Air-sea exchange of carbon dioxide is normally estimated by measurements of the difference of partial pressure across the air- water interface multiplied by an empirical exchange coefficient, which is a function of primarily wind speed and water temperature. Water chemistry i,e., the CO₂ buffer system, is normally not taken into account in the evaluation of the exchange coefficients and has often been regarded as a factor that can explain some of the uncertainties between the differences measurements and models. In this presentation we show that the effects of chemistry much to our surprise justify the use of the empirical exchange coefficient. The model is based on an assumption that for downward fluxes CO₂ in the buffer system in water near the surface may be described as a decaying tracer. This is justified by a timescale analysis of the CO₂ buffer system. An analytical expression for the water concentration profile is then developed. Based on the matching between the interface layer that consists of a very thin skin-layer, the water concentration profile and the profile of CO₂ in air we deduce an effective exchange coefficient to be used given a measurement at a certain height and depth. The model shows that the effective coefficient approaches the true exchange coefficient given at the surface. Furthermore the model predicts the depth where the gradients of CO_{2 in water vanish as function of wind speed which happen within 1 to 10 z0w (water roughness). The conclusion is thereby that gradients in water due to a downward flux of CO2 is practical impossible to measure.}