All Rights Reserved. Detecting Urban Emissions Changes and Events With a Near-Real-Time-Capable Inversion System

In situ observing networks are increasingly being used to study greenhouse gas emissions in urban environments. While the need for sufficiently dense observations has often been discussed, density requirements depend on the question posed and interact with other choices made in the analysis. Focusing on the interaction of network density with varied meteorological information used to drive atmospheric transport, we perform geostatistical inversions of methane flux in the South Coast Air Basin, California, in 2015–2016 using transport driven by a locally tuned Weather Research and Forecasting configuration as well as by operationally available meteorological products. We find total-basin flux estimates vary by as much as a factor of two between inversions, but the spread can be greatly reduced by calibrating the estimates to account for modeled sensitivity. Using observations from the full Los Angeles Megacities Carbon Project observing network, inversions driven by low-resolution generic wind fields are robustly sensitive (p < 0.05) to seasonal differences in methane flux and to the increase in emissions caused by the 2015 Aliso Canyon natural gas leak. When the number of observing sites is reduced, the basin-wide sensitivity degrades, but flux events can be detected by testing for changes in flux variance, and even a single site can robustly detect basin-wide seasonal flux variations. Overall, an urban monitoring system using an operational methane observing network and off-the-shelf meteorology could detect many seasonal or event-driven changes in near real time—and, if calibrated to a model chosen as a transfer standard, could also quantify absolute emissions.