Search for a command to run...
Accurate methane (CH 4 ) emission estimates are essential for tracking mitigation progress, yet regional-scale estimates remain uncertain. In South Korea, CH 4 emissions are predominantly anthropogenic. This study applies a Bayesian inverse modeling approach, coupled with the STILT (Stochastic Time-Inverted Lagrangian Transport) model, to quantify monthly and annual CH 4 emissions from 2010 to 2021 using continuous in-situ surface observations. Posterior emissions ranged from 1.51 Tg yr -1 in 2011 to 1.75 Tg yr -1 in 2020, consistently lower than prior estimates. Annual reductions relative to the prior inventory varied from 3.2% to 9.3%, with a 12-year mean posterior emission of 1.66 Tg yr -1 , reflecting a national average reduction of 6.1%. Although multiple sources contribute to total CH 4 emissions, the pronounced springtime corrections are primarily linked to the strong seasonal sensitivity of agricultural emissions and the limitations of inventory-based temporal allocation. To evaluate the robustness of our findings, we compared the STILT-based results with regional inversions from the National Institute for Environmental Studies (NIES, Japan), conducted under the CH 4 Inversion Inter-Comparison for Asia (MICA) project. Using the same prior emissions and in-situ data, both STILT and NIES inversions showed comparable reductions 3.2-9.2% (STILT) and 0.8-12.5% (NIES). Differences between the two systems fell within estimated posterior uncertainties, indicating good inter-model agreement. Furthermore, posterior emissions from the STILT-based inversion significantly improved the agreement between modeled and observed CH 4 mole fractions, effectively capturing seasonal and short-term variability. These results highlight the importance of continuous observations and inverse modeling for identifying and correcting biases in bottom-up inventories, thereby supporting more effective CH 4 mitigation strategies. • Top-down inversion constrains methane emissions over South Korea (2010-2021). • Posterior emissions are 3.2–9.2% lower than prior estimates. • Inversion yields mean emissions of 1.66 Tg yr -1 (2010-2021). • Agricultural emissions are likely overestimated in inventories, especially in spring. • Inversion improves methane mole fractions agreement and variability.