Truck-Based Methane Detection, Attribution, and Quantification in Upstream Oil and Gas: Controlled Release Validation and Field Case Study

The oil and gas (O&G) sector is a significant source of global anthropogenic methane (CH₄) emissions, prompting increased regulatory oversight and the rapid development of new methane measurement and mitigation technologies. While screening technologies such as optical gas imaging (OGI) are widely used for regulatory compliance due to their ability to visually identify component-level leaks, there is emerging evidence from regulatory effectiveness studies in Canada that OGI surveys do not detect all sources, with remote sensing surveys often identifying significantly higher site-level emissions. Complementary methods with low detection thresholds may be necessary to improve regulatory compliance and fully represent low-level emission distributions in measurement inventories. In this study, we characterize the performance of a truck-based measurement system using controlled release data, and present results from a field case study in which this method was applied alongside aerial LiDAR and quantitative OGI surveys.

Truck-based measurement systems are a relatively inexpensive and efficient option for site-level screening and emission quantification. This method integrates a vehicle-mounted gas analyzer, anemometer, and GPS to collect atmospheric CH₄ concentrations and wind characteristics along the driven route. This data is processed via an automated framework in which CH₄ plumes are identified, attributed to a source based on wind characteristics and source geometry, and quantified using a Gaussian plume dispersion model. We assess detection, attribution, and quantification performance using data collected by Eotrac Incorporated during controlled release experiments (0.025 - 11 kg/h) at test sites simulating realistic O&G emission scenarios. While release rates and locations were blind to the measurement team during testing, the analysis presented here was conducted after the releases were unblinded.

The truck-based system achieved a true positive detection rate exceeding 95 % with no false positives. We find that increasing the number of downwind measurement transects can significantly reduce the 90 % detection limit, from 0.45 kg/h with one transect to 0.03 kg/h with five transects. During single-source release scenarios, source attribution accuracy was 100 % at the facility level, 99.7 % at the equipment group-level, and 50 % at the individual source-level, indicating strong performance for identifying emitting equipment groups (7-15 m radius) despite challenges in pinpointing exact leak locations.

In the field case study, the site-level emission frequency was 74.3 % for the truck-based method, compared to 8.6 % for QOGI and 31.8 % for aerial LiDAR. This suggests that OGI misses a significant fraction of emitting sites. Truck-based methods therefore offer a reliable complement to existing detection approaches and have the potential to improve both regulatory compliance and the representation of low-level emitters in inventories.