A bloom of the coccolithophorid, E. huxleyi, occurred on the eastern Bering Sea shelf during September–October 2000. We examined the impact of this bloom on the CO2 system in the surface water. Drawdowns of total alkalinity (TAlk) from the values predicted by the TAlk-salinity conservative mixing relationship reached a maximum of 82.0 μmol kg−1, but was confined to latitudes 57.0°N–61.0°N. Surface water partial pressures of CO2 (pCO2) in excess of 400 μatm, depletion of nitrate + nitrite and low concentrations of silicate were also found, together with the TAlk drawdowns. The relationship between salinity-normalized TAlk and total CO2 suggests that the ratio of calcification to photosynthesis during the bloom was approximately 1.0, implying that any CO2 produced from calcification was balanced by photosynthesis. We discuss the possible cause of the observed high surface water pCO2 in the TAlk-drawdown (bloom) area.
 Coccolithophorids are a group of phytoplankton that produces plates of CaCO3 called ‘coccoliths’ surrounding their naked cells. Unlike non-CaCO3-producing-phytoplanktons, which exclusively produce organic matter by photosynthesis, blooms of coccolithophorids present a unique variation of the CO2 system in the ocean; in the open ocean, total alkalinity (TAlk) usually shows a linear relationship with salinity, and accordingly is controlled by the physical factors that regulate salinity (water mixing, precipitation and evaporation, etc.). However, if calcification occurs TAlk is reduced distinctly from the value expected from the linear relationship [e.g., Bates et al., 1996].
 During the MR00-K06 cruise (August–October 2000) by the R/V Mirai of the Japan Marine Science and Technology Center, we encountered aquamarine water in the eastern Bering Sea shelf, which often indicates a coccolithophorid bloom. Microscopic examination revealed that the dominant species of coccolithophorids was E. huxleyi (approx. 5,000,000 coccolithophorids L−1 by H. Okada, pers. comm.).
 The aim of the present study was to assess the impact of the bloom of E. huxleyi on the CO2 system in surface waters of the eastern Bering Sea shelf.
2. Materials and Methods
 Atmospheric and surface water partial pressure (pCO2), and surface water total CO2 (TCO2) content were measured continuously in the eastern Bering Sea shelf during September 1–7 (Line 1) and September 29–October 3 (Line 2), 2000 (Figure 1). These data were compared with pCO2 and TCO2 measured from 54.0°N, 176.1°W to 54.0°N, 161.9°W (Line 3; Figure 1) during August 14–15, 2000.
 The pCO2 was measured by a non-dispersive infrared analyzer (NDIR). For atmospheric pCO2, air from the bow of the ship was sampled into the NDIR. For surface water pCO2, air equilibrated with water taken from about 4.5 m depth within a showerhead-type equilibrator, was inputted into the NDIR. The calibration gases used were 240, 290, 310, 380 ppmv in a synthetic air, which are traceable to primary standard gases calibrated by Dr. C. D. Keeling of Scripps Institution of Oceanography (SIO). Values of TCO2 were calibrated against certified reference material (batches 45 or 48) provided by Dr. A. G. Dickson of SIO. Sea surface temperature (SST) and salinity (SSS) were measured continuously.
 TAlk was calculated from the 2° latitudinal or longitudinal averages of pCO2 and TCO2 along the ship's course.
 Along Line 1, nutrients (nitrate + nitrite, phosphate and silicate) were also sampled at 13 hydrocast stations (Figure 1) at depths between 0–5 m.
3.1. Distributions of the CO2-System Parameters
 Surface water pCO2 was highly variable ranging from 220 to 440 μatm along Lines 1 and 2, while atmospheric pCO2 was nearly constant (358 and 363 μatm for Lines 1 and 2, respectively) (Figure 2). Correspondingly, undersaturation (sink) and supersaturation (source) for atmospheric pCO2 appeared alternately with a small spatial scale. However, the distributions of sinks and sources were generally similar between Lines 1 and 2, despite the temporal and spatial differences between sampling (Figure 1); between latitudes 57°N–63°N, surface waters were supersaturated, while in the south and north, surface waters were undersaturated. The maximum level of supersaturation for both Lines 1 and 2 was +80 μatm, and the minimum level of undersaturation was approximately −130 μatm.
3.2. Drawdowns of TAlk Due to Calcification
 Calculated TAlk and salinity were significantly positively correlated along Line 3, outside the range of the coccolithophorid bloom (n = 60, R2 = 0.932; P = 0.05; Figure 3). While at higher salinities calculated TAlk along Lines 1 and 2 appeared to fit the regression line for Line 3, values for surface waters with salinities less than 30.6‰, deviated from the linear regression line (Figure 3), probably due to mixing with riverine water. The values of TAlk for Lines 1 and 2, in the area of the coccolithophorid bloom, between salinity range 31.0 and 32.0‰ were significantly smaller than the values expected from the linear regression line. Thus, it is inferred that calcification occurred in the salinity range, corresponding to latitudes from 57.0°N to 60.8°N and 56.8°N to 60.8°N along Lines 1 and 2, respectively. The maximum drawdowns of TAlk in the salinity range were 82 μmol kg−1 for Line 1 and 73 μmol kg−1 for Line 2.
3.3. Relationships Between the Coccolithophorid Bloom and Nutrients
 NOx was almost depleted (0.05–0.26 μmol kg−1) at latitudes north of 58°N along Line 1 (Figure 4a). The NOx depleted latitudes corresponded to the TAlk-drawdown latitudes (57.0°N–60.8°N). SiO4 showed a sharp decrease of about 20 μmol kg−1 between 55.0°N and 55.5°N, and was smaller than 10 μmol kg−1 at latitudes north of 58°N (Figure 4b). There was a significant linear relationship between salinity-normalized SiO4 and salinity-normalized NOx for latitudes south of 55°N (slope = 1.75 ± 0.24; intercept = 6.81 ± 1.18; n = 16; R2 = 0.948; P < 0.05). The distribution of PO4 was similar to that of SiO4 at latitudes 54°N–60°N, but increased more at latitudes 62.5°N–64°N (Figures 4b and 4c). Concentrations of PO4 at latitudes 58.2°N–61°N, close to the TAlk-drawdown latitudes, were >0.41 μmol kg−1.
3.4. Photosynthesis Versus Calcification
 Theoretically, if photosynthesis occurs simultaneously with calcification, TAlk and TCO2 change together, and the changing ratio of TAlk to TCO2, which equals the ratio of calcification to photosynthesis, stays between 1.0 and 2.0 [Robertson et al., 1994].
 Linear regression analysis revealed that TAlk normalized by salinity (nTAlk) changed against TCO2 normalized by salinity (nTCO2) at a ratio of 1.1 ± 0.3 (n = 18, R2 = 0.818, P < 0.05) between 57.0°N–60.8°N on Line 1 and 1.1 ± 0.2 (n = 21, R2 = 0.851, P < 0.05) between 56.8°N–60.8°N on Line 2 (the areas of TAlk-drawdown (bloom)). Statistically, these ratios can be regarded as 1.0, implying that the ratio of calcification to photosynthesis was about 1.0. The influence of NOx consumption on TAlk [Brewer and Goldman, 1976] was ignored.
 In the TAlk-drawdown (bloom) area, we found surface water pCO2 exceeded 400 μatm, supersaturated with atmospheric CO2. The elevated pCO2 in summer contrasts with the data of Dr. T. Takahashi (Columbia University; home page address: http://ingrid.ldgo.columbia.edu/), which indicate undersaturation, although other studies [Kelley and Hood, 1971; Kelley et al., 1971] have reported high pCO2 levels. We also observed elevated pCO2 along the north-south transect lines in the eastern Bering Sea shelf in 1998 and 1999 (unpublished data). It is worthy to note that extensive coccolithophorid blooms occurred in these years [Vance et al., 1998; Hunt et al., 1999]. During this previous study calculated TAlk significantly decreased with elevated pCO2, consistent with the results of the present study. We thus conclude that coccolithophorid blooms in the eastern Bering Sea shelf result in increased surface water pCO2 levels.
 In this study, the predicted ratio of calcification to photosynthesis was approximately 1.0, implying little possibility of the first mechanism. However, from theoretical considerations of diurnal net calcification and net photosynthesis, and nocturnal respiration of E. huxleyi, Crawford and Purdie  pointed out that increases in pCO2 can occur without invoking diurnal net calcification: net photosynthesis >1.0. With the present available data, we cannot examine this possibility. Nevertheless, pCO2 in excess of 400 μatm could not be obtained via this process. Since there was no evidence of a rise in SST in the bloom area compared to SST in the ambient area, the second mechanism also seems unlikely. Robertson et al.  used the buffer or Revelle factor to assess the relative impact of photosynthesis versus calcification on the CO2-system equilibrium. They showed that the buffer factor calculated from pCO2 corrected to a constant temperature and TCO2 normalized to a constant salinity decreases with increasing calcification. We attempted to estimate the buffer factor from a slope of a log-log plot of the corrected pCO2 and the normalized TCO2 [Takahashi et al., 1985] in the TAlk-drawdown (bloom) area. However, the plots were so scattered that any significant linear relationship could not be obtained for both Lines 1 and 2. Thereby the possibility of the third mechanism (CO2-system equilibrium by removal of HCO3−) seems to be small.
 None of the previously proposed mechanisms adequately explained the elevated pCO2 in the eastern Bering Sea shelf. However, depletion of NOx and low concentrations of SiO4 (section 3.3) suggest that activity of diatoms was not dominant, which is advantageous to coccolithophorid blooms [Brown and Yoder, 1994]. Previous studies have reported that activities of diatoms are dominant in the eastern Bering Sea shelf [e.g., Sambrotto et al., 1986]. Diatoms can reduce levels of surface water pCO2 by photosynthesis, just as found in latitudes 55°N–57°N along the Line 1. In contrast, the pCO2 would remain almost unconsumed if activity of diatoms was weak. Thus, conditions conducive to coccolithophorid blooms, but not to diatoms, may have limited the reduction of surface water pCO2 in the bloom area.
5. Concluding Remarks
 In the present study, we found that surface water pCO2 exceeded 400 μatm in the coccolithophorid bloom area. As a result, the area acted as a source for atmospheric CO2 (supersaturation = +80 μatm). The eastern Bering Sea shelf has been known to act as a sink for atmospheric CO2 in summer due to photosynthesis [Gordon et al., 1973]. Since the supersaturation as a result of the bloom was localized, the bloom of E. huxleyi seems to not have a large impact on the planetary-scale carbon budget. However, such an extensive bloom of E. huxleyi is not a phenomenon which occurred only in 2000 [Vance et al., 1998; Hunt et al., 1999]. Now it is inferred that the blooms are a result of climate change, with possible ecosystem-scale effects in the eastern Bering Sea shelf [Vance et al., 1998; Hunt et al., 1999]. Further study is required to investigate what aspects of climate change triggered the blooms.
 We wish to thank the officers and crew of the R/V Mirai for exceptional support during the cruises. We also give special thanks to the staff of Marine Works Japan, who worked as physical and chemical oceanography marine technicians onboard the Mirai.