Hierarchical analysis of rainfall variability across Nigeria

Abstract

Rainfall in Nigeria is subjected to wide variability both in time and space. This variability has assumed a more pronounced dimension as a result of climate change. In this paper, cluster analyses were used to study rainfall variability in Nigeria. Rainfall data in 20 locations spread across the geopolitical and ecological zones of Nigeria were subjected to hierarchical cluster analysis and analysis of time series and coefficient of variation for over periods spanning 30 years. Maps of spatial variations of mean annual rainfall and mean rainfall coefficient of variation were produced using ARCGIS 10.1. Furthermore, a better understanding of temporal variation of rainfall was sought by an investigation into the time series of rainfall coefficients of variation. It was found that the southern parts of the country were given to more severe rainfall variability/fluctuations than the northern parts. The north central parts exhibited more similarity to the southern parts than the other northern locations. The relationship between average annual rainfall and the coefficient of rainfall variation was found to follow a power law with R2 value approximately 0.7. With respect to variability of annual rainfall, three zones emerged as follows: a linear relationship (R2 = 0.90) exists between coefficient of variation and average annual rainfall for the rainforest zone of the southsouth; a power law (R2 = 0.86) exists between coefficient of variation and average annual rainfall for all rainforest and derived guinea savannah zones of the southeastern and southwestern states; and a logarithmic relationship (R2 = 0.54) exists between coefficient of variation and average annual rainfall for all northern states regardless of ecological zone. Generally, in-year rainfall variability increases from the northwest to the southwest; while between-year (yearly) rainfall variability increases from the north central to the southeast. This study further confirms that rainfall variability over time follows a spatial trend within a certain arbitrary boundary.