Evaluating gas transfer velocity parameterizations using upper ocean radon distributions

Sea-air fluxes of gases are commonly calculated from the product of the gas transfer velocity (k) and the departure of the seawater concentration from atmospheric equilibrium. Gas transfer velocities, generally parameterized in terms of wind speed, continue to have considerable uncertainties, partly because of limited field data. Here we evaluate commonly used gas transfer parameterizations using a historical data set of 222Rn measurements at 105 stations occupied on Eltanin cruises and the Geosecs program. We make this evaluation with wind speed estimates from meteorological reanalysis products (from National Centers for Environmental Prediction and European Centre for Medium-Range Weather Forecasting) that were not available when the 22Rn data were originally published. We calculate gas transfer velocities from the parameterizations by taking into account winds in the period prior to the date that 222Rn profiles were sampled. Invoking prior wind speed histories leads to much better agreement than simply calculating parameterized gas transfer velocities from wind speeds on the day of sample collection. For individual samples from the Atlantic Ocean, where reanalyzed winds agree best with observations, three similar recent parameterizations give k values for individual stations with an rms difference of ∼40% from values calculated using 222Rn data. Agreement of basin averages is much better. For the global data set, the average difference between k constrained by 222Rn and calculated from the various parameterizations ranges from −0.2 to +0.9 m/d (average, 2.9 m/d). Averaging over large domains, and working with gas data collected in recent years when reanalyzed winds are more accurate, will further decrease the uncertainties in sea-air fluxes.

Duke Scholars

Author Nicolas Cassar Earth and Climate Sciences