Biogeosciences Insignificant enhancement of export flux in the highly productive subtropical front , east of New Zealand : a high resolution study of particle export fluxes based on 234 Th : 238 U disequilibria

We evaluated the export fluxes of Particulate Organic Carbon (POC) in the Subtropical Frontal zone (STF) of the SW Pacific sector of the Southern Ocean. The site is characterized by enhanced primary productivity, which has been suggested to be stimulated through so-called natural iron fertilization processes where iron-depleted subantarctic water (SAW) mixes with mesotrophic, iron-replete subtropical water (STW). We adopted the small-volume 234Th method to achieve the highest possible spatial sampling resolution in austral late autumn-early winter, MayJune, 2008. Inventories of chlorophylla, particulate234Th and POC observed in the upper 100 m were all elevated in the mid-salinity water type (34.5 < S< 34.8), compared with low salinity waters (S< 34.5) which were of SAW origin with high macronutrients and high (S > 34.8) salinity waters which were of STW origin with low macronutrients. However, Steady-State 234Th fluxes were similar across the salinity gradient being, 25 ± 0.78 ((1.5± 0.047)× 103) in the mid-salinity, and 29± 0.53 ((1.8± 0.032)×103) and 22± 1.1 Bq m−2 d−1 ((1.3± 0.066)× 103 dpm m−2 d−1) in the high and low salinity waters respectively. Bottle POC/Th ratios at the depth of 100 m were used to convert 234Th fluxes into POC export fluxes. The derived POC flux did not appear to be enhanced in mid-salinity waters where the primary productivity was inferred to be the highest at the time of sampling, with a flux of 11± 0.45 mmol C m−2 d−1, compared to 14± 0.39 mmol C m−2 d−1 in high salinity waters and 8.5± 0.66 mmol C m−2 d−1 in low salinity waters. This study thus implied that natural iron fertilization does not necessarily lead to an enhancement of POC export in STF regions.


Introduction
The Subtropical Front (STF) (Fig. 1a) is a circum-global oceanographic feature, typically between about 35 • S and 45 • S, where cold, high macro-nutrient, iron-limited subantarctic waters (SAW) mix with warm, low macro-nutrient, subtropical waters (STW) (Longhurst, 1998;Orsi et al., 1995).A number of studies have shown enhanced year-round chlorophyll concentrations and primary production (PP) in the STF region (Behrenfeld and Falkowski, 1997;Comiso et al., 1993;Murphy et al., 2001).The observed PP in the STF, east of New Zealand, can be as high as 22 mmol C m −2 d −1 in winter, which may be more than 4 and 2-times higher than in the adjacent SAW and STW, respectively (Bradford-Grieve et al., 1997).Similarly elevated integrated production rates have been observed in the STF around 152 • E off Australia in summer (Clementson et al., 1998) and in the South African sector at 20 • E in winter (Froneman et al., 1999).Such enhancement in PP in the STF zone has been suggested to be induced by natural iron fertilization processes, with the iron sourced from atmospheric deposition, shelf boundary exchange processes and/or mixing with iron-replete subtropical waters (Boyd et al., 1999(Boyd et al., , 2004;;Pollard et al., 2009).(Nodder, 1997) (b); the location of sampling sites during the May-June 2008 research cruise TAN0806 (c); and the stations with salinities >34.8 (high), 34.5-34.8 (mid), and <34.5 (low) marked in black, red, and blue, respectively.The blue line represents the transect where stations were sampled throughout the water column (Transect M).The bathymetry of the study area is also shown to emphasize the location of the Chatham Rise.
iron levels (>0.2 nmol l −1 ), and low macronutrient concentrations (e.g.PO 4 <0.3 µmol l −1 ), to high nutrient-low chlorophyll (HNLC) SAW to the south, with typically low temperatures (summer <14 • C; winter <10 • C), low salinities (<34.6) and low iron levels (<0.1 nmol l −1 ), but high macronutrient concentrations (e.g.PO 4 >0.9 µmol l −1 ) (Boyd et al., 1999;Nodder, 1997).Regardless of seasonal and spatial variations, phytoplankton biomass and biological productivity are generally elevated over the Chatham Rise (Bradford-Grieve et al., 1997, 1999;Gall et al., 1999).In winter and spring, average chlorophyll a (Chla) concentrations, integrated down to 100 m, can reach up to ∼80 mg m −2 , and the corresponding PP, integrated to 1 % light level, can be more than 50 mmol C m −2 d −1 (Bradford-Grieve et al., 1997).In comparison, the average inventory of Chl-a and PP may be only 13 mg m −2 and 12 mmol C m −2 d −1 , respectively, in SAW and 33 mg m −2 and 46 mmol C m −2 d −1 in STW.Such perennially high PP levels in the STF support a diverse planktonic and benthic ecosystem (Probert and McKnight, 1993;Bradford-Grieve et al., 1999) that seems to be translated to many fish species, with New Zealand's richest deep-water fisheries, primarily in blue grenadier (known locally as hoki, Macruronus novaezelandiae) and orange roughy (Hoplostethus atlanticus), occurring on the Chatham Rise (New Zealand Ministry of Fisheries, 2009).However, little is known thus far about the spatio-temporal variability and magnitude of the downward export of particulate organic carbon (POC), or the fate of the PP in this frontal zone.Nodder (1997) used free-floating cylindrical sediment traps in the vicinity of the STF showing that the POC flux was less than 3 mmol C m −2 d −1 , which is in the same order of magnitude as that found in low export oligotrophic oceans.Indirect evidence from benthic studies, however, indicates that organic fluxes are enhanced on the crest and southern flank of the Chatham Rise (Probert and McKnight, 1993;Nodder et al., 2003).It is thus clear that information on the spatial variation of POC export within the STF is required if we are to evaluate whether the enhanced primary production in the frontal zone also leads to enhanced export production.
In the present study, we utilized a particle-reactive radionuclide, 234 Th (t 1/2 = 24.1 d), as a tracer for particle export from the upper ocean.The technique has been applied widely in many oceanographic settings to examine processes occurring on time-scales of days to weeks (e.g.Coale andBruland, 1985, 1987;Buesseler, 1992Buesseler, , 1998;;Cochran and Masque, 2003;Dai and Benitez-Nelson, 2001;Waples et al., 2006).A further technological advantage is the use of the recently developed small-volume technique that enables high resolution sampling (Benitez-Nelson et al., 2001a;Buesseler et al., 2001a;Cai et al., 2006a).This method is essential in order to capture the particle dynamics and export flux variability in such regions as the STF, which are characterized by dynamic hydrography (e.g., Cai et al. 2008;Buesseler et al., 2009).Indeed, with high resolution sampling, Cai et al. (2008) observed significant variations of 234 Th deficit in the South China Sea, with enhanced Th and POC fluxes along the western and southern boundaries of this marginal sea.Buesseler et al. (2009) even observed high spatial variability in the northern subtropical gyre at ALOHA station off Hawaii, where a homogenous spatial distribution of POC flux from the surface ocean would have been expected.
The present study aims to examine the spatial distribution, magnitude and variability of upper ocean POC export in the STF over the Chatham Rise to the east of New Zealand by using high spatial resolution sampling of 234 Th.We show that 234 Th-based POC export is not significantly enhanced in the frontal zone, despite high Chl-a and inferred PP levels, compared to the adjacent SAW and STW.

Sample collection
Samples were collected in late austral autumn-early winter from 23 May to 12 June in 2008 on board R/V Tangaroa, operated by the National Institute of Water and Atmospheric Research (NIWA) Ltd, New Zealand (NIWA voyage TAN0806).High spatial resolution sampling at 23 stations, covering a surface area of about 25 000 km 2 , was conducted during the cruise (Fig. 1).Water samples were collected using 10 l Niskin bottles mounted on a rosette sampler attached to a Seabird SBE9/11plus conductivity-temperature-depth (CTD) sensor.A sub-sample of 4 l seawater was used to determine the total 234 Th activity (see below) and another 8 l was filtered onto a 25-mm diameter Quartz Microfiber (QMA, norminal pore size ∼1.0 µm) for particulate 234 Th and POC measurements.Samples were collected at 5 depths in the upper 100 m (normally 10, 20, 50, 70, and 100 m), except along transect M where sampling was conducted at a finer depth resolution throughout the water column to better define the vertical structure of 234 Th over the Chatham Rise (shown as a blue line in Fig. 1).Since this was the first 234 Th study over the Chatham Rise, such an intensive sampling strategy enabled a robust description of both the vertical and spatial distribution of 234 Th.

234 Th analysis
We used the small-volume MnO 2 co-precipitation technique for our total 234 Th analyses, as initially developed by Benitez-Nelson et al. (2001a) and Buesseler et al. (2001a), and further modified by Cai et al. (2006a).Four litre seawater samples were immediately acidified with 6 ml concentrated HNO 3 and spiked with ∼166.7 mBq 230 Th after collection.The samples were then mixed vigorously and allowed to stand for 12 h for isotopic equilibration.The pH was then brought up to 8.00-8.20 and thorium isotopes were co-precipitated with MnO 2 by adding 0.25 ml KMnO 4 (3.0 g l −1 ) and 0.25 ml MnCl 2 (8.0 g MnCl 2 q 4H 2 O l −1 ).The formation of the MnO 2 precipitate was accelerated by heating to approximately 90 • C in a water bath for 2 h.The precipitate was then filtered onto a 25 mm QMA filter after samples were cooled to room temperature.The QMA filter was baked at about 100-200 • C until dryness, and then mounted under a layer of Mylar and two layers of aluminum foil (total density ∼7.2 mg m −2 ) for beta-counting at sea by a gas-flow proportional low-level RISØ beta-counter (Model GM-25-5, RISØ National Laboratory, Denmark).These samples were re-counted for background levels 6 months after the cruise (i.e., >5 half-lives of 234 Th).
The QMA filter used for particulate 234 Th determination was dried at 50 • C in an oven for 24 h, and then mounted and beta-counted as above for the total 234 Th.The average background levels for total and particulate 234 Th were 0.45 and 0.32 counts per minute (cpm), respectively.
We used an alpha spectrometric method for our total 234 Th recovery analysis (Cai et al., 2006a).Samples were demounted after background counting, spiked with ∼166.7 mBq 228 Th ( 232 U-228 Th solution) and then digested by adding 10 ml concentrated HNO 3 , 1 ml HF and 1 ml H 2 O 2 .Thorium isotopes were purified through iron precipitation and a classic anion column.Eluents were then evaporated onto a 25 mm stainless steel disc after extraction using 1.5 ml of 0.25 mol l −1 theonyl trifluoroacetone(TTA)/benzene solution.The disc was counted by alpha spectrometry (Octete TM PC) until both thorium isotopes ( 230 Th and 228 Th) reached more than 2,500 counts.Most of the final recoveries for 230 Th were between 80 % and 103 %.The average of the recoveries was 90.2 ± 1.4 % (mean ± 1 standard deviation).The errors associated with 234 Th activity determination are propagated from the counting errors on the first counts, background measurements and recovery analyses.The precision of the final 234 Th activity was better than 5 %.
Instead of undertaking specific sample analyses, we used the relationship: A U (mBq l −1 ) = 1.1801 × salinity to estimate the 238 U activity (A U ) in the seawater according to Chen et al. (1986).The uncertainty of 238 U activity was ∼3 % which was also included in the error estimates associated with the 234 Th fluxes.

POC and Particulate Nitrogen (PN) analysis
After beta-counting for particulate 234 Th, the samples were demounted for POC and PN analysis.The filters were placed in Petri dishes and fumed using concentrated hydrochloric acid for 24 h to remove the carbonate phase.POC concentrations were then determined by a PE-2400 SERIES II CHNS/O analyzer, according to JGOFS protocols (Knap et al., 1996).Replicate procedural C blanks from sampling to instrumental carbon determination have been tested before (Chen, 2008), and were all less than 6 µg C and 2 µg N, which typically accounted for less than 10 % of the sample POC and PN, respectively.The precision for our POC measurements were always better than 10 % (Cai et al., 2006a;Chen, 2008).

Other ancillary parameters
The temperature and salinity vertical profile data were obtained from the Seabird SBE9/11plus CTD.A Wetlabs fluorometer, interfaced with the CTD, was used to determine the fluorescence at an excitation and emission wavelength of 470 and 695 nm.In order to calibrate the fluorometer, discrete seawater samples (1-2 l) were collected for the measurement of Chl-a.Briefly, seawater was filtered through a Whatman GF/F filter, and Chl-a on the filter was then extracted with 90 % acetone and analyzed using a spectrofluorometer.A linear relationship was found between fluorescence and Chl-a (Chl-a (µg l −1 ) = 0.639 × Fluorescence, R 2 = 0.87, n = 50), which was then applied to convert the fluorescence values into Chl-a concentrations.Uncertainties for the Chl-a determination were <10 %.

Hydrography
The complex hydrological characteristics of the STF are shown by the relationship between potential temperature and salinity in Fig. 2. In the upper ocean, distinctly different water characteristics were found between stations, representing different degrees of mixing.The surface temperature and salinity changed dramatically over the sampled area from lows of 9.4 • C and 34.2, respectively, in surface waters with SAW affinities, to highs of 15.2 • C and 35.2 in surface waters associated with STW.
Figure 3 shows the spatial distributions of temperature and salinity at 2 m and 100 m water depths.The gradients of salinity and temperature were much greater in the region between 43 • S and 44 • S, coinciding with the crest and upper southern flank of the Chatham Rise.Vertically, the surface 150 m of all stations sampled in the STF was well stratified (Fig. 4a, b, and c).However, the depth of the surface isothermal mixed-layer, defined by the density surface with a 0.5 • C temperature difference from the reference depth (Kara et al., 2000), was not uniform between stations, ranging from 54 m to 178 m.Typically and as illustrated in Fig. 4, the depth of the mixed-layer at high salinity stations was deeper than at mid-and low salinity stations (see next section for definitions of water types).

Chl-a, macronutrient, and POC distribution
Vertically, Chl-a was highest within the upper mixed-layer.Below the mixed-layer, the Chl-a concentration decreased to background, effectively zero (Fig. 4a, b).Regionally, Chl-a concentrations were elevated where salinities were between 34.5 and 34.8 (Fig. 2b).In such cases, the average Chl-a concentration in the upper 100 m was as high as 0.79 µg l −1 compared to 0.40 µg l −1 at salinities >34.8 or 0.37 µg l −1 at salinities <34.5.Accordingly, for the ease of discussion, we have divided the STF waters of our study area into three different water types: low (S < 34.8), mid-(34.5 < S < 34.8) and high salinity (S > 34.8), according to the vertical and horizontal distribution of Chl-a.
Both DIN and PO 4 concentrations were generally at replete levels during the cruise.In the upper 100 m, DIN concentrations ranged from 3.6 µmol l −1 to 13.9 µmol l −1 in mid-salinity waters, with an average of 8.9 µmol l −1 .In high salinity waters, it was lowest, ranging from 1.1 µmol l −1 to 12.3 µmol l −1 , with an average of 4.7 µmol l −1 .In low salinity waters, DIN concentration varied from 7.8 µmol l −1 to 14.9 µmol l −1 , with an average of 11.3 µmol l −1 , which was the highest among the three water types.PO 4 shared the similar distribution pattern with that of DIN.In mid salinity waters, the average of PO 4 concentration in the upper 100 m was 0.71 µmol l −1 , compared to 0.42 µmol l −1 in high salinity waters and 0.87 µmol l −1 in low salinity waters.Si(OH) 4 was only measured at stations C1 through C7 covering high and mid-salinity waters.Based on the limited Si(OH) 4 data set, its concentration in the upper 100 m varied from 0.39 to 3.43 µmol l −1 in mid-salinity stations (C5 and C6), which is comparable to the range in the high salinity stations (C1, C2, C3, C4 and C7, 0.91-3.11µmol l −1 ).Inventories of these macronutrients are listed in Table 2.
POC concentrations varied from 0.40 µmol C l −1 to 6.1 µmol C l −1 across the study area (Table 1).In the upper mixed-layer, POC distributions generally followed that of Chl-a, indicating a relationship between POC and phytoplankton biomass.Regionally, POC concentration was also relatively enhanced in mid-salinity waters.The average POC concentration in the upper 100 m was 3.9 µmol C l −1 in this water type, compared to 2.4 µmol C l −1 and 2.6 µmol C l −1 in high and low salinity waters, respectively.In a departure  from the vertical distribution of Chl-a, POC concentrations were often higher near the bottom of the water column, which may be indicative of near-bottom sediment resuspension (e.g., Nodder, 1997;Nodder et al., 2007).
The C/N ratio was quite stable in the upper ocean, ranging from 5.2 to 8.5, with an average of 6.6 ± 1.5 (n = 146) (mean ± 1 standard deviation), which is almost identical to the Redfield ratio of 6.63 (Redfield et al., 1963).No obvious changes in C/N ratio were found in association with changes in salinity, suggesting that the particles in the study area were predominantly biogenic in origin.

Vertical profiles of 234 Th
Particulate and total 234 Th activities are listed in Table 1.
Profiles of three representative stations from low, mid-and high salinity waters are shown in Fig. 4. The activities of total 234 Th varied from 24 ± 1.0 mBq l −1 to 42 ± 2.4 mBq l −1 .Particulate 234 Th activities on suspended particles were variable, ranging from 1.7 ± 0.15 mBq l −1 to 12 ± 0.29 mBq l −1 .The 234 Th activity profile in the upper 100 m was related to the vertical distribution of Chl-a.As illustrated in Fig. 4, total 234 Th activity was lowest within the isothermal mixed layer.Below the mixed-layer, the activity of total 234 Th began to increase as the Chl-a leveled off, until it reached secular equilibrium with 238 U at the base of the euphotic zone (Ez) where Chl-a reached its minimum.At stations with depths of Ez greater than 100 m, the total 234 Th activity was in deficit with respect to 238 U from 0-100 m.The vertical distribution of particulate 234 Th also generally resembled that of Chl-a and POC, in that its activity was homogenous within and decreased below the mixed-layer.Not surprisingly, higher particulate 234 Th activities were related to higher levels of Chl-a, indicating intensive scavenging of biogenic particles in high Chl-a waters.However, total 234 Th activities seemed to be independent of Chl-a and/or biomass.As shown in Fig. 4, for example, the Chl-a concentration at the mid-salinity station C15 was twice at the low salinity station C17, but the activities of total 234 Th were similar at each station.

Meridional distribution of 234 Th
To better describe the spatial distribution pattern of 234 Th in the STF across the Chatham Rise, transect M, consisting of stations C5, C6, C7, C17, C18, and C20, was sampled throughout the water column (shown in Fig. 5).Generally, total 234 Th was in deficit in upper and near-bottom waters, and was in equilibrium with 238 U over the mid-water column.
No meridional trends were found in the upper water column along transect M. Particulate 234 Th was higher within the Ez and in bottom waters than in the mid-water column.Interestingly, at mid-depths of 300-600 m, a deficit of total 234 Th was found in waters on both sides of the Chatham Rise crest.This observation may reflect the horizontal transport of the re-suspended particles along the Chatham Rise flanks, consistent with the near-bottom increases in particle fluxes from sediment trap (Nodder, 1997;Nodder and Northcote, 2001) and benthic studies (Nodder et al., 2007) in the area.

Regional distribution of 234 Th
To provide a composite view of the regional distribution of 234 Th in relation to different water types, the total and particulate 234 Th activities are superimposed on a T/S diagram (Fig. 6).Similar to Chl-a and POC, the activities of particulate 234 Th in the upper ocean were enhanced in mid-salinity waters, where the average of particulate 234 Th activities increased up to 7.5 ± 0.063 mBq l −1 , which was two times higher than in the high salinity waters (3.3 ± 0.035 mBq l −1 ).Noticeably, however, high activities of particulate 234 Th were not necessarily correlated with low total 234 Th activities.Such a distribution pattern also held for total 234 Th among the three water types.The activities of total 234 Th in mid-salinity waters varied from 24 ± 3.0 mBq l −1 to 42 ± 2.4 mBq l −1 , with an average of 33 ± 0.23 mBq l −1 , and similarly from 23 ± 0.94 mBq l −1 to 41 ± 1.3 mBq l −1 in the high salinity waters, with an average of 32 ± 0.16 mBq l −1 .In comparison, total 234 Th activity ranged from 27 ± 1.0 to 39 ± 1.3 mBq l −1 in low salinity waters, with an average of 34 ± 0.32 mBq l −1 .

234 Th export flux estimates
The export flux of 234 Th at a specific depth horizon can be estimated by using the following equation: where δA Th δt is the rate of change of total 234 Th activity, A U is the 238 U activity estimated from the U-S relationship (Chen et al., 1986), A Th is the total 234 Th activity, λ Th is the decay constant for 234 Th (0.02876 d −1 ), P is the net removal flux of 234 Th on particles, and V is the sum of contributions from advection and diffusion.b For those stations where Ez > 100 m, 234 Th was assumed to be in equilibrium with dissolved 238 U at the base of the Ez as most of the stations were only sampled in the upper 100 m, and 234 Th flux was then integrated down to the base of the Ez.c Bottle C/Th ratios@100 m are used in these calculations, such that the estimated POC fluxes should be considered as upper limits for the Chatham Rise region at the time of sampling.The Steady State (SS) model is applicable when little temporal change ( δA Th δt ) occurs in 234 Th activities or SS 234 Th flux is low (Savoye et al., 2006).However, when there are rapid changes in 234 Th activities, for example, during algal blooms or within physically dynamic regions, such as the STF, a non-steady state (NSS) 234 Th flux model is generally necessary (Buesseler et al., 1992).To implement a NSS 234 Th flux model, however, time-series observations from the same water mass are needed.In a practical sense, since there are difficulties in tracing specific water masses in the ocean, most studies have adopted a protocol to visit the same station at least twice during a particular study period (Benitez-Nelson et al., 2001b;Coppola et al., 2005;Kawakami and Honda, 2007).NSS 234 Th flux can then be calculated using the following equation (Buesseler et al., 1992): where P NSS is the NSS 234 Th export flux, and A Th1 and A Th2 are the 234 Th activities during the first and second visit, respectively.During our cruise, two stations (C3 and C4) were visited twice.The profiles of the temperature, salinity and 234 Th activity for these stations are shown in Fig. 7.At C3, differences in temperature and salinity of the upper 100 m between the two visits were 0.5 • C and 0.02, respectively.
In contrast, at C4, larger differences were found: 2.1 • C for temperature and 0.37 for salinity, which clearly indicated that different water masses were present at this location between the two visits.Due to the relatively minor changes in hydrography, however, we assume the same water mass was Physical processes may also influence the estimates of downward 234 Th fluxes.Previous studies indicate that currents over the Chatham Rise can be strong, but variable, with alternating zones of convergence and divergence, although net zonal flows dominate along the northern and southern edges of Chatham Rise and predominantly meridional flows occur over the rise itself (Chiswell, 1996;Sutton, 2001).Given the little change observed in total 234 Th activities along the salinity gradients as shown in Fig. 6, however, we assumed that the horizontal contribution to the 234 Th fluxes was small compared to the downward vertical component of 234 Th export.Therefore, the V term in Eq. ( 3) can be neglected in our 234 Th flux estimates.
Based on the above discussion, the P term in Eq. ( 1) can then be solved as follows: In our case, the steady state (SS) downward flux of total 234 Th from the depth horizon of 100 m can be integrated by: The calculated 234 Th flux results for the upper 100 m of the water column are listed in Table 2. SS 234

Bottle POC/ 234 Th ratios
Bottle POC/ 234 Th ratios are listed in Table 1.This ratio was quite variable, ranging from 0.15 mmol C Bq −1 to 1.7 mmol C Bq −1 .All POC/ 234 Th ratios were separated into three groups based on the previously defined salinity criteria, as shown in Fig. 8. Consistent with many prior studies, POC/ 234 Th was higher and more variable in the upper ocean, compared to the deep ocean (Buesseler et al., 2006).Interestingly, the ratio was generally lower in mid-and low salinity waters than in high salinity waters, reflecting different biological effects on carbon and thorium partitioning.This difference disappeared at and below 100 m.At the export horizon of 100 m, bottle POC/ 234 Th ratios varied from 0.27 ± 0.030 mmol C Bq −1 to 0.82 ± 0.11 mmol C Bq −1 , with an average of 0.46 ± 0.013 mmol C Bq −1 (Table 1).Similar to 234 Th fluxes, no difference in POC/ 234 Th ratio was found among the three water types.In high salinity waters, the POC/ 234 Th ratio ranged from 0.27 ± 0.030 mmol C Bq −1 to 0.82 ± 0.11 mmol C Bq −1 , with an average of 0.48 ± 0.018 mmol C Bq −1 .In midsalinity waters, it varied from 0.31 ± 0.038 mmol C Bq −1 to 0.66 ± 0.090 mmol C Bq −1 , with an average of 0.66 ± 0.022 mmol C Bq −1 .

POC/ 234 Th ratios and POC export flux
POC export fluxes can be estimated by multiplying 234 Th fluxes by the POC/ 234 Th ratios of sinking particles (Buesseler et al., 1992).To better constrain the POC export from the upper ocean using the 234 Th method, knowledge of the POC/ 234 Th ratios in sinking particles should be examined.However, in the present study, only bottle POC/ 234 Th data were collected during the cruise.
Many studies have shown that the bottle POC/ 234 Th ratio is typically higher than in pump and trap samples (Buesseler et al., 2006;Cai et al., 2008;Kawakami and Honda, 2007).Although bottle POC/ 234 Th ratios from particles in bottle samples are not expected to be particularly representative of sinking particles, it is reasonable to apply these measurements as an upper limit for the actual POC/ 234 Th ratio in sinking particles (Cai et al., 2008).A major advantage of this approach is that bottle filtrations enabled us to undertake high spatial resolution sampling of POC/ 234 Th, compared to the spatially limited deployment of in situ pumps and sediment traps.
Due to the lack of an effective methodology for determining downward POC export in such a physically dynamic area, few similar studies have been carried out in the STF over the Chatham Rise or even in other STF zones globally.Nodder (1997) attempted to test the hypothesis that the STF is a region of elevated export production in austral autumn via the deployment of a free-floating, surfacetethered, cylindrical sediment trap array on the northern side of the Chatham Rise.Mean POC fluxes were estimated to be ∼2.5 mmol C m −2 d −1 at 200 m water depth, which is similar in magnitude to other locations in oligotrophic waters.Nodder and Alexander (1998) showed that spring particulate phosphorus fluxes at ∼100 m in the STF were almost double those in winter 1993.These earlier studies, however, were limited in their spatial coverage, which restricts the reliability of these results when extrapolated to the entire STF.Long-term moored sediment traps were deployed for a year at 300 m and 1000 m in the STF on the flanks of the Chatham Rise to investigate the seasonal variation of particle fluxes (Nodder and Northcote, 2001).POC export reached its maximum in spring (14.8 mmol C m −2 d −1 at 300 m in the North and 4.8 mmol C m −2 d −1 in the South), and annual average POC fluxes were calculated to be 10.1 mmol C m −2 d −1 in the North, compared to 4.1 mmol C m −2 d −1 in the South.POC export was also determined in frontal zones (including the STF and Polar Front) in the Indian Ocean sector of the Southern Ocean in summer using 234 Th methods (Coppola et al., 2005).At an export depth horizon of 100 m, POC export varied from 0.1 to 2.5 mmol C m −2 d −1 , which is very low compared to other observations in the Southern Ocean (Buesseler et al., 2001b;Rutgers van der Loeff et al., 1997, 2002).The low diatom dominance in the water column was hypothesized to result in these lower than expected export fluxes.However, direct comparisons between these studies are difficult due to differences in the regional oceanography, observation duration, the chosen export depth horizon and applied methodologies.The export fluxes presented here for the STF on the Chatham Rise, are also only applicable to the season in which they were collected (i.e. late autumnearly winter), and it is expected that POC fluxes in the highly productive spring might be significantly enhanced in comparison (Nodder and Alexander, 1998;Nodder and Northcote, 2001).Nevertheless, the calculated POC fluxes using the 234 Th method are at least similar in magnitude to previous measurements of POC export in the same area, and provide us with the first appreciation of the substantial degree of spatial variability that could be expected in POC flux to the seafloor on the Chatham Rise.The POC flux estimated at 100 m in the present study decreased from west to east and from north to south across the rise (Fig. 9f).Benthic biomass and activity is generally higher on the southern flank of the Chatham Rise, compared to the crest and northern flank (Nodder et al., 2003(Nodder et al., , 2007;;Probert and McKnight, 1993), which is somewhat incongruous with the lower POC fluxes suggested by this study and previous sediment trap results (Nodder and Northcote, 2001).Thus, it might be the quality, rather than the quantity, of POC supply to the benthos that has the most influence in structuring seafloor communities in this region (Nodder et al., 2003(Nodder et al., , 2007)).T -tests with unequal variance were carried out to compare the differences between the mid-and high salinity waters, and showed that P values (α = 0.05) for Chl-a, POC and particulate 234 Th were 0.0031, 0.0021, and 0.000014, respectively (Table 3).P values would be 0.00053, 0.020, and 0.030, respectively for the same comparisons between mid-and low salinity waters.Such analyses indicate that the mid-salinity waters were statistically different from the other two water types.Note that the biological enhancement observed in the mid-salinity waters was also observed for primary production measured in a parallel study on the same cruise (Jill Schwarz, NIWA, personal communication).It is interesting that the enhancement of particulate 234 Th activity in mid-salinity waters was much stronger than for Chl-a and POC.
The high Chl-a biomass and/or PP levels in the STF have been well-defined in previous studies.For example, phytoplankton biomass in winter and spring were 4 and 6 times higher, respectively, in the STF than in adjacent low salinity SAW (Hall et al., 1999).The mean PP rate in winter could be as high as 22 mmol C m −2 d −1 , which was 4 times higher than that in the SAW (Bradford-Grieve et al., 1999).Remotesensing satellite data also show that the STF is characterized by year-round heightened pigment concentrations (Comiso et al., 1993;Murphy et al., 2001).
As described above, both DIN and PO 4 were replete.In comparison, Si(OH) 4 concentrations were in the range 0.39-3.4µmol l −1 , which might indicate inhibition of diatom growth according to Chang and Gall (1998).Nevertheless, we tend to believe that Si(OH) 4 was not yet limiting the PP because PP was significantly high at mid-salinity as compared to high salinity water regime despite the Si(OH) 4 concentrations being similar at these sites.
Moreover, other studies have proposed that the elevated iron levels caused by mixing induced by the shallow bathymetry of Chatham Rise and/or the atmospheric deposition from Australian dust may lead to such high biomass/PP levels in the STF to the east of New Zealand, a process frequently termed "natural iron fertilization".Indeed, Boyd et al. (2004) determined the dissolved iron concentrations and potential iron sources in a transect across the STF.Dissolved iron concentrations in frontal surface waters reached the highest values of 0.8 nmol l −1 (above the stress level of 0.2 nmol l −1 ) at about 43 • S, which is coincident geographically with the crest of the Chatham Rise.Boyd et al. (2004) further noted that iron concentrations dropped dramatically to less than 0.2 nmol l −1 within 1 degree of latitude to the north of this location.
In contrast to the enhancement of PP within the STF, little difference in POC and 234 Th fluxes were found among the three water types identified in the present study.A t-test comparing the POC and 234 Th fluxes at the 100m export horizon between high and mid-salinity waters resulted in P values (α = 0.05) of 0.097 and 0.10, respectively (Table 3).The same parameters were compared between the mid-and low salinity waters, and similar observations were apparent (see Table 3).These relationships still held when all parameters were integrated to the bottom of the Ez (instead of to a fixed norminal depth such as 100 m), which is not surprising given the fact that sinking particles originate mainly within the Ez and shallow re-mineralization has frequently been found just below this depth (e.g.Buesseler and Boyd, 2009).Thus, there was no difference in POC and 234 Th fluxes at 100 m and/or the base of the Ez between the three water types.
The reasons why the elevated PP levels in the STF frontal zone, especially in our mid-salinity waters, did not lead to an increase in POC export are still unclear.Note that our study area is characterized by abundant diatom production in most seasons (Boyd et al., 1999;Bradford-Grieve et al., 1997), which should have driven high POC export fluxes as in many oceanic regimes (Michaels and Silver, 1988).Therefore we suggest that there are other factors limiting POC export here.Similar scenarios of this decoupling between PP and POC export were also observed in most of the artificial iron fertilization experiments (see reviews by Boyd et al., 2007), and the limited export response was attributed to the consequence of complex functioning of the planktonic community structure (Buesseler et al., 2004) and/or bacterial re-mineralization and grazing pressure (dominance of microzooplankton grazing over mesozooplankton) (Boyd and Newton, 1995).In the present study, the most plausible reason causing the little enhancement of the export flux would be related to microzooplankton grazing.Indeed, Hall et al. (1999) demonstrated that >78 % of daily PP can be grazed by microzooplankton in the STF in austral spring and winter.In contrast, mesozooplankton grazing is likely to be in the order of only 1-2 % of daily PP (Bradford-Grieve et al., 1998).Unlike mesozooplanton, the fecal pellets produced by microzooplankton are smaller and readily remineralized in the upper ocean and may not contribute significantly to export flux (Boyd and Newton, 1995;Michaels and Silver, 1988), as also suggested by other studies in the STF region (Nodder and Gall, 1998;Zeldis et al., 2002).

Conclusions
The present study applied a high resolution 234 Th sampling technique to define the magnitude and distribution of POC export in the STF region, which revealed with greater confidence that the POC export fluxes were on the order of 5.6 ± 0.81 to 30 ± 3.7 mmol C m −2 d −1 , with an overall average of 12 ± 0.41 mmol C m −2 d −1 (n = 25).There was little spatial variation among low, mid-and high salinity waters within the STF in austral autumn-winter, despite differences in biological particle production, as inferred from fluorescence/Chl-a profiles.The present study, on the other hand, confirmed that the STF region is characterized by elevated PP, in particular, in the mid-salinity waters (34.5 < S < 34.8), presumably stimulated by so-called natural iron fertilization processes (Boyd et al., 1999(Boyd et al., , 2004;;Pollard et al., 2009).
The present study, therefore, implies that natural iron fertilization does not necessarily lead to the enhancement of POC export in STF regions.It must be pointed out that, compared to other natural/ artificial iron experiments (Boyd et al., 2007), the present study was carried out in a different season (late autumn-early winter cf.summer) and latitude (43-44 • cf.>50-60 • ).Therefore, we anticipate that variations in temperature, latitude, season and oceanographic region will induce different ecosystem responses.

Fig. 1 .
Fig. 1.Hemispherical map of the Southern Ocean showing the approximate location of the Subtropical Front as a red line (a), with the map of the study area showing the mean circulation and surface water masses (Nodder, 1997) (b); the location of sampling sites during the May-June 2008 research cruise TAN0806(c); and the stations with salinities >34.8 (high), 34.5-34.8(mid), and <34.5 (low) marked in black, red, and blue, respectively.The blue line represents the transect where stations were sampled throughout the water column (Transect M).The bathymetry of the study area is also shown to emphasize the location of the Chatham Rise. Fig.2

Table 2 .
Inventories of dissolved inorganic nitrogen (DIN), Si(OH) 4 , particulate organic carbon (POC), particulate 234 Th and Chl-a, Steady-State (SS) 234 Th flux, POC/ 234 Th ratios and POC flux in the upper 100 m and euphotic zone (Ez) within the Subtropical Front in May-June 2008.Bq m −2 d −1 Bq m −2 d −1 mmol C Bq −1 mmol C m −2 d −1 determined a The depth of euphotic zone (Ez) is estimated to be where fluorescence reaches its minimum.

Fig. 8 .
Fig.8.Profiles of bottled POC/Th ratios for the three water types identified by their salinity differences (see Fig.2).

Table 1 .
Temperature, Salinity, Particulate and Total 234 Th activities, 238 U activities, 234 Th: 238 U and POC: 234 Th ratios for all stations in the Subtropical Front, Chatham Rise, New Zealand, measured in May-June 2008 (NIWA cruise TAN0806).

Table 3 .
P values derived from t-tests with unequal variance on inventories of POC, Particulate 234 Th and Chl-a, Steady-State (SS) 234 Th fluxes and POC fluxes at 100 m and euphotic zone (Ez) water depths for mid-vs.high salinity water and mid-vs.low salinity water.