© Lekia Nkpordee, Yusuf Abass Aleshinloye, Ejidokun Adekunle Olugbenga, Ikpotokin Osayomore. Article available under the CC BY-SA 4.0 licence
ARTICLE
ABSTRACT
This study develops and evaluates an ARIMA-LSTM hybrid model for forecasting urban temperature dynamics in selected Ugandan cities, with the goal of capturing both linear trends and nonlinear fluctuations within a unified and interpretable framework. Monthly temperature data from 2017 to 2023 obtained from the Uganda National Meteorological Authority, alongside long-term urban population data from the World Bank, were used to support robust urban climate analysis. Data quality was verified through systematic preprocessing and outlier assessment, providing reliable inputs for model estimation. The proposed hybrid approach applies ARIMA to explicitly model-linear and seasonal temperature structures, while an LSTM network learns the remaining nonlinear patterns embedded in residuals. The model’s performance was evaluated against a wide range of benchmark models, including standalone statistical models, deep learning architectures, machine learning methods, and Facebook Prophet used strictly for comparison. The evaluation relied on multiple accuracy and goodness-of-fit measures such as RMSE, MAE, MAPE, SMAPE and R squared, complemented by visual diagnostics and classification-based performance analysis. Results consistently show that the ARIMA-LSTM hybrid outperforms all competing models, achieving smaller forecast errors, stronger explanatory power, better classification of temperature and more reliable prediction interval coverage. Forecasts generated for major Ugandan cities show persistent spatial differences in temperature patterns, with northern cities remaining warmer than highland regions. Overall, the findings demonstrate that hybrid modeling offers a reliable and practical tool for urban temperature forecasting, with clear relevance for urban climate planning, adaptation strategies, and evidence-based decision-making in Uganda.
KEYWORDS
ARIMA-LSTM hybrid model, urban temperature forecasting, temperature dynamics, time series modeling, machine learning algorithms.
REFERENCES
Chen, X., Liu, Y., (2018). Integrating ARIMA and LSTM for improved time series forecasting. IEEE Transactions on Neural Networks and Learning Systems, Vol. 29, No. 5, pp. 1923–1935.
Cheng, S., Shi, J., Cheng, Q., Zhou, X. and Zeng, S., (2025). Hybrid model for mediumterm load forecasting in urban power grids. Energies, Vol. 18, No. 16, p. 4378. https://doi.org/10.3390/en18164378.
Garcia, M., Silva, R., Pereira, L., Santos, A. and Rocha, D., (2019). Urban climate resilience: Statistical modeling of temperature variability in Brazilian cities. Climate Resilience Reviews, Vol. 15, No. 1, pp. 150–165.
Gomez, F., Sanchez, M., Patel, R., Li, Q. and Zhao, Y., (2020). Hybrid statisticalmachine learning models for environmental forecasting. Journal of Environmental Science and Technology, Vol. 18, No. 4, pp. 456–470.
Johnson, A., Smith, B. and Williams, C., (2016). Urban temperature dynamics: ARIMA modeling for climate forecasting. Journal of Climate Change, vol. 8, no. 3, pp. 210–225.
Jones, E., Brown, D., (2017). Limitations of ARIMA models in urban temperaturę forecasting. Urban Climate Dynamics Review, Vol. 12, No. 3, pp. 210–225.
Kiarie, J., Mwalili, S., Mbogo, R., Mutinda, J. and Langat, A. (2025). Statistical, machine learning, and deep learning models for COVID-19 forecasting in Kenya. Computational and Mathematical Biophysics, 13(1), 20250026. https://doi.org/10.1515/cmb-2025-0026.
Lee, X., Chen, Y., Huang, Z., Wang, M. and Zhang, L., (2017). Forecasting urban temperature dynamics: A comparison of Facebook Prophet and traditional models. Environmental Science and Technology, Vol. 12, No. 2, pp. 300–315.
Li, Z., Johnson, A., Perez, M., Brown, C. and Taylor, S., (2019). Urban climate resilience strategies: A review of recent literature. Climate Resilience Reviews, Vol. 15, No. 4, pp. 321–335.
Lynda, D., Logeswari, G., Tamilarasi, K. and Rakesh, S., (2025). Hybrid Bayesian deep learning model for predicting urban heat island intensity in African cities. Scientific Reports, Vol. 15, Article No. 31280. https://doi.org/10.1038/s41598-025-13492-4.
Mutinda, J. K., Geletu, A. (2025). Stock market index prediction using CEEMDANLSTM- BPNN-decomposition ensemble model, Journal of Applied Mathematics, 7706431, 32 pages, https://doi.org/10.1155/jama/7706431.
Nguyen, H., Smith, J., Wang, T., Chen, L. and Lee, X., (2021). Advancements in environmental modeling using hybrid models. Environmental Modeling and Software, Vol. 30, No. 2, pp. 89–102.
Nguyen, T., Huynh, Q., Tran, P., Le, M. and Vo, D., (2020). Machine learning for urban temperature forecasting: A comparative study with traditional models. Journal of Computational Statistics, Vol. 30, No. 3, pp. 500–515.
Smith, A., Jones, B. and Brown, C., (2018). Urban temperature dynamics: Challenges and opportunities. Journal of Climate Change, Vol. 5, No. 2, pp. 123–135.
Tang, S., Jiang, Y., Su, H., Lim, K. M., Zheng, Z. and Zhu, Y., (2025). Wear defect detection of hydraulic pump using a hybrid method of VGG and LSTM. International Journal of Hydromechatronics, Forthcoming Articles, https://doi.org/10.1504/IJHM.2025.10073233.
Taylor, S. J., Letham, B., (2017). Facebook Prophet: A forecasting tool for the R programming language. Journal of Computational Statistics and Data Visualization, Vol. 25, No. 3, pp. 301–315.
Wang, H., Liu, J., Zhao, F., Chen, Q. and Yang, S., (2018). Hybrid ARIMA-LSTM modeling for urban temperature forecasting. Journal of Environmental Science, Vol. 25, No. 4, pp. 400–415.
Wang, L., Thompson, D., Green, S., Kim, J. and Lee, Y., (2020). Machine learning applications in environmental science: A comprehensive review. Environmental Science: Processes & Impacts, Vol. 22, No. 7, pp. 1567–1583.