Air–water diffusive gas flux is commonly determined using measurements of gas concentrations and an estimate of gas transfer velocity (k600) usually derived from wind speed. The great heterogeneity of aquatic systems raises questions about the appropriateness of using a single wind-based model to predict k600 in all aquatic systems. Theoretical considerations suggest that wind speed to k600 relationships should instead be system-specific. Using data collected from aquatic systems of different sizes, we show that k600 is related to fetch and other measures of ecosystem size. Lake area together with wind speed provided the best predictive model of gas transfer velocity and explained 68% of the variability in individual k600 measurements. For a moderate wind speed of 5 m·s−1, predicted k600 varied from 6 cm·h−1 in a small 1 ha lake to over 13 cm·h−1 in a 100 km2 system. Wave height is also shown to be a promising integrative predictor variable. The modulating influence of system size on wind speed – gas transfer velocity relationships can have a large impact on upscaling exercises of gas exchange at the whole landscape level.
Heterogeneity Matters: Aggregation Bias of Gas Transfer Velocity Versus Energy Dissipation Rate Relations in Streams
