Tropical Cyclone Intensity Prediction Using Deep Convolutional Neural Network
In this study, deep convolutional neural network (CNN) models of stimulated tropical cyclone intensity (TCI), minimum central pressure (MCP), and maximum 2 min mean wind speed at near center (MWS) were constructed based on ocean and atmospheric reanalysis, as well Best Track of tropical hurricane da...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-05-01
|
| Series: | Atmosphere |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2073-4433/13/5/783 |
| _version_ | 1827670380262719488 |
|---|---|
| author | Xiao-Yan Xu Min Shao Pu-Long Chen Qin-Geng Wang |
| author_facet | Xiao-Yan Xu Min Shao Pu-Long Chen Qin-Geng Wang |
| author_sort | Xiao-Yan Xu |
| collection | DOAJ |
| description | In this study, deep convolutional neural network (CNN) models of stimulated tropical cyclone intensity (TCI), minimum central pressure (MCP), and maximum 2 min mean wind speed at near center (MWS) were constructed based on ocean and atmospheric reanalysis, as well Best Track of tropical hurricane data over 2014–2018. In order to explore the interpretability of the model structure, sensitivity experiments were designed with various combinations of predictors. The model test results show that simplified VGG-16 (VGG-16 s) outperforms the other two general models (LeNet-5 and AlexNet). The results of the sensitivity experiments display good consistency with the hypothesis and perceptions, which verifies the validity and reliability of the model. Furthermore, the results also suggest that the importance of predictors varies in different targets. The top three factors that are highly related to TCI are sea surface temperature (SST), temperature at 500 hPa (TEM_500), and the differences in wind speed between 850 hPa and 500 hPa (vertical wind shear speed, VWSS). VWSS, relative humidity (RH), and SST are more significant than MCP. For MWS and SST, TEM_500, and temperature at 850 hPa (TEM_850) outweigh the other variables. This conclusion also implies that deep learning could be an alternative way to conduct intensive and quantitative research. |
| first_indexed | 2024-03-10T03:21:15Z |
| format | Article |
| id | doaj.art-3535a32ea3754d75b1e2f85dd8e587c2 |
| institution | Directory Open Access Journal |
| issn | 2073-4433 |
| language | English |
| last_indexed | 2024-03-10T03:21:15Z |
| publishDate | 2022-05-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Atmosphere |
| spelling | doaj.art-3535a32ea3754d75b1e2f85dd8e587c22023-11-23T10:02:45ZengMDPI AGAtmosphere2073-44332022-05-0113578310.3390/atmos13050783Tropical Cyclone Intensity Prediction Using Deep Convolutional Neural NetworkXiao-Yan Xu0Min Shao1Pu-Long Chen2Qin-Geng Wang3School of the Environment, Nanjing University, Nanjing 210046, ChinaSchool of the Environment, Nanjing Normal University, Nanjing 210023, ChinaNet Zero Era (Jiangsu) Environmental Technology Co., Ltd., Suzhou 215000, ChinaSchool of the Environment, Nanjing University, Nanjing 210046, ChinaIn this study, deep convolutional neural network (CNN) models of stimulated tropical cyclone intensity (TCI), minimum central pressure (MCP), and maximum 2 min mean wind speed at near center (MWS) were constructed based on ocean and atmospheric reanalysis, as well Best Track of tropical hurricane data over 2014–2018. In order to explore the interpretability of the model structure, sensitivity experiments were designed with various combinations of predictors. The model test results show that simplified VGG-16 (VGG-16 s) outperforms the other two general models (LeNet-5 and AlexNet). The results of the sensitivity experiments display good consistency with the hypothesis and perceptions, which verifies the validity and reliability of the model. Furthermore, the results also suggest that the importance of predictors varies in different targets. The top three factors that are highly related to TCI are sea surface temperature (SST), temperature at 500 hPa (TEM_500), and the differences in wind speed between 850 hPa and 500 hPa (vertical wind shear speed, VWSS). VWSS, relative humidity (RH), and SST are more significant than MCP. For MWS and SST, TEM_500, and temperature at 850 hPa (TEM_850) outweigh the other variables. This conclusion also implies that deep learning could be an alternative way to conduct intensive and quantitative research.https://www.mdpi.com/2073-4433/13/5/783tropical cyclonedeep learningconvolutional neural networkinterpretability |
| spellingShingle | Xiao-Yan Xu Min Shao Pu-Long Chen Qin-Geng Wang Tropical Cyclone Intensity Prediction Using Deep Convolutional Neural Network Atmosphere tropical cyclone deep learning convolutional neural network interpretability |
| title | Tropical Cyclone Intensity Prediction Using Deep Convolutional Neural Network |
| title_full | Tropical Cyclone Intensity Prediction Using Deep Convolutional Neural Network |
| title_fullStr | Tropical Cyclone Intensity Prediction Using Deep Convolutional Neural Network |
| title_full_unstemmed | Tropical Cyclone Intensity Prediction Using Deep Convolutional Neural Network |
| title_short | Tropical Cyclone Intensity Prediction Using Deep Convolutional Neural Network |
| title_sort | tropical cyclone intensity prediction using deep convolutional neural network |
| topic | tropical cyclone deep learning convolutional neural network interpretability |
| url | https://www.mdpi.com/2073-4433/13/5/783 |
| work_keys_str_mv | AT xiaoyanxu tropicalcycloneintensitypredictionusingdeepconvolutionalneuralnetwork AT minshao tropicalcycloneintensitypredictionusingdeepconvolutionalneuralnetwork AT pulongchen tropicalcycloneintensitypredictionusingdeepconvolutionalneuralnetwork AT qingengwang tropicalcycloneintensitypredictionusingdeepconvolutionalneuralnetwork |