Parameterising Gas-Transfer Velocity Using Momentum Flux or Wind Speed? Implications for the Reporting of Measurements from Ships at Sea.

Wind forcing plays a dominant role in air-sea exchange processes. Therefore, field measurements of relevant parameters, such as gas-transfer velocity or momentum flux, are commonly reported as functions of the measured wind speed normalised to a height of 10 metres and neutral conditions (u_10N). Wind speed-based parameterisations are used for the forcing of physical models.

Many studies have shown that the wind speed measured from large structures such as research vessels can be affected by airflow distortion leading to over- or under-estimations in the order of 10 percent, depending on the relative direction of the ship to the wind field. If not appropriately corrected for, this can lead to invalid conclusions when comparing measurements made during one experiment, or from different experiments on different platforms. Numerical models have been developed to predict the influence of the measurement platforms on the wind field. However, these models require a fine resolution of the structures nearby the anemometer and become invalid once new structures are added or removed. The effect of airflow distortion on direct fluxes is less well studied. Landwehr et al. (2015) showed that inaccuracies in the methods of analysis could account for previously observed biases in ship-borne momentum and scalar flux measurements.

The present study uses data from two research cruises on two different vessels, R\V-Knorr in the North Atlantic and R\V-Tangaroa in the South Pacific. We compare wind speed and direct flux measurements of the friction velocity and the transfer velocities of two trace gases, CO₂ and DMS. Results show a better agreement between the derived k_gas to u_* correlations of the two experiments than for k_gas and u_10N, even when airflow distortion models are used to correct the measured wind speed. This implies that airflow distortion affects direct flux measurements less than wind speed measurements, or is at least consistent for momentum and scalar fluxes. We conclude that the direct u_* measurements should be preferred over u_10N for the use as independent variable to parameterise trace gas fluxes.