Analyzing and forecasting rainfall patterns in the Manga-Bawku area, northeastern Ghana: Possible implication of climate change.

Abstract
Understanding rainfall processes is crucial in addressing several hydrological challenges that have both positive and negative impacts on agriculture, climate change, and natural hazards including floods and droughts. Statis- tical modeling is a key instrument for studying these processes. This study presents the trends and forecasted rainfall patterns from 2017 to 2035 in the Manga-Bawku area, northeastern Ghana using rainfall data spanning from 1976 to 2016. The simple seasonal exponential smoothing and ARIMA (0,1,1) models were employed in this study while the R software and the Statistical Package for the Social Sciences were the modeling tools used in this study. The results obtained from the test of the efficiency of the forecast model showed a Stationary R- squared value of 0.698 and 0.669, Root-Mean-Square Deviation (RMSE) of 48.775 and 50.717, and normalized Bayesian Information Criteria (BIC) of 7.800 and 7.904 for the exponential smoothing and ARIMA (0,1,1) models respectively. This indicated that the models are a good fit for the time-series analysis. However, a line plot of the two models and observed data showed that the simple seasonal exponential smoothing model was a better fit. The Autocorrelation Function (ACF) plot showed that coefficient values were recorded in equal time lags. The peak recorded at lag 1 was similar to lags 12 and 24 which suggest that the seasonal component in the time series occurred within twelve months. The study further showed that the seasonal component was uniform on an annual basis which signifies that it was an additive impact instead of multiplicative effect. The rainfall forecast predicts a decline in rainfall levels over the next 19 years. This suggests the need for proper water resources planning, and the formulation of policies to curtail the factors that are likely to have debilitating impacts on the local hydrological system.
Description
This article is published by ELSEVIERS.COM, 2021 and is also available at https://doi.org/10.1016/j.envc.2021.100354
Keywords
Citation
https://doi.org/10.1016/j.envc.2021.100354
Collections