Estimation of point source fugitive emission rates from a single sensor time series: A conditionally-sampled Gaussian plume reconstruction

Abstract

Emerging mobile fugitive emissions detection and measurement approaches require robust inverse source algorithms to be effective. Two Gaussian plume inverse approaches are described for estimating emission rates from ground-level point sources observed from remote vantage points. The techniques were tested using data from 41 controlled methane release experiments (14 studies) and further investigated using 7 field studies executed downwind of oil and gas well pads in Wyoming. Analyzed measurements were acquired from stationary observation locations 18–106 m downwind of the emission sources. From the fluctuating wind direction, the lateral plume geometry is reconstructed using a derived relationship between the wind direction and crosswind plume position. The crosswind plume spread is determined with both modeled and reconstructed Gaussian plume approaches and estimates of source emission rates are found through inversion. The source emission rates were compared to a simple point source Gaussian emission estimation approach that is part of Draft EPA Method OTM 33A. Compared to the known release rates, the modeled, reconstructed, and point source Gaussian controlled release results yield average percent errors of −5%, -2%, and 6% with standard deviations of 29%, 25%, and 37%, respectively. Compared to each other, the three methods agree within 30% for 78% of all 48 observations (41 CR and 7 Wyoming).