Introduction

The partitioning of anthropogenic CO₂ between atmosphere, land biosphere and the ocean is an important issue for predicting the future levels of atmospheric CO₂. The CO₂ fugacity (fCO₂) of the surface waters has been measured in the world ocean for estimations of oceanic CO₂ uptake. Higher fCO₂ in the surface ocean than in the atmosphere indicates evasion of CO₂ from the ocean, and, conversely, lower fCO₂ indicates invasion of CO₂ into the ocean. One of the direct estimations of the global air sea flux of CO₂ is to deduce it from the spatial and temporal integration of the local flux of CO₂, taken as the product of air-sea CO₂ partial pressure difference and of the air-sea exchange coefficient K [Tans et al., 1990]. There has been debate regarding the accuracy of K, although it is now possible to estimate its global variability in time and in space using satellite wind field data sets [Boutin and Etcheto, 1997]. Estimates of fCO₂ however still require in-situ observation. The comprehensive data set now covers the global ocean [Takahashi et al., 1993, 1997], although, the seasonal coverage is not complete. There have been efforts to fill in undersampled regions and seasons by parameterizing ΔfCO₂ (ocean minus atmosphere) using functions of temperature, salinity and productivity parameters. In the North Pacific, improvement of coverage in recent data sets of CO₂ fugacity measurements have been attempted using correlation of fCO₂ with sea surface temperature (SST) on a basin scale [Stephens et al., 1995]. However, the fCO₂-SST correlation differs regionally in the tropical, subtropical and subarctic Pacific within the meridional data sets [Landrum et al., 1996], and significant difference in the fCO₂-SST relationship are observed between the western and eastern North Pacific, even in the same season.

Surface ΔfCO₂ could be described by function controlled by physical processes such as temperature change, gas exchange at the air-sea interface and the vertical mixing of the sub-surface water of higher partial pressure of CO₂ from respiration. Biological uptake of dissolved CO₂ in seawater, relating to utilization of nutrients supplied by the vertical mixing should also be included in the function. However, it is complicated function of the nutrient concentrations given by the mixing and also the light intensity controlling the photosynthesis. Neglecting these processes may over simplify fCO₂-SST correlation and its resultant surface map for the North Pacific from satellite observed SST is not totally successful [Stephens et al., 1995].

Obtaining sufficient time series coverage with measurements from conventional research vessels is very difficult even in a fixed location such as station P [Wong and Chan, 1991]. The use of ships-of-opportunity is an effective way to obtain fCO₂ data with complete seasonal coverage along a fixed transects of the ocean [Wong, et al., 1993, 1995]. From March 1995, we started a program to monitor fCO₂ in the northern North Pacific on the cargo carrier M/S Skaugran (Seaboard International Shipping Co., Vancouver, Canada) as a bilateral collaboration between the National Institute for Environmental Studies in Japan and the Institute of Ocean Sciences in Canada [Bychkov and Saino, 1998]. A suite of analytical equipment was installed aboard a cargo ship travelling routinely between western North America and Japan. Continuous measurements of the concentrations of CO₂ in air and seawater were made, and periodic measurements of nutrients, total dissolved inorganic carbon (DIC), 13C in DIC, total alkalinity (TA), and plankton pigments. Additionally, discrete samples were taken for atmospheric CO₂ analysis.

In this monitoring program, the fCO₂ observation system uses a bubbling equilibrator, which was designed for rapid response and continuous monitoring. Continuous measurements using the rapid response system indicate dramatically more fCO₂ variability for some regions than observed from hourly data. This result has important ramifications when sampling in regions of high biological production and/or regions with large gradients in temperature and salinity.

Although pCO₂ intercalibration studies have been conducted in the laboratory and at sea, this is the first report of repeat measurements using companion systems operated over a range of conditions. These results expand the conclusions of the laboratory [Dickson, 1994] and seagoing [Körtzinger et al., 1999] intercomparison studies and provide a means of assessing the range of results from two shipboard systems which agree reasonably well (within 2 μatm) under controlled, laboratory conditions.