Coccolith Carbonate Fluxes in the Northwest Pacific Ocean

133

Coccolith Carbonate Fluxes in the Northwest Pacific Ocean

Yuichiro TANAKA

Institute for Marine Resources and Environment, National Institute of AdvancedIndustrial Science and Technology, Higashi 1-1-1, Tsukuba, Ibaraki 305-8567, Japan

Abstract. The coccolithophore fluxes and calculated coccolith CaCO3 fluxeswere investigated in three time-series sediment traps, deployed for a one-yearperiod in the West Caroline Basin in the western equatorial Pacific Ocean andin the central equatorial Pacific Ocean. These sediment trap stations werelocated at 3°N, 134°E (Site NT1) and 4°N, 135°E (Site NT2) in the westernPacific warm water mass region, and at 0°N, 175°E (Site NT3) in the centralequatorial Pacific Ocean region, at water depths of approximately 1500 m. Thegreatest coccolithophore fluxes were observed at Site NT1 in September andDecember 1991, at Site NT2 from February to March 1992, and at Site NT3from September through to November 1993. The calculated calcium carbonatefluxes of these coccolithophores at these three sites were 1.6, 1.5, and 1.4g m–2year–1 at Sites NT1, NT2 and NT3, respectively. These calculated valuescorrespond to only 23.3% of the total CaCO3 flux at Site NT1, 9.5% at Site NT2,and 11.6% at Site NT3.

Florisphaera profunda, Gladiolithus flabellatus, Gephyrocapsa oceanica,Umbilicosphaera sibogae, Emiliania huxleyi, Umbellosphaera irregularis andOolithotus fragilis were the most abundant species in the coccolith flora,together comprising more than 85% of the total flora at the three Sites.Florisphaera profunda accounted for more than 50% of total flora throughoutall the periods at the three sites. On the other hand, of the mass CaCO3 fluxcalculated for the coccolithophores, O. fragilis accounted for about 40% ormore throughout the year at site NT3, and in winter at Site NT1. At Site NT3,Calcidiscus leptoporus was also an important carbonate-contributing species.The contribution of these species to the export production of carbonate appearsto reflect the nutrient supply of the western equatorial Pacific Ocean.

Keywords: coccolithophore, coccolith, sediment trap, Pacific, flux, carbonate

1. INTRODUCTION

The interaction of the atmosphere and the ocean in the tropical Pacific Ocean isan important factor in understanding the climate system of the earth. In thetropical Pacific Ocean, the strength and distribution of the upwelling changegreatly as the El Niño Southern Oscillation (ENSO) progresses (Yan et al., 1992).ENSO is closely related to the monsoon strength in Southeast Asia, and to thedevelopment of upwelling in the Southeast Asian coastal zone. In particular, theequatorial upwelling region in the Western Pacific Warm Pool is an area of highsea-surface temperatures throughout the year, and the eastern part is also important

Global Environmental Change in the Ocean and on Land, Eds., M. Shiyomi et al., pp. 133–146.© by TERRAPUB, 2004.

134 Y. TANAKA

to our understanding of global climatic variation.Coccolithophores are an important primary producer in the temperate oceans

and are a major transporter of inorganic carbon from the sea’s surface to the deepsea floor. An understanding of how changes in the geographical distribution andstanding crop of coccolithophores influence the flow of carbon dioxide betweenthe atmosphere and ocean is therefore extremely important in elucidating thecarbon cycle. Moreover, it is important for us to understand the contribution ofcoccolithophores to the total carbonate flux as a result of their calcium carbonateplates. A continuous sampling of the settling particle flux was facilitated bysediment-trap experiments, allowing analyses of the seasonal and geographicchanges in zooplankton and phytoplankton with plates and tests.

Recently, many researchers have estimated the settling particle fluxes of theocean water column. An accurate estimate of carbonate production bycoccolithophores is important in evaluating the role of carbonate in the materialcycles in the ocean. In previous sediment-trap research, the contribution of thecoccolith carbonate flux to the total carbonate flux has been estimated to liebetween 60% and 80% (Honjo, 1977; Deuser and Ross, 1989; Fabry, 1989). Theerror in the estimated values is large because the estimated carbonate values forseveral species of coccoliths were extrapolated to all coccoliths, although theamount of carbonate in coccoliths has hitherto been estimated by a variety ofmethods. Young and Ziveri (2000) designed a method to determine the volumeof carbonate in various coccoliths from the morphological analysis of individualspecies. The estimation of coccolith carbonate flux for the coccolithophore ofindividual species is now possible using this method.

The purpose of this study was to clarify: (1) the flux and carbonate changesin coccolithophores collected in a time-series of sediment traps deployed indifferent areas in the equatorial Pacific Ocean, and (2) the difference in theircarbonate flux contribution between the low- and mid-latitude regions in thenorthwest Pacific Ocean.

2. OCEANOGRAPHIC SETTING

Very high sea-surface temperatures were recorded in the western equatorialPacific Ocean during one year. Trap Site NT1 used in the present study waslocated in the West Caroline Basin off the coast of New Guinea, which isdominated by the South Equatorial Current (SEC). Trap Site NT2 was locatednorthwest of Site NT1, in the Equatorial Counter Current (ECC) (Fig. 1).

In the West Caroline Basin, the North Equatorial Current (NEC) has awestbound flow that can be observed between 25°N and 5°N near 170–180°E inFebruary; in August, this current does not flow in the south from 10°N. When theNEC reaches the western end of the Pacific, the current meets the continentalbarrier of the Philippines, and it divides into two branches. One branch turnssouth and feeds the eastbound the ECC (Tchernia, 1980). The New Guinea coastalcurrent flows northwestward with the southeasterly winds in the boreal summer,and flows southeastward with the northwesterly winds in the boreal winter.

Coccolith Carbonate Fluxes in the Northwest Pacific Ocean 135

The Indonesian archipelago is strongly influenced by a monsoon-typeclimate. Generally, the northwest monsoon lasts from about December to Februaryand the southeast monsoon from June to August. The rest of the year representstransition periods from northwest to southeast monsoons (March–May) and fromsoutheast to northwest monsoons (September–November). During the northwestmonsoon, the wind blows eastward and causes heavy rainfall throughout most ofthe western parts of the Indonesian archipelago (Miyama et al., 1996).

Salinity is reasonably low (34.6–33.9 psu) in September during the rainyseason, although in the area studied, the temperature of the photic zone remainsalmost constant (28–30°C) throughout the year (Levitus and Boyer, 1994).

Relative to the West Equatorial Current, the sea-surface water is comparativelycold and abundant nutrients flow in the equatorial central Pacific Ocean whereSite NT3 was located.

3. MATERIALS AND METHODS

In the equatorial Pacific Ocean, time-series sediment traps (PARFLUXMark 7G-21: opening 0.5 m2) were deployed at a water depth of 1592 m at SiteNT1 (2°59.8′ N, 135°1.5′ E); at 1769 m at Site NT2 (4°7.5′ N, 136°16.6′ E) in theWest Caroline Basin; and at 1357 m at Site NT3 (0°0.2′ N, 175°9.7′ E) in thecentral equatorial Pacific (Fig. 1). Twenty-one time-series samples were collectedat each site from June 1991 to April 1992 at Sites NT1 and NT2, and from June1992 to April 1993 at Site NT3 for 15-day intervals. Samples were treated withformalin in a 3% solution buffered with sodium borate.

In the central North Pacific Ocean, four time-series sediment traps weredeployed at a 1412 m water depth at Site NT8 (46°07.2′ N, 175°01.9′ E), 1482 mat Site NT7 (37°24.2′ N, 174°56.7′ E), 1342 m at Site NT5 (34°25.3′ N, 177°44.2′E), and 3872 m at Site NT6 (30°00.1′ N, 174°59.7′ E), respectively. Eachsediment tap performed continuous time-series sampling for 15-day intervals,

Fig. 1. Locations of the time-series sediment-trap moorings at Sites NT1, NT2, and NT3.

136 Y. TANAKA

except for samples 1 to 7, which represents the 30-day intervals of Sites NT8,NT5, and NT6 from June 1993 to July 1994.

Time-series sediment trap samples were first wet-sieved through a 1 mmmesh to remove large particles. Each sample of <1 mm was split into severalaliquots (1/256 to 1/1014) with a high-precision rotary splitter. Each split samplewas wet-sieved over a 32 µm mesh, and fecal pellets and aggregates weredestroyed using a soft brush. The fraction <32 µm was filtered through Milliporefilters (47 mm diameter, 0.45 µm pore size) using a low pressure vacuum pump.After filteration procedure, the filters were placed in plastic Petri dishes and airdried for one day, and then kept in closed boxes with silica gel. A small sectionof each filtered membrane was analyzed using a polarized-light optical microscopeat 1250 × magnification to determine the total coccolith and coccosphere fluxesand the coccolith fluxes of coccolith species. The total number of coccolithscounted in each sample was more than 1500. At least 150 coccospheres wereidentified per sample. Species identification of coccoliths was performed byscanning electron microscopy. At least 300 coccoliths were identified per sample.A JEOL JSM-5410LV scanning electron microscope (SEM) was used fortaxonomic identification, except for Florisphaera profunda and Gladiolithusflabellatus, the fluxes of which were determined by light microscopy. Thecounting procedure and calculation of coccolith flux followed the method ofZiveri et al. (1995). The coccolith and coccosphere fluxes of Site NT1 used databy Tanaka and Kawahata (2001).

Coccolith carbonate fluxes were calculated using estimates of carbonatemass per coccolith for coccolith species, based on the morphology such as theshape and size of coccolith of individual species, according to Young and Ziveri(2000). Coccosphere carbonate fluxes were first converted into coccolith fluxesand then transformed into coccolith carbonate fluxes.

4. RESULTS AND DISCUSSION

4.1 Coccolith and coccosphere fluxes in the equatorial Pacific Ocean

At Site NT1, the coccolith flux ranged from 490 × 106 to 3400 × 106

coccoliths m–2day–1; the mean flux was 1800 × 106 coccoliths m–2day–1 (Tanakaand Kawahata, 2001) (Fig. 2). Coccolith fluxes increased during September toearly October 1991, and during late November 1991. The highest fluxes wererecorded from late December 1991 to January 1992, with a maximum abundanceof 3600 × 106 coccoliths m–2day–1. Low fluxes occurred during late spring 1991(Fig. 2). The coccosphere flux pattern paralleled that of the total coccolith flux,with peak fluxes of 3.2 × 106 coccospheres m–2day–1 in late September 1991, and3.6 × 106 coccospheres m–2day–1 in late December 1991. The mean coccosphereflux was 1.9 × 106 coccospheres m–2day–1.

At Site NT2, the coccolith flux ranged from 71 × 106 to 500 × 106 coccolithsm–2day–1, and the peak coccolith flux was observed from January through toMarch, with the highest value in the latter half of February 1992 (Fig. 3). The peakcoccosphere flux was recorded from the latter half of February to the first half of

Coccolith Carbonate Fluxes in the Northwest Pacific Ocean 137

March 1992, and the highest value was 6 × 106 coccospheres m–2day–1. Neitherthe coccolith nor coccosphere fluxes changed from June through December 1991.Two peaks in the coccolith flux were recorded, in June and from the latter half ofSeptember to the first half of November 1991.

At Site NT3, the highest value was 800 × 106 coccoliths m–2day–1 in the latterhalf of September 1992. The changes in coccosphere flux were also similar tothose in the coccolith flux (Fig. 4).

4.2 Coccolith flux patterns of individual species in the equatorial Pacific Ocean

Forty-one taxa were identified among the material in the sediment traps.Seven species (Florisphaera profunda, Gladiolithus flabellatus, Gephyrocapsa

Fig. 2. Seasonal coccolith flux rates and comparison of total carbonate and calculated coccolithcarbonate fluxes at Site NT1 (after Tanaka and Kawahata, 2001).

138 Y. TANAKA

oceanica, Umbilicosphaera sibogae, Emiliania huxleyi, Umbellosphaerairregularis, and Oolithotus fragilis) dominated the coccolith fluxes, and comprisedmore than 85% of the total coccolith assemblages.

In the Site NT1 trap, the typical deep-water species, F. profunda, occurredat levels exceeding 50% of the total flora. In August, and from late December1991 to early January 1992, F. profunda comprised more than 60%, withparticularly high abundance among the coccolith fluxes (Tanaka and Kawahata,2001) (Fig. 5). Two seasonal maxima for G. flabellatus fluxes were observed,during August and early September 1991, and late December 1991 to January1992. Gephyrocapsa oceanica exhibited higher fluxes during summer and earlyautumn 1991, whereas E. huxleyi dominated the winter assemblages. Theabundance of U. sibogae was high in September 1991. In autumn, the occurrence

Fig. 3. Seasonal coccolith flux rates and comparison of total carbonate and calculated coccolithcarbonate fluxes at Site NT2.

Coccolith Carbonate Fluxes in the Northwest Pacific Ocean 139

of O. fragilis was much higher than in any other season. Umbellosphaerairregularis was commonly observed during the trap experiments, with peakfluxes in July and November 1991.

In the Site NT2 trap, F. profunda dominated, accounting for 50% or more ofthe total flora, as at Site NT1, and this increased to 80% especially in November1991. Gladiolithus flabellatus comprised 20% of the flora. The peak of G.oceanica occurred in the latter half of February 1992, and E. huxleyi increased inMarch 1992, but the abundances of those two species were low in other seasons(Fig. 5).

At Site NT3, the coccolithophore assemblages did not display remarkableseasonal changes. Florisphaera profunda was the most abundant species of thecoccolith flora, altogether comprising more than 50% of the total flora.Gephyrocapsa oceanica accounted for 10%, O. fragilis and U. sibogae 5–10%,and C. leptoporus 5%.

Fig. 4. Seasonal coccolith flux rates and comparison of total carbonate and calculated coccolithcarbonate fluxes at Site NT3.

140 Y. TANAKA

Fig. 5. Relative abundance of different species at Sites NT1, NT2, and NT3 in terms of (a) numericalcoccolith fluxes and (b) coccolith CaCO3 mass estimates.

Coccolith Carbonate Fluxes in the Northwest Pacific Ocean 141

4.3 Calculated coccolithophore carbonate in the equatorial Pacific Ocean

At Site NT1, calculated coccolith CaCO3 fluxes ranged between 6.8 and 49.6mg m–2day–1, with an average flux of 16.6 mg m–2day–1, and contributed 23.3%of the total CaCO3 flux at this Site in the West Caroline Basin (Tanaka andKawahata, 2001) (Table 1). The coccolith CaCO3 contribution to the total CaCO3flux was higher in late September and late December 1991.

At Site NT2, calculated coccolith CaCO3 fluxes ranged between 1.5 and 9.8mg m–2day–1, and the peak fluxes were observed in the latter half of February1992 and the latter half of May 1992. Coccolith CaCO3 contributed 9.5% of thetotal CaCO3 (Table 1).

At Site NT3, calculated coccolith CaCO3 fluxes ranged between 0.4 and 16.2mg m–2day–1, and peak fluxes were observed in the latter half of June 1992, andfrom the latter half of September to the latter half of October 1992. CoccolithCaCO3 contributed 11.6% of the total CaCO3 (Table 1).

4.4 Coccolithophore CaCO3 contribution in the equatorial Pacific Ocean

In general, coccolithophores are considered as main calcium-carbonateproducers in the ocean (e.g., Westbroek et al., 1994; Takahashi, 1994). Broerse(2000b) clarified the contribution of the coccolithophores to total carbonateproduction. Sediment-trap samples from various areas were examined, and thecontribution of the coccolith carbonate flux to the total carbonate flux wasbetween 5% and 30% (Broerse et al., 2000b; Ziveri et al., 2000; Broerse et al.,2000a; Broerse, 2000a), which is lower than the values reported here. However,the calculated carbonate flux was lower than the carbonate flux of the finefraction (<32 µm), which was derived from coccolithophores, and a cumulativeerror of about 50% is inherent in this method (Young and Ziveri, 2000). Thesource of this cumulative error is the presence of other biogenic carbonatesderived from juvenile and fragmented foraminifera, pteropod and other molluscfragments, and non-biogenic carbonates of detrital origin (Broerse et al., 2000a,

Table 1. Total CaCO3, coccolith CaCO3 mass estimates, and relative contribution of coccolithCaCO3 to total carbonate in the northwestern Pacific Ocean at each sediment trap site.

Site Location Total carbonate Calculated coccolithCaCO3

% coccolith-CaCO3

of total CaCO3

(mg m–2day–1) (mg m–2day–1)

NT8 46°07.2′ N, 175°01.9′ E 38.5 11.2 29.1

NT7 37°24.2′ N, 174°56.7′ E 41.5 13.7 33.1

NT5 34°25.3′ N, 177°44.2′ E 23.3 9.0 38.5

NT6 30°00.1′ N, 174°59.7′ E 30.2 11.2 37.1

NT3 0°00.2′ N, 175°09.7′ E 30.1 3.5 11.6

NT2 4°07.5′ N, 136°16.6′ E 41.4 3.9 9.5

NT1 2°59.8′ N, 135°01.5′ E 71.5 16.7 23.3

142 Y. TANAKA

b). Another possible source of error is the systematic undercalculation of thecoccolithophore or calcareous dinophyte CaCO3 flux. However, the greatadvantage in estimating the carbonate flux of coccolithophores allows us tocompare different regions or different depths, because individual estimates ofmost species, including the dominant species, can be calculated.

The contribution of the coccolith carbonate fluxes to the total carbonate fluxwas minor, with a mean of 23.3% at Site NT1, 9.5% at Site NT2, and 11.6% at SiteNT3 (Table 1). The contribution of coccolith carbonate at Site NT3 was slightlyhigher than that at Site NT2. The correlation between calculated coccolith CaCO3and the carbonate flux of the <63 µm fraction (Kawahata et al., 2002) was veryhigh at Sites NT1 (r = 0.90) and NT3 (r = 0.83). This result indicates that thevalues calculated for the coccolith CaCO3 flux were a large contribution by the<63 µm fraction. The carbonate fluxes, except for the coccolith fluxes, showedthe highest value in August 1991, whereas the next highest one appeared fromDecember 1991 through January 1992 at Site NT1. This tendency is similar to thechanges observed in the planktonic foraminifera flux (Kawahata, 2002) at samesediment trap. The calculated coccolith CaCO3 flux reached its highest value inDecember 1991, the next highest value being in September 1991 at Site NT1.Therefore, a difference in the seasonal pattern was identified between thecoccolithophore and planktonic foraminifera carbonate fluxes at Site NT1.McPhaden (1998) represented the oceanographic environmental change in thisregion, due to a weak ENSO in 1991–1992. That is, in the summer, the surfacecirculation around Site NT1 was dominated by the SEC and by the ECC. Thesetwo current systems contain high nutrient transported primarily by the southeastmonsoon off the coast of New Guinea. However, this region is influenced by thenortheast monsoon in winter. Thus, nutrients are provided to this area by inflowfrom the marginal seas and from the ocean. It is presumed that the differencebetween the direction of monsoon in summer and winter influences the observedchanges in biological production. At Site NT3, the coccolithophore flux washighest in the latter half of September 1992, and then decreased gradually,whereas the foraminifera flux was highest in the latter half of October 1992, sothe peak of the coccolithophore flux occurred one month ahead of that of theforaminifera flux.

At the three Sites (NT1, NT2 and NT3), F. profunda comprised 50% of thecoccolithophore community throughout the year, contributing 20% or less of theentire calculated coccolith CaCO3 flux (Fig. 5). Significant contributors to thetotal coccolith CaCO3 were O. fragilis, G. oceanica, and C. leptoporus at SiteNT1, and O. fragilis and C. leptoporus at Site NT3. Oolithotus fragilis isdominant at 75 m depth in the equatorial region (Okada and Honjo, 1973) andmoderately phosphate-rich waters in the Pacific Ocean (Roth and Coulbourn,1982). Gephyrocapsa oceanica and U. sibogae are most abundant in the highnutrient condition in the South Pacific (Roth and Berger, 1975). Gephyrocapsaoceanica dominates especially in the marginal seas of the Pacific Ocean (Okadaand Honjo, 1975), tropical to subtropical regions (Broerse et al., 2000b), and thePanama Basin (Roth and Coulbourn, 1982). Moreover, the presence of this

Coccolith Carbonate Fluxes in the Northwest Pacific Ocean 143

species was reported in both low-latitude upwelling regions (Winter, 1982, 1985;Mitchell-Innes and Winter, 1987; Hagino et al., 2000) and in warm, nutrient-richwaters (Houghton and Gupta, 1991). Sites NT1 and NT3 were located within theinfluence of the SEC, and the coastal current in particular strongly influences thewaters overlying Site NT1 (Fig. 1). Therefore, O. fragilis was dominant at bothsites because of the high nutrient supply from the surface waters. The relativecontributions of G. flabellatus, F. profunda, and G. oceanica were high at SiteNT2 (Fig. 5). Site NT2 is located in the ECC (Fig. 1). In this region, the supplyof nutrients to the sea surface is inferred to be low from evidence suggesting thatthe organic flux at same station was low throughout study period (Kawahata,2002). Therefore, the carbonate contribution of F. profunda and G. flabellatus,generally abundant within the lower photic zone, is high compared with the othertwo sites.

Broerse (2000b) suggested that the significant contributors to total coccolithCaCO3 at low latitudes are G. oceanica and G. flabellatus. In the easternequatorial Pacific, G. flabellatus comprises about 50% of the abundance of thecoccolithophore community, G. oceanica about 20%, and F. profunda 10%, atstations 5°S, 140°W and 12°S, 140°W (Broerse, 2000a); the frequency of O.fragilis was not reported. However, the assemblages at Site NT3 and the other twosites are obviously different.

4.5 Comparison of coccolith carbonate fluxes with other sea areas

The mid- and high-latitudes of the North Pacific Ocean are characterized bymany masses of subarctic, transition and central waters, and the KuroshioExtension Current; they constitute an interesting environment in terms ofoceanographic biological pump processes. The subarctic region is a distinctivereducer of CO2 partial pressure in the surface water during the spring due to asignificantly high primary production in the same region (Kawahata et al., 1998).The sediment-trap materials from these regions have previously been used tostudy the response of planktonic foraminifera to seasonal and regional changesin upper-ocean conditions (Eguchi et al., 1999). It has been suggested thatavailability of food is the most important factor in the foraminiferal production,and the high total foraminifera fluxes and high organic fluxes are observedbetween summer and autumn in the subarctic water mass (Eguchi et al., 1999). Onthe other hand, low total foraminifera flux is due to lower productivity underoligotrophic conditions in the subtropical water mass. Kawahata (2002) representedhis sediment trap results in the North Pacific for the evaluation of the regionalvariation of the biological production. In the central north Pacific, comparisonshowed that the opal fluxes of the high-latitude region were about three timesthose of the low-latitude region, although the total carbonate flux of the high-latitude region was only slightly higher than that of the low-latitude region(Kawahata, 2002).

In the central north Pacific, the calculated coccolith CaCO3 flux was 11.2 mgm–2day–1 (29.1% of the total carbonate flux) at Site NT8, 13.7 mg

144 Y. TANAKA

m–2day–1 (33.1%) at Site NT7, 9.0 mg m–2day–1 (38.5%) at Site NT5, and 11.2 mgm–2day–1 (37.1%) at Site NT6 (Table 1). The calculated coccolith CaCO3 flux washighest in the transition zone in the north central Pacific Ocean. The averagecalculated coccolith carbonate fluxes in the northeast Atlantic Ocean were 7.4 mgm–2day–1 at 48°N and 10.4 mg m–2day–1 at 34°N (47% of the total carbonate flux)(Broerse et al., 2000b), and 3.9 mg m–2day–1 at 29°N, 15°W (33.4% of the totalcarbonate flux) (Sprengel et al., 2000). A comparison between the north Pacific(Site NT5) and the northeast Atlantic at 34°N (Broerse et al., 2000b) indicatedthat the estimated coccolith CaCO3 fluxes and their contribution to the totalcarbonate fluxes were higher in the latter region.

In the Atlantic Ocean, the relative contribution of coccolith CaCO3 to totalcarbonate was high in the subtropical region and low in the temperate region(Broerse, 2000b). In the central equatorial Pacific, the calculated coccolithcarbonate fluxes reached values of 3.6 mg m–2day–1 (8.8%) at 5°S, 140°W and 2.3mg m–2day–1 (11.4%) at 12°S, 140°W (Broerse, 2000a). Therefore, in thenorthwest and central Pacific Ocean, the relative contribution of coccolith CaCO3to total carbonate was highest in the subarctic region, and low in the equatorialregion, excluding the West Caroline Basin, to which nutrients are abundantlysupplied. Thus, it has been shown semi-quantitatively that the role of thecoccolithophore carbonate flux differs considerably geographically.

5. CONCLUSIONS

The following conclusions were drawn from estimations of coccolithophorecarbonate made in the western equatorial Pacific Ocean.

(1) In the western equatorial Pacific Ocean, coccolithophores are minorcontributors to the CaCO3 flux. However, the levels of coccolith CaCO3 calculatedin the West Caroline Basin were comparatively high at low latitudes due to highnutrient supply.

(2) Oolithotus fragilis, Gephyrocapsa oceanica, and Calcidiscus leptoporuscontribute to the coccolith CaCO3 flux, although Florisphaera profundapredominates in the coccolithophore assemblages in the western equatorialPacific Ocean. Oolithotus fragilis appears to be an important species in theseasonal pattern of coccolithophore carbonate variation in the western equatorialPacific Ocean, because there is a strong correlation between O. fragilis andnutrient supply. Gephyrocapsa oceanica is similarly important in the westequatorial Pacific Ocean.

(3) The contribution of coccolithophores to the calcium-carbonate flux washighest in the transition zone and lowest in the equatorial Pacific Ocean.

Acknowledgements—The author is grateful to three anonymous reviewers whose commentshelped in substantially improving this manuscript. He thanks Dr. Hodaka Kawahata forhelpful advice. He also wishes to thank the scientists on the R/V Hakureimaru for theircooperation and valuable discussions, and the captain and crew for their generousassistance. This work was supported by the following research programs: 1) the NorthwestPacific Carbon Cycle Study, which was consigned to the Kansai Environmental EngineeringCenter Co., Ltd., by the New Energy and Industrial Technology Development Organization,

Coccolith Carbonate Fluxes in the Northwest Pacific Ocean 145

and 2) GCMAPS (Global Carbon Cycle and the Related Global Mapping Based onSatellite Imagery) program; the latter was promoted by MEXT of Japan.

REFERENCES

Broerse, A. T. C. (ed.) (2000a): Coccolithophore fluxes at Sites 5°S 139°W and 12°S 135°W in thecentral tropical Pacific: response to El Niño conditions. In: Broerse, A. T. C. ed., CoccolithophoreExport Production in Selected Ocean Environments: Seasonality, Biogeography and CarbonateProduction, 121–142.

Broerse, A. T. C. (2000b): Coccolithophore export production in selected ocean environments:seasonality, biogeography and carbonate production. Vrijr Univ. Academisch Proefschrift, 184pp.

Broerse, A. T. C., G.-H. A. Brummer and J. E. van Hinte (2000a): Coccolithophore export productionin response to monsoonal upwelling off Somalia (northwestern Indian Ocean). Deep-Sea Res.II, 47, 2179–2205.

Broerse, A. T. C., P. Ziveri, J. E. van Hinte and S. Honjo (2000b): Export production, speciescomposition, and coccolith-CaCO3 fluxes in the NE Atlantic (34N 21W and 48N 21W). Deep-Sea Res. II, 47, 1877–1905.

Deuser, W. G. and E. H. Ross (1989): Seasonally abundant planktonic foraminifera of the SargassoSea: succession, deep-water fluxes, isotopic compositions, and paleoceanographic implications.J. Foram. Res., 19, 268–293.

Eguchi, N., H. Kawahata and A. Taira (1999): Seasonal response of planktonic foraminifera tosurface ocean condition: sediment trap results from the central North Pacific Ocean. J.Oceanogr., 55, 681–691.

Fabry, V. J. (1989): Aragonite production by pteropod mollusks in the sub arctic Pacific. Deep-SeaRes., 36, 1735–1751.

Hagino, K., H. Okada and H. Matsuoka (2000): Spatial dynamics of coccolithophore assemblagesin the Equatorial Western–Central Pacific Ocean. Mar. Micropaleont., 39, 53–72.

Honjo, S. (1977): Biogenic carbonate particle in the ocean; do they dissolve in the water column?In: Anderson, N. R. and Malahoff, A., eds., The Fate of Fossil Fuel CO2 in the Oceans, PlenumPress, New York, 269–295.

Houghton, S. D. and M. V. Gupta (1991): Monsoonal and fertility controls on recent marginal seaand continental shelf coccolith assemblages from the western Pacific and northern Indianoceans. Mar. Geol., 97, 251–259.

Kawahata, H. (2002): Suspended and settling particles in the Pacific. Deep-Sea Res. II, 49, 5647–5664.

Kawahata, H., A. Suzuki and H. Ohta (1998): Sinking particles between the equatorial and subarcticregions (0N–46N) in the central Pacific. Geochem. J., 32, 125–133.

Kawahata, H., A. Suzuki and H. Ohta (2000): Export fluxes in the Western Pacific Warm Pool. Deep-Sea Res. II, 47, 2061–2091.

Kawahata, H., A. Nishimura and M. K. Gagan (2002): Seasonal change in foraminiferal productionin the western equatorial Pacific warm pool: evidence from sediment trap experiments. Deep-Sea Res. II, 49, 2783–2800.

Levitus, S. and T. Boyer (1994): World Ocean Atlas 1994 Volume 4: Temperature. NOAA AtlasNESDIS 4, U.S. Department of Commerce, Washington, D.C.

McPhaden, M. J. (1998): TOGA-TAO and the 1991–93 El Niño-Southern Oscillation Event. NOAAPublication, U.S. Department of Commerce, Washington, D.C.

Mitchell-Innes, B. A. and A. Winter (1987): Coccolithophores: a major phytoplankton componentin mature upwelled waters off the Cape Peninsula, South Africa in March, 1983. Mar. Biol., 95,25–30.

Miyama, T., T. Awaji, K. Akitomo and N. Imasato (1996): A lagrangian approach to the seasonalvariation of salinity in the mixed layer of the Indonesian Seas. J. Geophys. Res., 101, 12265–12285.

146 Y. TANAKA

Okada, H. and S. Honjo (1973): The distribution of oceanic coccolithoporids in Pacific. Deep-SeaRes., 20, 355–374.

Okada, H. and S. Honjo (1975): Distribution of coccolithophores in marginal seas along the westernPacific Ocean and in the Red Sea. Mar. Biol., 31, 271–285.

Roth, P. H. and W. H. Berger (1975): Distribution and dissolution of coccoliths in the south andcentral Pacific. In: Sliter, W. V., Bé, A. W. H. and Berger, W. H., eds., Cushuman Foundationfor Foraminiferal Research, 87–113.

Roth, P. H. and W. T. Coulbourn (1982): Floral and solution patterns of coccoliths in surfacesediments of the north Pacific. Mar. Micropaleont., 7, 1–52.

Sprengel, C., K.-H. Baumann and S. Neuer (2000): Seasonal and interannual variation ofcoccolithophore fluxes and species composition in sediment traps north of Gran Canaria (29°N15°W). Mar. Micropaleont., 39, 157–178.

Takahashi, K. (1994): Coccolithophorid biocoenosis: production and fluxes to the deep sea. In:Green, J. C. and Leadbeater, B. S. C., eds., The Haptophyte Algae, Systematic AssociationSpecial 51, Clarendon Press, Oxford, 335–350.

Tanaka, Y. and H. Kawahata (2001): Seasonal occurrence of coccoliths in sediment traps from WestCaroline Basin, equatorial West Pacific Ocean. Mar. Micropaleont., 43, 273–284.

Tchernia, P. (1980) Descriptive Regions Oceanography. Pergamon Marine Series, 3, PergamonPress, Oxford, 253 pp.

Westbroek, P., J. E. van Hinte, G.-J. Brummer, M. Veldhuis, C. Brownlee, J. C. Green, R. Harris andB. R. Heimdal (1994): Emiliania huxleyi as a key to biosphere-geosphere interaction. In: Green,J. C. and Leadbeater, B. S. C., eds., The Haptophyte Algae, Systematic Association Special 51,Clarendon Press, Oxford, 321–334

Winter, A. (1982): Paleoenvironmental interpretation of Quaternary coccolith assemblages from theGulf of Aqaba (Elat), Red Sea. Re. Esp. Micropaleontol., 14, 291–314.

Winter, A. (1985): Distribution of living coccolithophores in the California Current system,Southern California borderland. Mar. Micropaleont., 9, 385–393.

Yan, X.-H., C.-R. Ho, Q. Zheng and V. Klemas (1992): Temperature and size variabilities of thewestern Pacific warm pool. Science, 258, 1643–1645

Young, J. R. and P. Ziveri (2000): Calculation of coccolith volume and its use in calibration ofcarbonate flux estimates. Deep-Sea Res. II, 47, 1679–1700.

Ziveri, P., R. C. Thunell and D. Rio (1995): Export production of coccolithophores in an upwellingregion: Results from San Pedro Basin, Southern California Borderlands. Mar. Micropaleont.,24, 335–358.

Ziveri, P., A. T. C. Broerse, J. E. van Hinte, P. Westbroek and S. Honjo (2000): The fate of coccolithsat 48°N 21°W, northeastern Atlantic. Deep-Sea Res. II, 47, 1853–1875.

Y. Tanaka (e-mail: [email protected])