Quantification Error Model for Aerial LiDAR Methane Emission Rate Estimates

Cameron D. Dudiak, Devin B Goodwin, Dominic Altamura, Chris P Donahue, Michael Thorpe

Accurate characterization of methane emission rate quantification error (QE) is essential for building measurement-based emissions inventories that benchmark emissions, guide mitigation, and satisfy reporting frameworks such as OGMP 2.0. Previous studies have summarized QE using distributions of errors from controlled release experiments, but with limited consideration of environmental conditions or plume characteristics. We address this gap by developing a continuous QE model as a function of local wind speed and measured signal-to-noise ratio (SNR). Candidate models with diverse functional forms are fitted by maximum likelihood estimation and evaluated using Akaike Information Criterion on a dataset of 2,178 single- and double-blind controlled releases across six test sites scanned with dozens of Bridger Photonics’ Gas Mapping LiDAR (GML) 2.0 sensors. Results show that GML QE can be decomposed into two components, bias and uncertainty, where bias is primarily controlled by SNR, and uncertainty is jointly controlled by SNR and wind speed. Bias corrections derived from the model are reinforced by physical principles, providing confidence in their origin and applicability. Applying these corrections to U.S. basin inventories reduces systematic inventory bias by up to 17.5%, enabling more accurate, transparent, and comparable methane emission estimates.