eastern India using weather radar[J]. Climate Dynamics, 62(6): 4 749-4 773.
[31]	TIAN F Y, XIA K, SUN J H, et al, 2024. Ingredients-based methodology and fuzzy logic combined short-duration heavy rainfall short-range forecasting: An improved scheme[J]. Journal of Tropical Meteorology, 30(3): 241-256..
[32]	WANG C, WANG P, WANG D, et al, 2020. Nowcasting multicell short-term intense precipitation using graph models and random forests[J]. Monthly Weather Review, 148(11): 4 453-4 466.
[33]	XIAO M Y, WANG L, DONG Y C, et al, 2024. An early warning approach for the rapid identification of extreme weather disasters based on phased array dual polarization radar cooperative network data[J]. PLoS One, 19(1): e0296044. DOI:10.1371/journal.pone.0296044.
[34]	ZHAO G X, WANG Y J, LIAN Z L, 2022. Characteristics of radar echo parameters and microphysical structure simulation of a short-time heavy precipitation supercell[J]. Journal of Tropical Meteorology, 28(4): 388-404.