Comparing Continuous Methane Monitoring Technologies for High-Volume Emissions: A Single-Blind Controlled Release Study

Abstract

Methane emissions from oil and gas operations are a primary concern for climate change mitigation. While traditional methane detection relies on periodic surveys that yield episodic data, continuous monitoring solutions promise to offer consistent insights and a richer understanding of emission inventories. Despite this promise, the detection and quantification ability of continuous monitoring solutions remain unclear. To address this uncertainty, our study comprehensively assessed 8 commercial continuous monitoring solutions using controlled release tests to simulate high-volume venting (e.g., uncontrolled tanks, pneumatics, unlit flares), which accounts for a significant fraction of total emissions from oil and gas systems. The performance of each team varied: when comparing reported results on a second-by-second basis, all teams reported false positive rates below 10\%. For true positive rates, 4 out of 8 systems exceed 80%. In the field test where continuous monitoring solutions identified and reported an emission event, all systems' reliability of identification surpassed 70%. When systems reported there was no emission event, the reliability of non-emission identification varied from 29.37% to 96.15%. Among 5 systems tested for quantifying the daily average emission rate released by the Stanford team, all underestimated by an average of 74.38% emissions. This indicates that their application in emissions reporting or regulation may be premature. The variability in monitor performance underscores the importance of understanding systems' strengths and limitations before their broader adoption in methane mitigation approaches or regulatory frameworks.

Methane Quantification Performance of the Quantitative Optical Gas Imaging (QOGI) System Using Single-Blind Controlled Release Assessment

“…All three groups had positively skewed (mean > median) distributions that were significantly (statistically using the Mann-Whitney U test) different (p < 0.05) with mean errors inflated by outliers as shown in 4a (also see Supplementary Materials Section S3.1). This type of positive skewness has been seen in several other studies of next-generation leak quantification solutions [71][72][73][74]. As controlled release rate increased, there was observed improvement in quantification performance; mean error decreased, the interquartile error range narrowed, and the number and size of outliers dropped.…”

Estimating methane emission durations using continuous monitoring systems

“…Current commercially available CMS solutions have large quantification errors on controlled releases, [15][16][17] but their detection capabilities show promise, especially for larger emissions. 17 Therefore, while quantification capabilities evolve, CMS can complement snapshot measurements by bounding the duration of detected emissions.…”

“…Current commercially available CMS solutions have large quantification errors on controlled releases, [15][16][17] but their detection capabilities show promise, especially for larger emissions. 17 Therefore, while quantification capabilities evolve, CMS can complement snapshot measurements by bounding the duration of detected emissions. The current reporting threshold for large emissions is 100 kg/hr under the proposed EPA rule, but bounding the duration of smaller emissions may become important in the near future, which becomes increasingly challenging for satellites as the size of the relevant emissions decreases.…”