Sea-to-air methane fluxes in heterogeneous shallow-water environments – how should we assess the integrated flux?

Coastal waters with depths less than 40 m contribute at least 90% to the marine methane flux to the atmosphere and these emissions partially offset the carbon burial efficiency of shallow-water inshore environments. Northern high-latitude shallow-water environments have summer productivity on a similar scale to many well-investigated tropical shallow-water ecosystems, but reliable spatial and temporal assessments of methane emissions have been difficult due to the high habitat diversity and the large seasonal variability in temperature and productivity. Here we report on methane fluxes from shallow bays in the archipelago seas of the central Baltic with a focus on both environmental and methodological factors controlling methane emissions for the period August 2020 to May 2021. Three methods were used to determine methane fluxes: floating chambers (FC), eddy covariance (EC), and thin-film boundary layer models (BL). We present 263 repeated FC flux measurements with corresponding BL calculations, and 3013 EC 30-minute flux periods and related these to environmental controlling factors in three different shallow-water ecotypes. The results showed that vegetation density and sediment type were poor predictors for methane fluxes during the period of our study, while eutrophication influences were clearly detectable. Water depth and distance to shore at the scale of <50 meters were not found to be statistically significant when determining methane flux, whereas the day hour of sampling influenced the results. Wind velocity and temperature have commonly been used to predict methane fluxes, but our results showed that wind was only influential for exposed bays and temperature did not appear to have a direct relationship with methane fluxes. The BL method underestimated the gas transfer at low wind speeds and the EC method showed a low signal to noise ratio, with the majority of the methane fluxes below the detection limit. Overall the three methods showed relatively good agreements, but in terms of sensitivity and correlation with environmental factors the FC method was the most suitable method for spatiotemporal scaling of methane fluxes in these complex inshore habitats.