FROM COMPLIANCE AUDIT TO PREVENTIVE APPROVAL: A GIS-MCDA DECISION FRAMEWORK FOR SITING FUTURE PETROL STATIONS IN KEFFI, NIGERIA

Retrospective compliance audits will tend to discover poorly sited petrol stations after the money has been invested and before there has been public exposure. This methodology paper translates evidence from a survey of 43 petrol stations in Keffi, Nigeria into a preventative GIS-MCDA framework for screening station proposals for future investments. The paper differs from previous reporting of station clustering and compliance and analysis of stakeholder-claimed locational choices and hazards. The framework segregates land that is legally or technically infeasible from gradient suitability. Stage 1 applies exclusionary criteria for sensitive public uses, minimum inter-station spacing, reserved road space, infeasible land uses and parcels without sufficient area or emergency access. Stage 2 evaluates feasible land using access, traffic demand, road class, junction safety, population/ service need, terrain, drainage, environmental potential, utility access and corridor saturation. Criteria weights will be elicited from Nigerian petroleum regulators, planning authorities, fire and environmental agencies, transport bodies, operators and community groups in a transparent decision process, using the Analytic Hierarchy Process. Testing for internal logical consistency is mandatory, and sensitivity analysis advisable. Stage 3 site-specific validation, traffic impacts, environmental assessment, emergency response appraisal and a multi-agency decision are mandatory. The parent evidence justifies the framework: stations were largely clustered, 60.5% occurred on 3 corridors, compliance reporting was highly inconsistent for different criteria, and commercial locational decisions were biased towards corridors and junctions. The model prevents an agentically attractive criterion from cancelling mandatory safety exclusion. In addition, demand and service equity metrics are provided to prevent further stations from contributing solely to already operate corridors. My paper offers a data schema, workflow map, decision matrix, and validation protocol and governance framework. My paper does not offer a final suitability map due to unavailability of current land use, traffic, and parcel and hazard data. The model is ready for empirical implementation once these datasets and current governance practices are documented.