Abstract
Prediction of Tropical cyclones (TCs), particularly intensity prediction, has always been challenging for climate researchers due to the complicated physical mechanisms in TC dynamics and the way it interacts with upper-ocean and atmospheric circulation. Furthermore, the available data set over the North Indian Ocean (NIO) is also very limited for Machine Learning (ML) model development. Here, we demonstrated a simple yet robust hybrid architecture leveraging a Convolutional Neural Network for automated prediction of the intensity of the cyclone based on IR satellite imagery of 2000–2022. The model comprises a binary classifier, a multiclass classifier, a YOLOv3 based cyclone detector and a regression module. The paper also highlights the discrepancy between the results of independent testing wherein training is done on 2000 to 2019 dataset and tested on 2020 to 2022 dataset, as well as the outcomes of a stratified train-test split performed over the entire dataset using a 70:15:15 ratio for training, validation and testing, respectively. The model is tuned for the NIO region with a binary classification accuracy score of 98.4% (± 0.003), multiclass classification accuracy of 63.83% (± 1.3) and RMSE of 16.2 (± 0.9) knots on stratified split. The results highlight the careful interpretation of the DL model’s performance when applied to time series problems. Additionally, it discusses the limitations stemming from the dataset's small size and the challenges posed by the 5 kt resolution of the best track intensity estimation from the Indian Meteorological Department (IMD). The internal representations learned by the model through feature maps analysis were studied, shedding light on the model’s decision-making process. The study underscores the need for further data accumulation and highlights avenues for enhancing model performance in the future.
Similar content being viewed by others
References
Bandyopadhyay, S., Dasgupta, S., Khan, Z. H., & Wheeler, D. (2021). Correction to: spatiotemporal analysis of tropical cyclone landfalls in Northern Bay of Bengal, India and Bangladesh. Asia-Pacific Journal of Atmospheric Sciences, 57, 865–875. https://doi.org/10.1007/s13143-021-00227-4
Bonfanti, C., Stewart, J., Hall, D., & Govett, M. (2020). Tropical and extratropical cyclone detection using deep learning. Journal of Applied Meteorology and Climatology, 59(12), 1971–1985. https://doi.org/10.1175/JAMC-D-20-0117.1
Chandan, Roy Rita, Kovordányi (2012) Tropical cyclone track forecasting techniques ― A review Atmospheric Research 104-10540-69 10.1016/j.atmosres.2011.09.012
Chaudhari, S., Dutta, D., Goswami, S., & Middey, A. (2013). Intensity forecast of tropical cyclones over North Indian Ocean using multilayer perceptron model: Skill and performance verification. Natural Hazards, 65, 97–113. https://doi.org/10.1007/s11069-012-0346-7
Chen, B.-F., Chen, B., Lin, H.-T., & Elsberry, R. L. (2019). Estimating tropical cyclone intensity by satellite imagery utilizing convolutional neural networks, vol. 34, issue 2, pp. 447–465, 1 Apr 2019. https://doi.org/10.1175/WAF-D-18-0136.1.
Devaraj, J., Ganesan, S., Elavarasan, R. M., & Subramaniam, U. (2021). A novel deep learning based model for tropical intensity estimation and post-disaster management of hurricanes. Applied Sciences, 11, 4129. https://doi.org/10.3390/app11094129
Emanuel, K. A. & Nolan, D. S., (2004). Tropical cyclone activity and global climate. In Proceedings of the 26th Conference Hurricanes Tropical Meteorology, Miami, FL, USA, vol. 10, pp. 240–241. https://ams.confex.com/ams/26HURR/techprogram/paper_75463.htm .
Kar, C., Banerjee, S. (2021). Tropical cyclones intensity estimation by feature fusion and random forest classifier using satellite images. Journal of the Indian Society of remote Sensing, pp. 689–700. https://doi.org/10.1007/s12524-021-01477-5
Kar, C., Kumar, A., & Banerjee, S. (2019). Tropical cyclone intensity detection by geometric features of cyclone images and multilayer perceptron. Springer Nature Appl. Sci., 1, 1099. https://doi.org/10.1007/s42452-019-1134-8
Koepken, C. (2004). Solar stray light effects in meteosat radiances observed and quantified using operational data monitoring at ECMWF. Journal of Applied Meteorology, 43, 28–37. https://doi.org/10.1175/1520-0450(2004)043%3c0028:SSLEIM%3e2.0.CO;2
Kumawat, S., & Jaiswal, J. (2021). Cyclone detection and forecasting using deep neural networks through satellite data. Proceedings of International Conference on Innovations in Software Architecture and Computational Systems. https://doi.org/10.1007/978-981-16-4301-9_2
Lee, Y.-J. (2021). Interpretable tropical cyclone intensity estimation using Dvorak-inspired machine learning techniques. Engineering Applications of Artificial Intelligence, 101, https://doi.org/10.1016/j.engappai.2021.104233.
Lee, C. Y., Tippett, M. K., Camargo, S. J., & Sobel, A. H. (2015). Probabilistic multiple linear regression modeling for tropical cyclone intensity. Mon. Wea. Rev., 143, 933–954. https://doi.org/10.1175/MWR-D-14-00171.1
Liefhebber, F., Lammens, S., Brussee, P. W. G., Bos, A., John, V. O., Rüthrich, F., Onderwaater, J., Grant, M. G., & Schulz, J. (2020). Automatic quality control of the Meteosat First Generation measurements. Atmos. Meas. Tech., 13, 1167–1179.
Matsuoka, D., Nakano, M., Sugiyama, D., & Uchida, S. (2018). Deep learning approach for detecting tropical cyclones and their precursors in the simulation by a cloud-resolving global nonhydrostatic atmospheric model. Progress in Earth and Planetary Science, 5(1), 1–16. https://doi.org/10.1186/s40645-018-0245-y
Meer, M. S., Mishra, A. K., & Nagaraju, V. (2024). Investigation of meteorological characteristics of tropical supercyclone shaheen insights from high-resolution satellite observations. J Indian Soc Remote Sens. https://doi.org/10.1007/s12524-024-01857-7
Mooers, G., Pritchard, M., Beucler, T., Srivastava, P., Mangipudi, H., Peng, L., Gentine, P., & Mandt, S. (2023). Comparing storm resolving models and climates via unsupervised machine learning. Scientific Reports, 13(1), 22365. https://doi.org/10.1038/s41598-023-49455-w
Moon, I. J., Ginis, I., Hara, T., & Thomas, B. (2007). A Physics-based parameterization of air-sea momentum flux at high wind speeds and its impact on hurricane intensity predictions. Monthly Weather Review, 135, 2869–2878. https://doi.org/10.1175/MWR3432.1
Nair, A., Srujan, K., Kulkarni, S., Alwadhi, K., Jain, N., Kodamana, H., Sandeep, S., John, V. (2022) A deep learning framework for the detection of tropical cyclones from satellite images. IEEE Geoscience and Remote Sensing Letters, 19. https://doi.org/10.1109/LGRS.2021.3131638
Olander, T. L. & Velden, C. S. (2016). The current status of the UW-CIMSS Advanced Dvorak Technique (ADT). In: Confernece on Hurricanes and Tropical Meteorology.
Olander, T. L., & Velden, C. S. (2019). The advanced Dvorak technique (ADT) for estimating tropical cyclone intensity: Update and new capabilities. Weather Forecasting, 34(4), 905–922. https://doi.org/10.1175/WAF-D-19-0007.1
Olander, T., Wimmers, A., Velden, C., & Kossin, J. (2021). Investigation of machine learning using satellite-based advanced dvorak technique analysis parameters to estimate tropical cyclone intensity. AMS Weather and Forecasting, 36, 2161–2186. https://doi.org/10.1175/WAF-D-20-0234.1
Pineros, M. F., Ritchie, E. A., & Tyo, J. S. (2011). Estimating tropical cyclone intensity from infrared image data. Weather Forecasting, 26(5), 690–698. https://doi.org/10.1175/WAF-D-10-05062.1
Redmon, J., & Farhadi, A ., (2018). YOLOv3: an incremental improvement, April 2018, ar**v:1804.02767.
Ritchie, E. A., & Valliere, K. G. (2012). Tropical cyclone intensity estimation in the North Atlantic basin using an improved deviation angle variance technique. Weather Forecasting, 27(5), 1264–1277. https://doi.org/10.1175/WAF-D-11-00156.1
Shakya, S., Kumar, S., & Goswami, M. (2020). Deep learning algorithm for satellite imaging based cyclone detection. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, vol 13, pp. 827–839. https://https://doi.org/10.1109/JSTARS.2020.2970253.
Sharma, N., Ali, M. M., Knaff, J. A., & Chand, P. (2013). A soft-computing cyclone intensity prediction scheme for the western North Pacific Ocean. RMetS Atmospheric Science Letters, 14, pp. 187–192. https://doi.org/10.1002/asl2.438 .
Tian, W., Huang, W., Yi, L., Wu, L., & Wang, C. (2020). A CNN-based hybrid model for tropical cyclone intensity estimation in meteorological industry. IEEE Access, 8, 59158–59168. https://doi.org/10.1109/ACCESS.2020.2982772
Wang, C., Zheng, G., Li, X., Xu, Q., Liu, B., & Zhang, J. (2021). Tropical cyclone intensity estimation from geostationary satellite imagery using deep convolutional neural networks. IEEE Transactions on Geoscience and Remote Sensing., Early access, Mar 26, 202. https://doi.org/10.1109/TGRS.2021.3066299.
Zhang, Q., & Liguang, Q. (2009). Tropical cyclone damages in China 1983–2006. Bulletin of the American Meteorological Society, 90(4), 489. https://doi.org/10.1175/2008BAMS2631.1
Acknowledgements
The CIMSS Tropical Cyclone Archive for providing open access data. Author Saurabh Das also thankfully acknowledge the financial support received under SERB project MTR/2019/001581.
Funding
The research is funded by SERB project MTR/2019/001581.
Author information
Authors and Affiliations
Contributions
The authors confirm contribution to the paper as follows: Study conception and design: MM, SD. Data Collection and Analysis: MM. Interpretation of results: MM, SD. Draft Manuscript Preparation: MM. All authors reviewed the results and approved the final version of the manuscript.
Corresponding author
Ethics declarations
Conflict of interest
The authors declared that they have no conflict of interest.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix
Appendix
Precision is the measure of a positive predicted value and is defined as
Recall or True Positive Rate is defined as
True Positive (TP) is the outcome where the model correctly predicts the positive class. False Positive (FP) is the outcome where the model incorrectly predicts the positive class.
F1 score is the harmonic mean of precision and recall and is defined as
False positive rate (FPR) is the ratio of negative events wrongly classified as positive and the total number of actual negative events.
About this article
Cite this article
Mawatwal, M., Das, S. An End-to-End Deep Learning Framework for Cyclone Intensity Estimation in North Indian Ocean Region Using Satellite Imagery. J Indian Soc Remote Sens (2024). https://doi.org/10.1007/s12524-024-01929-8
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s12524-024-01929-8