Particle-reactive radionuclides ( 234 Th , 210 Pb , 210 Po ) as tracers for the estimation of export production in the South China

Introduction Conclusions References


Introduction
It is known that carbon dioxide sequestration depends, in part, on the magnitude of carbon removal via particle settling from the surface to the deep layer of the ocean.Thus, the quantification of export flux from the euphotic zone is of great importance to the understanding of carbon cycling.The SEATS site, a time-series station regularly visited by Taiwanese oceanographers since 1999, is located at 18 • N 116 • E in the Introduction

Conclusions References
Tables Figures

Back Close
Full central basin of the South China Sea (Wong et al., 2007).Although being in a marginal sea, the site shows open-ocean characteristics with low productivity.The SEATS is an ecosystem dominated by picoplankton with a relatively low primary productivity ranging between 300 mg-C m −2 d −1 in summer to 550 mg-C m −2 d −1 in winter (Chen et al., 2004;Chen and Chen, 2005).The South China Sea has a thin mixed layer of 30-80 m and a constant euphotic depth of 100 m, which make the diffusive flux of nutrients an important mechanism of supporting the primary production in this oligotrophic system (Tseng et al., 2005).Despite the fact that previous estimates of the export flux of particulate organic carbon (POC) at the SEATS in the South China Sea were attempted by various approaches, e.g., biogeochemical modeling (Liu et al., 2002), carbon budgeting (Chou et al., 2006), and new productivity measurements (Chen and Chen, 2005;Chen et al., 2004), the magnitude of POC export from the euphotic zone has never directly been measured.Previously, short-lived particle-reactive radionuclides, 234 Th (t 1/2 = 24.1 d), 210 Pb (t 1/2 = 22.4 a), and 210 Po (t 1/2 = 138 d), have been found to be very useful for the estimation of export production in the ocean.Among the three, 234 Th has been widely used as a powerful tracer for POC and provided estimates of the export production (see review of Buesseler, 1998).There is only a few studies that used 210 Pb to estimate the export production in the ocean, although a strong correlation of 210 Pb removal with the POC flux was previously demonstrated (Moore and Dymond, 1988).On the other hand, due to its high affinity with biological particles (Cherry et al., 1975), 210 Po is also used as a powerful tracer for particulate matter sinking out of surface ocean (Murray et al., 2005;Friedrich and Rutgers van der Loeff, 2002;Verdeny et al., 2009;Kim and Church, 2001;Buesseler et al., 2008).
Here we report the results of POC, PIC, and PN export fluxes directly measured by floating traps deployed in six cruises to the SEATS during October 2006-December 2008.The export fluxes were also estimated from the distributions of the three particlereactive radionuclides in the euphotic layer.The goals of this study therefore are (1) to investigate the temporal variability of particle scavenging and removal processes, (2) to Introduction

Conclusions References
Tables Figures

Back Close
Full compare the geochemical behavior of 234 Th, 210 Pb, and 210 Po, and (3) to estimate the export flux of POC, PIC, and PN by 234 Th/ 238 U, 210 Po/ 210 Pb, and 210 Pb/ 226 Ra disequilibria in the euphotic layer of the northern South China Sea.

Material and methods
Seawater samples were collected by multiple cruises to the SEATS site in the South China Sea on board the R/V Ocean Researcher I or III (refer to Fig. 1 for the timeline of the cruises).

Seawater
Seawater was collected using either 10 l (R/V Ocean Researcher III) or 20 l (R/V Ocean Researcher I) Teflon-coated Go-Flo bottles mounted on a Sea-Bird ® CTD (SBE 9/11) rosette system.Except for the first cruise (ORI-812), from which only 234 Th activity in unfiltered seawater was measured, filtration of large-volume seawater for the determination of 234 Th, 210 Pb, and 210 Po activities in filtrate and particles was carried out for all samples.At each sampling depth, 40 l seawater was collected and divided into one 20 l and two 10 l subsamples.The 20 l sample was used to determine 234 Th and the two 10 l were used to determine 210 Pb and 210 Po, respectively.Seawater was immediately pressure-filtered by compressed air through a pre-weighed 142 mm Nuclepore filter (0.45 µm) mounted in a Plexiglas filter holder.Filtrate and particles retained on the filter were processed and analyzed for 234 Th, 210 Pb, and 210 Po activities following the procedures described in Wei et al. (2009).Auxiliary standard parameters (e.g., nutrients, primary productivity, etc.) were measured by technical personnel of SEATS program.Nutrients were determined following the methods of Strickland and Parsons (1984).Primary productivity was determined by Na 14 HCO 3 incubation of seawater samples collected from 6-8 depths in the upper 100 m.Introduction

Conclusions References
Tables Figures

Back Close
Full

Floating trap
Particle interceptor trap (Wei et al., 1994) was deployed for the measurement of vertical fluxes of total mass (F M ), 234 Th (F Th ), 210 Po (F Po ), 210 Pb (F Pb ), particulate carbon (F PC ), particulate nitrogen (F PN ), and particulate organic carbon (F POC ), at three depths (30 m, 100 m, 160 m).At each depth, a set of sediment traps consisting of eight Plexiglas tubes filled with trap solution prepared from the mixture of filtered seawater and formaldehyde were snapped on a polypropylene cross frame, which is secured to a mooring line.It should be noted that radionuclide release into overlying waters in unpoisoned traps (even when deployed for less than 1 day) has been documented (Hung et al., 2010;Xu et al., 2011), which can make radionuclide estimates a minimum.
The mooring array was left free drifting for 37-48 h.Upon recovery, the upper layer of seawater in the tube was siphoned off and the remaining trap solution was filtered through a pre-weighed 90 mm Nuclepore filter (0.45 µm pore size).The swimmers, mainly macrozooplanktons, were picked under a microscope.The samples then were rinsed by deionized water to remove salts.The filters were then dried in a desiccator and weighed for total mass flux calculation.The filters were analyzed for 234 Th, 210 Pb, and 210 Po activities following the procedures as particulate samples described in Wei et al. (2009).
The samples in tubes used for carbon and nitrogen analyses were filtered through pre-combusted Whatman 25 mm GF/F filters.After swimmers were picked, the filters were wrapped firmly into tin boats and loaded into the autosampler of EA elemental analyzer (EuroEA3000) for total carbon and nitrogen content for F PC and F PN calculations.The filters for organic carbon analysis were fumed for 24 h with concentrated HCl to remove calcium carbonate, then analyzed for carbon content for F POC calculation.Fluxes of particulate inorganic carbon (F PIC ) were calculated by difference of F PC and F POC .The overall procedural error are better than ±2 % for both carbon and nitrogen determinations.Introduction

Conclusions References
Tables Figures

Back Close
Full

Sinking fluxes
Sinking fluxes of total particulate matter or mass (F Mass ), POC (F POC ), PIC (F PIC ), PN (F PN ), 234 Th (F Th ), 210 Pb (F Pb ), and 210 Po (F Po ) at the three deployment depths (30, 100, and 160 m) are shown in Fig. 5a-g, respectively.The average, range, and standard deviation of sinking fluxes of various parameters are summarized in Table 1.In Introduction

Conclusions References
Tables Figures

Back Close
Full general, all fluxes decreased with depth.Although the fluxes for all components, except 210 Pb, show the largest variability with a RSD of >70 % at 30 m, the correlations between fluxes are noticeably high, with correlation coefficients of >0.9 between all combinations of flux parameters.Except for the F PIC , and F Po , all fluxes varied within a relatively small range, with a RSD of ∼40 % at 100 m, the approximate euphotic depth at SEATS.

Deficiencies and fluxes of 234 Th, 210 Po, and 210 Pb
There are very few previous open ocean studies that report 234 Th, 210 Pb, and 210 Po data determined from the same site.Among those few, Sarin et al. (1994) reported three vertical profiles of dissolved activity concentrations of 234 Th, 210 Pb, and 210 Po in the northeastern Arabian Sea.Wei andMurray (1991, 1994) compared the geochemical behavior of 234 Th, 210 Pb, and 210 Po in the Black Sea.Shimmield et al. (1995) measured the three radionuclides from the same seawater samples collected in the upper 500 m from the marginal ice zone in the Antarctica.Kim and Church (2001) presented dissolved and particulate 234 Th, 210 Pb, and 210 Po data determined on the same samples collected from the Sargasso Sea.Murray et al. (2005) reported a complete 234 Th, 210 Pb, and 210   (Buesseler et al., 2008).The remineralization phenomenon is also shown as a layer of excess 210 Po found in all cruises, except in June and December 2008, corroborating Introduction

Conclusions References
Tables Figures

Back Close
Full the results from other regions (Bacon et al., 1976;Thomson and Turekian, 1976).As a result of atmospheric deposition, excess 210 Pb was commonly observed in the upper 200 m during all cruises, consistent with the findings of Obata et al. (2004) and Chung and Wu (2005).
Assuming steady-state and negligible physical transport, the flux of 234 Th out of the euphotic depth can be calculated by Eq. ( 1).
where F Th is the removal flux, λ Th is radioactive decay constant (=0.0387 d −1 ) and V is the net flux of 234 Th from physical transport.To justify the assumption of steadystate, the contribution of temporal changes in 234   2006) also concluded that advection/diffusion is not a significant term in the scavenging model.
The flux of 210 Pb out of the euphotic depth can be calculated by Eq. ( 2).
where F Pb is the removal flux, λ Pb is radioactive decay constant of 210 Pb Introduction

Conclusions References
Tables Figures

Back Close
Full The flux of 210 Pb out of the euphotic depth is calculated by Eq. ( 3), where F Po is the removal flux, λ Pb is radioactive decay constant of 210 Po (=0.005 d −1 ) and I Po is the atmospheric 210 Po deposition flux, which is assumed to be 10 % of the atmospheric 210 Pb flux (Turekian et al., 1977).Since the spatial variation of 210 Po in the surface water of the South China Sea is small (Yang et al., 2006;Wei et al., 2011), as is the case for 234 Th, horizontal transport should be negligible.However, similar to the nutrients, the vertical profile of 210 Po t shows evidence for an increase in the thermocline layer, vertical diffusion may play a significant role in the mass balance of 210 Po in the euphotic layer (Sarin et al., 1994).If an eddy diffusion coefficient of 0.55 cm 2 s −1 (Nozaki and Yamamoto, 2001) and mean vertical gradient of 210 Po between 0 and 300 m were applied to the mass balance equation, the vertical fluxes of 210 Po due to upward diffusion would be equivalent to 2-8 % of the total flux, lower than that in oligotrophic North Pacific reported by Verdeny et al. (2008).Hence, the contribution of the vertical diffusion does not need to be included in the flux estimation.It is known that the sediment traps can show trapping efficiencies different from 1 (or 100 %), due to hydrodynamic and swimmer effects (Buesseler, 1991;Buesseler et al., 2007), as well as selective dissolution effects (Hung et al., 2010;Xu et al., 2011).
To evaluate the trapping efficiency of the floating trap, the 234   3).A large range of trapping efficiency for floating traps, with the ratio of measured and modeled fluxes ranging from 0.1 to 20, is found (Buesseler, 1991;Stewart et al., 2007;Wei and Murray, 1992;Buesseler et al., 2007).A trapping efficiency close to unity and a small variability for the traps Introduction

Conclusions References
Tables Figures

Back Close
Full deployed at 100 m increase the reliability of the export flux assessment (Buesseler et al., 2007).While it is noted that the traps deployed at 30 m, the mixed depth, generally give significantly higher radionuclide fluxes than expected values and reveal a large variability, they also experience greater hydrodynamic effects, which may be the cause for these phenomena.It is also worth noting that, except for the 100 m trap of July 2007, all trapping efficiencies estimated by F Po and F Po are greater than unity, which may be related to the fast regeneration rate of 210 Po due to particle remineralization in the euphotic layer.Compared with the concentrations of trace metal in the same samples (Ho et al., 2010), it was found 210 Po content is highly correlated (r > 0.8, n = 11) with biophilic elements, P, S, Ca, Zn, Mo, and Cd.Previous studies have shown that 210 Po, a sulfur analog found mostly associated with proteins (Fisher et al., 1983), shows higher affinity toward organic particles than 210 Pb (Bacon et al., 1976;Shannon et al., 1970), and thus, 210 Po is more sensitive to particle decomposition processes than 234 Th (Stewart et al., 2007), which is mostly surface-bound (Santschi et al., 2006).

Export fluxes of POC, PIC, and PN estimated by radionuclide tracers
Export fluxes of POC, PIC, and PN can be indirectly estimated if the ratio of the concentration of carbon and nitrogen to the radionuclides in sinking particles are known.
Since the first study in the Panama Basin (Murray et al., 1989), 234 Th has been extensively used in the past two decades as a powerful tracer for estimating the export production (i.e., POC flux) from the euphotic layer of the ocean (Buesseler et al., 2006 and refereces therein).For the same purpose, 210 Pb and 210 Po, as additional tracers, were used to constrain the determination of the export production in the euphotic layer (Shimmield et al., 1995;Stewart et al., 2007;Murray et al., 2005;Friedrich and Rutgers van der Loeff, 2002;Kim and Church, 2001;Buesseler et al., 2008).In the context of treating particle-reactive radionuclides as proxies for POC flux, export production or sinking flux of PIC and PN from the euphotic layer can also be estimated by multiplying the flux of radionuclides, which is obtained from the mass balance of daughter radionuclide relative to its parent radionuclide in the euphotic layer, by the ratio of POC (or PIC, 9681 Introduction

Conclusions References
Tables Figures

Back Close
Full PN) and radionuclide in sinking (or sinkable) particles (Murray et al., 2005;Buesseler, 1998;Bacon et al., 1996), i.e., where RN is the abbreviation of radionuclide ( 234 Th, 210 Pb, 210 Po), F POC-RN , F PIC-RN , and F PN-RN are fluxes of POC, PIC, and PN estimated by RN as proxy, respectively, F RN is the flux of radionuclide calculated from the deficiencies of RN in the euphotic layer by Eqs. ( 1)-( 3), and POC/RN , PIC/RN, and PN/RN is the ratio of POC, PIC, and PN and the radionuclides in sinking (or sinkable) particles collected at the euphotic depth.
Equations ( 4) and ( 6) were previously used for 234 Th data to estimate the export flux of POC and PN from the euphotic layer of the Atlantic Ocean (Buesseler et al., 1992;Amiel et al., 2002) and the Pacific Ocean (Murray et al., 1996).Limited by the scarcity of sediment trap data on radionuclide to PIC ratios, very few applications of Eq. ( 5) on PIC export exist.The only application of using 234 Th as a PIC proxy is Bacon et al. (1996), who estimated PIC export fluxes based on PIC/ 234 Th ratios in filtered particles in the Equatorial Pacific.
The POC/RN, PIC/RN, and PN/RN in sinking particles collected at the three depths are shown in Fig. 7.The POC/ 234 Th ratios in sinking particles decreased with depth, with average values of 23.9 ± 16.3, 8.7 ± 3.2, and 4.8 ± 1.8 µmol dpm −1 at 30, 100, and 160 m, respectively.The decreasing trend with depth can be attributed to preferential remineralization of organic carbon when particles settle through the water column (Rutgers van der Loeff et al., 2002;Buesseler et al., 2006).Decreasing of POC/ 234 Th ratio with depth were also found for suspended particles in the northern South China Introduction

Conclusions References
Tables Figures

Back Close
Full Sea by Cai et al. (2008) and Chen et al. (2008).A larger variability of this ratio was found at the mixed layer depth, which may be caused by a more active biological activity in the surface layer.It is noted here that the POC/ 234 Th ratio in sinking particles collected at the euphotic depth falls in a small range and values are similar to those reported for more productive regions, e.g., Northern Atlantic (Buesseler et al., 1992) and higher than in the oligotrophic open ocean, e.g., HOTS and BATS (see the compilation of Buesseler et al., 2006).Due to the low scatter of POC/ 234 Th ratio at the euphotic depth, the export production based on Eq. ( 4) tends to be less variable (Buesseler et al., 2006).Except during July 2007, the POC/ 210 Pb ratio in sinking particles also shows a decreasing trend with depth (Fig. 7b).The average POC/ 210 Pb ratios are 1088 ± 403, 686 ± 352, and 321 ± 131 µmol dpm −1 at 30, 100, and 160 m, respectively.These values are comparable to POC/ 210 Pb ratios of 80-268 µmol dpm −1 in sinking particles collected by sediment traps at 150 m off the east coast of the United States (Biscaye et al., 1988) and 50-489 µmol dpm −1 in the northwestern Mediterranean (Tateda et al., 2003).The POC/ 210 Po ratios in the sinking particles are highly variable, ranging from 120 to 1200 µmol dpm −1 , similar to ratios in the sinking particles collected in the coastal Mediterranean Sea (Tateda et al., 2003;Stewart et al., 2007).However, compared to open ocean values (Verdeny et al., 2008 and references therein), our data fall on the high end of these POC/ 210 Po ratios.Different from 234 Th and 210 Pb, no systematic trends with depth were found for POC/ 210 Po ratios.
Values of the PIC/ 234 Th (Fig. 7d) and PN/ 234 Th (Fig. 7g) ratios show temporal and vertical variations similar to those of POC/ 234 Th, suggesting no preferential association of 234 Th with different components of biological particles.The average PIC/ 234 Th ratios are 8.9 ± 6.3, 2.8 ± 1.1, and 1.9 ± 2.4 µmol dpm −1 at 30, 100, and 160 m depth, respectively.These ratios are comparable to the limited published sediment trap data, e.g., 0.6 µmol dpm −1 at 300 m in the Northwest Mediterranean (Szlosek et al., 2009).
The PN/ 234 Th ratios in the lower euphotic zone range between 0.6 and 1.4 µmol dpm −1 , which is similar to 0.7-1.3µmol dpm −1 in the Northern Atlantic (Buesseler et al., 1992).

Conclusions References
Tables Figures
Using the ratios of POC, PIC, and PN to the three radionuclides in sinking particles of the 100 m trap, the export fluxes of organic carbon, inorganic carbon, and nitrogen via particle settling from the euphotic layer can be calculated by Eqs. ( 4)-( 6) and plotted for different cruises in Fig. 8. Along with the estimated fluxes based on the three radionuclide proxies, directly measured fluxes of POC, PIC, and PN by sediment traps deployed at 100 m are also shown in the figure.The export fluxes of POC, PIC, and PN from the euphotic layer of the northern South China Sea determined by various approaches are discussed in the following section.

Comparison of export fluxes by different approaches
As can be seen from They attributed the negative values at some stations in the study area to large input of particulate matter by horizontal transport.Although 210 Pb has been found useful in estimating the POC sinking flux in the deep ocean (Moore and Dymond, 1988), there are very few studies using the 210 Pb/ 226 Ra disequilibrium to estimate the export production in the euphotic layer because a reliable atmospheric 210 Pb flux is not available to better constrain the source term in Eq. ( 2), which is much greater than the input from the radioactive decay of 226 Ra.Nonetheless, assuming the atmospheric 210 Pb flux of 0.6 dpm cm −2 a −1 (Feichter et al., 1991;Xu et al., 2010), the F POC-Pb of 7.2-21.3mmol-C m −2 d −1 can be estimated, which is within the range of F POC and F POC-Th values.Since a constant atmospheric 210 Pb flux is assumed, the F Pb ' values show no seasonal variation (Table 2).Hence, the temporal variations of the F POC-Pb are resulting from the POC/ 210 Pb ratio in sinking particles, which may be affected by particle composition.The F POC-Po based on the 210 Po- 210 Pb system showed the largest range of POC export, ranging from 1.8 to 20.3 mmol-C m −2 d −1 , among the three pairs of disequilibria.Verdeny et al. (2009) recently presented a review of previous studies using 234 Th and 210 Po as dual tracers to estimate the export production in the ocean.A discrepancy between the POC fluxes estimated by 210 Po- 210 Pb and 210 Po- 210 Pb disequilibria was commonly found.For example, the export production deduced from the 210 Po deficiency was two folds higher than that estimated from the 234 Th-238 U disequilibrium in the Atlantic (Sarin et al., 1994).On the other hand, Shimmield et al. (1995) found the export production estimated by the F POC-Po was an order of magnitude lower than that  2008) used 234 Th and 210 Po proxies to estimate POC export fluxes from the euphotic layer of the Sargasso Sea and found the F POC-Po is about two folds higher than the F POC-Th .Hence, considering the intrinsic difference in half-life, source function, and geochemical characteristics among the three radionuclides and the uncertainty associated with the scavenging model, the POC fluxes estimated by these proxies in our work can be reliably accepted as representative of the export production in the northern South China Sea.
There is some literature that estimated the export production by various other approaches in the northern South China Sea.These approaches include a 228 Ra/NO 3 coupled model (Nozaki and Yamamoto, 2001), biochemical modeling (Liu et al., 2002), carbon budgeting (Chou et al., 2006), 15 N incubation measurements of new production (Chen andChen, 2004, 2005), and phosphorus flux measurements by sediment traps (Ho et al., 2009).The export fluxes by these approaches are summarized in Table 4.
Among these studies, based on the P fluxes measured by the sediment traps deployed at the depth underlying the euphotic layer, Ho et al. (2009)  production (Buesseler, 1998), which showed a large range between 0.04 and 2.17 with an average ratio of 0.53.The unreasonably high ratios greater than unity were found for ORIII-1239, from which the primary production was suspected to be underestimated based on the vertical profile of primary productivity.It is found that most of data fall in the range of 0.2 and 0.7, which implies that export efficiency is high in the South China Sea.
The values of F PIC range from 2.1 to 9.5 mmol-C m −2 d −1 (Fig. 8b), which is significantly higher than in other regions (Berelson et al., 2007).The temporal variation of the F PIC is generally predicted by the fluxes estimated from the other proxies.Based on the 234  Atlantic (Martin et al., 1993) and the northeast Pacific (Wong et al., 1999).

Conclusions
Temporal data of dissolved and particulate concentrations and fluxes of the particlereactive radionuclides, 234 Th, 210 Pb, and 210 Po, in the upper water column of the time-series station, SEATS (18 Full  Full    Full Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Po p in the upper 200 m at SEATS station, and related TSM concentrations from the filtration of the 210 Po samples are shown in Fig. 4.
Discussion Paper | Discussion Paper | Discussion Paper | Po data set measured from seawater samples and settling particles collected by floating traps deployed in the Equatorial Pacific.Vertical profiles showing the daughter/parent nuclide ratios of 234 the upper 200 m obtained from the six cruises are given in Fig. 6.Relative to their parent radionuclides, deficiencies of 234 Th and 210 Po in the euphotic layer were commonly found, indicating evidence for an efficient removal process.A layer with excess 234 Th underlying the euphotic layer was found in January 2006, October 2007 and June 2007, which is the result of remineralization Discussion Paper | Discussion Paper | Discussion Paper | Th activity concentrations in the mass balance needs to be assessed.It is found that the temporal change is only equivalent to 0.1-3.1 % of the contribution by radioactive decay, which is negligible in the flux calculation.The F Th caused by physical transport can be evaluated based on the unpublished 234 Th data obtained from 6 stations in March and July 2000, which covered 3 × 3 degree region of the study area.Following the procedures of Liang et al. (2003), the mean current velocity was estimated by averaging the archived shipboard ADCP data collected during 1991-2006.The results showed that horizontal transport by surface currents contributes only 1.5 % of 234 Th in the region.Negligible contribution of 234 Th from horizontal advection at the SEATS site corroborates to the work conducted by Chen et al. (2008) in the northern shelf of the South China Sea.Based on the distributions of 234 Th and 228 Th in the upper 500 m, Cai et al. ( Discussion Paper | Discussion Paper | Discussion Paper | (=8.48 × 10 −5 d −1 ) and I Pb is the atmospheric 210 Pb deposition flux.The I Pb is about 0.4 dpm cm −2 a −1 according to the model by Feichter et al. (1991).Based on 210 Pb chronology in ornithogenic sediments, Xu et al. (2011) estimated the atmospheric 210 Pb flux of 0.76 dpm cm −2 a −1 in the southern South China Sea.Here we assume an average I Pb of 0.6 dpm cm −2 a −1 (16.4 dpm m −2 d −1 ).
The inventories, deficiencies, and removal fluxes of234 Th,210 Pb, and210 Po in the euphotic layer are summarized in Table2.Daughter nuclide deficiencies from their parent nuclides resulted in removal fluxes of 1.1 × 10 3 -1.8× 10 3 dpm m −2 d −1 and 7.1-40.2dpm m −2 d −1 for 234 Th and 210 Po, respectively, from the euphotic layer.Due to atmospheric input, an excess of 210 Pb relative to 226 Ra is commonly observed in the upper water column.As is apparent, the atmospheric input is the dominant source term in Eq. (2), thus, the F Pb is close to the constant atmospheric 210 Pb flux, Discussion Paper | Discussion Paper | Discussion Paper | Th/ 238 U disequilibria in the upper water column are commonly used to assess the trapping efficiency of the floating trap.The 234 Th flux measured by the sediment trap deployed at specific depth should be equal to the removal flux calculated from the deficiency in the water column above the depth.Ideally, following the same concept, the comparison of measured with estimated 210 Pb and 210 Po fluxes can also be used as an indicator of trapping efficiency of the floating traps.Hence, we have calculated the trapping efficiencies of the floating traps by the ratios of measured and predicted fluxes of 234 Th, 210 Pb, and 210 Po at the three depths.The trapping efficiencies for the six cruises and the average values at the three deployment depths are shown in Table 3.The trapping efficiencies for the traps deployed at 100 m, the euphotic depth, range between 1.06 ± 0.2 and 3.99 ± 3.23 based on 234 Th and 210 Po data, respectively.Considering the uncertainties associated with sediment trap deployments, these trapping efficiencies are acceptable, as most of the values at 100 and 160 m are close to 1 (Table Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | temporal and vertical variations similar to those of POC/ 210 Pb.The PIC/ 210 Pb ratios are 433 ± 252, 279 ± 51, and 108 ± 96 µmol dpm −1 at 30, 100, and 160 m, respectively. Fig. 8, values of F POC-Th display a remarkable resemblance to F POC values, both in terms of magnitude and temporal variation.The F POC-Th ranges from 9.6 mmol-C m −2 d −1 to 21.0 mmol-C m −2 d −1 and the F POC ranges from 9.8 to 18.5 mmol-C m −2 d −1 .The temporal variation of both F POC and F POC-Th generally corroborates with the seasonal variation of primary productivity at SEATS (Tseng et Discussion Paper | Discussion Paper | Discussion Paper | al., 2005).In the region further south to SEATS,Cai et al. (2008), using a 3-D scavenging model, estimated the POC export ranged from −10.7 to 12.6 mmol-C m and POC/ 234 Th ratios in pump-collected particles.
estimated from the F POC-Th values in the marginal ice zone of the Antarctica.In the Equatorial Pacific, Murray et al. (2005) reported a generally higher and more variable F POC-Po than F POC-Th values.Stewart et al. (2007) obtained the F POC-Th and F POC-Po the northwestern Mediterranean.Recently, Buesseler et al. ( obtained the largest range of export fluxes (4.2-70.8mmol-C m −2 d −1 ) at the SEATS.Some anomalously high fluxes reported byHo et al. (2009) may be caused by over-trapping of their sediment traps, which were deployed at 160 m on a long mooring line.It can thus be concluded that the export production can be reasonably estimated by using234 Th,210 Pb, and210 Po as carbon proxies in sinking particles.In a more recent paper, Ho et al. (2011) evaluated the POC fluxes estimated from particulate phosphorus fluxes from the moored traps deployed at 120 m and found consistent results with the observations by the floating traps.The correlation of POC exports obtained by various methods and integrated primary production during the six cruises is shown in Fig. 9.The primary production varied from 9.5 mmol-C m −2 d −1 in July 2007 to 46.1 mmol-C m −2 d −1 in October 2007.The export efficiency can be indicated by the ratio of the POC exports and the primary Discussion Paper | Discussion Paper | Discussion Paper | Th deficiency and the PIC/ 234 Th ratio in pump-collected large particles, Bacon et al. (1996) estimated a PIC export of 0.45-0.80mmol-C m −2 d −1 at 150 m and 0.54-0.71mmol-C m −2 d −1 at 200 m in the Equatorial Pacific.High F PIC values at SEATS may be attributed to the dominant coccolithophores in the algal community during the winter season in the South China Sea (Chen and Chen, 2005), which enhances the sinking flux due to a ballast effect.Indeed, microscopic examination of trap particles shows that skeletal remains of coccolithophorids and diatoms are abundant during the winter cruises (ORI-821 and ORI-887) (Ho et al., 2010).Calculated from the temporal change of the inventory of alkalinity in the euphotic layer during October 2006 and December 2008, a PIC production ranging from 0.4 to 11.4 mmol-C m −2 d −1 can be estimated.Direct measurements of Ca concentration from the same samples shows F PIC ranging from 1.6 to 6.2 mmol-C m −2 d −1 (Ho et al., 2010).All the F PIC , F PIC-Th , F PIC-Pb , and F PIC-Po values are thus within the PIC production estimates, indicating the PIC export flux was reliably estimated.The F PN values measured at 100 m of the SEATS ranged from 1.5 mmol-N m −2 d −1 in July 2007 to 4.1 mmol-N m −2 d −1 in December 2008, very close to the flux of ∼2 mmol-N m −2 d −1 derived from a 228 Ra/NO 3 coupled model at the site about 300 km to the south of our station (Nozaki and Yamamoto, 2001).Except for the F PN-Po in October 2006 and January 2007, the values of F PN-Th , F PN-Pb , and F PN-Po are close to the F PN within a factor of three.The export of nitrogen via particle settling in the northern South Discussion Paper | Discussion Paper | Discussion Paper | China Sea is also close to the measured PN flux of 1.2-2 mmol-N m −2 d −1 in the North Discussion Paper | Discussion Paper | Discussion Paper | Bacon, M. P., Cochran, J. K., Hirschberg, D., Hammar, T. R., and Fleer, A. P.: Export flux of carbon at the equator during the EqPac time-series cruises estimated from 234 Th measurements, Deep-Sea Res.Pt.II, 43, 1133-1153, 1996.Berelson, W. M., Balch, W. M., Najjar, R., Feely, R. A., Sabine, C., and Lee, K.: Relating estimates of CaCO 3 production, export, and dissolution in the water column to measurements of Discussion Paper | Discussion Paper | Discussion Paper | from time series of sediment traps at Ocean Station P, 1982-1993: relationship to changes in subarctic primary productivity, Deep-Sea Res.Pt.II, 46, 2735-2760, 1999.Wong, G. T. F., Ku, T. L., Mulholland, M., Tseng, C. M., and Wang, D. P.: The SouthEast Asian time-series study (SEATS) and the biogeochemistry of the South China Sea -An overview, Deep-Sea Res.Pt.
Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper |

Fig. 1 .Fig. 3 .Fig. 5 .Fig. 6 .Fig. 7 .
Fig. 1.Location of SEATS station in the South China Sea.Inserted box shows the timeline of the six cruises to the station.

Table 3 .
Trapping efficiencies estimated by the ratios of measured and modeled fluxes of 234 Th,210Pb, and210Po from the three depths for all cruises.Average value, standard deviation, and relative standard deviation for each depth are listed.Depth Oct 2006 Jan 2007 Jul 2007 Oct 2007 Jun 2008 Dec 2008 Av.