Cascade Deep Learning With Physics Guidance for Typhoon Intensity Prediction

Zhengya Sun , Bojie Fan , Bo Yin , Jie Nie , Anan Liu

CAAI Transactions on Intelligence Technology ›› 2026, Vol. 11 ›› Issue (5) : 1345 -1359.

PDF (2603KB)
CAAI Transactions on Intelligence Technology ›› 2026, Vol. 11 ›› Issue (5) :1345 -1359. DOI: 10.1049/cit2.70160
ORIGINAL RESEARCH
research-article
Cascade Deep Learning With Physics Guidance for Typhoon Intensity Prediction
Author information +
History +
PDF (2603KB)

Abstract

Typhoons pose a significant threat to both human safety and economic stability. As a crucial metric for assessing their destructive potential, typhoon intensity (TI) prediction has become an important research focus, with numerous methods developed. However, effectively combining two-dimensional typhoon structure domain-expert knowledge (2D-TSDK) with three-dimensional typhoon structure data-driven knowledge (3D-TSDK) for accurate TI prediction remains a challenge. In this paper, we propose a novel cascaded deep learning framework that combines domain expertise with data-driven refinement for typhoon intensity prediction. The two-stage architecture first generates initial intensity estimates by fusing 2D-TSDK features with established physical wind-pressure relationships via a specialised loss function. These estimates are subsequently refined by a vision transformer that extract fine-grained spatial patterns from 3D wind field reanalysis data. Both stages incorporate temporal attention mechanisms with dual (absolute and relative) position encoding, providing comprehensive modelling of typhoon evolution dynamics. Extensive experiments on CMA–BST and ERA-interim datasets demonstrate consistent performance improvements over baseline methods, confirming the framework's effectiveness for TI prediction.

Keywords

physics-guided neural networks / spatio-temporal sequence / typhoon intensity prediction

Cite this article

Download citation ▾
Zhengya Sun, Bojie Fan, Bo Yin, Jie Nie, Anan Liu. Cascade Deep Learning With Physics Guidance for Typhoon Intensity Prediction. CAAI Transactions on Intelligence Technology, 2026, 11 (5) : 1345-1359 DOI:10.1049/cit2.70160

登录浏览全文

4963

注册一个新账户 忘记密码

References

[1]

D. Swain, “ Tropical Cyclones and Coastal Vulnerability: Assessment and Mitigation,” in Geospatial Technologies for Land and Water Resources Management (2022), 587-621.

[2]

M. A. Bender, I. Ginis, and Y. Kurihara, “ Numerical Simulations of Tropical cyclone-ocean Interaction With a High-Resolution Coupled Model,” Journal of Geophysical Research: Atmospheres 98, no. D12 (1993): 23245-23263, https://doi.org/10.1029/93jd02370.

[3]

M. E. Kozar, M. E. Mann, S. J. Camargo, J. P. Kossin, and J. L. Evans, “ Stratified Statistical Models of North Atlantic Basin-Wide and Regional Tropical Cyclone Counts,” Journal of Geophysical Research: Atmospheres 117, no. D18 (2012), https://doi.org/10.1029/2011jd017170.

[4]

L. M. Ma, “ Research Progress on China Typhoon Numerical Prediction Models and Associated Major Techniques,” Progress in Geophysics 29 (2014): 1013-1022, https://doi.org/10.6038/pg20140305.

[5]

B. Pan, X. Xu, and Z. Shi, “ Tropical Cyclone Intensity Prediction Based on Recurrent Neural Networks,” Electronics Letters 55, no. 7 (2019): 413-415, https://doi.org/10.1049/el.2018.8178.

[6]

S. Yuan, C. Wang, B. Mu, F. Zhou, and W. Duan, “ Typhoon Intensity Forecasting Based on Lstm Using the Rolling Forecast Method,” Algorithms 14, no. 3 (2021): 83, https://doi.org/10.3390/a14030083.

[7]

S. Jiang and L. Tao, “ Classification and Estimation of Typhoon Intensity From Geostationary Meteorological Satellite Images Based on Deep Learning,” Atmosphere 13, no. 7 (2022): 1113, https://doi.org/10.3390/atmos13071113.

[8]

M. Higa, S. Tanahara, Y. Adachi, et al., “ Domain Knowledge Integration Into Deep Learning for Typhoon Intensity Classification,” Scientific Reports 11, no. 1 (2021): 12972, https://doi.org/10.1038/s41598-021-92286-w.

[9]

P. Yin, H. Chen, H. Huang, H. Su, Q. Wu, and Q. Wan, “ A Region-Based Convolutional Fusion Network for Typhoon Intensity Estimation in Satellite Images,” Engineering Applications of Artificial Intelligence 134 (2024): 108671, https://doi.org/10.1016/j.engappai.2024.108671.

[10]

Z. Zhang, X. Yang, L. Shi, et al., “ A Neural Network Framework for Fine-Grained Tropical Cyclone Intensity Prediction,” Knowledge-Based Systems 241 (2022): 108195, https://doi.org/10.1016/j.knosys.2022.108195.

[11]

S. Jiang, H. Fan, and C. Wang, “ Improvement of Typhoon Intensity Forecasting by Using a Novel Spatio-Temporal Deep Learning Model,” Remote Sensing 14, no. 20 (2022): 5205, https://doi.org/10.3390/rs14205205.

[12]

R. Chen, X. Wang, W. Zhang, X. Zhu, A. Li, and C. Yang, “ A Hybrid CNN-LSTM Model for Typhoon Formation Forecasting,” GeoInformatica 23, no. 3 (2019): 375-396, https://doi.org/10.1007/s10707-019-00355-0.

[13]

G. Xu, K. Lin, X. Li, and Y. Ye, “ Saf-Net: A Spatio-Temporal Deep Learning Method for Typhoon Intensity Prediction,” Pattern Recognition Letters 155 (2022): 121-127, https://doi.org/10.1016/j.patrec.2021.11.012.

[14]

C. Wang, X. Li, and G. Zheng, “ Tropical Cyclone Intensity Forecasting Using Model Knowledge Guided Deep Learning Model,” Environmental Research Letters 19, no. 2 (2024): 024006, https://doi.org/10.1088/1748-9326/ad1bde.

[15]

D. Ma, L. Wang, S. Fang, and J. Lin, “ Tropical Cyclone Intensity Prediction by Inter-and Intra-Pattern Fusion Based on Multi-Source Data,” Environmental Research Letters 18, no. 1 (2023): 014020, https://doi.org/10.1088/1748-9326/aca9e2.

[16]

W. Li, Y. Li, Y. Yang, D. Song, Z. Wei, and A. A. Liu, “ Multimodal Adversarial Fusion for Typhoon Intensity Forecasting,” IEEE Transactions on Geoscience and Remote Sensing 62 (2024): 1-12, https://doi.org/10.1109/tgrs.2024.3367962.

[17]

M. DeMaria, C. R. Sampson, J. A. Knaff, and K. D. Musgrave, “ Is Tropical Cyclone Intensity Guidance Improving?,” Bulletin of the American Meteorological Society 95, no. 3 (2014): 387-398, https://doi.org/10.1175/bams-d-12-00240.1.

[18]

K. Emanuel and F. Zhang, “ On the Predictability and Error Sources of Tropical Cyclone Intensity Forecasts,” Journal of the Atmospheric Sciences 73, no. 9 (2016): 3739-3747, https://doi.org/10.1175/jas-d-16-0100.1.

[19]

P. Hewage, M. Trovati, E. Pereira, and A. Behera, “ Deep Learning-Based Effective Fine-Grained Weather Forecasting Model,” Pattern Analysis & Applications 24, no. 1 (2021): 343-366, https://doi.org/10.1007/s10044-020-00898-1.

[20]

F. Molteni, R. Buizza, T. N. Palmer, and T. Petroliagis, “ The Ecmwf Ensemble Prediction System: Methodology and Validation,” Quarterly Journal of the Royal Meteorological Society 122, no. 529 (1996): 73-119, https://doi.org/10.1002/qj.49712252905.

[21]

A. Latto, Y. J. Kim, M. B. Natoli, and E. M. Schaefer, “ Preliminary Reporting on the Case Evaluation of Forecasters’ Needs for the Global Forecast System (Gfs),” in 104th Annual AMS Meeting 2024 (2024), 427911.

[22]

J. G. Powers, J. B. Klemp, W. C. Skamarock, et al., “ The Weather Research and Forecasting Model: Overview, System Efforts, and Future Directions,” Bulletin of the American Meteorological Society 98, no. 8 (2017): 1717-1737, https://doi.org/10.1175/bams-d-15-00308.1.

[23]

R. S. Reddy, R. Karim, L. White, and A. Thorp, “ Mesoscale Modeling Investigation Using Penn State/NCAR MM5 Model for Weather Simulation and Prediction,” Journal of the Mississippi Academy of Sciences 55 (2010): 171-180.

[24]

N. M. Leonardo and B. A. Colle, “ Verification of Multimodel Ensemble Forecasts of North Atlantic Tropical Cyclones,” Weather and Forecasting 32, no. 6 (2017): 2083-2101, https://doi.org/10.1175/waf-d-17-0058.1.

[25]

D. J. Halperin, H. E. Fuelberg, R. E. Hart, J. H. Cossuth, P. Sura, and R. J. Pasch, “ An Evaluation of Tropical Cyclone Genesis Forecasts From Global Numerical Models,” Weather and Forecasting 28, no. 6 (2013): 1423-1445, https://doi.org/10.1175/waf-d-13-00008.1.

[26]

B. R. Jarvinen, and C. J. Neumann, “ Statistical Forecasts of Tropical Cyclone Intensity for the North Atlantic Basin,” NOAA Technical Memorandum NWS NHC-10 (National Hurricane Center, National Weather Service, 1979).

[27]

M. DeMaria and J. Kaplan, “ A Statistical Hurricane Intensity Prediction Scheme (Ships) for the Atlantic Basin,” Weather and Forecasting 9, no. 2 (1994): 209-220, https://doi.org/10.1175/1520-0434(1994)009<0209:aships>2.0.co;2.

[28]

M. DeMaria, M. Mainelli, L. K. Shay, J. A. Knaff, and J. Kaplan, “ Further Improvements to the Statistical Hurricane Intensity Prediction Scheme (Ships),” Weather and Forecasting 20, no. 4 (2005): 531-543, https://doi.org/10.1175/waf862.1.

[29]

C. Y. Lee, M. K. Tippett, A. H. Sobel, and S. J. Camargo, “ Autoregressive Modeling for Tropical Cyclone Intensity Climatology,” Journal of Climate 29, no. 21 (2016): 7815-7830, https://doi.org/10.1175/jcli-d-15-0909.1.

[30]

R. Jing and N. Lin, “ Tropical Cyclone Intensity Evolution Modeled as a Dependent Hidden Markov Process,” Journal of Climate 32, no. 22 (2019): 7837-7855, https://doi.org/10.1175/jcli-d-19-0027.1.

[31]

R. Pradhan, R. S. Aygun, M. Maskey, R. Ramachandran, and D. J. Cecil, “ Tropical Cyclone Intensity Estimation Using a Deep Convolutional Neural Network,” IEEE Transactions on Image Processing 27, no. 2 (2017): 692-702, https://doi.org/10.1109/tip.2017.2766358.

[32]

B. F. Chen, B. Chen, H. T. Lin, and R. L. Elsberry, “ Estimating Tropical Cyclone Intensity by Satellite Imagery Utilizing Convolutional Neural Networks,” Weather and Forecasting 34, no. 2 (2019): 447-465, https://doi.org/10.1175/waf-d-18-0136.1.

[33]

J. Lee, J. Im, D. H. Cha, H. Park, and S. Sim, “ Tropical Cyclone Intensity Estimation Using Multi-Dimensional Convolutional Neural Networks From Geostationary Satellite Data,” Remote Sensing 12, no. 1 (2019): 108, https://doi.org/10.3390/rs12010108.

[34]

B. Chen, B. F. Chen, and Y. N. Chen, “ Real-Time Tropical Cyclone Intensity Estimation by Handling Temporally Heterogeneous Satellite Data,” in Proceedings of the AAAI Conference on Artificial Intelligence (2021), 14721-14728.

[35]

B. Tong, X. Wang, J. Fu, P. Chan, and Y. He, “ Short-Term Prediction of the Intensity and Track of Tropical Cyclone via Convlstm Model,” Journal of Wind Engineering and Industrial Aerodynamics 226 (2022): 105026, https://doi.org/10.1016/j.jweia.2022.105026.

[36]

W. Yang, J. Fei, X. Huang, J. Ding, and X. Cheng, “ Enhancing Tropical Cyclone Intensity Estimation From Satellite Imagery Through Deep Learning Techniques,” Journal of Meteorological Research 38, no. 4 (2024): 652-663, https://doi.org/10.1007/s13351-024-3186-y.

[37]

P. Nandal, P. Mann, N. Bohra, G. Aldehim, A. A. H. Elnour, and R. Allafi, “ Tropical Cyclone Intensity Estimation Based on YOLO-NAS Using Satellite Images in Real Time,” Alexandria Engineering Journal 113 (2025): 227-241, https://doi.org/10.1016/j.aej.2024.10.072.

[38]

H. Chen, P. Yin, H. Huang, Q. Wu, R. Liu, and X. Zhu, “ Typhoon Intensity Prediction With Vision Transformer,” preprint, arXiv: 2311.16450 (2023), https://doi.org/10.48550/arXiv.2311.16450.

[39]

Y. Tian, W. Zhou, P. K. Y. Cheung, and Z. Liu, “ Vision Transformer for Extracting Tropical Cyclone Intensity From Satellite Images,” Advances in Atmospheric Sciences 42, no. 1 (2025): 79-93, https://doi.org/10.1007/s00376-024-3191-1.

[40]

S. R. Katragadda, “ Tropical Cyclone Intensity Prediction Using Optimized Vision Transformer Dynamic Graph Convolutional Network,” Knowledge-Based Systems 330 (2025): 114603, https://doi.org/10.1016/j.knosys.2025.114603.

[41]

J. Lian, J. Shao, H. Yu, et al., “ Multimodal Fusion Learning for Predicting Tropical Cyclone Intensity Over Western North Pacific,” IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 18 (2025): 7048-7063, https://doi.org/10.1109/jstars.2025.3544865.

[42]

A. Karpatne, G. Atluri, J. H. Faghmous, et al., “ Theory-Guided Data Science: A New Paradigm for Scientific Discovery From Data,” IEEE Transactions on Knowledge and Data Engineering 29, no. 10 (2017): 2318-2331, https://doi.org/10.1109/tkde.2017.2720168.

[43]

E. McGowan, V. Gawade, and W. G. Guo, “ A Physics-Informed Convolutional Neural Network With Custom Loss Functions for Porosity Prediction in Laser Metal Deposition,” Sensors 22, no. 2 (2022): 494, https://doi.org/10.3390/s22020494.

[44]

H. Li, L. Gou, H. Li, and Z. Liu, “ Physics-Guided Neural Network Model for Aeroengine Control System Sensor Fault Diagnosis Under Dynamic Conditions,” Aerospace 10, no. 7 (2023): 644, https://doi.org/10.3390/aerospace10070644.

[45]

B. Kapusuzoglu and S. Mahadevan, “ Physics-Informed and Hybrid Machine Learning in Additive Manufacturing: Application to Fused Filament Fabrication,” JOM 72, no. 12 (2020): 4695-4705, https://doi.org/10.1007/s11837-020-04438-4.

[46]

A. Daw, A. Karpatne, W. D. Watkins, J. S. Read, and V. Kumar, “ Physics-Guided Neural Networks (PGNN): An Application in Lake Temperature Modeling,” in Knowledge Guided Machine Learning (Chapman and Hall/CRC, 2022), 353-372.

[47]

W. Gu, S. Primatesta, and A. Rizzo, “ Physics-Informed Neural Network for Quadrotor Dynamical Modeling,” Robotics and Autonomous Systems 171 (2024): 104569, https://doi.org/10.1016/j.robot.2023.104569.

[48]

G. J. Holland, “ The Maximum Potential Intensity of Tropical Cyclones,” Journal of the Atmospheric Sciences 54, no. 21 (1997): 2519-2541, https://doi.org/10.1175/1520-0469(1997)054<2519:tmpiot>2.0.co;2.

[49]

J. A. Knaff and R. M. Zehr, “ Reexamination of Tropical Cyclone Wind-Pressure Relationships,” Weather and Forecasting 22, no. 1 (2007): 71-88, https://doi.org/10.1175/waf965.1.

[50]

G. Holland, “ A Revised Hurricane Pressure-Wind Model,” Monthly Weather Review 136, no. 9 (2007): 3432-3445, https://doi.org/10.1175/2008mwr2395.1.

[51]

J. A. Knaff, C. R. Sampson, and M. DeMaria, “ An Operational Statistical Typhoon Intensity Prediction Scheme for the Western North Pacific,” Weather and Forecasting 20, no. 4 (2005): 688-699, https://doi.org/10.1175/waf863.1.

[52]

Q. Jin, X. Fan, J. Liu, Z. Xue, and H. Jian, “ Using Extreme Gradient Boosting to Predict Changes in Tropical Cyclone Intensity over the Western North Pacific,” Atmosphere 10, no. 6 (2019): 341, https://doi.org/10.3390/atmos10060341.

[53]

N. M. Foumani, C. W. Tan, G. I. Webb, and M. Salehi, “ Improving Position Encoding of Transformers for Multivariate Time Series Classification,” Data Mining and Knowledge Discovery 38, no. 1 (2024): 22-48, https://doi.org/10.1007/s10618-023-00948-2.

[54]

B. Harper, Tropical Cyclone Parameter Estimation in the Australian Region , Vol. 83 (Systems Engineering Australia Pty Ltd for Woodside Energy Ltd, 2002).

[55]

X. Lu, H. Yu, M. Ying, et al., “ Western North Pacific Tropical Cyclone Database Created by the China Meteorological Administration,” Advances in Atmospheric Sciences 38, no. 4 (2021): 690-699, https://doi.org/10.1007/s00376-020-0211-7.

[56]

D. P. Dee, A. J. Uppala, P. Berrisford, et al., “ The Era-Interim Reanalysis: Configuration and Performance of the Data Assimilation System,” Quarterly Journal of the Royal Meteorological Society 137, no. 656 (2011): 553-597, https://doi.org/10.1002/qj.828.

[57]

J. H. Chu, “ A Regression Model for the Western North Pacific Tropical Cyclone Intensity Forecast,” Technical Report (1994).

[58]

T. Chen and C. Guestrin, “ Xgboost: A Scalable Tree Boosting System,” in Proceedings of the 22nd Acm Sigkdd International Conference on Knowledge Discovery and Data Mining (2016), 785-794.

[59]

C. Wang, G. Zheng, X. Li, Q. Xu, B. Liu, and J. Zhang, “ Tropical Cyclone Intensity Estimation From Geostationary Satellite Imagery Using Deep Convolutional Neural Networks,” IEEE Transactions on Geoscience and Remote Sensing 60 (2021): 1-16, https://doi.org/10.1109/tgrs.2021.3066299.

PDF (2603KB)

0

Accesses

0

Citation

Detail

Sections
Recommended

/

〈 〉