Constraining the Seasonal and Interannual Variability of Global Sectoral Methane Emissions in the 2010s using GOSAT XCH4 data

Atmospheric methane (CH₄), the second most important greenhouse gas, poses substantial uncertainties with its global emission inventory. We use inverse modeling analyses with Greenhouse Gases Observing Satellite (GOSAT) XCH₄ data to reduce those uncertainties and obtain improved quantitative estimates of sectoral monthly methane emissions from January 2010 to December 2019. We first conducted GEOS-Chem simulations with global emission inventories, including GFEIv2, EDGARv8, and WetCHARTs. The model with the global emission inventories showed a cumulative negative bias of approximately -1% per year compared to the GOSAT data, primarily due to the underestimation of tropical wetland emissions. Simulated monthly mean methane concentrations with the Kalman filter were used to optimize monthly variations of different sectoral CH₄ emissions over the decade, focusing on anthropogenic sources often assumed to be aseasonal in previous studies. Our inverse analyses resulted in increases of the global CH₄ emission trend of 3.86 Tg yr^-1 from 2.55 Tg year^-1, driven mainly by increases of agricultural and waste management sources. The seasonality of global methane emissions is more prominent in our top-down emission estimates than bottom-up emission, mainly driven by increased agricultural emissions in the Northern Hemisphere and tropical regions during June, July, and August. Furthermore, the top-down estimates of waste management emissions exhibited a significant summer peak in the Northern Hemisphere, indicating its temperature sensitivity, which was previously not recognized. The inverse analysis of methane emissions significantly reduced the spatiotemporal biases of the GEOS-Chem model compared to TCCON XCH₄, demonstrating the robustness of the inversion.