Ecology and biogeochemistry of contrasting trophic environments in the South East Paciﬁc by carbon isotope ratios on lipid biomarkers

The distribution of lipid biomarkers and their carbon isotope composition was inves-tigated on suspended particles from di ﬀ erent contrasting trophic environments at six sites in the South East Paciﬁc. High algal biomass with diatom-related lipids was characteristic in the upwelling zone, whereas haptophyte lipids were proportionally most 5 abundant in the nutrient-poor settings of the centre of the South Paciﬁc Gyre and on its easter edge. Dinoﬂagellate–sterols were minor contributors in all of the studied area and cyanobacteria-hydrocarbons were at maximum in the high nutrient low chlorophyll regime of the subequatorial waters at near the Marquesas archipelago. The taxonomic and spatial variability of the relationships between carbon photosyn- thetic fractionation and environmental conditions for four speciﬁc algal taxa (diatoms, haptophytes, dinoﬂagellates and cyanobacteria) was also investigated. The carbon isotope fractionation factor ( ε p ) of the diatom marker varied over a range of 16‰ along the di ﬀ erent trophic systems. In contrast, ε p of dinoﬂagellate, cyanobacteria and alkenone markers varied only by 7–10‰. The low fractionation factors and small variations be- tween the di ﬀ erent phytoplankton markers measured in the upwelling area likely re-veals uniformly high speciﬁc growth rates within the four phytoplankton taxa, and/or that transport of inorganic carbon into phytoplankton cells may not only occur by di ﬀ usion but by other carbon concentrating mechanisms (CCM). In contrast, in the oligotrophic zone, i.e. gyre and eastgyre, relatively high ε p values, especially for the diatom marker, 20 indicate di ﬀ usive CO 2 uptake by the eukaryotic phytoplankton. At these nutrient-poor sites, the lowest ε p values for haptophytes, dinoﬂagellates and cyanobacteria infer higher growth rates compared to diatoms.

EGU to these sources (Falkowski, 1984;Falkowski et al., 1998;Irwin et al., 2006;Litchman et al., 2006). The variability of these factors and their role in biogeochemical processes emerges from the properties of the surface mixed layer. Hence, eutrophic areas, such as upwellings, with continuous nutrient supply to the euphotic zone differ strongly from areas with a permanently nutrient-depleted surface layer. In these areas, e.g. oceanic 5 gyres, a deep pycnocline prevents surface waters from a supply with deeper, more nutrient-rich waters.
Photosynthesis is a major biogeochemical process where carbon dioxide and water are converted into organic carbon with the presence of light. The fate of this organic carbon is therefore intimately linked to the conditions of its synthesis and depends 10 strongly on the composition of the phytoplankton assemblages. For understanding the global marine carbon cycle, efforts should be made to derive information relative to these ecosystems. In addition to the now widely used chlorophyll and pigment analysis (Barlow et al., 1993;Claustre et al., 2004;Mackey et al., 1996), field data about phytoplankton diversity might also be acquired by the distribution of accessory lipid biomark- 15 ers. They provide important information on the phytoplankton composition complementary to that of algal pigments, as well as on the relative importance of carbon sources from heterotrophic bacteria and other zooplankton (Dijkman and Kromkamp, 2006;Pinturier-Geiss et al., 2002;Tolosa et al., 2004).
Isotopic characterization of marine organic matter can provide insight into the condi- 20 tions under which carbon fixation occurs contributing to the understanding of the global marine carbon cycle. The carbon isotopic composition (δ 13 C) of any photosynthetic material and their derived carbon isotope fractionation factor (ε p ) reflects the δ 13 C of the carbon source utilized and the modification by the processes and environmental variables involved in its production: growth rate, temperature, dissolved CO 2 , cell ge- 25 ometry, irradiance, etc. (Burkhardt et al., 1999a;Burkhardt et al., 1999b;Eek et al., 1999;Hayes, 1993;Laws et al., 1997;Popp et al., 1998b;Rau et al., 1996). For example, high growth rates of the phytoplankton producers are linked to high δ 13 C values,

Printer-friendly Version
Interactive Discussion EGU become depleted in 13 C with increasing the concentration of dissolved CO 2 (Burkhardt et al., 1999a;Riebesell et al., 2000). This latter linkage seems to be limited in its extent, and differences of δ 13 C higher than 1-2‰ cannot be explained any more by the change in the CO 2 concentration, but they are rather accounted for by differences in the growth rates of the phytoplankton producers and carbon uptake mechanisms (Bidigare et al., 5 1997;Burkhardt et al., 1999b, Benthien et al., 2007. Since particulate organic carbon (POC) is a complex mixture of autotrophs, heterotrophs, and detritus, potentially clearer relationships between δ 13 C values and environmental conditions can be obtained using specific biomarkers derived from particular species or taxonomic groups (Bidigare et al., 1999;Pancost et al., 1999;Pancost et al., 1997), compared to the δ 13 C 10 of the bulk POC (Rau et al., 2001;Woodworth et al., 2004). This has become possible with compound-specific isotope-ratio mass spectrometry (Freeman et al., 1990;Hayes et al., 1990). Our field study uses molecular and isotopic carbon ratios of specific lipid biomarkers to evaluate the sources of the organic matter and to explore variations in the biogeo- 15 chemistry of the particulate organic matter at the different hydrodynamic and trophic environments from the South East Pacific. These include the eutrophic upwelling area off the Chilean coast, the mesotrophic and HNLC regions south of the equatorial current and the oligotrophic South Pacific Gyre, which was the major focus of this work. The major questions being addressed are: 20 1) What is the spatial distribution of lipid biomarkers of phyto-, zooplankton and bacteria in the contrasting trophic environments of the South East Pacific?
2) What is the variability of the carbon isotope fractionation (ε p ) and growth rates of different phytoplankton taxa in relation to the environmental conditions characterized by different nutrient, dissolved inorganic carbon concentrations and produc-(MAR) and a high nutrient low chlorophyll zone (HNL) upstream of the Islands. Second, an extremely oligotrophic area, very poor in nutrients, located in the centre of the South Pacific Gyre (GYR) and a less oligotrophic site in the east of the gyre (EGY). At the end of the transect, we studied a eutrophic zone highly enriched in nutrients and associated to the upwelling off the Chilean coast (UPW and UPX). UPW station was farther from 15 the coast than UPX and exhibited a more important water stratification. "Challenger Oceanics" in-situ pumps were used to filter large volumes (400 to 900 l) of water from the upper 300 m of the water column, to collect suspended particles through a Nitex screen of 70 µm and a precombusted (550 • C) Microquartz filter (QMF, Sartorius) of 1 µm pore size. 20

EGU
Total CO 2 (C T ) and total alkalinity (A T ) of water samples were measured by potentiometry (Azouzi et al., 2007) at all sites except UPW. The dissolved CO 2 concentration was calculated from C T , alkalinity, temperature, salinity and the concentrations of silicate and phosphate using the CO2SYS program developed for CO 2 system (Lewis and Wallace, 1998). This program is based on equations of the seawater CO 2 system 5 (DOE, 1994) and the dissociation constants of Goyet and Poisson (1989). For the UPW site, we assumed the same concentrations as UPX, since vertical profiles for aqueous CO 2 interpolated from the WOCE cruise P06 in GLODAP three-dimensional gridded data at the two stations indicate similar concentrations.

Lipid extraction 15
Filters containing the suspended particles were spiked with internal standards (n-C 24 D 50 , anthracene-d 10 , pyrene-d 10 , perylene-d 12 , friedeline, 5α-androstan-3β-ol and cholanic acid), and extracted by microwave oven with 40 ml of a mixture with CH 2 Cl 2 /MeOH (3:1) at 70 • C for 15 min. Isolation of the neutral and acid lipid fractions were done following the method of Tolosa and de Mora (2004). Extractable lipids 20 were saponified using 1 ml KOH 6% in methanol/water (80:20) plus 1 ml of Milli-Q water (80 • C, 1 h). Then the neutral fraction was recovered with n-hexane and subject to fractionation by HPLC on a normal phase column (Nucleosil column,20 cm×0.4 cm i.d. 5 µm) to isolate the aliphatic hydrocarbons (F1), polycyclic aromatic hydrocarbons (F2), ketone compounds (F3) and sterol and alcohol fraction (F4). Saponified solutions (200 µl, 70 • C, 1 h) to convert the alcohols and sterols to their corresponding trimethylsilyl ethers. The acid fraction was derivatized by transesterifying the lipid extract with 500 µl of 20% BF 3 in methanol at 80 • C for 1 h. Gas chromatography (GC) was performed with a Hewlett Packard HP5890 series II equipped with a flame ionization detector and split/splitless injector. Two fused 10 silica capillary columns were employed: (A) a DB-5 fused silica capillary column (30 m×0.25 mm i.d.; film thickness 0.25 µm) for neutral compounds and fatty acids and (B) a BPX-70 (SGE, 60 m×0.32 mm×0.5 um) for the fatty acids. Helium was the carrier gas (1.2 ml min −1 ). The oven temperature for the DB-5 was programmed from 60 • C (0.5 min hold) to 290 • C at 6 • C min −1 . The GC oven for the BPX-70 column was 15 programmed from 60 • C (0.5 min hold) to 250 • C at 6 • C min −1 . Injector and detector temperatures were 270 • C and 320 • C, respectively. Aliphatic hydrocarbons, ketones, sterols and fatty acids were quantified by internal standards (n-C 24 D 50 , friedeline, 5α-androstan-3β-ol, and cholanic acid, respectively). Confirmation of peak identity was obtained using GC with mass spectrometric detection 20 (GC-MS) (Hewlett-Packard 5889B MS "Engine") operated in the electron impact at 70 eV.

Compound-specific isotope analysis
The lipid biomarkers were analyzed for their stable carbon isotope composition using an HP 5890 GC equipped with a HP 7673 autoinjector and interfaced through a Introduction
The GC/C/IRMS was equipped with a 100% methylpolysiloxane fused silica column (Ultra-1, 50 m×0.32 mm i.d.; 0.5 µm film thickness) pre-connected with a press-fit connector (Supelco, France) to a 0.32 mm i.d. deactivated fused silica capillary retention gap of 5 m. Injections of 2 µl in isooctane were made via an on-column injector. The 5 GC oven was programmed from 60 to 100 • C at 10 • C min −1 , then to 310 • C at 4 • C min −1 and maintained at 310 • C for 40 min. Values reported were determined by at least three replicates to calculate the average and standard deviation. All δ 13 C values are reported in the delta notation relative to the Pee Dee Belemnite (PDB) standard as follows: δ 13 C=[( 13 C/ 12 C) sample /( 13 C/ 12 C) PDB −1] × 10 3

10
Corrections for the isotopic change introduced in the derivatisation of sterols, fatty alcohols, and fatty acids were determined through derivatisation of standards of known isotopic composition and applying the equation of Jones et al. (1991). Cholesterol, methanol, 18:0 fatty acid and 18:0 FAME of known isotopic carbon composition (measured by elemental analyser coupled to isotope ratio mass spectrometer), were used to 15 calibrate the GC/C/IRMS and correct the isotopic change introduced by the derivatization. The surrogate standards, 5α-androstan-3β-ol, cholanic acid and the GC internal standard friedelin of known isotopic composition served as internal isotopic standards. The precision (standard deviation) for most analytes with GC-C-IRMS signals higher than 0.5 V (m/z 44) was comparable to the instrument specifications (0.5‰). 20 2.6 Calculations of carbon isotope fractionation (ε p ) Molecular ε p was determined following the general equation (1) outlined in Freeman and Hayes (1992): where CO 2 is its dissolved phase in the water column and C pp the primary photosyn- 25 thate. 4660 Printer-friendly Version Interactive Discussion

EGU
In this study, direct measurement of δ 13 CO 2 was not available. Therefore δ 13 CO 2 was calculated according to the Eq. (2) of Mook (1974): where ε b is the temperature-dependent carbon isotope fractionation of dissolved CO 2 with respect to bicarbonate, T is the absolute temperature in Kelvin, and the refer-5 ence value of δ 13 for bicarbonate (b) in sea surface water was taken as +2.2‰ (Craig, 1970). δ 13 C pp (primary photosynthate) for eukaryotic organisms was estimated by using a constant isotopic fractionation between photosynthetic lipids and algal biomass, which has been estimated at 4.2‰ for alkenones (Popp et al., 1998a) This value is assumed to be the same for the rest of eukaryotic photosynthetic compounds (Hayes, 1993). For prokaryote, δ 13 C pp was estimated from the n-heptadecane using a constant isotopic fractionation between photosynthetic lipids and algal biomass of 8.4‰ (Sakata et al., 1997).

15
Carbon isotopic fractionation for phytoplankton (ε p ) which obtain CO 2 by passive diffusion is summarized by the expression of Popp et al. (1998b): where ε f is the fractionation associated with the enzyme-catalyzed carbon fixation step, β is a constant, µ is the specific growth rate, V and S are the volume and surface area  EGU growth rate and reflects the intracellular carbon demand. This b-value was calculated following the Eq. (4) of Bidigare et al. (1997): with ε f values of 25‰ for eukaryotic algae utilizing Rubisco and β-carboxylase enzymes (Bidigare et al., 1997) and [CO 2 ] aq calculated as described in Sect. 2.2.

5
Specific growth rates (µ, (d −1 )) of alkenone producing haptophytes were estimated with the following equation found by Bidigare et al. (1997) in laboratory chemostat culture experiments of Emiliania huxleyi: and applying the corrections for the effects of day length and respiration on growth rate where µ is the 24-h average growth rate, t p is day length or photoperiod in hours, and the factor 0.8 adjusts the growth rate for dark respiration.

Results and discussion
The analytical scheme used in this study identified and quantified ∼60 individual com- 15 pounds in the neutral lipid fraction and ∼40 acid compounds in the acid fraction. A summary of selected lipid biomarkers discussed in this study together with their sources is shown in Table 3. Concentrations of the selected lipid biomarkers are summarized in Table 4  ited a similar distribution as phytol concentrations (Table 4), except in the Gyre where alkenones and dinosterol peaked at lower depths than phytol and diatom markers. Highest concentrations of long-chain C 37 and C 38 alkenones were measured at the eutrophic UPW site (27 ng l −1 ) but also at the two gyre sites EGY (20 ng l −1 ) and GYR (23 ng l −1 ). These peak values corresponded to the depth of chlorophyll maximum except for the GYR site where the peak was situated above this maximum, at 125 m depth. A much lower concentration (6.5 ng l −1 ) was recorded at the MAR site. These values are much lower than those reported for suspended particles from the Bering Sea after blooms of Emiliana Huxleyi and ranging from 0.15 to 3.12 µg l −1 (Harada et al., 2003), but similar to concentrations observed in suspended particles collected  EGU consistent with the abundance distribution of prokaryotic phototrophic organisms, with high abundances at the HNL and EGY sites (Grob et al., 2007). In all samples, the n-alkanols were dominated by the short-chain fatty alcohols of even carbon number (n-C14, n-C16 and n-C18), which are associated to zooplankton markers (Sargent et al., 1977). With the exception of the two gyre sites (GYR and 5 EGY), maximum concentrations of lineal alcohols were found below the depth of chlorophyll maximum, at the upwelling sites (UPW, UPX) at 300 m and at the Marquesas sites (MAR, HNL) at 100 m depth. In the Gyre, maximum concentrations of n-alcohols coincided with the phytol maximum, though another peak (11.6 ng l −1 ) was registered at the surface of the GYR site. The fatty acids C 20:1 and C 22:1 , typical markers of herbivorous mesozooplankton (Lee et al., 2006), exhibited the highest concentrations at the upwelling sites (UPW, UPX) at the depth of chlorophyll maximum, but also below the euphotic zone (300 m) at the MAR and UPX sites.
Similar to the phytol distribution, the concentrations of bacterial biomarkers, such as branched fatty acids, were highest at the UPW site (Table 3), whereas the concen- 15 trations of zooplankton markers, such as cholesterol and n-alcohols exhibited higher abundance at UPX. In general, heterotrophic bacterial populations seemed to be associated to diatom biomass, which is supported by the positive correlation between the δ 13 C of the branched fatty acid (i -C 15 FA) and the δ 13 C of the C 20:5 FA (r=0.81, p<0.05). 20 Some more insight into the phytoplankton distribution may be gained by comparing the relative contribution of the biomarkers within the total neutral lipids, which are illustrated in Fig. 1. The percentage of phototrophic biomarkers generally followed the chlorophyll distribution except at UPX, where phytol and diatoms markers peaked at 100 m depth. Also at HNL, phytol showed relatively high percentages down to 100 m 25 depth. This may be related to the highest relative importance of diatom sterol observed in the euphotic layer ( Fig. 1b) and to the presence of diatoms which formed "balls of needles" or clusters (Gómez et al., 2007). In contrast to the diatom biomarkers, the percentage of total alkenones was by far highest in the gyre especially above the chloro-Introduction EGU phyll maximum (Fig. 1c), where also concentrations had been among the highest of all sites (Table 4). Prymnesiophytes were likely to be the major constituents of eukaryotic phytoplankton in the mixed surface layer of the gyre, suggesting that haptophytes are well adapted to the low nitrate concentrations prevailing in the oligotrophic zone of the Pacific Gyre. The C 28 ∆ 5,24(28) sterol/alkenones ratio provides us with information about 5 the relative contribution of diatoms to prymnesiophytes. Highest ratios were obtained at the UPX site, and in particular below the euphotic zone. Haptophytes predominated over diatoms at the GYR site, especially above the chlorophyll maximum whereas the diatom signal showed a deeper maximum at 175 m corresponding to the maximum of both phytol and chlorophyll a. Dinosterol showed overall low percentages (<3%), 10 which indicated a minor contribution of dinoflagellates in the algal mixture of these Pacific waters, as it was also confirmed by pigment analyses (Ras et al., 2007). Figure 2 illustrates other diagnostic biomarkers indices to evaluate the relative dominance of zooplankton and bacterial sources within each site as well as the state of the particulate material in the different zones. The cholesterol/phytosterol ratio which indi-15 cates dominance of zooplankton over phytoplankton increased with depth at all sites, and in particular at UPX. This was consistent with the substantial concentrations of n-alcohols and zooplanktonic C 20:1 and C 22:1 fatty acids found at 300 m depth. At most of the sites, the relative importance of bacterial fatty acids was higher below than at the depth of chlorophyll maximum. A contrasting image was observed at UPX where 20 this relative importance was lower. In fact, high bacterial production and a negative net community production were reported from the euphotic zone of UPX (Van Wambeke et al., 2007), which suggests rather "decomposing" conditions compared to "productive" conditions at UPW. At all sites, the ratio of phytosterols/phytol indicated more degraded phytoplankton material at depths below the chlorophyll maximum, but also above, at 25 the site of the gyre. This ratio showed a slightly more degraded material in the euphotic zone of the UPX site compared to UPW, but fresher material at 300 m depth of UPX.
The polyunsaturation index of C 16 fatty acids (PUFA % of C 16 ) is an indicator of the ecophysiological state of marine diatom populations because storage lipids, mainly Interactive Discussion EGU C 16:0 and C 16:1 FA, are synthesized during senescence, rather than during logarithmic growth (Shin et al., 2000). The high indices observed in the euphotic zone of the UPW site and at 300 m depth of UPX suggest that these PUFA originated from diatoms at logarithmic growth. Moreover, the carbon isotope ratios of lipid biomarkers in the euphotic zone were generally more enriched at UPW compared to UPX (Fig. 3), likely 5 indicating higher growth rates at UPW than at UPX. However, below the euphotic zone of UPX, δ 13 C values identified higher growth rates at depth compared to the surface.
All these parameter point out that post-bloom conditions with high concentrations of animal-derived detritus prevailed at the surface of the UPX site, whereas the important signal of zooplankton and diatom markers below the euphotic layer indicated the 10 presence of zooplankton feeding on phytoplankton produced during bloom conditions. These findings are supported by the highest stock and flux measured at UPX compared to the UPW site (Miquel et al., 2007 2 ) and the high concentration of detritus and senescent colonial diatoms observed by microscope in samples from the euphotic zone at the UPX site (Gómez, personal communication, 2006). These conclusions contrast, 15 however, with those derived from pigment biomarkers (Ras et al., 2007) where UPW site was characteristic of a typically mature bloom of diatoms and the phytoplankton at site UPX was probably at an early stage of development. Also, much higher nutrient concentrations at the surface of UPX site indicated a recent upwelling of deep water. Overall, these contrasted observations might be accounted for by the different turnover 20 and lability between pigment and lipid compounds which represent different pools of the particulate matter. Pigments are relatively more labile and associated with the living material. In contrast, lipids are included in both, the living and detrital particulate pools of the matter.
3.2 Vertical distribution of biomarkers in the center of the gyre 25 Depth profiles of selected accessory lipid biomarkers in suspended matter from the center of the Gyre are presented in Fig. 4. Phytol concentrations showed very low surface values and they were increasing progressively with depth, with maximum concen-4666 abundance. Branched fatty acids which are derived from heterotropic bacteria exhibited the maximum concentration between 150 and 200 m depth, following the same trend as planktonic biomass. This feature indicates that the bacterial population is associated with the major planktonic biomass. Other more specific phototrophic biomarkers, such as sterol markers for diatoms exhibited a similar profile as phytol with two maximum at 150 and 200 m depth. Only few macro diatom species, such as Nitzschia and Dactyliosolen were observed between 200 and 300 m depth and a significant number of Bacteriastrum associated to a cyanobacteria symbiont was also observed around 140 m depth (Gómez, personal communication, 2006). In contrast, the sterol marker for dinoflagellates exhibited a The C 37 alkenones, which are specific markers for some algae of the class Haptophyceae/Prymnesiophyceae, including the coccolithophorid species such as Emiliana huxleyi, exhibited the maximum concentration above the deep chlorophyll maximum, at 25 125 m depth. Cell densities of different coccolitophorid taxa showed, however, different depth profiles, with a maximum peak between 150-200 m for Emiliana huxleyi whereas other taxa peaked at shallower depths (∼100 m) (Beaufort et al., 2007). These findings indicate that (i) other noncalcifying haptophytes might synthetize alkenones at shal- EGU lower depths, (ii) that alkenones are not associated with the integrity of coccospheres and/or (iii) that cellular alkenone concentrations varied with the physiological status and species composition of the coccolithiphorid assemblage. On the other hand, the concentration of alkenones and the accessory carotenoid 19'Hexanoyloxyfucoxanthin (19'HF) (Ras et al., 2007), characteristics of prymnesiophytes, also exhibited differ-5 ent depth distributions (Fig. 5). Such discrepancy reflects that alkenone-producers in these waters are small contributors to the 19'HF stock, and that the habitat of alkenone synthesizers diverges from that of the major phytoplankton taxa contributing to the 19'HF distribution. An analogous feature was observed at station ALOHA from the oligotrophic North Pacific Subtropical Gyre (Prahl et al., 2005) and other studies showed 10 that 19 ′ HF abundance was generally not tightly correlated with that of coccolithophorids (Dandonneau et al., 2006). Figure 6 illustrates the carbon isotope composition of the diunsaturated alkenone together with the total concentrations of C 37 alkenones. More enriched δ 13 C values which infer higher growth rates were obtained for those haptophytes inhabiting the depth of the chlorophyll maximum. However, the higher con-15 centrations of alkenones found at 125 m depth were associated to lowest growth rates which indicates a higher concentration of cellular alkenones. Similar observations were reported in nitrate-limited cultures where a small increase in the concentration of cellular alkenones occurred with decreasing growth rates (Conte et al., 1998;Parrish et al., 1998). 20 The growth unsaturation index (U K ′ 37 ) which is widely used as a proxy of sea surface temperature (Prahl and Wakeham, 1987) was calculated as the relative proportion of di-and triunsaturated C 37 alkenones. This index was converted to a measure of temperature by the commonly used empirical calibration equation T =(U EGU sis. Similar observations were reported in winter at ALOHA station (Prahl et al., 2005). They were explained by simply biogeographical variations observed in the alkenone vs. temperature relationship in natural waters, which may reflect differences in genetic and physiological status of the local alkenone-synthesizing populations. Since haptophytes have a low phosphorous requirement (critical concentrations of 0.2 µM), 5 nitrogen limitation seems likely since inorganic N concentrations at the Gyre stations were well below the half-saturation constant (K s , the concentration supporting an uptake rate one-half the maximum rate) determined for E. Huxley (≤0.5 µM) (Eppley et al., 1969). Overall and according to batch cultures of haptophytes (Epstein et al., 1998), the observed increase in the U K ′ 37 values with the consequent overestimation of the tem-10 peratures, might indicate that this marine phytoplankton taxon is under nutrient-limited "stationary growth" conditions. This overestimation can also be explained by a change in the ecology of alkenone-producing algae (Popp et al., 2006) and by autoxidation of alkenones in these highly irradiated waters (Rontani et al., 2006), especially when residence times of particles are long.

Biogeochemical implications from carbon isotope fractionation
Stable carbon isotope differences between the inorganic carbon source and that of organic carbon synthesized by autotrophic organisms known as photosynthetic carbon fixation (ε p ), can assist in distinguishing between the different CO 2 fixation pathways ( Table 5). Maximum carbon isotope fractionation of photoautotrophic organisms using 20 the Calvin cycle, like micro-algae and cyanobacteria is in the range of 20 to 27‰ (Popp et al., 1998b;Sakata et al., 1997). However, the ε p expected for biomarkers derived from eukaryotes can vary between 5 and 25‰ depending basically on [CO 2 ], growth rate and the ratio of cellular surface area to volume (Bidigare et al., 1997a;Popp et al., 1998). In contrast, ε p for prokaryotes (cyanobacteria) range between 16 and 22‰ 25 because the large surface-to-volume ratio guarantees a large CO 2 supply relative to the cellular demand. Others pathways, apparently restricted to other bacteria, such as anoxygenic phototrophic bacteria, are the reversed tricarboxylic acid cycle and the 4669 We observed higher taxonomic variations in ε p for eukaryotic algae growing in the oligotrophic areas (variations of ∼10 ‰) compared to the eutrophic sites of the upwelling (variations of 3 to 7‰). The variation of the carbon isotope fractionation for 5 the diatom marker covered a range of ∼16‰ along the different trophic systems. In contrast, ε p of dinoflagellate and alkenone markers varied much less ca. 10 and 7‰, respectively.
The carbon isotope fractionation of diatom markers exhibited a significant, negative logarithmic correlation with measured silicate concentrations (ε p =−4.3 ln[SiOH 4 ] 10 +23.57, n=16, r=0.96, p<0.01, Fig. 8), whereas lower correlations were found with the other nutrients, nitrates and phosphates. Lower correlations were also obtained with the ε p of dinoflagellates and no correlation was observed between ε p of alkenone and nutrients. These findings confirm (Pancost et al., 1999) that silicate is likely one of the essential nutrients which control the ε p and growth rates of diatoms in the Pacific 15 ocean.
The carbon isotope fractionation of diatom and dinoflagellate markers showed a negative, linear correlation with [CO 2 ] aq (r diatom =0.94 (Fig. 9a), r dinoflagellate =0.91, n=16, p<0.01). These relationships deviate from the previously reported general oceanic trend (Rau et al., 2001) and culture studies (Burkhardt et al., 1999a) where carbon 20 isotope fractionation increases (δ 13 C decrease) when [CO 2 ] aq increases. However, this apparent deviation has also been observed in Peruvian upwelling waters where it was suggested that a diatom carbon concentrating mechanism (CCM) was likely the cause of the lower ε p of diatoms in these waters with high [CO 2 ] aq . A much lower correlation (r=0.75, n=14, p< 0.01) and a slope close to zero was observed for the ε p 25 of alkenone (Fig. 9b) which agrees with other studies that showed the small effect of [CO 2 ] aq on isotope fractionation in E. Huxleyi compared to potential changes of ε p due to growth rate and carbon uptake mechanisms (Benthien et al., 2007;Bidigare et al., 1997). No correlation was found between [CO 2 ] and the ε p of n-C17 (r CO2 =0.37, n=6, 4670 EGU p=0.46) which is also consistent with a previous work (Popp et al., 1998b) who found for Synechococcus that ε p is independent of the concentration of dissolved CO 2 , likely because its cell geometry guarantees a large CO 2 supply.
Overall, the lower isotope fractionation factors and smaller variations between the different eukaryotic markers measured in the upwelling area might indicate uniformly high 5 growth rates for the three phytoplankton taxa, diatoms, haptophytes and dinoflagellates and/or that phytoplankton may employ carbon concentrating mechanisms (CCM) other than diffusion, which actively transport inorganic carbon into cells. Similar findings were reported by other authors (Pancost et al., 1999;Pancost et al., 1997;Rau et al., 2001;Werne and Hollander, 2004) who invoked that an active transport of bicarbonate 10 into the cell may still play a role in the carbon isotope fractionation by phytoplankon in upwelling areas with high concentrations of CO 2 .
In contrast, the GYR and EGY sites exhibited the highest carbon isotope fractionation factors for eukaryotic algae and in particular for the diatom marker. Their values reached 25-26‰ which is close to the maximum isotope fractionation of eukaryotic 15 algae utilizing Rubisco and β-carboxylase enzymes (Goericke et al., 1994;Laws et al., 1997). Such high ε p values cannot be obtained by bicarbonate uptake and are indicative of diffusive CO 2 uptake. Moreover, active uptake of carbon in oligotrophic sites is rather unlikely because of the higher metabolic energy required and CCM may be inhibited by the low oceanic concentrations of certain trace metals (Morel et al., 1994). 20 In these oligotrophic sites, lowest ε p values were measured for the haptophytes and dinoflagellates, which might infer higher growth rates for these organisms compared to diatoms.
In the mesotrophic areas of the Marquesas Islands (MAR and HNL sites), ε p values for alkenone producers and dinosterol were also lower than those for diatoms, but still 25 higher than those estimated in the upwelling area. The similar ε p values for alkenones measured in the poor-nitrated waters of the Gyre and in the high nutrient waters of the HNL site suggest that nitrogen and phosphate are not the limiting nutrients affecting carbon isotope fractionation by the prymnesiophyte algae. This finding contrasts with EGU a study from the NE Pacific where nitrogen starvation seemed to affect the ε p values for alkenones (Eek et al., 1999) and adds further support to the "trace-metal-growthrate" hypothesis (Bidigare et al., 1997), which suggested that micronutrients control growth regardless of the concentrations of PO 4 . Overall, the relatively low ε p values for alkenones from the oligo-and mesotrophic waters seem to indicate the use of a CCM 5 other than diffusion. However, recent studies provided clear evidence that haptophytes have developed an inefficient but regulated CCM, with a direct uptake of HCO − 3 (Rost et al., 2003). The highest ε p for dinoflagellates at the GYR and EGY sites are likely associated to lower growth rates and might be explained by the low N:P ratios since optimum dinoflagellate growth occurs at ratios ranging between 6 and 15 (Hodgkiss 10 and Ho, 1997).
The carbon isotope fractionation derived from n-C 17 alkane reached the values of 17-19‰ in the mesotrophic waters of the HNL and MAR sites, which are within the range of 16-22‰ reported for cyanobacteria biomass (Sakata et al., 1997). In contrast, the low carbon isotope fractionations (8 to 12‰) obtained for n-C17 in the oligotrophic waters of the GYR and EGY sites as well as in the eutrophic waters of the upwelling area are suggestive of microorganisms which use CO 2 -concentrating mechanisms. Despite substantial advances over the past few years, in the understanding of the mechanism and genes involved in cyanobacterial CCMs (Badger and Price, 2003), the induction of the CCM mechanism remains to be determined (McGinn et al., 2003;20 Woodger et al., 2005). On the other hand, since similar ε p values were measured in the upwelling area for biomarkers derived from eukaryotic plankton, it cannot be ruled out that n-C 17 in the upwelling sites is mainly derived from some eukaryotic source.
The b-values and growth rates for alkenone-producing haptophytes varied almost tenfold, ranging from 65 to 500‰µmol kg −1 and from 0.21 to 1.7 d −1 , respectively. The   (Bidigare et al., 1997;Harada et al., 2003;Laws et al., 2001) but they are slightly higher than those reported from the Peru upwelling zone (197-397 µmol kg −1 and 0.5-1 d −1 ) (Bidigare et al., 1997). This is probably related to the strength of the upwelling as indicated by the higher nutrient and CO 2 concentrations of our samples.

5
The b-values for the alkenone synthesizer phytoplankton correlated with concentrations of silicate, nitrate and phosphate concentrations. However, these correlations might not indicate that the nutrients themselves are the growth limiting factor, since a natural correlation between concentrations of dissolved CO 2 and nutrients occurred. Instead, specific growth rate may be controlled by some trace micronutrient that co-10 varies with CO 2 , silicate, nitrates and phosphate concentrations (Bidigare et al., 1997;Shaked et al., 2006).

Summary and conclusions
Our results along the different trophic systems, showed that source-specific algal biomarkers and compound specific isotope analyses largely responded to the com-15 position of the phytoplankton and to the different processes of carbon acquisition. It has become clear that natural variability in carbon isotope fractionation among algal taxa, is a consequence of multiple factors. Although further studies are needed to clearly define the relationship between ε p and the specific-growth rates of the different phytoplankton taxa, our field study illustrates the large shift in isotopic fractionation by 20 the different groups of phytoplankton taxa which should reflect the high variability in growth rates and carbon assimilation mechanisms. As a summary, we identified the upwelling zone as the one with highest algal biomass, most developed carbon concentration mechanisms and highest growth rates. In contrast, the oligotrophic area of the Gyre was characterized by the lowest concentration levels of phyto, zoo-and bacterial 25 markers, extremely deep concentration maxima, relatively high contribution of haptophytes and high carbon isotope fractionation factors indicative of low growth rates and 4673 EGU higher diffusive CO 2 uptake. P. A., and Wakeham, S. G.: Consistent fractionation of C in nature and in the laboratory: growth-rate effects in some haptophyte algae, Global Biogeochem. Cy., 11, 279-292, 1997. Bidigare, R. R., Hanson, K. L., Buesseler, K. O., Wakeham, S. G., Freeman, K. H., Pancost, R. D., Millero, F. J., Steinberg, P. A., Popp, B. N., Latasa, M., Landry, M. R., and Laws, E. A.: Iron-stimulated changes in 13 C fractionation and export by equatorial Pacific phytoplankton: 5 Toward a paleogrowth rate proxy, Paleoceanography, 14, 589-595, 1999. Burkhardt, S., Riebesell, U., and Zondervan, I.: Effects of growth rate, CO 2 concentration, and cell size on the stable carbon isotope fractionation in marine phytoplankton, Geochim. Cosmochim. Ac., 63, 3729-3741, 1999a. Burkhardt, S., Riebesell, U., and Zondervan, I.: Stable carbon isotope fractionation by marine 10 phytoplankton in response to daylength, growth rate, and CO 2 availability, Mar. Ecol.-Progr. Ser., 184, 31-41, 1999b.  , 62, 69-77, 1998b. 4678 occurrence, formation, transformation and ultimate fates, Mar. Chem., 5, 573-584, 1977. Shaked, Y., Xu, Y., Leblanc, K., and Morel, F. M. M.: Zinc availability and alkaline phosphatase activity in Emiliania huxleyi: Implications for Zn-P co-limitation in the ocean, Limnol. Oceanogr., 51, 299-309, 2006. Shin, K. H., Hama, T., Yoshie, N., Noriki, S., and Tsunogai, S Table 3. Summary of the lipid biomarkers discussed in this study.