T h-234 Scavenging in the Water Column off Southwestern Taiwan

A station off southwestern Taiwan (22° 25'N 120° 08'E) was selected to carry out detailed seawater sampling for the measurements of dissolved and particulate 234Th. The profiles showed that 234Th is subject to scavenging and removal processes by biological particles in the euphotic zone and by resuspended particles in the deep layer. Estimated by box-model, the residence times of dissolved and particulate 234Th in the euphotic layer are in the ranges of 90-110 days and 12-28 days, respectively. A vertical 234Th flux of 1220 dpm m-2 d-1 out of the euphotic layer is estimated. Horizontal transport of resuspended particles in the deep and bottom layer is evident based on morphological examination of particles but is difficult to quantify. Resuspended particles act as an effective scavenger for Th in the deep layer and cause a radioactive disequilibrium of 234Th with respect to its progenitor, 238U. Residence times of dissolved and particulate 234Th in this layer are 112 days and 35 days, respectively. A lower limit of 234Th flux of 1600 dpm m-2 d-1 ihto the bottom sediment can be estimated.


INTRODUCTION 183
It is well-known that 'particle-reactive' elements are subject to scavenging onto parti cle surface and removed from the ocean via particle settling (Goldberg, 1954).• As radionuclides are .measured in both dissolved and particulate phases, we are able to estimate not only how fast the elements are scavenged but also the settling velocity of particulate matter in the water column.The power of radionuclides originates from its definite decay rate which provides a 'clock' for the processes which govern the distribu tion of elements in the ocean.
Compared to the open ocean, coastal waters have enhanced scavenging properties and more dynamic 'charact�risrics of:particle cycling,.,This is be .cause tei;restrial input of particulates and resuspended sed,iment� play a significant role in scavenging processes.In this paper, we investigated the �ca�enging processes of 234Th in a station about 22.5 km off Kaohsiung Harbor.We have determined detailed vertical.profiles of dissolved and particulate 234Th in the water column.Scavenging and particle removal rates will •be estimated using a box model calculation.
About 20• liters of seawater were collected using 20 l Go-Flo bottles mounted on hydrowire.Seawater was' immediately pressure-filtered by compressed air through a pre weighed • 142 mtn Nuclepore:filter (0.45 µm) mounted in a Plexiglas fi lter holder.
After filtration, the filter was rinsed with about 10 ml deionized.distilled water and stored in a petri dish and returned to the laboratory.Preconcentration and.separation of uranium and thorium from the filtered seawater samples were carried out on board ship.Detailed procedures for 234Th sample handling can be found in Wei andMurray (1991, 1992) and Wei and Hung (1992).During this cruise, parallel sample� fo� diss�lv�d and particulate �10Pb and 210Po determinations were alr;o collected.Particulates collected by Nuclepore filters.were examined' using the scanning.electron microscope (JSM�5400).The 210Pb, and 210Po :results along with particulate composition will bei reported in a separate paper (Wei et al., in prep.).
: The activities of 234Th were counted by a low background ( <0.3 cpm) anticoindden: ce counter (Tertnelec LB-5100) via its P-emitting daughter 234Pa.The purity of samples was checked by counting through a time span of about two months to confirm that the radioactive decay curve 'Was followed (Wei, 1991).Blank determinations of all reagents and fi lter paper used for this study indicated negligible contribution.Chemical : yield of thorium was •estimated by counting spiked 230Th using silicon.surface-barrier detectors '(EG&G Ortec 576).The counting efficiencies of the a detectors were calibrated against NIST traceable 241Am (Isotope Products Laboratory 387-67-2-2 ) and 230Th (Iso tope Products Laboratory 3 87-67-3) standard plates.Activities of 234Th reported here were corrected back to the sampling time after the ingrowth of 234Th from 238U was subtracted.

RESULTS
Depth, salinity, calculated 238U activities, concentrations of total suspended matter (TSM), dissolved (DTh) al1d p�iculate (PTh) 2 3 4T.h activities are g{ven in Table 1.Uncer tainties of all radioisotope data listed were estimated according to, the propagation of counting error (±lo} Hydrographic and nutrients data are not listed but are available upon request.(1977).
It is evident from interleaving salinity structure (Figure 1) that the water colunm is subject to complicated w�te • r mixing.A minimum of34.2psu and a maximum of 34.8 psu were found at the depth of 40 m and 125 m, respectively.Several interfingering structures were observed in the t.!pper 200 m.The salinity maximum is caused by intrusion of the Kuroshio (Shaw,19S9).Main pycnocline of0.05 G0 m-1 was in the layer of 95-125 m and a secondary pycnocline of -0.007 Ge m-1 was in the layer of 125-325 m.
The most commonly used model for 234Th scavenging in the water column of the ocean is the irreversible .
scavenging model (Coale andBruland, 1985, 1987;Wei and Murray, 199 1, 1992).As pointed out by Wei and Murray (1992), the scavenging and particle removal rates can be calcµlat�d either py a point-by-point approach or by dividing the water column into a number of boxes then carrying out a mass balance calculation.
Here we adopted the box model approach.The water column was divided into five boxes based on the vertical stability.The four boundaries between boxes were 20 m, I 00 m, 220 m, and 345 m.In each box, inventories of dissolved and particulate 234Th were calcu- Residence times for dissolved (-r0Th) and particulate (-rPT h) 214-'fh relative to scavenging and particulate fluxes. in each box can be calculated by : DTh. 'l"DThi=J The results of box model calculation are shown in Figure 5. (3) (4) J' 1.02 /' 0.17

Implication of particle dynamics in the euplwtic layer
Euphotic layer is a unique• system in which biological materials are produced, recy cled and then transported into the interior of the ocean.One of the primary goals of JGOFS (Joint Global Ocean Flux Study) is to quantify the export flux of carbon out of the euphotic zone.Th-234 has been identifi ed to be a superior tracer for particulate organic carbon (Eppley, 1989; U.S. JGOFS Report# 12, 1990) to estimate the export production.The rationale of this recommendation is based on the findings that 234Th scavenging responds to primary productivity (Coale and Bruland, 1985) and to organic particulate flux (Bruland and Coale, 1986).Furthermore, the residence time of POC with respect to its production rate in .theeuphoric layer is similar to that of 234Th, which makes 234Th a potential tracer for biological particles (Wei and Murray, 1992).Recently, Beals and Bruland (in press) found that dissolved 234Th residence time inversely correlates with fluorescence in the Northeast Pacific Ocean.All of these findings warrant obtaining more 234Th data on both spatial and temporal bases.
Nutrient measurements showed that the surface water at the site was a N03-limiting system (Figw .•e 2).N03 concentration in the mixed layer was below the detection limit (0.3 µM) then started to increase at the base of the mixed layer.Other nutrients (P04 and Si02) showed a similar increasing trend but the surface concentrations never diminished to detection limit values.At this specific region, the euphotic depth was at 100 m, about 75 m deeper than the mixed depth.Such euphotic system (euphoric depth> mixed depth) is commonly found in meso-and oligotrophic pelagic oceans and is called stratified two layer system (Dugdale, 1967).
The impact of the two-layer euphotic structure on 234Th scavenging was studied by Coale and Bruland (1987).In the mixed layer off Mexico, the productivity was fueled by regenerated nutrients, and planktonic community was dominated by pico-or nanoplanktons.In this oligotrophic mixed layer, 234Th was scavenged onto particles but not readily transported out of the mixed layer because the particles in it were small in size, and decomposition rate of partides might be relatively fast compared to the long residence time of particles.On the other hand, the lower euphoric layer was a lit system with relatively high nutrients.Thus, the productivity was mainly supported by nitrate and the planktons which resided in it were larger in size� The export fluxes of 234Th out of this layer were high because of fast scavenging and repackaging rates .
. ' At our sampling station, dissolved 234Th in the mixed layer were not significantly higher than• those in the lower euphoric layer.Particulate 234Th was higher in the mixed layer than those in the lower euphotic zone by �0.2 dpmkg-1• The sum of dissolved and particulate 234Th (total 234Th) between two layers was not as contrasted as the case found off Mexico• (Coale and Bruland, I 987).Lack of clear two-layered Th structure at the station implies that 234Th dynamics in the euphotic zone was not as simple as we thought.
A clear picture of trophic dynamics in the euphotic zone in different oceanographic re gimes must be known.
A subsurface chlorophyll maximum (SCM) was observed at the base of the mixed layer.At the SCM, dissolved 234Th showed evident minimum values and TSM concentra tions showed maximal values.The association of SCM and dissolved 234Th minimum was consistent with the case found in the Northeast Pacific Ocean (Beals and Bruland, in press), which confirmed the close relationship of 234Th scavenging and biological activi ties.Residence time and fluxes of 234Th in the euphoric layer can be calculated from the degree of disequilibrium and will be presented in the following section.

Implication fr om excessive 211 Th in the deep layers
It is noted that four (150,200,400, and 500 m) out of thirteen samples had excessive total 234Th relative to mu (Figure 4).Excessive 234Th relative to mu was also observed by others.In the profiles of Coale and Bruland (1987), excessive total 234Th over 238U was found below the euphotic layer at two stations in the North Pacific Ocean.In the Panama Basin, total 234Th activities exceed 238U at certain depths (Murray et al., 1989).However, the extent of excess was too small, usually in the range of uncertainty, to confirm if the excess signal was real.
Judging from the relatively large difference between total 234Th and mu at our sam pling station, we believe the excess signal is a real feature.The excessive 234Th over mu at aforementioned depths may be due to particle regeneration and/or retardation of particle settling.It can be seen from Figure 4, dissolved 234Th approaches 238U at these depths, implying that regeneration of 234Th is a slow process relative to the radioactive tum-over rate of234Th.Although representing a small pool (-10%), particulate phase plays a major role in the exchange of 234Th between different layers.It is interesting that the particulate 234Th above the 234Th-excess layers are usually smaller than those in other depths.Accord ingly, the layers with enriched 234Th may be caused by 'trapping' effect on settling parti cles originated from the overlying layers.

Importance of horizontal transport of resuspended particles
A signifi cant deficiency of 234Th with respect to mu by -0.5 dpm kg-1 was observed at depths of 250 and 300 m.In these depths, TSM concentration reached a maximum of 0.15 mg kg-1 (Figure 3).We believe that the disequilibrium of 234Th and mu was caused by horizontal injection of resuspended particles from the continental slope.Because of the short radioactive lifetime of 234Th, the tum-over rate of 234Th is relatively fast so that the disequilibrium below the euphotic layer is seldom found in the open ocean.An exception of this is coastal regime (i.e., Dabob Bay, Washington, USA, Wei and Murray, 1992) where the water depth is shallow and resuspension of bottom sediments is effective in removing 234Th from seawater.
Direct evidence of the horizontal transport of resuspended particles was detrital com position from SEM examination of filtered particles collected at these depths.Over 95% of particulates at these depths were detritus material (Wei et al., in prep).The specific activity of 234Th in the bottom sediments may be lower than that in the suspended parti cles, hence, acting as an effective Th scavenger when resuspended in the water column (Wei and Murray, 1992).Comparing with the other area, residence time of dissolved 234Th, 100 days, in the euphotic layer of the station is long.In the coastal California Current, the 'torh ranges from 6 to 47 days (Coale and Bruland, 1985).Recently, Wei and Hung (1992) found that 't'orh in the surface water of the Bashi Chann el and the Luzon Strait is also short, mostly shorter than 50 days.The•surface water of the Bashi Channel has the same source of water as the study area (Shaw, 1989).The abnormally long 't'orh in the area should be related to low local biological productivity.Unfortunately, there is no primary productivity data to con firm the interpretation.It is informative to compare the magnitudes of scavenging flux (J) and vertical flux (F).From boxes 1 to 5, the ratios of J and F are 1.81, 1.12, 0.8 1, 0.81, and 0. 31, respectively.The ratios indicate that most of the vertical 234Th fluxes in the euphotic boxes (box 1 and 2) originate from scavenging of dissolved 234Th onto particulate materials, while only a small proportion of vertical 234Th flux in the bottom box (box 5) is contrib uted from scavenging.

SUMMARY
A detailed study of dissolved and particulate 234Th in the water column of a station off southwest Taiwan has enabled us to assess the scavenging and particle removal processes.The water column under investigation has a stratified two-layer euphotic layer and a deep layer subjected to the influence of active horizontal input of resuspended sediments.Pronounced deficiencies of 234Th relative to 238U in the euphoric and deep layers were observed.On the other hand,• a significant amount of � xcessive 234Th was found below the euphoric zone and in the bottom water, which is caused by regeneration and/or trapping effect of settling particles.We adopted a box-model to estimate the residence times of dissolved and particulate 234Th.Furthermore, vertical fluxes of 234Th out of five compart ments were also estimated.
The vertical distributions of hydrographic data (salinity, temperature, and potential density), fluorescence and TSM concentration, and dissolved oxygen and nutrients (P04, N0 3 , N0 2 , and -Si02) at the sampling station are shown in Figures 1-3, respectively.Dissolved anctparticulate 234Th with counting error bars are shown in Figure 4. Total 234Th activities as the •sum of dissolved and particulate 234Th are also induded in the figure.A dotted line drawn at 2.39 dpm kg-1 •in the figure represents 238U activity calculated from the S-238U relationship from Ku et al.

Figure 4
Figure4shows that all dissolved 234Th except at 400 m were deficient relative to the secular equilibrium values.Maximal deviations of dissolved 234Th from 238U were found in the photic zone and in the�• d�ep TSM maximum layer.Particulate 234Th in the euphotic layer showed mirror image of dissolved 234Th.A minimum of particulate 234Th of only 0. 1 dpm kg-1 was located at S minimum layer.Total 23�Th profile was similar to dissolved profile because dissolved :form was the dominant phase of 234Th in seawater.

Fig. 4 .
Fig. 4. Vertical profiles of total, dissolved and particulate 234Th.The vertical dashed line at 2.39 dpm kg-• represents 238U activity calculated from S-238U relationship.Error bars repre sent uncertainties based on propagated counting error.The bases of mixed depth and euphotic zone are shown as the horizontal dashed and solid lines, respectively.Deficiencies and excesses of total 234Th relative to 238U are shown as hetched and grey area, respectively.

Fig. 5 .
Fig. 5. Results of box-model calculation from equations (1) to (4).Inventories of dissolved (blank box) and particulate (grey box) 234Th are written as outlined charactors in unit of 1000 dpm m-2• Numbers in bracket are residence times in day.Horizontal arrows into the dissolved (particulate) 234Th boxes are production (scavenging) fluxes from 238U (dissolved 234Th).Dashed arrows are fluxes of radioactive decay.Vertical arrows represent fluxes of 234Th via particle settling from overlying boxes.All fluxes.are in unit of 1000 dpm m-2 d-1•

5. 4
Inf erences fr om box-model The water column was divided into five boxes and can be envisioned as upper euphotic, lower euphotic, intermediate and two bottom boxes.Each box has its own characteristics which determine the scavenging and removal rates.For example, the upper euphotic box (box 1) is characterized by low nutrient, high phytoplankton population, high regenerated productivity, and low POC flux.The lower euphoric box (box 2) is character ized by high nutrients, lower phytoplankton population, high new productivity, and high POC flux.In the two bottom boxes (boxes 4 and 5) suspended particles are dominated by detritus materials which may impose a rather different scavenging force for 234Th.

Figure 5
Figure 5 summarized the fate of 234Th in the water column.The common feature shared by different boxes is that 'tDTh is always longer than 'tPTh• implying that uptake/ adsorption rate controls the non-radioactive sink for 234Th at the station.In eVident 234Th deficiency layers (box 1, 2, 4), t0Th is about 100 days while in other layers t0Th is too long to be resolved by the disequilibrium.tPTh ranges from 12 days associated with the lower euphotic box (box 2) to 183 days associated with the intermediate box (box 3), giving an average settling velocity from 0. 7 m d-1 to 6. 7 m d-1• Comparing with the other area, residence time of dissolved 234Th, 100 days, in the Vertical fluxes of 234Th estimated from the box-model range from 210 to 2580 dpm m-2 d-1, and generally increase with depth except the bottom box.It should be pointed out that the estimated 234Th fluxes from the two bottom boxes are lower limits.This is because a hori .zontal input of resuspended sediments, evident from dominant detrital materials on the filters (Wei et al., in prep), was not considered.If this component was known and included in the mass balance equation (eq.(2)), a higher vertical 234Th flux should be expected.Unfortunately, a quantitative determination of the input was difficult to make; hence, we put a question mark on the box and interpreted the results with caution.