Improving Methane Emission Estimates in Gippsland: Insights From Bayesian Inversion and In Situ Observations

Abstract This study refines methane emission estimates for southeastern Australia's Gippsland region in 2019 through high‐resolution atmospheric inverse modeling. Methane observations from three in situ monitoring sites were assimilated into a Bayesian four‐dimensional variational inversion framework coupled with the Community Multiscale Air Quality model, employing prior emissions from the Open Methane Database (2024). Posterior results revealed increased emissions in the eastern region, particularly near gas infrastructure and wastewater facilities, suggesting that these sources may be underrepresented in prior inventories. In contrast, a consistent decline in methane emissions was observed across all months in the central‐western part of the domain, where coal mining and power generation are concentrated. This decline is likely linked to the predominance of brown coal in the region, which emits substantially less methane than black coal, a distinction not clearly captured in the study inventory, potentially leading to overestimated emissions. Additionally, the most significant reduction in the entire studied domain occurred in February 2019, with emissions decreasing by approximately 30% relative to prior estimates. Sensitivity analyses and comparisons with the Emissions Database for Global Atmospheric Research inventory support the robustness of these findings. While total emissions across data sets were broadly similar, differences in spatial distribution led to notable variations in posterior outcomes. These results highlight the importance of incorporating local knowledge into emission inventories and the need for more detailed classification of emission sources, particularly coal type, to improve methane emission estimates at regional scales.