The ocean is an important sink for carbon and heat, yet high-resolution measurements of biogeochemical properties relevant to global climate change are being made only sporadically in the ocean at present. There is a growing need for automated, real-time, long-term measurements of CO2 in the ocean using a network of sensors, strategically placed oh ships, moorings, free-drifting buoys and autonomous remotely operated vehicles. The ground-truthing of new sensor technologies is a vital component of present and future efforts to monitor changes in the ocean carbon cycle and air-sea exchange of CO2. A comparison of a moored Carbon Interface Ocean Atmosphere (CARIOCA) buoy and shipboard fugacity of CO2 (fCO(2)) measurements was conducted in the western North Atlantic during two extended periods (> I month) in 1997. The CARIOCA buoy was deployed on the Bermuda Testbed Mooring (BTM), which is located 5 km noah of the site of the US Joint Global Ocean Flux Study (JGOFS) Bermuda Atlantic Time-series Study (BATS). The high frequency of sampling revealed that temperature and fCO(2) responded to physical forcing by the atmosphere on timescales from diurnal to 4-8 days. Concurrent with the deployments of the CARIOCA buoy, frequent measurements of surface fCO(2) were made from the R/V Weatherbird II during opportunistic visits to the BTM and BATS sites, providing a direct calibration of the CARIOCA buoy fCO(2) data. Although, the in situ ground-truthing of the CARIOCA buoy was complicated by diurnal processes, sub-mesoscale and fine-scale variability, the CARIOCA buoy fCO(2) data was accurate within 3+/-6 mu atm of shipboard fCO(2) data for periods up to 50 days. Longer-term assessments were not possible due to the CARIOCA buoy breaking free of the BTM and drifting into waters with different fCO(2)-temperature properties. Strategies are put forward for future calibration of other in situ sensors. (C) 2000 Elsevier Science B.V. All rights reserved.