Original Framework for Uncertainty Quantification in Environmental Risk Assessment of Produced Water Discharge in Nigeria’s Coastal Zone

Environmental risk assessment of produced water discharge to Nigerian coastal waters is currently conducted without rigorous uncertainty quantification, resulting in regulatory decisions based on point estimates that systematically underestimate risk at the tails of plausible outcome distributions. This study develops an original probabilistic framework for uncertainty quantification in environmental risk assessment of produced water discharge, applied to three shallow offshore discharge operations in Rivers State, Akwa Ibom State, and Ondo State coastal zones. The framework integrates Monte Carlo uncertainty propagation through a fate and transport model for produced water constituents with a probabilistic species sensitivity distribution model for marine ecological risk characterisation. An original contribution is the development of a Bayesian hierarchical model that pools sparse ecotoxicological data for West African coastal marine species with globally available datasets using phylogenetic distance-based relevance weighting, overcoming the data sparsity problem that has historically prevented rigorous species sensitivity distribution construction for Nigerian coastal species. Probabilistic environmental risk quotients are presented as full posterior distributions rather than point estimates, enabling regulators to characterise 95th percentile risk scenarios. Application to the three case study discharge sites reveals that while median risk quotients for all three are below unity, 95th percentile risk quotients exceed unity for dissolved barium and total petroleum hydrocarbons at two sites, identifying conditions invisible to conventional deterministic risk assessment. The study establishes a standard of practice for probabilistic environmental risk assessment in Nigeria's offshore petroleum regulatory framework. Keywords: uncertainty quantification, produced water, environmental risk, Bayesian modelling, Nigeria coastal