Methane retrieval from MethaneAIR using the CO2 Proxy Approach: A demonstration for the upcoming MethaneSAT mission

Reducing methane (CH$_4$) emissions from the oil and gas (O\&G) sector is key to mitigating climate change in the near-term. MethaneSAT is an upcoming satellite mission designed to monitor basin-wide O\&G emissions globally, providing estimates of emission rates and helping identify the underlying processes leading to methane release to the atmosphere. MethaneSAT data will help advocacy and policy efforts to help track methane reduction commitments and targets made by countries and industry. Here we introduce the CH$_4$ retrieval algorithm for MethaneSAT based on the CO$_2$ proxy method. We apply the algorithm to observations from the maiden campaign of MethaneAIR, an airborne precursor to the satellite with similar instrument specifications. The campaign was conducted during winter 2019 and summer 2021 over three major US oil and gas basins. Analysis of the MethaneAIR data shows that measurement precision is typically better than 2\% for $20\times20$~m$^2$ pixel resolution, with no strong dependence on geophysical variables such as surface reflectance. We show that detector focus drifts over the course of each flight likely due to thermal gradients that develop across the optical bench. The impacts of this drift on retrieved CH$_4$ can mostly be mitigated by including a parameter that squeezes the laboratory tabulated instrument spectral response function in the spectral fit. Validation against coincident EM27/SUN retrievals shows that MethaneAIR values are generally within 1\%. MethaneAIR retrievals were also intercompared with those of TROPOMI; the latitudinal gradients for the two datasets are in good agreement, with a 2.5~ppb mean bias between instruments. We evaluate the accuracy of MethaneAIR estimates of point source emissions using observations made over the Permian O\&G basin, based on the integrated mass enhancement approach coupled with a plume-masking algorithm based on total variational denoising. We estimate that the median point source detection threshold is 100-150~kg~h$^{-1}$ at the aircraft's nominal 12km above-surface observation altitude, based on an ensemble WRF large eddy simulations used to mimic the campaign conditions with the threshold for quantification about 2$\times$ the detection threshold. Retrievals from repeated basin surveys indicate the presence of both persistent and intermittent sources, and we highlight an example from each case. For the persistent source we infer emissions from a large O\&G processing facility, and estimate a leak rate between 1.6 and 2.1~\%, higher than any previously-reported emission from a facility of its size. We also identify a ruptured pipeline that alone would constitute 2~\% of estimated basin emissions, two weeks before it was found by its operator, highlighting the importance of regular monitoring from the future satellite mission. The results showcase the capability of MethaneAIR to make highly accurate, precise measurements of methane dry-air mole fractions in the atmosphere, with fine spatial resolution over large swaths on the ground. The results provide confidence that MethaneSAT can make such measurements at unprecedentedly fine scales from space ($\sim 130\times400$~m$^2$), thereby delivering quantitative data on basin-wide methane emissions.