Optimizing rainfall prediction in coastal and inland areas: a comparative analysis of forecasting models in eThekwini district, South Africa

Ntokozo Xaba and Ajay Kumar Mishra
Additional contact information
Ntokozo Xaba: Durban University of Technology
Ajay Kumar Mishra: Durban University of Technology

International Journal of Business Ecosystem & Strategy (2687-2293), 2025, vol. 7, issue 1, 180-197

Abstract: While floods and droughts are natural occurrences in the earth’s hydrological cycle, their escalating frequency and intensity have become a major concern for governments throughout the globe. Developing nations, such as South Africa, are weary of these extreme weather events because they understand they lack the necessary resources and infrastructure to deal with them. The eThekwini Municipality serves as a prime example of how vulnerable developing nations' regions are to the devastating effects of floods and droughts, as multiple floods have devastated the area, resulting in fatalities, damaging public infrastructure, and demolishing houses. The scale of the damage from the floods reveals that significant gaps exist in disaster preparedness in the eThekwini Region. Rainfall forecasting is a vital tool that has been underutilised that can be used preemptively to manage or mitigate flooding and enhance water resource management in the region. Machine learning models in particular are very useful in rainfall forecasting; hence, the goal of this study was to evaluate the most efficient models for forecasting precipitation in the eThekwini northern and central regions, which are coastal and inland areas, respectively. Rainfall data spanning 32 years was obtained from meteorological stations in both regions, and the SARIMA, ARIMA, and ETS machine learning models were used for rainfall forecasting and evaluated based on their ability to capture seasonal patterns, handle non-stationarity, and provide accurate predictions. Model performance was analysed, and comparisons were made using the root mean squared error (RMSE), mean absolute error (MAE), and mean absolute scaled error (MASE) as evaluation metrics. The study's findings indicate that the most effective models for both the northern and central regions were SARIMA (0,0,0) (2,0,0) [12] and SARIMA (1,0,0) (1,0,0) [12]. These findings provide valuable insights for meteorologists, hydrologists, and policymakers involved in regional climate modelling and water resource management. Key Words: Rainfall Prediction, ARIMA, SARIMA, ETS, Climate Modelling

Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
https://www.bussecon.com/ojs/index.php/ijbes/article/view/640/408 (application/pdf)
https://doi.org/10.36096/ijbes.v7i1.640 (text/html)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:adi:ijbess:v:7:y:2025:i:1:p:180-197

Access Statistics for this article

International Journal of Business Ecosystem & Strategy (2687-2293) is currently edited by Umit Hacioglu

More articles in International Journal of Business Ecosystem & Strategy (2687-2293) from Bussecon International Academy Bussecon International Academy, School of Business, IHU, Ordu cad. F-05 Blok No 3, 34480 Basaksehir, Istanbul, Turkey. Contact information at EDIRC.
Bibliographic data for series maintained by Umit Hacioglu ().