Iranian Journal of Geophysics

Iranian Journal of Geophysics

Impact of upper-level atmospheric data on the accuracy of predicting different ENSO phases using hybrid machine learning models in the southern provinces of Iran

Document Type : Research Article

Authors
Mahdi Malekian 1
Ali Akbar Sabziparvar 2
1 Ph.D. Candidate of Agricultural Meteorology, Department of Water Science Engineering, Faculty of Agriculture, Bu-Ali Sina University, Hamedan, Iran
2 Professor of Meteorology, Department of Water Science Engineering, Faculty of Agriculture, Bu-Ali Sina University, Hamedan, Iran
10.30499/ijg.2025.520217.1694
Abstract
The El Niño-Southern Oscillation (ENSO), as one of the world's most important Teleconnection phenomena, has widespread impacts on meteorological patterns and extreme events in various regions, including southern Iran. This study aims to enhance the predictive capabilities for ENSO phases, leveraging advanced machine learning techniques to address climate variability in vulnerable regions. This research investigates the impact of upper-air Parameters/variables on the prediction accuracy of ENSO phases using a hybrid ConvLSTM2D machine learning model in the southern provinces of Iran. The results indicate the key role of upper-level atmospheric data in improving the prediction of ENSO phases. Meteorological data including geopotential height, temperature, relative humidity, specific humidity, wind speed, and wind direction were collected from the ERA5 (ECMWF Reanalysis v5) reanalysis dataset at four pressure levels (1000, 850, 700, and 500 hPa) for the period 1994 to 2023. These atmospheric variables were meticulously processed to capture intricate Spatio-Temporal relationships critical for accurate climate forecasting. The proposed model was designed by combining ConvLSTM2D, LSTM, and Dense layers to extract complex Spatio-Temporal features, and the outcomes were evaluated using metrics such as R², MAE, and MSE. Results showed that inclusion of upper-air data (500 and 700 hPa), compared to near-surface levels, could significantly increase prediction accuracy. The highest model performance was noted at the 500 hPa level with a three month lead time (R² = 0.93, MAE = 0.14). This also shows the 500 hPa level of the atmosphere does not change its behavior, which in turns suggests why there is a delay during change in the atmosphere. This finding underscores the importance of mid-tropospheric dynamics in long-term climate predictions. Feature Importance analysis also showed the key role of dynamic variables (geopotential height and horizontal wind components) and thermodynamic variables (temperature and specific humidity) in the mid-level atmosphere. Furthermore, adding the suggest hybrid model in addressing non-linear relationships also makes it more applicable to complex climate systems. This research emphasizes the necessity of using a wide range of atmospheric data and optimizing the time horizon in climate predictions. The results can contribute to improving water resource management and reducing atmospheric hazards in southern Iran, although challenges such as model interpretability require further research. Future studies could explore integrating multi-source datasets to further refine predictive accuracy. In order to enhance prediction accuracy and improve performance in practical applications, it is recommended that hybrid models which incorporate Attention Mechanisms enabling focusing on salient features be developed.
Keywords
Forecast
teleconnection
artificial intelligence
hybrid models
feature importance

Subjects

Amini, M., Ghadami, M. R., Fathian, F., & Modarres, R. (2020). Teleconnections between oceanic–atmospheric indices and drought over Iran using quantile regressions. Taylor & Francis. https: //doi.org/10.1080/02626667.2020.1802029
Anochi, J. A., Almeida, V. A. D., & Velho, H. F. D. C. (2021). Machine Learning for Climate Precipitation Prediction Modeling over South America. Multidisciplinary Digital Publishing Institute. https: //doi.org/10.3390/rs13132468
Bonavita, M. & Laloyaux, P. (2020). Machine learning for model error inference and correction. Wiley. https: //doi.org/10.1029/2020ms002232
Cao, X., Yanan, G., Liu, B., Peng, K., Wang, G., & Mei, G. (2020). ENSO prediction based on long short-term memory (LSTM). None. https: //doi.org/10.1088/1757-899X/799/1/012035
Chapman, D., Cane, M. A., Henderson, N., Lee, D. E., & Chen, C. (2015). A Vector Autoregressive ENSO Prediction Model. American Meteorological Society. https: //doi.org/10.1175/jcli-d-15-0306.1
Chattopadhyay, A., Subel, A., & Hassanzadeh, P. (2020). Data-Driven Super-Parameterization Using Deep Learning: Experimentation With Multiscale Lorenz 96 Systems and Transfer Learning. Wiley. https: //doi.org/10.1029/2020ms002084
Chen, C., Cane, M. A., Henderson, N., Lee, D. E., Chapman, D., Kondrashov, D., & Chekroun,
 
M. D. (2015). Diversity, nonlinearity, seasonality, and memory effect in ENSO simulation and prediction using empirical model reduction. American Meteorological Society. https: //doi.org/10.1175/jcli-d-15-0372.1
Chen, L., Han, B., Wang, X., Zhao, J., Yang, W., & Yang, Z. (2023). Machine learning methods in weather and climate applications: a survey. Applied Sciences. https: //doi.org/10.20944/preprints202309.1764.v2
Cheng, Y., Tang, Y., Jackson, P. L., Chen, D., & Deng, Z. (2010). Ensemble Construction and Verification of the Probabilistic ENSO Prediction in the LDEO5 Model. American Meteorological Society. https: //doi.org/10.1175/2010jcli3453.1
Cifuentes, J., Marulanda, G., Bello, A., & Reneses, J. (2020). Air temperature forecasting using machine learning techniques: a review. Multidisciplinary Digital Publishing Institute. https: //doi.org/10.3390/en13164215
Donnelly, J., Daneshkhah, A., & Abolfathi, S. (2023). Forecasting global climate drivers using Gaussian processes and convolutional autoencoders. Elsevier BV. https: //doi.org/10.1016/j.engappai.2023.107536
Feng, J., Wang, L., Chen, W., Fong, S. K., & Leong, K. C. (2010). Different impacts of two types of pacific ocean warming on southeast Asian rainfall during boreal winter. American Geophysical Union. https: //doi.org/10.1029/2010jd014761
Gao, C., Wu, X., & Zhang, R. (2016). Testing a four-dimensional variational data assimilation method using an improved intermediate coupled model for ENSO analysis and prediction. Springer Science+Business Media. https: //doi.org/10.1007/s00376-016-5249-1
Gao, Y., Tang, Y., & Liu, T. (2023). Reducing Model Error Effects in El Niño–Southern Oscillation Prediction Using Ensemble Coupled Data Assimilation. Remote Sensing. https: //doi.org/10.3390/rs15030762
Geng, H. & Wang, T. (2021). Spatiotemporal Model Based on Deep Learning for ENSO Forecasts. Multidisciplinary Digital Publishing Institute. https: //doi.org/10.3390/atmos12070810
Ghasemi, A. (2019). Influence of northwest Indian Ocean sea surface temperature and El Niño–Southern Oscillation on the winter precipitation in Iran. IWA Publishing. https: //doi.org/10.2166/wcc.2019.274
Gholizadeh, M. H. (2015). Evaluation of Relation between Rainfall and El Nino Phenomena in Iran. https: //doi.org/10.51611/iars.irj.v5i1.2015.42
Ibebuchi, C. C. & Richman, M. B. (2024). Deep learning with autoencoders and LSTM for ENSO forecasting. https: //doi.org/10.1007/s00382-024-07180-8
Kumar, R., & Shrivastava, A. (2025). Integrating CNN-LSTM models for improved climate change forecasting. International Journal of Sciences and Innovation Engineering, 2(6), 394–400. https: //doi.org/10.70849/IJSCI
Luo, J., Ling, F., Ham, Y., & Kim, J. (2020). Seasonal-to-multiyear prediction of ENSO using machine deep learning. https: //doi.org/10.5194/egusphere-egu2020-21603
Mu, B., Qin, B., & Yuan, S. (2021). ENSO-ASC 1.0.0: ENSO deep learning forecast model with a multivariate air–sea coupler. Copernicus Publications. https: //doi.org/10.5194/gmd-14-6977-2021
Naz, F., She, L., Sinan, M., & Shao, J. (2024). Enhancing radar echo extrapolation by ConvLSTM2D for Precipitation nowcasting. Sensors, 24(2), 459. https: //doi.org/10.3390/s24020459
Nikraftar, Z. & Khaniani, A. S. (2018). Assessing the impact of cold and warm ENSO on drought over Iran. None. https: //doi.org/10.22059/EOGE.2018.257714.1022
NOAA Climate.gov. (2009). Climate variability: Oceanic Niño Index. NOAA Climate.gov. https: //www.climate.gov/news-features/understanding-climate/climate-variability-oceanic-nino-index
Qin, B., Yang, Z., Mu, M., Wei, Y., Cui, Y., Fang, X., Dai, G., & Yuan, S. (2024). The first kind of predictability problem of El Niño predictions in a multivariate coupled data-driven model. Quarterly Journal of the Royal Meteorological Society. https: //doi.org/10.1002/qj.4882
Sabziparvar, A. A., Mirmasoudi, S. H., Tabari, H., Nazemosadat, M. J., & Maryanaji, Z. (2010). ENSO teleconnection impacts on reference evapotranspiration variability in some warm climates of Iran. International Journal of Climatology , 31 (1), DOI: 10.1002/joc.2187
Santos, M. A. D. C., Vega Oliveros, D. A., Zhao, L., & Berton, L. (2020). Classifying El Nino-southern oscillation combining network science and machine learning. Institute of Electrical and Electronics Engineers. https: //doi.org/10.1109/access.2020.2982035
Santos, M. A. D. C., VegaOliveros, D. A., Zhao, L., & Berton, L. (2020). Classifying El Niño-Southern Oscillation Combining Network Science and Machine Learning. Institute of Electrical and Electronics Engineers. https: //doi.org/10.1109/access.2020.2982035
Slater, L., Arnal, L., Boucher, M., Chang, A., Moulds, S., Murphy, C., Nearing, G., Shalev, G., Shen, C., Speight, L., Villarini, G., Wilby, R., Wood, A., & Zappa, M. (2023). Hybrid forecasting: blending climate predictions with AI models. Hydrology and Earth System Sciences. https: //doi.org/10.5194/hess-27-1865-2023
Tang, Y., Zhang, R., Liu, T., Duan, W., Yang, D., Zheng, F., Ren, H., Lian, T., Gao, C., Chen, D., & Mu, M. (2018). Progress in ENSO prediction and predictability study. Oxford University Press. https: //doi.org/10.1093/nsr/nwy105
VicenteSerrano, S. M., LpezMoreno, J. I., Gimeno, L., Nieto, R., MornTejeda, E., LorenzoLacruz, J., Beguera, S., & Azorn-Molina, C. (2011). A multiscalar global evaluation of the impact of ENSO on droughts. American Geophysical Union. https: //doi.org/10.1029/2011jd016039
Wang, C., Ren, B., Li, G., Zheng, J., Jiang, L., & Zhang, Z. (2023). Why could ENSO directly affect the occurrence frequency of arctic daily warming events after the late 1970s?. Environmental Research Letters. https: //doi.org/10.1088/1748-9326/acb06f
Wang, L., Ren, H., Zhu, J., & Huang, B. (2020). Improving prediction of two ENSO types using a multi-model ensemble based on stepwise pattern projection model. https: //doi.org/10.1007/s00382-020-05160-2
Yeh, S., Cai, W., Min, S., McPhaden, M. J., Dommenget, D., Dewitte, B., Collins, M., Ashok, K., An, S., Yim, B., & Kug, J. (2018). ENSO atmospheric teleconnections and their response to greenhouse gas forcing. Wiley. https: //doi.org/10.1002/2017rg000568
Yin, Z., Zhou, B., Duan, M., Chen, H., & Wang, H. (2023). Climate Extremes become increasingly fierce in china. Elsevier BV. https: //doi.org/10.1016/j.xinn.2023.100406
Zhai, P., Yu, R., Guo, Y., Li, Q., Ren, X., Wang, Y., Xu, W., Liu, Y., & Ding, Y. (2016). The Strong El Nino of 2015/16 and Its Dominant Impacts on Global and China's Climate. https: //doi.org/10.1007/s13351-016-6101-3
Zhang, W., Li, H., Stuecker, M. F., Jin, F., & Turner, A. G. (2015). A new understanding of El Nino impact over east Asia: dominance of the ENSO combination mode. American Meteorological Society. https: //doi.org/10.1175/jcli-d-15-0104.1
Zheng, Z., Hu, Z., & LHeureux, M. (2016). Predictable components of ENSO evolution in real-time multi-model predictions. Nature Portfolio. https: //doi.org/10.1038/srep35909
Zhou, L. & Zhang, R. (2023). A self-attention–based neural network for three-dimensional multivariate modeling and its skillful ENSO predictions. Science Advances. https: //doi.org/10.1126/sciadv.adf2827
Ziv, S. Z., Garfinkel, C. I., Davis, S., & Banerjee, A. (2022). The roles of the Quasi-Biennial Oscillation and El Niño for entry stratospheric water vapor in observations and coupled chemistry–ocean CCMI and CMIP6 models. Copernicus Publications. https: //doi.org/10.5194/acp-22-7523-2022
 
 
Iranian Journal of Geophysics
Volume 19, Issue 4
September and October 2025
Pages 175-199