Trend Analysis and Forecasting of Precipitation for Southern Nigeria

Abstract

The current study is carried out to determine the potential trend in annual precipitation for Southern Nigeria with a view to assessing its significance, and to model the annual rainfall for future prediction to enhance policy implementation. For this study, ten stations within the southern Nigeria region with continues data series of thirty years and above were subjected to homogeneity analysis using SNHT and Pettitt's test. The widely used Mann-Kendall and Sen's estimator tests were run at 5% significance level on time series data for each of the homogenized stations for the time period, 1971 to 2012. Finally, Holt's exponential smoothing model was developed which allow predictions on future evolutions. These analyses were executed using the R i386 3.4.3 programming language. Three out of the ten stations (Benin, Ikeja and Onitsha) failed both homogeneous tests and were excluded from further analysis. The Mann-Kendall tests indicated that Akure, Enugu, and Port Harcourt stations do not indicate statistically significant results, while the other states show statistically significant increasing trends, Sen's estimator revealed a negative slope for Akure. Holt's Exponential smoothing revealed that the annual rate of increase in rainfall were; -0.514mm/year for Akure, 2.4mm/year for Calabar, 15.232mm/year for Enugu, 40.03mm/year for Ibadan, 27.189mm/year for Owerri, -20.643mm/year for PH and 162.708mm/year for Uyo respectively.