Amplified Arctic warming could substantially increase methane (CH₄) fluxes from high-latitudes by accelerating soil thaw and releasing soil organic carbon into the atmosphere. Yet, the magnitude and spatiotemporal patterns of current and future high-latitude CH₄ fluxes remain highly uncertain. We estimate growing season (May to September) CH₄ fluxes in Alaska from 2012 to 2017 using atmospheric observations from two NASA campaigns: the Carbon in Arctic Reservoirs Vulnerability Experiment (CARVE) and Arctic-Carbon Atmospheric Profiles (Arctic-CAP). We couple these observations with an atmospheric transport model and a geostatistical inverse model to estimate the geographic distribution, seasonality, and inter-annual variability of CH₄ fluxes. As part of this approach, we evaluate how well different state-of-the-art, process-based flux models compare to the inverse estimate of CH₄ fluxes. We find that estimated statewide fluxes are a factor of 1.3 to 2.9 higher compared to different ensembles of process-based models. This discrepancy is largest across the North Slope of Alaska – up to a factor of 6.5 when compared to one of the model ensembles. We do not find any statistically significant inter-annual variability (IAV) during years 2012–2017, and the IAV in our estimate is generally consistent with process-based models. In addition, our analysis reveals that differences in inundation fraction mapping contribute to the magnitude discrepancies in current process-based models, and future improvements in inundation maps are essential to reduce uncertainties in CH₄ flux estimates across high latitudes. These results provide an improved atmospheric constraint on regional CH₄ flux estimates from northern wetlands under accelerated Arctic warming.