DSpace at VNU: Thermoluminescence ages for a reworked coastal barrier, southeastern Vietnam: a preliminary report

DSpace at VNU: Thermoluminescence ages for a reworked coastal barrier, southeastern Vietnam: a preliminary report tài li...

Thermoluminescence ages for a reworked coastal barrier,

southeastern Vietnam: a preliminary report

Colin V Murray-Wallacea,*, Brian G Jonesa, Tran Nghib, David M Pricea, Vu Van Vinhc,

a School of Geosciences, University of Wollongong, Wollongong, NSW, 2522, Australia

b Department of Lithology and Marine Geology, Vietnam National University, Hanoi, Viet Nam

c Quaternary Geomorphological Center, Mineral and Geological Survey of Vietnam, South Vietnam Geological Mapping Division,

14/5 Ky Dong St., Dist 3, Ho Chi Minh City, Viet Nam Received 5 October 2000; revised 26 February 2001; accepted 27 March 2001

Abstract

Thermoluminescence dating of quartz sand (90±125 mm) from the coastal barrier successions between Phan Thiet and Tuy Phong, southeastern Vietnam, reveals that a substantial component was deposited during the last interglacial sensu lato (Oxygen Isotope Stage 5) between 108 and 85 ka The barrier successions have subsequently, in places, experienced multiple phases of aeolian reworking during the last glacial cycle, and in particular between 62 and 19 ka, possibly related to reduced vegetation cover and landscape instability caused by climatic change The dif®culties of applying the thermoluminescence (TL) method in areas of intense tropical weathering are also examined

q 2002 Elsevier Science Ltd All rights reserved

Keywords: Coastal barrier deposition; Thermoluminescence dating; Quaternary; Vietnam

1 Introduction

The southeastern coastline of Vietnam between Phan

Thiet and Tuy Phong is dominated by extensive sandy

coastal barrier successions of presumed Late Pleistocene

and Holocene age (Nghi, 1996; Fig 1) Although these

successions have been examined for their economic mineral

potential (Nguyen et al., 1981), little has been done to

estab-lish their age and origin The general absence of organic

materials within these successions has precluded the

appli-cation of radiocarbon dating or other geochronological

methods for determining their age

At a regional scale two prominent groups of coastal

barriers can be identi®ed along parts of the central and

southern coastlines of Vietnam; outer barriers of

unconso-lidated white sand and inner barriers of predominantly red

sands with, in places, cores of older white sand The former

are represented by numerous bay head barriers, spits and

tombolos (Eisma, 1985) with some active dune®elds

Although the older, inner barrier sands may be

unconsoli-dated, they generally show signs of moderate to strong induration, are characteristically red and may contain, or

be capped by duricrusts, suggestive of a diagenetic origin for the red colouration of the sands Similarly, the presence

of clay cutans provides further evidence for the diagenetic modi®cation of the red sand units The inner and outer barrier successions merge over extensive parts of the coast-line giving rise to composite features In addition to the above, an older marginal marine sand succession onlaps the Annamite Mountains, separated from the former barriers

by an extensive back-barrier mud basin (Fig 1)

The dichotomy of an outer barrier of white sand of presumed Holocene age and an older, more landward Pleis-tocene barrier of predominantly red sand does not apply to coastal areas of Vietnam subject to active erosion At Hon

Da Chau (Fig 1), for example, the succession of diageneti-cally strongly modi®ed Pleistocene sand crops out along the modern shoreline

The working hypothesis of this pilot project is that the outer barrier white sand was deposited during the Holocene following the culmination of the post-glacial marine trans-gression, that in this region, is generally considered to have occurred about 7 ka ago (Fontaine and Delibrias, 1974; Pirazzoli, 1991; Nguyen et al., 2000), and that the inner

1367-9120/02/$ - see front matter q 2002 Elsevier Science Ltd All rights reserved.

PII: S1367-9120(01)00040-2

www.elsevier.com/locate/jseaes

* Corresponding author Tel.: 161-2-4221-4419; fax: 161-2-4221-4250.

E-mail address: colin_murray-wallace@uow.edu.au (C.V

Murray-Wallace).

barrier is associated with an earlier sea level high-stand,

possibly the last interglacial maximum (Oxygen Isotope

Substage 5.5) following the isotopic chronology of Bassinot

et al (1994)

In this work, a summary is presented of our reconnaisance

®eld investigations and thermoluminescence (TL) dating of

coastal barrier sands in southeastern Vietnam This report

represents part of a larger project examining the Late

Quaternary coastal evolution of southeastern Vietnam and

further ®eldwork and geochronological studies are planned

Several representative sections through the coastal barrier

successions at Hon Da Chau, Suoi Tien Village, Mui Ne

Beach west of Hon Rom Mountain, Phu Lac near Tuy

Phong, De Them II south of the River Luy and the disused

military airport at Phan Thiet were investigated in this study

(Fig 1) In addition, the Holocene white sand succession at

Bau Gang was also examined

2 Geographical setting

The study area is situated in Binh Thuan Province

between Phan Thiet and Tuy Phong (Fig 1) Extensive

accumulations of barrier sand occur along this coastline

Broad alluvial coastal plains cover extensive back-barrier

lagoonal mud basins, now heavily cultivated, situated

land-ward of the barriers and extending up to the Annamite

Mountains The continental shelf of the South China Sea

is shallow, and up to 14 km off the south Vietnamese

coast-line water depth does not exceed 20 m, giving rise to a

sheltered, low wave-energy coastline Apart from a narrow band of detrital sediment parallel to the modern coastline, much of the Sahul shelf directly offshore from the study area

is covered with a veneer of largely relict ®ne-grained sand, minimal carbonate and some ferruginous concretions and lateritic fragments (Emery and Niino, 1963; Astakhov et al., 1990) Tides are diurnal with a strong semidiurnal component during the period of equinoxes (Eisma, 1985)

3 Methods 3.1 Sample collection Deeply dissected stream cuttings through the coastal barrier successions of southeastern Vietnam were examined

to delineate lateral facies relationships Sections were measured using a laser theodolite Samples were collected for particle-size and petrographic analysis, and for TL dating Two methods were used in the collection of TL samples from homogeneous aeolian dune sediments Those from vertical sections were obtained by cleaning the face to expose the original stratigraphy, into which a steel sampling tube was inserted, extracted with the sample and wrapped in black plastic Where vertical sections were not present, samples were obtained by hand augering followed by transferring the samples to black plastic bags

by inserting the auger barrel into the bag and then releasing the sediment thus avoiding exposure to sunlight In the case

of some cohesive units, a block of sediment was cut from the

Fig 1 Location map of thermoluminescence dating sample sites in the Binh Thuan Province of southeastern Vietnam between Phan Thiet and Tuy Phong.

face and wrapped in black plastic, the inner unexposed part

of the block being used for dating All samples were

collected from depths exceeding 1 m Corrections for the

variation of cosmic radiation with burial depth were made

in accordance with Prescott and Stephan (1982) Standard

petrographic analysis was carried out to determine mineral

composition, shape and grain coatings The quantity of

K-feldspar is of particular importance since it has the potential

to in¯uence the radiation dose and accordingly, the TL ages

3.2 Thermoluminescence dating

TL samples were analysed using the combined additive

and regenerative method on the 90±125 mm quartz fraction,

separated from the bulk sediment sample In view of the

dif®culty of ®nding an appropriate modern analogue,

correc-tions were not made for surface residual TL The TL starting

point at the time of deposition is, therefore, assumed to be

that level reached following a 24 h exposure under a

labora-tory ultraviolet lamp (Philips MLU300W)

The combination of the regenerative and additive

meth-ods has the advantage of assessing the possibility of TL

sensitivity changes due to laboratory procedures The

method used here ®ts the mean natural TL output to the

regenerated growth curve and then plots the additive data

with respect to this value (Price et al., 1999) If there is no

change in TL sensitivity the two data plots coincide

A minimum of 28 sample aliquots, each containing

around 5 mg of quartz of the desired grain size, were utilised

in the comparison of palaeodose values All ®rst glow TL

output values were normalised using a second glow

proce-dure, thus correcting for possible variations between sample

purity nitrogen atmosphere and TL output signals were

detected using an EMI 9635QBphotomultiplier ®tted with

a Corning 7-57 blue transmission ®lter Sample aliquots

were irradiated using a Littlemore Automatic Irradiator

Sample annual radiation dose (mGy/a) values were

deter-mined by thick source alpha counting using a ®nely crushed

portion of the received sample, previously dried to

deter-mine moisture content Potassium concentrations were

measured by atomic absorption spectroscopy and

correc-tions were made for the cosmic radiation contribucorrec-tions

and moisture content in the ®nal radiation dose

computa-tion A full account of the laboratory procedures adopted

here are given by Shepherd and Price (1990), Nanson et al

(1991) and Price et al (1999)

4 Field sites

4.1 General remarks

The coastal barrier successions of southeastern Vietnam,

and in particular between Phan Thiet and Tuy Phong,

repre-sent particularly large sand accumulations The average

height of the barrier system over much of its length between Phan Thiet and Tuy Phong is about 100±140 m above present sea level (APSL), with the highest portion reaching

160 m APSL to the south of River Luy near De Them II At its widest near Hon Moc, the sands extend some 20 km inland (Fig 1)

Stratigraphically, the barrier succession closest to the modern coastline appears to be a composite feature comprising an inner core of Pleistocene white and red sand and, in more restricted pockets, an outer unconsoli-dated barrier of white Holocene sand The landward limit

of the barrier sands appears to be constrained by the occur-rence of large inselbergs of Cretaceous rhyolite and tuff (e.g Nui Ta Dom, Nui Gieng Xo, Nui Re and Hon Moc), and ¯uvial activity as de®ned by the course of the Rivers Cai, Can and Luy (Fig 1)

4.2 Phan Thiet airport The coastal red sand succession on which the disused military airport at Phan Thiet was constructed has several steeply dissected gullies on the seaward side of the airstrip Gully erosion was accentuated due to increased surface water runoff since airport construction TL sample VN#45 was collected 7 m below the ground surface from a vertical exposure of unconsolidated, dark red (2.5YR 4/8) medium-grained quartz sand, near the head of a major cutting

4.3 Bau Gang

An extensive sheet of unconsolidated white dune sand

erosion as attested by an extensive de¯ation lag of gravel, stone and pottery artefacts, and the presence of numerous isolated vegetated knolls of dune sand, the crests of which indicate a former dune surface The dune sediment comprises clean, medium-grained, pale yellow (2.5Y 8/2) quartz sand with no visible cross-bedding The unoxidized and unconsolidated nature of these sands suggests that they were more recently deposited than the indurated Pleistocene successions Although the origin of these sands is not clear, they may represent source-bordering dunes that formed adjacent to an extensive back-barrier lagoon, an area now under extensive cultivation, or represent a product of reworking of a more extensive dune sheet seaward of this locality TL sample VN#44 was collected 3 m below the present dune surface, from the vertical face of a small borrow pit

4.4 Suoi Tien

A thick section of sand is exposed within a deeply dissected stream cutting approximately 1 km inland from the presently eroding shoreline of Suoi Tien village

modern stream channel reveals a 70 m thick accumulation

of sand (Figs 2 and 3) that at stream level is light bluish

grey (8B8/1) medium- to ®ne-grained, strongly indurated,

with little discernible primary bedding This unit is overlain

by a medium- to ®ne-grained free ¯owing white quartz sand

The white sand has a sharp basal contact with the underlying

indurated sand The top of the white sand, however, has a

gradational boundary at about 6 m above the channel ¯oor,

and small mottles and rootlets extend down into this unit

from the overlying red sand The red sand comprises free

¯owing, medium-grained, reddish yellow quartz (5YR 6/8)

The sand particles are rounded with frosted surfaces or

surfaces coated with hematite The stream is actively

erod-ing the indurated grey sand cropperod-ing out at the base of the

channel resulting in the red cover sand avalanching into the

modern channel (Fig 2)

Samples for TL dating were collected from each of the

three sedimentary units exposed in the Suoi Tien stream

cutting (Fig 3) Sample VN#14, from the basal grey sand

was collected 2.5 m above the channel ¯oor From the

over-lying white sand TL sample VN#37 was collected 1 m below the gradational boundary with the overlying red sand, the latter sampled from within a deeply dissected runnel approximately 15 m above the channel ¯oor (VN#15)

4.5 4.5 Mui Ne Beach near Hon Rom Mountain

A composite section was constructed based on the measurement of two sections through a predominantly red sand succession at Mui Ne Beach, west of Hon Rom Moun-tain A very thick succession of red sand exposed within a steeply dissected amphitheatre-like cutting 3 km west of

and a cyclic succession, in ascending order, of red, yellow, white, yellow and red sand, exposed within a small road-side runnel approximately 1 km south of the amphitheatre at Mui Ne Beach, provide the basis for the composite section (Fig 4)

The base of the exposed composite section is represented

by a lower red sand which comprises a 5.5 m thick (mini-mum thickness as the lower bounding surface is not exposed) coarse-grained, yellowish red (5YR 5/8) quartz sand The base of the exposure is 9.5 m APSL A poorly developed ferricrete occurs in the upper portion of the unit The lower red sand is overlain by a thin unit (0.6 m) of medium- to coarse-grained, yellow (10YR 7/6) quartz

Fig 2 View looking northwest at the section ca 500 m upstream from Suoi

Tien Village of the Late Pleistocene red sand and underlying white sand.

Location of thermoluminescence samples; (a) red sand; (b) white sand The

underlying grey sand was sampled from 2.5 m above the channel ¯oor.

Fig 3 Schematic section of the red and white sands at Suoi Tien showing

the location of thermoluminescence samples.

Fig 4 Composite measured section at Mui Ne Beach west of Hon Rom Mountain.

sand The lower bounding surface has a mottled contact and

the upper bounding surface is represented by a sharp contact

with the overlying white sand, and a thin Fe-rich layer

sepa-rates the two units The overlying white sand unit is 2.6 m

thick and is in turn overlain by a 2.7 m thick unit of yellow

sand An intermediate red sand unit some 26 m thick

over-lies the yellow sand and contains tectites in its basal portion

The intermediate red sand unit comprises planar

cross-bedded, moderately well-indurated, medium- to

®ne-grained, yellowish red (5YR 5/8) quartz sand and is overlain

by 41.5 m of reddish yellow (5YR 6/8) free ¯owing,

medium- to coarse-grained quartz sand (upper red sand;

Fig 4)

Samples for TL dating were collected from the

intermedi-ate and upper red sand units exposed within the

amphithea-tre Sample VN#31, from the upper red sand, was collected

58.8 m APSL and 12.8 m above the contact with the

under-lying intermediate red sand unit Sample VN#32 was collected from near the top of the intermediate red sand and about 1 m below the prominent truncation surface which separates these two units Over a large area of the amphitheatre exposure, the truncation surface which sepa-rates the intermediate and upper red sand units has been exhumed due to erosion of the upper red sand In places this gives rise to a well-de®ned seaward dipping bench, the presence of which appears to be related to a weakly developed ferricrete immediately beneath the truncation surface

4.6 De Them II

An extensive accumulation of coastal barrier sand occurs

at least 140 m APSL and is likely to be younger than the extensive, indurated ¯uvial sand and gravel succession which crops out immediately adjacent to the River Luy The aeolian succession comprises yellowish red (5YR 5/8) medium- to coarse-grained quartz sand The sediment is indistinctly mottled and is free ¯owing under a case hardened surface

Samples for TL dating were collected from the upper portion of the barrier structure Sample VN#29 is from a cutting within a dune and was collected below the ground surface in the lee of the barrier Sample VN#30 was collected towards the crest of the barrier and approximately

20 m above sample VN#29

4.7 Hon Da Chau

A strongly dissected coastal section at Hon Da Chau

a prominent granite headland, reveals a well-developed duricrust overlying a 19 m thick succession of red and white sand (Figs 5 and 6) The coastal cliffs are cut through

Fig 5 Photograph of the measured section at Hon Da Chau The section

was measured from the top of the residual located slightly to the left of the

centre of the photograph down to high tide level See Fig 6 for further

stratigraphical details as well as the location of the thermoluminescence

samples.

Fig 6 Shore-normal transect of the measured section through the Pleistocene sands at Hon Da Chau.

a Trˆ

moderately to strongly indurated sand and a series of

promi-nent pillars with rillen occur throughout a small

amphithea-tre, attesting to the indurated nature of these sediments

The basal unit at modern beach level comprises some

6.7 m of very strongly indurated, medium- to ®ne-grained

white (1 8/N) quartz sand The base is not exposed Towards

the top are faint traces of bedding parallel with the overlying

unit The white sand is overlain by a 2.4 m thick unit of

massive, yellow sand which is, in turn, overlain by a

0.9 m unit of well-laminated yellow sand The yellow

sand is overlain by a 5.7 m succession of red sand capped

by a prominent duricrust The red sand is here mapped as

two sub-units, de®ned as a lower and upper red sand

separated by a disconformity The lower unit is 2.4 m

thick and comprises very strongly indurated, yellowish red

(5YR 5/6±5YR 4/6) ®ne- to medium-grained quartz sand,

slightly mottled towards the top and with iron preferentially

concentrated along fractures Separated by an irregular

disconformity, the upper unit comprises 3.3 m of

medium-to coarse-grained, homogeneous, yellowish red (5YR 5/6)

quartz sand with bedding only discernible in weathered

exposures The sediment is moderately indurated and

jointed, but without mottles

The upper red sand is overlain by a well-developed

duri-crust some 2.6 m thick consisting of moderate to strongly

indurated pisolitic ferricrete Strong brown (7.5YR 5/6)

subrounded nodules occur in a white to light grey (1 8/N±

1 7/N) matrix The section in places is capped by a modern

cover sand of strongly bioturbated, medium- to ®ne-grained,

reddish yellow (5YR 6/8) quartz sand

Samples for TL analysis were collected from the upper

red sand (VN#18), the top of the lower red sand (VN#20),

the well-laminated yellow sand unit (VN#22) and from the

top of the white sand (VN#24; Table 1; Fig 6)

4.8 Phu Lac

Further samples of red sand for TL dating were obtained

from a small road-side borrow pit adjacent to highway 1 at

1) The site is located within an elongate, coast-parallel sand

sheet TL sample VN#12 was collected 2.5 m below the

ground surface from the wall of the borrow pit, within

light to yellowish red to reddish brown (5YR 5/6±5YR 4/

4) medium-grained sand The sediment reveals faint traces

of trough cross-bedding and evidence of weak bioturbation

by plant roots in the upper-most part of the section The

quartz grains are coated with pellicles of iron oxide and

approximately 1% of the sediment comprises lithic

frag-ments A tan coloured soil, 10±20 cm thick, has developed

on the dune facies

A further sample was collected from an auger hole taken

within the ¯oor of the borrow pit TL sample VN#12Bwas

collected 9 m below the dune surface The very slightly

indurated sediment comprises reddish brown (5YR 4/4)

medium-grained quartz sand

5 Results 5.1 Petrology All sediment samples analysed in this study have a petro-logically similar grain composition dominated by mono-crystalline quartz (Table 1) Mono- and polymono-crystalline quartz grains are generally more common than feldspar or lithic grains and hence the majority of samples can be clas-si®ed as quartzarenite Feldspar (exclusively orthoclase) forms up to 5.5% of the detrital grains and is generally partly to extensively altered and replaced by secondary clay minerals Rock fragments also form a very minor component (3%) and are dominated by siltstone fragments although chert, granite, basalt, metaquartzite and other metamorphic rock fragments are also present Most non-quartzose lithic grains are partly altered and weathered In many of the measured sections the stratigraphically lower samples tend to have higher orthoclase feldspar contents than the upper samples, possibly a result of diagenetic alteration and dissolution by ¯uctuating groundwaters in the near surface environment or a function of reworking

of the upper sands, although in the Hon Da Chau area it is

a function of the adjacent granitoid provenance Accessory minerals are commonly present in trace amounts and include opaque grains, zircon, tourmaline, garnet, monazite and ferromagnesian minerals (epidote?) Virtually no detri-tal matrix was recorded in these dune samples

Secondary minerals include rare micritic carbonate in two samples from the Hon Da Chau area, while clay minerals and/or iron oxides are common in most samples In all the red and yellow sands the iron oxides and clay form thin to thick grain coatings, some of which cement the smaller grains In several of the samples these secondary coatings have the appearance of soil cutans These iron oxide and clay coatings form up to 38% of the sample by volume on a porosity-free basis The clay consists predominantly of poorly crystalline kaolinite

No clear systematic petrological trends are apparent from the analysed samples and the quantity of cement present probably re¯ects local and ¯uctuating groundwater table positions under the tropical weathering regime In general, the amount of iron oxides and clay determined from the petrological study exceeds the amount of ®ne material seen in the particle size analysis This indicates that not all the secondary cements were disaggregated during parti-cle size analysis which, in turn, re¯ects slightly upon the accuracy of the particle size analyses, especially for skew-ness and kurtosis

5.2 Particle size analysis Particle size analysis was carried out on all the samples analysed by TL One of the major limitations of the particle size analysis is the abundance of authigenic iron oxide and clay in many samples All the yellowish and red sand

Plateau region

Palaeodose (Gy)

a Atomi

b TL

samples contain moderately thick iron oxide grain coatings

that mask the true grain surface and cause a slight increase

in the particle size of most sand grains In addition to the

iron oxide, many of the oxidized samples also contain very

poorly crystalline kaolinite cutans Similar clay cutans are

also present in the white sands, being particularly abundant

in the Mui Ne area When the samples are disaggregated

using ultrasonic methods most of the clay cutans and some

of the less stable altered feldspar and lithic grains break

down to produce a ®ne fraction that does not appear to be

indicative of depositional matrix Thus, the total particle

size analysis (Table 1) probably relates at least to some

extent to the severity of diagenetic alteration and soil

form-ing processes in the tropical environment, thereby maskform-ing

a realistic indication of mode of origin The reduced mean

grain size, poor to moderate sorting and ®ne skewness all

re¯ect the amount of diagenetic alteration

If just the sand and coarser silt size particles are

consid-ered (which petrographic analysis indicates mainly consist

of individual grains and iron oxide coated grains) the

samples show a fairly uniform mean ®ne sand size (0.22±

0.25 mm) and are moderately to moderately-well and

well-sorted (Table 1) Two samples are medium-grained and

moderately sorted Although many of the grain surface

textures are disguised by authigenic grain coatings, the

samples from the base of the Mui Ne and Hon Da Chau

sections have the greatest proportions of grains with dull

surfaces These are possibly of marine origin In contrast, samples from topographically higher sections generally have more frosted and polished grains This probably indi-cates an aeolian environment with a greater amount of reworking of the detrital grains

5.3 Thermoluminescence dating With the exception of the TL data that re¯ect saturation, and therefore represent minimum ages, all the TL results indicate Late Pleistocene ages for the barrier sands (Table 2) The TL data indicate that saturation is attained over potentially varying intervals ranging from approximately 48±204 ka and point to the dif®culty of obtaining resolvable luminescence ages 200 ka within the study area, owing to the speci®c luminescence characteristics of the quartz (e.g apparent early TL saturation), high annual radiation doses and the intense weathering conditions experienced during diagenesis, resulting in the migration of radionuclides down pro®le by in situ leaching

The data highlight, in part, the dif®culty of applying the

TL method in environments where consistently high reac-tion temperatures and ¯uctuating groundwaters may mobi-lize radioisotopes and thus potentially compromise the veracity of the method (Aitken, 1985; Forman et al 2000) The results for Suoi Tien Village illustrate this point The basal grey sand yielded a result of 204 ka

Fig 7 Regenerated TL growth curve at 3758C for sample VN#37 (W-2531) The natural TL (N) is minimized using a laboratory ultraviolet lamp (UV) and incrementally irradiated using a calibrated 90 Sr plaque beta radiation source ( b ) The laboratory-induced TL (N 1 UV 1 b ) TL is shown plotted (A) against increasing laboratory irradiation dose The mean (n ˆ 8) natural TL value is shown plotted (1) upon the growth curve The additive TL data (N 1 b ) is shown plotted (x) with reference to the mean N value and thus provides a check for possible TL sensitivity changes Inset shows the temperature plateau comparison derived by plotting the ratio of the natural TL (N) to that induced in the laboratory, following TL minimization, against temperature where the laboratory-induced signal approximates the natural value The plateau region in this case extends between 300 and 5008C.

(VN#14; W-2211) while the overlying white sand in the

mottled zone beneath the red sand yielded an age of

48 ^ 6 ka (VN#37; W-2531), and the red sand in the

upper part of the succession yielded an age of 85 ^ 9 ka

(VN#15, W-2212; Fig 3) The younger age for TL sample

VN#37 (W-2531) is problematic Although the

post-deposi-tional concentration of radioisotopes of the elements K, Th

and U would potentially increase the environmental dose for

this sediment and hence yield a younger apparent TL age, it

is noted that of the three TL samples from Suoi Tien, sample

VN#37 actually revealed the lowest annual radiation dose

(Table 2) Although 2.5% of the sediment in sample VN#37

comprises orthoclase feldspar which may have in¯uenced

the radiation dose, the measured annual dose remains the

lowest of the Suoi Tien samples Sample VN#37 (W-2531)

yielded a well-de®ned growth curve with an extended

plateau between 300 and 5008C and does not indicate

saturation (Fig 7) The basis for this age reversal is still

being investigated, particularly in the context of the

possi-bility of radioactive disequilibrium

The TL results for the red sand at Hon Da Chau con®rm

the Late Pleistocene age for the upper part of this succession

and further illustrate the problem of obtaining TL ages on

strongly diagenetically modi®ed sediments, as revealed

from lower within the succession (Fig 6) Sample VN#18

from the upper red sand yielded an age of 52 ^ 7 ka

(W-2525) and sample VN#20 from the top of the lower red sand

yielded an age of 103 ^ 11 ka (W-2526; Fig 6) Farther down pro®le, however, the TL results show the combined effects of TL saturation and partial bleaching before deposi-tion, the latter indicated by the truncated growth curve for sample VN#24 (Fig 8) Samples VN#22 from the well-laminated yellow sand unit and VN#24 from the top of the white sand yielded minimum ages of 48 ka and 53 ka respectively and represent conservative minimum ages (Table 2) Although sample VN#24 displays a trun-cated plateau indicative of partial bleaching (Fig 8), and would otherwise yield an older apparent age, the effects of

TL saturation in this particular sample far exceed the former effect, thus deriving a younger minimum age Sample VN#24 comprises 5.5% orthoclase feldspar derived from the local weathering of the granite at Hon Da Chau, and would account for the particularly high annual radiation dose for this sample, resulting in the earlier attainment of

TL saturation (Tables 1±2)

The two TL ages for the upper and lower red sand units from the Mui Ne Beach succession near Hon Rom Moun-tain, yielded results consistent with respect to depth down pro®le Sample VN#31 yielded an age of 101 ^ 17 ka (W-2533) and VN#32 an age of 108 ^ 49 ka (W-2534) sugges-tive of deposition during Stage 5 sensu lato of the Oxygen Isotope record (Bassinot et al., 1994)

The red sand from Phu Lac near Tuy Phong collected from 9 m below the dune surface (VN#12B) yielded an

Fig 8 Regenerated TL growth curve at 3758C for sample VN#24 (W-2528) The natural TL (N) is minimized using a laboratory ultraviolet lamp (UV) and incrementally irradiated using a calibrated 90 Sr plaque beta radiation source ( b ) The laboratory-induced TL (N 1 UV 1 b ) TL is shown plotted against increasing laboratory irradiation dose (A) Due to apparent TL saturation there is no discernible increase in TL beyond a palaeodose value of 165 Grays This conservative minimum palaeodose corresponds with a minimum age of 53 ka The mean (n ˆ 8) natural TL value (6.77 ^ 0.91) has not been reliably induced

in this case Inset shows the temperature plateau comparison derived by plotting the ratio of the natural TL (N) to that induced in the laboratory, following TL minimization, against temperature where the laboratory-induced signal approximates the natural value The plateau region in this case appears truncated which, combined with a 13% uncertainty in the mean natural TL value, is suggestive of unreliable TL data.