Methane Hazard in a Post-Mining Urban Environment: A Probabilistic Risk Assessment for Lupeni, Romania

Abandoned mine methane can remain active beneath post-mining cities long after mine closure, creating environmental and public-safety challenges when urban works reactivate subsurface gas-migration pathways. This study assesses methane hazard in Lupeni, Romania, where unexpected gas emissions were detected during commissioning tests for a new natural gas pipeline in August 2024. The analysis uses a monitoring database of 41 fixed points surveyed during three field campaigns (123 observations in total). Descriptive statistics showed a strongly right-skewed concentration field, with an overall mean of 5.65% vol., a median of 1% vol., and a maximum of 54% vol. Campaign-to-campaign correlations (Pearson r = 0.746–0.923) indicated short-term recurrence of hotspot behaviour over the monitored period. To translate these observations into decision-relevant risk information, a fully reproducible monitoring-driven probabilistic screening model was implemented. Consequence severity was mapped from measured methane concentration, while the probability component was derived from alert-threshold recurrence across the three campaigns using a Beta posterior and Monte Carlo resampling (50,000 iterations). The mean point-level risk was 6.97, indicating that the typical monitored location was not uniformly critical. However, a hotspot-envelope indicator defined as the 95th percentile of point-level risk across monitoring points had a mean of 19.40, and the model-derived probability that it exceeded the critical threshold of 17 was 97.6%. The findings suggest that the Lupeni hazard was localized but severe, supporting targeted monitoring, excavation control, and hotspot-focused intervention rather than generalized alarmism. More broadly, the workflow may provide a practical screening approach for other post-mining urban areas where repeated monitoring data are available but mechanistic subsurface characterization remains incomplete.