An Event-Based Framework for Estimating Annual Methane Emissions and Managing Emissions Data from Upstream Oil and Gas Facilities

Mozhou Gao, Zahra Ashena, Steve H.L. Liang, Sina Kiaei, Saeedi Sara

Accurate reporting of annual site-level methane emissions is increasingly required under emerging regulatory and voluntary frameworks in the oil and gas (O&G) sector. In this study, we present an event-based framework for estimating annual methane emissions from upstream O&G facilities. The framework applies the Emission Event Data Model (EEDM) to spatiotemporally aggregate multi-scale emissions data into discrete events using the concept of Allen’s interval algebra and spatial proximity. Following the creation of events, emissions are categorized into three groups: resolved (with a known emission rate and duration), partially resolved (with a known emission rate but an unknown duration), and unresolved (with an unknown emission rate and duration) to facilitate various management and emissions estimation approaches. To estimate emissions and its associated uncertainties using events, we developed three Monte Carlo-based approaches, which are (1) estimating durations for partially resolved events using null detections and leak generation and natural repair processes; (2) estimating emissions from unresolved events based on the minimum detection limit of deployed technologies; and (3) estimating emissions from unresolved events using probabilistic occurrence and best-fit rate and duration distributions. This framework enables emissions to be reported and verified at a uniform scale rather than at the individual observation scale. To demonstrate the estimation of emissions using this framework, we created two case studies. In both studies, we performed emissions estimation using emission observations synthesized from real emissions data from an upstream O&G site. The proposed framework and methodologies can be implemented in voluntary initiatives such as Veritas 2.0 and the Oil & Gas Methane Partnership 2.0 and the event data model can be applied as a data management framework for the Measurement, Monitoring, Reporting, and Verification (MMRV) framework.