Surface seawater pCO2 and related parameters were measured at high frequency onboard the volunteer observing ship M/V Falstaff in the North Atlantic Ocean between 36¿ and 52¿N. Over 90,000 data points were used to produce monthly CO2 fluxes for 2002/2003. The air-sea CO2 fluxes calculated by two different averaging schemes were compared. The first approach used gas transfer velocity determined from wind speed retrieved at the location of the ship and called colocated winds, while for the second approach a monthly averaged gas transfer velocity was calculated from the wind for each grid pixel including the variability in wind. The colocated wind speeds determined during the time of passage do not capture the monthly wind speed variability of the grid resulting in fluxes that were 47% lower than fluxes using the monthly averaged wind products. The Falstaff CO2 fluxes were in good agreement with a climatology using averaged winds. Over the entire region they differed by 2--5%, depending on the time-dependent correction scheme to account for the atmospheric in increase in pCO2. However, locally the flux differences between the ship measurements and the climatology were greater, especially in regions north of 45¿N, like the eastern sector. A comparison of two wind speed products showed that the annual CO2 sink is 4% less when using 6 hourly NCEP/NCAR wind speeds compared to the QuikSCAT wind speed data.