Correction of simulation biases in stratospheric methane concentrations for the inverse analysis of satellite column observations

Global chemical transport models often overestimate stratospheric methane concentrations due to inaccuracies in simulating stratospheric circulation. If uncorrected, these biases can distort inverse analyses of satellite methane column observations (e.g., GOSAT), which encompass contributions from both the troposphere and stratosphere, and lead to erroneous estimates of surface emissions and tropospheric sinks. In this study, we assess the impact of stratospheric biases on the global inversion of satellite column observations. We implemented several correction methods, including empirically derived polynomial corrections, age-of-air proxies, and both offline and online replacements of stratospheric methane fields with independent observations. Correcting for stratospheric biases on average resulted in a 22 Tg a^-1 increase in inferred global methane emissions from GOSAT data, with notable latitudinal shifts. The correction increased extratropical emissions by 46 Tg a^-1 but decreased tropical emissions by 24 Tg a^-1. Different correction methods varied in their impact on the inversion estimates of emissions and OH concentrations, underscoring uncertainties in bias correction. Our results also indicate that stratospheric biases can induce tropospheric biases in the simulation through stratosphere-troposphere exchange, potentially affecting the analysis of surface methane observations.