Forecasting Monthly Air Temperature in Northeastern Libya Using Regularised Regression Models.
Keywords:
Time series forecasting, Temperature prediction, Regularized regression, Fourier seasonality, NASA POWER, Climate modeling, Lag-based modelsAbstract
This study addresses the need for accurate temperature forecasting in semi-arid Mediterranean regions to support climate adaptation, energy management. The objective is to develop a structured and computationally efficient forecasting framework for monthly air temperature in northeastern Libya (32.1167°N, 20.0667°E) using NASA POWER data spanning 2000–2024. To reduce high-frequency variability and enhance signal stability, daily meteorological observations were aggregated into monthly averages. Forecasting was formulated as a supervised learning problem using engineered lag features of the target variable alongside exogenous atmospheric drivers, including relative humidity (RH2M), wind speed (WS2M), and surface solar radiation (ALLSKY_SFC_SW_DWN). Seasonal dynamics were explicitly modeled using Fourier terms to capture annual periodicity. Three regularised regression models—Ridge, Lasso, and Elastic Net—were evaluated under a strict chronological framework comprising training (2000–2018), validation (2019–2021), and independent testing (2022–2024), with an expanding-window walk-forward strategy for one- and three-month-ahead forecasting. The results demonstrate that the Lasso-based model achieves the best overall performance, with a Root Mean Square Error (RMSE) of approximately 1.02°C for one-month-ahead forecasts and comparable accuracy for three-month-ahead predictions. While the seasonal naïve model remains competitive due to strong annual periodicity, the proposed framework consistently outperforms naïve benchmarks. Diebold–Mariano statistical tests confirm that improvements over the naïve persistence model are statistically significant (p < 0.001), although differences relative to the seasonal naïve model are not statistically significant. Regularised regression with structured lag features and explicit seasonal representation provides a robust, interpretable, and computationally efficient alternative to more complex nonlinear forecasting models for regional climate applications.
References
[1] M. A. Farahani et al., “Time-series forecasting in smart manufacturing systems: An experimental evaluation of state-of-the-art algorithms,” Reliable Computing in Industrial Manufacturing, 2025.
[2] X. Zhang, “Research on time series forecasting method based on ARIMA–ZKF prediction model,” Sustainability, vol. 17, no. 7, p. 2993, 2025.
[3] R. D. T. de Lima et al., “Time series forecasting with exogenous variables: A literature review and comparative perspectives,” Principia (IFPB), 2025.
[4] E. H. Houssein, “Artificial intelligence and classical statistical models for time series forecasting,” Journal of Big Data Analytics, 2025.
[5] H. Mahajan, “Comparative study of ARIMA, LSTM, and Prophet models,” Preprints, 2026.
[6] F. Yao et al., “CaReTS: A multi-task framework unifying classification and regression,” arXiv preprint.
[7] D. Shah and M. Thaker, “A review of time series forecasting methods,” Int. J. Research and Analytical Reviews, vol. 11, 2024.
[8] M. Kolambe and S. Arora, “Forecasting the future: A comprehensive review,” Journal of Electrical Systems, vol. 20, 2024.
[9] J. Kim et al., “A comprehensive survey of time series forecasting,” 2024.
[10] G. E. P. Box and G. M. Jenkins, “Time series analysis: Forecasting and control,” JRSS B, vol. 33, no. 2, 1971.
[11] C. Chatfield, Time-series forecasting. Chapman & Hall/CRC, 2000.
[12] D. C. Montgomery et al., Introduction to time series analysis and forecasting. Wiley, 2008.
[13] H. Akaike, “A new look at the statistical model identification,” IEEE Trans. Autom. Control, vol. 19, no. 6, 1974.
[14] P. Mondal et al., “Study of effectiveness of ARIMA in forecasting stock prices,” IJ CSEA, vol. 4, no. 2, 2014.
[15] Z. Hajirahimi and M. Khashei, “Hybrid structures in time series modeling,” Eng. Applications of AI, vol. 86, 2019.
[16] A. Tealab et al., “Forecasting of nonlinear time series using ANN,” Future Computing and Informatics Journal, vol. 2, 2017.
[17] P. H. B. de Melo et al., “COVID-19 forecasting using MLP,” Procedia Computer Science, vol. 181, 2021.
[18] V. N. Vapnik, The Nature of Statistical Learning Theory, 2nd ed. Springer, 2000.
[19] M. A. Ahamed and Q. Cheng, “TimeMachine: A time series is worth four Mambas,” arXiv preprint, 2024.
[20] D. M. Ahmed et al., “Deep sequential models for forecasting time series,” Applied Computational Intelligence, 2022.
[21] J. M. L. Alcaraz and N. Strodthoff, “Diffusion-based time series forecasting,” arXiv preprint, 2022.
[22] T. Alghamdi et al., “Forecasting traffic congestion using ARIMA,” Proc. IWCMC, 2019.
[23] A. F. Ansari et al., “Chronos: Learning the language of time series,” arXiv preprint, 2024.
[24] G. Athanasopoulos et al., “The tourism forecasting competition,” Int. J. Forecasting, vol. 27, no. 3, 2011.
[25] A. Behrouz et al., “Chimera: Modeling multivariate time series,” arXiv preprint, 2024.
[26] K. Benidis et al., “Deep learning for time series forecasting,” ACM Computing Surveys, vol. 55, no. 6, 2022.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Jamila Omar Aldrogi
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
Published by: 