Vehicle-Based Methane Detection, Attribution, and Quantification in the Upstream Oil and Gas Sector: Method Overview and Controlled Release Validation

Robyn Latimer , Evelise Bourlon, Khalil El Hachem, Jacob Johnson, Chukwuemeka Nwokoye, David Risk

Vehicle-based measurement systems occupy an important niche in the multi-scale methane measurement landscape between component-level technologies (e.g. OGI) and aerial-based methods. Performance evaluations are often limited to one dimension (e.g. quantification), while methodological opacity limits interpretation and inter-comparisons between technologies. Here, we develop a vehicle-based measurement system and processing framework incorporating a novel weighted likelihood algorithm for automated source attribution. Using controlled release experiments at two dedicated testing facilities, we characterize system performance across three dimensions – detection, attribution, and quantification. The system achieved a 96.1 % true positive detection rate with no false positives and a 90% detection limit as low as 0.024 kg/h with repeat transects. Attribution accuracy was high at the source-group level, reaching 99.7 % in single-source releases and 86 % in multi-source scenarios. Quantification was unbiased for low-to-moderate complexity releases (2 source groups; slope=0.99, R2=0.81) but overestimated under highly complex conditions (5 source groups; slope=1.45, R2=0.91). These results provide a transparent performance benchmark, demonstrating that these systems are well-suited to survey the large populations of low-to-moderate complexity sites (e.g. single wells and batteries), as a screening layer in LDAR applications and to quantify the low-end of emission distributions in inventories, reducing reliance on generic emission factors.