AccScience Publishing / NSCE / Online First / DOI: 10.36922/NSCE026250025
Submit to NSCE
Cite this article
6
Download
159
Views
Related Info Links
Google Scholar
More by Authors Links
Mohammed Gadafi Tamimu
Journal Browser
Volume | Year
Issue
Search
News and Announcements
View All
RESEARCH ARTICLE

PHyST-Net: A hybrid seasonal-temporal deep learning framework for climate forecasting and extreme rainfall risk assessment in semi-arid Ghana

Toufic Seini1 ,  Daniel Makinde Oluwole2 ,  Ibrahim Yakubu Seini1 ,  Mohammed Gadafi Tamimu3,4*
Show Less
1 Department of Mathematics, Faculty of Physical Sciences, University for Development Studies, Tamale, Northern Region , Ghana
2 Faculty of Military Science, Stellenbosch University, Stellenbosch, Western Cape , South Africa
3 College of International Communication and Cooperation, Sichuan University of Culture and Arts, Mianyang, Sichuan , China
4 School of Economics and Management, Center for West African Studies, University of Electronic Science and Technology of China, Chengdu, Sichuan , China
NSCE, 026250025 https://doi.org/10.36922/NSCE026250025
Received: 17 June 2026 | Revised: 20 August 2026 | Accepted: 1 September 2026 | Published online: 15 September 2026
© 2026 by the Author(s). This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution -Noncommercial 4.0 International License (CC-by the license) ( https://creativecommons.org/licenses/by-nc/4.0/ )
Download PDF
XML
Cite
Abstract

Accurate forecasting of temperature and rainfall is critical for climate adaptation, agricultural planning, water-resource management, and disaster risk reduction, particularly in semi-arid regions where climate variability poses significant socio-economic challenges. Climate prediction in data-limited regions remains difficult due to nonlinear temporal interactions, seasonal variability, and the rarity of extreme rainfall events. This study proposes the physics-aware spatio-temporal network (PHyST-Net), a hybrid seasonal-temporal deep learning framework that integrates 1D convolutional neural networks (CNNs) and bidirectional long short-term memory (LSTM) networks for multivariate climate forecasting and extreme rainfall risk assessment in semi-arid Ghana. Unlike conventional statistical models, PHyST-Net combines local temporal feature extraction through CNN layers with long-range sequential dependency modeling through bidirectional LSTM layers. It incorporates climate-informed temporal variables, including seasonal month and long-term year indicators, and employs various diagnostic tools, uncertainty estimation, and gradient-based saliency analysis to improve model interpretation. The framework is evaluated using monthly temperature and rainfall observations obtained from the Ghana Meteorological Agency. A chronological 70%–15%–15% training–validation–testing strategy was adopted, using a 36-month historical window for prediction. Experimental results show that PHyST-Net outperformed the persistence, autoregressive integrated moving average, and standalone LSTM models. For temperature forecasting, PHyST-Net achieved a root mean square error (RMSE) of 3.64 °C and mean absolute error (MAE) of 3.08 °C, while rainfall forecasting achieved an RMSE of 19.7 mm and MAE of 14.9 mm. Extreme rainfall detection was formulated as a rare-event classification problem using a statistical rainfall threshold. The classifier provided probabilistic estimates of extreme rainfall occurrence, while calibration, precision–recall analysis, and probability-based interpretation were used because extreme rainfall events are highly imbalanced. Overall, PHyST-Net offers a robust and interpretable framework for climate forecasting and extreme-event risk assessment in data-limited regions.

Keywords
Climate forecasting
Convolutional neural network–bidirectional long short-term memory
Extreme rainfall detection
Time-series forecasting
Deep learning
Climate risk assessment
Ghana
Funding
None.
Conflict of interest
The authors declare they have no competing interests.
References
  1. Bruzzone OA, Perri DV, Easdale MH. Vegetation responses to variations in climate: A combined ordinary differential equation and sequential Monte Carlo estimation approach. Ecol Inform. 2023;73:101913. https://doi.org/10.1016/j.ecoinf.2022.101913.
  2. Singh P. Crop models for assessing impact and adaptation options under climate change. J Agrometeorol. 2023;25(1):18-33. https://doi.org/10.54386/jam.v25i1.1969
  3. Nicholson SE. The West African Sahel: A review of recent studies on the rainfall regime and its interannual variability. ISRN Meteorol. 2013:1-32. https://doi.org/10.1155/2013/453521
  4. Antwi-Agyei P, Fraser EDG, Dougill AJ, Stringer LC, Simelton E. Mapping the vulnerability of crop production to drought in Ghana using rainfall, yield and socioeconomic data. Appl Geogr. 2012;32(2):324-334. https://doi.org/10.1016/j.apgeog.2011.06.010
  5. Giorgi F, Mearns L. Introduction to special section: Regional climate modeling revisited. J Geophys Res Atmos. 1999;104(D6):6335-6352. https://doi.org/10.1029/98JD02072
  6. Palmer T. Towards the probabilistic earth-system simulator: A vision for the future of climate and weather prediction. Q J R Meteorol Soc. 2012;138(665):841-861. https://doi.org/10.1002/qj.1923
  7. Maraun D, Wetterhall F, Ireson AM, et al. Precipitation downscaling under climate change: Recent developments to bridge the gap between dynamical models and the end user. Rev Geophys. 2010;48(3). https://doi.org/10.1029/2009rg000314
  8. Shi X, Chen Z, Wang H, Yeung DT, Wong W, Woo W. Convolutional LSTM network: A machine learning approach for precipitation nowcasting. In: Advances in Neural Information Processing Systems 28. Red Hook, NY, USA: Curran Associates, Inc.;2015:802-810.
  9. Ojo OS, Ogunjo ST. Machine learning models for prediction of rainfall over Nigeria. Sci Afr. 2022;16:e01246. https://doi.org/10.1016/j.sciaf.2022.e01246
  10. Adjei CO, Tian W, Onzo BM, Chen S, Kedjanyi EAG, Darteh OF. Rainfall forecasting in sub-Sahara Africa-Ghana using LSTM deep learning approach. IJERTORG. 2021;10(3):464-470. https://doi.org/10.5281/zenodo.18629903
  11. Jing L, Gulcehre C, Peurifoy J, et al. Gated Orthogonal Recurrent Units: On Learning to Forget. Neural Comput. 2019;31(4):765-783. https://doi.org/10.1162/neco_a_01174
  12. Sheriff R, Meer MS, Tariq A. Remote sensing-based assessment of vegetation and land surface temperature effects on nitrogen dioxide concentrations in Chennai and Bengaluru using Google Earth Engine. Eng Rep. 2026;8(1). https://doi.org/10.1002/eng2.70436
  13. Tamimu MG. Sustainable AI-driven predictive modeling in agricultural supply chains: AI, IoT, and blockchain integration for climate adaptation in Ghana. IJEBM. 2026;3(3). https://doi.org/10.33552/ijebm.2026.03.000564
  14. Seneviratne SI, Nicholls N, Easterling D, et al. Changes in climate extremes and their impacts on the natural physical environment. In: Field CB, Barros V, Stocker TF, Dahe Q, eds. Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation. Cambridge, UK: Cambridge University Press; 2012:109-230. https://doi.org/10.1017/cbo9781139177245.006
  15. Kasei R, Diekkrüger B, Leemhuis C. Drought frequency in the Volta Basin of West Africa. Sustain Sci. 2009;5(1):89-97. https://doi.org/10.1007/s11625-009-0101-5
  16. Tamimu MG, Liang D. Sustainable mining development in Ghana: An integrated AHP-TOPSIS-Manhattan distance approach. Resour Policy. 2025;110:105757. https://doi.org/10.1016/j.resourpol.2025.105757
  17. Tamimu MG, Zhao S, Xu Q, Zhang J. Attention-based deep learning hybrid model for cash crop price forecasting: Evidence from global futures markets with implications for West Africa. Appl Sci. 2026;16(3):1600. https://doi.org/10.3390/app16031600
  18. Senjaliya J, Vaghasia V, Shah SM. A comprehensive review of machine learning and deep learning approaches for rainfall forecasting: current progress, challenges, and future directions. Theor Appl Climatol. 2026;157(3):177. https://doi.org/10.1007/s00704-026-06051-y
  19. Vandal T, Kodra E, Ganguly S, Michaelis A, Nemani R, Ganguly AR. Generating high resolution climate change projections through single image super-resolution: An Abridged Version. In: Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence. International Joint Conferences on Artificial Intelligence Organization; 2018:5389-5393. https://doi.org/10.24963/ijcai.2018/759
  20. Reichstein M, Camps-Valls G, Stevens B, et al. Deep learning and process understanding for data-driven Earth system science. Nature. 2019;566(7743):195-204. https://doi.org/10.1038/s41586-019-0912-1
  21. Raissi M, Perdikaris P, Karniadakis GE. Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. J Comput Phys. 2019;378:686-707. https://doi.org/10.1016/j.jcp.2018.10.045
  22. Yang Y, Perdikaris P. Adversarial uncertainty quantification in physics-informed neural networks. J Comput Phys. 2019;394:136-152. https://doi.org/10.1016/j.jcp.2019.05.027
  23. Chen Y, Yang Q, Chen Z, Yan C, Zeng S, Dai M. Physics-informed neural networks for building thermal modeling and demand response control. Build Environ. 2023;234:110149. https://doi.org/10.1016/j.buildenv.2023.110149
  24. Lu L, Meng X, Mao Z, Karniadakis GE. DeepXDE: A deep learning library for solving differential equations. SIAM Rev. 2021;63(1):208-228. https://doi.org/10.1137/19m1274067
  25. Bonavita M, Geer A, Laloyaux P, Massart S, Chrust M. Data assimilation or machine learning? ECMWF Newsl. 2021;167:17-22. https://www.ecmwf.int/en/newsletter/167/meteorology/data-assimilation-or-machine-learning
  26. Gadafi TM, Mohammed AMA, Ma J, Muhammad SR, Darko AP, Liang D. BiLSTM-based climate and agricultural supply chain resilience modeling using time series forecasting in Ghana. In: Proceedings of the 2nd International Conference on Image Processing, Machine Learning, and Pattern Recognition. New York, NY, USA: Association for Computing Machinery; 2025:225-230. https://doi.org/10.1145/3759928.3759967
  27. Yamba EI, Fink AH, Badu K, Asare EO, Tompkins AM, Amekudzi LK. Climate drivers of malaria transmission seasonality and their relative importance in sub-Saharan Africa. GeoHealth. 2023;7(2):e2022GH000698. https://doi.org/10.1029/2022gh000698
  28. Nwazelibe VE, Agbasi JC, Arunachalam KP, Usman AG, Abba SI, Egbueri JC. Monthly rainfall forecasting for southern Nigeria using 2D deep learning and implications for climate adaptation and sustainability: A CNN-LSTM+Attention approach. Phys Chem Earth Parts A/B/C. 2026;144:104577. https://doi.org/10.1016/j.pce.2026.104577
  29. Karpatne A, Atluri G, Faghmous JH, et al. Theory-guided data science: A new paradigm for scientific discovery from data. IEEE Trans Knowl Data Eng. 2017;29(10):2318-2331. https://doi.org/10.1109/tkde.2017.2720168
  30. Pradhan R, Aygun RS, Maskey M, Ramachandran R, Cecil DJ. Tropical cyclone intensity estimation using a deep convolutional neural network. IEEE Trans Image Process. 2018;27(2):692-702. https://doi.org/10.1109/tip.2017.2766358
  31. Aghenda M, Labbaci A, Bouchaou L, et al. Flood prediction using machine learning and deep learning models: a systematic review. Med Geosc Rev. 2025;7(4):1149-1167. https://doi.org/10.1007/s42990-025-00201-6
  32. Stengel K, Glaws A, Hettinger D, King RN. Adversarial super-resolution of climatological wind and solar data. Proc Natl Acad Sci USA. 2020;117(29):16805-16815. https://doi.org/10.1073/pnas.1918964117
  33. Mwendera EJ. Rural water supply and sanitation (RWSS) coverage in Swaziland: Toward achieving millennium development goals. Phys Chem Earth Parts A/B/C. 2006;31(15-16):681-689. https://doi.org/10.1016/j.pce.2006.08.040
  34. Asare-Kyei D, Forkuor G, Venus V. Modeling flood hazard zones at the sub-district level with the rational model integrated with GIS and remote sensing approaches. Water. 2015;7(7):3531-3564. https://doi.org/10.3390/w7073531
  35. Owusu K, Waylen PR. The changing rainy season climatology of mid-Ghana. Theor Appl Climatol. 2012;112(3-4):419-430. https://doi.org/10.1007/s00704-012-0736-5
  36. Afriyie R, Bessah E, Attoh EMNAN, et al. Climate information services and farm-level decision-making in the Pra River Basin of Ghana. Discov Environ. 2026;4:335. https://doi.org/10.1007/s44274-026-00655-x
  37. Dunee D, Dagadu PP, Ayimadu ET, et al. Impact of climate change on water resources and its implications on biodiversity, flood disasters, and food security in Ghana: a review. GeoJournal. 2025;90(3). https://doi.org/10.1007/s10708-025-11348-y
  38. Unami K, Abagale FK, Yangyuoru M, Badiul Alam AHM, Kranjac-Berisavljevic G. A stochastic differential equation model for assessing drought and flood risks. Stoch Environ Res Risk Assess. 2009;24(5):725-733. https://doi.org/10.1007/s00477-009-0359-2
Share
Back to top
Nonlinear Science and Control Engineering, Published by AccScience Publishing
We use cookies on our website to ensure you get the best experience.
Read more about our cookies here
Accept