Suitability assessment of CMA multi-source precipitation analysis products for short-term heavy rainfall monitoring in Shaanxi

doi:10.11755/j.issn.1006-7639(2024)-02-0263

Abstract

Abstract:

Based on the precipitation data from Shaanxi Province from 2016 to 2021 and real-time products of two-source and three-source precipitation from the CMA Multi-source Precipitation Analysis System (CMPAS-V2.1) during 2018-2021, the spatial and temporal characteristics of short-term heavy rainfall in Shaanxi Province over the past 6 years were analyzed. Furthermore, statistical tests were applied to evaluate the accuracy of the multi-source precipitation products, with the aim of providing reference for correction of multi-source precipitation products in short-term heavy rainfall processes. The results are as follows:(1) In Shaanxi Province, the short-term heavy rainfall frequency peak was at 19:00 pm, with heavy precipitation primarily occurring from 16:00 to 02:00 in a day and predominantly in the months of June to August. The diurnal extreme value of heavy rainfall shows relatively higher from 17:00 to 01:00 and from 04:00 to 07:00 in the morning. Short-term heavy rainfall is more frequent in southern Shaanxi compared to Guanzhong and northern Shaanxi. The regions with relatively larger extreme values include northern Shaanxi, eastern Guanzhong, and central-eastern of southern Shaanxi. (2) Both two types of precipitation products tend to underestimate precipitation compared to observed values. The mean absolute errors of the three-source product exhibit smaller in southern of northern Shaanxi, most part of Guanzhong and southern of southern Shaanxi, while the two-source product performs relatively better in other region.The mean absolute error increases with precipitation increase. For heavy precipitation ranging from 20 to 50 mm, three-source product is recommended, while two-source product is more suitable for heavy precipitation above 50 mm. (3) The diurnal variations in mean absolute errors for both two types of multi-source products are relatively larger from 13:00 to 19:00, 23:00 to 01:00 and from 04:00 to 06:00, and relatively smaller from 08:00 to 12:00, 20:00 to 22:00 and from 02:00 to 03:00. The three-source product outperforms two-source product from May to August, while two-source product performs better from September to October. (4) The accuracy rate of multi-source precipitation products increases as the threshold of short-term heavy rainfall decrease. Both mean absolute error and accuracy rate indicate that the three-source product outperforms two-source product. Multi-source precipitation products exhibit higher accuracy from September to October compared to May to August.

Key words: multi-source precipitation product, short-term heavy rainfall, assessment, Shaanxi

摘要：

使用2016—2021年陕西省降水数据以及2018—2021年中国区域融合降水分析系统（CMA Multi-source Precipitation Analysis System，CMPAS-V2.1）的二源和三源融合降水分析实时产品，分析陕西省近6 a短时强降水的时空特征，并通过统计检验方法评估融合降水产品的准确性，为融合产品在短时强降水过程中的订正提供参考依据。结果表明：（1）陕西省短时强降水的最高频次出现在19：00（北京时，下同），强降水主要集中在16：00至次日02：00，且更多发生在6—8月。强降水极值在17：00至次日01：00和04：00—07：00相对较大。陕南地区短时强降水频次明显高于关中和陕北地区；陕北北部、关中东部以及陕南中东部的强降水极值相对更大。（2）两类降水产品的降水量均偏小。三源产品的平均绝对误差在陕北南部、关中大部和陕南南部地区较小，而其余地区二源产品较优。平均绝对误差随着降水量的增加而增大，对于20~50 mm的短时强降水，三源产品更为可靠，而对于50 mm以上的强降水，二源产品表现更佳。（3）两类产品平均绝对误差在13：00—19：00、23：00至次日01：00和04：00—06：00较大，在08：00—12：00、20：00—22：00和02：00—03：00较小，5—8月三源产品表现优于二源产品，而在9—10月二源产品较为可信。（4）降水量阈值降低时融合产品的准确性增加。三源融合产品在性能上优于二源；9—10月两类融合产品的降水量均优于5—8月。

关键词: 融合产品, 短时强降水, 评估, 陕西

CLC Number:

P426

LIU Juju, ZHAO Qiang, JING Yu, ZHANG Weiran, DAI Changming. Suitability assessment of CMA multi-source precipitation analysis products for short-term heavy rainfall monitoring in Shaanxi[J]. Journal of Arid Meteorology, 2024, 42(2): 263-273.

刘菊菊, 赵强, 井宇, 张蔚然, 戴昌明. CMPAS融合产品在陕西短时强降水监测中的适用性评估[J]. 干旱气象, 2024, 42(2): 263-273.

Figures/Tables 14

Fig.1 The diurnal variation of accumulated frequency (a) and extreme value (b) of short-term heavy rainfall in Shaanxi from 2016 to 2021

Fig.2 Monthly (a) and annual (b) variation of short-term heavy rainfall frequency in Shaanxi from 2016 to 2021

Fig.3 Spatial distribution of the frequency (a, Unit: times) and extreme value (b, Unit: mm) of short-time heavy rainfall in Shaanxi from 2016 to 2021

Fig.4 Mean absolute error distribution of two-source (a) and three-source (b) precipitation products during short-term heavy rainfall processes in Shaanxi from 2018 to 2021 (Unit: mm)

Fig. 5 The spatial distribution of mean error of two-source (a) and three-source (b) precipitation products (Unit: mm) and absolute value of two-source mean error minus absolute value of three-source mean error (c) during short-term heavy rainfall in Shaanxi from 2018 to 2021

Fig.6 Variation of two-source (a) and three-source (b) precipitation products mean absolute error with precipitation intensity and the mean absolute errors corresponding to the two types of products for different precipitation intervals (c) during short-term heavy rainfall processes in Shaanxi from 2018 to 2021

Fig.7 Diurnal variations of multi-source precipitation products mean absolute error (a) and mean error (b) during short-term heavy rainfall processes in Shaanxi from 2018 to 2021

Fig.8 Monthly variations of multi-source precipitation products mean absolute error (a) and mean error (b) during short-time heavy rainfall processes in Shaanxi from 2018 to 2021

Tab.1 Average annual accuracy rate of multi-source precipitation products under different short-term heavy rainfall thresholds

产品	降水阈值/mm	准确率/%
产品	降水阈值/mm	年平均	2018年	2019年	2020年	2021年
二源	20	35	18	26	42	50
	15	60	37	62	68	72
	10	84	62	88	90	94
三源	20	36	20	32	45	47
	15	63	43	60	78	72
	10	86	74	80	93	94

Tab.2 Monthly grade accuracy rate of multi-source precipitation products under different short-term heavy rainfall thresholds

产品	降水阈值/mm	准确率/%
产品	降水阈值/mm	5月	6月	7月	8月	9月	10月
二源	20	33	26	29	38	56	50
	15	44	53	56	61	78	90
	10	78	77	84	84	93	100
三源	20	44	30	32	38	44	60
	15	56	59	58	64	81	90
	10	78	82	83	88	89	100

Fig.9 Diurnal variation of two-source (a) and three-source (b) precipitation products accuracy rate under different short-term heavy rainfall thresholds in Shaanxi from 2018 to 2021

Fig.10 Box line plot of multi-source precipitation products mean absolute error during three short-term heavy rainfall processes

Fig.11 Spatial distribution of two-source product (a, d, g), three-source product (b, e, h) and observed precipitation (c, f, i) in Shaanxi during three short-term heavy rainfall processes at 16:00 on August 21, 2018 (a, b, c), 00:00 on August 21, 2021 (d, e, f) and 12:00 on October 3, 2021 (g, h, i) (Unit: mm)

Fig.12 The topographic height of Guanzhong (a, Unit: m; the black dots indicate the stations where the error values of both types of multi-source precipitations are less than -5 during short-term heavy rainfall processes, the triangle indicates the location of the radar station in Xi'an) and the detection range diagram of 0.5° elevation of the C-band radar station in Xi'an combined with the topographic height map (b, the black indicates the part obscured by the terrain)

References 24

[1]	黄勤, 龙亚星, 李亚丽, 2021. 多源融合降水产品在陕西的适用性评估[J]. 气象水文海洋仪器, 38(4): 108-111.
[2]	孔玉寿, 章东华, 2000. 现代天气预报技术[M]. 北京: 气象出版社.
[3]	李超, 唐千红, 陈宇, 等, 2017. 多源数据融合系统LAPS的研究进展及其在实况数据服务中的应用[J]. 气象科技进展, 7(2): 32-38.
[4]	李显风, 周自江, 李志鹏, 等, 2017. 基于江西省水文资料对中国融合降水产品的质量评估[J]. 气象, 43(12): 1 534-1 546.
[5]	李超, 李华宏, 杨素雨, 等, 2021. 国家级多源融合降水产品在云南的适用性评估[J]. 高原山地气象研究, 41(3): 108-114.
[6]	刘菊菊, 陈小婷, 肖贻青, 等, 2022. 降水融合格点产品在陕西2019年暴雨过程中的检验[J]. 陕西气象(2): 1-9.
[7]	刘晓阳, 杨洪平, 李建通, 等, 2010. 新一代天气雷达定量降水估测集成系统[J]. 气象, 36(4): 90-95.
[8]	李萍云, 赵强, 王楠, 等, 2019. 2005—2018年陕西短时强降水时空分布特征[J]. 陕西气象, (5): 34-39.
[9]	潘旸, 谷军霞, 徐宾, 等, 2018a. 多源降水数据融合研究及应用进展[J]. 气象科技进展, 8(1): 143-152.
[10]	潘旸, 沈艳, 宇婧婧, 等, 2012. 基于最优插值方法分析的中国区域地面观测与卫星反演逐时降水融合试验[J]. 气象学报, 70(6): 1 381-1 389.
[11]	潘旸, 谷军霞, 宇婧婧, 等, 2018b. 中国区域高分辨率多源降水观测产品的融合方法试验[J]. 气象学报, 76(5): 755-766.
[12]	任芝花, 赵平, 张强, 等, 2010. 适用于全国自动站小时降水资料的质量控制方法[J]. 气象, 36(7): 123-132.
[13]	宋雯雯, 龙柯吉, 黄晓龙, 等, 2022. 融合格点降水产品在四川盆地西部一次极端暴雨过程中的评估分析[J]. 中低纬山地气象, 46(3): 9-15.
[14]	孙靖, 程光光, 黄小玉, 2021. 中国地面气象要素格点融合业务产品检验[J]. 高原气象, 40(1): 178-188. DOI
[15]	沈晓燕, 申燕玲, 权晨, 等, 2022. 不同方法对青海2020年强降水模式产品预报性能的检验对比[J]. 干旱气象, 40(2): 333-343. DOI
[16]	吴薇, 黄晓龙, 徐晓莉, 等, 2021. 融合降水实况分析产品在四川地区的适用性评估[J]. 沙漠与绿洲气象, 15(4): 1-8.
[17]	吴薇, 黄晓龙, 徐晓莉, 等, 2023. 四川省降水实况分析产品影响因素综合评估[J]. 干旱气象, 41(1): 143-151. DOI
[18]	许冠宇, 李琳琳, 田刚, 等, 2020. 国家级降水融合产品在长江流域的适用性评估[J]. 暴雨灾害, 39(4): 400-408.
[19]	俞剑蔚, 李聪, 蔡凝昊, 等, 2019. 国家级格点实况分析产品在江苏地区的适用性评估分析[J]. 气象, 45(9): 1 288-1 298.
[20]	宇婧婧, 沈艳, 潘旸, 等, 2013. 概率密度匹配法对中国区域卫星降水资料的改进[J]. 应用气象学报, 24(5): 544-553.
[21]	朱乾根, 林锦瑞, 寿绍文, 等, 2000. 天气学原理和方法: 第4版[M]. 北京: 气象出版社: 341-342.
[22]	赵强, 王建鹏, 王楠, 等, 2017. 2012年夏季秦巴山区暴雨过程的地形作用诊断[J]. 气象科技, 45(1): 139-147.
[23]	JACOBS C M J, MOORS E J, TER MAAT H W, et al, 2008. Evaluation of European Land Data Assimilation System (ELDAS) products using in situ observations[J]. Tellus Series A: Dynamic Meteorology & Oceanography, 60(5): 1 023-1 037.
[24]	JOYCE R J, JANOWIAK J E, ARKIN P A, et al, 2004. CMORPH: A method that produces global precipitation estimates from passive microwave and infrared data at high spatial and temporal resolution[J]. Journal of Hydrometeorology, 5(3): 487-503.