设为首页
加入收藏
Filling in the Dual Polarization Radar Echo Occlusion Based on Deep Learning
-
摘要: 广州S波段双偏振天气雷达低仰角多方位存在遮挡，高仰角也存在部分遮挡。基于卷积神经网络等深度学习方法，构建垂直填补（vertical echo-filling，VEF）和水平填补（horizontal echo-filling，HEF）网络架构，基于两种架构，利用无遮挡区的反射率因子ZH、差分反射率ZDR，差传播相移率KDP构建训练集，填补遮挡区的ZH和ZDR。针对仅0.5°仰角存在遮挡的区域，基于VEF架构，利用上层多个仰角、径向、距离库的三维数据，分距离段训练垂直填补模型。针对遮挡仰角较高的区域，则基于HEF架构，利用同一仰角左右相邻的多个径向、距离库的数据，分遮挡径向训练水平填补模型。根据解释方差、平均绝对偏差和相关系数3个指标和3个个例，对模型效果进行评估。结果表明：ZH填补模型的解释方差最大为0.92，平均绝对偏差最小为1.69 dB，相关系数最高为0.96；ZDR填补模型的解释方差最大为0.92，平均绝对偏差最小为0.12 dB，相关系数最高为0.96。利用该研究构建的深度学习填补架构，可有效填补偏振雷达遮挡区域回波，提高雷达数据质量。Abstract: Radar beam blockage is an important error source that affects the quality of weather radar data. The S-band dual-polarization radar in Guangzhou has multi-azimuth occlusion at low elevation and is partially occluded at high elevation. Based on deep learning methods such as convolutional neural network, two echo filling networks, i.e., VEF(vertical echo-filling) and HEF(horizontal echo-filling) are constructed. Based on this architecture, echoes from the unblocked area are used to construct training datasets and fill the reflectivity ZH and differential reflectivity ZDR in the occlusion area. For the area with only 0.5° elevation occlusion, multi-modal modeling is carried out based on VEF architecture by using 3D data from multiple upper elevations, radial directions and gates. Considering that the radar beam broadens with distance and to avoid the influence of the melting layer, the radar beam is divided into four sections according to the oblique distance of 0.5° elevation, and the vertical echo-filling model is trained respectively. For the area with high occlusion elevation, multi-mode modeling is carried out based on HEF architecture using the data of multiple adjacent radial directions and gates with the same elevation. According to the number of occlusion radial, two types of horizontal echo-filling models, three radials echo-filling model and five radials echo-filling model are constructed respectively. Finally, the models are evaluated by three cases and three indicators:Explained variance, mean absolute error and correlation coefficient. The maximum explained variance of ZH vertical echo-filling model is 0.91, the minimum mean absolute error is 1.72 dB, and the maximum correlation coefficient is 0.96. The maximum explained variance of ZDR vertical echo-filling model is 0.87, the minimum mean absolute error is 0.12 dB, and the maximum correlation coefficient is 0.92. The maximum explained variance of ZH horizontal fill model is 0.92, the minimum mean absolute error is 1.69 dB, and the maximum correlation coefficient is 0.96. The maximum explained variance of ZDR horizontal echo-filling model is 0.92, the minimum mean absolute error is 0.12 dB, and the maximum correlation coefficient is 0.96. The deep learning echo-filling model can be used to correct the echoes of Guangzhou S-band dual-polarization radar occlusion area effectively, and the quality of weather radar data is improved.
-
图 1 通过地理高程数据计算的广州CINRAD/SAD雷达0.5°仰角(a)和1.5°仰角(b)遮挡系数
(相邻距离圈相距30 km)
Fig. 1 Occlusion coefficient of 0.5° elevation(a) and 1.5° elevation(b) of Guangzhou CINRAD/SAD calculated from geographic elevation data
(the distance between adjacent rings is 30 km)
图 3 ZH和ZDR垂直填补模型不同距离训练模型的估算值和真实值散点图
(色标为高斯核密度估计值)
Fig. 3 Plots of predicted and observed echo reflectivity of each section with ZH and ZDR vertical echo-filling models
(color mark denotes Gaussian kernel density estimation)
图 4 ZH和ZDR水平填补模型不同填补径向数的模型估算值和真实值散点图
(色标为高斯核密度估计值)
Fig. 4 Plots of predicted and observed echo reflectivity of different fill radial numbers with ZH and ZDR horizontal echo-filling models
(color mark denotes Gaussian kernel density estimation)
图 5 2019年8月25日11：36模型填补效果对比0.5°仰角PPI图
(红色椭圆内为明显遮挡区域，下同)
(a)ZH真实值，(b)ZH估算值，(c)ZDR真实值，(d)ZDR估算值Fig. 5 PPI images of echo-filling result comparison of 0.5° elevation at 1136 UTC 25 Aug 2019
(red ellipse denotes the obvious blocked area, the same as in after)
(a)observed ZH, (b)predicted ZH, (c)observed ZDR, (d)predicted ZDR
图 6 2020年6月6日08：54模型填补效果对比0.5°仰角PPI图
(a)ZH真实值，(b)ZH估算值，(c)ZDR真实值，(d)ZDR估算值
Fig. 6 PPI images of echo-filling result comparison of 0.5° elevation at 0854 UTC 6 Jun 2020
(a)observed ZH, (b)predicted ZH, (c)observed ZDR, (d)predicted ZDR
图 7 2019年5月8日03：54模型填补效果对比0.5°仰角PPI图
(a)ZH真实值，(b)ZH估算值，(c)ZDR真实值，(d)ZDR估算值
Fig. 7 PPI images of echo-filling result comparison of 0.5° elevation at 0354 UTC 8 May 2019
(a)observed ZH, (b)predicted ZH, (c)observed ZDR, (d)predicted ZDR
表 1 ZH和ZDR垂直填补模型不同距离段评估结果
Table 1 Evaluation of each section in ZH and ZDR vertical echo-filling models
填补量 距离段/km 解释方差 平均绝对偏差/dB 相关系数 ZH [20, 54) 0.9116 1.7947 0.9572 [54, 68) 0.9155 1.7251 0.9634 [68, 88) 0.8956 1.7210 0.9526 [88, 120] 0.8206 2.2630 0.9208 ZDR [20, 54) 0.8762 0.1520 0.9240 [54, 68) 0.8564 0.1204 0.8971 [68, 88) 0.8538 0.1228 0.9083 [88, 120] 0.8308 0.1440 0.8683 
下载: 导出CSV
表 2 ZH和ZDR水平填补模型不同填补径向数的评估结果
Table 2 Evaluation of different fill radial numbers in ZH and ZDR horizontal echo-filling models
填补量 遮挡径向数 解释方差 平均绝对偏差/dB 相关系数 ZH 3 0.9228 1.6973 0.9615 5 0.8342 2.1143 0.9227 ZDR 3 0.9254 0.1189 0.9639 5 0.8542 0.1275 0.9333
下载: 导出CSV
表 3 2019年8月25日11：36无遮挡区回波填补不同距离段评估结果
Table 3 Evaluation of echo-filling of each section in the unblocked area at 1136 UTC 25 Aug 2019
填补量 距离段/km 解释方差 平均绝对偏差/dB 相关系数 ZH [20, 54) 0.9031 1.0245 0.9567 [54, 68) 0.9094 1.0267 0.9632 [68, 88) 0.8250 2.0311 0.9102 [88, 120] 0.7851 2.1401 0.8816 ZDR [20, 54) 0.9015 0.0315 0.9211 [54, 68) 0.9242 0.0309 0.9012 [68, 88) 0.8240 0.0353 0.8743 [88, 120] 0.7789 0.0328 0.8674
下载: 导出CSV
表 4 2020年6月6日08：54无遮挡区回波填补不同距离段的评估结果
Table 4 Evaluation of echo-filling of each section in the unblocked area at 0854 UTC 6 Jun 2020
填补量 距离段/km 解释方差 平均绝对偏差/dB 相关系数 ZH [20, 54) 0.9311 1.4654 0.9457 [54, 68) 0.9594 2.0431 0.9341 [68, 88) 0.8950 2.1143 0.9019 [88, 120] 0.8854 2.3496 0.8942 ZDR [20, 54) 0.9592 0.0312 0.9732 [54, 68) 0.9321 0.0207 0.9643 [68, 88) 0.9034 0.0236 0.9325 [88, 120] 0.8811 0.0413 0.9078
下载: 导出CSV
表 5 2019年5月8日03：54无遮挡区回波填补不同距离段的评估结果
Table 5 Evaluation of echo-filling of each section in the unblocked area at 0354 UTC 8 May 2019
填补量 距离段/km 解释方差 平均绝对偏差/dB 相关系数 ZH [20, 54) 0.9330 1.0309 0.9681 [54, 68) 0.9121 1.0331 0.9616 [68, 88) 0.9345 2.0154 0.9473 [88, 120] 0.8708 2.3480 0.9126 ZDR [20, 54) 0.9865 0.0333 0.9618 [54, 68) 0.9618 0.0243 0.9679 [68, 88) 0.9097 0.0251 0.9402 [88, 120] 0.8602 0.0381 0.9231
下载: 导出CSV
-
[1] 黄兴友, 马雷, 杨敏, 等.利用空间相关性进行天气雷达波束阻挡数据的识别和订正.气象科学, 2019, 39(4):532-539. https://www.cnki.com.cn/Article/CJFDTOTAL-QXKX201904011.htmHuang X Y, Ma L, Yang M, et al. Recognition and correction of beam blockage of weather radar based on spatial correlation. J Meteor Sci, 2019, 39(4): 532-539. https://www.cnki.com.cn/Article/CJFDTOTAL-QXKX201904011.htm [2] 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 [3] Shakti P C, Maki M, Shimizu S, et al. Correction of reflectivity in the presence of partial beam blockage over a mountainous region using X-band dual polarization radar. J Hydrometeor, 2013, 14(3): 744-764. doi: 10.1175/JHM-D-12-077.1 [4] Germann U, Galli G, Boscacci M, et al. Radar precipitation measurement in a mountainous region. Quart J Roy Meteor Soc, 2006, 132(618): 1669-1692. doi: 10.1256/qj.05.190 [5] 吴翠红, 万玉发, 吴涛, 等. 雷达回波垂直廓线及其生成方法. 应用气象学报, 2006, 17(2): 232-239. doi: 10.3969/j.issn.1001-7313.2006.02.014Wu C H, Wan Y F, Wu T, et al. Vertical profile of radar echo and its determination methods. J Appl Meteor Sci, 2006, 17(2): 232-239. doi: 10.3969/j.issn.1001-7313.2006.02.014 [6] Tabary P. The new French operational radar rainfall product. Part Ⅰ: Methodology. Wea Forecasting, 2007, 22(3): 393-408. doi: 10.1175/WAF1004.1 [7] Battan L J. Radar Observation of the Atmosphere. Chicago: The University of Chicago Press, 1973. [8] 罗丽, 井高飞, 郭佳, 等. 北京气象局天气雷达回波阻挡订正技术研究. 科学技术与工程, 2016, 16(12): 12-19. doi: 10.3969/j.issn.1671-1815.2016.12.003Luo L, Jing G F, Guo J, et al. Study on weather radar echo block correction technology of Beijing Meteorological Bureau. Sci Technol Eng, 2016, 16(12): 12-19. doi: 10.3969/j.issn.1671-1815.2016.12.003 [9] Lee J, Jung S H, Kim H L, et al. Improved rainfall estimation based on corrected radar reflectivity in partial beam blockage area of S-band dual-polarization radar. Korean Meteor Soc, 2017, 27(4): 467-481. [10] 杨洪平, 张沛源, 程明虎, 等. 多普勒天气雷达组网拼图有效数据区域分析. 应用气象学报, 2009, 20(1): 47-55. doi: 10.3969/j.issn.1001-7313.2009.01.006Yang H P, Zhang P Y, Cheng M H, et al. The valid mosaic data region of the CINRAD network. J Appl Meteor Sci, 2009, 20(1): 47-55. doi: 10.3969/j.issn.1001-7313.2009.01.006 [11] Andrieu H, Creutin J D, Delrieu G, et al. Use of a weather radar for the hydrology of a mountainous area. Part Ⅰ: Radar measurement interpretation. J Hydrol, 1997, 193: 1-25. doi: 10.1016/S0022-1694(96)03202-7 [12] Creutin J D, Andrieu H, Faure D. Use of a weather radar for the hydrology of a mountainous area. Part Ⅱ: Radar measurement validation. J Hydrol, 1997, 193: 26-44. doi: 10.1016/S0022-1694(96)03203-9 [13] 杨泷, 刘黎平, 王红艳. 利用反射率因子垂直廓线填补雷达波束遮挡区的方法研究. 气象科技, 2015, 43(5): 788-793. doi: 10.3969/j.issn.1671-6345.2015.05.003Yang L, Liu L P, Wang H Y. Filling-up of occultation regions using vertical profile of reflectivity factor. Meteor Sci Technol, 2015, 43(5): 788-793. doi: 10.3969/j.issn.1671-6345.2015.05.003 [14] Ryzhkov A V, Giangrande S E, Melnikov V M, et al. Calibration issues of dual-polarization radar measurements. J Atmos Oceanic Technol, 2005, 22(8): 1138-1155. doi: 10.1175/JTECH1772.1 [15] Lang T J, Nesbitt S W, Carey L D. On the correction of partial beam blockage in polarimetric radar data. J Atmos Oceanic Technol, 2009, 26(5): 943-957. doi: 10.1175/2008JTECHA1133.1 [16] Zhang P, Zrnic D, Ryzhkov A. Partial beam blockage correction using polarimetric radar measurements. J Atmos Oceanic Technol, 2013, 30(5): 861-872. doi: 10.1175/JTECH-D-12-00075.1 [17] 蔡金圻, 谭桂容, 牛若芸. 基于迁移CNN的江淮持续性强降水环流分型. 应用气象学报, 2021, 32(2): 233-244. doi: 10.11898/1001-7313.20210208Cai J Q, Tan G R, Niu R Y. Circulation pattern classification of persistent heavy rainfall in Jianghuai Region based on the transfer learning CNN model. J Appl Meteor Sci, 2021, 32(2): 233-244. doi: 10.11898/1001-7313.20210208 [18] 孙健, 曹卓, 李恒, 等. 人工智能技术在数值天气预报中的应用. 应用气象学报, 2021, 32(1): 1-11. doi: 10.11898/1001-7313.20210101Sun J, Cao Z, Li H, et al. Application of artificial intelligence technology to numerical weather prediction. J Appl Meteor Sci, 2021, 32(1): 1-11. doi: 10.11898/1001-7313.20210101 [19] 赵琳娜, 卢姝, 齐丹, 等. 基于全连接神经网络方法的日最高气温预报. 应用气象学报, 2022, 33(3): 257-269. doi: 10.11898/1001-7313.20220301Zhao L N, Lu S, Qi D, et al. Daily maximum air temperature forecast based on fully connected neural network. J Appl Meteor Sci, 2022, 33(3): 257-269. doi: 10.11898/1001-7313.20220301 [20] 金子琪, 王新敏, 鲍艳松, 等. 基于卷积神经网络的飑线识别算法. 应用气象学报, 2021, 32(5): 580-591. doi: 10.11898/1001-7313.20210506Jin Z Q, Wang X M, Bao Y S, et al. Squall line identification method based on convolution neural network. J Appl Meteor Sci, 2021, 32(5): 580-591. doi: 10.11898/1001-7313.20210506 [21] Chen H, Chandrasekar V, Tan H, et al. Rainfall estimation from ground radar and TRMM precipitation radar using hybrid deep neural networks. Geophys Res Lett, 2019, 46(17/18): 10669-10678. [22] 郭瀚阳, 陈明轩, 韩雷, 等. 基于深度学习的强对流高分辨率临近预报试验. 气象学报, 2019, 77(4): 715-727. https://www.cnki.com.cn/Article/CJFDTOTAL-QXXB201904009.htmGuo H Y, Chen M X, Han L, et al. High-resolution nowcasting experiment of strong convection based on deep learning. Acta Meteor Sinica, 2019, 77(4): 715-727. https://www.cnki.com.cn/Article/CJFDTOTAL-QXXB201904009.htm [23] Singh S, Sarkar S, Mitra P. A Deep Learning Based Approach with Adversarial Regularization for Doppler Weather Radar ECHO Prediction. 2017 IEEE International Geoscience and Remote Sensing Symposium(IGARSS), 2017: 5205-5208. [24] 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 [25] Yin X Y, Hu Z Q, Zheng J F, et al. Study on radar echo-filling in an occlusion area by a deep learning algorithm. Remote Sens, 2021, 13(9): 1779. doi: 10.3390/rs13091779 [26] 米前川, 高西宁, 李钥, 等. 深度学习方法在干旱预测中的应用. 应用气象学报, 2022, 33(1): 104-114. doi: 10.11898/1001-7313.20220109Mi Q C, Gao X N, Li Y, et al. Application of deep learning method to drought prediction. J Appl Meteor Sci, 2022, 33(1): 104-114. doi: 10.11898/1001-7313.20220109 [27] 刘黎平, 吴林林, 杨引明. 基于模糊逻辑的分步式超折射地物回波识别方法的建立和效果分析. 气象学报, 2007, 65(2): 252-260. https://www.cnki.com.cn/Article/CJFDTOTAL-QXXB200702010.htmLiu L P, Wu L L, Yang Y M. Development of fuzzy-logical two-step ground clutter detection algorithm. Acta Meteor Sinica, 2007, 65(2): 252-260. https://www.cnki.com.cn/Article/CJFDTOTAL-QXXB200702010.htm [28] 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. http://qikan.camscma.cn/article/id/20120203Li 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. http://qikan.camscma.cn/article/id/20120203 [29] 文浩, 刘黎平, 张扬. 多普勒天气雷达地物回波识别算法改进. 高原气象, 2017, 36(3): 736-749. https://www.cnki.com.cn/Article/CJFDTOTAL-GYQX201703014.htmWen H, Liu L P, Zhang Y. Improvements of ground clutter identification algorithm for Doppler weather radar. Plateau Meteor, 2017, 36(3): 736-749. https://www.cnki.com.cn/Article/CJFDTOTAL-GYQX201703014.htm [30] 曹西娟. 模糊时间序列在钻井风险概率中的预测及应用. 成都: 西南石油大学, 2018.Chao X J. Prediction and Application of Fuzzy Time Series in Drilling Risk Probability. Chengdu: Southwest Petroleum University, 2018. [31] 王红艳. 新一代天气雷达组网估算降水的覆盖能力分析法研究. 北京: 中国气象科学研究院, 2015.Wang H Y. Assessment of Coverage Ability and Research of Methods for Regional Mosaic QPE of CINRAD. Beijing: Chinese Academy of Meteorological Sciences, 2015. [32] LeCun Y, Boser B, Denker J S, et al. Backpropagation applied to handwritten zip code recognition. Neural Computation, 1989, 11(4): 541-551. [33] 黄丽梅, 王武, 林琼斌, 等. 基于核密度估计分类器的变换器故障诊断方法. 电网技术, 2019, 43(6): 2204-2210. https://www.cnki.com.cn/Article/CJFDTOTAL-DWJS201906044.htmHuang L M, Wang W, Lin Q B, et al. Fault diagnosis method of converter based on kernel density estimation classifier. Power System Technol, 2019, 43(6): 2204-2210. https://www.cnki.com.cn/Article/CJFDTOTAL-DWJS201906044.htm -
计量
- 摘要浏览量: 1837
- HTML全文浏览量: 230
- PDF下载量: 210
- 被引次数: 0
