Variation of initial
230
Th/
232
Th and limits of high precision
U–Th dating of shallow-water corals
Chuan-Chou Shen
a,*, Kuei-Shu Li
a, Kerry Sieh
b, Danny Natawidjaja
b, Hai Cheng
c,
Xianfeng Wang
c, R. Lawrence Edwards
c, Doan Dinh Lam
d, Yu-Te Hsieh
a,
Tung-Yung Fan
e, Aron J. Meltzner
b, Fred W. Taylor
f, Terrence M. Quinn
f,g,
Hong-Wei Chiang
a, K. Halimeda Kilbourne
g,haDepartment of Geosciences, National Taiwan University, Taipei, Taiwan, ROC bTectonics Observatory, California Institute of Technology, Pasadena, CA, USA cDepartment of Geology and Geophysics, University of Minnesota, MN, USA dInstitute of Geology, Vietnamese Academy of Science and Technology, Hanoi, Vietnam
eNational Museum of Marine Biology and Aquarium, Pingtung, Taiwan, ROC
fInstitute for Geophysics, John A. and Katherine G. Jackson School of Geosciences, University of Texas at Austin, Austin, TX, USA gCollege of Marine Science, University of South Florida, St. Petersburg, FL, USA
h
Physical Sciences Division R/PSD1, NOAA, Earth System Research Laboratory, Boulder, CO, USA Received 4 September 2007; accepted in revised form 5 June 2008; available online 24 June 2008
Abstract
One hundred eighty U–Th data, including 23 isochrons on 24 pristine modern and Holocene corals and 33 seawater sam-ples, were analyzed using sector-field mass spectrometry to understand the variability of initial230Th/232Th (230Th/232Th0).
This dataset allows us to further assess the accuracy and precision of coral230Th dating method. By applying quality control,
including careful sampling and subsampling protocols and the use of contamination-free storage and workbench spaces, the resulting low procedural blanks give an equivalent uncertainty in age of only ±0.2–0.3 yr for 1–2 g of coral sample. Using site-specific230Th/232Th0values or isochron techniques, our study demonstrates that corals with an age less than 100 yrs can be 230
Th-dated with precisions of ±1 yr. Six living subtidal coral samples were collected from two continental shelf sites, Nanwan off southern Taiwan in the western Pacific and Son Tra off central Vietnam in the South China Sea; one coral core was drilled from an open-ocean site, Santo Island, Vanuatu, in the western tropical Pacific; and modern and fossil intertidal coral slabs, 17 in total, were cut from six sites around the islands of Simeulue, Lago, North Pagai and South Pagai of Sumatra in the eastern Indian Ocean. The results indicate that the main source of thorium is the dissolved phase of seawater, with variation of230Th/232Th0depending on local hydrology. With intense input of terrestrial material, low
230
Th/232Th0atomic ratios of
4.9 106and 3.2 106with a 10% variation are observed in Nanwan and Son Tra, respectively. At the Santo site, we find
a value of 5.6 106at 4 horizons and one high value of 24 106in a sample from AD 1974.6 ± 0.5, likely due to the upwelling of cold water during a La Nin˜a event between AD 1973 and 1976. The natural dynamics of230Th/232Th0recorded
in the intertidal corals at sites in the Sumatran islands are complicated so that this value varies significantly from 3.0 to 9.4 106. Three of the 141 modern coral230Th ages differ from their true ages by23 to +4, indicating the presence of
detri-tal material with anomalous230Th/232Th values. Duplicate measurement of coeval subsamples is therefore recommended to verify the age accuracy. This improved high precision coral230Th dating method raises the prospects of refining the age mod-els for band-counted and tracer-tuned chronologies and of advancing coral paleoclimate research.
Ó 2008 Elsevier Ltd. All rights reserved.
0016-7037/$ - see front matterÓ 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.gca.2008.06.011
* Corresponding author. Fax: +886 2 3365 1917. E-mail address:[email protected](C.-C. Shen).
www.elsevier.com/locate/gca
Available online at www.sciencedirect.com
1. INTRODUCTION
Geochemical and isotopic compositions in shallow-water coralline aragonite have been widely used as proxies to reconstruct paleoclimatic conditions (e.g., Wellington and Dunbar, 1995; Shen et al., 1996; Beck et al., 1997; Gagan et al., 1998; McCulloch et al., 1999, 2003; Hendy et al., 2002; Cutler et al., 2003; Shen et al., 2005a; Thomp-son and Goldstein, 2005). Since the 1990s, weekly to monthly resolution coral climate records over decadal, centennial, or millennial time scales have been published (e.g.,Gagan and Chivas, 1995; Crowley et al., 1997; Evans et al., 1998; Linsley et al., 2000; Urban et al., 2000; Hendy et al., 2002; Cobb et al., 2003a; McCulloch et al., 2003; Kilbourne et al., 2004; Sun et al., 2004; Nyberg et al., 2007). For centuries-old modern and young fossil corals, age uncertainties of ±1–5 yrs limit our ability to accu-rately determine the timing of climatic events and the pre-cisely cross-correlate coral records to other high resolution proxy records.
Coral chronology can be determined by bulk annual density band counting (e.g.,Dodge and Brass, 1984; Cole and Fairbanks, 1990; Crowley et al., 1997; Linsley et al., 1999; Kilbourne et al., 2004), or by tuning tracer records using methods adapted from dendrochronology and sedi-ment core chronostratigraphy (e.g., Gagan and Chivas, 1995; Linsley et al., 1999; Cobb et al., 2003b; Hendy et al., 2003). Difficulty arises when using these methods for precise absolute chronological control in the following cases: (1) living corals with banding discontinuities caused by climatic and/or tectonic anomalies (e.g., Sieh et al., 1999; Hendy et al., 2003; Natawidjaja et al., 2004), (2) cor-als living in the equatorial oceans with intrinsically small seasonal climatic cycles (as low as 0–2°C of seasonal tem-perature change), and (3) dead coral heads and fossils with-out absolute ages and/or mortality event-related age markers (Glynn et al., 1983; Yu et al., 2004).
The230Th dating method provides an ideal absolute chro-nological tool for coral because of high uranium levels of 2–3 parts per million (ppm) in their skeletons (Bateman, 1910; Barnes et al., 1956; Broecker and Thurber, 1965; Kaufman and Broecker, 1965; Edwards et al., 1987, 1988, 2003). The
230
Th age equation, including initial230Th/232Th, is reported
byEdwards et al. (2003)(modified fromBateman (1910) and
Broecker (1963)) 230Th 238U ¼ 1 þ 232Th 238U 230Th 232Th i 1 ek230t þd 234U m 1000 k230 k230 k234 1 eðk234k230Þ; ð1Þ
where all isotope ratios are activity ratios, the k’s are decay constants, t is the230Th age, and (230Th/232Th)iis the initial 230
Th/232Th ratio. The234U/238U ratio has been formulated into d-notation, which denotes the fractionational enrich-ment in the234U/238U ratio at secular equilibrium in parts
per thousand. The observed value is given by d234Um=
{[(234U/238U)
m/(234U/238U)eq] 1} 103where (234U/238U)m
is the measured activity ratio and (234U/238U)eqis the ratio
of 1 at secular equilibrium (Edwards et al., 1987). As
(232Th/238U) can be measured, an accurate 230Th age can be determined if (230Th/232Th)iis known.
The first high accuracy young coral230Th dating tech-niques with precisions of ±3 yrs were developed using ther-mal ionization mass spectrometry byEdwards et al. (1988). Their study suggested the initial230Th (230Th0) is not
signif-icant within errors of ±3 yrs in typical shallow-water corals with 232Th levels less than 100 parts per trillion (ppt) and low initial230Th/232Th (230Th/232Th) of 2–4 106(atomic ratio, hereafter). However, development of a coral 230Th dating method with a precision better than ±1–2 yrs is difficult with uncertainties mainly affected by the thorium levels and the uncertainty of the correction for 230Th0
(Zachariasen et al., 1999; Cobb et al., 2003b). For example, uncertainties of 25%, 50%, and 100% for a 230Th/232Th0
value of 4 106, give approximate age errors of ±3, ±5, and ±10 yrs, respectively, for a decades-centuries-old coral with 2.4 ppm uranium and 1000 ppt thorium.
The initial thorium in the coralline structure is influ-enced both by scavenging processes between seawater and the coral skeleton, and by thorium association with detrital material incorporated into the crystal matrix (Cobb et al.,
2003b; Edwards et al., 2003).Cobb et al. (2003b)proposed
different potential sources of coral 230Th0, including (1)
wind-blown dust with a 230Th/232Th ratio of 4 106; (2) surface seawater containing dissolved and particulate tho-rium with a 230Th/232Th ratio of 5–10 106; (3) deep seawater Th with a 230Th/232Th ratio of up to 2 104 (Moran et al., 2002); and (4) carbonate sands with
230Th/232Th ratios as high as 1 102. Different sources
and the non-conservative property of thorium in seawater (e.g., Broecker et al., 1973) could result in diverse coral
230
Th/232Th0 ratios. The range of 230Th/232Th0 values
observed in intertidal corals at different sites near Sumatra is 6.5 ± 6.5 106(Zachariasen et al., 1999). High values of 10–20 106are documented in Palmyra living corals (Cobb et al., 2003b). Modern Bahamas corals have 60–80 106of230
Th/232Th0values, intermediate between
those of normal surface water and of bank-top water (Robinson et al., 2004). Clearly, the 230Th0 incorporated
into a growing aragonitic skeleton needs to be well-constrained or its uncertainty may lead to significant error in the230Th ages of young corals.
The precision and accuracy of coral230Th dating
tech-niques are affected by not only232Th content and variabil-ity of 230Th
0, but also by the 230Th introduced during
sample collection and treatment. The previously published lowest procedural 230Th blank, 0.0015 ± 0.0015 femto-mole (fmol) (Cheng et al., 2000), is equivalent to an uncertainty of ±0.4 yr. Coupled with instrumental constraints (Goldstein and Stirling, 2003), the best coral U–Th dating procedures yielded an accuracy and preci-sion of ±2–3 yrs (after Edwards et al., 1987; Edwards et al., 2003).
Four novel approaches were taken in this study to better understand the natural variability of coral230Th/232Th0in
the oceans and to establish a high precision 230Th dating method. First, chemical and analytical techniques were re-fined. The procedural 230Th blank was reduced to as low as 0.0008 fmol, corresponding to an error of ±0.2 yr for a
1-g coral sample. Analytical uncertainty in 230Th using inductively coupled plasma sector-field mass spectrometry (ICP-SF-MS) techniques (Shen et al., 2002) was reduced to an equivalent age uncertainty of about ±0.2–0.3 yr. Sec-ond, concentrations and isotopic compositions of dissolved and particulate thorium were analyzed to distinguish the contributions of the two phases to the coral skeletal matrix. Third, the variability of230Th/232Th0at different
hydrolog-ical settings was evaluated. Subtidal and intertidal modern and fossil corals, 24 in total, from the continental shelf sites of Nanwan in the western Pacific and Son Tra in the South China Sea, the open-ocean site of Santo in the western tropic Pacific, and six sites among the Sumatran islands in the eastern Indian Ocean, were analyzed to investigate the influences of terrestrial input,230Th ingrowth from the underlying substrate, and different water masses on
230
Th/232Th0. Finally, comparisons of the corals’230Th ages
and independently-derived absolute ages, as well as isochron plots, were used to constrain spatial and temporal variations of230Th/232Th0 values on annual to millennial
timescales at the selected sites.
2. MATERIALS AND METHODS 2.1. Field collection
Seawater and massive Porites coral samples were collected at sites with different hydrological settings from Nanwan, Taiwan, Son Tra Island, central Vietnam, southern Espiritu Santo Island, Vanuatu, and Sumatran islands (Figs. 1, A1–A4).
2.1.1. Nanwan
Nanwan (21°550N, 120°470E) is a semi-enclosed basin on
the southern tip of Taiwan (Fig. A1). Two subtidal coral heads, NW0310 and NW0402, each 30–40 cm in diameter, were collected from the water intake channel of the nuclear power plant located at the northeastern corner of Nanwan in October 2003 and February 2004 at depths of 4 and 2 m, respectively.
A hydrological diagram of seawater Sr/Ca versus (vs.) salinity shows that Nanwan water is composed mainly of three endmembers: 25% offshore surface water, 75% tidally-induced upwelled subsurface cold water from depths of 100–200 m offshore, and an additional 0–2.5% fresh water in the summer (Lee et al., 1997; Shen et al., 2005b). In order to understand the influences of upwelled seawater and tide height on the local seawater232Th level and230Th/232Th
ra-tio, 21 consecutive 500-ml samples of filtered seawater were collected from the water intake channel of the Third Nuclear Power Plant, Nanwan, on July 31 and August 1, 2004, at a depth of 1 m (Fig. A1andTable A1). Seawater was filtered with an acid-cleaned 0.45-lm acetate cellulose filter and stored acidified in an acid-cleaned polyethylene bottle by adding 0.5–1 g 14 N HNO3in the field (Fig. A1). Eight
sam-ples of suspended particulate matter, filtered from 5 L seawa-ter with 0.45-lm acetate cellulose filseawa-ters, were collected between October 28 and 30, 2004, at the same position where filtered seawater samples were collected (Fig. A1andTable A2). Suspended particulate matter samples were stored in acid-cleaned polyethylene bags at 5°C in a refrigerator be-fore chemical analysis. The same seawater sampling process was applied at other sites. One calibrated underwater
Fig. 1. Map of sample collection sites. Modern and fossil corals were collected from Nanwan, Taiwan, Son Tra Island, Vietnam, southern Espiritu Santo Island, Vanuatu, and the Sumatran islands. Surface seawater samples were collected from Nanwan, Son Tra, and North Pagai of the Sumatran islands.
thermometer with precision of 0.05°C was placed next to the NW0402 coral sampling site to record sea surface tempera-ture and monitor the timing of cold upwelled water intrusion. Tide height data monitored at Houbihu, 2 km south of the sample collection site (Fig. A1), are from the Central Weath-er Bureau.
2.1.2. Son Tra and Santo
One living 35-cm Porites coral head, ST0506, and 2 fil-tered 1-L seawater samples from a depth of 4 m below sea le-vel were collected at Son Tra Island, a near-shore island in central Vietnam (16°130N, 108°120E) on June 14, 2005
(Fig. A2). The site is located at the north tip of Vung Da Nang Bay with terrestrial material influxes mainly delivered to the southern corner of the bay by the Han River. Two 1-L filtered seawater samples were also collected from a depth of 2 m at the same site.
The Santo subtidal coral core, 92MC, is 130 cm in length and 8 cm in diameter. It was drilled from a Porites lutea head living at a depth of 1.5 m at an open-ocean site, with notable lack of riverine influence, in the passage be-tween Malo Island and Espiritu Santo Island (15°420S,
167°120E), in the western tropical Pacific Ocean, in June
1992 (Kilbourne et al., 2004) (Fig. A3). 2.1.3. Sumatran islands
The Sumatran islands are located along the Sunda sub-duction zone, an area associated with frequent earthquakes (e.g.,Briggs et al., 2006; Subarya et al., 2006). The subduction zone and Sumatran fault separate the Indian and Australian plates from the Southeast Asian plate (Fitch, 1972; McCaffrey, 1991). Modern and fossil Porites slabs with thicknesses of 7–9 cm were cut with a chain saw from microatolls on shallow reefs within the intertidal zone at sites near Lewak (2°560N,
95°480E) and Gusong Bay (2°230N, 96°200E) on Simeulue
Island; Lago (0°030N, 94°320E) in the Batu islands; North
Pagai Island (2°350S, 100°060E); and Bulasat (3°07.60S,
100°18.70E), and Saumang (3°07.70S, 100°18.60E) on South
Pagai Island (Fig. A4). One slab, LWK05, was collected from Lewak, and two slabs, GSG05YNG and GSG05OLD, were collected from Gusong Bay between the 1st and 3rd of June 2005; the two sites are 80 km apart, on the northern tip and the southern southwest coast of Simeulue Island, respec-tively. Two contiguous slabs, LG99A1 and LG00A1, were collected from the northeastern flank of Lago in 1999 and 2000, respectively (Natawidjaja et al., 2004). The Lago site is a landward pool characterized by a wide intertidal flat with coral rubble on a carbonate-enriched sandy substrate. Two slabs, NP00A1 and NP00A2, were collected from the eastern coast of North Pagai. Nine samples were collected from microatolls at Bulasat and Saumang on the southwestern coast of South Pagai. Following the same sampling proce-dure as at Nanwan, two 500-ml filtered seawater samples were collected at depths of 1 and 2 m at the coral collection site in North Pagai in January 2004.
2.2. Sample storage
General storage rooms for Quaternary carbonates can have230Th blanks higher than 0.002 fmol, which could bias
dates for coral samples by as much as 30 yrs (Table A3). This cross-contamination, mainly from fine powder of Qua-ternary samples with high230Th content, may not be easy to remove from a porous coral skeleton, even using ultrasonic cleaning methods. Subsequently, we now bag coral samples individually and isolate young and old corals from each other in our storage rooms. In addition, we isolate coral samples from other types of natural carbonate samples, such as speleothem and tufa. These procedures effectively keep storage-related 230Th background levels to 2 104
fmol or less.
2.3. Coral subsampling
Modern corals were subsampled in a separate class-100 laminar-flow clean working bench. For Nanwan, Son Tra, and Santo coral slabs, subsamples 0.5–1.5 cm in width were cut using a surgical blade, along growth layers, based on X-ray photographs. The widths are equivalent to 0.5–1 yr of coral growth. For Sumatran coral slabs, subsamples were collected by either cutting with a blade or drilling with a hole-drill bit 1 cm in diameter. The drilled cylindrical subs-amples were 0.8 cm in diameter. Weights of coral subsam-ples ranged from 0.3 to 3 g (Tables A3–A12). All were ultrasonicated with deionized water 4–5 times until there was no visible powder or detrital material, and then dried at 70°C. To avoid possible cross-contamination in the pre-treatment process, all steps were performed on class-100 laminar-flow benches.
2.4. Chemistry
U–Th chemistry was performed in the High-precision Mass Spectrometry and Environment Change Laboratory (HISPEC) of the Department of Geosciences, National Taiwan University (NTU) and in the Minnesota Isotope Laboratory of the Department of Geology and Geophysics, University of Minnesota (UMN). Chemistry for seawater samples was conducted following the procedures described
inMoran et al. (2002) and Shen et al. (2003). Coral
subsam-ples were prepared with chemical methods similar to those described by Edwards et al. (1988) and Shen et al. (2002,
2003). Samples were spiked with a229Th–233U–236U tracer.
Uranium and thorium were separated with Fe coprecipita-tion and anion-exchange chromatography. The uranium and thorium aliquots were dissolved in 1% HNO3+
0.005 N HF for instrumental measurements (Shen et al., 2002).
2.5. Instrumental analysis
Determinations of uranium and thorium isotopic com-positions and concentrations were performed using two ICP-SF-MS: a Thermo Electron ELEMENT II, housed at the Department of Geosciences, NTU, and a Finnigan ELEMENT, at the Minnesota Isotope Laboratory, UMN, with analytical techniques described by Shen et al. (2002). Spiked NBL-112A standard solution was measured to correct for multiplier intensity bias every day (Cheng et al., 2000). Pairs of the separated thorium and uranium
fractions were sequentially analyzed by ICP-SF-MS. Both thorium and uranium mass biases were normalized to a
236
U/233U atomic ratio of 1.01057 ± 0.00050 (Cheng et al., 2000). The accuracy of measurements of different tho-rium standards within error of accepted values suggests that there is no significant mass bias difference between thorium and uranium (Shen et al., 2002). Typical230Th+ion beam
intensities were just tens of counts per second (cps) for cor-als younger than 20 yrs and for filtered seawater and sus-pended particulate material. Sources of background noise include multiplier dark noise, memory blanks, tailing of
238
U+and232Th+, and isobaric spectral interferences. The dark noise of the multiplier (0.05–0.1 cps), instrumental memory blanks (typically 0.1–0.2 cps at m/z = 230, for exam-ple), and background from the238U+and232Th+tails were subtracted during off-line data processing. Spectral interfer-ences of 0–10 cps at m/z = 229–237 could be generated from polyatomic organics and/or complexes, which cause a high noise/signal ratio of 20–50% for low230Th analyses. This contamination was effectively removed by oxidation treat-ment with perchloric acid in chemistry, reducing the interfer-ences to less than 0.1 cps at m/z = 230 (Shen et al., 2002). Resultant analytical uncertainty in 230Th corresponded to
an age uncertainty of 0.2–0.3 yrs. All errors given are two standard deviations (2r) unless otherwise noted.
2.6. Blanks
Procedural blanks were 0.02 ± 0.01 pmol 238U, 0.003 ± 0.003 pmol232Th, and 0.0008 ± 0.0008 fmol230Th
for coral. The low230Th procedural blank corresponds to an age uncertainty of ±0.1–0.2 yr. The isotopic composi-tion in the spike solucomposi-tion was carefully re-quantified. The value of the230Th/229Th ratio in the229Th–233U–236U spike solution is 0.000050 ± 0.000002. The uncertainty corre-sponds to an error of ±0.1–0.2 yr for 1–2-g coral samples. The improved chemical procedure results in an overall age error of only ±0.2–0.3 yr. Procedural blanks of 238U,
232Th, and230Th for filtered seawater samples had the same
as the values for corals. For the particulate fraction, the use of an acetate cellulose filter caused a relatively high proce-dural232Th blank of 0.06 ± 0.03 pmol.
3. RESULTS AND DISCUSSION 3.1. Uranium and thorium data
U–Th data for all seawater and coral samples are listed
inTables A1–A12. Uncertainties in the 235U,234U,232Th,
and230Th data are calculated at the 2r level and include corrections for blanks, multiplier dark noise, abundance sensitivity, and errors associated with quantifying the isoto-pic composition in the spike solution. The238U level was calculated from measurement of235U and the assumed
nat-ural238U/235U atomic ratio of 137.88 (Cowan and Adler, 1977; Steiger and Jager, 1977).
The decay constants used are 9.1577 106yr1 for 230
Th and 2.8263 106yr1 for 234U (Cheng et al.,
2000), and 1.55125 1010yr1 for 238U (Jaffey et al.,
1971). Ages are corrected for 230Th0 by estimating
230
Th/232Th0 with 3-D (232Th/238U-230Th/238U–234U/238U)
isochron techniques and with seawater data for coral sam-ple ST0506 from Son Tra. The initial d234U value is calcu-lated from the measured d234U value using the corrected
230Th age. Intercepts, isochron ages, and isochron plot
errors were calculated with an Excel macro, Isoplot 3.00, by K.R. Ludwig of the Berkeley Geochronology Center, California, USA (Ludwig and Titterington, 1994; Ludwig, 2003).
3.2. Seawater uranium and thorium at Nanwan and other sites
Dissolved 238U concentration is 3.1 ± 0.1 ppb for
Nanwan seawater (Fig. 2andTable A1) with a salinity of 33.8–34.0, measured by H.J. Lee of the Department of Marine Environmental Informatics, National Taiwan Ocean University. Our observations support the conserva-tive behavior of the238U concentration in this coastal ocean (Chen et al., 1986; Robinson et al., 2004), although Taiwan experiences a high erosion rate of 3–6 mm/yr and supplies 300–500 Mt/yr of suspended sediment to the ocean (Dadson et al., 2003). Particulate238U concentration ranges
from 0.6 to 2.0 ppt (Fig. 3 and Table A2), representing 0.02–0.06% of the respective dissolved concentrations.
For Nanwan water, the measured dissolved d234U of 147.4 ± 2.0 (Fig. 2andTable A1) is within the open-ocean interval of 146 ± 2 (Chen et al., 1986). d234U of suspended particulate matter is 34 ± 36 in Nanwan seawater (Fig. 3
andTable A2). The calculated d234U of bulk (or unfiltered)
232 Th (ppt) 0 4 8 T ( OC) 24 26 28 30 Tide Height (m) -1.2 0.0 1.2 δ 234 U 138 144 150 156 238 U (ppb) 2.8 3.0 3.2 Time (hour) 0 6 12 18 24 30 36 42 48 230 Th/ 232 Th (x 1 0 -6) 1 2 3 4 5 6 3.1± 0.1 ppb 147.4 2.0± 2.4 2.9 ppt± 4.0 0.5 x 10± -6
Nanwan: Filtered seawater a c b f e d
Fig. 2. Time series of (a) tide height, (b) temperature, (c) 238U concentration, (d) d234U, (e) 232Th concentration, and (f) 230Th/232Th ratio for filtered Nanwan seawater collected from July 31, 06:00 AM to August 1, 10:00 PM, 2004. Means and 2r errors of U–Th data are given.
seawater is 146.1–146.5, which is not significantly different from dissolved d234U values. With an analytical precision of 1–2&, the influence of suspended particulate matter on the dissolved d234U value is not observable. The dissolved d234U is 146.6 in North Pagai and 145.0 in Son Tra (Table
A1), both within the range observed in the open ocean (Chen et al., 1986).
Consecutive measurements of dissolved 232Th range from 4–7 ppt at low tide to 0.7–3 ppt at high tide (Fig. 2
andTable A1), 10–100 times higher than the open-ocean
bulk values from the central Pacific (Roy-Barman et al.,
1996). The averaged particulate 232Th of 3.6 ppt (Fig. 3)
is 150% of the dissolved fraction. This proportion is higher, by 10–30%, than those at open-ocean sites in the Labrador Sea and Atlantic (Moran et al., 2002). This is likely the result of high thorium fluxes associated with terrestrial material input at Nanwan. The dissolved
230
Th/232Th ratio, 4.0 ± 0.5 106, is slightly higher than that (3.0 ± 0.7 106) of the particulate fraction (Fig. 3). It indicates that thorium is coming not only from terrestrial sources, but also from the open ocean which has high
230
Th/232Th (Roy-Barman et al., 1996) or is derived from the ingrowth of230Th by uranium decay (e.g., Robinson et al., 2004).
There is no observable periodic trend of230Th/232Th ra-tios for either the dissolved fraction or suspended particu-late matter (Figs. 2 and 3). For neither fraction do
230Th/232Th ratios correlate with tide height or cold water
intrusion, indicating that near-shore water 230Th/232Th is
not affected by either tide height or upwelled subsurface cold water. Plots of 230Th/232Th vs. 1/232Th for both the dissolved and suspended particulate fractions support the observation of no additional distinguishable sources of tho-rium at Nanwan (Fig. 4).
3.3. Uranium in corals
238
U concentrations range from 2.4 to 2.7 ppm for mod-ern Porites coral samples from Nanwan, Son Tra, and San-to (Tables A3–A6). Porites coral skeletal Ca content is 38.2–38.5% by mass in the South China Sea and western Pacific (Sun et al., 1999). The calculated molar coral U/Ca ratio, [U/Ca]coral, is 1.0–1.2 106. Seawater Ca
con-centration is 10.3 mmol/g at a salinity of 35 (Chen, 1990; Shen et al., 1996, 2005b). We calculate a molar U/Ca ratios for Nanwan water to be 1.3 106. The distribution coefficient of U between Porites coral and seawater at the three sites is estimated as: D[U/Ca] = [U/Ca]coral/
[U/Ca]sea= 0.80–0.90, consistent with the values reported
by Swart and Hubbard (1982) and Shen and Dunbar
(1995). For all corals from Nanwan, Son Tra, Santo, and the Sumatran islands, molar U/Ca ratios are within 30%, agreeing with previous observations (Min et al., 1995; Shen and Dunbar, 1995).
For Nanwan corals, d234U averages 147.7 ± 3.2, match-ing the local dissolved value of 147.6 ± 2.7 (Fig. 2). d234U in corals of Son Tra and North Pagai also matches the seawa-ter value of 145–147 (Table A1). There is no distinguishable difference of coral d234U from dissolved value at the sites with tremendous input of terrestrial material, which sup-ports the conclusion that marine uranium is incorporated into the coral skeleton without isotopic fractionation and that the initial d234U is a reliable parameter for estimating diagenesis (Edwards et al., 2003).
3.4. Thorium in corals 3.4.1. Distribution coefficient
Coral skeletal 232Th content is a function of
oceano-graphic setting. For Porites corals, it increases from
0 40 80 0 5 10 6 12 18 24 30 36 42 48 1 2 3 4 5 22 24 26 28 -0.6 0.0 0.6 0.0 0.8 1.6 2.4 232 Th (ppt) T ( OC) Tide Height (m) δ 234 U 238 U (ppt) 230 Th/ 232 Th (x 1 0 -6 ) Time (hour) 1.1 1.0 ppt 34 36 3.6±3.7 ppt 3.0 0.7 x 10-6 Nanwan: Suspended particulate matter
± ± ±
Fig. 3. Time series of (a) tide height, (b) temperature, (c)238U concentration, (d) d234U, (e) 232Th concentration, and (f) 230Th/232Th ratio for the suspended particulate matter of Nanwan seawater collected from October 28, 7:00 AM to October 30, 1:00 AM, 2004. Average values and 2r errors of U–Th data are listed by record. 1/[232 Th] (ppt-1 ) 0.0 0.5 1.0 1.5 3.0 4.0 230 Th/ 232 Th (x 1 0 -6) 1 2 3 4 5
Nanwan, filtered seawater Nanwan, suspended material North Pagai, filtered seawater Son Tra Island, filtered seawater
Fig. 4. A plot of230Th/232Th vs. 1/232Th for the dissolved fraction (solid cubes) and suspended particulate matter (gray cubes) of Nanwan seawater, dissolved fraction of Son Tra seawater (hollow circles), and dissolved fraction of intertidal seawater of North Pagai (hollow triangles).
10 s–100 s ppt in the open ocean, such as at the Santo site, to 100 s–1000 s ppt on the continental shelf, such as at the Son Tra and Nanwan sites, characterized by input of high-232Th terrestrial material. For the intertidal corals at sites in the Sumatran islands with similar geological set-tings, skeletal 232Th level varies from 100 s–10,000 s ppt. Even at the same site, such as Gusong Bay,232Th content
in Porites GSG05OLD is 3000–7800 ppt, 10 times higher than that in Porites GSG05YNG (Table A8). Tenfold var-iation in skeletal232Th levels of 1000 s–10,000 s ppt is also observed in eight modern and Holocene corals from Bula-sat, South Pagai. Unlike uranium that has a long residence time of 300–500 kyr (Dunk et al., 2002) and displays con-servative behavior (Ku et al., 1977; Chen et al., 1986), tho-rium is non-conservative and a residence time of 0.1–0.7 yr in surface water (Broecker et al., 1973; Okubo, 1982). These features, along with the fact that different sources of tho-rium have various230Th/232Th ratios (Cobb et al., 2003b), result in a wide range of coral 232Th levels in different hydrological settings.
Thorium partitioning, similar to uranium (Cross and Cross, 1983), appears to be species dependent. The molar Th/Ca ratio is 1.0 ± 1.2 109 in the dissolved fraction of seawater and 2.2 ± 2.2 1010in Porites from Nanwan.
A distribution coefficient, D[Th/Ca], of0.2 suggests Por-ites excludes thorium during growth (Edwards et al., 2003). Based on our measurements and previous reports (Edwards et al., 1987; Zachariasen, 1998; Zachariasen et al., 1999, 2000; Cobb et al., 2003b; Edwards et al., 2003; Natawidjaja et al., 2004; Robinson et al., 2004), Acropora corals have high D[Th/Ca] values =1. Thorium is effectively excluded by Goniastrea with an estimated D[Th/Ca] of 0.02 or less. 3.4.2. Dissolved phase of seawater and detrital materials
Three isochrons with ages of 4.1 ± 1.2, 5.1 ± 3.0 and 15.56 ± 0.56 yrs, for two Nanwan coral slabs, NW0310 and NW0402, are shown inFig. 5a. All regression y-intercept val-ues, with uncertainties of 1–2 106, overlap with each other at 5 106(Fig. 5a). An excellent fit to straight lines
illus-trates that there is no significant difference of the isotopic composition of associated thorium in the skeletal lattice be-tween the three growth bands of the two coral heads in the same hydrographic environment. All intercepts are consis-tent with a value of 4.0 ± 0.5 106in the dissolved fraction
of seawater (Fig. 2) and higher than the value of 3.0 ± 0.7 106 in the suspended particulate matter
(Fig. 3). A230Th/232Th0value of 4.7 ± 1.0 106for growth
banding of a modern coral, NP00A1 (Fig. 5f), from North Pagai captures the dissolved values of 4.0–4.3 106at that site. At the Son Tra site, using a dissolved230Th/232Th value
of 3.20 ± 0.32 106as the initial thorium isotopic
compo-sition, the230Th ages of three bands of the Son Tra Porites,
ST0506, match the absolute ages (Table A5).
Robinson et al. (2004)showed the seawater230Th/232Th
value should be used for230Th0correction with coral and
bulk seawater data. The similarity between the dissolved
230Th/232Th values and the respective initial ratios in corals
from Nanwan, North Pagai, and Son Tra, further supports that the idea that230Th
0is mainly from the dissolved
frac-tion of seawater. The data from all three sites show that the
230
Th age should be corrected for230Th0using the dissolved 230
Th/232Th ratio. Precise and accurate correction for
230
Th0content in coral can be achieved by understanding
the spatial and temporal variability of 230Th/232Th0; this
can be accomplished with isochron techniques as described in Sections3.5 and 3.6.
Three discordant ages were observed from isochrons (Fig. 5) and the 1:1 line of the230Th age vs. growth band age plot (Fig. 6). The 230Th age of one subsample, NW0310-5#3 (Table A4), is 3.8 ± 1.0 yrs older than the other three coeval subsamples. Two230Th/232Th–234U/232Th data points, 92MC-1#4 (Table A6) and NP00A1–1#5 (Table
A10), do not lie on isochrons (Fig. 5b and f). The230Th age of
92MC-1#4 is 5.0 ± 1.4 yrs older than the isochron age of 14.0 ± 1.1 yrs, constructed with the other three coeval subs-amples. The230Th age of NP00A1–1#5 is 23 ± 6 yrs younger
than the isochron age of 8.1 ± 1.9 yrs, established with four coeval subsamples (Table A10). The discordant ages are pro-posed to be caused by detrital materials with different tho-rium concentrations and 230Th/232Th ratios that were incorporated into the growing lattice during crystallization. The positive age bias for NW0310-5#3 and 92MC-1#4 could be due to carbonate detritus with low thorium concentration and a high230Th/232Th ratio. Terrestrial particulates with a high thorium level and a low 230Th/232Th ratio of 1– 2 106could result in the negative age bias of NP00A1–
1#5.
3.5. Variability of230Th/232Th0in coral skeletons
3.5.1. Subtidal corals at continental shelf sites: Nanwan and Son Tra
3.5.1.1. Nanwan. The 230Th/232Th0 values derived from
three isochrons of two modern Nanwan corals, 5.2 ± 1.1 106and 4.86 ± 0.27 106for NW0310 and
5.1 ± 1.9 106 for NW0402, collected 60 m apart, are consistent with each other (Fig. 5a). Plots of230Th age, cal-culated with the 230Th/232Th0 value of 4.86 ± 0.27 106
from NW0310-6, vs. banding age for three subsamples of NW940617 and five layers of NW0310, including bulk and three crushed fractions of <840, 840–1410, and 1410– 2380 lm (Table A4), are shown inFig. 6. All points, except for NW0310-5#3 with high230Th/232Th-detritus, plot
with-in error of a 1:1 lwith-ine for both bulk subsamples and crushed fractions, indicating that the230Th/232Th0ratios on
differ-ent growth layers vary less than ±0.27 106. A single 230
Th/232Th0 value can be used to determine different 230
Th ages on annual to decadal timescales for Nanwan samples. The results and measurements of seawater tho-rium isotopic composition at Nanwan (Figs. 2 and 3) indi-cate that inter-annual variability of the230Th/232Th ratio in the dissolved fraction is not resolvable, even though there is a large daily intrusion of cold upwelled subsurface water mass at the Nanwan site (Lee et al., 1997; Shen et al., 2005b).
3.5.1.2. Son Tra. All six subsamples of three layers of the modern Son Tra coral lie on a 1:1 line inFig. 6using a sea-water230Th/232Th value of 3.20 ± 0.32 106as the initial
value, indicating no significant annual change of the initial thorium isotopic composition at this site. There is a distin-guishable difference in 230Th/232Th0 values of
1.66 ± 0.42 106 between Son Tra and Nanwan (Fig. 7), suggesting that a terrestrial source with low tho-rium isotopic composition and high 232Th levels of 1000 s ppt is more dominant at Son Tra.
Located on the continental shelf with a large influx of inland material, the Nanwan and Son Tra sites in the western Pacific yield isochron-inferred230Th/232Th0 ratios
of 3–5 106(Fig. 5), consistent with an average crustal value (Richards and Dorale, 2003). Results of seawater
analyses and isochrons, and the consistency of230Th and absolute ages, show that the variability of 230Th/232Th
ratios in the dissolved fraction of seawater and initial values in corals is ±0.25–0.35 106, which is within 10% of the mean value at the two local sites. These small temporal and spatial variations suggest that the crust-derived mean value of 4–5 106with an arbitrary uncertainty of 50%
or 100% is practical for the continental shelf region in the western Pacific. The observation of resolvable differences in the 230Th/232Th0values between Nanwan and Son Tra
indicates that a site-specific230Th/232Th
0ratio is important
for high precision and accurate230Th dating.
0.0 0.1 0.2 0.3 0.4 0.5 0 10 20 30 230 Th/ 232 Th (x 10 -6 ) 234 U/232Th NW03 10-6 (15.56 ±0.5 6) NW0402-2 ( 5.1 ±3.0) NW0310-2 (4.1 ± 1.2) 0.0 0.5 1.0 1.5 2.0 0 50 100 150 92MC-1 (14.0 ± 1.1) GSG05YN G-2 (13.3± 3.2 ) GSG05YNG-1 (6.0 ± 1. 4) GS G0 5O LD -4 (2 42 ± 18 ) GS G0 5O LD -3 (26 9± 12 ) 0.00 0.02 0.04 0.06 0.08 0 5 10 15 LWK05 -4 (39. 9 ±3.8 ) 0.00 0.05 0.10 0.15 0 25 50 75 100 NP 00A 2-1 (476 .0± 5.9) NP00A1-1 (8.1 ± 1.9) BLS02A 3 (670 ± 40) BLS02A1 (43 8 ± 13) BLS02A5 (42 ± 11) BLS0 2A4 ( 1530 ±78 ) BLS0 2A2 ( 1680 ±38 ) BLS 02A 8-3a (241 3± 21) BLS 02A 7(2 436 ±17 ) 0.000 0.005 0.010 0.015 0.020 0 5 10 15 20 25 SMG0 2A1 (2 89± 2 1) 0.0 0.2 0.4 0.6 0.8 0 50 100 150 LG99A1-1 (8.3 ± 0.7) LG00 A1-1 (69. 3± 0.9) 92MC -2(2 9.2± 1.3) 0 5 10 0 5 10 0 5 20 25 30 0 5 10 0 5 10 0.00 0.02 0.04 0.06 0.08 0 100 200 300 400 500 0 5 10 BLS 02A 8-1b (248 0± 25) 0 5 10 0 5 10 0.0 0.1 0.2 0.3 0.4 0 20 40 60
a
d
b
c
e
h
g
f
Lago North Pagai Bulasat Saumang Nanwan Santo Lewak Gusong BayFig. 5. Isochron plots of230Th/232Th vs.234U/232Th (atomic ratios) for 23 coeval sets of subsamples from (a) three growth bands of two Nanwan coral slabs, NW0310 and NW0402, (b) two growth bands of the Santo 92MC coral, and (c–h) 18 horizons of corals from the Sumatran islands. Isochron-inferred230Th/232Th
3.5.2. Subtidal corals at open-ocean site: Santo
Data of230Th age vs. growth band age for subsamples of the Santo 92MC coral, except for subsample 1#4, and
subs-amples 2#1 through 2#4 (Table A6), reside on a 1:1 line using the isochron-inferred 230Th/232Th0 value of
5.62 ± 2.05 106 from the coeval subsamples on layer
92MC-1 (Fig. 5b). This initial value at this open-ocean site, with minor terrestrial influence, is higher than 3.2 106 measured at the continental shelf site of Son Tra (Fig. 5). The discordant age of subsample 1#4 was discussed earlier in Section3.4.2. The isochron plot of layer 92MC-2 shows a much higher 230Th/232Th0 ratio of 23.8 ± 4.7 106 and
the isochron age of 29.2 ± 1.3 yrs matches the absolute date (Table A6). If we use the initial value inferred from the layer 92MC-1, however, the 230Th ages of subsamples of 2#1–2#4 are 2.9–8.7 yrs older than the absolute age of 30.0 ± 0.5 yrs.
The high 230Th/232Th0 value of layer 92MC-2 in AD
1974.6 ± 0.5 (growth band age) could be attributed to sub-stantial upwelling of cold water in the eastern Pacific during a predominant La Nin˜a episode between AD 1973 and 1976. In terms of the magnitude of the Southern Oscillation Index, the three La Nin˜a years, AD 1974, 1975, and 1976, are historically ranked as the 1st, 15th, and 4th, respec-tively, during the interval from 1951 to 1996 (Clark et al., 2001). A monthly sea surface temperature anomaly (SSTA) of 2 °C over the Nin˜o-3.4 region (5°S–5°N, 170–120°W;
http://www.cpc.noaa.gov/data/indices/) in the central
equatorial Pacific was observed in both La Nin˜a episodes of AD 1973/74 and 1975/76. During a similar SSTA during the 1998/99 La Nin˜a event, shoaling of the thermocline in the eastern equatorial Pacific was induced by a strong west-ward near-surface (0–15 m) current of 1 m/s (Grodsky and Carton, 2001; Bonjean and Lagerloef, 2002). A high seawa-ter230Th/232Th value of 20–200 106was measured at a depth of 25 m at the Aloha Station in the central Pacific (22°450N, 158°000W) in September 1994 (normal El Nin˜o/
Southern Oscillation (ENSO) condition; Roy-Barman et al., 1996). Notably, in the 1944 La Nin˜a event, a high
230Th/232Th
0 ratio of 22–25 106, calculated with U–Th
data from the Modern-2 coral in Table 2 of Cobb et al. (2003b), was also observed at Palmyra Island, in the central tropical Pacific (5°510N, 162°80W). The high 230
Th/232Th0
value of 24 106found in layer 92MC-2 of the Santo cor-al could likely be attributed to cold upwelled water deliv-ered by the persistent westward current. An alternate possibility might be local changes in current circulation and/or wind strength and direction at the Santo site that were tied to ENSO.
In Santo coral 92MC, an initial isotopic ratio of 5.6 ± 2.1 106 is observed in 4 of 5 bands and a high value of 23.8 ± 4.7 106 observed in AD 1974–1975
during a strong La Nin˜a episode. The fluctuation of
230
Th/232Th0values at Santo, an open-ocean site in the
wes-tern tropical Pacific, could be related to mixing of water masses. Also taking into consideration the case at Palmyra Island in the central tropical Pacific (Cobb et al., 2003b), isochron techniques are suggested for determining accurate and precise 230Th dates. High-230Th/232Th thorium, deliv-ered by the upwelled water, can be taken up by the coral skeleton. The magnitude of230Th/232Th0variation at these
regions should therefore be understood in advance for high precision coral230Th dating.
Growth band age (yrs) 0 20 40 60 230 Th age (y rs) 0 20 40 60 0 5 10 15 20 0 5 10 15
a
b
High 230Th/232Th-detritus NW940617 ST0506 NW030525 NW0310 92MC LG99 LG00 GSG05YNG LWK05 ContaminationFig. 6. 230Th age vs. growth band age plots of over 50 data points for coral subsamples of Nanwan, Son Tra, Santo, and the Sumatran islands. Eight of the 230Th ages are inferred from isochron techniques and the other ages are calculated with different 230Th/232Th
0ratios (see text andAppendices). Note that subsample NP00A1–1#5, which yielded a230Th age 23 ± 6 yrs younger than the isochron age of 8.1 ± 1.9 yrs, is not shown on this plot. Data for subsamples that may have acquired high230Th in the storage room or during the subsampling process are given in black.
1/232Th (ppt-1)
0 2e-4 4e-4 6e-4
230 Th/ 232 Th 0 (x 10 -6 ) 0 2 4 6 8 10 12 230Th age (ka) 0.0 0.5 1.0 1.5 2.0 2.5 230 Th/ 232 Th 0 (x 10 -6 ) 0 2 4 6 8 10 12
8 Bulasat Porites corals
Variation of isochron-inferred initial 230Th/232Th ratios
Fig. 7. (a) Temporal variation of isochron-inferred230Th/232Th0 ratios for eight corals collected from Bulasat of the Sumatran islands over 2.5 ka. (b) An endmember plot of 230Th/232Th0 vs. 1/232Th shows a mixing trend (gray line) with an intercept of 4.4 ± 2.3 106at the 95% confidence (dashed lines).
3.5.3. Intertidal corals at sites in the Sumatran islands 3.5.3.1. Gusong Bay and Lewak of Simeulue. At the Gusong Bay site of southwestern Simeulue, there is no significant dif-ference between four isochron-inferred230Th/232Th0values,
7.35 ± 0.65 106 and 7.24 ± 0.70 106 from modern coral GSG05YNG, and 6.59 ± 0.97 106 and 6.97 ±
2.86 106from fossil GSG05OLD (Fig. 5d). Using an ini-tial value of 7.35 ± 0.65 106, six additional230
Th age vs. band counting age points lie on a 1:1 line (Fig. 6andTable
A8). Gusong Bay data show that the230Th/232Th0ratio
re-mains steady at this site. A similar case of a steady but lower
230
Th/232Th0value of 3.01 ± 0.47 106, inferred from a
LWK05 coral isochron (Fig. 5c andTable A7) and supported by the concordance between the banding ages and the230Th
ages calculated using the initial value (Fig. 6), is observed at the Lewak site in northern Simeulue, 80 km from the Gusong Bay site. The discrepancy between the initial values at the two sites, on the same island, clearly shows the230Th/232Th0ratio
is dependent on local hydrology.
3.5.3.2. Lago. A 230Th/232Th0ratio of 9.4 ± 1.2 106 is
observed for band LG00A1-1 of coral LG00A1, from the northeastern side of Lago island (Fig. 5e). An age of AD 1935.1 ± 0.9, determined with this initial value, is identical to a band age of AD 1935.0 ± 0.5 on the slab (Fig. 5e). Cal-culated with the same 230Th/232Th0 value, a date of AD
1942.1 ± 1.6 is calculated, identical to the absolute age (AD 1942.5 ± 0.5), for horizon LG99A1-3 of a contiguous coral slab, LG99A1, from the same coral head (Table A9). This indicates that the high 230Th/232Th0 value did not
change after the earthquake in December 1935 (Mw =
7.7;Rivera et al., 2002). The carbonate-enriched sandy
sub-strate could supply 230Th and a high 230Th/232Th ratio
source to coral skeletons (Robinson et al., 2004). A different initial value of 7.08 ± 0.86 106 is found using an iso-chron from an AD 1996 layer, LG99A1-1. This low initial value could likely be attributed to either a different thorium isotopic composition in seawater or a lower proportion of
230
Th to the coral skeleton.
3.5.3.3. North Pagai. The isochron-derived 230Th/232Th 0
ratios for one modern coral, NP00A1, and one 475-year-old fossil, NP00A2, sampled near Simanganya village along the northeastern coast of North Pagai are indistinguishable (Fig. 5f). The value of 4.7 ± 1.0 106is identical to that of the dissolved fraction of ambient seawater, 4.0– 4.3 106, supporting the finding that the dissolved
seawa-ter thorium is the main source of thorium to the coral skel-eton. Dissolved 232Th concentrations at the North Pagai site range from 9.5 to 13.8 ppt, which is even higher than those at the Nanwan site. Coupled with low230Th/232Th ra-tios of 4.0–4.3 106, this indicates that the site in North Pagai also experiences a high flux of terrestrial matter. 3.5.3.4. Bulasat and Saumang of South Pagai. The Bulasat and Saumang sites have similar hydrological settings and are 10 km apart on the southwestern coast of South Pagai. The 230Th/232Th0 ratios of two modern coral slabs,
BLS02A5 and SMG02A1, are 4.9 ± 1.4 106 and 5.02 ± 0.65 106, respectively (Figs. 5g and h). The
iden-tical230Th/232Th0values of the two slabs suggest that there
is no spatial variation of thorium isotopic composition at the sites in South Pagai. Temporal variation of
230
Th/232Th0values at the Bulasat site is characterized with
eight isochrons going back 2.5 thousand years (Figs. 5g and
7a). Despite 2r uncertainties of 1.4–3.8 106, the graphs
show that the means of the initial values ranged from 7– 8 106at 2.5 thousand years ago (ka) to 3.6 106 at
1.5 ka, with intermediate values of 4.8–4.9 106 during the past 0.7 ka. An endmember plot of 230Th/232Th0 vs.
1/232Th displays a mixing trend with an intercept of 4.4 ± 2.3 106 at the 95% confidence level (Fig. 7b), which agrees with the crust-derived mean value of 4.5– 4.7 106(Richards and Dorale, 2003). Terrestrial sources with low 230Th/232Th dominated in the four coral heads since 1.5 ka. High230Th/232Th
0ratios recorded in the other
four fossil corals, older than 1.5 ka, can be attributed to dif-ferent sources such as open ocean water masses and carbon-ate-enriched substrate with high 230Th (Robinson et al., 2004).
3.5.3.5. 230Th/232 Th0 variation between sites. The natural
dynamics of 230Th/232Th
0values in the different intertidal
zones at sites of the Sumatran islands are more complicated and governed by variability of four sources: (1) terrestrial influx, (2) in-situ ingrowth 230Th, (3) open-ocean surface seawater, and (4) upwelled cold seawater. Terrestrial sources appear to be dominant at the Lewak and North Pagai sites with low 230Th/232Th values of 3–4 106. The Lago site, away from sources of significant terrestrial input and with a carbonate sandy substrate, shows high
230
Th/232Th0 values of 7–9 106. The regional seawater 230
Th/232Th value is affected by the upwelling of cold sea-water, which has been observed in the eastern Indian Ocean. For example, three strong positive Indian Ocean Dipole events, in 1877, 1994 and 1997, caused a 4°C drop of sea surface temperature in the Mentawai reefs (Abram et al., 2003). The offshore islands have also experienced frequent rupture of active faults (Zachariasen et al., 1999; Sieh et al., 1999; Natawidjaja et al., 2004; Briggs et al., 2006). The proportion of the different thorium sources could plausibly be altered by uplift or subsidence associated with seismic displacements. A wide isochron-inferred initial range of 3.0–9.4 106 is shown inFig. 5. The observed spatiotemporal variation of the means of the isochron-in-ferred 230Th/232Th
0 ratios is 6.2 ± 3.2 106at the sites,
which sprawl across 800 km of the Sumatran islands (Figs. 5c–h). The range, consistent with the previous value of 6.5 ± 6.5 106 published by Zachariasen et al. (1999),
can be applied to 230Th dating of intertidal corals in the Sumatran islands. However, high precision and accurate
230
Th dates will require well-constrained230Th/232Th0
val-ues specific to each site.
3.6. Limitations of high precision coral230Th dating 3.6.1. Prerequisites
High precision coral 230Th dating with a precision of ±1 yr requires an appropriate quality control with a careful sampling procedures, cross-contamination-free sample
storage, and proper subsampling methods. The level and uncertainty of230Th blank in the chemical procedure and instrumental analysis should be effectively reduced. In this study, the procedural230Th blank of 0.0008 ± 0.0008 fmol
230Th corresponds to an uncertainty of only ±0.1–0.2 yr
for 1-g coral samples.
3.6.2. Usage of appropriate230Th/232Th0values
A bulk Earth crustal Th/U atomic ratio of 3.6–3.8 (Taylor and McLennan, 1985, 1995) and an assumed secular equilibrium between230Th and 238U in the
terres-trial upper or bulk continental crust is generally used to determine a 230Th/232Th value of 4–5 106 (Richards and Dorale, 2003); this, coupled with an arbitrary uncer-tainty of 50% or 100%, is often considered as the initial value for the age calculation in the 230Th dating equa-tion (Eq. (1)). This study of modern and fossil coral and seawater samples indicates that the dissolved frac-tion of seawater is the primary thorium source for the coral skeleton. Variability of seawater230Th/232Th values
may result from mixing of sources with different thorium isotopic compositions. The use of the continental crust-derived values is valid where the dominant source of thorium is terrestrial. However, a positive age bias can be generated if the crustal value is used for sites with high 230Th/232Th sources. Isochron techniques can offer an accurate 230Th/232Th
0value and a date with precision
as good as ±1 yr for corals younger than 100 yrs (Fig. 5). A high precision date can also be achievable using site-specific 230Th/232Th0 ratios (Fig. 6 and Tables
A3–A12).
3.6.3. Duplicate measurements
Three discordant ages in 141 U–Th points indicate additional 230Th and 232Th sources, presumably high-230Th/232Th carbonate sands and low-230Th/232Th ter-restrial materials, which cause occasional biases of 4–23 yrs from the true age (Fig. 6andTables A4, A6, A10). Coeval subsamples with different dating results indicate that heter-ogeneous coprecipitation places limits on high precision
230
Th dating. In our study, these discordant ages were relatively rare (2.1% of the 141 total subsamples). Duplicate measurement of coeval subsamples is suggested to verify the accuracy of high precision dates if230Th/232Th0can be well
estimated.
This issue is not critical for Quaternary samples, for which a230Th dating precision of ±30–50 yrs is satisfactory.
For young corals, the occasional problem of detrital mate-rial with230Th/232Th ratios significantly different from the dissolved value in seawater can be resolved by duplicate measurements of230Th ages at 2–3 coeval loci on a single growth band as a concordance test to rule out the influence of detrital material contributing an anomalous230Th/232Th value.
4. CONCLUSIONS
Measurements of U–Th isotopic compositions in sus-pended particulate material and the dissolved fraction of
seawater and in modern and fossil corals from sites in the western Pacific Ocean, South China Sea, and eastern Indian Ocean were performed to understand the natural variation of230Th/232Th0in shallow seawater corals and its effects on
coral 230Th dating. To approach this objective, sample
preparation procedures and chemical procedures have been refined, resulting in an equivalent age uncertainty of only ±0.2–0.3 yr.
Coral isochron-derived 230Th/232Th0 ratios,
compari-son of 230Th ages and absolute ages, and thorium analy-ses in seawater samples demonstrate that coral skeletal thorium originates mainly from the dissolved fraction of seawater. The 230Th/232Th0 value is strongly influenced
by local hydrological setting. 230Th/232Th
0 values are
low (3–5 106) at Son Tra and Nanwan, two
continen-tal shelf sites with intense terrestrial material input. The variation of initial values at each site is10%. The crus-tal U–Th composition-inferred 230Th/232Th values can be used for age calculation. At the open-ocean site of Santo in the western Pacific, a value of 5.6 106 is observed at 4 horizons, and one high value of 24 106 is docu-mented during a La Nin˜a event in AD 1973–1976. The initial values recorded in the intertidal corals at the sites in the Sumatran islands vary significantly from 3.0 to 9.4 106. Accordingly, isochron techniques should be applied for high precision dating at these sites with a variety of 230Th/232Th0 ratios. Detrital material with
anomalously low or high 230Th/232Th ratios causes biases ranging from23 to +4 yrs in this study. Duplicate mea-surement of coeval subsamples is recommended to verify the age accuracy.
Band-counted and tracer-tuned chronologies are usually characterized by compounded age errors of at least ±2–3 yrs by AD 1800–1900. Using site-specific 230Th/
232
Th0 values or isochron techniques, our study
demon-strates that a coral 230Th dating method with a precision as high as ±1 yr is achievable for corals with ages less than 100 yrs. The ability to obtain high precision and accuracy ages on young coral will be useful in diverse fields within the broad area of global change, including oceanography, tectonic evolution, anthropogenic pollution history, paleo-climate and paleo-environment. This230Th dating method-ology can also be applied to different carbonate samples, such as speleothem and tufa.
ACKNOWLEDGMENTS
We thank Y. Lin and Y.-G. Chen for field assistance and C.-C. Chang for laboratory work. Constructive comments of C.-A. Huh and D.-C. Lee of the Institute of Earth Sciences, Academia Sinica, S. Luo of the Department of Earth Sciences, National Cheng Kung University, K. Ludwig of the School of Oceanography, University of Washington, and K.R. Ludwig of the Berkeley Geochronology Center, are appreciated. Construc-tive and comprehensive reviews by J. Rubenstone and one anon-ymous reviewer significantly improved this paper. Funding for this study was provided by Taiwan ROC Grants (94-2116-M002-012; 2116-M002-015; 2752-M002-012-011-PAE, 95-2752-M002-012-PAE for CCS) and USA NSF Grants (NSF 0537953, EAR-9628301, EAR-9804732, EAR-9903301, EAR-0208508, and EAR-0538333).
Fig. A1. Map of collection sites of corals (stars) and seawater samples (square) at Nanwan, Taiwan.
Fig. A2. The collection site of a living Porites coral (star) and seawater (square) at Son Tra Island, Vietnam.
APPENDIX A
Fig. A3. The coral coring site of sample 92MC (star) on southern Espiritu Santo Island of Vanuatu.
Fig. A4. The collection sites of modern and fossil corals (stars) and seawater (square) in the Sumatran islands.
Table A1
U–Th data for dissolved fractions of seawater samples from Nanwan, North Pagai and Son Tra
Sample IDa Weight (g) 238U (ppb) 232Th (ppt) d234U (230Th/238U) activity ratio 230Th/232Th 106
NS1 536.88 3.061 ± 0.005 2.492 ± 0.006 146.2 ± 1.8 0.00020 ± 0.00002 4.04 ± 0.38 NS2 539.24 3.098 ± 0.004 2.251 ± 0.005 147.0 ± 1.4 0.00018 ± 0.00002 4.13 ± 0.37 NS3 536.74 3.069 ± 0.003 3.557 ± 0.008 148.5 ± 1.2 0.00029 ± 0.00002 4.12 ± 0.27 NS4 534.31 3.094 ± 0.004 2.460 ± 0.006 148.7 ± 1.2 0.00020 ± 0.00002 4.17 ± 0.42 NS5 541.01 3.090 ± 0.004 2.490 ± 0.006 148.3 ± 1.6 0.00021 ± 0.00002 4.22 ± 0.34 NS6 539.33 3.099 ± 0.004 2.561 ± 0.005 147.6 ± 1.3 0.00020 ± 0.00001 3.94 ± 0.29 NS7 537.08 3.090 ± 0.004 2.064 ± 0.004 146.5 ± 1.5 0.00018 ± 0.00001 4.34 ± 0.35 NS8 534.77 3.096 ± 0.005 0.716 ± 0.002 147.2 ± 1.7 0.00005 ± 0.00001 3.89 ± 0.78 NS9 533.95 3.078 ± 0.004 0.671 ± 0.002 148.1 ± 1.6 0.00005 ± 0.00001 3.91 ± 0.79 NS10 536.91 3.092 ± 0.005 1.899 ± 0.005 147.7 ± 1.9 0.00014 ± 0.00001 3.81 ± 0.38 NS11 538.77 3.045 ± 0.004 1.456 ± 0.004 148.6 ± 1.6 0.00012 ± 0.00001 4.17 ± 0.47 NS12 539.74 3.034 ± 0.005 2.803 ± 0.006 148.0 ± 1.9 0.00019 ± 0.00002 3.31 ± 0.30 NS13 535.66 3.077 ± 0.004 2.371 ± 0.005 148.5 ± 1.7 0.00017 ± 0.00001 3.57 ± 0.30 NS14 543.20 2.942 ± 0.005 7.004 ± 0.019 146.0 ± 2.0 0.00054 ± 0.00003 3.72 ± 0.22 NS15 538.38 3.093 ± 0.005 3.953 ± 0.008 146.2 ± 1.9 0.00032 ± 0.00002 4.17 ± 0.25 NS16 487.17 3.146 ± 0.005 4.463 ± 0.009 147.3 ± 1.7 0.00034 ± 0.00002 3.98 ± 0.22 NS17 534.88 3.124 ± 0.010 2.020 ± 0.004 148.1 ± 3.4 0.00015 ± 0.00001 3.92 ± 0.36 NS18 538.08 3.099 ± 0.004 1.861 ± 0.004 146.5 ± 1.6 0.00014 ± 0.00001 3.95 ± 0.40 NS19 537.00 3.104 ± 0.004 1.955 ± 0.005 148.2 ± 1.6 0.00017 ± 0.00002 4.34 ± 0.45 NS20 539.29 3.104 ± 0.004 0.991 ± 0.003 147.6 ± 1.5 0.00008 ± 0.00001 3.99 ± 0.57 NS21 542.62 3.116 ± 0.005 1.145 ± 0.003 145.2 ± 1.8 0.00010 ± 0.00001 4.53 ± 0.56 NPSW1 487.42 3.026 ± 0.005 9.503 ± 0.029 147.7 ± 1.8 0.00081 ± 0.00005 4.27 ± 0.29 NPSW2 539.59 3.052 ± 0.009 13.755 ± 0.057 145.5 ± 3.1 0.00109 ± 0.00007 4.00 ± 0.26 VTSW1 991.20 2.872 ± 0.015 0.282 ± 0.001 146.3 ± 2.9 0.000019 ± 0.000005 3.14 ± 0.84 VTSW2 1012.01 3.371 ± 0.009 0.318 ± 0.001 143.6 ± 1.7 0.000018 ± 0.000005 3.23 ± 0.85
a NS1–NS21: Nanwan samples; NPSW1–NPSW2: North Pagai samples; VTSW1–VTSW2: Son Tra Island samples. Nanwan samples were
collected every 2 h from July 31, 6:00 AM to August 1, 10:00 PM, 2004. North Pagai samples were collected in January 2004, and Son Tra Island samples on June 14, 2005.
Table A2
U–Th data in the suspended particulate matter of 5-L Nanwan seawater samples
Sample IDa 238U (ppb) 232Th (ppt) d234U (230Th/238U) activity ratio 230Th/232Th 106
SP1 1.276 ± 0.008 4.98 ± 0.05 23.9 ± 5.3 0.60 ± 0.09 2.52 ± 0.37 SP2 0.576 ± 0.007 1.77 ± 0.01 62.3 ± 8.1 0.57 ± 0.06 3.06 ± 0.34 SP3 0.696 ± 0.008 2.37 ± 0.01 22.3 ± 6.1 0.70 ± 0.08 3.38 ± 0.39 SP4 0.660 ± 0.007 1.99 ± 0.01 47.8 ± 9.3 0.60 ± 0.07 3.27 ± 0.38 SP5 1.319 ± 0.008 3.63 ± 0.09 36.2 ± 3.8 0.49 ± 0.06 2.91 ± 0.36 SP6 0.988 ± 0.008 3.28 ± 0.03 38.0 ± 4.6 0.53 ± 0.08 2.66 ± 0.41 SP7 1.035 ± 0.008 2.99 ± 0.02 40.0 ± 4.3 0.62 ± 0.08 3.52 ± 0.43 SP8 2.048 ± 0.010 7.41 ± 0.05 2.6 ± 3.7 0.60 ± 0.06 2.75 ± 0.28
U–Th data and230Th ages for subsamples of Nanwan NW030525, NW940101, and NW940617 corals Coral IDa Subsample IDb Weight (g) 238 U (ppb) 232 Th (ppt) d234U (230Th/238U) activity ratio 230 Th/232Th 106 d234Uinitial corrected 230 Th age uncorrected 230 Th age correctedc 230 Th date AD Banding date AD Chemistry date AD NW030525 1#1 1.046 2538 ± 5 470.6 ± 2.1 147.1 ± 2.0 0.000047 ± 0.000027 4.2 ± 2.4 147.1 ± 2.0 4.5 ± 2.6 0.7 ± 2.6 2004.3 ± 2.6 2002.9 ± 0.5 2003.58 1#2 2.578 2336 ± 6 148.2 ± 0.4 146.0 ± 2.4 0.000043 ± 0.000003 11.1 ± 0.9 146.0 ± 2.4 4.1 ± 0.3 2.3 ± 0.3 2001.3 ± 0.3 2002.9 ± 0.5 2003.58 2 0.966 2851 ± 5 380.7 ± 1.8 148.1 ± 1.9 0.000078 ± 0.000026 9.6 ± 3.2 148.1 ± 1.9 7.4 ± 2.4 3.7 ± 2.5 1999.9 ± 2.5 1999.9 ± 0.5 2003.58 3 0.985 2649 ± 5 388.2 ± 1.3 146.6 ± 2.1 0.000104 ± 0.000008 11.7 ± 1.0 146.6 ± 2.1 9.9 ± 0.8 5.8 ± 0.8 1997.8 ± 0.8 1997.9 ± 0.5 2003.58 NW940101 1#1 1.003 2363 ± 5 227.2 ± 0.9 145.4 ± 2.0 0.000176 ± 0.000008 30.2 ± 1.4 145.4 ± 2.0 16.8 ± 0.8 14.1 ± 0.8 1989.5 ± 0.8 1993.5 ± 0.5 2003.58 1#2 2.505 2411 ± 5 263.3 ± 0.3 146.7 ± 2.2 0.000193 ± 0.000004 29.1 ± 0.7 146.7 ± 2.2 18.4 ± 0.4 15.3 ± 0.5 1988.3 ± 0.5 1993.5 ± 0.5 2003.58 2 0.795 2640 ± 5 489.8 ± 1.4 145.1 ± 1.9 0.000497 ± 0.000011 44.3 ± 0.9 144.3 ± 1.9 47.4 ± 1.0 42.2 ± 1.0 1961.4 ± 1.0 1991.5 ± 0.5 2003.58 3 0.850 2524 ± 5 388.0 ± 1.2 144.9 ± 2.0 0.000439 ± 0.000011 47.1 ± 1.2 144.9 ± 2.0 41.9 ± 1.1 37.5 ± 1.1 1966.0 ± 1.1 1989.5 ± 0.5 2003.58 NW940617 1 1.459 2442 ± 5 328 ± 1 146.3 ± 2.1 0.000136 ± 0.000008 16.7 ± 1.0 146.3 ± 2.1 13.0 ± 0.8 9.2 ± 0.8 1994.4 ± 0.8 1994.0 ± 0.5 2003.58 2 0.844 2403 ± 4 1109 ± 2 145.6 ± 1.9 0.000247 ± 0.000011 8.8 ± 0.4 145.6 ± 1.9 23.6 ± 1.1 10.6 ± 1.3 1993.0 ± 1.3 1992.2 ± 0.5 2003.58 3 0.847 2535 ± 5 888 ± 2 145.0 ± 2.0 0.000240 ± 0.000011 11.3 ± 0.5 145.0 ± 2.0 22.9 ± 1.0 13.0 ± 1.2 1990.6 ± 1.2 1990.0 ± 0.5 2003.58
a Samples, NW030525 and NW940617, were stored in plastic bags and NW940101 was stored uncovered with Quaternary samples.
b
The subsamples, 1#2 of NW030525 and all of NW940101 shown in italic, cut on a class-100 bench in a general Quaternary carbonate preparation room, were contaminated.
c
Age corrected using an initial230Th/232Th atomic ratio of 4.86 ± 0.27 106, inferred from the isochron with subsamples 6#1–6#4 of NW0310 (seeA8).
Table A4
U–Th data and230Th ages for subsamples of Nanwan NW0310 and NW0402 corals
Coral ID Subsample ID Crushed size (lm) Weight (g) 238U (ppb) 232Th (ppt) d234U (230Th/238U) activity ratio 230Th/232Th 106 234U initial corrected 230Th age uncorrected 230Th age correcteda 230Th date AD Banding date AD Chemistry date AD NW0310 1 Bulk 0.696 2446 ± 4 320.3 ± 1.8 145.6 ± 1.5 0.000070 ± 0.000012 8.8 ± 1.5 145.6 ± 1.5 6.6 ± 1.1 3.0 ± 1.1 2001.6 ± 1.1 2002.0 ± 0.3 2004.54 2#1 Bulk 2.326 2620 ± 9 442.1 ± 1.4 147.1 ± 2.2 0.000097 ± 0.000006 9.2 ± 0.6 147.1 ± 2.2 9.2 ± 0.6 4.5 ± 0.7 2000.5 ± 0.7 2000.8 ± 0.3 2004.96 2#2 Bulk 2.258 2646 ± 9 519.9 ± 1.5 151.7 ± 2.4 0.000106 ± 0.000006 8.9 ± 0.5 151.7 ± 2.4 10.0 ± 0.6 4.6 ± 0.6 2000.4 ± 0.6 2000.8 ± 0.3 2004.96 2#3 Bulk 3.066 2596 ± 9 388.3 ± 1.2 150.0 ± 2.0 0.000091 ± 0.000005 10.0 ± 0.6 150.0 ± 2.0 8.6 ± 0.5 4.4 ± 0.5 2000.5 ± 0.5 2000.8 ± 0.3 2004.96 2#4 Bulk 2.868 2647 ± 10 497.9 ± 1.5 148.9 ± 2.2 0.000103 ± 0.000006 9.0 ± 0.5 148.9 ± 2.2 9.8 ± 0.5 4.5 ± 0.6 2000.4 ± 0.6 2000.8 ± 0.3 2004.96 2#5 Bulk 2.356 2520 ± 9 637.9 ± 1.4 150.8 ± 2.4 0.000124 ± 0.000005 8.1 ± 0.3 150.8 ± 2.4 11.7 ± 0.5 4.7 ± 0.6 2000.3 ± 0.6 2000.8 ± 0.3 2004.96 3#1 840– 1410 0.807 2662 ± 5 321.3 ± 1.2 147.5 ± 1.5 0.000084 ± 0.000007 11.5 ± 1.0 147.5 ± 1.5 8.0 ± 0.7 4.6 ± 0.7 1999.3 ± 0.7 1999.0 ± 0.3 2003.95 3#2 1410– 2380 0.836 2664 ± 7 328.2 ± 1.2 147.2 ± 2.2 0.000085 ± 0.000007 11.4 ± 0.9 147.2 ± 2.2 8.1 ± 0.7 4.6 ± 0.7 1999.3 ± 0.7 1999.0 ± 0.3 2003.95 3#3 Bulk 0.758 2622 ± 5 440.8 ± 2.5 146.7 ± 1.6 0.000107 ± 0.000013 10.5 ± 1.3 146.7 ± 1.6 10.2 ± 1.2 5.5 ± 1.3 1999.0 ± 1.3 1999.0 ± 0.3 2004.54 4#1 <840 2.324 2626 ± 5 465.5 ± 0.7 145.9 ± 1.6 0.000149 ± 0.000004 13.9 ± 0.3 145.9 ± 1.6 14.2 ± 0.4 9.3 ± 0.5 1994.7 ± 0.5 1995.0 ± 0.3 2003.95 4#2 840–1410 0.742 2657 ± 5 340.5 ± 1.4 146.9 ± 2.0 0.000134 ± 0.000009 17.3 ± 1.2 146.9 ± 2.0 12.8 ± 0.9 9.2 ± 0.9 1994.8 ± 0.9 1995.0 ± 0.3 2003.95 4#3 1410–2380 0.733 2621 ± 6 334.1 ± 1.5 146.7 ± 2.2 0.000133 ± 0.000008 17.2 ± 1.1 146.7 ± 2.2 12.7 ± 0.8 9.1 ± 0.8 1994.9 ± 0.8 1995.0 ± 0.3 2003.95 4#4 Bulk 0.622 2727 ± 5 427.7 ± 1.7 146.7 ± 1.7 0.000148 ± 0.000011 15.6 ± 1.2 146.7 ± 1.7 14.1 ± 1.1 9.7 ± 1.1 1994.8 ± 1.1 1995.0 ± 0.3 2004.54 5#1 <840 2.009 2534 ± 5 395.9 ± 1.1 149.6 ± 2.1 0.000183 ± 0.000006 19.3 ± 0.6 149.6 ± 2.1 17.3 ± 0.6 13.0 ± 0.6 1991.0 ± 0.6 1991.0 ± 0.3 2003.95
(continued on next page)
Initial 230 Th/ 232 Th and high precision coral U–Th dating 4215
Table A4 (continued) Coral ID Subsample ID Crushed size (lm) Weight (g) 238 U (ppb) 232 Th (ppt) d234U (230Th/238U) activity ratio 230 Th/232Th 106 234 Uinitial corrected 230 Th age uncorrected 230 Th age correcteda 230 Th date AD Banding date AD Chemistry date AD 5#2 840–1410 0.655 2543 ± 6 331.9 ± 2.0 147.8 ± 2.6 0.000171 ± 0.000009 21.6 ± 1.2 147.8 ± 2.6 16.3 ± 0.9 12.6 ± 0.9 1991.3 ± 0.9 1991.0 ± 0.3 2003.95 5#3 1410–2380 0.664 2545 ± 6 348.0 ± 1.3 146.8 ± 2.4 0.000214 ± 0.000010 25.9 ± 1.2 146.8 ± 2.4 20.4 ± 0.9 16.6 ± 1.0 1987.4 ± 1.0 1991.0 ± 0.3 2003.95 5#4 Bulk 0.956 2588 ± 5 317.7 ± 1.1 145.8 ± 1.7 0.000171 ± 0.000008 23.0 ± 1.1 145.8 ± 1.7 16.9 ± 0.8 13.5 ± 0.8 1991.1 ± 0.8 1991.0 ± 0.3 2004.54 6#1 <840 1.964 2508 ± 6 485.3 ± 1.5 147.3 ± 2.6 0.000218 ± 0.000008 18.6 ± 0.7 147.7b± 2.7 20.7 ± 0.8 15.56c± 0.56 1988.4 ± 0.6 1988.0 ± 0.3 2003.95 6#2 840–1410 0.985 2445 ± 6 356.9 ± 1.7 148.3 ± 2.5 0.000202 ± 0.000009 22.9 ± 1.0 19.3 ± 0.9 1988.0 ± 0.3 2003.95 6#3 1410– 2380 1.090 2497 ± 6 347.3 ± 1.5 147.6 ± 2.5 0.000213 ± 0.000009 25.2 ± 1.1 20.2 ± 0.9 1988.0 ± 0.3 2003.95 6#4 Bulk 0.991 2443 ± 6 4547 ± 15 148.3 ± 2.1 0.000713 ± 0.000028 6.3 ± 0.2 67.8 ± 2.7 1988.0 ± 0.3 2003.95 NW0402 2#1 Bulk 2.932 2435 ± 9 677.5 ± 1.7 147.2 ± 2.3 0.000137 ± 0.000006 8.1 ± 0.4 147.3 ± 2.3 13.1 ± 0.6 5.3 ± 0.7 1999.7 ± 0.7 No clear banding 2004.96 2#2 Bulk 2.495 2443 ± 8 752.9 ± 1.8 147.9 ± 2.2 0.000148 ± 0.000006 7.9 ± 0.3 147.9 ± 2.2 14.1 ± 0.5 5.5 ± 0.7 1999.5 ± 0.7 2004.96 2#3 Bulk 3.179 2468 ± 8 629.4 ± 1.3 147.5 ± 2.1 0.000130 ± 0.000005 8.4 ± 0.3 147.5 ± 2.1 12.3 ± 0.5 5.2 ± 0.6 1999.8 ± 0.6 2004.96 2#4 Bulk 3.173 2513 ± 10 618.2 ± 1.7 145.3 ± 2.1 0.000131 ± 0.000006 8.8 ± 0.4 145.3 ± 2.1 12.5 ± 0.6 5.6 ± 0.7 1999.4 ± 0.7 2004.96
a Age corrected using an initial230Th/232Th atomic ratio of 4.86 ± 0.27
106, inferred from the isochron with subsamples 6#1–6#4 of NW0310.
bIsochron-derived initial d234U value.
cIsochron age for the subsamples, 6#1, 6#2, 6#3, and 6#4, of NW0310.
Table A5
U–Th data and230Th ages for subsamples of the ST0506 coral slab from Son Tra Island, Vietnam
Subsample ID Weight (g) 238U (ppb) 232Th (ppt) d234U (230Th/238U) activity ratio 230Th/232Th 106 d234U initial corrected 230Th age uncorrected 230Th age correcteda 230Th date AD Banding date AD Chemistry date AD 1 1.573 2501 ± 10 1246 ± 12 146.4 ± 2.5 0.00012 ± 0.00002 3.8 ± 0.7 146.4 ± 2.5 11.0 ± 1.9 1.8 ± 2.1 2003.7 ± 2.1 2004.1 ± 0.3 2005.56 2#1 2.333 2427 ± 11 1062 ± 3 143.3 ± 2.6 0.00015 ± 0.00001 5.6 ± 0.4 143.3 ± 2.6 14.1 ± 1.1 6.0 ± 1.4 1999.6 ± 1.4 2000.4 ± 0.5 2005.56 2#2 1.913 2587 ± 6 1191 ± 5 147.5 ± 2.5 0.00015 ± 0.00001 5.3 ± 0.5 147.5 ± 2.5 14.0 ± 1.3 5.5 ± 1.5 2000.1 ± 1.5 2000.4 ± 0.5 2005.56 2#3 2.208 2461 ± 10 1102 ± 3 146.3 ± 2.4 0.00015 ± 0.00001 5.4 ± 0.4 146.3 ± 2.4 14.1 ± 1.0 5.8 ± 1.3 1999.8 ± 1.3 2000.4 ± 0.5 2005.56 2#4 2.376 2405 ± 11 1045 ± 3 147.2 ± 2.6 0.00015 ± 0.00001 5.7 ± 0.4 147.2 ± 2.6 14.2 ± 1.1 6.1 ± 1.4 1999.4 ± 1.4 2000.4 ± 0.5 2005.56 3 2.273 2410 ± 11 1388 ± 5 148.7 ± 2.5 0.00025 ± 0.00001 7.2 ± 0.4 148.7 ± 2.5 23.8 ± 1.2 13.1 ± 1.6 1992.4 ± 1.6 1992.6 ± 0.6 2005.56
a Age corrected using an initial230Th/232Th atomic ratio of 3.20 ± 0.32
106, estimated from seawater thorium data.
C.-C. Shen et al. / Geochi mica et Cosmochim ica Acta 72 (2008) 4201–42 23
