Patterns of coccolithophore pigment change under global acidification conditions based on in-situobservations at BATS site between July 1990–Dec 2008

Abstract

Coccolith production is an important part of the biogenic carbon cycle as the largest source of calcium carbonate on earth, accounting for about 75% of the deposition of carbon on the sea floor. Recent studies based on laboratory experiment results indicated that increasing anthropogenic CO₂ in the atmosphere triggered global ocean acidification leading to a decrease of calcite or aragonite saturation and calcium carbonate, and to decreasing efficiency of carbon export/pumping to deep layers. In the present study, we analyzed about 20 years of field observations of coccolithophore pigment, dissolved inorganic carbon (DIC), nutrients, and temperatures from the Bermuda Atlantic Time-series Study (BATS) site and satellite remote sensing to investigate the variable tendency of the coccolithophore pigment, and to evaluate the influence of ocean acidification on coccolithophore biomass. The results indicated that there was a generally increasing tendency of coccolithophore pigment, coupled with increasing bicarbonate concentrations or decreasing carbonate ion concentration. The change of coccolithophore pigment was also closely associated with pH, nutrients, mixed layer depth (MLD), and temperature. Correlation analyses between coccolithophores and abiotic parameter imply that coccoliths production or coccolithophore pigment has increased with increasing acidification in the recent 20 years.

Keywords

dissolved inorganic carbon / BATS / MLD / coccolithophore pigments / the Bermuda

Introduction

Marine biogeochemical processes are largely controlled by phytoplankton because phytoplankton forms the basis of the food chain and affects sea surface CO₂ through photosynthesis. The biological carbon pump is one of the most important processes reducing CO₂ from the atmosphere through uptake of CO₂ and deposition of organic carbon to the deep ocean. It is very important to investigate the influence of increasing CO₂ on phytoplankton production and deposition of particle inorganic carbon (PIC) and particle organic carbon (POC) in the context of ocean acidification. Coccolithophores (or coccolithophorids) are single-cell algae, protists, and phytoplankton. They are distinguished by special calcium carbonate plates (or scales) called coccoliths, which are main sources forming microfossils. Coccolithophores are widely distributed in oceans and are observed in large numbers throughout the surface euphotic zone of the oceans (Heimdal, 1983). Thus coccolithophore plays an important role in the biogenic carbon cycle (Westbroek et al., 1985), with its production accounting for about 75% of the deposition of carbon on the sea floor having a marine origin, and as the largest source of calcium carbonate on earth (Honjo, 1986, 1990; Groom and Holligan, 1987). From the mid-Mesozoic, coccolithophores have been major calcium carbonate producers in the world’s oceans, today accounting for about two-thirds of the total marine CaCO₃ production.

With increasing anthropogenic CO₂ in the world, ocean acidification, and global warming in response to rising atmospheric CO₂, partial pressures have attracted considerable attention from people, governments, scientists, and international organizations. It is widely expected that ocean acidification will reduce calcification or production of phytoplankton. Recent evidence suggests that the increased absorption of CO₂ by the oceans, as a result of anthropogenic CO₂ release, will result in decreased calcification by corals, foraminifera, and coccolithophores (Bijma et al., 1999; Kleypas et al., 1999; Riebesell et al., 2000; Zondervan et al., 2001; Delille et al., 2005; Zeebe et al., 2008). The above viewpoint is rather popular in the world. However, Iglesias-Rodriguez et al. (2008) proved also that calcification and net primary production in the coccolithophore species Emiliania huxleyi (E. huxleyi) were significantly increased by high CO₂ partial pressures, based on their special laboratory evidence and field evidence from the deep oceans. Moreover, recent studies have already shown that different coccolithophore species exhibit different calcification responses: under increasedpCO₂, a decrease in calcification for E. huxleyi and Gephyrocapsa oceanica (Bijma et al., 1999; Kleypas et al., 1999; Zondervan et al., 2001; Delille et al., 2005), a negligible calcification change with rising pCO₂ for Coccolithus pelagicus; and an increase followed by a decrease in calcification with rising pCO₂ for Calcidiscus leptoporus (Trimborn et al., 2007).

Up to now, scientists have not reached an agreement on responses of Coccolithophere to increasing oceanic acidification. Further, previous studies (Marshall, 1968; Haidar and Thierstein, 2001; Sprengel et al., 2002) are mainly based on laboratory experiments and sparse field surveys with short cultivation periods or relatively short-termin-situ observation (Hulburt et al., 1960; Hulburt, 1990; Haidar and Thierstein, 2001) in relatively small regions, respectively. Satellite measurements have been also used in the study of coccolith variations (Strong and Eadie, 1978;Brown, 2000; Cokacar et al., 2004; Oguz and Ediger, 2006), and some interesting results have been presented in recent years. However, few studies were focused on longer-term change in coccolithophore/calcite of the northwest Atlantic Ocean near BATS in the sea surface water, and the possible mechanism of its responses to pH, dissolved inorganic carbon (DIC), nutrients, mixed layer depth (MLD), and sea surface temperature (SST).

Some pigments are exclusive for specific phytoplankton groups and can be used as taxonomic indicators (Jeffrey et al., 1997; Silva et al., 2013). For example, 19^’-Hexanoyloxyfucoxanthin (Hex-fuco) (Silva et al., 2013) was used in this study as an indicator of coccolithophores with good positive correlation between Hex-fuco and coccolithophores cell counts (r=0.61, p<0.05) in the middle North Atlantic. In the present study, we plan to investigate seasonal/annual patterns of coccolithophore pigment (CP) using Hex-fuco as an indicator associated with ocean acidification, nutrients, winds, and SST in the Sargasso Sea of the northwest Atlantic Ocean, through analyzing pH, nutrients and MLDs based on relatively longer-term time-series observation data from BATS. Interesting results such as the seasonal and inter-annual variation of CP will be presented, and its responses to ocean acidification will also be discussed.

Study area, data and methods

Study area

Our study area, the Bermuda Atlantic Time-series Study station (BATS) (B in Fig. 1) located 82 km southeast of the island of Bermuda (31°40′N, 64°10′W) in the central northwestern Atlantic Ocean (Fig. 1(a)), belongs to one typical subtropical oceanic climate zone with mild weather all year round. Bottom depth at the BATS deployment area is ~4,680 m. The water column in the study area is stratified in summer, but well mixed in the winter (e.g.,Joyce and Robbins, 1996) when mixing depths normally reach 100–200 m (Michaels, 1995). There are large seasonal temperature variations in the study area, generally with highest SST in August (~27.6 °C) and lowest SST in March (~19.1°C). The BATS core cruises consist of a single 4–5 day cruise monthly. On each core cruise, conductivity-temperature-depth profiles and water collections are made at 36–48 depths during a series of 4–7 casts from the surface to 4,200 m. Since June 1994, continuous surface seawater measurements of temperature, salinity, fluorescence, andpCO₂ have been made from a flow-through seawater system (Bates et al., 1998). Between January and April, an extra “bloom cruise” is added to increase sampling frequency to semimonthly during the spring-bloom period. A lot ofin-situ hydrological-biological measurements have been carried out at BATS. In view of our study area being a typical subtropical region, our study possesses important significance in the investigation of the global carbon sink.

Data and observations

In-situ vertical data of pH, nutrients, DIC, total alkalinity, Hex-fuco (associated with coccolithophore), sediment fluxes, and algae blooms are derived from the BATS. In the study, we use Hex-fuco as a proxy of coccolithophore concentrations. The observation frequencies of data were>2 times per month. Measurements and sampling means have been described in previous works (Gardner, 1977; Knap et al., 1993, 1995; Dore et al., 1996; Bates et al., 1998; Hansell and Carlson, 2001); please refer to http://bats.bios.edu/ for detail.

SeaWiFS-derived Photosynthetic Available Radiation (PAR) data with 9 km resolution are obtained from the Distributed Archive Center of NASA (ftp://oceans.gsc.nasa.gov), which are available from 15 September 1997 to 10 March 2010. Here, monthly PAR data are used to display the depths of monthly CP peaks. ModisA-derived sea surface temperature data are also used with a high resolution (4 km) (ftp://modis.gsfc.nasa.gov/).

Methods

Time series data and power spectral Analysis

Due to missing values or sparse vertical resolution, data are interpolated into monthly time periods and 10 m vertical intervals. Time series of coccolithophore pigments and temperature were averaged for the depths from 10–150 m. Because of many values with default data or zero between 0–60 m, the time series of nutrients are averaged for 70–150 m. The values for pH, carbonate, and bicarbonate are derived from salinity, DIC, and alkalinity and temperature. We use only surface data of pH, carbonate, and bicarbonate to produce time series. MLD are determined by searching down the water column from 10 m below the sea surface until the sea water density has increased by 0.03 kg·m^–3. At the same time, we use spectrum average power analysis to display seasonal and annual patterns of CP.

Seasonal pattern

Time series data were further processed into monthly climatologies of CP and abiotic factors based on averages of the data for 1990–2008. The seasonality of CP and abiotic factors were displayed here. The relations between CP and abiotic factors were discussed through analysis of the climatologies.

Time series profiles

Monthly vertical profiles during 1990–2008 were also presented based on CP, CTD temperature for the surface to 150 m, and nutrients for 60–150 m because most nutrients data such as nitrate and phosphate are missing or low in the upper 50 m layers. Due to relatively low concentrations of phosphate and numerous missing values, we only used one nitrate-nitrite profile as an index of nutrients.

Correlation analysis

In order to remove seasonal and periodic signals, and probe into the possible mechanism of CP responses to abiotic factors, we first calculated the differences of all above parameters from monthly climatologies, and then the correlation between CP and other parameters was calculated and analyzed respectively. In the analyses, we limited nutrients data for the upper 60 m layers, because dominant photosynthesis appeared generally in the upper 60 m layer. Finally, based on particulate inorganic carbon (PIC) data derived from MODISA, we also investigated the correlation between the monthly time series of satellite PIC averaged for the small region of 64°W–63°W, 31°N–32°N near BATS andin-situ Hex-fuco.

Results

Monthly coccolithophore pigment

The time series (Fig. 2) averaged for the upper 150 m during 1990–2008 indicated evidently the seasonality of CP with peak concentrations in spring (March–April) and lowest in autumn (September–November). An overall trend of increase was observed from 1990–2008. The concentrations increased from 1990–1995, decreased from 1995–2001, and increased again from 2001 to present. Highest values occurred in the spring of 1990 (0.105  µg·kg^–1), 1995 (0.15 µg·kg^–1), and 2006 (0.11 µg·kg^–1), respectively, with lowest concentrations in 2001–2003. Analysis of the power spectrum (Fig. 2(b)) indicated dominant CP variations at one-year and ten-year cycles. In the following subsection, the abiotic time series data (Fig. 3) will be presented to investigate the trend in CP.

The monthly climatologies (Fig. 4) indicated the characteristics of high (low) CP concentration coincided with high (low) bicarbonate, low (high) pH, high (low) DIC, low (high) calcite saturation (Ω), low (high) temperature, and high (low) nitrate. Unlike the relation between CP and nitrate, that between CP and phosphate is not good. In addition, the seasonal peak of phosphate lagged one month compared with nitrate and MLDs.

Dissolved inorganic carbon, nutrients, and physical conditions

DIC concentration (Fig. 3(a)) shows a trend of increase (Fig. 3(a), left) with a strong seasonal and inter-annual variation (Fig. 3(a), right). There is positive coincidence between DIC and CP: higher DIC concentrations associated with higher CP in late winter–spring, and lower DIC with lower CP in boreal autumn, especially in October. DIC also shows a strong 1-year cycle (Fig. 3(a), right), closely coinciding with that of CP.

Seasonal and annual cycles of MLDs, temperature, and salinity (Figs. 3(b)–3(d)) were evident, with highest temperature, highest salinity, and shallowest MLDs in late summer–early autumn, and the lowest temperature and deepest MLDs in late winter–early spring. Generally, there were higher nutrients with deeper MLDs.

Nutrients (i.e., P, N) revealed approximately an annual cycle and seasonal change (Fig. 3(e)). Especially, the N concentration coincided closely with that of CP with roughly 1 month lag of CP. The ratio of N:P was generally over 16:1. The concentration of P is relatively low, approaching zero in most of the measurement period. However, results of the power spectrum showed that the cycle of 1 year was much stronger than the cycles over 1 year.

CP, nitrate, and temperature

CP, nitrate, and temperature (Fig. 5) also showed strong vertical difference/variation in the annual cycle from 1990–2008. Monthly peak values (Figs. 4(a), 4(b) and 5) of CP did not appear generally at the surface, but at 40– 100 m. Peak values in spring (Figs. 4 and 5) are closer to surface layers. Nitrate is generally very low except in late winter–early spring. Temperature is higher near the surface, except that during the winter–spring period there was uniformly low temperature in the upper layer.

Analysis of correlation

Through analyses of correlation, CP has a positive relationship (Fig. 6) with bicarbonate (r= 0.32, p<0.005), DIC (r= 0.28, p<0.05) and nitrate (r=0.24, p<0.05). In contrast, it has a negative relationship with carbonate (r=0.24, p<0.05), pH and Ω (r=0.24, p<0.05). The bicarbonate has a more significant correlation with CP than other abiotic factors.

Discussion

Seasonality and annual tendency in CP

As shown in Figs. 2 and 3, significant seasonal signals were presented in CP, temperature, light, nutrients (here, P and N as the indicators), salinity, pH, and\mathrm{HC}{\mathrm{O}}_{\mathrm{3}}^{-} and \mathrm{C}{\mathrm{O}}_{\mathrm{3}}^{\mathrm{2}-}. The CP pattern was regulated possibly by seasonal cycles of DIC, nutrients, light, and temperature. Though there was roughly a similar seasonal tendency between CP and nutrients (Figs. 2–4), it was observed that monthly mean nutrients for the upper layer were generally high in winter, in contrast with low CP. The results may be attributed to low sea water temperature and weak light irradiance (Fig. 7) and strong mixing (Figs. 4 and 5), which were probably unfavorable to (restraint) growth of phytoplankton (Haidar and Thierstein, 2001). High nutrients, increasing light irradiance, and increasing stable stratification induced by elevated SST from winters to springs triggered frequent occurrences of seasonal CP peaks in later winters or early springs, implying bio-physical influence of abiotic factors on coccolithophore. The present result is consistent with the results of Haidar and Thierstein (2001). Nutrients (Figs. 4(a) and 4(i)) for the upper 60 m are possibly a limiting factor (e.g., [N]<0.03 μmol·kg^–1) from late spring to early winter versus (>0.1 μmol·kg^–1) in winter–spring, leading to monthly CP peak value occurring in the deeper layer outside of the winter–spring transit season. Due to generally low nutrients (Fig. 5) and strong light PAR (Fig. 7) at the surface layer, the monthly peak lies at the deeper layer compared with that in late winter–early spring. Thus, the change in depth of the monthly CP peaks may be attributed to a compromise of nutrients, light irradiance, and sea water temperature.

Considering the fact that there are no strong mesoscale eddies controlling regional dynamics around BATS (Steinberg et al., 2001), we can ignore the influence of upwelling or downwelling triggered by mesoscale eddies. Thus, MLDs triggered by perturbation can be generally treated as a good index of the upper-layer mixing and the uptake of nutrients (Figs. 3(c), 3(e), 4(h), and 4(i);Conte et al., 2001; Haidar and Thierstein, 2001; Steinberg et al., 2001), which was supported by the good correlation (r=0.31, p<0.001) between MLDs and CP (one month lag of CP). Better coincidence between CP and MLDs may reflect the uptake and entrainment of nutrients from below the eutrophic layer. However, due to consumption of phytoplankton photosynthesis, present nutrients in stock cannot represent synchronous phytoplankton biomass, leading to relatively weak correlation (r=0.24, p<0.001) between CP and nutrients (nitrate and nitrite) compared with that between CP and MLDs.

Influence of DIC, light, temperature, and nutrients on CP

Recent evidence has shown that increased absorption of CO₂ by the oceans as a result of anthropogenic CO₂ release will result in decreased calcification by coccolithophores (Riebesell, et al., 2000; Zondervan et al., 2001; Delille et al., 2005). However, Iglesias-Rodriguez et al. (2008) suggested that some coccolithophore species increased significantly due to high CO₂ partial pressures or ocean acidification. Our results showed that a decreasing pH value could trigger a high biomass of coccolithophore consistent with the conclusion of Iglesias-Rodriguez et al. (2008). The previous studies about change of coccolithophore species with CO₂ were mainly laboratory based experiments which had obvious differences with real oceanic conditions. In particular, coccolithophore species experienced more periods to adapt to ocean acidification or pH changes in real oceans, compared with short culture times and sharp ocean acidification conditions. Furthermore, experimental designs did not consider the adaptation to changing carbonate chemistry. In contrast, the experimental methods of Iglesias-Rodriguez et al. (2008) were close to real ocean structures of dissolved inorganic carbon reported in recent studies. Thus gradual ocean acidification will possibly trigger increasing coccolithophore production according to our present data.

Langer et al. (2006) proved that coccolith morpholoyies did not present any evidence of significantly incomplete or malformed coccolith in the last glacial maximum and Holocene sediments according to analysis of sediment records samples before the lysocline in the Atlantic Ocean, suggesting adaptation of coccolithophore calcification to changing CO₂ concentrations in natural upper oceans. In other words, some species of coccolithophores have adapted their calcification mechanism to changes in carbonate chemistry which occurred since the last glacial maximum. The significant discrepancy of coccolith morphology between experimental and geological results (Langer et al., 2006) also indicated the experimental flaws can’t reflect the adaptation of coccolithophore and possible products in real oceans. This indicated the importance ofin-situ observations in the oceans. There were generally higher DIC concentrations during high CP periods. However, it was inconsistent in 1995 when there were low DIC and high CP. This is the reason that Haidar and Thierstein (2001) found there was a negative relation between DIC and CP at the same BATS using 3-years in-situ observations, opposite to our findings.

In highly calcifying cells of E. huxleyi\mathrm{H}\mathrm{C}{\mathrm{O}}_{\mathrm{3}}^{-} is the substrate for calcification (Nimer and Merrett, 1992) that can be schematically represented as follows:

\mathrm{2}\mathrm{H}\mathrm{C}{\mathrm{O}}_3^{-}+\mathrm{C}{\mathrm{a}}^{2+}\leftrightarrow {\mathrm{H}}_{\mathrm{2}}\mathrm{O}+\mathrm{C}\mathrm{a}\mathrm{C}{\mathrm{O}}_{\mathrm{3}}+\mathrm{C}{\mathrm{O}}_{\mathrm{2}}.

This reaction results in intracellular production of CO₂, which is closely coupled (i.e., ratio of C:P ≈ 1.0) to photosynthetic assimilation within the cell (Crawford and Purdie, 1997). The increase of [\mathrm{H}\mathrm{C}{\mathrm{O}}_{\mathrm{3}}^{-}] with ocean acidification induces possibly the above chemical equilibrium moving the right side, producing more calcium carbonate, consistent with our preliminary results (Fig. 6(b)). The speculation was also supported by the experimental results of Langer et al. (2006). Studies (Riebesell et al., 2000; Zondervan et al., 2001; Delille et al., 2005) from experimental results which demonstrated that different species of coccolithophores or species in different regions have different responses to changes in ocean acidification. Thus, our study may only represent the characteristics of coccolithophore pigment change near BATS, not globally.

In order to further investigate the influence of oceanic acidification, we compared the pigments concentration in the same period. The results also showed that higher [\mathrm{H}\mathrm{C}{\mathrm{O}}_{\mathrm{3}}^{-}] corresponded with higher coccolithophore pigments. Furthermore, the fact that seasonal changes of DIC (Fig. 3(a)) were stronger than the inter-annual signal can also suggest that phytoplankton may probably adapt to the environment of ocean acidification and changes of anthropogenic CO₂ at the present rate due to relative slow inter-annual changes compared with seasonal changes in this region.

19'-Hexanoyloxyfucoxanthin exists widely in pelagophyceae and prymnesiophyceae including coccolithophore, which may lead to wrong results if 19'-Hexanoyloxyfucoxanthin is used as an indicator of coccolithophore. However, phytoplankton from pelagophyceae (Lochhead et al., 2001; Silva et al., 2013) is not dominant in our study area; the influence of the phytoplankton pigment out of coccolithophore can be neglected in our study. Moreover, the good correlation (r=0.65, p<0.05) presented between the MODIS-derived particulate inorganic carbon data and Hexfuco (Fig. 8) suggested that Hexfuco can be used as a good indicator of coccolithophore in our region. In addition, we calculated the CaCO₃ flux using 200-m fluxes of mass, particulate organic carbon (POC), and particulate organic nitrogen (PON) according to the Redfield ratio of C:N:P (106:16:1) and the ratio of C:O:H (12:16:2), given negligible contribution of other elements than the above to the total mass fluxes. This is reasonable because diatom algae are not a dominant species in the region. The results calculated showed a similar tendency (Fig. 9) (r=0.21, p<0.001) to 19'-Hexanoyloxyfucoxanthin pigment, implying that it as the indicator of coccolithophore is acceptable. Considering the complexity of seawater chemistry and factors influencing oceans, our speculation needs to be investigated further in the future.

Summary

Coccolithophore pigment presented evident seasonal and inter-annual cycles with the highest concentrations in early spring and the lowest concentrations in early autumn. Overall, an increasing trend of coccolithophore pigments was observed over 1990–2008. The variation of coccolithophore pigments was closely associated with DIC and nutrients. [\mathrm{H}\mathrm{C}{\mathrm{O}}_{\mathrm{3}}^{-}] of DIC species showed the most significant correlation with coccolithophore pigments, implying that [\mathrm{H}\mathrm{C}{\mathrm{O}}_{\mathrm{3}}^{-}] regulated the coccolithophore pigments more efficiently and significantly. Due to an increase in [\mathrm{H}\mathrm{C}{\mathrm{O}}_{\mathrm{3}}^{-}] triggered by ocean acidification or increasing anthropogenic CO₂, ocean acidification will probably enhance the calcification or coccolithore production in the study area.

References

[1] Bates N R, Takahashi  T, Chipman D W ,  Knap A H  (1998). Variability of pCO₂ on diel to seasonal timescales in the Saragasso Sea near Bermuda. J Geophys Res, 103(C8): 15567–15585

[2] Bijma J, Spero  H J, Lea  D W, Bemis  B E (1999). Reassessing foraminiferal stable isotope geochemistry: impact of the oceanic carbonate system (experimental results). In: Fischer  G, Wefer G , eds. Use of Proxies in Paleoceanography: Examples from the South Atlantic. Berlin: Springer Science & Business Media, 489–512

[3] Brown C W (2000). Spatial and temporal variability of Emiliania huxleyi blooms in Sea WiFS imagery. Paper presented at American Geophysical Union, Ocean Sciences Meeting, San Antonio, Texas (USA), January, 24–28

[4] Crawford D, Purdie  D (1997). Increase of pCO₂ during blooms of Emiliania Huxleyi: theoretical considerations on the asymmetry between acquisition of  HCO3−  and respiration of free CO₂. Limnol Oceanogr, 42(2): 365–372

[5] Cokacar T, Oguz  T, Kubilay N  (2004). Satellite-detected early summer coccolithophore blooms and their interannual variability in the Black Sea. Deep Sea Res Part I Oceanogr Res Pap, 51(8): 1017–1031

[6] Conte M H, Ralph  N, Ross E H  (2001). Seasonal and interannual variability in deep ocean particle fluxes at the Oceanic Flux Program (OFP)/Bermuda Atlantic Time Series (BATS) site in the western Sargasso Sea near Bermuda. Deep Sea Res Part II Top Stud Oceanogr, 48(8–9): 1471–1505

[7] Delille B, Harlay  J, Zondervan I ,  Jacquet S ,  Chou L, Wollast  R, Bellerby R G J ,  Frankignoulle M ,  Borges A V ,  Riebesell U ,  Gattuso J P  (2005). Response of primary production and calcification to changes of pCO₂ during experimental blooms of the coccolithophorid Emiliania huxleyi. Global Biogeochem Cycles, 19(2): GB2023

[8] Dore J E, Houlihan  T, Hebel D V ,  Tien G, Tupas  L, Karl D M  (1996). Freezing as a method of sample preservation for the analysis of dissolved inorganic nutrients in seawater. Mar Chem, 53(3–4): 173–185

[9] Gardner W D (1977). Incomplete extraction of rapidly settling particles from water samplers. Limnol Oceanogr, 22(4): 764–768

[10] Groom S B, Holligan  P M (1987). Remote sensing of coccoltihophore blooms. Proceedings XXVI COSPAR Meeting, Toulouse, France

[11] Haidar A T, Thierstein  H R (2001). Coccolithophore dynamics of Bermuda (N. Atlantic). Deep Sea Res Part II Top Stud Oceanogr, 48(8–9): 1925–1956

[12] Hansell D A, Carlson  C A (2001). Biogeochemistry of total organic carbon and nitrogen in the Sargasso Sea: control by convective overturn. Deep Sea Res Part II Top Stud Oceanogr, 48(8–9): 1649–1667

[13] Heimdal B R (1983). Phytoplankton and nutrients in the waters north-west of Spitsbergen in the autumn of 1979. J Plankton Res, 5(6), 901–918

[14] Honjo S (1986). Oceanic particles and pelagic sedimentation in the western North Atlantic Ocean. The Geology of North America, 1000: 469–478

[15] Honjo S (1990). Particle fluxes between 47 N and 34 N 20 W stations between April 3 to September 26, 1989. EOS Trans AGU, 71: 81

[16] Hulburt  E  M   (1990). Description  of  phytoplankton  and nutrient  in spring  in  the  western  North  Atlantic  Ocean. J  Plankton  Res, 12(1): 1–28

[17] Hulburt E M, Ryther  J H, Guillard  R (1960). The phytoplankton of the Sargasso Sea off Bermuda. Journal du Conseil, 25(2): 115–128

[18] Iglesias-Rodriguez M D ,  Halloran P R ,  Rickaby R E M ,  Hall I R ,  Colmenero-Hidalgo E ,  Gittins J R ,  Green D R H ,  Tyrrell T ,  Gibbs S J ,  von Dassow E ,  Rehm E, Armbrust  E V, Boessenkool  K P (2008). Phytoplankton calcification in a high-CO₂ world. Science, 320(5874): 336–340

[19] Jeffrey S W, Mantoura  R F C, Bjørnland  T (1997). Data for the identification of 47 key phytoplankton pigments. In: Jeffrey  S W, Mantoura  R F C, Wright  S W, eds. Phytoplankton Pigments in Oceanography: Guidelines to Modern Methods. Unesco Monographs on Oceanographic  Methodology,  vol  10.  UNESCO,  Paris,  449–559

[20] Joyce T M, Robbins  P (1996). The long-term hydrographic record at Bermuda. J Clim, 9(12): 3121–3131

[21] Kleypas J A, Buddemeier  R W, Archer  D, Gattuso J P ,  Langdon C ,  Opdyke B N  (1999). Geochemical consequences of increased atmospheric carbon dioxide on coral reefs. Science, 284(5411): 118–120

[22] Knap A H, Michaels  A F, Dow  R L, Johnson  R J, Gundersen  K, Sorensen J C ,  Close A R ,  Howse F ,  Hammer M ,  Bates N ,  Doyle A ,  Waterhouse T  (1993). BATS Methods Manual, Version 3. U.S. JGOFS Planning O$ce, Woods Hole, MA

[23] Knap A H, Michaels  A F, Steinberg  D, Bahr F ,  Bates N ,  Bell S, Countway  P, Close A ,  Doyle A ,  Howse F ,  Gundersen K ,  Johnson R ,  Little R ,  Orcutt K ,  Parsons R ,  Rathbun C ,  Sanderson M ,  Michaels A F ,  Knap A H  (1995). Overview of the U.S. JGOFS BATS and Hydrostation S program. Deep-Sea Res, 43(2–3): 157–198

[24] Langer G, Geisen  M, Baumann K H ,  Kläs J ,  Riebesell U ,  Thoms S ,  Young J R  (2006). Species-specific responses of calcifying algae to changing seawater carbonate chemistry. Geochem Geophys Geosyst, 7(9): Q09006

[25] Lochhead V C, Lomas  M W, Lethaby  P J (2001). Long-term variability of phytoplankton community structure at the Bermuda Atlantic Time-series Study (BATS) site based on pigment analyses using the “CHEMTAX” matrix (Abstract). Workshop “Pigments as a Tool to Estimate the Biomass of Different Phytoplankton Groups”. Barcelona, 2001. 25

[26] Marshall H (1968). Coccolithophores in the northwest Sargasso Sea. Limnol Oceanogr, 13(2): 370–376

[27] Michaels A F (1995). Ocean time series research near Bermuda: the Hydrostation S time series and the Bermuda Atlantic time series study (BATS). In:  Powell T M ,  Steele J H , eds. Ecological Time Series. New York: Chapman and Hall, 181–208

[28] Nimer N A, Merrett  M J (1992). Calcification and utilization of inorganic carbon by the coccolithophorid Emiliania huxleyi Lohmann. New phytol, 121(2): 173–177

[29] Oguz T, Ediger  D (2006). Comparision of in situ and satellite-derived chlorophyll pigment concentrations, and impact of phytoplankton bloom on the suboxic layer structure in the western Black Sea during May–June 2001. Deep Sea Res Part II Top Stud Oceanogr, 53(17–19): 1923–1933

[30] Riebesell U, Zondervan  I, Rost B ,  Tortell P D ,  Zeebe R E ,  Morel F M  (2000). Reduced calcification of marine plankton in response to increased atmospheric CO₂. Nature, 407: 365–367

[31] Silva A, Brotas  V, Valente A ,  Sá C ,  Diniz T ,  Patarra R F ,  Álvaro N V ,  Neto A I  (2013). Coccolithophore species as indicators of surface oceanographic conditions in the vicinity of Azores islands. Estuar Coast Shelf Sci, 118: 50–59

[32] Sprengel C, Baumann  K H, Henderiks  J, Henrich R ,  Neuer S  (2002). Modern coccolithophore and carbonate sedimentations along a productivity gradient in the Canary Islands region: seasonal export production and surface accumulation rate. Deep Sea Res Part II Top Stud Oceanogr, 49(17): 3577–3598

[33] Steinberg D K ,  Carlson C A ,  Bates N R ,  Johnson R J ,  Michaels A F ,  Knap A H  (2001). Overview of the US JGOFS Bermuda Atlantic Time-series Study (BATS): a decade-scale look at ocean biology and biogeochemistry. Deep Sea Res Part II Top Stud Oceanogr, 48(8–9): 1405–1447

[34] Strong A E, Eadie  B J (1978). Satellite observations of calcium carbonate precipitations in the Great Lakes. Limnol Oceanogr, 23(5): 877–887

[35] Trimborn S, Langer  G, Rost B  (2007). Effect of varying calcium concentrations and light intensities on calcification and photosynthesis in Emiliania huxleyi. Limnol Oceanogr, 52(5): 2285–2293

[36] Westbroek P, Jong  V D, Walder  P V, Borman  A H, Vrind  J P (1985). Biopolymer-mediatedcalcium and manganese accumulation and biomineralization. Geol Mijnb, 64: 5–15

[37] Zeebe R E, Zachos  J C, Caldeira  K, Tyrrell T  (2008). Oceans: carbon emissions and acidification. Science, 321(5885): 51–52

[38] Zondervan I, Zeebe  R E, Rost  B, Riebesell U  (2001). Decreasing marine biogenic calcification: a negative feedback on rising atmospheric pCO₂. Global Biogeochem Cycles, 15(2): 507–516

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag Berlin Heidelberg