Forecasting Drought in Rwanda Using Time Series Approach Case Study: Bugesera District

Drought is a global phenomenon that occurs almost in all landscapes causing significant damage both in natural environment and in human lives. Early detection of droughts helps to implement drought mitigation strategies and measures, before they occur. This study finds it necessary to use mathematical and statistical techniques to build models that will be used to forecast occurrences of drought in Bugesera district which is the area in Rwanda hardest hit by droughts. The developed ARIMA model was applied to forecast droughts using monthly rainfall data of Bugesera district, Eastern province of Rwanda from January 1970 to December 2015. The ARIMA generator has proven to be relatively portable across different systems, provide a good source of practically strong random data on most systems. Using the computational analysis of ARIMA approaches, better and accurate results were obtained from non-stationary monthly rainfall data. The forecasted results were compared with the observed results and the results show reasonably good approximation. So the obtained ARIMA model can be used to forecast droughts up to5 months of lead-time with reasonably accuracy as it produce the lower error value. Furthermore, the study revealed that the model can improve the current forecast future conditions, which will allow the communities to plan ahead the water management activities during droughts.