Gaussian plume and backward Lagrangian stochastic modeling of methane emissions: an experiment at METEC

Abstract

Methane (CH4) is a potent greenhouse gas, and emissions from the oil and gas (O&G) sector are a crucial contributor to it. Accurate quantification of CH4 emission from the O&G sector is important. This study investigates the performance of Gaussian plume (GP) and backward Lagrangian stochastic (bLS) dispersion models through controlled release experiments conducted at the METEC facility at Colorado State University for release rates of 0.5 to 6 kg h-1. Experiments were performed in near-field conditions (<100 m) with a 5-minute sampling period across varying release heights and atmospheric conditions. Results reveal that performance from the bLS model was better than GP, as indicated by the factor of 2 (FAC2;35% for GP and 52% for bLS), geometric mean bias (MG; 0.08 for GP and 0.63 for bLS), and mean factor of error (MFoE; 3.5 for GP and 1.9 for bLS). Models' performance slightly improved when the release height was closer to the sampling height and in moderately unstable atmospheric conditions. This study suggests that the bLS model is more suitable than GP for mock near-field O&G facilities like METEC, when experiments are conducted under a single controlled release.