Uncertainty and Ambiguity in Air-sea Gas Fluxes; a Flexible System in Response to a Debate on Transfer velocity
Woolf, David1; Goddijn-Murphy, Lonneke2; Prytherch, John3; Yelland, Margaret3; Nightingale, Phil4; Land, Peter4; Shutler, Jamie4; Piolle, Jean-Francois5; Hanafin, Jenny5; Chapron, Bertrand5
1Heriot Watt University, UNITED KINGDOM; 2Environmental Research Institute, NHC-UHI, UNITED KINGDOM; 3National Oceanography Centre, UNITED KINGDOM; 4Plymouth Marine Laboratory, UNITED KINGDOM; 5IFREMER, FRANCE

The Oceanflux Greenhouse Gases project is concerned with the calculation of the air-sea flux of greenhouse gases, especially carbon dioxide, using an air-sea flux equation. Within the flux equation, fluxes are proportional to a transfer velocity, k. We can use Earth observation or in situ data for the calculation of the fluxes, only if we adopt a suitable algorithm that typically describes the dependence of k on wind speed and water temperature. Unfortunately, appropriate algorithms for the transfer velocity of carbon dioxide are hotly debated. The debate is not limited to an uncertainty in one or more parameters, ''parameter uncertainty'', but extends to a structural uncertainty arising from both conflicting data and rival theories. Measurements and parameterisations of gas transfer velocities are critically reviewed. Two substantially different empirical algorithms for k are arrived at by separate approaches. The most established algorithm is based on a simple Schmidt number dependence (and thus temperature dependence) and a simple quadratic dependence on wind speed. An alternative ''cubic'' algorithm shares the same temperature dependence, but predicts much higher transfer velocities in high wind speeds. Those high values in strong winds are primarily, but not solely, supported by micrometeorological measurements of carbon dioxide flux and are assumed to be the result of a roughly cubic wind speed dependence of wave breaking. Each empirical algorithm has strong proponents, but is contradicted by some data and criticised by other experts. Since there is a large body of data, but the perceived information is contradictory, more data is unlikely to resolve the debate swiftly. This ambiguity poses a dilemma for estimation of fluxes and their uncertainty. A twofold approach is adopted. Firstly, we examine critically both algorithms and pursue an alternative more mechanistic approach. Secondly, we construct an open system for the calculation of air-sea fluxes that enables a very broad range of algorithms to be applied.