The importance of integrating wave breaking for predicting air-water gas exchange in a large lake

Abstract:

Although there is a crucial need to assess lakes CO₂ emissions at a global scale, these fluxes are almost never directly measured. Instead, fluxes are often estimated from a restricted number of CO₂ concentration measurements in water (mostly during daytime) combined with modeled piston velocity using forcing data averaged by day, week or month. Yet, in large lakes, the short-term variability in surface CO₂ can be substantial enough to generate major inaccuracies in estimated fluxes. Besides, models for piston velocity integrate only a limited number of the physical and chemical mechanisms that drive the air/water gas exchanges. Their performance, although rarely tested, might vary depending on the seasonal contribution of wind shear, convection and wave breaking.

Here, we compared direct measurements of CO₂ fluxes in Lake Geneva, a large hardwater lake, from an automated (forced diffusion) flux chamber to computed values based on high frequency CO₂ measures and different models of piston velocity (k) of increased complexity (progressive integration of wind shear stress, convective mixing and wave breaking). Surveys were conducted at different time periods of the year on the new LéXPLORE platform in order to cover distinct weather conditions and surface CO₂ concentrations. We evaluated the performance of the different models, and identified the importance of considering wave breaking during wind events to improve the CO₂ flux estimation in large lakes. Altogether, we show how crucial the choice of k-models and the high-frequency of data are for CO₂ fluxes computations.