Machine Learning Models for Climate Prediction: A Comparative Study with Classical Statistical Methods
Keywords:
ARIMA, Climate change, Climate prediction, Deep learning, LSTM, Machine learning, Statistical models, Time series forecastingAbstract
This work investigates the role of advanced modeling techniques in improving the accuracy of climate prediction, which is vital for understanding environmental change and guiding mitigation efforts. Conventional statistical models, including Autoregressive Integrated Moving Average (ARIMA) and Seasonal ARIMA (SARIMA), have been extensively used due to their clarity, efficiency, and effectiveness in handling linear and stationary time series data. However, the growing complexity of climate systems has encouraged the adoption of machine learning approaches that can better capture nonlinear relationships and long-term dependencies. Techniques such as Long Short-Term Memory (LSTM) networks, Convolutional Neural Networks (CNN), and ensemble learning models have shown strong potential in extracting meaningful patterns from large and complex climate datasets. In this study, a detailed comparison is conducted between traditional statistical methods and modern machine learning models for forecasting key variables such as temperature, rainfall, and atmospheric CO₂ concentration. Performance is assessed using widely accepted evaluation metrics, including Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and the coefficient of determination (R²), along with considerations of computational requirements and model interpretability. The findings suggest that machine learning models, especially LSTM and hybrid approaches, generally provide more accurate predictions for nonlinear and large-scale data. Nevertheless, statistical models remain reliable for short-term forecasts and relatively stable datasets, highlighting the value of combining both approaches for enhanced climate prediction.
References
P. B. Gibson, W. E. Chapman, A. Altinok, L. Delle Monache, M. J. DeFlorio, and D. E. Waliser, “Training machine learning models on climate model output yields skillful interpretable seasonal precipitation forecasts,” Communications Earth & Environment, vol. 2, no. 1, pp. 1–13, Aug. 2021.
O. Watt‐Meyer et al., “Correcting Weather and Climate Models by Machine Learning Nudged Historical Simulations,” Geophysical Research Letters, vol. 48, no. 15, pp. 1-10, Aug. 2021.
C. Song, X. Chen, P. Wu, and H. Jin, “Combining time varying filtering based empirical mode decomposition and machine learning to predict precipitation from nonlinear series,” Journal of Hydrology, vol. 603, pp. 126914, Dec. 2021.
L. Chen et al., “A machine learning model that outperforms conventional global subseasonal forecast models,” Nature Communications, vol. 15, no. 1, Jul. 2024.
C. C. Ibebuchi and M. B. Richman, “Deep learning with autoencoders and LSTM for ENSO forecasting,” Climate Dynamics, vol. 62, pp. 5683–5697, Mar. 2024.
H. Zhang, J. Chen, Y. Wang, J. Han, and Y. Xu, “Improving 2 m temperature forecasts of numerical weather prediction through a machine learning-based Bayesian model,” Meteorology and Atmospheric Physics, vol. 137, no. 1, pp. 1-17 Jan. 2025.
X. Zhong et al., “FuXi-ENS: A machine learning model for efficient and accurate ensemble weather prediction,” Science Advances, vol. 11, no. 44, Oct. 2025.
M. M. Uddin Quresh, A. B. Ahmed, A. Dulmini, M. Mahboob, and R. Rois, “Developing a seasonal-adjusted machine-learning-based hybrid time series model to forecast heatwave warning,” Scientific Reports, vol. 15, no. 1, pp. 1-17, Mar. 2025.
M. S. Islam et al., “Explainable deep learning for rainfall prediction: A CNN-XGBoost hybrid approach in the northern region of Bangladesh,” Neural Computing and Applications, vol. 37, pp. 28125–28160, Oct. 2025.
F. M. A and E. A. Barnes, “Multi-Year-to-Decadal Temperature Prediction using a Machine Learning Model-Analog Framework,” arXiv.org, pp. 1-14, 2025.
M. C. Agana Navarro, G. Li, T. Wolf, and M. Pérez-Ortiz, “Do machine learning climate models work in changing climate dynamics?” arXiv.org, pp. 8, 2025.
R. Yang et al., “Interpretable Machine Learning for Weather and Climate Prediction: A Survey,” arXiv.org, pp. 26, 2024.
M. Panja et al., “Forecasting precipitation in the Arctic using probabilistic machine learning informed by causal climate drivers,” arXiv.org, vol. 35, no. 11, 2025.
