1 INTRODUCTION

About 30% of anthropogenic CO₂ is taken up bythe ocean，especially the Arctic and Southern oceans，owing to high biological production in the large oceanmargin areas and low temperatures(Cai et al.，2010).The Chukchi Shelf is one of the most productivecontinental shelves in the Arctic Ocean and isimportant in carbon sequestration(Woodgate et al.，2005; Yu，2010). The magnitude of atmospheric CO₂absorbed in these areas has been of concern tooceanographers worldwide(Baskaran et al.，1995;Anderson et al.，1998; Pipco et al.，2002; Bates et al.，2006; Liu et al.，2007; Else et al.，2008).

As an interface for exchange of CO₂ between thesurface ocean and interior ocean，the euphotic zoneplays a key role in the growth，removal and cycling ofbiomass. Fluxes within the euphotic zone of carbon，nutrients and other associated elements involved in biogeochemical cycles are very important in the studyof global CO₂(Yang et al.，2004; He et al.，2008).Radioactive isotopes，especially natural particlereactiveradionuclides(e.g.，²³⁴Th and ²¹⁰Po)provideanother possible means for quantifying export fl ux ofparticulate organic carbon(POC)from the surfaceocean at various time scales，because of their specifi chalf-lives(Verdeny et al.，2009; Stewart et al.，2011).

Both ²¹⁰Po and ²¹⁰Pb are progeny of the ²³⁸U decaychain，and are naturally occurring particle-reactiveradionuclides. ²¹⁰Pb(t_1/2 =22.3 a)is produced in situ bydecay of its longer-lived gr and parent ²²⁶Ra and fromatmospheric deposition，whereas ²¹⁰Po is supplied almost exclusively in situ by the decay of itsgr and parent ²¹⁰Pb，with a minor additional sourcefrom atmospheric deposition(Baskaran，2011).Although ²¹⁰Pb and ²¹⁰Po are particle-reactiveradionuclides，the preferential uptake of ²¹⁰Po byplankton and its transfer in the food chain can be usedto quantify organic carbon fl uxes(Carvalho，2011;Kim et al.，2012). Furthermore，export on a timescaleof several months can be derived by ²¹⁰Po/ ²¹⁰Pbdisequilibrium，longer than the several weeks of²³⁴Th/ ²³⁸U disequilibrium(Friedrich et al.，2002).

In this study，we estimated ²¹⁰Po-derived exportfl uxes of POC from the euphotic zone at differentstations during the 4th Chinese National ArcticResearch Expedition(CHINARE-4).2 MATERIAL AND METHOD 2.1 Location and sample collection

During CHINARE-4 from July 1 to September 20，2010，a conductivity，temperature and depth(CTD)rosette(SEABIRD 911/17)was used to collect 10-Lseawater samples at various depths in the upper0–100 m water column at four stations in the ArcticOcean，for analysis of ²¹⁰Po and ²¹⁰Pb. At stations R03 and R09，sampling was done during July 21–24，whereas that at R12 and R20 was during August24–29. A map of the study area and stations isprovided in Fig. 1.

One liter of each 10-L seawater sample was fi lteredthrough a pre-combusted(450°C，2 h)0.45-μm QMAmembrane immediately after sampling，and thenparticles and the fi lter were sealed inside tin foil and stored frozen for subsequent POC analysis. Theremaining nine liters of seawater from each sample were fi ltered through a 0.45-μm membrane fi lter(120 mm diameter)，to obtain dissolved samples and particulate samples. After fi ltration，dissolved sampleswere acidifi ed to pH 1 with concentrated HCl. Allsamples were stored until laboratory analysis of POC，²¹⁰Po and ²¹⁰Pb.2.2 Radionuclide analysis

For dissolved and particulate ²¹⁰Po and ²¹⁰Pb，theanalytical methods used followed the modifi cationsof Nozaki(1986) and Masque et al.(2002). Briefl y，fi lter samples were digested using sequential aquaregia，HNO₃ and HCl，with HClO₄ and HF，and thedissolved material was eventually taken up in1.5 mol/L HCl. During digestion，samples werespiked with ²⁰⁹Po as a yield tracer and stable lead as arecovery tracer. The dissolved ²¹⁰Po was coprecipitatedwith Fe(OH)₃ and redissolved with1 mol/L HCl. Po was plated onto nickel discs usingstirring hot plates and counted on a low backgroundalpha spectrometer(Canberra 7200-08; CT，USA).The plating solution was then replated to remove anyresidual ²¹⁰Po and ²⁰⁹Po，then re-spiked with ²⁰⁹Po and retained for at least 6 months for ingrowth of ²¹⁰Po todetermine ²¹⁰Pb. Pb yields were determined throughmeasurement of stable Pb by inductively coupledplasma mass spectrometry(ICP-MS)，and recoverywas generally between 80% and 90%. The ²¹⁰Po and ²¹⁰Pb activities were decay-corrected to the time ofsampling according to Fleer and Bacon(1984). 3 RESULT AND DISCUSSION 3.1 Distribution of particulate and dissolved ²¹⁰Po and ²¹⁰Pb

The activity of ²¹⁰Po and ²¹⁰Pb in different phases and the activity ratio ²¹⁰Po/ ²¹⁰Pb)_A.R are given inTable 1. Vertical profi les of ²¹⁰Po/ ²¹⁰Pb)_A.R in differentwater columns are shown in Fig. 2. The averagedissolved ²¹⁰Pb of various stations shows an increasingtrend from 1.12 to 1.78 Bq/m³ northward from July toAugust，which may have resulted from atmosphericdeposition of ²¹⁰Pb during ice melt. The averageparticulate ²¹⁰Pb(2.22 Bq/m³)was higher than thedissolved ²¹⁰Pb(1.49 Bq/m³)，indicating intenseparticle scavenging of ²¹⁰Pb(Bacon et al.，1976;Shimmield et al.，1995).

The vertical profi le of POC in different watercolumns and relationship between POC content and particulate Po are shown in Fig. 3. The average contentof POC decreased northward. However，the highest value for surface water appeared at R09，perhapscaused by ice melt，which resulted in biomass bloomfrom the bottom ice(Yu et al.，2012). Positivecorrelation(R² =0.027)between POC content and particulate Po confi rms that Po may be used as a tracerof POC export.

3.2 Total Po and Pb，residence time of ²¹⁰Po，watercolumndisequilibrium and flux

The half-life of ²¹⁰Po constrains the seasonal timescale of disequilibrium of ²¹⁰Po/ ²¹⁰Pb for quantifyingthe export fl uxes of POC. Boundary scavenging，especially advective processes，is important for ²¹⁰Pb，owing to its half-life of 22.3 years(Friedrich et al.，2002; Kim et al.，2012). The spatially comparableactivity of ²¹⁰Po and ²¹⁰Pb further constrains the fl uxesbecause of advective processes，which are minorrelative to vertical fluxes.

An irreversible，steady state model of ²¹⁰Po wasbuilt without advective and diffusive items and atmospheric deposition. To estimate the rates of ²¹⁰Po and ²¹⁰Pb scavenged by suspended particles and theirremoval via sinking particles，the following equations were used for the disequilibrium between ²¹⁰Po and ²¹⁰Pb in seawater(Bacon et al.，1976):

where J is the scavenging flux of dissolved ²¹⁰Po; λ_Pois the decay constant(0.005/d)of ²¹⁰Po; I DPo and I DPbare the inventory of dissolved ²¹⁰Po and ²¹⁰Pb，respectively; P is the export fl ux of particulate ²¹⁰Po;I_PPo and I_PPb are the inventory of particulate ²¹⁰Po and ²¹⁰Pb. The residence times of particulate and dissolved²¹⁰Po are given byBased on Eqs.1–4，J，P，τ_DPo，and τ_PPo are presentedin Table 2. This shows that mean residence times ofparticulate and dissolved ²¹⁰Po in the euphotic zone ofdifferent stations were almost all negative，which maybe caused by regeneration of particles(He et al.，2008). However，the residence time of particulate²¹⁰Po was higher than that of dissolved ²¹⁰Po，indicatingrapid biogeochemical cycling in the study area(Rutgers van der Loeff et al.，1995). 3.3 Distribution of particulate organic carbonconcentrations

Concentrations of POC in the water column aregiven in Table 1. POC concentration was 9–601 μgC/L in the study area，with average 106 μg C/L. Thehighest concentration of 601 μg C/L were found atR03 during the fi rst sampling period，this decreasednorthwards and offshore during the second samplingperiod，consistent with the distribution of chlorophyll-a(Yu et al.，2012). POC concentrations in surfacewater were lower than in deeper water，owing tosediment resuspension(Yu et al.，2012).3.4 POC export fl uxes from euphotic zoneThe distribution of POC export fl uxes in the ArcticOcean has signifi cant temporal and regional variability(Yu et al.，2012). Export fl uxes of particulate ²¹⁰Po and its residence time in the euphotic zone were calculatedin the irreversible scavenging model(Yang，2005).Furthermore，POC export fl uxes were calculated withthe ratio of POC content to activity of ²¹⁰Po，per theequation of Buesseler et al.(1992) and Yang(2005):

where f_Po is the export flux of particulate ²¹⁰Po and r isthe ratio of content of POC to activity of ²¹⁰Po at thebottom of the euphotic zone. With this equation，wehypothesized that the carriers of ²¹⁰Po and POC werethe same particle but not that they have similargeochemical behaviors. Apparently，the formerhypothesis was easily satis fied in the study area. Theexport fl uxes and the residence times of ²¹⁰Po arelisted in Table 2，and the POC fluxes(Method B，marked with B_POC-flux)are given in Table 3. At stationR03，the residence time of ²¹⁰Po was the shortest，only0.38 a，and the POC export fl ux was the highest，at1 848 mmol C/(m² ∙a). The latter may have beencaused by the upwelling in the area(Yu，2010).

Another equation to calculate POC export flux atsteady state was developed by Eppley(1979)，which assumes the residence time of ²¹⁰Po and POC as

where the f_POCinv is export flux of the POC inventory inthe euphotic zone. The results(Method E，marked byE_POC-flux)derived from Eq.6 are listed in Table 3.Compared with the export fl uxes derived from methodB，the results calculated by different methods arealmost equal(within 40%; He et al.，2008).Considering errors for the activities of ²¹⁰Po and content of POC，we assume that the results derived bythe above two methods are consistent.

Average POC export fl uxes 945 mmol C/(m² ∙a)(Method E) and 1 848 mmol C/(m² ∙a)(Method B)inthe euphotic zone at R03 were much higher than thoseat R20，126 mmol C/(m² ∙a)(Method E) and 109 mmolC/(m² ∙a)(Method B). This shows a decreasing trendnorthward，comparable to prior work in the sameregion(Yu et al.，2012). Compared with other reportedPOC export fl uxes(Table 3)in certain areas of theworld ocean based on various approaches，the fl uxeson the Chukchi Shelf in summer were very high. It isconcluded that there is high export production on theChukchi Shelf in summer. Because available lightlimits marine primary production at high latitude，thisproduction is increased by sea ice melt during summer.Seawater in the Chukchi Sea is derived from theBering Sea，with substantial nutrients to supportprimary production. Consequently，the Chukchi Shelfhas a signifi cant role as CO₂ sink during summer.4 CONCLUSION

In the present study，average dissolved ²¹⁰Pb atdifferent stations had an increasing trend northwardfrom 1.12 to 1.78 Bq/m³ during July and August，and average particulate ²¹⁰Pb 2.22 Bq/m³ was higher th and issolved ²¹⁰Pb，at 1.49 Bq/m³ . Residence time ofparticulate ²¹⁰Po was greater than that of dissolved ²¹⁰Po，indicating rapid biogeochemical cycling in the studyarea. The highest concentration of POC(601 μg C/L)was at station R03 during the fi rst sampling period，which decreased northward during the second samplingperiod，consistent with the distribution of chlorophyll-a .Average POC export fl uxes in the euphotic zone werederived by two methods with an irreversible scavengingmodel. These fl uxes had a decreasing trend northward and were comparable to prior work in the study area.This suggests an effi cient biological pump in the ArcticOcean and confi rms relatively high new production inthat ocean during summer，indicative of a signifi cantrole as CO₂ sink. The results may exp and knowledge ofthe carbon cycle in the western Arctic Ocean.5 ACKNOWLEDGMENT

We thank the captain and crew members of R/VXuelong and members of the oceanographic group foronboard assistance during CHINARE-4.