Reconciling Bottom–Up and Top–Down Approaches to Quantify Sub-Regional Methane Emissions with Improved Inventory and Three-Year High-Resolution Satellite Measurements

Inverse modeling of methane concentrations helps optimize bottom–up emission inventories, but its effectiveness at the subregional scale is challenged by insufficient observations and poor prior emission estimates. Here, we integrate bottom–up and top–down approaches to optimize methane emissions in the Guangdong-Hong Kong-Macao Greater Bay Area (GBA), the most populous and developed region in China. We use a process-based biogeochemical model with satellite-constrained rice distributions to estimate rice emissions and integrate newly available inventories for other sectors to compile an improved bottom–up inventory for the GBA. We then utilize three-year (2019–2021) TROPOMI observations to optimize GBA methane emissions at 0.25° × 0.3125°, with the improved bottom–up inventory and multiple compilations from other widely used inventories as prior estimates. Our results demonstrate that the uncertainty of posterior total emissions over the GBA (2.43–2.80 Tg a–1, 15%) is reduced by 57% compared to the prior range (1.77–3.04 Tg a–1, 72%), highlighting the strength of our approach. The best estimate of total emissions averages 2.70 Tg a–1 in 2019–2021. The inversion corrects the overestimated rice emissions over the Pearl River Estuary, where satellite observations reveal limited rice paddies, and prioritizes waste treatment (1.13 Tg a–1) as the largest anthropogenic source over the GBA, which is likely underestimated in most bottom–up inventories.