Marine storm impacts on bubbles and air-sea exchange of gases

Starting from September 2002, two years of data from the Canadian SOLAS Northeast Pacific mooring deployed at station P (50N, 145W) have been recovered. The instrumentation includes a vertical array of dissolved gas sensors (pCO2, O2, N2), an upward looking echo-sounder for bubble measurements, an ambient sound recorder for wind and precipitation, and a variety of other sounders. This allows studies of air-sea gas exchange processes in the open ocean, especially in periods with high winds and sea states. These data are quite rare, in terms of collections of field data during storms.

In an analysis of time-series data of dissolved CO2 measurements during Hurricane Gustav in the NW Atlantic, Perrie et al. (2004, GRL) showed that the storm-induced CO2 flux can vary by a factor of two, comparing three recent gas transfer formulations (Wanninkhof and McGillis, 1999, GRL; Zhao et al. 2003, Tellus; Fairall et al. 2000, Bound. Layer Meteor.). But lack of any vertical resolution in the measurements meant that they could not rule out water advection and entrainment, rather than bubble-mediated air injection, as the cause of the increased flux.

A 7-day long storm case, with winds (10 m reference height) greater than 15 m/s lasting about 2 days during the 2003–2004 mooring period, is studied in this paper. As suggested by Woolf (1997), given a time series of gas concentrations, wave breaking and bubble cloud statistics and other environmental measurements are important considerations. We present our progress in terms of working to establish the dependence of direct, bubble-mediated gas transfer velocities, and the fractional supersaturation induced by the bubbles and related environmental factors. As a preliminary step in better understanding the gas exchange under high wind conditions, with the contributions from bubbles included, the bubble distribution and the possible dependence of bubble–induced gas transfer velocity on wave breaking and other environmental factors are included in our studies and simulations co-located with in situ observations.