Advancements in seasonal rainfall forecasting: A seasonal auto-regressive integrated moving average model with outlier adjustments for Ghana's Western Region
Accurate rainfall forecasting is essential for agricultural planning, water resource management, and disaster preparedness, particularly in regions with variable weather patterns, such as Ghana's Western Region. This study advances existing research by developing and applying a Seasonal Auto-Re...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-06-01
|
| Series: | Scientific African |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2468227625001024 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Accurate rainfall forecasting is essential for agricultural planning, water resource management, and disaster preparedness, particularly in regions with variable weather patterns, such as Ghana's Western Region. This study advances existing research by developing and applying a Seasonal Auto-Regressive Integrated Moving Average (SARIMA) model, specifically SARIMA(1,0,2)(2,0,0)[12], tailored to the unique seasonal rainfall patterns of the region. The methodology incorporates robust outlier detection and adjustment techniques, including the interquartile range (IQR) and winsorization, to increase the model's resilience against extreme weather events while preserving critical data characteristics. The stationarity of the dataset, comprising 84 months of rainfall records from 2017 to 2023, was rigorously tested via augmented Dickey–Fuller (ADF) and KPSS tests, confirming its suitability for time series analysis. Diagnostic plots validated the model's performance, and metrics such as the R-squared (99.02 %), mean absolute percentage error (MAPE, 5.97 %), and Theil's U statistic (0.05) demonstrated its accuracy and reliability. Compared with other models, the Seasonal Auto-Regressive Integrated Moving Average consistently outperformed alternatives in capturing seasonal trends while moderating the influence of anomalies. This study's methodology and findings contribute to enhanced rainfall prediction capabilities, assisting local stakeholders in climate-resilient agricultural planning and water resource management. The approach also offers a replicable framework for similar regions facing climatic variability and extreme weather challenges. |
|---|---|
| ISSN: | 2468-2276 |