[1] The influence of wind stress, small-scale waves, and surface films on air-sea gas exchange at low to moderate wind speeds (<10 m s À1 ) is examined. Coincident observations of wind stress, heat transfer velocity, surface wave slope, and surface film enrichments were made in coastal and offshore waters south of Cape Cod, New England, in July 1997 as part of the NSF-CoOP Coastal Air-Sea Chemical Fluxes study. Gas transfer velocities have been extrapolated from aqueous heat transfer velocities derived from infrared imagery and direct covariance and bulk heat flux estimates. Gas transfer velocity is found to follow a quadratic relationship with wind speed, which accounts for $75-77% of the variance but which overpredicts transfer velocity in the presence of surface films. The dependence on wind stress as represented by the friction velocity is also nonlinear, reflecting a wave field-dependent transition between limiting transport regimes. In contrast, the dependence on mean square slope computed for the wave number range of 40-800 rad m À1 is found to be linear and in agreement with results from previous laboratory wind wave studies. The slope spectrum of the small-scale waves and the gas transfer velocity are attenuated in the presence of surface films. Observations over largescale gradients of biological productivity and dissolved organic matter show that the reduction in slope and transfer velocity are more clearly correlated with surface film enrichments than with bulk organic matter concentrations. The mean square slope parameterization explains $89-95% of the observed variance in the data and does not overpredict transfer velocities where films are present. While the specific relationships between gas transfer velocity and wind speed or mean square slope vary slightly with the choice of Schmidt number exponent used to scale the heat transfer velocities to gas transfer velocities, the correlation of heat or gas transfer velocity with mean square slope is consistently better than with wind speed.

[1] GasEx-2001, a 15-day air-sea carbon dioxide (CO 2 ) exchange study conducted in the equatorial Pacific, used a combination of ships, buoys, and drifters equipped with ocean and atmospheric sensors to assess variability and surface mechanisms controlling air-sea CO 2 fluxes. Direct covariance and profile method air-sea CO 2 fluxes were measured together with the surface ocean and marine boundary layer processes. The study took place in February 2001 near 125°W, 3°S in a region of high CO 2 . The diurnal variation in the air-sea CO 2 difference was 2.5%, driven predominantly by temperature effects on surface solubility. The wind speed was 6.0 ± 1.3 m s À1 , and the atmospheric boundary layer was unstable with conditions over the range À1 < z/L < 0. Diurnal heat fluxes generated daytime surface ocean stratification and subsequent large nighttime buoyancy fluxes. The average CO 2 flux from the ocean to the atmosphere was determined to be 3.9 mol m À2 yr À1 , with nighttime CO 2 fluxes increasing by 40% over daytime values because of a strong nighttime increase in (vertical) convective velocities. The 15 days of air-sea flux measurements taken during GasEx-2001 demonstrate some of the systematic environmental trends of the eastern equatorial Pacific Ocean. The fact that other physical processes, in addition to wind, were observed to control the rate of CO 2 transfer from the ocean to the atmosphere indicates that these processes need to be taken into account in local and global biogeochemical models. These local processes can vary on regional and global scales. The GasEx-2001 results show a weak wind dependence but a strong variability in processes governed by the diurnal heating cycle. This implies that any changes in the incident radiation, including atmospheric cloud dynamics, phytoplankton biomass, and surface ocean stratification may have significant feedbacks on the amount and variability of air-sea gas exchange. This is in sharp contrast with previous field studies of air-sea gas exchange, which showed that wind was the dominating forcing function. The results suggest that gas transfer parameterizations that rely solely on wind will be insufficient for regions with low to intermediate winds and strong insolation.

No abstract

An oscillating grid-stirred tank was studied for flow repeatability and the existence of secondary circulations. For the particular tank studied, results indicate that mean flow values may vary by up to 25% and turbulent fluctuations may vary by up to 15% from one run to another. This result was observed to exhibit a potential grid geometry dependence. More importantly, there is evidence of significant flow field sensitivity to initial conditions. Particle image velocimetry results were used to reveal secondary mean flows in the grid-stirred tank. Because these characteristics are believed to be intrinsic to grid-stirred tanks, studies using such tanks must recognize and consider these effects.Oscillating grid-stirred tanks have long been used in experimental studies of turbulent mixing and transport processes in a variety of fluid systems (e.g., Refs. 1-5). The flow produced by an oscillating grid is the result of interactions between the individual jets and wakes created by the motion of the grid bars. At sufficient distance from the grid, these jets and wakes interact and break down into turbulence that is carried away from the grid by the jet motions. In concept, no mean flow exists and the turbulence is considered near isotropic and homogeneous in planes parallel to the driving grid.The oscillating grid-stirred tank examined here was used as part of a larger study of free-surface air-water gas transfer in the presence of subsurface turbulence. 6 As part of that investigation, several diagnostic experiments were conducted with the grid tank to quantify its particular characteristics. Among these were flow repeatability experiments and studies of secondary circulations. The findings from these investigations revealed some subtle characteristics of grid-stirred tanks that may have important consequences.The tank studied had dimensions 45ϫ 45ϫ 57 cm 3 and was fitted with a vertically oscillating grid, making it similar to tanks used by those cited above. Two stirring grids were used. Grid A was a 7 ϫ 7 square grid of solid bars. The cross section of each bar was d = 1.3 cm square, and the bars were spaced equally in both directions by M = 6.4 cm. Grid A closely resembled the grid used by Brumley and Jirka. 2 Grid B was a 5ϫ 5 square array of the same bars, spaced 8.9 cm apart in both directions. The grid was rigidly attached at its center point to a stainless steel shaft that passed through the tank floor with a watertight seal. A gearmotor-driven reciprocating mechanism beneath the tank vertically oscillated the shaft and grid. All experiments were performed with a water depth of 51 cm. The mean vertical position of the grid was located midway between the tank bottom and the free surface. For both grids, the distance from the extreme edge of the grid bar ends to the tank side wall was 4 mm. Several different grid stroke S and frequency f combinations were explored and are summarized in Table I. Other parameters appearing in Table I were determined using the empirical expressions of Hopfinger and Toly, 4 who fou...

No abstract