Spatial and temporal variability of CO2 and CH4 gas transfer velocities and quantification of the CH4 microbubble flux in mangrove dominated estuaries

Gas transfer velocities (k) of CO2 and CH4 were determined from 209 deployments of a newly designed floating chamber in six mangrove dominated estuaries in Australia and the United States to estimate mangrove system specific k. k600-CO2 and k600-CH4 (k normalized to the Schmidt number of 600) varied greatly within and between mangrove creeks, ranging from 0.9 cm h−1 to 28.3 cm h−1. The gas transfer velocity correlated well with current velocity at all study sites suggesting current generated turbulence was the main driver controlling k. An empirical relationship that accounts for current velocity and a linearly additive contribution of wind speed and water depth was a good predictor of k600-CO2 (R2 = 0.67) and k600-CH4 (R2 = 0.57) in the mangrove creeks in Australia. In a side-by-side study, good agreement was found between kdetermined from this new floating chamber and a 3He/SF6 dual tracer release experiment (∼5% discrepancy). k600-CH4 correlated well with k600-CO2 (R2 = 0.81), however, k600-CH4 was on average 1.2 times higher than k600-CO2, most likely reflecting a microbubble flux contribution. The microbubble flux contributed up to 73% of the total CH4 flux and was best predicted by a model that included CH4 supersaturation, temperature, and current velocity. A large overestimation was found for both CO2 and CH4 fluxes when calculated using empirically derived k models from previous studies in estuaries. The high temporal and spatial variabilities of kCO2 and kCH4 highlights the importance of site specific transfer velocity measurements in dynamic ecosystems such as mangrove estuaries