Limnol. Oceanogr., 37(2), 1992, 296-314
0 1992, by the American Society of Limnology and Oceanography, Inc.
Temporal variations of 234Thactivity in the water column of
Dabob Bay: Particle scavenging
Ching-Ling Wei’ and James W. Murray
School of #Oceanography,WB-10, University of Washington, Seattle 98 195
Monthly measurements of total, dissolved, and particulate 234Th,along with fluxes of total mass
and particulate 234Th, were made throughout 1987 in Dabob Bay, Washington. Vertical distri-
butions of total 234Thshowed a large deficiency relative to 238Uthroughout the water column during
all sampling periods.
Total 234Thactivities in surface waters were generally bigher in winter and lower in summer, in
response to seasonal biological activity. In spite of this variation, the steady state assumption seems
valid for modeling 234Thof the bay because the magnitude of variation is insignificant in the mass
balance calculation. The range of residence times of total, dissolved, and particulate 234Thestimated
horn an irreversible scavenging model are 4-70, <: l-10, and 3-60 d, respectively. The residence
time of 234This controlled by the suspended particles whose distribution is determined by biological
production in the surface layer and resuspension of bottom sediments.
Vertical fluxes of 234Thmeasured by sediment traps also showed wide seasonal variation and
generally increased monotonically from 200-l ,000 dpm me2 d-l at 20 m to 2,000-4,000 at 100
m. The flux of 234Thpredicted with the steady state scavenging model tends to be about twice the
directly measured 234Th flux. This difference may be due either to low trap efhciency for 234Th-
containing particles or to boundary scavenging at the sides of the basin. The residence times of
particulate 234Th and total suspended matter agree well, and both are influenced by biological
Because of their high particulate matter elide tracers. Among those radionuclides,
concentrations, coastal waters have en- the 238U-234Thpair is e!;pecially suitable to
hanced scavenging properties relative to the study dynamic systems j n which scavenging
open ocean. Consequently, ocean margins time scales range from #days months. to
play an important role in controlling the 234Th (half-life - 24.1 d) has been used
geochemical cycling of many elements, es- successfully to study sca.venging in the open
pecially those that are particle reactive. Be- ocean (e.g. Bhat et al. 1!>69;Bacon and An-
cause particles serve as a conveyor belt for derson 1982; Bruland and Coale 1986) and
many elements, it is important to under- in coastal waters (e.g. K.aufman et al. 198 1;
stand the processesthat control their cycling McKee et al. 1984, 1986; Huh and Beasley
and removal in coastal waters. Knowledge 1987). Coastal systems are more compli-
of scavenging is based largely on distribu- cated because scavenging particles originate
tions of naturally occurring radioactive nu- from both biological processes and sedi-
ment resuspension. Both of these processes
are usually seasonal and not necessarily in
--- phase. The particles also have different
l Present address: Institute of Oceanography, Na- points of origin. The importance of these
tional Taiwan University, P.O. Box 23-l 3, Taipei, Tai- different particle sources varies depending
wan, Republic of China on location. For example, Santschi et al.
Acknowledgments (1979) estimated that the residence time of
This research was supported by NSF grant OCE 85- dissolved 234Th with respect to scavenging
11580. C.-L. Wei acknowledges support from the Na- in Narragansett Bay ir; only 2-22 d. Nar-
tional Science Council of the Republic of China (NSC-
80-0209-M002a- 13). ragansett Bay is shallow and dominated by
We th;ank B. Paul for help in all aspects of this project sediment resuspension, Extensive studies of
and A. Devol, R. Citterman, M. Peterson, W. Clark, 234Th scavenging in Funka Bay, a semi-
and H. Jannasch for help in sample collection. Jennifer closed coastal bay, have also been reported
Young also made valuable contributions. We thank P. (Minagawa and Tsunogai 1980; Tsunogai et
Crawford and R. McQuin for ship operation. Reviews
of the manuscript by L. Balistrieri, G. Shen, and M. al. 1980, 1986; Tanaka et al. 1983). Funka
Bacon are gratefully acknowledged. Bay differs from Narragansett Bay in that
Dabob Bay 234Thscavenging 297
the seasonal biological effects are more im- tern like Dabob Bay where the assumption
portant than sediment resuspension. of steady state needs to be evaluated. There
The residence times of 234Th in Funka are very few previous data sets that include
Bay vary seasonally and range between 30 analyses of both the water column and sed-
and 210 d (Tanaka et al. 1983). The first iment traps. Thus, the third goal of our study
objective of this study was to use 234Th to was to compare measured fluxes with those
study scavenging in a coastal location (Da- predicted by scavenging models.
bob Bay) known to be influenced by both In order to address the goals listed above
biological particle production and sediment we conducted an intensive, monthly sam-
resuspension. Previous work in Dabob Bay pling program for 234Th in the bay (Fig. 1)
has indicated that biological production oc- from January to December 1987. This
curs primarily in spring and summer coastal marine environment presented the
(Welschmeyer and Lorenzen 198 5; Downs opportunity to study the sensitivity of 234Th
and Lorenzen 1985), while sediment resus- to time-dependent variables like primary
pension is most important in the fall and productivity and sediment resuspension.
winter period of strong wind and high rain- Dabob Bay is a coastal fjord that has
fall (Bennett 1980; Hedges et al. 1988a). served as an ideal natural laboratory for bi-
In seasonally variable systems, the steady ological oceanographers and chemists to
state assumption used in modeling 234Th study coastal ocean processes (e.g. Welsch-
scavenging may not be valid. It is important meyer 1982; Welschmeyer et al. 1984;
to evaluate this assumption at each location Hedges et al. 1988a,b). A sill located at the
before estimating vertical fluxes of 234Th southern end of the bay makes it a semi-
from 234Th/238Udisequilibria in the water closed system with a negligible horizontal
column. Buesseler et al. (in press) showed component of advective transport. Gener-
that a nonsteady state model including the ally, productivity maxima are observed in
time rate of change of Th was necessary to spring and fall (Welschmeyer and Lorenzen
model the 234Thdata from the U.S. JGOFS 1985; Downs and Lorenzen 1985). Our
North Atlantic Bloom Experiment. In a sea- chosen station was in the northern end of
sonal study of 234Thin Funka Bay, Tanaka the bay at an average water depth of 115 m
et al. (1983) analyzed total 234Th and thus (Fig. 1). Detailed description of the study
studied total removal rates from the water area can be found elsewhere (Bennett 1980).
column. There have been few studies of the
seasonal dynamics of Th scavenging that Methods
include dissolved and particulate samples. Water-column samples -Seawater was
The advantage of this kind of sampling is collected from six depths (1, 10, 20,40, 70,
that Th scavenging can be broken into its 110 m) for measurements of salinity, nu-
component steps of uptake by particles and trients, total suspended matter concentra-
removal of the particles from the water col- tion, and total, dissolved, and particulate
umn. The second objective of our study was 234Th. Eighty liters of seawater were col-
to sample for dissolved and particulate 234Th lected with 30-liter Niskin samplers. Two
in Dabob Bay in order to evaluate the sea- 20-liter subsamples were used for total, dis-
sonality of their residence times. solved, and particulate 234Th analysis. The
Scavenging models have been developed rest of the sample was used for 210Pb/210Po
to calculate dissolved, particulate, and total and 226Ra measurements to be discussed
residence times of 234Th(e.g. Coale and Bru- elsewhere. Upon recovery, samples were
land 1985, 1987). Many of the assumptions drained into cubitainers, salinity and nutri-
commonly used in such scavenging models ent samples were taken, and the cubitainers
can be tested with sediment trap measure- were transported back to the laboratory.
ments of the particulate flux of 234Th. In Nutrients (NO,-, Si02, and P043-) were
addition, the predicted 234Th flux can be measured by autoanalyzer. The average pre-
compared with that measured to calculate cisions of N03-, Si02 and P043- concentra-
efficiency of the traps. Such evaluation is tions obtained from duplicate samples were
especially important in a time-varying sys- 3, 7, and 2%.
298 Wei and Murray
Fig. 1. Dabob Bay and the Puget Sound area. Sampling station (0) is shown in the insert.
The total 234Thsamples were weighed and passing the sample through three anion ex-
acidified with -50 ml of 12 N HCl and change columns preconditioned with 8 N
spiked with 230Thand 236Uyield tracer along HN03. The sample was evaporated down
with 60 mg of Fe carrier (as FeCl,). An iso- to one drop and was thlen ready for extrac-
topic equilibration time of at least 5 h was tion. 234Th and the yield tracer 230Th were
allowed. About 50 ml of concentrated extracted into a TTA ([thenoyltrifluoroace-
NH40H was then added to raise the pH to tone)-benzene solution and stippled on a
precipitate Fe(OH)3, which was separated stainless steel disk. 238Uin the sample along
from the supernatant by centrifugation and with the yield tracer 21W was plated on a
dissolved in concentrated HCl. stainless steel disk by electrodeposition.
Sample digestion and purification pro- Samples were weighed and then separated
cedures were modified slightly from An- into dissolved and particulate fractions by
derson and Fleer (1982). The sample was filtration through preweighed 0.45~pm Nu-
digested with perchloric acid to break down clepore filters. Dissolved samples were pro-
organic material and hydrofluoric acid to cessedin the same way as the total samples,
dissolve silicate minerals. After digestion, except that the digesting procedure was ex-
the sample was reprecipitated with concen- cluded. Filters were rinsed with -30 ml of
trated NH40H in a SO-ml centrifuge tube. deionized distilled water, dried in a desic-
The precipitate was then rinsed three times cator at room temperature, and then weighed
by repeating centrifugation, decanting, and to estimate the concentration of total sus-
resuspension with deionized distilled water. pended material. They were soaked in con-
The rinsed precipitate was dissolved in con- centrated NH,OH for a few hours to break
centrated HCl to make a 9 N HCl solution. them down and then digested and purified
Separation of U and Th was accom- following the same procedures as for total
plished by passing the solution through an 234Th samples. In order to minimize in-
anion ‘exchangecolumn ( 1X-8 Dowex resin) growth of 234Th from 238U, filtration and
preconditioned with 9 N HCl. Th passed Fe(OH), precipitate separation were typi-
through the column while U and Fe were cally completed within 1 d after the sample
retained by the resin. U was then eluted with was taken.
0.1 N HCl into a separate Teflon beaker. 234Th activity was counted with a low-
Th anld U samples were purified further by level, anti-coincidence beta counter with a
evaporating the sample on a hot plate and background of 0.2-0.3 cpm. Counting effi-
Dabob Bay 234Thscavenging 299
Table 1. 234Thactivity (dpm kg-‘) of multiple samples collected on selected cruises.
Cruise (m) Xl x2 x3 x4 ?I s SD(W
Mar/Apr 110 0.812 0.957 0.910 - 3 0.893 0.074 (8)
Ott 0 0.395 0.286 0.320 - 3 0.334 0.056 (17)
Nov 0 0.322 0.289 0.304 0.289 4 0.301 0.016 (5)
ciency was determined by counting a 234Th drained through 47-mm, 0.45~pm pore-size
standard plate prepared from 238U. The Nuclepore filters mounted on the bottom of
counting efficiencies of two detectors were the traps and then rinsed with -30 ml of
monitored throughout the experiment and deionized distilled water to remove salts.
showed no significant variation. Selected Zooplankton “swimmers” were found in all
234Th samples were also counted through the traps and were removed by hand pick-
time to confirm that the decay curve was ing. The 234Th associated with the zoo-
followed. Alpha activity of 230Th was ob- plankton that were hand picked in April
tained with silicon surface-barrier detectors samples was determined and found to be
(EG&G Ortec 576). U assay of the total negligible compared to total 234Thactivities
sample was also obtained with the same of the trap samples. The filters were dried
detectors. The counting efficiency was cal- in a desiccator at room temperature and
ibrated against an 241Am standard disk. weighed to estimate total mass flux.
Total counts for the samples were generally Analytical procedures used to measure
> 1,000 to achieve satisfactory counting sta- 234Thactivity on these filters were the same
tistics (SE = 3%). The chemical yield of Th as those for particulate 234Th.Blank values
was determined by 230Thactivity on the disk for mass flux and 234Th flux were deter-
and was typically in the range of 30-60%. mined with two dummy traps left on board
Blank determinations for Nuclepore fil- ship during deployment of the other traps.
ters and reagents were carried out for each Most of the blank values of the mass and
batch of samples and were generally negli- 234Th fluxes were < 7 and 1% of the total
gible. Beta contamination from the Nucle- mass and 234Th fluxes. These blank values
pore filters was found, however, in partic- were subtracted from measured values. Pre-
ulate samples of January, February, and cision is expressed as + 1 SD and was ob-
March-April. The measured particulate tained by propagating counting error.
234Thdata for these months were discarded
and particulate 234Thvalues were estimated Results
by subtracting dissolved 234Th from total The precision obtained from multiple
234Th. measurements (Table 1) averaged 10%. The
Sediment-trap samples-Modified MLML sum of the dissolved (DTh) and the partic-
sediment traps (Knauer et al. 1979; ulate (PTh) 234Th activities are not exactly
Martin et al. 1987) were deployed at five equal to total 234Thactivities (TTh) (Tables
depths on an anchored mooring line near 2 and 3). Most of the discrepancies between
the water-column sampling station. Each has (DTh + PTh) and TTh are < lo%, however
a trapping area of -50 cm2. The deepest (i.e. within the the average precision of the
traps were 5 m from the anchor weight. Each analysis). The least-squares fit of (DTh +
set was composed of two or four traps, and PTh) and TTh gives a slope of 0.9 1-to.07
sets were separated by 20 m. Typical de- (n = 48).
ployment lasted -3 d. The trap solutions and
Salinity ranges from 19.83 to 3 l.O2o/oo
were prepared by adding - 1,000 g of NaCl increases with depth (Fig. 2). The 3OcroO iso-
and -60 ml of 37% formaldehyde to filtered haline fluctuates from 25-40 m in winter to
surface seawater from the bay. This trap 90 m in summer, reflecting river runoff and
solution was used to minimize resuspension the degree of seawater exchange with Hood
losses and microbial degradation of trapped Canal. In periods of high wind, like Decem-
particles. ber 1987, the water column was homoge-
Upon recovery, trap solutions were neous and overall salinity was higher.
300 Wei and Murray
Table Z!. Salinity, nutrients, total suspended material, 238U,and total, dissolved, and particulate 234Thfrom
Dabob B;ly for January-December 1987. Standard deviations are based on propagated counting error (f 1 SD).
PO,‘- SiO, NO,- 238” TTh DTh PTh
Depth Salinity _ TSM ---
Cm) (73 @mol liter ‘) (mg kg-‘) (dpm-
1 28.832 2.51 73.28 27.57 0.48 0.589t-0.017 0.130f0.005
10 29.279 2.33 70.03 25.03 0.45 0.626+-0.018 0.1231k0.004
20 29.952 2.64 68.52 28.46 0.31 0.856kO.029 0.399kO.012
40 30.444 2.79 67.87 28.68 0.17 1.37720.049 0.137~0.005
70 30.580 2.48 69.67 28.28 0.44 0.895aO.023 0.110f0.004
110 30.804 2.42 78.57 26.45 0.96 0.910f0.026 0.113~0.004
1 24.170 0.61 67.30 2.75 4.19 0.392t-0.009 0.08 1k 0.003
10 29.142 2.31 69.95 25.71 0.96 0.552-t-0.030 0.283 k 0.007
20 29.859 2.51 73.89 28.44 1.03 0.770+-0.016 0.156k 0.004
40 30.132 2.56 75.03 28.90 0.72 l.O02t-0.038 0.150* 0.005
70 30.435 2.72 70.63 29.60 1.25 1.0121c_O.O26 0.131~0.004
110 30.744 3.20 83.25 30.37 0.84 0.934a0.023 0.086 2 0.003
1 26.553 1.89 0.375 -co.009 0.097 5 0.003
10 29.181 0.44 0.5701t0.017 0.192? 0.006
20 29.496 0.35 0.774-cO.020 0.170~.0.006
40 29.625 0.39 1.097-c0.030 0.137+.0.004
70 30.841 1.02 0.997f0.032 0.101 kO.004
110 30.188 1.07 0.893-eO.027 0.141 kO.003
1 27.487 0.75 7.81 1.11 1.20 0.317-c0.011 0.08 1kO.003 0.225kO.005
10 28.938 1.39 20.87 7.36 0.66 0.401 kO.012 0.148 kO.006 0.361 +O.O14
20 29.394 2.17 43.20 18.22 0.97 0.573kO.019 0.175+:0.007 0.404+0.0 10
40 29.802 2.37 63.73 28.07 0.42 0.723zkO.013 0.214kO.007 0.487kO.O 15
70 29.936 2.54 67.65 28.57 1.08 0.695 kO.023 0.156k0.005 0.589kO.017
90 29.948 2.75 71.60 30.24 3.70 0.750a0.029 0.055 k0.002 0.494+0.016
1 28.350 0.39 4.85 0.72 1.15 1.820 0.254-tO.007 0.09 1kO.003 0.226+0.006
10 28.744 1.05 16.62 5.48 2.06 1.910 0.474-to.013 0.104~I0.003 0.293kO.011
20 29.533 2.39 45.94 22.70 1.05 1.970 0.48 1kO.020 0.140~r0.005 0.248 zkO.008
40 29.834 2.56 55.03 26.43 0.51 2.030 0.8631kO.032 0.129 kO.003 0.579kO.018
70 29.97 1 2.70 63.53 28.88 0.82 2.010 0.839a0.029 0.078kO.002 0.701 kO.021
110 30.145 3.07 74.8 1 31.42 3.15 1.940 0.824a0.031 0.040~:0.00 1 0.759f0.024
0.62 34.64 0.07 0.68 1.750 0.363a0.013 0.154-1~0.007 0.134~0.005
10 1.21 38.26 5.80 1.37 1.990 0.365kO.010 0.074~~0.003 0.272+0.012
20 2.48 53.00 25.76 0.57 1.900 0.61OkO.019 0.131~~0.005 0.369+0.010
40 2.70 60.30 29.00 0.37 1.760 0.896a0.036 0.1541t0.008 0.509-tO.021
70 2.78 62.35 29.29 0.61 2.040 0.929 kO.038 0.0541kO.002 0.807kO.033
110 3.20 78.75 33.33 0.40 1.740 0.782-tO.025 0.062kO.003 0.79 1kO.027
1 29.069 0.82 33.86 0.08 1.28 1.970 0.278+0.011 0.08 1kO.003 0.214+0.009
10 29.768 2.19 43.41 15.07 0.76 2.003 0.5251kO.019 0.131:t0.006 0.223 kO.008
20 29.924 2.69 45.19 24.75 0.31 1.898 0.525kO.019 0.128k0.006 0.417+0.015
40 30.095 2.85 54.94 28.08 0.45 2.015 0.933-1-0.037 0.116:t0.004 0.743kO.025
70 30.342 2.62 53.9 I 28.53 0.61 2.037 0.738zkO.028 0.113:+0.005 0.6001kO.019
110 30.369 2.79 58.86 27.88 0.78 1.889 0.6241kO.020 0.086:+0.004 0.615+0.019
1 29.224 0.53 20.17 0.57 2.00 1 0.191f0.008 0.006.kO.000 0.145+0.005
10 29.840 1.oo 2.58 7.77 1.39 2.04 1 0.321+0.012 0.090~+0.003 0.244+0.011
20 30.132 2.73 57.61 29.44 0.84 1.665 0.843 kO.026 0.185 kO.006 0.830+0.026
Dabob Bay 234Thscavenging 301
Table 2. Continued.
PO,‘- SiO, NO,- 23qJ TTh DTh PTh
Depth Salinity TSM
(4 (W &mol liter-‘) (mg kg-‘) @pm kg-‘)
40 30.461 2.66 56.70 28.80 0.28 2.028 0.646kO.019 0.020*0.001 0.497zk0.015
70 30.479 2.58 58.59 28.08 0.10 2.262 0.517kO.016 0.107~0.004 0.55O-cO.018
110 30.65 1 2.19 49.74 21.79 0.44 2.193 0.438+0.014 0.043+0.001 0.497kO.026
1 29.27 1.07 9.91 2.05 0.93 1.792 0.395f0.014 0.121-tO.004 0.223+_0.008
10 30.27 2.68 65.92 21.89 0.32 1.898 0.473kO.014 0.24OkO.009 0.124+0.004
20 30.59 2.60 67.83 22.57 0.20 1.923 0.694kO.017 0.133~0.004 0.685f0.024
40 30.71 2.72 51.82 24.68 0.20 1.924 0.6161kO.013 0.197f0.007 0.419+0.015
70 30.79 2.69 55.17 24.82 0.35 1.946 0.599+0.015 0.070f0.002 0.436f0.014
110 30.81 2.58 45.00 20.38 0.59 2.038 0.532-eO.018 0.083+_0.003 0.479kO.016
1 19.380 2.32 46.64 20.63 0.25 1.787 0.413+0.011 0.180~0.007 0.146+0.006
10 23.602 2.20 19.37 9.95 0.31 1.831 0.503f0.021 0.281 kO.011 0.165f0.005
20 25.164 2.37 40.00 18.18 0.02 1.811 0.328kO.008 0.209 kO.008 0.155Ik0.005
40 29.860 2.55 35.96 17.62 0.42 1.745 0.469kO.015 0.023 t-O.00 1 0.399+0.015
70 29.925 2.76 51.15 21.69 0.78 1.978 0.968kO.033 0.029+_0.001 0.77320.026
110 30.7 10 2.43 52.55 20.61 1.14 1.778 0.545kO.020 0.0061kO.000 0.482f0.013
1 28.277 2.56 63.67 25.52 1.97 1.828 0.55920.02 1 0.037f0.001 0.317kO.012
10 28.258 2.60 61.99 25.82 1.88 1.829 0.4131kO.015 0.028~0.001 0.356kO.012
20 30.220 2.77 60.94 27.19 0.59 1.901 0.428f0.018 0.087 zkO.004 0.468-eO.016
40 3 1.022 2.77 59.71 27.34 0.70 1.963 0.735kO.023 0.039+0.002 0.473+0.018
70 31.005 2.71 59.65 27.39 0.75 2.149 0.7OlkO.022 0.068kO.003 0.502kO.018
110 31.022 2.69 65.55 26.94 0.93 2.204 0.696kO.025 dO.001 0.68 1kO.026
Nutrient concentrations showed large vari- weight basis are 9-22 and 0.5-5.4%, re-
ations, depending on physical conditions spectively. Only one observation was
and biological activity. Generally, in winter available to show vertical variation in the
and early spring the nutrient content of the composition of suspended matter. A par-
bay is higher and more homogeneous than ticulate sample taken from 50 m in winter
in other seasons.In early fall (e.g. September contained significantly lower (50% lower)
and October) broad maxima of all three ma- organic C and higher (60% higher) Al than
jor nutrients are found at middepths. the sample taken from 5 m.
The vertical variation of total suspended 238U activities in the unfiltered samples
matter (TSM) is characterized by a mini- were determined only from June to Decem-
mum at middepth, which is sandwiched be- ber. Despite poor recovery efficiencies
tween maxima resulting from production by (< 5%) the average value of the normalized
biological activity in the surface layer and 238Uto 35%0salinity is 2.3320.22 dpm kg-‘,
resuspension of bottom sediments in deep consistent with the open-ocean value of 2.43
water (Fig. 3). Bennett (1980) gave a slightly dpm kg-’ at 35sloo salinity (Ku et al. 1977).
different picture of the vertical distribution Thus, we assume a conservative distribu-
of TSM (i.e. generally constant values in the tion of 238Uin the bay with the same 238U-
upper 50 m increasing toward bottom). He salinity correlation as the open ocean, 238U
pointed out, however, that his TSM mea- (dpm liter-l) = 0.07081 X salinity (?&).
surements may have been biased by the The 234Thdistribution in the bay (Fig. 4)
small-volume (50- 100 ml) samples used. is different from that of open and coastal
Bennett determined the compositions of oceans in several ways. First, total 234Th
suspended matter collected on four cruises below the euphotic zones of open and coast-
in the bay. He found that the concentration al oceans is generally in secular equilibrium
of organic C and particulate Al on a dry- with 238U, but in Dabob Bay total 234Th
302 Wei and Murray
Table 3. Total mass and 234Thfluxes determined Table 3. Continued.
from sediment-trap samples collected in Dabob Bay C
from February to December 1987. Standard deviations Depth Mass flux 234Thflux *-Th
are based on propagated counting error (+ 1 SD). Spe- b-4 (gm-*d-l ) (dpm rnd2d-l) @pm mg-9
cific 234Thactivities are calculated by the ratio of 234Th 68 3.36 1,1601t32 0.35
flux to mass flux. 88 5.39 2,751+83 0.51
-- 108 8.17 3,642+ 125 0.45
Depth Mass flux 234Thflux 234Th
(ml (g nl-2 d-l) (dpm IT-~ d-l) @pm mg-‘) December
February 25 22.85 2,146+73 0.09
20 2.98 693k 18 0.23 45 23.80 3,442+82 0.14
50 2.76 1,181&36 0.43 65 25.55 4,343+83 0.17
90 10.02 3,105+110 0.31 85 36.40 7,473f 175 0.21
105 42.85 9,832-+ 170 0.23
19 0.96 365+ 14 0.38
39 1.13 352+6 0.31
59 0.94 1,229&27 1.31
79 5.03 1,835+49 0.36 activities are below the level of secular equi-
99 15.48 3,513+115 0.23
librium with 238Uthroughout the water col-
May umn. 238Uactivity in Dabob Bay waters is
25 1.41 168+5 0.12 - 1.9-2.0 dpm kg-l and is relatively ho-
45 1.12 29028 0.26
6!j 1.36 4492 17 0.33 mogeneous throughout the water column.
85 3.89 907+36 0.23 Hence, the 234Th: 238Uratio of total samples
1O!j 16.14 1,945-e37 0.12 ranges from -0.2 in the surface layer to
June -0.5 in the deep layer, and the vertical dis-
23 4.82 635-t-24 0.13 tribution of 234Th*238Uis similar to the 234Th
43 4.15 977+-34 0.24 activity profiles. Second, 234Th in the open
63 4.91 1,248+37 0.25 and coastal ocean is primarily in dissolved
t3:3 5.59 1,790+45 0.32 form, but in Dabob Bay most of the 234Th
103 11.61 2,722+- 100 0.23
(>90%) is incorporated into filterable par-
23 2.86 722223 0.25 The partition coefficient, &, of 234Th
4.3 2.76 1,23St-47 0.45
6.3 2.21 l,266t-43 0.57 ranges from 1O6to 1O8ml g- ’ . Generally, Kd
83 5.10 1,657t65 0.33 in the euphotic zone is about an order of
103 12.36 3,722f 133 0.30 magnitude less than that below the euphotic
August zone, and a broad maximum in Kd is found
20 1.98 16817 0.08 in the intermediate layer of Dabob Bay. This
40 2.08 97Of25 0.47 pattern is due to lower concentration and
60 2.52 1,676f60 0.67 longer residence time of TSM at interme-
81) 4.81 2,197-t84 0.46
1010 11.39 3,060&85 0.27 diate depth relative to the surface and bot-
tom layers. Third, the vertical variation in
September 234Th is more complicated than that of ei-
24 3.44 98Ok30 0.28 ther the open or coastal ocean. Total 234Th
44 2.09 999+33 0.48
64 2.52 1,1s3+47 0.47 profiles in the bay have their lowest values
84 7.53 1,706-t54 0.23 at the surface and a subsurface maximum,
104 15.45 1,74 1f47 0.11 usually at 20-40 m. Dissolved 234Th (Fig.
October 4b) activities are low, CO.2 dpm kg-‘, and
25 1.26 632f15 0.50 show less vertical variation than those of
4.5 1.51 1,160+39 0.77 total 234Th.A slight m.aximum of dissolved
65 2.42 1,239a49 0.5 I 234Th is also found at intermediate depth
85 2.86 2,042+55 0.72
105 5.97 2,553-t43 0.43 for most of the samp’ling months. Particu-
late 234Thactivities (Fig. 4c) are much high-
November er than the dissolved fraction and show a
;!8 2.39 358k 12 0.15
48 1.95 683+22 0.35 vertical distribution similar to total 234Th.
The range of total mass fluxes measured
Dabob Bay 234Thscavenging 303
Salinity, %bo TSM, mg kg”
.-- Fig. 3. As Fig. 2, but of total suspended matter.
J F M A M J J A S 0 N D tal mass flux at 55-60 m in 1977 and 1978
and found temporal variability of up to a
factor of three (2-5 g m-2 d-l). This range
Lo-. . . . . . . . . . .
of variation was reproduced by our 60-m
a - H trap if the anomalous December data were
fb- excluded. The anomalous mass fluxes ob-
served in December were as high as 43 g
m-2d-1 at 100 m. These fluxes were possibly
a result of active resuspension of bottom
sediment by strong winds. The consistency
between our data and others indicates that
J F M A M J J A S 0 N D
our trap deployment period of only 3 d was
sufficient to reproduce the seasonal patterns
found by Bennett and Hedges et al. ( 19886)
(15-30-d deployments). The agreement be-
tween mass fluxes measured by duplicate
traps on the same trap holder is excellent
(flux A = -0.5 + 1.05 x flux B, r = 0.99,
n = 78).
Mass fluxes obtained from our sediment
traps were relatively constant or decreased
slightly in the upper 50 m. Below 50 m a
SlO,, pm01 Iltei’
J F M A M J J A S 0 N D dramatic increase was observed. The mass
fluxes for our deepest trap (N 100 m) were
up to a factor of 5 higher than in the shal-
lowest trap (-20 m). The increasing mass
flux with depth was probably the result of
added resuspended sediment to our traps
deeper than 50 m. This resuspension phe-
nomenon is supported by the findings of
Cowie and Hedges (1984), who showed that
the carbohydrate composition of particles
in their deepest traps (90 m) was the same
Fig. 2. Vertical distributions of salinity, nitrate, as in underlying sediments. Furthermore,
phosphate, and silicate vs. time. The regions with high
and low concentrations are represented by H and L. all mass fluxes deeper than 60 m were higher
than the average sedimentation rate of
3.2OkO.6 1 g mm2d-l (Carpenter et al. 1985;
from February to November was 0.5-16.0 Furlong 1986). Resuspension must be an
g m-2 d-l (Fig. 5a), consistent with previous important factor controlling the distribu-
studies in the bay (Hedges et al. 1988b; Ben- tion of particles in the deep water of Dabob
nett 1980). Bennett (1980) measured the to- Bay.
Dabob Bay 234Thscavenging 305
pared with a measured trap flux. In this way, J Th = xTh(u - DTh), (7)
234Th can be used to test the efficiency of
sediment-trap collection. DTh
T,, = - (8)
In the irreversible scavenging model J Th ’
adopted from Coale and Bruland (1985),
diffusive and advective transport rates are PTh = xTh(u - TTh), (9)
ignored because the turnover of 234Th in
seawater is rapid as a result of its brief mean and
lifetime (35 d) and fast scavenging onto par-
ticles. If advection and diffusion terms are rp = - (10)
ignored, then the mass-balance equations PTh ’
for total, dissolved, and particulate 234Th
are written as The scavenging model can be applied to
a given data set by different, but equally
valid approaches. The choice depends on
dTTh the purpose of the model. Values of X,, .&,,
- = X,,U - &TTh - X,TTh, (1)
dt PTh, and T can be calculated on a point-by-
point basis using U and Th activities in units
- = XT& - &,DTh - JTh, (2) of dpm kg-l. These parameters can also be
dt calculated for the water column for one or
a number of compartments or boxes, with
the activities in each box expressed as in-
dPTh ventories in dpm m-2.
dt= JTh - X,,PTh - PTh (3) Coale and Bruland (1985, 1987) applied
the steady state version of this model to
where TTh is the total 234Th (dpm kg-l), vertical profiles of water-column and sedi-
DTh the dissolved 234Th (dpm kg-‘), PTh ment-trap data. Values of JTh and PTh were
the particulate 234Th (dpm kg-l), U the 238U calculated point-by-point through the water
(dpm kg-‘), &, the decay constant of 234Th column. The resulting values of residence
(d-l), X, the first-order scavenging rate con- time tended to be erratic. In order to cal-
stant for 234Th (d-l), and JTh and PTh the culate a Th flux (dpm m-2 d-l) that could
rate of removal of dissolved and particulate be compared with the measured trap flux
(and total) 234Th, respectively, by processes they integrated the values of PTh (in dpm
other than radioactive decay (dpm kg-l d-l). kg-’ d-l) from the sea surface to the trap
The removal rate constants for total 234Th depth.
and the nonradiogenic production-con- Buesseler et al. (in press) obtained a time
sumption of dissolved and particulate 234Th series during the UWJGOFS North Atlan-
are easily calculated by assuming steady tic Bloom Experiment. They divided the
state: water column into a series of boxes that
corresponded to the floating sediment-trap
in each box. Changes in the inventories be-
-c-=-c dDTh dPTh o tween sampling times were used to calculate
dt dt dt ’ (4)
depths and calculated the inventories (dpm
m-2) of dissolved, particulate, and total 234Th
The total, dissolved, and particulate 234Th time-rate-of-change terms in Eq. 1, 2, and
residence times (TT, TP, and TP) are calcu- 3. The resulting values of JTh and P-,-hhave
lated as units of dpm m-2 d-‘. The calculated flux
at the bottom of each box was equal to the
x = A U-J - TTh) PTh produced in that box plus the PTh input
TTB ’ (5)
from the overlying box. Only the surface
box has no PTh input from above. This ap-
1 proach is useful to determine the flux of
TT = -, (6)
xc 234Th produced in different parts of the wa-
306 Wei and Murray
ter column. If the goal is simply to predict Table 4. Residence times, (in days) of total and dis-
the flux of 234Th at some depth, the multi- solved 234Thin the surface water calculated from the
irreversible scavenging model with the assumptions of
box approach is not necessary and an in- steady state and nonsteady state.
tegration from. the surface to the trap depth
will give the correct value. The only data
necessary to predict a 234Th flux at a given
depth are the activities of total 238U and
234Th (‘seeEq. I and 9). Because disequili-
14 Jan 14.6 2.4
bria represent removal, the model flux (PTh) 18 Feb 10.7 8.4
is the net of all processes, including aggre- 1 Apr 8.9 10 2.1
7.2 1.9 1.6
gation and disaggregation. This approach 20 May 6.8
5.6 1.6 1.8
assumes that all the 234Th is collected in 24 Jun 5.6 1.8
29 Jul 9.1 8.7 3.2
sampling total Th (i.e. that no significant 6.6 3.4 1.7
26 Aug 5.6 1.5
particulate pool is excluded). The residence 23 Sep 3.7 3.7 0.1
9.9 0.1 2.5
times for the single- or multibox approach 26 Ott 9.8 2.5
10 Nov 10.5 10.6 3.9
are calculated by dividing the inventories 17.3 3.9
7 Dee 15.3 0.7 0.7
of 234Th by JTh and I’Th as appropriate.
Tanaka et al. ( 1983) doubted the validity
of the steady state assumption in the coastal
environment and retained dTThldt in Eq. state is an acceptable assumption for all
1 to solve for TTh analytically: depths in the bay.
In Funka Bay, the mean residence time
of total 234Th was l-7 months rather than
Th2 5-30 d as observed in Dabob Bay. The lon-
Xc + b-h
ger residence times (or slower removal rates)
of234Th in Funka Bay make dTThldt more
significant in the mass-balance equation. In
some months, dTTh/dt was even greater
than the particulate sinking term of 234Th.
-=I+(& + h,)Ar] (11) Hence, the steady state assumption is in-
appropriate for conditions in Funka Bay.
where Thi is the total 234Th activity from Dabob Bay is different from Funka Bay in
successive observations, and AT the inter- two aspects: the removal rate of 234Th is
val between sampling time of Th, and Th2. faster and temporal variations of 234Th are
smaller. Radioactive decay and other re-
Discussion moval processes are fast enough to balance
Steady state vs. nonsteady state- We test- the production rate of 234Th. Accordingly,
ed the steady state assumption in Dabob the steady state assumption is applicable for
Bay by comparing steady state and non- our Dabob Bay data. The difference be-
steady state approaches. We used total and tween the two bays may be due to physical
dissolved 234Th data of the surface water for mixing. Active exchange of water masses in
this test because those samples showed the Funka Bay has been reported (Ohtani and
largest variations in 234Th activity with time. Akiba 1970). On the other hand, the pres-
Total and dissolved 234Th should bracket ence of a sill at the southern end of Dabob
particulate 234Th in terms of activity and Bay inhibits water exchange with the main
removal rate. It is clear (Table 4) that there body of Puget Sound, which is also some-
is good agreement between the two ap- what isolated from the North Pacific Ocean
proaches. The same calculation was applied by a sill.
to the: subsurface total and dissolved 234Th Residence time of 234Th in Dabob Bay-
data but is not shown. There was no differ- One of the goals of this study was to utilize
ence between the residence times calculated our dissolved and particulate analysis of
from steady state and nonsteady state ap- 234Th to evaluate the seasonal variation in
proac.hes. These results suggest that steady dissolved Th uptake by particles and re-
Dabob Bay 234Thscavenging 307
moval of particulate Th from the water col-
umn. The range of residence times for total
(TT), dissolved (TD), and partidate (7~) 234Th
are 4-70, < l-10, and 3-60 d (Fig. 7). There 2
tends to be a maximum in 73 between 30
and 70 m. Minagawa and Tsunogai (1980)
and Tsunogai et al. (1986) observed a sim-
ilar subsurface residence-time maximum in
Funka Bay. They attributed this distribu-
tion to removal from the surface layer, o-
followed by disaggregation and decreased
settling velocity of particles, without regen- loo- >20 20 (20
eration of nutrients, at middepth. We be- . . . . . . .
lieve this combination of processes is also *(a
the cause of most of the vertical distribution TD(d)
Jan Feb Mar May Jun Jul Aug Sep Ott Nov Dee
of TT in Dabob Bay. Remineralization of
particulate material with nutrient regener-
ation may be important in September and
October when a broad maximum develops 2
in nutrient distributions at intermediate
depths. Another possible cause of the max- z 4
imUm in TT iS reSUSpenSiOn Of bottom sed- 2
iment into waters below the maximum, per- 5 6o
haps from the sides ofthe bay. Resuspended c1 - ’ ’ l ’ ’ ’ l l l ’ ’
particles contain less 234Th per gram than *O-
particles found in the overlying water (Car-
penter et al. 1984). When these sediments
- . . . . . . . . . . l
are resuspended into the overlying water, (bl
they may act as efficient scavengers for Th
in seawater. TP (d)
Jan Feb Mar May Jun Jul Aug Sep Ott Nov Dee
The residence time of dissolved 234Th with 0 I1 I I I I I
I I I I I,
. . .
I, I , , ,
respect to the net scavenging rate is very . . . . . . . . . . .
brief (3 d) throughout the year. Extremely 20
short TD relative to TT and rp implies that
the uptake rate is much faster than the re- 40
moval rate of particulate 234Th. A maxi- 2
mum of TD was observed below the euphotic 5 60
zone at the depth of the TSM minimum. n
- . . . .
The coexistence of the TSM minimum and 80
the TD maximum is not surprising because
fewer adsorption sites are available in that 100 >20
layer. Honeyman et al. (1988) reviewed - . . . . . . . . . . .
234Th field data and found that a positive 120
correlation exists between TSM and the
Fig. 7. Vertical distributions of residence time of
scavenging rate constant. Vertical and tem- total, dissolved, and particulate 234Thvs. time as es-
poral variations of Tp were similar to rT. timated by the irreversible scavenging model.
This finding implies that the removal rate
of particulate matter is the determinant of
total residence time of 234Th in Dabob Bay. (Welschmeyer 1982; Downs and Lorenzen
234Thin the surface waters-Average pri- 1985). Welschmeyer and Lorenzen ( 198 5)
mary production in the euphotic zone of the estimated that >80% of the chlorophyll
bay is -600 mg C m-2 d-’ with a seasonal produced in the euphotic zone is removed
variation of one order of magnitude by settling in the form of fecal pellets pro-
308 Wei and Murray
Tot. ing the annual cycle of 1987. Dissolved 234Th
0.5 -’ fluctuates between 0 and 0.2 dpm kg-l and
particulate 234Th activity generally shows a
pattern similar to that of total 234Th (i.e.
0.3 -’ Part. high in winter and early spring and low in
The temporal variations of the residence
a1 .& times of surface 234Th are similar to those
Diss. of the 234Th activity (Fig. 8b). The ranges
I I -I
o.o- ’ ’ ’ ’ ’ ’ ’ ’ of the residence times for total, dissolved,
and particulate 234Th are 3-15, -K1-6, and
Month 3-l 1 d. Residence times for total and par-
ticulate 234Th show strong seasonal varia-
Tot. tions, while those of dissolved 234Th re-
Seasonal variations in the activity and
residence time of 234Th in the surface layer
Part. .are closely related to biological processes.
Primary productivity data provided by A.
Devol show phytoplankton blooms in May,
July, and September. 13ecause these 14C in-
J cubation measurements were not carried out
‘J FMAMJJASOND at exactly the same d,ates as our Th sam-
Month pling, the data can serve only in qualitative
comparison with Th data. In Dabob Bay,
T 60- w microphytoplankton biomass is dominated
by diatoms (Hedges et al. 19883), which
$j may explain the two minima in SiO, con-
= centration in late spring and late fall (Fig.
3. 8~). It is interesting that SiO, concentration
; seems to vary with total 234Th activity. Slight
8 20- offsets of the 234Th and SiO, peaks from
bd, primary productivity are probably due to
O- the role of secondary ,producers.
J FMAMJJASOND The grazing activities of copepods in the
Month genera Calanus and Pseudocalanus play an
important role in the dynamics of algal bio-
Fig. 8. Variation of 234Thactivities and residence
times in total, dissolved, and particulate samples, and mass in the bay (Shaman 1978; Bennett
SiO, concentrations in the surface waters vs. time. 1980; Welschmeyer :and Lorenzen 1985;
Welschmeyer et al. 1984). A simulation
model describing the relationship between
phytoplankton biomass and filter-feeding
ducecl by herbivorous zooplankton or by the zooplankton biomass (Frost 1980) suggest-
direct sinking of algal cells. Only a small ed a time lag of - 1 month between the
proportion of fixed C stays in the euphotic maxima of phytoplankton biomass and
zone and is recycled. These biologically pro- zooplankton biomass. From careful exam-
duced particles may serve as a conveyor belt ination of the pigment budget at the same
for Th and other particle-reactive metals. site, Welschmeyer and Lorenzen ( 198 5)
Accordingly , 234Th activity in surface wa- concluded that zooplankton grazing is the
ters should give some seasonal signals that major process that controls phytoplankton
reflect biological activity. Generally, the ac- abundance. Hence, a reasonable picture is
tivity of 234Th is higher in winter than in that herbivorous zooplankton biomass (and
summer and fall (Fig. 8a). A threefold vari- activity) starts to increase, in response to
ation in 234Th activities was observed dur- increasing phytoplankton biomass, and
Dabob Bay 234Thscavenging 309
reaches a maximum - 1 month later than
the primary productivity peak. Although
zooplankton grazing activity increases,
downward fluxes of egested fecal pellet also
increase and reach a maximum at - 1 month
after the primary productivity peak.
A lag between the primary productivity
peak and the pheopigment flux was not ob-
served by Welschmeyer and Lorenzen
(198 5). It should be noted that their trap
data were from 60 to 80 m and thus the JFMAMJ JASOND
pheopigment fluxes may not be represen- Month
tative of that out of the euphotic zone. Close Fig. 9. Variation of total, dissolved, and particulate
examination reveals that the pheopigment 234Thinventory for the O-l 10-m depth range vs. time.
fluxes did show significant vertical varia-
tion, up to eightfold, between 20 and 40 m ever, provided a large amount of particulate
during some of the months of their sampling matter to scavenge 234Th out of the water
year. In this study, we observed maxima in column. As a result, the inventory of total
the mass and 234Th fluxes at 20 m in June 234Th in December was lower than in Jan-
and September. Hence, the total 234Th ac- uary.
tivity in surface water in these 2 months The dissolved 234Th inventory is rela-
was probably a minimum as a result of en- tively constant throughout the year. Partic-
hanced 234Th transport out of the euphotic ulate 234Th inventory follows the annual
zone during those periods. variation of total 234Th inventory. More than
Inventory of 234Th-The highest total 70% of the 234Th in Dabob Bay is associated
234Th standing stock was found in January with filterable particles (Fig. 9).
(1.08 x 1O5 dpm m-2), and minima, on the Th
234 activity on particles-The activity
order of 0.70 x lo5 dpm m-2, were found of 234Th per unit of weight of particles is
in May and September (Fig. 9). Standing called the specific activity and is a function
stock of 238U above 110 m is -2.35 x 1O5 of the concentration and residence time of
dpm m-2 and thus the daily radiogenic pro- particles. In Dabob Bay the specific activity
duction of 234Th is 6,800 dpm m-2. Part of of 234Th in filterable particles ranges from
this production (30-45%) disappears 0.15 to 2.5 dpm mg-l and generally shows
through radioactive decay. If steady state is a maximum at middepth (Fig. 1Oa). This
maintained, the remaining 55-70% must be range brackets the values found by others
removed by processes other than radioac- (e.g. Minagawa and Tsunogai 1980). Unlike
tive decay. Annual variations in the inven- the findings of Minagawa and Tsunogai
tory of total 234Th are generally similar to (1980), however, the correlation of partic-
variations in the 234Th activity in surface ulate 234Th and TSM in our data set is poor.
waters (Fig. 8) with two exceptions: the min- The maximum specific activities of 234Th
imum in the total and particulate 234Th in- are found on particles at middepth due to
ventory was found in May instead of in lower TSM concentrations and longer res-
June and the total 234Th inventory of De- idence times of particles in that layer. The
cember is only -70% of the January inven- dilution effect caused by resuspension of
t0I-y. sediment with low 234Th activity into the
Causes of the variations in the inventory bottom layers may also result in a relative
are more complicated than that of the ac- maximum of specific 234Th activity at mid-
tivities in the surface layer. Sediment re- depth. The dilution effect was especially ev-
suspension needs to be considered as well ident in late winter when the resuspension
as biological activity. For example, primary was active, resulting in low specific 234Th
productivity in December is low and, as a activity in filterable particles throughout the
result, the 234Th activity in surface waters water column.
is as high as the January value. Active re- The specific activity of 234Th in sediment-
suspension caused by strong winds, how- trap material on a salt-free, dry-weight basis
310 Wei and Murray
=Th, dpm mg’ The difference between the scavenging abil-
ity of these two kinds of particles has also
been reported for other elements. Bennett
( 1980) found that trap :particles have higher
Al content and lower Mn and organic C
content than do suspended particles. The
difference in the 234Th activity of these two
types of particles is a reflection of the dif-
ference in their size spectrum, composition,
and history. Tsunogai e:t al. (1974) proposed
that most of the particles collected by fil-
P4Th, dpm mg” tration from seawater do not contribute to
JFMAMJJASOND the vertical settling flux. The particles caught
by a trap originate in the euphotic zone.
Bacon and Anderson (19 82) and Nozaki
et al. ( 1987) differentiated particles into two
reservoirs, fine particles with negligible set-
tling velocity and 1arg;e particles with fast
settling velocity, to model the distribution
of 230Th (half-life = 7.5 x 1O4 yr) in the
ib) open ocean. Sinking particles are generally
larger in size and have a smaller surface-
Fig. 10. As Fig. 2, but of 234Thactivity in filterable area-to-volume ratio, which results in fewer
particles (a) and trap particles (b) vs. time.
adsorption sites for Th and other elements.
Bennett (1980) found that 86 111% of trap
was 0.1-0.5 dpm mg--’ (Table 3 and Fig. particles on a dry-weight basis are > 10 pm
lob). This value is similar to that observed in size and that suspencled particles are much
for suspended matter in the euphotic zone. smaller. He argued that the enrichment of
Our data bracket the 234Th specific activities fine (< 1 pm) MnO, particles in the sus-
obtained at the same station by Carpenter pended pool was explained by the inability
et al. I(1984), who used Lorenzen-type sed- of Pseudocalanus to feed on particles small-
iment traps (Lorenzen et al. 198 1). A max- er than 4 pm. Murray ( 1975a, b) and Murray
imum in the activity of 234Th in sediment- and Brewer (1977) found that Mn oxides or
trap materials at 60 m was found in all hydroxides can be efficient scavengers of
months, although the maximum was not as dissolved trace metals. Hunter et al. (1988)
pronounced as that for filterable particles. extrapolated their laboratory results to con-
These: maxima are probably the net result clude that significant adsorption of Th by
of adsorption of 234Th by particles settling Mn oxides can be expected in the coastal
through the water column and dilution with environment.
resuspended particles having lower specific The existence of fine Mn02 particles in
234Th activities. The highest activity of ex- the suspended material may explain why
cess 234Th observed in surface sediments in the activity of 234Th in suspended particles
the miain basin of Puget Sound is 0.05 dpm is higher than that of trap particles. Another
mg-l (Carpenter et al. 1984), which is lower possible explanation for the lower 234Th ac-
than lthe value of either our trap samples or tivity on trap particles is their “older” age.
filterable particles. Radioactive decay of Almost all the particulate matter caught by
234Th while particles travel through water sediment traps is fecal1pellet material egest-
column is probably insignificant because the ed by zooplankton a:nd those fecal pellets
settling velocity of fecal pellets egested by must have undergone recycling before sink-
zoopjtankton is high (143 m d-l, Shuman ing (Shuman 1978; Bennett 1980). Bennett
1978; Lorenzen and Welschmeyer 1983). (1980) attributed this recycling to coproph-
234’rh activity on trap particles is lower agy which is intensified when the population
than that of particles collected by filtration. of zooplankton is high.
Dabob Bay 234Thscavenging 311
?l~ Flux (1000 dpm m”d -3
oO 12 3 4 5 6 0 123 4: 6 012 9 4 26 O! 2 3 4 56 01 .2 Cj 4 5 6
a b A&
MulApr May Jun Jd
A’w od NW Dee
Fig. 11. Vertical profiles of the 234Thfluxes measured from sediment-trap data (0) and calculated from the
irreversible scavenging model (0).
Measured vs. calculated 234Th fluxes- depth and calculating the inventories of to-
Minagawa and Tsunogai (1980) observed tal 238U and 234Th. Both fluxes increase with
that measured 234Th fluxes were lower than depth and, with the exception of December,
fluxes derived from an irreversible scav- the trap flux is always greater than the model
enging model. They attributed the discrep- flux. Overall the ratio of the trap flux to
ancy to poor trapping efficiency and exper- model flux is -56%, and there are no sys-
imental error. Tsunogai et al. (1986) applied tematic trends in this ratio with water depth.
a nonsteady state scavenging model to the There are some variations from month to
Funka Bay data and concluded that the dif- month (Table 5). In February, the ratio is
ference between calculated and measured close to 1 and in December it is > 1. In May
234Th fluxes was caused by variations in sta-
bility of the water column. They observed
that the calculated and measured fluxes co- Table 5. The monthly averages of the ratio of the
measured trap flux of 234Thto that predicted with the
incided in summer and diverged in other irreversible scavenging model.
seasons. In open-ocean sutiace waters, Coale
and Bruland ( 198 5) found that the mea- Month Trap/model
sured 234Th flux agreed with the calculated Feb 0.84
234Th flux to within 20-40%. Mar/Apr 0.6 1
In the irreversible scavenging model (Eq. May 0.24
3 and 9), PTh is defined as the removal rate Jun 0.55
of particulate 234Th by processes other than Aug 0.57
radioactive decay. For comparison with trap Sw 0.55
data (Fig. 11) we calculated the fluxes of Ott 0.52
234Th (PTh in dpm m -2 d-l) by defining a Nov 0.52
box from the surface to each successive trap _
312 Wei and Murray
u” 25 -
Feb Mar May Jun Jul Aw Sep Ott Nov Dee
Fig. 12. Residence times of particulate 234Th and total suspended matter in the euphotic zone vs. time.
Residence time of particulate 234Thgiven by the ratio of particulate 234Thabove the shallowest trap and the PTh
term defined in the irreversible scavenging model-Cl; residence time of particulate z34Thgiven by the ratio of
inventory of PTh above the shallowest trap (19-28 m) and 234Thflux measured al: that depth-O; residence
time of suspended particles given by the ratio of inventory of TSM above the shallowest trap (19-28 m) and
total mass flux measured at that depth-O.
the ratio averages 24%. For all the rest of solve these alternatives with the present data
the year the ratio is - 5 5%. May is a period set.
of maximal grazing following the spring Particle dynamics in the euphotic zone-
bloom (Downs and Lorenzen 1985) and it The depth of the euphotic zone, judging from
may be that the low ratio during that month a Secchi disk, ranged between 10 and 25 m.
reflects poor trapping efficiency for the type We have taken the depth of the shallowest
of sinking particles produced at that time. trap (19-28 m) as the lower boundary of the
December is anomalous because the mea- euphotic zone. Particulate 234Th residence
sured fluxes are larger than the modeled times in the euphotic zone (Fig. 12) were
fluxes. The predicted flux is based on water- calculated by two approaches: dividing the
column measurements made at the time of inventory of particulate 234Th above the
trap deployment. After those samples were shallowest trap by the 234Th flux (ThF) de-
collected there were strong winds that may termined from sediment-trap data at that
have :resulted in a pulse of resuspended sed- depth and dividing the inventory by the flux
iment. This possibility is supported by the predicted by the irreversible scavenging
low values of the specific: 234Th activities model (Eq. 10). A similar approach was used
observed (Table 3). The value of 55% ob- to calculate residence time of TSM by di-
served for the rest of the year may represent viding the inventory of TSM above the trap
the efficiency of our sediment traps for 234Th- by the measured mass flux (A@‘).
containing particles. This low ratio may be In general, the residence time of partic-
due either to hydrodynamic effects or to the ulate 234Th calculated by trap and model
nature of the sinking particles. An alterna- fluxes agree to within a factor of 3. Partic-
tive explanation is that because Dabob Bay ulate 234Th residence times calculated from
is a fjord, boundary scavenging, as de- the trap data are anolmalously long in May
scribed by Anderson et al. (1983) may be and August. These points are at the time of
an important removal process. The bound- the spring and fall blooms and the partic-
aries of the bay would be a sink for 234Th ulate 234Th residence times are long because
and, as a result, the vertical flux would be the trap fluxes for those two months were
less than expected. It is n.ot possible to re- unusually low. The connection between
Dabob Bay 234Thscavenging 313
these low fluxes and the bloom conditions DOWNS,J. N., AND C. J. LORENZEN. 1985. Carbon:
is not clear. The residence times of TSM pheopigment ratios of zooplankton fecal pellets as
an index of herbivorous feeding. Limnol. Gcean-
and particulate 234Th agree well, suggesting ogr. 30: 1024-1036.
that 234Th is a good tracer for the mass flux FROST,B. W. 1980. The inadequacy of body size as
at this location. During 1977 and early 1978 an indicator of niches in the zooplankton. Am.
the residence time of particulate organic C Sot. Limnol. Oceanogr. Spec. Symp. 3: 742-753.
(POC) ranged between 8 and 36 d with lon-
FURLONG, T. 1986. Sediment geochemistry of pho-
ger residence times in spring and early sum- tosynthetic pigments in oxic and anoxic marine
mer (Bennett 1980). Similarities in the time and lacustrine sediments: Dabob Bay, Saanich In-
scale and seasonal variation pattern be- let, and Lake Washington. Ph.D. thesis, Univ.
tween the residence time of particulate 234Th Washington. 2 14 p.
and POC implies that 234Th may also be a HEDGES, J.I.,W.A. CLARK,ANDG.C.COWIE. 1988a.
Organic matter sources to the water column and
useful tracer for organic C in coastal envi- surficial sediments of a marine bay. Limnol.
ronments like Dabob Bay. Oceanogr. 33: 1116-l 136.
- -,AND-. 1988b. Fluxes and reac-
References ti;ities of organic matter in a coastal marine bay.
ANDERSON,R. F., M. P. BACON,AND P. G. BREWER. Limnol, Oceanogr. 33: 1137-l 152.
1983. Removal ofz30Th and 231Pa from the ocean HONEYMAN, B. D., L. S. BALISTRIERI,AND J. W.
margins. Earth Planet. Sci. Lett. 66: 73-90. MURRAY. 1988. Oceanic trace metal scavenging:
-, AND A. P. FLEER. 1982. Determination of The importance of particle concentration. Deep-
natural activities and plutonium in marine particle Sea Res. 35: 227-246.
material. Anal. Chem. 54: 1142-l 147. HUH, C.-A., AND T. M. BEASLEY. 1987. Profiles of
BACON, M. P., AND R. F. ANDERSON. 1982. Distri- dissolved and particulate thorium isotopes in the
bution of thorium isotopes between dissolved and water column of coastal southern California. Earth
particulate forms in the deep sea. J. Geophys. Res. Planet Sci. Lett. 85: l-10.
87: 2045-2056. HUNTER,K.A.,D.J. HAWKE,AND L.K. CHOO. 1988.
BENNETT,J. T. 1980. The biogeochemical signifi- Equilibrium adsorption of thorium by metal ox-
cance of zooplankton fecal material in a biologi- ides in marine electrolytes. Geochim. Cosmo-
cally productive, temperate fjord. Ph.D. thesis, chim. Acta 52: 627-636.
Univ. Washington 258 p. KAUFMAN, A.,Y.-H. LI, AND K.K. TUREKIAN. 1981.
BHAT, S. G., S. KRISHNASWAMI, LAL, RAMA, AND
D. The removal rates of 234Thand 228Thfrom waters
W. S. MOORE. 1969. 234Th/23*U ratios in the of the New York Bight. Earth Planet. Sci. Lett. 54:
ocean. Earth Planet Sci. Lett. 5: 483-491. 385-392.
BRULAND, K. W., AND K. H. COALE. 1986. Surface KNAUER,G. A.,J.H. MARTIN,AND K. W. BRULAND.
water 234Th/238Udisequilibria: Spatial and tem- 1979. Fluxes of particulate carbon, nitrogen and
poral variations of scavenging rates within the Pa- phosphorus in the upper water column of the
cific Ocean, p. 159-172. In J. D. Burton et al. northeast Pacific. Deep-Sea Rcs. 26: 97-108.
[eds.], Dynamic processes in the chemistry of the Ku,T.L.,K.G. KNAUSS,ANDG.G.MATHIEW. 1977.
upper ocean. Plenum. Uranium in the open ocean: Concentration and
BUESSELER, V.,M. P. BACON,J. K. COCHRAN,AND isotopic composition. Deep-Sea Res. 24: 1005-
H. D. LIVINGSTON. In press. Carbon and nitrogen 1017.
export during the JGOFS North Atlantic Bloom C.
LORENZEN, J., F. R. SHUMAN,AND J. T. BENNETT.
Experiment estimated from 234Th:238U disequili- 198 1. In situ calibration of a sediment trap. Lim-
bria. Deep-Sea Res. 39. nol. Oceanogr. 26: 580-585.
R.,M. L. PETERSON, J.T. BENNETT. -, ANDN. A. WELSCHMEYER.1983. The in situ
1985. 210Pb-derived sediment accumulation and sinking rates of herbivore fecal pellets. J. Plankton
mixing rates for the greater Puget Sound region. Res. 5: 929-933.
Mar. Geol. 64: 29 l-3 12. McKEE,B. A.,D.J. DEMASTER,ANDC. A. NITTROUER.
--- AND B. L. K. SOMAYAJULU. 1984. The use of 234Th/238Udisequilibrium to
1984. Mixing and’ cycling of uranium, thorium examine the fate of particle-reactive species on the
and 210Pbin Puget Sound sediments. Geochim. Yangtze continental shelf. Earth Planet. Sci. Lett.
Cosmochim. Acta 48: 1949-l 963. 68: 43 l-442.
COALE,K. H., AND K. W. BRULAND. 1985. 234Th:238U - -,AND-. 1986. Temporal vari-
disequilibria within the California Current. Lim- ability in the partitioning of thorium between dis-
nol. Oceanogr. 30: 22-33. solved and particulate phases on the Amazon shelf:
-, AND - . 1987. Oceanic stratified eu- Implications for the scavenging of particle-reac-
photic zone as elucidated by 234Th:238U disequi- tive species. Cont. Shelf Rcs. 6: 87-106.
libria. Limnol. Oceanogr. 32: 189-200. MARTIN, J.H.,G. A. KNAUER, D.M. KARL, AND W.
COWIE, G. L., AND J. I. HEDGES. 1984. Carbohydrate W. BROENKOW.1987. VERTEX: Carbon cycling
sources in a coastal marine environment. Geo- in the northeast Pacific. Deep-Sea Res. 34: 267-
chim. Cosmochim. Acta 48: 2075-2087. 285.
314 Wei and Murray
MINAGAWA, M., AND S. TSDIOGAI. 1980. Removal TSIJIVOGAI, Y. NOZAKI, AND M. MINAGAWA. 1974.
of 234Th from a coastal sea: Funka Bay, Japan. Behavior of heavy metals and particulate matters
Earth Planet. Sci. Lett. 47: 51-64. in seawater expected from that of radioactive nu-
MURRAY, J. W. 1975~. The interaction of metal ions elides. J. Oceanogr. Sot. Jpn. 30: 251-259.
at the manganese dioxide-solution interface. Geo- --, K. TAGUCHI, AND K. HARADA. 1986. Sea-
chim.. Cosmochim. Acta 39: 505-5 19. sonal variation in the difference between observed
-. 11975b. The interaction of cobalt with hy- and calculated particulate lluxes of Th-234 in Fun-
drous manganese dioxide. Geochim. Cosmochim. ka Bay, Japan. J. Oceanogr. Sot. Jpn. 42: 91-98.
Acta 39: 635-647. ---, M. UEMATSU,N. TANAKA, AND K. HARADA.
AXVD G. BREWER. 1977. The mechanisms 1980. A sediment trap experiment in Funka Bay,
of rejmoval of iron, manganese and other trace Japan: “Upward flux” of particulate matter in sea-
metals from seawater, p. 29 l-326. Zn G. P. Glasby water. Mar. Chem. 9: 321-334.
[ccl.], Marine manganese deposits. Elsevier. WEIXHMEYER, N. A. 1982. The dynamics of phy-
NOZAKI, Y., H.-S. YANG, AND M. YAMADA. 1987. toplankton pigments: Imlplications for zooplank-
Scavenging of thorium in the ocean. J. Geophys. ton grazing and phytoplankton growth. Ph.D. the-
Res. ‘92: 772-778. sis, Univ. Washington. 1’77 p.
OHTANI, K., AND Y. AKIBA. 1970. Studies on the ---, M.
A. E. COPPING, VERNET, ANDC. J. LORENZEN.
change of the hydrographic conditions in the Fun- 1984. Diel fluctuation in :zooplankton grazing rate
ka Bay. Bull. Fat. Fish. Hokkaido Univ. 20: 303- as determined from the downward vertical flux of
312. pheopigments. Mar. Biol. 83: 263-270.
SANTSCHI, H., Y.-H. LI, AND J. BELL. 1979. Natural
P. ----> AND C. J. LORENZEN. 1985. Role of herbivory
radionuclides in the water of Narragansett Bay. in controlling phytoplank.ton abundance: Annual
Earth Planet. Sci. Lett. 45: 201-213. pigment budget for a temperate marine fjord. Mar.
SHUMAN,F. R. 1978. The fate of phytoplankton chlo- Biol. 90: 75-86.
rophyll in the euphotic zone: Washington coastal
waters. Ph.D. thesis, Univ. Washington. 250 p.
TANAKA, N., Y. TAKEDA, AND S. TXJNWAI. 1983. Submitted 21 April 1989
Biological effect on removal of Th-234, PO-~10
and Pb-210 from surface water in Funka Bay, Ja- Accepted: 7 November 1990
pan. Geochim. Cosmochim. Acta 47: 1783-1790. Revised: 2 October 1991