A reconciliation of empirical and mechanistic models of the air-sea gas transfer velocity

Models of the air-sea transfer velocity of gases may be either empirical or mechanistic. Extrapolations of empirical models to an unmeasured gas or to another water temperature can be erroneous if the basis of that extrapolation is flawed. This issue is readily demonstrated for the most well-known empirical gas transfer velocity models where the influence of bubble-mediated transfer, which can vary between gases, is not explicitly accounted for. Mechanistic models are hindered by an incomplete knowledge of the mechanisms of air-sea gas transfer. We describe a hybrid model that incorporates a simple mechanistic view—strictly enforcing a distinction between direct and bubble-mediated transfer—but also uses parameterizations based on data from eddy flux measurements of dimethyl sulphide (DMS) to calibrate the model together with dual tracer results to evaluate the model. This model underpins simple algorithms that can be easily applied within schemes to calculate local, regional, or global air-sea fluxes of gases.

A reconciliation of empirical and mechanistic models of the air-sea gas transfer velocity

Models of the air-sea transfer velocity of gases may be either empirical or mechanistic. Extrapolations of empirical models to an unmeasured gas or to another water temperature can be erroneous if the basis of that extrapolation is flawed. This issue is readily demonstrated for the most well-known empirical gas transfer velocity models where the influence of bubble-mediated transfer, which can vary between gases, is not explicitly accounted for. Mechanistic models are hindered by an incomplete knowledge of the mechanisms of air-sea gas transfer. We describe a hybrid model that incorporates a simple mechanistic view—strictly enforcing a distinction between direct and bubble-mediated transfer—but also uses parameterizations based on data from eddy flux measurements of dimethyl sulphide (DMS) to calibrate the model together with dual tracer results to evaluate the model. This model underpins simple algorithms that can be easily applied within schemes to calculate local, regional, or global air-sea fluxes of gases.

Dimethylsulfide gas transfer coefficients from algal blooms in the Southern Ocean

Air–sea dimethylsulfide (DMS) fluxes and bulk air–sea gradients were measured over the Southern Ocean in February–March 2012 during the Surface Ocean Aerosol Production (SOAP) study. The cruise encountered three distinct phytoplankton bloom regions, consisting of two blooms with moderate DMS levels, and a high biomass, dinoflagellate-dominated bloom with high seawater DMS levels (> 15 nM). Gas transfer coefficients were considerably scattered at wind speeds above 5 m/s. Bin averaging the data resulted in a linear relationship between wind speed and mean gas transfer velocity consistent with that previously observed. However, the wind-speed-binned gas transfer data distribution at all wind speeds is positively skewed. The flux and seawater DMS distributions were also positively skewed, which suggests that eddy covariance-derived gas transfer velocities are consistently influenced by additional, log-normal noise. A flux footprint analysis was conducted during a transect into the prevailing wind and through elevated DMS levels in the dinoflagellate bloom. Accounting for the temporal/spatial separation between flux and seawater concentration significantly reduces the scatter in computed transfer velocity. The SOAP gas transfer velocity data show no obvious modification of the gas transfer–wind speed relationship by biological activity or waves. This study highlights the challenges associated with eddy covariance gas transfer measurements in biologically active and heterogeneous bloom environments.

Air-sea exchange of methanol and acetone during HiWinGS: Estimation of air phase, water phase gas transfer velocities

The air-sea fluxes of methanol and acetone were measured concurrently using a proton-transfer-reaction mass spectrometer (PTR-MS) with the eddy covariance (EC) technique during the High Wind Gas Exchange Study (HiWinGS) in 2013. The seawater concentrations of these compounds were also measured twice daily with the same PTR-MS coupled to a membrane inlet. Dissolved concentrations near the surface ranged from 7 to 28 nM for methanol and from 3 to 9 nM for acetone. Both gases were consistently transported from the atmosphere to the ocean as a result of their low sea surface saturations. The largest influxes were observed in regions of high atmospheric concentrations and strong winds (up to 25 m s(-1)). Comparison of the total air-sea transfer velocity of these two gases (K-a), along with the in situ sensible heat transfer rate, allows us to constrain the individual gas transfer velocity in the air phase (k(a)) and water phase (k(w)). Among existing parameterizations, the scaling of k(a) from the COARE model is the most consistent with our observations. The k(w) we estimated is comparable to the tangential (shear driven) transfer velocity previously determined from measurements of dimethyl sulfide. Lastly, we estimate the wet deposition of methanol and acetone in our study region and evaluate the lifetimes of these compounds in the surface ocean and lower atmosphere with respect to total (dry plus wet) atmospheric deposition.

Air-sea transfer of gas phase controlled compounds

Gases in the atmosphere/ocean have solubility that spans several orders of magnitude. Resistance in the molecular sublayer on the waterside limits the air-sea exchange of sparingly soluble gases such as SF6 and CO2. In contrast, both aerodynamic and molecular diffusive resistances on the airside limit the exchange of highly soluble gases (as well as heat). Here we present direct measurements of air-sea methanol and acetone transfer from two open cruises: the Atlantic Meridional Transect in 2012 and the High Wind Gas Exchange Study in 2013. The transfer of the highly soluble methanol is essentially completely airside controlled, while the less soluble acetone is subject to both airside and waterside resistances. Both compounds were measured concurrently using a proton-transfer-reaction mass spectrometer, with their fluxes quantified by the eddy covariance method. Up to a wind speed of 15 m s-1, observed air-sea transfer velocities of these two gases are largely consistent with the expected near linear wind speed dependence. Measured acetone transfer velocity is ~30% lower than that of methanol, which is primarily due to the lower solubility of acetone. From this difference we estimate the "zero bubble" waterside transfer velocity, which agrees fairly well with interfacial gas transfer velocities predicted by the COARE model. At wind speeds above 15 m s-1, the transfer velocities of both compounds are lower than expected in the mean. Air-sea transfer of sensible heat (also airside controlled) also appears to be reduced at wind speeds over 20 m s-1. During these conditions, large waves and abundant whitecaps generate large amounts of sea spray, which is predicted to alter heat transfer and could also affect the air-sea exchange of soluble trace gases. We make an order of magnitude estimate for the impacts of sea spray on air-sea methanol transfer.

Air-sea transfer of gas phase controlled compounds

Gases in the atmosphere/ocean have solubility that spans several orders of magnitude. Resistance in the molecular sublayer on the waterside limits the air-sea exchange of sparingly soluble gases such as SF6 and CO2. In contrast, both aerodynamic and molecular diffusive resistances on the airside limit the exchange of highly soluble gases (as well as heat). Here we present direct measurements of air-sea methanol and acetone transfer from two open cruises: the Atlantic Meridional Transect in 2012 and the High Wind Gas Exchange Study in 2013. The transfer of the highly soluble methanol is essentially completely airside controlled, while the less soluble acetone is subject to both airside and waterside resistances. Both compounds were measured concurrently using a proton-transfer-reaction mass spectrometer, with their fluxes quantified by the eddy covariance method. Up to a wind speed of 15 m s-1, observed air-sea transfer velocities of these two gases are largely consistent with the expected near linear wind speed dependence. Measured acetone transfer velocity is ~30% lower than that of methanol, which is primarily due to the lower solubility of acetone. From this difference we estimate the "zero bubble" waterside transfer velocity, which agrees fairly well with interfacial gas transfer velocities predicted by the COARE model. At wind speeds above 15 m s-1, the transfer velocities of both compounds are lower than expected in the mean. Air-sea transfer of sensible heat (also airside controlled) also appears to be reduced at wind speeds over 20 m s-1. During these conditions, large waves and abundant whitecaps generate large amounts of sea spray, which is predicted to alter heat transfer and could also affect the air-sea exchange of soluble trace gases. We make an order of magnitude estimate for the impacts of sea spray on air-sea methanol transfer.