Use of Statistical Methods in Environmental Impact Assessment for Oil Spill Risk in Bayelsa State

Environmental impact assessments for oil infrastructure in Bayelsa State require rigorous statistical quantification of spill risk, contamination spread probability, and ecosystem damage magnitude to support regulatory decision-making by the Department of Petroleum Resources and NESREA. This study applies statistical methods to quantify oil spill risk parameters using historical spill incident data from NOSDRA and soil and water contamination measurements from field sampling in Yenagoa and Ekeremor LGAs. Spill frequency was modelled by Poisson distribution, and spill volume was modelled by log-normal distribution. Risk indices were computed as the product of estimated probability and consequence severity scores. Spatial concentration of contamination was tested by spatial autocorrelation analysis using Moran's I. The mean annual spill frequency was 14.3 incidents per year (Poisson lambda = 14.3, 95% CI: 12.1 to 16.8). Spill volumes were best described by the log-normal distribution (KS p-value = 0.22). Pipeline corrosion was the statistically dominant cause (48 percent), followed by equipment failure (31 percent) and third-party interference (21 percent). Moran's I confirmed significant spatial clustering of high contamination sites near aging pipeline corridors (I = 0.41, p = 0.003). Risk indices were highest in communities within 2 km of pipeline routes over 20 years old. The study recommends risk-stratified pipeline maintenance scheduling and community monitoring participation based on statistically derived risk index maps. Keywords: oil spill risk, statistical risk assessment, Bayesa State, Poisson model, environmental statistics