Machine Learning-Driven Predictive Modelling of Toxic Hydrocarbon Emissions and Public Health Risks in The Port Harcourt Refining Corridor: Integrating Data Governance, Ethical AI, and Intellectual Property Protection Frameworks

This paper uses machine learning (ML), geospatial, and public-health risk assessment to describe toxic hydrocarbons emissions in the Port Harcourt Refining Corridor (PHRC), a large industrial belt in Nigeria. The analysis demonstrates extreme exceedances of the WHO air quality guidelines, with average PM_2.5 (78.4 0g/m³) and benzene (23.8 0g/m³) levels 5 and 4 times, respectively, and a 99% frequency of PAH exceedances. The low wind speed, high humidity, and stable boundary-layer conditions allow the use of spatial distribution maps to identify hotspots of persistent pollution in the Trans-Amadi, Eleme petrochemical areas, and the Okrika artisanal refining clusters. Gradient-boosting ML models also showed high predictive scores (R² = 0.92 for PM_2.5; R² = 0.91 for benzene). The superior analysis based on SHAP showed that the most significant predictors were lag concentrations, wind speed, temperature, and humidity, with strong interactions between emissions and meteorology. Essentially, health risk analyses reveal high risks, where the non-carcinogenic hazard index (HI) of PM_2.5 and benzene are above 1 in all of the population groups (non-carcinogenic), and lifetime risks of cancer (1.85 x 10^-4) of benzene are very significant in comparison with tolerable levels. Risk contour maps also highlight industry hotspots as areas of concern due to exposure. The evaluation of data governance indicated very high privacy protection and access controls, and moderate transparency and auditability. In contrast, the IP analysis indicated medium-high ownership risk and model inversion vulnerability. The paper has shown that ML-based predictive modelling integrated into ethical governance/IP paradigms offers a highly promising avenue for real-time environmental surveillance, regulatory decision-making, and people-health defence in high-risk industrial areas.