ALBERT

Description: We present field measurements of air-sea gas exchange by the radon deficit method that were carried out during JASIN 1978 (NE Atlantic) and FGGE 1979 (Equatorial Atlantic). Both experiments comprised repeated deficit measurements at mixed position over periods of days or longer, using a previously described precise and fast-acquisition, automatic radon measuring system. The deficit time series exhibit variations that only partly reflect the expected changes in gas transfer. By evaluating averages over each time series we deduce the following average gas transfer velocities (average wind velocity and water temperature in parentheses): JASIN phase 1: 1.6 ± 0.8 mid (at -6 mis, 13 °C) JASIN phase 2: 4.3 ± 1.2 mid (at -8 mis, 13 °C) FGGE: 1.2 ± 0.4 mid (at -5 mis, 28 °C) 0.9 ± 0.4 mid (at -7 mis, 28 °C) 1.5 ± 0.4 mid (at -7 mis, 28 °C) The large difference between the JASIN phase 2 and FGGE values despite quite similar average wind velocity becomes even larger when the values are corrected to a common temperature. Both values are, however, fully compatible with the range of gas transfer velocities observed in laboratory experiments and the conclusion is suggested that their difference is caused by the highly different wind variability in JASIN and FGGE. We conclude that in gas exchange parameterization it is not sufficient to consider wind velocity only. A comparison of our observations with laboratory results outlines the range of variation of air-sea gas transfer velocities with wind velocity and sea state. We also reformulate the radon deficit method, in the light of our observed deficit variations, to account explicitely for non-stationarity and horizontal inhomogeneity in the near-surface radon deficit layer (i.e., mixed-layer and upper thermocline). We show that neglection of non-stationarity and horizontal inhomogeneity in previous radon work introduces considerable uncertainty in deduced gas transfer velocities. We furthermore discuss the observational requirements that have to be met for an adequate exploitation of the radon deficit method, of which an observation area of minimum horizontal inhomogeneity and monitoring of the remaining inhomogeneities are thought to be the most stringent ones.