Irregularities in rainfall behaviour has been characteristically visible in semi-arid climate particularly in the tropics. Trends and variations have been used as proxies in detecting the presence of irregularities in climatic variables such as rainfall. Thirty six years (1980-2015) of data obtained from Nigerian Meteorological Agency (NIMET) were used for the present study. Multivariate statistics such as ANOVA and Cluster Analysis were used for assessing variations between and within the data mean. However, Mann-Kendall trend test was applied for trend detection. The ANOVA results showed significant variation in rainfall [F (3, 140) = 67.012, P < .05], between the stations. The cluster analysis produces two classes for rainfall indicating that rainfall is less variable spatially. Mann-Kendall trend test result shows an insignificant annual increasing trend in rainfall. However, the monthly series showed varied trend results consisting of significant and insignificant increasing and decreasing trends. The trend results was spatially interpolated using inverse distance weightage (IDW) for easy comprehension of the spatial distribution of rainfall across the river catchment Even though the annual trend results was statistically insignificant, the fact that the area is fragile and sensitive to minor climatic changes, the result is still crucial for the planning and management of rainfall related activities especially water supply and agriculture, and for the preparation against weather extreme events such as floods and droughts in the area. Finally, it is suspected that other factors temperature and land use changes may aggravates the impacts of rainfall irregularities and are thus recommended for future climate variability studies. Key words: Rainfall, Climate variability, Mann-Kendall, Cluster Analysis, Semi-Arid, Nigeria.