Statistical Evaluation of Changes and Periodicity in Rainfall Over East Africa During the Period 1960–2017

East African countries (Uganda, Kenya, Tanzania, Rwanda, and Burundi) are adversely impacted by droughts and floods which occasionally result from rainfall variability and periodicity. For this study, rainfall datasets from 30 rainfall stations were selected from the eight homogeneous rainfall zones of East Africa (EA) after preliminary quality assessment and tests, including outliers, normality, and homogeneity tests. The Morlet wavelet technique was used to establish the periodicity of monthly rainfall for typical March–May (MAM) and October–December (OND) seasons. Mann–Kendall's (MK) statistical test and Sen’s slope estimator (SE) were used to examine the trend and the magnitude of change in seasonal and annual rainfall over EA, respectively. Wavelet transform approaches were used to show the linkages between the seasonal rainfall and the two atmospheric indices: El Niño–Southern Oscillation (ENSO) and Indian Ocean Dipole (IOD). Results revealed a 1-year band as a dominant period of variability over EA. MK results indicate that 97% of the stations had a decreasing trend during MAM rainfall seasons, while nearly 60% of the stations presented a positive trend for OND and annual rainfall. A statistically significant negative correlation was established between ENSO and MAM rainfall, indicating a strong link between ENSO/La Niña and drought events during 1973/1974 and 1983/1984. Meanwhile, a statistically significant positive correlation between IOD and OND seasonal rainfall coincided with the 1972/1973 and 1997/1998 El Niño events. This established relationship could be used in drought and flood early warning systems to predict extreme events like drought and floods over EA. Besides, for the highly variable rainfall in the region, tailor-made climate information for planning economic programs like agriculture, energy, and disaster preparedness is highly recommended.