Deep Learning for Polarimetric Radar Quantitative Precipitation Estimation during Landfalling Typhoons in South China
Heavy rain associated with landfalling typhoons often leads to disasters in South China, which can be reduced by improving the accuracy of radar quantitative precipitation estimation (QPE). At present, raindrop size distribution (DSD)-based nonlinear fitting (QPE<sub>DSD</sub>) and tradi...
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-08-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/13/16/3157 |
| _version_ | 1797522152276099072 |
|---|---|
| author | Yonghua Zhang Shuoben Bi Liping Liu Haonan Chen Yi Zhang Ping Shen Fan Yang Yaqiang Wang Yang Zhang Shun Yao |
| author_facet | Yonghua Zhang Shuoben Bi Liping Liu Haonan Chen Yi Zhang Ping Shen Fan Yang Yaqiang Wang Yang Zhang Shun Yao |
| author_sort | Yonghua Zhang |
| collection | DOAJ |
| description | Heavy rain associated with landfalling typhoons often leads to disasters in South China, which can be reduced by improving the accuracy of radar quantitative precipitation estimation (QPE). At present, raindrop size distribution (DSD)-based nonlinear fitting (QPE<sub>DSD</sub>) and traditional neural networks are the main radar QPE algorithms. The former is not sufficient to represent the spatiotemporal variability of DSDs through the generalized Z–R or polarimetric radar rainfall relations that are established using statistical methods since such parametric methods do not consider the spatial distribution of radar observables, and the latter is limited by the number of network layers and availability of data for training the model. In this paper, we propose an alternative approach to dual-polarization radar QPE based on deep learning (QPENet). Three datasets of “dual-polarization radar observations—surface rainfall (DPO—SR)” were constructed using radar observations and corresponding measurements from automatic weather stations (AWS) and used for QPENet<sub>V1</sub>, QPENet<sub>V2</sub>, and QPENet<sub>V3</sub>. In particular, 13 × 13, 25 × 25, and 41 × 41 radar range bins surrounding each AWS location were used in constructing the datasets for QPENet<sub>V1</sub>, QPENet<sub>V2</sub>, and QPENet<sub>V3</sub>, respectively. For training the QPENet models, the radar data and AWS measurements from eleven landfalling typhoons in South China during 2017–2019 were used. For demonstration, an independent typhoon event was randomly selected (i.e., Merbok) to implement the three trained models to produce rainfall estimates. The evaluation results and comparison with traditional QPE<sub>DSD</sub> algorithms show that the QPENet model has a better performance than the traditional parametric relations. Only when the hourly rainfall intensity is less than 5 mm (<i>R</i> < 5 mm·h<sup>−1</sup>), the QPE<sub>DSD</sub> model shows a comparable performance to QPENet. Comparing the three versions of the QPENet model, QPENet<sub>V2</sub> has the best overall performance. Only when the hourly rainfall intensity is less than 5 mm (<i>R</i> < 5 mm·h<sup>−1</sup>), QPENet<sub>V3</sub> performs the best. |
| first_indexed | 2024-03-10T08:25:25Z |
| format | Article |
| id | doaj.art-7251ddd0b334455d8fe5b15abcdeec32 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-10T08:25:25Z |
| publishDate | 2021-08-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-7251ddd0b334455d8fe5b15abcdeec322023-11-22T09:32:54ZengMDPI AGRemote Sensing2072-42922021-08-011316315710.3390/rs13163157Deep Learning for Polarimetric Radar Quantitative Precipitation Estimation during Landfalling Typhoons in South ChinaYonghua Zhang0Shuoben Bi1Liping Liu2Haonan Chen3Yi Zhang4Ping Shen5Fan Yang6Yaqiang Wang7Yang Zhang8Shun Yao9School of Geographical Sciences, Nanjing University of Information Science & Technology, Nanjing 210044, ChinaSchool of Geographical Sciences, Nanjing University of Information Science & Technology, Nanjing 210044, ChinaState Key Laboratory of Severe Weather, Chinese Academy of Meteorological Sciences, Beijing 100081, ChinaDepartment of Electrical and Computer Engineering, Colorado State University, Fort Collins, CO 80523, USAGuangdong Meteorological Public Service Center, Guangzhou 510641, ChinaGuangdong Emergency Early Warning Release Center, Guangzhou 510641, ChinaGuangdong Technology Support Center of Flood Control, Guangzhou 510635, ChinaInstitute of Artificial Intelligence for Meteorology, Chinese Academy of Meteorological Sciences, Beijing 100081, ChinaState Key Laboratory of Severe Weather, Chinese Academy of Meteorological Sciences, Beijing 100081, ChinaDepartment of Electrical and Computer Engineering, Colorado State University, Fort Collins, CO 80523, USAHeavy rain associated with landfalling typhoons often leads to disasters in South China, which can be reduced by improving the accuracy of radar quantitative precipitation estimation (QPE). At present, raindrop size distribution (DSD)-based nonlinear fitting (QPE<sub>DSD</sub>) and traditional neural networks are the main radar QPE algorithms. The former is not sufficient to represent the spatiotemporal variability of DSDs through the generalized Z–R or polarimetric radar rainfall relations that are established using statistical methods since such parametric methods do not consider the spatial distribution of radar observables, and the latter is limited by the number of network layers and availability of data for training the model. In this paper, we propose an alternative approach to dual-polarization radar QPE based on deep learning (QPENet). Three datasets of “dual-polarization radar observations—surface rainfall (DPO—SR)” were constructed using radar observations and corresponding measurements from automatic weather stations (AWS) and used for QPENet<sub>V1</sub>, QPENet<sub>V2</sub>, and QPENet<sub>V3</sub>. In particular, 13 × 13, 25 × 25, and 41 × 41 radar range bins surrounding each AWS location were used in constructing the datasets for QPENet<sub>V1</sub>, QPENet<sub>V2</sub>, and QPENet<sub>V3</sub>, respectively. For training the QPENet models, the radar data and AWS measurements from eleven landfalling typhoons in South China during 2017–2019 were used. For demonstration, an independent typhoon event was randomly selected (i.e., Merbok) to implement the three trained models to produce rainfall estimates. The evaluation results and comparison with traditional QPE<sub>DSD</sub> algorithms show that the QPENet model has a better performance than the traditional parametric relations. Only when the hourly rainfall intensity is less than 5 mm (<i>R</i> < 5 mm·h<sup>−1</sup>), the QPE<sub>DSD</sub> model shows a comparable performance to QPENet. Comparing the three versions of the QPENet model, QPENet<sub>V2</sub> has the best overall performance. Only when the hourly rainfall intensity is less than 5 mm (<i>R</i> < 5 mm·h<sup>−1</sup>), QPENet<sub>V3</sub> performs the best.https://www.mdpi.com/2072-4292/13/16/3157polarimetric radarquantitative precipitation estimationdeep learningconvolutional neural networklandfalling typhoons |
| spellingShingle | Yonghua Zhang Shuoben Bi Liping Liu Haonan Chen Yi Zhang Ping Shen Fan Yang Yaqiang Wang Yang Zhang Shun Yao Deep Learning for Polarimetric Radar Quantitative Precipitation Estimation during Landfalling Typhoons in South China Remote Sensing polarimetric radar quantitative precipitation estimation deep learning convolutional neural network landfalling typhoons |
| title | Deep Learning for Polarimetric Radar Quantitative Precipitation Estimation during Landfalling Typhoons in South China |
| title_full | Deep Learning for Polarimetric Radar Quantitative Precipitation Estimation during Landfalling Typhoons in South China |
| title_fullStr | Deep Learning for Polarimetric Radar Quantitative Precipitation Estimation during Landfalling Typhoons in South China |
| title_full_unstemmed | Deep Learning for Polarimetric Radar Quantitative Precipitation Estimation during Landfalling Typhoons in South China |
| title_short | Deep Learning for Polarimetric Radar Quantitative Precipitation Estimation during Landfalling Typhoons in South China |
| title_sort | deep learning for polarimetric radar quantitative precipitation estimation during landfalling typhoons in south china |
| topic | polarimetric radar quantitative precipitation estimation deep learning convolutional neural network landfalling typhoons |
| url | https://www.mdpi.com/2072-4292/13/16/3157 |
| work_keys_str_mv | AT yonghuazhang deeplearningforpolarimetricradarquantitativeprecipitationestimationduringlandfallingtyphoonsinsouthchina AT shuobenbi deeplearningforpolarimetricradarquantitativeprecipitationestimationduringlandfallingtyphoonsinsouthchina AT lipingliu deeplearningforpolarimetricradarquantitativeprecipitationestimationduringlandfallingtyphoonsinsouthchina AT haonanchen deeplearningforpolarimetricradarquantitativeprecipitationestimationduringlandfallingtyphoonsinsouthchina AT yizhang deeplearningforpolarimetricradarquantitativeprecipitationestimationduringlandfallingtyphoonsinsouthchina AT pingshen deeplearningforpolarimetricradarquantitativeprecipitationestimationduringlandfallingtyphoonsinsouthchina AT fanyang deeplearningforpolarimetricradarquantitativeprecipitationestimationduringlandfallingtyphoonsinsouthchina AT yaqiangwang deeplearningforpolarimetricradarquantitativeprecipitationestimationduringlandfallingtyphoonsinsouthchina AT yangzhang deeplearningforpolarimetricradarquantitativeprecipitationestimationduringlandfallingtyphoonsinsouthchina AT shunyao deeplearningforpolarimetricradarquantitativeprecipitationestimationduringlandfallingtyphoonsinsouthchina |