Carbon isotopic composition of suspended and sinking particulate organic matter in the northern South China Sea—From production to deposition

(1)

Deep-Sea Research II 54 (2007) 1504–1527

Carbon isotopic composition of suspended and sinking

particulate organic matter in the northern South China

Sea—From production to deposition

Kon-Kee Liu

a,

, Shuh-Ji Kao

b

, Han-Chieh Hu

a

, Wen-Chen Chou

c

,

Gwo-Wei Hung

d

, Chun-Mao Tseng

e

a_{Institute of Hydrological Sciences, National Central University, 300 Jhongda Road, Jhongli, Taoyuan 32001, Taiwan, ROC} b_{Research Center for Environmental Changes, Academia Sinica, Taiwan, ROC}

c_{Institute of Marine Environmental Chemistry and Ecology, National Taiwan Ocean University, Taiwan, ROC} d_{General Education Center, National Penghu University, Taiwan, ROC}

e_{Institute of Oceanography, National Taiwan University, Taiwan, ROC}

Accepted 12 May 2007 Available online 23 July 2007

Abstract

Between May 2004 and March 2005, samples of suspended particulate matter (SPM) were collected from the top 200 m on ﬁve cruises to the South-East Asia time-series study (SEATS) Station. Isotopic and elemental analyses of the organic matter in these samples gave d13_{C values ranging from 25.2% to 21.3% with a decreasing trend downward, and C/N}

ratios ranging from 5.5 to 11.4 with a weighted mean value of 6.74, which is very close to the Redfield ratio, suggesting a predominantly marine origin. The temporal isotopic variation in the surface layer has been successfully simulated with the algorithm based on diffusion-controlled carbon uptake during photosynthesis. The calculation of the carbon isotopic composition of phytoplankton was based on observed values of hydrographic, isotopic and chemical variables. It is noted that variations in the biological parameters, including the specific growth rate, enzymatic isotope fractionation during carbon fixation, cell size, and cell wall permeability, within the normal ranges may have contributed significantly to the observed isotope variability. According to simulation using the same algorithm, isotopically very light particulate organic carbon (POC) could be produced in the subsurface euphotic zone due to the much reduced specific growth rate, but the contribution of the subsurface production to the sinking flux was probably not significant. Sediment traps deployed from September 2001 to May 2002 in the northern South China Sea (SCS) provided samples for isotopic and elemental analyses of the organic matter. The measurements gave d13C values ranging from 25% to 20.8% and C/N ratios ranging from 5.5 to 18. The isotopic variation of organic carbon in the sediment trap samples was successfully explained by the mixing of terrigenous organics (d13C ¼ 25.5% and C/N ¼ 22) and marine organics (d13C ¼ 22.171.1% and C/N ¼ 6.6371). The latter composition is very close to the weighted mean composition of the suspended particulate organic matter (POM) from the top 20 m, implying the surface water as the major source of organic matter in sinking particles. Compared to previously reported results of samples collected from the seafloor in the SCS, the inferred d13_{C values of the marine}

organics in the sinking ﬂux are lower than those (d13C ¼ 22.9% to 20.1%) of the POC in the nepheloid layer, which are in turn lower than those of the organics (d13C ¼ 21.5% to 18.8%) in surﬁcial sediments. The progressively heavier POC below the top 200 m is contrary to the trend of decreasing d13CPOCin the top 200 m. We have demonstrated that the Suess

ARTICLE IN PRESS

www.elsevier.com/locate/dsr2

Corresponding author. Tel.: +886 3 422 3354; fax: +886 3 422 2894. E-mail address:kkliu@ncu.edu.tw (K.-K. Liu).

(2)

effect and the elevated concentration of aqueous CO2in the surface water due to the increasing atmospheric CO2partial

pressure may cause depletion of 13_{C more than enough to account for the observed depression of d}13_{C values in the}

progressively younger POM. In the past, diagentic isotopic alteration has been proposed as the process responsible for both the decreasing trend of d13_C

POCin the surface layer and the increasing trend in the subsurface layer. Although the

diagenetic effect cannot be ruled out, this study shows that other processes are sufﬁcient to explain the observed trends of isotopic variation of POC.

Keywords: Isotope fractionation; Phytoplankton growth; SEATS; Suess effect; Sediment traps; Sediments

1. Introduction

The South China Sea (SCS) is the largest marginal sea in the world aside from the ice-covered Arctic Ocean (Sverdrup et al., 1942). It has a deep basin reaching a maximum depth of 5006 m off southwest Luzon (NGDC, 2001) and is connected to the open ocean mainly through the Luzon Strait for effective water exchange with the western North Pacific. The circulation and, consequently, the biogeochemistry of the SCS are subject to forcing of the alternating East Asian monsoons (Shaw and Chao, 1994; Liu et al., 2002). Inter-annual varia-tions in the SCS are mostly associated with modulation of the monsoon strength by El Ninõ-Southern Oscillation events (Chao et al., 1996;Wu et al., 1999). On a longer time scale, variations in sedimentary records, such as organic carbon con-tents, during glacial–interglacial cycles were attrib-uted to climate changes manifested as variations in monsoon strengths. The responsiveness of the SCS to climate changes has inspired active sampling of the SCS sediments (Chen et al., 1998; Prell et al., 1999). Subsequently, paleoceanographic studies on the SCS have flourished in recent years (e.g.,Wang et al., 2004).

Considerable attention has been given to the isotopic records of sedimentary organic matter in the SCS (Kienast et al., 2001; Higginson et al., 2003). Because the isotopic composition of sedi-mentary organic matter has been widely used to estimate the CO2partial pressure in the geological

history (Rau et al., 1989), a critical evaluation was performed on the usefulness of carbon isotopic records of sedimentary organic matter from the SCS and other low-latitude oceanic sites as a proxy indicator of aqueous CO2 concentration (Kienast et al., 2001), which is related to the CO2 partial

pressure in the atmosphere. The authors concluded that better constraints must be imposed on the mechanisms governing the isotopic composition of

marine particulate organic carbon (POC). There-fore, it is critical to better understand the processes controlling the isotopic variation from the produc-tion to the deposiproduc-tion of POC in the ocean. For the SCS, there has been one study on the isotopic composition of dissolved inorganic carbon (DIC) (Lin et al., 1999), which was based on field observations conducted in 1995 and none on that of organic carbon in the SCS proper. In fact, there have been few, if any, studies, in which the isotopic compositions of POC in the entire vertical profile from the surface ocean to the nepheloid layer and the surficial sediment were examined; let alone a detailed mechanistic understanding of the processes responsible for the isotopic composition of depos-ited organic carbon on seafloor.

The South-East Asian Time-series study (SEATS) project (Karl et al., 2003; Tseng et al., 2005) has allowed us to conduct seasonal samplings of suspended particular organic matter (POM) at the SEATS Station (18.01N, 115.61E) (Fig. 1A). Sedi-ment traps deployed in the northern SCS pro-vided samples of sinking particles for comparison (Fig. 1A). The sampling of the suspended POM covered all four seasons of an annual cycle, while the trap deployments covered three seasons. These samples together with previously reported d13C data (Kienast et al., 2001) of organic matter from the ﬂuff samples in the nepheloid layer and surﬁcial sediments (Fig. 1B) provide us a unique opportunity to examine d13CPOC from the entire water column.

Although many localities of observations discussed in this study are not close to the SEATS Station, all sites are within the oligotrophic regime of the subtropical SCS with average sea-surface chlorophyll concentration within 0.1–0.3 mg m3 (Fig. 1B). It is reasonable to presume that their biogeochemical characteristics are similar, following the widely accepted principles of using ocean color data for oceanic ecological characterization (e.g.,

Longhurst et al., 1995;Ducklow, 2003).

ARTICLE IN PRESS

(3)

ARTICLE IN PRESS

Fig. 1. Locations of sampling sites in the northern South China Sea. SEATS indicates the main site of the SEATS Station. M1S and M2S indicate sites of sediment trap deployments (Table 2). (A) Bathymetry map. (B) Average Chl-a distribution derived from SeaWiFS data. The polygon indicates the sampling area of surﬁcial sediments and the ﬂuff materials (Kienast, 2000;Kienast et al., 2001). See text for detail.

K.-K. Liu et al. / Deep-Sea Research II 54 (2007) 1504–1527 1506

(4)

It is noted that other subprojects of SEATS provided ancillary data of hydrography and carbo-nate chemistry, which enabled us to simulate the d13C values of POC produced by phytoplankton during the study period. The simulation in turn permitted us to understand better the isotopic variation of phytoplankton in the euphotic zone and to delineate relationships between the POM produced in the upper water column and the organic matter in particles sinking through the water column, resuspended in the nepheloid layer and accumulated in surﬁcial sediments. The purpose of this study is to improve our knowledge on the factors controlling the d13C value of organic matter deposited on the seaﬂoor.

2. Materials and methods

2.1. Sampling

Water samples were collected for sampling of suspended particulate matter (SPM) on five cruises between May 2004 and March 2005 at the SEATS Station (Table 1). The CTD-Rosette sampler fitted with 20-L go-flo bottles were used for the sampling on cruises on board R/V Ocean Researcher I. Ten-liter go-flo bottles were used on the FR1-SC33 cruise. Twelve to 22 samples were collected on 1–3 hydrocasts depending on the schedule of the cruise. For collection of SPM samples, about 10 L of seawater were first stored by filling a opaque plastic bottle and then filtered through a Whatman quartz filter, 25 mm in diameter with mean pore size of 2.2 mm, on board the research vessel shortly after collection. Each of the opaque plastic storage bottles was calibrated for its volume before use. The quartz filters were preheated at 500 1C in an oven for the removal of organic matter and wrapped in aluminum foil before use. After filtra-tion, the filters were folded and wrapped again in aluminum foil and stored in a freezer at 20 1C until analysis.

The French-made time-series sediment traps, Technicap PPS 3/3, were used in this study to collect settling particulate samples. This type of trap has an aspect ratio of 2.5 and a collection area of 0.125 m2. The traps were deployed at three to four depths at two sites (Table 2) near the SEATS Station in the northern SCS (Fig. 1A). Each of the two sites was occupied individually during one period between September 2001 and May 2002 (Table 2). Each collection cup was ﬁlled with a buffered (5% or 1.7 M) formalin solution (7.5opHo8.0) to poison ‘‘swimmers’’ and prevent the trapped particles from degradation. Nine to 12 samples were obtained from each trap for this study. Collectively they covered roughly an annual cycle.

Upon recovery of the traps, the samples were immediately stored in a cold-room below 4 1C until they were brought back to the laboratory. After removal of the supernatant, the trapped particles were wet-sieved through 1 mm nylon mesh with 0.2-mm filtered seawater to remove large organisms. Zooplankton swimmers that passed through the sieve were manually removed under a microscope. The sample was then divided into sub-samples by a perimatic dispenser (Jencons Ltd.). One of the sub samples was filtered with a pre-weighed 47-mm Millipore filter (0.45-mm pore size) and the filtered sample was rinsed with deionized distilled water to remove the sea salt. The desalted sample was dried at 40 1C for 48 h. Then the sample was scraped off from the filter paper, ground to powder and stored in an auto-desiccator until it was analyzed.

2.2. Isotopic and chemical analyses

The filters bearing SPM were dried at 70 1C in an oven, and then acidified for the removal of carbonate. The acidification was carried out by wetting the filter with 1 mL of 1 N solution of reagent grade HCl. Afterwards the filters were dried at 60 1C in an oven for 48 h. Dried sediment trap samples were weighed into plastic centrifuge tubes

ARTICLE IN PRESS

Table 1

Sampling cruises for the collection of suspended particulate matter for carbon and nitrogen isotope analysis

Cruise Date No. of hydrocasts Depth range (m) No. of samples

OR1-717 May 3–8, 2004 1 0200 10

OR1-726 August 2–6, 2004 1 0200 12

OR1-736 November 5–11, 2004 2 0200 22

OR1-743 January 20–24, 2005 3 0300 20

FR1-SC33 March 26–April 2, 2005 2 0200 22

(5)

and treated with 2 mL of 1N HCl solution for the removal of carbonates. The centrifuge tube bearing the sample and acid solution was dried in an oven at 50 1C for 48 h. Then the acidiﬁcation procedure was repeated once more to ensure complete removal of carbonates.

The decarbonated samples were analyzed for carbon isotopic composition and carbon/nitrogen contents in a continuous flow elemental analyzer-isotope ratio mass spectrometer system. The filter and trap samples were combusted in an elemental analyzer (Flash EA-1100 NC, Thermo-Finnigan) to produce CO2and N2, which flowed through a GC

column for separation, then into a mass spectro-meter (Thermo Finnigan Deltaplus Advantage) for measurements of carbon and nitrogen contents and the carbon isotopic composition. The isotopic composition is expressed in the d-notation deﬁned as follows:

d13C ¼ ð1 Rsam=RrefÞ1000%, (1)

where Rsamis the 13

C/12C ratio of the samples and Rref is the ratio of the reference, which is Pee Dee

Belemnite (Hoefs, 1997). USGS 40 (L-glutamatic

acid) and IAEA-CH6 with certiﬁed value of 26.3% and 10.4%, respectively, were used as working standards. The long-term reproducibility of carbon isotope determination for the two international standards and ﬁeld samples is better than70.2%.

Measurements of DIC and titration alkalinity (TA) for seawater samples followed the standard operating procedures (DOE, 1994). The coulometric method was used for DIC measurements with a precision of 70.1%. TA was determined by the potentiometric titration method in an open cell with a precision of70.2%. Seawater references provided by A.G. Dickson of the Scripps Institution of Oceanography had been used for calibration and accuracy assessments throughout this study. Differ-ences between the certiﬁed values and our measure-ments are less than 2 and 3 mmol kg1for DIC and TA, respectively. The concentration of aqueous

CO2was computed from DIC and TA data using

the standard procedure (Lewis and Wallace, 1998). For the analysis of d13C of DIC, 40 mL of seawater were acidiﬁed with 2 mL 85% phosphoric acid in a pre-evacuated vessel, and the liberated CO2

gas was then trapped with liquid nitrogen under vacuum. Isotopic analysis was performed with a VG Optima mass spectrometer. The overall procedural error for d13CDICanalyses was better than70.05%.

3. Results

We ﬁrst present isotopic compositions of POC collected from the ﬁve cruises and ancillary data obtained concurrently. Then we describe results of isotopic and elemental analyses of sediment trap samples retrieved from the two mooring sites in the northern SCS near the SEATS Station.

3.1. Isotopic compositions of suspended particulate carbon

The ﬁve cruises between May 2004 and March 2005 that provided samples for this study covered roughly an annual cycle. The temperature of seawater in the surface layer of 10–50 m (Fig. 2A) faithfully reﬂected the seasonal variation of hydro-graphic conditions as reported before (Tseng et al., 2005). The uniform temperature distribution in the top 50 m observed in November 2004 and January 2005 was a result of the deepest mixed layer, reaching 100 m, in late autumn and early winter. The concentration of POC (Fig. 2B) varied in a nearly reverse relationship with the temperature change. This is consistent with the reverse rela-tionship found between sea surface temperature and chlorophyll (Tseng et al., 2005). By contrast, the d13C value of POC showed some variation (Fig. 2C), but there was no clear correlation with either of the other two parameters.

The hydrographic, chemical and isotopic proﬁles showed some consistent patterns of vertical varia-tion as demonstrated by two sets of data from the

ARTICLE IN PRESS

Table 2

Durations and locations of sediment traps deployed in the northern South China Sea and number of samples obtained for this study Site Duration Location Interval Trap depths (m) [no. of samples] M1S Sept. 22, 2001–Mar. 21, 2002 21130.90_{N, 119127.8}0_E _{15 days} _{374 [11], 925 [11], 1925 [12], 2700 [12]}

M2S Dec. 16, 2001–May 14, 2002 19100.00_{N, 117129.5}0_E _{15 days} _{447 [9], 1248 [10], 3250 [10]}

(6)

top 200 m at the SEATS Station, one from summer (Fig. 3) and the other from winter (Fig. 4). The summer condition showed a rather thin mixed layer of about 35 m (Fig. 3A). The d13C values of POC in the mixed layer were quite uniform, while the concentration of POC showed a surface maximum (Fig. 3B). Also plotted are the concentrations of aqueous CO2and the d13C values of DIC observed

on the same cruise (Fig. 3C). Both parameters were quite uniform in the mixed layer. Below the mixed

layer, the concentration of POC decreased, while the DIC concentration increased. The d13C values of both POC and DIC decreased, but the ranges of variation were quite different. The isotopic compo-sition of POC varied in the range from 24.2% to 21.5%, while that of DIC varied in a very narrow range from 0.1% to 0.9%.

The winter condition showed a much thicker mixed layer, approaching 100 m (Fig. 4A). All carbon parameters showed rather uniform values

ARTICLE IN PRESS

Fig. 2. Time-series of temperature and the concentration and the d13_{C value of POC in the upper 50 m at the SEATS Station observed on}

the ﬁve cruises of this study. The samples were taken every 10 m starting at 10 m below surface. The crosses mark the mean values. The mean d13C values are the concentration-weighted averages.

(7)

in the mixed layer except a surface maximum of POC concentration (Fig. 4B, C). Below the mixed layer, the trends were similar to those found in summer. It is noted that the isotopic composition of POC in winter showed a narrower range from 23.2% to 22.3%, while that of DIC showed a slightly narrower range from 0.1% to 0.8%. The decreasing trend of the d13CDICvalue with depth is

consistent with previous observations in the SCS (Lin et al., 1999), but the range (0.1–0.9%) of all observed values in the top 200 m differs from the previous one (0.5–1.6%) with the current values lower by about 0.5% on average. The shift could be attributed to the 10 years’ difference between the two observations due to the Suess effect, i.e., the intrusion of the isotopically light CO2 from fossil

fuel burning into the ocean (Quay et al., 2003). However, a systematic difference in the isotopic

measurements cannot be ruled out, since certiﬁed reference materials for quality control of d13CDIC

measurements are not yet available. It is noted that an inter-laboratory calibration of d13CDIC

measure-ments was conducted during the course of this study. No systematic error was found between the values obtained in the laboratory supporting this study and the one at University of California, Davis (Chou, 2004).

The vertical trends of POC variation on all ﬁve cruises are summarized inFig. 5. The concentration of POC decreased with depth with mean concentra-tions of 2.3 mM for the top 20 m, 1.5 mM for 21–100 m and 0.8 mM for 101–200 m. The d13C value of POC also decreased with depth with concentration weighted mean values of 22.1% for the top 20 m, 23.0% for 21–100 m and 24.0% for 100–200 m.

ARTICLE IN PRESS

Fig. 3. Hydrography and selected carbon chemistry profiles observed at the SEATS Station on Cruise OR1-726 in August 2004. (A) Temperature and salinity profiles. (B) Profiles of POC concentration and its d13_{C value. (C) Profiles of aqueous CO}

2and the d13C

value of dissolved inorganic carbon.

(8)

3.2. Organic carbon isotopic compositions of sinking particles

The organic carbon isotopic compositions of sinking particles collected in the sediment traps deployed at four depths at Site M1S are plotted in

Fig. 6and those for Site M2S are plotted inFig. 7. Also plotted are the corresponding organic carbon contents and the C/N ratio by atom in the same samples.

The sediment trap mooring at Site M1S was deployed from September 22, 2001 to March 20, 2002 (Table 2). The d13C value varied in the range from 23.2% to 21.0% (Fig. 6). There was no consistent trend in the temporal variations recorded by the four traps. In terms of vertical variation, top two traps recorded a stronger variability, while the bottom two traps showed more uniform values. The organic carbon contents of the sediment trap

samples decreased consistently with depth (Fig. 6) from 10% to 22% at 374 m to 2–3% at 2700 m. There appeared to be a weak temporal trend of a minimum occurring in winter. The C/N ratio varied in the range from 5.5 to 10.7 with most variability occurring in the top two traps; the bottom two traps showed quite uniform values from 7.7 to 9.4.

The sediment trap mooring at Site M2S was deployed from December 16, 2001 to May 14, 2002 (Table 2). Compared to samples from Site M1S, those from M2S had d13C values varying in a wider range from 25.0% to 21.5% (Fig. 7). Similarly most of the variability occurred in samples from the shallowest trap at 447 m. There was no consistent trend in the temporal variations recorded by the three traps. The organic carbon contents of the sediment trap samples also decreased consistently with depth (Fig. 6) from 29–48% at 447 m to 4–6% at 3250 m. Although the shallowest trap at M2S was

ARTICLE IN PRESS

Fig. 4. The same asFig. 2except for OR1-743 in January 2005.

(9)

deeper than that at M1S, its trapped samples had considerably higher organic carbon contents. The range of C/N ratios recorded in the shallowest trap was also wider with a lower bound of 5.9, similar to that recorded at M1S, but a higher upper bound of 18, which corresponded to a rather low d13C value of 24.9% (Fig. 7), suggesting a predominantly terrestrial origin. The bottom two traps showed lesser variability in the C/N ratio with time.

4. Discussion

4.1. Origin of POM

The origin of POM may be indicated by the C/N ratio and the carbon isotopic composition (Jeffrey et al., 1983; Hayes, 1993; Middelburg and Nieu-wenhuize, 1998). The concentrations of POC and

particulate nitrogen as SPM collected at the SEATS Station are highly correlated (Fig. 8). The equation of linear regression is as follows:

POC ¼ ð6:36 0:16Þ PN þ 0:09 0:04; R2_¼_0:94.

(2) The slope of the regression line is quite close to the Redﬁeld C/N ratio of 6.63, suggesting that most of the variation of POM in the upper water column was attributable to marine organic matter.

In contrast to the very coherent relationship between POC and PN, the relationship between organic carbon and nitrogen contents in sinking particles collected by sediment traps is more scattered (Fig. 9). Even so, the data points are scattered about the line of Redﬁeld ratio with obvious deviation towards the higher C/N ratio, indicating contribution of terrigenous materials.

ARTICLE IN PRESS

Fig. 5. Histogram of occurrences of concentrations and the d13C values of POC in the upper 50, 51–100 and 101–200 m observed at the SEATS Station on the ﬁve cruises. The arrows indicate the mean values of concentration and the concentration weighted mean d13C values of POC.

(10)

A simple mixing model based on C/N ratio and d13C of the POM is used to explore the contribution of terrigenous organic matter in sediment trap samples. It is assumed that the organic matter in sediment trap samples is a mixture of marine and terrigenous organic matter, which have distinctive C/N ratios, Rm and Rt, respectively, and d13C

values, namely, dm and dt, respectively. Then both

the C/N ratio (R) and the d13C value (d) of the mixture may be expressed as a function of the fraction of terrigenous organic matter (f) as follows: R ¼ Rtf þ ð1 f ÞRm, (3)

d ¼ ðfRtdtþ ð1 f ÞRmdmÞ=R. (4)

It is noted that the fraction of terrigenous organic matter is based on nitrogen contents so that it matches the C/N ratio to make the calculation more straightforward.

The modeled mixing curves are shown in

Fig. 10A. For the model calculation, the terrigenous end-member is assumed to have a uniform composi-tion because of the relatively long distances of the two mooring sites from land, which may make terrigenous materials from different sources well

mixed during transportation. For the marine end-member, both the d13C value and the C/N ratio are assumed to ﬂuctuate within a range from a representative value. The compositions of the end-members are so chosen that the mixing curves encompass most of the data points. The terrigenous end-member is assumed to have a C/N ratio of 22 and d13C value of 25.5%. The C/N ratio of marine end-member is assumed to be 6.6371 and the d13

C value is assumed to be 22.171.1%. The inferred ﬁeld of marine end-member composition coincides with the compositions of suspended POM to have the highest d13C values (Fig. 10B). It is worth noting that the minimum C/N ratio of the inferred marine end-member also coincides with the minimum of observed C/N ratios of the suspended POM (Table 3,Fig. 10B).

The distribution of data points (Fig. 10A) shows characteristic differences between the two sites. Samples from Site M2S appeared to contain more terrigenous organic matter as indicated by the higher C/N ratios and lower d13C values. Samples from the deeper traps were more uniform in composition, but the same bias still existed. The site differences could be attributed to the circulation

ARTICLE IN PRESS

Fig. 6. Time-series of the d13_{C value, the organic carbon content and the C/N ratio of the decarbonated materials of samples collected by}

sediment traps deployed at depths of 374, 925, 1925 and 2700 m at Site M1S (Fig. 1) from September 22, 2001 to March 20, 2002. The sampling interval was 15 days per cup.

(11)

pattern, which may have controlled the dispersion of re-suspended sediments. The higher C/N ratios, implying stronger terrigenous signals, all occurred in winter, namely, from late December 2001 to early March 2002 (Fig. 7), coinciding with the trap deployment at both sites. The strong northeast monsoon (Liu et al., 2002) might have induced sediment re-suspension on the continental shelf northwest of the mooring sites of sediment traps and sustained seaward export of ﬁne-grained sedi-ments rich in terrigenous organic matter. However, the lack of the terrestrial signal in the trap samples from Site M1S, which was near the Luzon Strait, could be due to the intrusion of the Kuroshio (Wu and Chiang, 2007). In spite of the characteristic distinction of samples from the two sites, all data points follow the same mixing trend well (Fig. 10A).

It is also clear from the distribution of sea surface Chl-a (Fig. 1B) that the two sites belong to the same biogeochemical regime (Fig. 1B). Therefore, we believe that the trap samples are representative of biogeochemical characteristics in the northern SCS proper and may reﬂect the biogeochemical processes occurring at the SEATS Station.

Applying the same mixing model to the composi-tion of the suspended POM, we found that mixing of the terrigenous end-member with the inferred compositions of marine end-members can explain the compositional variations of only part of the samples. More than half of the POM data points cannot be accounted for by the mixing curves constructed for the trap samples. It is necessary to extend the lower limit of the d13C value of the marine end-member to as low as 26% to account

ARTICLE IN PRESS

Fig. 7. The same asFig. 6except for traps deployed at 447, 1248 and 3250 m at Site M2S (Fig. 1) from December 16, 2001 to May 14, 2002.

(12)

for the entire observed compositional ﬁeld. Addi-tional controlling factors other than contribution of terrigenous organic matter are needed to explain the carbon isotopic variation.

4.2. Control of carbon isotopic composition in phytoplankton

The processes that control the carbon isotopic composition of POM produced by phytoplankton

ARTICLE IN PRESS

Fig. 8. Scatter plot of POC concentration vs. PN concentration obtained on the ﬁve cruises listed inTable 1. The line indicates the result of linear regression.

Fig. 9. Scatter plot of organic carbon content vs. nitrogen content in sediment trap samples. The line is a reference line indicating the Redﬁeld ratio by weight.

Fig. 10. (A) Scatter plot of the d13_{C value of organic carbon in}

trap samples vs. C/N ratio. The curves represent mixing between a terrigenous end-member (d13C ¼ 25.5% and C/N ¼ 22) and marine end-members of different compositions, which are listed inTable 3for different cases (see text). (B) The same as (A) except for the SPM samples.

Table 3

Compositions of the marine end-member used in the mixing model (see text)

Case C/N d13C (%) A 6.63 21.5 B 5.63 21.2 C 5.63 22.5 D 7.63 21.5 E 5.63 24.0 F 6.63 26.0

(13)

in the surface water have been extensively discussed (Rau et al., 1982, 1996; Hofmann et al., 2000). The fractionation between the carbon dioxide dissolved in the ambient aqueous medium, termed as aqueous CO2, and the product of photosynthesis

is expressed as

aPaq ¼ ð1 þ dP=1000Þ=ð1 þ daq=1000Þ, (5)

where aP is the isotope fractionation factor, daq is

the d13C value of aqueous CO2, which is usually the

major source of carbon during photosynthesis in aquatic environments, and dP is the d13C value of

phytoplankton. The isotope fractionation factor is often replaced by the % fractionation deﬁned as Paq¼1000ðaPaq1Þ. (6)

It is noted that the convention of the isotope fractionation parameters deﬁned in this study follows that of Hofmann et al. (2000), which is different from that used byRau et al. (1996), though the essence is the same.

The % fractionation, eP-aq, has been found to be

dependent mainly on the isotope fractionation during carbon ﬁxation and the diffusive process in the aqueous medium (Rau et al., 1996). The higher the concentration of aqueous CO2 is, the stronger

the fractionation gets. A simpliﬁed formulation is derived from previously proposed formulas as follows (Rau et al., 2001):

Paq¼PCO2 fðPCO2þdÞmi=½CO2ðaqÞgðg=SÞ=P,

(7) where PCO2 is the % fractionation during

enzy-matic carbon ﬁxation, ed is the diffusive %

fractionation of CO2in seawater, miis the

instanta-neous speciﬁc growth rate (s1), [CO2(aq)] is the

ambient concentration of aqueous CO2, g is the

carbon content per cell (mol), S is the surface area of the phytoplankton cell (m2), and P is the cell wall permeability (m s1). The instantaneous speciﬁc growth rate is related to the 24 h speciﬁc growth rate (mphy) by the following relationship:

mi¼ ðL þ DÞmphy=L, (8)

where L and D are the lengths of light and dark periods in a day. The length of the light period from sunrise to sunset at the SEATS Station varies from 11.07 to 13.13 h with an average of 12.0570.8 h. For practical reason, we assume the ratio (L+D)/L to be 2.0. The carbon content per cell (g, in mol C) is

related to the cell volume (V, in mm3) by the following relationship:

g ¼ 3:154 1014V0:758_. ₍₉₎

The surface area of the phytoplankton cell is calculated from the radius (r) of the cell, which is assumed to be spherical:

S ¼ 4pr2. (10)

Because it is difﬁculty to determine the dCO2value

directly, it may be calculated from dDIC, the d13C

value of DIC, and the isotope fractionation factor between the aqueous CO2 and DIC. The isotope

fractionation is deﬁned as

aaqDIC¼ ð1 þ daq=1000Þ=ð1 þ dDIC=1000Þ. (11)

The % fractionation, which is deﬁned as 1000

(aaqDIC–1), was found to be a function of

temperature and the second ionization fraction of carbonic acid (Zhang et al., 1995; Hofmann et al., 2000).

aqDIC¼ ð0:0144½CO23 =½DIC 0:1119ÞT 11:84,

(12) where T is temperature (1C).

The isotopic composition of the POC produced by phytoplankton may be calculated from that of the DIC and the overall % fractionation.

dP¼ f½ð1 þ dDIC=1000ÞaPaqaaqDIC 1g1000

dDICþPaqþaqDIC. ð13Þ

Therefore, the d13C value of POC may be calculated by Eq. (13), which in turn needs input from Eqs. (7) and (12). The overall calculation depends on values of dDIC, temperature (T), aqueous CO2

concentra-tion ([CO2(aq)]) and the second ionization fraction

of carbonic acid ([CO32]/[DIC]) and a suite of

variables listed inTable 4.

4.3. Temporal variation of d13CPOCin surface water

Following the previous approach to explore isotopic variation of POC produced by phytoplank-ton in surface seawaters (Rau et al., 1997), we used Eq. (13) to calculate the d13C value of POC in the surface water at SEATS under the hydrographic and chemical conditions observed on the ﬁve cruises. The purpose is to estimate the parameters of various processes controlling the isotopic com-position of POC and to determine the major factors contributing to the observed isotopic variation of POC.

ARTICLE IN PRESS

(14)

In order to account for the co-variation of ambient conditions in the surface water, which is deﬁned as the top 20 m, we constructed time-series of temperature, [CO2]aq, dDICand [CO32]/[DIC] by

ﬁtting polynomials to the observed data (Fig. 11). Using the time-series data, we calculated the d13CPOC during the study period from May 2004

to March 2005 with different combination of parameter values for the calculation (Fig. 12). The base values and ranges of various parameters are listed inTable 4. The base value of % fractionation during enzymatic carbon ﬁxation (PCO2) was set to

be 23% followingHofmann et al. (2000). The base value of the cell wall permeability of phytoplankton (P) was set to 5 105ms1 following Rau et al. (1997). The base value of cell radius was set to 1.5 mm, which is considerably smaller than those used for the average oceanic conditions (5–20 mm), because phytoplankters smaller than 3 mm ac-counted for 77.6–90.1% of total Chl-a concentra-tion in the basin water of the northern SCS (Chen, 2005). The base value of mphy was set to 0.95 d1,

which is very close to the value of 1 d1used byRau et al. (1997).

The curve of calculated d13CPOC values based on

the base values of the parameters appears to follow the trend of observed values (Fig. 12), but the amplitude of the observed variation is more pronounced than the calculated results. By varying the parameter values, one at a time, we were able to produce curves bracketing the observed values. The speciﬁc growth rate (mphy) ranging from 0.8 to

1.1 d1. In order to compare the range with observations, we plotted values of primary produc-tion vs. Chl-a concentraproduc-tion in the top 20 m (Fig. 13) observed at a station (18.0N, 115.5E) very close to the SEATS Station (Chen, 2005). The

speciﬁc growth rate (mphy) may be calculated from

the following relationship: mphy¼PP=½POCphy

¼PP=f½ChlaðPOC=ChlaÞ_phyg, ð14Þ where PP is primary production in units of mg C m3d1, [POCphy] is the phytoplankton

bio-mass in units of mg m3, and (POC/Chl-a)phy is

the mean organic carbon to Chl-a ratio of phyto-plankton. Using the (POC/Chl-a)phy ratio of

50 mg C mg1 Chl-a (Sarmiento et al., 1993), we calculated the speciﬁc growth rates from the data plotted inFig. 13. The histogram of the calculated mphy values is shown inFig. 13B. The range of the

speciﬁc growth rate (0.8–1.1 d1) that produced d13CPOC values bracketing the observed values

coincides with the most frequently occurring ob-served values, lending support to the derived mphy

values.

Aside from the speciﬁc growth rate, other para-meters also might contribute to the observed iso-topic variability. It is demonstrated in Fig. 12B–D

that similar curves bracketing the observed values may be produced by varying the other para-meters. The sensitivities of the calculated d13CPOC

value to the variation of the four parameters (Table 4) are similar to those found byRau et al. (1997).

The sensitivity of the calculated d13CPOCvalue to

the variation of the concentration of aqueous CO2

([CO2]aq) is explored by constructing a d13CPOC

curve corresponding to [CO2]aq in the range

observed in the surface water at the SEATS Station with all parameters ﬁxed at the base value and other variables ﬁxed at the mean values in the top 20 m (Fig. 14). The constructed curve appears to follow

ARTICLE IN PRESS

Table 4

Values of parameters used for the calculation of d13_{C values of POM produced by phytoplankton (see text)}

Parameter Description Value Range Units Sensitivity

(this study)

Sensitivity (Rau et al., 1997) mphy Twenty-four hour speciﬁc growth

rate of phytoplankton

0.95 0.8–1.1 d1 8% d 7% d

PCO2 Per-mil fractionation during

enzymatic carbon ﬁxation

23 25 to 21 0.6% 0.7%

ed Diffusive isotope % fractionation

of CO2in seawater

0.7 0.7

r Mean cell radius of

phytoplankton

1.5 0.85–2.6 mm 1.4% mm 1.1% mm

P Cell wall permeability of

phytoplankton 5 105 4.3 105 –6 105 m s1 1.5 105 % m1 s 1.3 105 % m1 s

(15)

the trend manifested by the observed values except those from August 2004. The trend gives a sensitivity of 0.75% mM1, which is slightly higher than that (0.6% mM1) reported by Rau et al. (1997). It is clear that the variation in [CO2]aq

contributes only about 0.5% to the overall varia-bility of d13CPOC.

4.4. Vertical variation of d13CPOCin the upper water

column

The vertical isotopic variation of POC in the SCS from the surface water to the surﬁcial sediments is illustrated in Fig. 15. The accompanying distribu-tions of the C/N ratio of the suspended and the

ARTICLE IN PRESS

Fig. 11. Time-series of (A) temperature, (B) [CO2]aq, (C) d13CDICand (D) the second ionization fraction of carbonic acid in the top 20 m

observed at the SEATS Station on the five cruises. The curves are polynomial fits of the data. The fit for d13CDICis based on the time-series

ﬁt of temperature and the temperature-d13CDICrelationship (not shown), which has a R2value of 0.68, for optimization.

(16)

sinking POM in the northern SCS are shown in

Fig. 16. The suspended POC in the upper 200 m showed a decreasing trend of the d13C value. The concentration-weighted mean d13C value decreased from 22.1% in the top 20 m to 23.0% in the 21–100 m layer to 24.0% in the 101–200 m layer. As shown inFig. 16, the ranges of the C/N ratio is quite narrow in the top 100 m with the concentra-tion weighted mean values, 6.2 for 0–21 m and 6.5 for 21–100 m, very close to the Redﬁeld ratio of 6.67. Deeper than 100 m the C/N ratio tends to increase, suggesting the presence of terrigenous POC, but the weighted mean, 7.3 for 101–200 m, is only slightly higher than the Redﬁeld ratio, indicat-ing limited terrestrial contribution. We have

demon-strated earlier (seeFig. 10) that the decreasing trend of d13CPOC with depth cannot be satisfactorily

explained by contribution of terrigenous POC. What has caused the vertical variation of isotopic composition of suspended POC in the upper water column is worth further discussion.

In the subsurface water from 21 to 200 m, some of the observed d13CPOC values were signiﬁcantly

lower than those observed in the surface water. It has been reported before that d13CPOC proﬁles

usually show a minimum in the pycnocline (Jeffrey et al., 1983). The depths of the pycnocline at the SEATS Station were in the range of 50–150 m, as exempliﬁed by the hydrographic proﬁles inFigs. 2 and 3, where the d13CPOC values were indeed the

ARTICLE IN PRESS

Fig. 12. Comparison between observed data and modeled time-series curves of the d13_C

POCvalue from the top 20 m at the SEATS Station

on the ﬁve cruises. (A) The solid curve is the modeled results calculated from time-series of variables shown inFig. 11and base values of parameters listed inTable 4. The dashed curves are the same as the solid curves except for different mphyvalues. (B) The same as (A) except

for different PCO2values. (C) The same as (A) except for different phytoplankton radii. (D) The same as (A) except for different cell

membrane permeability.

(17)

lowest. The depletion of 13C is attributed to biological reworking of POC in the pycnocline (Jeffrey et al., 1983), where the settling of the particles may slow down due to increasing density of the seawater and respiration may be enhanced. The trend appears to agree with the diagenetic isotopic change observed for the residual of aquatic vascular plants, namely, Spartina grass, which showed relatively large decreases, up to 4%, of d13Corg values during degradation (Benner et al., 1987). The isotopic alteration is attributed to preferential preservation of lignin, which is isotopi-cally light, in the plant detritus. It also agrees with results of incubation experiments on lacustrine biomass (Lehmann et al., 2002), which showed a decrease of d13CPOCup to 1.5% during degradation.

However, none of these experiments were conducted for algal materials from the open ocean. In other words, direct evidence of carbon isotopic shift toward lighter values during degradation of bio-genic POM in the water column of the open ocean is lacking. Therefore, there could be other explanation for the decreasing trend of d13CPOCobserved in the

water column at SEATS and other oceanic regions. It is proposed here that the changing isotope fractionation during production of POC at different depths, instead of biological degradation, may have contributed to the decreasing trend of d13CPOC

in the subsurface water column. Because of the

ARTICLE IN PRESS

Fig. 13. (A) Scatter plot of primary production vs. Chl-a concentration observed at a station very close to the SEATS Station (Chen, 2005). These data allow us to calculate the specific growth rates (see text). The line indicates the specific growth rate of 1 d1. (B) The histogram of the specific growth rates calculated form the data shown in (A).

Fig. 14. The scatter plot of d13_C

POCvs. [CO2]aqobserved at the SEATS Station on the ﬁve cruises. The line indicates the modeled results

calculated by Eq. (13) using the mean values of all variables in the surface layer except the [CO2]aqvalue, which varied in the plotted range,

and the base values of the parameters listed inTable 4.

(18)

oligotrophic condition in the SCS (Liu et al., 2002), the euphotic zone is pretty thick. For the euphotic zone depth deﬁned as the depth at which the photosynthetically active radiance is reduced to 0.6% of the surface value, it was found in the

northern SCS basin to be about 100 m in winter (Chen et al., 2004). It is also known that the Chl-a concentration in the SCS in winter is generally higher than in other seasons (Liu et al., 2002;Tseng et al., 2005; Chen, 2005). Therefore, the euphotic

ARTICLE IN PRESS

Fig. 15. Histogram of the d13_{C values of organic carbon in different types of samples from different depths in the SCS. The solid arrows}

indicate the concentration-weighted mean values (see text).

(19)

zone depths in other seasons should be deeper than 100 m. The speciﬁc growth rates in the top 100 m, which were calculated in the same manner described earlier, decreased rapidly with depth below the top 20 m (Fig. 17A). Its depth dependency may be

expressed as the following equation: m_phy¼3:0 exp½0:07ðz 2Þ

f1 exp½0:08ðz þ 4Þg þ 0:1, ð15Þ where mphyis in units of d1and z is depth (m).

ARTICLE IN PRESS

Fig. 16. Same asFig. 15except for C/N ratio. It is noted that sedimentary samples are not included due to lack of data. K.-K. Liu et al. / Deep-Sea Research II 54 (2007) 1504–1527

(20)

Assuming the same mechanism represented by Eq. (13) also governed the isotopic fractionation in the subsurface water column, we calculated the d13CPOC proﬁle in the top 100 m (Fig. 17B). For

the calculation, we employed Eq. (15) to estimate the speciﬁc growth rate, adopting the base values listed inTable 4for all other parameters and using mean values of variables in the surface layer for all depths. In other words, the calculated results reﬂect only the effect of changing mphy on the

isotopic composition of POC produced by phyto-plankton. It is clear that very light POC may be produced in the subsurface water column due to decreasing mphy. In fact, the calculated d

13

CPOC

values in the subsurface layer are much lower than the observed values, suggesting that the subsurface POC was a mixture of in situ produced POC and the POC produced in the surface water, where primary productivity is the highest (Chen, 2005). The decreasing trend of d13CPOCin the subsurface water

could be mainly due to increasing contribution of isotopically light POC produced in the lower euphotic zone.

4.5. Relationship between suspended and sinking POM

To extend our discussion further down the water column, we explore the isotopic variation of the sinking and sedimentary organic matter with respect to that of the suspended POM. The sedimentary records from the SCS were reported by Kienast et al. (2001). Though the sampling localities of the ﬂuff materials and the surface sediments distribute in a large area of the SCS (Fig. 1B), they all fall within the region of rather low average sea-surface Chl-a level, which also includes the SEATS Station and the two sediment trap sites. In light of the widely applied biogeochemical characterization based on ocean color (e.g., Longhurst et al., 1995), we have good reason to believe that the average biogeochemical conditions of these sites are similar and the comparison of the isotopic compositions of organic carbon from these sites is meaningful.

Contrary to the decreasing trend with depth in the top 200 m, the d13C values of particles collected in sediment traps deployed at depths deeper than

ARTICLE IN PRESS

Fig. 17. (A) Calculated speciﬁc growth rates from previously reported data of primary production and Chl-a concentration at a station very close to the SEATS Station (Chen, 2005). The curve is the ﬁt of the data. (B) Observed and modeled d13_C

POCvalues in the top 100 m

at SEATS (see text). The curve is the modeled results, which were based on the speciﬁc growth rates represented by the ﬁtted curve in (A). K.-K. Liu et al. / Deep-Sea Research II 54 (2007) 1504–1527 1523

(21)

200 m became higher than those values observed in the subsurface water (Fig. 15). On the other hand, the trend of increasing C/N ratio in the top 200 m continued in the deeper water as recorded by the sediment traps reaching a maximum in the depth range of 374–1248 m (Fig. 16). As demonstrated earlier (Figs. 7 and 10A), the trap samples, especially those from the intermediate depths of 447–1248 m at Site M2S, contained a significant amount of terrigenous organic matter. The signifi-cant contribution of terrigenous organic matter in the intermediate depths is consistent with the notion of seaward export of re-suspended materials from the continental shelf. Intermediate nepheloid layers between 600 and 1440 m have been observed in the Bay of Biscay during the OMEX project (McCave et al., 2001). In spite of the terrigenous contribution, we have successfully derived the carbon isotopic composition of the marine end-member for the organics in sediment trap samples (Fig. 10A). Its inferred range of the d13C value, 22.5% to 21.3%, overlaps significantly with the range observed in the surface layer (0–20 m) and the mid value (22.1%) of the inferred range coincides with the concentration-weighted mean of d13C value in the surface layer (Fig. 15). This indicates that the sinking POC is mainly contributed by the highly productive surface layer.

It has been shown that the organic carbon in fluff samples above seafloor in the SCS (Kienast et al., 2001) is more enriched in 13C with respect to the suspended POC in the surface water and the inferred marine end-members in sinking particles (Fig. 15); moreover, the surface sediments in turn have higher d13Corg than the fluff samples. Such

progressive enrichment in 13C may be attributed to the preferential removal of 12C during de-gradation of POM (Westerhausen et al., 1993). However, observations in the Cariaco Basin did not show signiﬁcant changes in the d13CPOC values

between sediment traps deployed at different depths, suggesting insigniﬁcant isotopic fractiona-tion during diagenesis in the water column ( Wood-worth et al., 2004). On the other hand, it has been proposed that the lower d13CPOC values in ﬂuff

samples than the sedimentary organics is attributed to the Suess effect, i.e. the invasion of the iso-topically light anthropogenic CO2 into the ocean

(Kienast et al., 2001). The Suess effect observed in the SCS may cause the d13CDICvalue to change by

0.026% for an increase of DIC by 1 mmol kg1 -SW (Chou, 2004;Chou et al., 2007). The increase of

DIC concentration in the surface water of the SCS was estimated to be about 60 mmol kg1-SW on average (Chou, 2004; Chou et al., 2007) and the average change in d13CDIC was estimated to be

about 1.6%.

In addition to the Suess effect, the elevated concentration of aqueous CO2in the surface water

resulting from increasing CO2 partial pressure

may also further depress the d13CPOCin the present

day ocean. For the sake of discussion, we esti-mated concentrations of DIC and carbonate species during pre-industrial era in the SCS surface water corresponding to the ﬁve sampling conditions obtained in this study by assuming the same total alkalinity for each case but a reduced CO2

partial pressure of 280 matm (Petit et al., 1999). The calculation showed that concentrations of aqueous CO2 in the pre-industrial era had been

only 75% of the current values on average and the second ionization fractions of carbonic acid had been about 23% higher than the current values due to higher pH. Then we calculated the d13CPOC

values by using the same values of the variables and parameters for the modern cases except the carbonate species and the d13CDIC values which

were assigned with the aforementioned values. The resultant pre-industrial d13CPOC values had

been higher than the current values by 3.5–4.3%. Qualitatively these estimates support the notion that the isotopic differences between the currently produced POC and the organic carbon in surficial sediments reflect the actual changes in time; quantitatively, the estimated differences are signifi-cantly larger than the observed differences of 2–3% (Fig. 15).

There are several factors that may cause the discrepancy. Firstly, the dissolved CO2in the SCS

surface water was assumed to be equilibrated with the atmospheric CO2 with an average partial

pressure of 280 matm (Petit et al., 1999), but the actual situation could have been slightly super-saturated due to the input of terrigenous organic carbon. Secondly, the sea surface temperatures were assumed to be the same as the present time, but the pre-industrial era, especially from 15 to 19th century, was cooler than today (Folland et al., 2001). According to a previous study (Rau et al., 1997), the temperature has a positive effect on the d13CPOC of 0.4% 1C1. Slight supersaturation of

CO2 and/or lower sea surface temperature may

reduce the estimated differences to the observed level.

ARTICLE IN PRESS

(22)

5. Summary and conclusions

On the ﬁve cruises from May 2004 to March 2005 to the SEATS Station, the observed d13C value of suspended POC varied from 24.2% to 21.3% in the top 20 m, from 25.1% to 21.5% in the 21–100 m layer, and from 25.9% to 23.2% in the 101–200 m layer. The organic carbon in sediment trap samples collected from September 2001 to May 2002 in northern SCS had d13C values ranging from 25.0% to 21.0% for traps deployed between 374 and 1248 m and from 23.7% to 21.9% for traps deployed between 1925 and 3250 m. The C/N ratios of sediment trap samples reached as high as 18, indicating contribution of terrigenous organic mat-ter, which occurred mainly in the depth range of 450–1300 m. A mixing model depicting the relation-ship between the d13C value and C/N ratio of POM revealed that the organic matter in sediment trap samples were mixtures of terrigenous organics with d13C value of 25.5% and C/N ratio of 22 and marine organics with d13C values from 22.5% to 21.3% and C/N ratios from 5.6 to 7.6. The same mixing model revealed that the suspended POM from the SEATS Station contained minor amount of terrigenous organic matter, and the observed range of d13CPOCvalues reﬂected the actual isotopic

variability of marine organic matter.

Simulation of the carbon isotopic composition of phytoplankton in the surface water based on observed values of hydrographic and chemical variables may match the observed range of d13CPOC

values in the top 20 m, while variations in the biological parameters, including the specific growth rate, enzymatic isotope fractionation during carbon fixation, cell size and cell wall permeability, within the normal ranges may have contributed signifi-cantly to the observed isotope variability. If the same processes controlling the carbon isotopic composition of phytoplankton in the surface water worked in the entire euphotic zone of about 100 m thick, isotopically very light POC with d13CPOC as

low as 28% could have been produced in the subsurface layer due to the exponentially decreasing speciﬁc growth rate. Therefore, the observed de-creasing trend of d13CPOC below the surface layer

down to 200 m could have resulted from addition of the isotopically very light POC produced in the subsurface water to the isotopically normal POC produced in the surface layer.

The marine component in the sinking POM collected in sediment traps had a mid d13C value

(22.1%) coinciding with the concentration weighted mean d13CPOC value in the top 20 m,

suggesting the surface layer as the dominant source of sinking POC. However, the inferred d13C values of the marine end-member of the sinking POM are significantly lower than those (22.8% to 20.1%) of the POM in the nepheloid layer above seafloor in the SCS, which are in turn lower than those (21.3% to 19.0%) of organic matter in surficial sediments from the SCS. It cannot be excluded that preferential removal of 12C during degradation of POM accounts for the progressive enrichment of

13

C during settling and deposition of organic matter. However, according to a calculation, the relative depletion of13C in the more recent samples, namely the POM in the surface water and the nepheloid layer, with respect to the organic carbon in surﬁcial sediments is attributable to two factors: (1) the drop of 1.6% in the isotopic composition of DIC attributed to the Suess effect, and (2) the increasing concentration of aqueous CO2and other

related changes in carbonate species that favor isotopic fractionation against incorporation of 13C during carbon ﬁxation. The overall effect is more than enough to account for the observed isotopic differences. Therefore, this study supports the notion that the isotopic composition of deposited organic carbon in the SCS can faithfully reﬂect the CO2 chemistry in the surface water, but

consider-able isotopic variation may be induced by changes in hydrographic and biological conditions. In other words, isotope fractionation during degradation of POM may not be ruled out as a possible mechanism for the observed vertical variation in the d13CPOC

value, but its importance in controlling the d13C value of deposited organic carbon is not warranted. More studies focused on the production of POM in the surface water and relationship between the freshly produced POM in the upper water column and the sinking/deposited organics are warranted for a better understanding of the factors controlling the isotopic composition of the sedimentary organic carbon.

Acknowledgments

This study was supported by the Grant NSC 94-2611-M-008-002 from the National Science Council of the Republic of China. We thank the captains and crew of the R/V Ocean Researcher I and R/V FR-1 for sampling assistance. We thank the Sediment Trap Laboratory (NCOR) for providing

ARTICLE IN PRESS

(23)

trap samples obtained under the Grant NSC 90-2119-M-110-003. This is NCU-IHS Contribution #53.

References

Benner, R., Fogel, M.L., Hodson, R.E., Sprague, E.K., 1987. Depletion of C-13 in lignin and its implications for stable carbon isotope studies. Nature 329, 708–710.

Chao, S.Y., Shaw, P.T., Wu, S.Y., 1996. Deep-water ventilation in the South China Sea. Deep-Sea Research I 43, 445–466.

Chen, M.T., Beaufort, L., Shipboard Scientiﬁc Party of the IMAGES III/M 106-IPHIS Cruise (Leg_II), 1998. Exploring Quaternary variability of the East Asia monsoon, Kuroshio Current, and Western Paciﬁc warm pool systems: high-resolution investigations of paleoceanography from the IMAGES III (MD106)-IPHIS Cruise. Terrestrial Atmo-spheric Oceanic Sciences 9, 129–142.

Chen, Y.F.L., 2005. Spatial and seasonal variations of nitrate-based new production and primary production in the South China Sea. Deep-Sea Research I 52 (2), 319–340.

Chen, Y.L.L., Chen, H.Y., Karl, D.M., Takahashi, M., 2004. Nitrogen modulates phytoplankton growth in spring in the South China Sea. Continental Shelf Research 24 (4–5), 527–541.

Chou, W.C., 2004. Seasonal variability of CO2 specis and

d13CTCO2in the mixed-layer and their vertical distributions

at the SEATS site. Ph.D. Dissertation, National Sun Yat-sen University, Kaohsiung, Taiwan, 211pp.

Chou, W.C., Sheu, D.D., Lee, B.S., Tseng, C.M., Chen, C.T.A., Wang, S.L., Wong, G.T.F., 2007. Depth distributions of alkalinity, TCO2 and d13CTCO2 at SEATS time-series site,

northern South China Sea: controlling processes and anthro-pogenic CO2 inﬂuence. Deep-Sea Research II, this issue

[doi:10.1016/j.dsr2.2007.05.002].

DOE, 1994. Handbook of methods for the analysis of the various parameters of the carbon dioxide system in seawater. US Department of Energy CO2Science Team Report, vol. 2, US

Department of Energy.

Ducklow, H.W., 2003. Biogeochemical provinces: towards a JGOFS synthesis (Chapter 1). In: Fasham, M.J.R. (Ed.), Ocean Biogeochemistry: A New Paradigm. Global Change — The IGBP Series. Springer, New York, pp. 3–18.

Folland, C.K., et al., 2001. Observed climate variability and change. In: Houghton, J.T., et al. (Eds.), Climate Change 2001: The Scientiﬁc Basis. Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, UK, p. 881pp.

Hayes, J.M., 1993. Factors controlling C-13 contents of sedimentary organic- compounds-principles and evidence. Marine Geology 113, 111–125.

Higginson, M.J., Maxwell, J.R., Altabet, M.A., 2003. Nitrogen isotope and chlorin paleoproductivity records from the Northern South China Sea: remote vs. local forcing of millennial- and orbital-scale variability. Marine Geology 201 (1–3), 223–250.

Hoefs, J., 1997. Stable Isotope Geochemistry. Springer, Berlin, 201pp.

Hofmann, M., Wolf-Gladrow, D.A., Takahashi, T., Sutherland, S.C., Six, K.D., Maierreimer, E., 2000. Stable carbon-isotope distribution of particulate organic-matter in the ocean-A model study. Marine Chemistry 72, 131–150.

Jeffrey, A.W.A., Brooks, J.M., Pﬂaum, R.C., Sackett, W.M., 1983. Vertical trends in particulate organic-carbon C-13-C-12 ratios in the upper water column. Deep-Sea Research 30, 971–983.

Karl, D.M., et al., 2003. Temporal studies of biogeochemical processes in the world’s oceans during the JGOFS era. In: Fasham, M.J.R. (Ed.), Ocean Biogeochemistry: The Role of the Ocean Carbon Cycle in Global Change. IGBP Book Series. Springer, Berlin, pp. 239–267.

Kienast, M., 2000. Unchanged nitrogen isotopic composition of organic-matter in the South-China-Sea during the last climatic cycle-global implications. Paleoceanography 15, 244–253.

Kienast, M., Calvert, S.E., Pelejero, C., Grimalt, J.O., 2001. A critical review of marine sedimentary delta C-13(org)-pCO(2)

estimates: new palaeorecords from the South China Sea and a revisit of other low-latitude d13_C

(org)-pCO(2)records. Global

Biogeochemical Cycles 15 (1), 113–127.

Lehmann, M.F., Bernasconi, S.M., Barbieri, A., McKenzie, J.A., 2002. Preservation of organic matter and alteration of its carbon and nitrogen isotope composition during simulated and in situ early sedimentary diagenesis. Geochimica et Cosmochimica Acta 66 (20), 3573–3584.

Lewis, E., Wallace, D., 1998. Program Developed for CO2

System Calculation. Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, Oak Ridge, Tennessee, USA, 31pp.

Lin, H.L., Wang, L.W., Wang, C.H., Gong, G.C., 1999. Vertical-distribution of d13_{C of dissolved inorganic carbon in the}

northeastern South China Sea. Deep-Sea Research I 46 (5), 757–775.

Liu, K.K., Chao, S.Y., Shaw, P.T., Gong, G.C., Chen, C.C., Tang, T.Y., 2002. Monsoon-forced chlorophyll distribution and primary production in the South China Sea: observations and a numerical study. Deep-Sea Research I 49 (8), 1387–1412.

Longhurst, A., Sathyendranath, S., Platt, T., Caverhill, C., 1995. An estimate of global primary production in the ocean from satellite radiometer data. Journal of Plankton Research 17 (6), 1245–1271.

McCave, I.N., et al., 2001. Distribution, composition and ﬂux of particulate material over the European margin at 471–501N. Deep-Sea Research II 48, 3107–3139.

Middelburg, J.J., Nieuwenhuize, J., 1998. Carbon and nitrogen stable isotopes in suspended matter and sedi-ments from the Schelde Estuary. Marine Chemistry 60, 217–225.

NGDC, 2001. ETOPO2 Global 2-Minute Gridded Elevation Data Volume E1. National Geophysical Data Center, National Oceanographic and Atmospheric Agency, USA. Petit, J.R., et al., 1999. Climate and atmospheric history of the

past 420,000 years from the Vostok ice core, Antarctica. Nature 399, 429–436.

Prell, W.L., Wang, P., Blum, P., 1999. Ocean Drilling Program Leg 184 Preliminary Report: South China Sea. Texas A&M University, College Station, TX.

Quay, P.D., Sonnerup, R., Westby, T., Stutsman, J., McNichol, A., 2003. Changes of the 13_C/12_{C of dissolved inorganic}

ARTICLE IN PRESS

(24)

carbon in the ocean as a tracer of anthropogenic CO2uptake.

Global Biogeochemical Cycles 17 (10), 1029.

Rau, G.H., Kaplan, I.R., Sweeney, R.E., 1982. Plankton13_C–12_C

ratio changes with latitude-differences between northern and southern oceans. Deep-Sea Research 29, 1035–1039. Rau, G.H., Riebesell, U., Wolf-Gladrow, D., 1996. A model of

photosynthetic 13_{C fractionation by marine-phytoplankton}

based on diffusive molecular CO2 uptake. Marine Ecology

Progress Series 133, 275–285.

Rau, G.H., Riebesell, U., Wolf-Gladrow, D., 1997. CO2aq

-dependent photosynthetic 13C fractionation in the ocean-A model versus measurements. Global Biogeochemical Cycles 11 (2), 267–278.

Rau, G.H., Chavez, F.P., Friederich, G.E., 2001. Plankton

13

C/12C variations in Monterey Bay, California: evidence of non-diffusive inorganic carbon uptake by phytoplankton in an upwelling environment. Deep-Sea Research I 48 (1), 79–94. Rau, G.H., Takahashi, T., Des Marais, D.J., 1989. Latitudinal variations in plankton d13C: implications for CO2 and

productivity in past oceans. Nature 341, 516–518.

Sarmiento, J.L., Slater, R.D., Fasham, M.J.R., Ducklow, H.W., Toggweiler, J.R., Evans, G.T., 1993. A seasonal three-dimen-sional ecosystem model of nitrogen cycling in the North Atlantic euphotic zone. Global Biogeochemical Cycles 7, 417–450. Shaw, P.T., Chao, S.Y., 1994. Surface circulation in the South

China Sea. Deep-Sea Research I 41, 1663–1683.

Sverdrup, H.U., Johnson, M.W., Fleming, R.H., 1942. The Oceans. Prentice Hall, New York, 1087pp.

Tseng, C.M., Wong, G.T.F., Lin II, Wu, C.R., Liu, K.K., 2005. A unique seasonal pattern in phytoplankton biomass in low-latitude waters in the South China Sea. Geophysical Research Letters 32 (8), L08608.

Wang, P.X., Tian, J., Cheng, X.R., Liu, C.L., Xu, J., 2004. Major Pleistocene stages in a carbon perspective: the South China Sea record and its global comparison. Paleoceanography 19 (4), PA4005.

Westerhausen, L., Poynter, J., Eglinton, G., Erlenkeuser, H., Sarnthein, M., 1993. Marine and terrigenous origin of organic-matter in modern sediments of the Equatorial East Atlantic-The d13C and molecular record. Deep-Sea Research I 40 (5), 1087–1121.

Woodworth, M., Goni, M., Tappa, E., Tedesco, K., Thunell, R., Astor, Y., Varela, R., Diaz-Ramos, J.R., Muller-Karger, F., 2004. Oceanographic controls on the carbon isotopic compo-sitions of sinking particles from the Cariaco Basin. Deep-Sea Research I 51 (12), 1955–1974.

Wu, C.-R., Chiang, T.-L., 2007. Mesoscale eddies in the northern South China Sea. Deep-Sea Research II, this issue [doi:10.1016/j.dsr2.2007.05.008].

Wu, C.R., Shaw, P.T., Chao, S.Y., 1999. Assimilating altimetric data into a South China Sea model. Journal of Geophysical Research 104, 29987–30005.

Zhang, J., Quay, P.D., Wilbur, D.O., 1995. Carbon-isotope fractionation during gas-water exchange and dissolution of CO2. Geochimica Et Cosmochimica Acta 59,

107–114.