Motion correlated flow distortion of ship-based eddy covariance measurements

Eddy covariance measurements made from a ship are influenced by platform motion and airflow distortion. The direct effects of ship motion on the turbulence measurements are usually corrected by using a co-mounted motion sensor, but motion-correlated contamination often persists in the turbulence spectra. Contamination is likely to be caused by airflow distortion that is correlated to periodic ship motion. We developed a novel method of correcting the turbulent wind series for motion-correlated flow distortion. A linear fit between the turbulence and ship motion is removed from the turbulence series; this is repeated using several ship motion series until no correlations between turbulence and motion remain.

The results for bow-on flow are validated by comparison to cospectra that were interpolated over the ship motion frequency range; interpolation being a simpler correction method, albeit one with limitations of its own. The average results are a near-perfect match to the interpolated results above relative wind speeds of 10 metres per second. However they are biased high by 15% on average when records were sampled during low relative wind speeds and significant rolling motions. In such cases there were too many motion types correlated to the turbulence, and to each other, for a convergent solution to be found. 10 metre neutral drag coefficients calculated from the corrected momentum fluxes were also compared to the results of the Coupled Ocean-Atmosphere Response Experiment bulk flux algorithm version 3.0; a good match was found when the 10 metre true mean wind speed was above 10 metres per second. Two types of flow distortion were identified: one correlated to the vertical velocity of the bow; another correlated to the orientation of the bow.