Abstract
One of the most important aspects of tropical cyclone (TC) forecasting and disaster warning/management is accurately calculating TC intensity. The Dvorak technique—as well as a few upgraded versions—has been used by forecasters worldwide for more than 40 years to estimate temperature intensity. Nevertheless, the operational Dvorak techniques that are mainly employed by different agencies have a number of drawbacks, including intrinsic subjectivity that results in conflicting intensity estimates across different basins. CNN model gives more training parameters to enhance the overall training process and this model is enhanced with channel attention (CA) and spatial attention (SA) layer to gain valuable results. SA layer tries to pay extra consideration to the semantic-interrelated regions, instead of considering all image area similarly. SA layer does not contain the visual feature to calculate the weight, so the CA layer presents the visual feature
Keywords: Tropical cyclone intensity, Channel attentive dense convolutional neural network, DenseNet-121 CNN, Channel attentive layer, Spatial attentive layer
References
1. Jin, Qingwen, Xiangtao Fan, Jian Liu, Zhuxin Xue, and Hongdeng Jian. "Estimating
tropical cyclone intensity in the South China Sea using the XGBoost Model and FengYun Satellite images." Atmosphere 11, no. 4 (2020): 423.
2. Velden, Christopher S., and Derrick Herndon. "A consensus approach for estimating tropical cyclone intensity from meteorological satellites: SATCON." Weather and Forecasting 35, no. 4 (2020): 1645-1662.
3. Zhuo, Jing-Yi, and Zhe-Min Tan. "Physics-augmented deep learning to improve tropical cyclone intensity and size estimation from satellite imagery." Monthly Weather Review 149, no. 7 (2021): 2097-2113.
4. Chen, Rui, Weimin Zhang, and Xiang Wang. "Machine learning in tropical cyclone forecast modeling: A review." Atmosphere 11, no. 7 (2020): 676.
5. Asif, Amina, Muhammad Dawood, Bismillah Jan, Javaid Khurshid, Mark DeMaria, and Fayyaz ul Amir Afsar Minhas. "PHURIE: hurricane intensity estimation from infrared satellite imagery using machine learning." Neural Computing and Applications 32 (2020): 4821-4834.
6. Charrua, Alberto Bento, Rajchandar Padmanaban, Pedro Cabral, Salomão Bandeira, and Maria M. Romeiras. "Impacts of the tropical cyclone idai in mozambique: A multi-temporal landsat satellite imagery analysis." Remote Sensing 13, no. 2 (2021): 201.
7. Yu, Peng, Johnny A. Johannessen, Xiao-Hai Yan, Xupu Geng, Xiaojing Zhong, and Lin Zhu. "A study of the intensity of Tropical Cyclone Idai using dual-polarization Sentinel-1 data." Remote Sensing 11, no. 23 (2019):2837.
8. Lu, Xiaoqin, Hui Yu, Xiaoming Yang, and Xiaofeng Li. "Estimating tropical cyclone size in the Northwestern Pacific from geostationary satellite infrared images." Remote Sensing 9, no. 7 (2017): 728.
9. Mohan, Preeya, and Eric Strobl. "The short-term economic impact of tropical Cyclone Pam: an analysis using VIIRS nightlightsatellite imagery." International journal of remote sensing 38, no. 21 (2017): 5992-6006.
10. Hoque, Muhammad Al-Amin, Stuart Phinn, Chris Roelfsema, and Iraphne Childs. "Tropical cyclone disaster management using remote sensing and spatial analysis: A review." International journal of disaster risk reduction 22 (2017): 345-354.
11. Zhang, Caiyun, Sara Denka Durgan, and David Lagomasino. "Modeling risk of mangroves to tropical cyclones: A case study of Hurricane Irma." Estuarine, Coastal and Shelf Science 224 (2019): 108-116.
12. Kim, Minsang, Myung-Sook Park, Jungho Im, Seonyoung Park, and Myong-In Lee. "Machine learning approaches for detecting tropical cyclone formation using satellite data." Remote Sensing 11, no. 10 (2019): 1195.
13. Maskey, Manil, Rahul Ramachandran, Muthukumaran Ramasubramanian, Iksha
Gurung, Brian Freitag, Aaron Kaulfus, Drew Bollinger, Daniel J. Cecil, and Jeffrey Miller. "Deepti: Deep-learning-based tropical cyclone intensity estimation system." IEEE journal of
selected topics in applied Earth observations and remote sensing 13 (2020): 4271-4281.