Topâ down estimate of methane emissions in California using a mesoscale inverse modeling technique: The San Joaquin Valley

Abstract

We quantify methane (CH4) emissions in California’s San Joaquin Valley (SJV) by using 4â days of aircraft measurements from a field campaign during Mayâ June 2010 together with a Bayesian inversion method and a mass balance approach. For the inversion estimates, we use the FLEXible PARTicle dispersion model (FLEXPART) to establish the sourceâ receptor relationship between sampled atmospheric concentrations and surface fluxes. Our prior CH4 emission estimates are from the California Greenhouse Gas Emissions Measurements (CALGEM) inventory. We use three meteorological configurations to drive FLEXPART and subsequently construct three inversions to analyze the final optimized estimates and their uncertainty (one standard deviation). We conduct May and June inversions independently and derive similar total CH4 emission estimates for the SJV: 135â ±â 28â Mg/h in May and 135â ±â 19â Mg/h in June. The inversion result is 1.7 times higher than the prior estimate from CALGEM. We also use an independent mass balance approach to estimate CH4 emissions in the northern SJV for one flight when meteorological conditions allowed. The mass balance estimate provides a confirmation of our inversion results, and these two independent estimates of the total CH4 emissions in the SJV are consistent with previous studies. In this study, we provide optimized CH4 emissions estimates at 0.1° horizontal resolution. Using independent spatial information on major CH4 sources, we estimate that livestock contribute 75â 77% and oil/gas production contributes 15â 18% of the total CH4 emissions in the SJV. Livestock explain most of the discrepancies between the prior and the optimized emissions from our inversion.Key PointsEstimate of methane emissions in the San Joaquin Valley using inverse modeling and a mass balance approachMethane emissions are estimated to be greater than the bottomâ up inventory by a factor of 1.7Livestock largely account for differences between the optimized and prior methane emission estimates