Enhanced Multimodal Deep Learning Framework for Accurate El Niño and La Niña Forecasting

Document Type : Original Article

Authors

1 Department of Computer Engineering, Shiraz University of Technology, Shiraz, Iran

2 Jacksonville University, Jacksonville, Florida

3 Department of Computer, Sharif University of Technology, Tehran, Iran

4 Department of Civil Engineering, Faculty of Engineering, University of Zanjan, Zanjan, Iran

10.48308/ijce.2025.241830.1016

Abstract

Reliable long-lead forecasting of the El Niño–Southern Oscillation (ENSO) remains a long-standing challenge in climate science. The previously developed Multimodal ENSO Forecast (MEF) model uses 80 ensemble predictions by two independent deep learning modules: a 3D Convolutional Neural Network (3DCNN) and a time-series Informer module. In their approach, outputs of the two modules are combined using a weighting strategy wherein one is prioritized over the other as a function of global performance. Separate weighting or testing of individual ensemble members did not occur, however, which may have limited the model to optimize the use of high-performing but spread-out forecasts. In this study, we propose a novel framework that employs graph-based analysis to implicitly find the best ensemble. By constructing an undirected graph whose vertices are ensemble outputs and whose weights on edges measure similarity, we identify and cluster the best prediction. This method improves the forecast skill through noise removal and emphasis on ensemble coherence. Interestingly, our graph-based selection shows robust statistical characteristics among top performers, offering new ensemble behavior insights. The approach is model-agnostic too, suggesting that it can be applied directly to other forecasting models with gargantuan ensemble outputs, such as statistical, and physical models.

Keywords

References

Ali, M., Richter, S., & Ertürk, A. (2025). Graph neural networks learn emergent tissue properties from spatial molecular profiles. Nature Communications, 16, 8419. https://doi.org/10.1038/s41467-025-63758-8
Andrej, K. (2016). Generative models. OpenAI.
Blöschl, G., Hall, J., Viglione, A., et al. (2019). The changing climate both increases and decreases European river floods. Nature, 573, 108–111.
Cai, W., Borlace, S., & Lengaigne, M. (2014). Increasing frequency of extreme El Niño events due to greenhouse warming. Nature Climate Change, 4, 111–116. https://doi.org/10.1038/nclimate2100
Cai, W., McPhaden, M. J., & Grimm, A. M. (2020). Climate impacts of the El Niño–Southern Oscillation on South America. Nature Reviews Earth & Environment, 1, 215–231. https://doi.org/10.1038/s43017-020-0040-3
Castangia, M., Grajales, L., Aliberti, A., & Rossi, C. (2023). Transformer neural networks for interpretable flood forecasting. Environmental Modelling & Software, 160, 105581.
Erdmann, M., Glombitza, J., & Quast, T. (2019). Precise simulation of electromagnetic calorimeter showers using a Wasserstein generative adversarial network. Computing and Software for Big Science, 3(4). https://arxiv.org/abs/1807.01954
Ganji, S., Labibzadeh, A. R., Hassani, A., & Naisipour, M. (2025). Leveraging GNN to enhance the MEF method in predicting ENSO. arXiv preprint arXiv:2508.07410. https://doi.org/10.48550/arXiv.2508.07410
Ganji, S., Naisipour, M., Hassani, A., & Adib, A. (2025). Distillation of CNN ensemble results for enhanced long-term prediction of the ENSO phenomenon. arXiv preprint arXiv:2509.06227. https://doi.org/10.48550/arXiv.2509.06227
Gao, C., & Zhang, R. H. (2017). The roles of atmospheric wind and entrained water temperature in the secondary cooling of the 2010–12 La Niña event. Climate Dynamics, 48, 597–617.
Goodfellow, I. J., et al. (2004). Elasto-plastic fracturing model for wellbore stability using non-penetrating fluids. ARXIV preprint arXiv:1406.2661. https://doi.org/10.48550/arXiv.1406.2661
Ham, Y., Kim, J., & Luo, J. J. (2019). Deep learning for multi-year ENSO forecasts. Nature, 573, 568–572.
Hassani, A., & Javadi, M. (2025a). When AI replaces developers: Graph neural networks reveal paths from computer to management occupations. Preprints, 2025, 2025120212. https://doi.org/10.20944/preprints2025120212.v1
Hassani, A., & Javadi, M. (2025b). Leveraging artificial intelligence to study and forecast its impacts on employment management and wages in China’s key cities. Spectrum of Engineering and Management Sciences, 1, 1–12. https://doi.org/10.31181/sems4120266oh
Hassani, A., Javadi, M., & Naisipour, M. (2025). The time series Informer model for stock market prediction. Preprints, 2025101842. https://doi.org/10.20944/preprints202510.1842.v1
Ho, J., & Ermon, S. (2016). Generative adversarial imitation learning. Advances in Neural Information Processing Systems, 29, 4565–4573. https://arxiv.org/abs/1606.03476
Isola, P., Zhu, J., Zhou, T., & Efros, A. A. (2017). Image-to-image translation with conditional adversarial nets. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 1125–1134.
Labibzadeh, M., et al. (2015). Efficiency test of the discrete least-squares meshless method in solving heat conduction problems using error estimation. Sharif Civil Engineering, 31(2), 31–40. https://sid.ir/paper/127886/en
L’Heureux, M. L., Levine, A. F., Newman, M., Ganter, C., Luo, J. J., Tippett, M. K., et al. (2020). ENSO prediction. In El Niño–Southern Oscillation in a Changing Climate (pp. 227–246). John Wiley & Sons. https://doi.org/10.1002/9781119548164.ch10
Luc, P., et al. (2016). Semantic segmentation using adversarial networks. Advances in Neural Information Processing Systems, 29. https://arxiv.org/abs/1611.08408
Luo, J. J., Masson, S., Behera, S. K., & Yamagata, T. (2008). Extended ENSO predictions using a fully coupled ocean–atmosphere model. Journal of Climate, 21, 84–93.
McPhaden, M. J., Zebiak, S. E., & Glantz, M. H. (2006). ENSO as an integrating concept in Earth science. Science, 314, 1740–1745.
Moazezi Khah, Tehran, A., Hassani, A., Mohajer, S., Darvishan, S., Shafiesabet, A., & Tashakkori, A. (2025). The impact of financial literacy on financial behavior and financial resilience with the mediating role of financial self-efficacy. International Journal of Industrial Engineering and Operational Research, 7(2), 38–55.
Mohamed, S., & Lakshminarayanan, B. (2016). Learning implicit generative models. arXiv preprint arXiv:1610.03483.
Mustafa, M., et al. (2019). CosmoGAN: Creating high-fidelity weak lensing convergence maps using generative adversarial networks. Computational Astrophysics and Cosmology, 6(1), 1. https://arxiv.org/abs/1706.02390
Naisipour, M., Afshar, M. H., Hassani, B., & Zeinali, M. (2009). An error indicator for two-dimensional elasticity problems in the discrete least-squares meshless method. Proceedings of the 8th International Congress on Civil Engineering, Shiraz, Iran. https://civilica.com/doc/68292
Naisipour, M., Saeedpanah, I., & Adib, A. (2025a). Multimodal deep learning for two-year ENSO forecast. Water Resources Management, 39, 3745–3775. https://doi.org/10.1007/s11269-025-04128-3
Naisipour, M., Saeedpanah, I., & Adib, A. (2025b). Metrics matter: A deep assessment of deep learning CNN method for ENSO forecast. Atmospheric Research, 108545. https://doi.org/10.1016/j.atmosres.2025.108545
Naisipour, M., Saeedpanah, I., & Adib, A. (2024). Novel deep learning method for forecasting ENSO. Journal of Hydraulic Structures, 14, 25–42. https://doi.org/10.22055/jhs.2024.47965.1322
Nicolas, C., Francisco, M., Gabriel, S., & Nicolas, U. A. (2020). End-to-end object detection with transformers. In Proceedings of the European Conference on Computer Vision (pp. 213–229).
Paganini, M., de Oliveira, L., & Nachman, B. (2017). Learning particle physics by example: Location-aware generative adversarial networks for physics synthesis. Computing and Software for Big Science, 1(4). https://doi.org/10.1007/s41781-017-0004-6
Reiser, P., Neubert, M., & Eberhard, A. (2022). Graph neural networks for materials science and chemistry. Nature Communications Materials, 3, 93. https://doi.org/10.1038/s43246-022-00315-6
Rives, A., Meier, J., Sercu, T., & Goyal, S. T. (2016). Biological structure and function emerge from scaling unsupervised learning to 250 million protein sequences. Proceedings of the National Academy of Sciences, 118(15), e2016239118.
Salamani, D., et al. (2018). Deep generative models for fast shower simulation in ATLAS. In IEEE 14th International Conference on e-Science (pp. 348–348).
Salimans, T., et al. (2016). Improved techniques for training GANs. arXiv preprint arXiv:1606.03498.
Scaife, A. A., et al. (2010). Impact of ENSO on European climate. In the ECMWF Seminar on Predictability in the European and Atlantic Regions, ECMWF, Shinfield Park, Reading.
Schawinski, K., et al. (2017). Generative adversarial networks recover features in astrophysical images of galaxies beyond the deconvolution limit. Monthly Notices of the Royal Astronomical Society: Letters, 467(1), L110–L114.
Schwaller, P., Laino, T., Gaudin, T., Bolgar, P., Hunter, C., Bekas, C., & Lee, A. (2019). Molecular transformer: A model for uncertainty-calibrated chemical reaction prediction. ACS Central Science, 5(9), 1572–1583. https://doi.org/10.1021/acscentsci.9b00576
Taccari, M. L., Wang, H., & Nuttall, J. E. (2024). Spatial–temporal graph neural networks for groundwater data. Scientific Reports, 14, 24564. https://doi.org/10.1038/s41598-024-75385-2
Talaat, F. M., & Gamel, S. A. (2022). RL-based hyper-parameter optimization algorithm (ROA) for a convolutional neural network. Journal of Ambient Intelligence and Humanized Computing, 13(10), 4637–4651.
Vaswani, A., Shazeer, N., & Parmar, N. (2017). Attention is all you need. In Advances in Neural Information Processing Systems (pp. 5998–6008).
Wang, G. G., Cheng, H., Zhang, Y., & Yu, H. (2023). ENSO analysis and prediction using deep learning: A review. Neurocomputing, 520, 216–229.
Wu, S. (2025). Application of an artificial intelligence graph convolutional network in classroom grade evaluation. Scientific Reports, 15, 32044. https://doi.org/10.1038/s41598-025-17903-4
Ye, F., Hu, J., Huang, T., You, L., Weng, B., & Gao, J. (2022). Transformer for El Niño–Southern Oscillation prediction. IEEE Geoscience and Remote Sensing Letters, 19, 1–5.
Zhou, H., Zhang, S., & et al. (2021). Informer: Beyond efficient transformer for long-sequence time-series forecasting. In Proceedings of the AAAI Conference on Artificial Intelligence (Vol. 35, pp. 11106–11115).
Interdisciplinary Journal of Civil Engineering
Volume 1, Issue 3 - Serial Number 3
September 2025
Pages 256-281

Files

Share

How to cite

Statistics