Air-Sea Turbulent Fluxes From a Wave-Following Platform During Six Experiments at Sea

Turbulent fluxes at the air‐sea interface are estimated with data collected in 2011 to 2017 with a low‐profile platform during six experiments in four regions. The observations were carried out with moderate winds (2‐10 m s‐1) and averaged wave heights of 1.5 m. Most of the time, there was a swell, with an averaged wave age (the ratio between wave phase speed and wind speed) being equal to 2.8±1.6. Three flux calculation methods are used, namely the eddy‐covariance (EC), the inertial‐dissipation (ID), and the bulk methods. For the EC method, a spectral technique is proposed to correct wind data from platform motion. A mean bias affecting the friction velocity (u*) is then evaluated. The comparison between EC u* and ID u* estimates suggests that a constant imbalance term (ϕimb) equal to 0.4 is required in the ID method, possibly due to wave influence on our data. Overall, the confidence in the calculated u* estimates is found to be on the order of 10%. The values of the drag coefficient (CD) are in good agreement with the parameterizations of Smith (1988) in medium‐range winds and of Edson et al. (2013) in light winds. According to our data, the inverse wave age varies linearly with wind speed, as in Edson et al. (2013), but our estimates of the Charnock coefficient do not increase with wind speed, which is possibly related to sampling swell‐dominated seas. We find that the Stanton number is independent from wind speed. Plain language summary A small wave‐following platform was deployed in 2011‐2017 across four oceanic regions. The data are used to estimate turbulent fluxes, which are physical quantities that describe the exchanges of heat and momentum through the air‐sea interface. In weather models, simplified representations of the fluxes are used, which themselves depend on coefficients named drag coefficient for momentum exchange and Stanton number for temperature exchange, respectively. In this study, we evaluate these coefficients. First, we compare the flux estimates from the three main available methods. We adjust the parameters in the methods to reach the best possible agreement between the calculated fluxes. Two types of corrections are proposed, depending on the method considered, because turbulence data are modified by the motion of the platform and by the proximity of waves. Data are corrected by applying a mean bias to the fluxes and by accounting for a non‐measured term in the turbulence equations. Then, we analyze the wind dependence of the estimated drag coefficient and Stanton number. We find that drag is slowly increasing with wind speed, in agreement with existing models. Estimates of the Stanton number have biases, but which do not depend on wind speed