Comparison of the Deliberate Tracer Method and Eddy Covariance Measurements to Determine the air/sea Transfer Velocity of CO2

During the international Air Sea Gas Exchange / MAGE (ASGAMAGE) field experiment in the Southern North Sea in 1996, deliberate tracer (DT) releases and simultaneous eddy covariance(EC) measurements of the CO2 flux were carried out to determine air/sea gas transfer rates. When expressed as a function of wind speed, EC-derived values of gas transfer velocity (k) are a factor of 2-2.5 higher than k-values from the DT experiment. Uncertainties in k, especially for the EC-derived values, are large. Possible reasons for the systematic differences and data scatter were investigated using 1-D and 2-D models of the upper-ocean boundary layer and atmospheric surface layer (ASL). Model outputs suggest that the DT method may underestimate k by 10-30%, due to near-surface vertical concentration gradients. The required correction depends on environmental conditions and sampling position relative to the centre of the tracer patch. Chemical bufferingof CO 2 renders the results from EC measurements much less sensitive to vertical gradients. However, horizontal heterogeneities in the aqueous concentration field can cause flux divergence in the ASL. At patch scales of less than a kilometre, this flux divergence might account for scatter of 20-50% in the true flux. Additional scatter of a similar magnitude might arise from instationarity of the aqueous CO 2 concentration, if the sources and sinks that cause the instationarity are located predominantly near the air-sea interface.