Abstract
Monitoring methane (CH(4)) emissions from terrestrial ecosystems is essential for assessing the relative contributions of natural and anthropogenic factors leading to climate change and shaping global climate goals. Fires are a significant source of atmospheric CH(4), with the increasing frequency of megafires amplifying their impact. Global fire emissions exhibit large spatiotemporal variations, making the magnitude and dynamics difficult to characterize accurately. In this study, we reconstruct global fire CH(4) emissions by integrating satellite carbon monoxide (CO)-based atmospheric inversion with well-constrained fire CH(4) to CO emission ratio maps. Here we show that global fire CH(4) emissions averaged 24.0 (17.7-30.4) Tg yr(-1) from 2003 to 2020, approximately 27% higher (equivalent to 5.1 Tg yr(-1)) than average estimates from four widely used fire emission models. This discrepancy likely stems from undetected small fires and underrepresented emission intensities in coarse-resolution data. Our study highlights the value of atmospheric inversion based on fire tracers like CO to track fire-carbon-climate feedback.