Elsevier

Marine Geology

Volumes 295–298, 15 February 2012, Pages 20-27
Marine Geology

Indian Ocean tsunami recurrence from optical dating of tsunami sand sheets in Thailand

https://doi.org/10.1016/j.margeo.2011.11.012Get rights and content

Abstract

Sand sheets deposited by the 2004 Indian Ocean tsunami and three Holocene tsunami predecessors in western Thailand have been dated by optically stimulated luminescence (OSL). Ages from the modern Indian Ocean tsunami sand sheet allow an assessment of the suitability of OSL to date tsunami-deposited sediment in tropical beach-ridge plain depositional environments. They show that, provided appropriate precautions are taken to correct for incomplete bleaching of sediments, OSL can provide a robust chronology of tsunami recurrence. Ages from the tsunami-deposited sand sheets underlying the 2004 tsunami deposits provide a chronology of the largest tsunamis in western Thailand over the late Holocene. Three sand sheets, likely deposited by tsunamis of similar magnitude to the 2004 event, were dated by luminescence to 380 ± 50, 990 ± 130–1410 ± 190, and 2100 ± 260 years before AD 1950 (at 1-sigma precision). The youngest of these compares with previous radiocarbon ages of detrital bark high in buried soils (Jankaew et al., 2008), which suggest that the most recent large-scale predecessor to the 2004 tsunami occurred soon after 550–700 cal BP. The OSL ages from the lower two sand sheets, provide constraint on tsunami recurrence that was not previously available from radiocarbon dating. The ages imply that between 350 and 1100 years separate successive tsunamis on the Andaman coast of Thailand, with an average tsunami recurrence interval of around 550 years. These results contribute to the growing body of research on tsunami recurrence in the Indian Ocean region and show that OSL can provide independent estimates of tsunami recurrence for hazard analysis, particularly in areas where suitable material for radiocarbon dating is unavailable.

Highlights

► OSL dating provides a chronology of the largest Holocene tsunamis in Thailand. ► The average tsunami recurrence interval is 550 years in Western Thailand. ► OSL can provide independent estimates of tsunami recurrence for hazard analysis. ► These results improve the regional picture of large tsunamis in the Indian Ocean.

Introduction

There was no historical precedent for an event the size of the 2004 Indian Ocean tsunami along the Aceh-Andaman subduction zone. Determining how often tsunamis occur and their magnitude is important for understanding tsunami hazard, which in turn informs how countries can prepare for such disasters. Studying the geological signatures of past tsunamis can extend the record of tsunami inundation and run-up by thousands of years. Over the past two decades, the identification of tsunami deposits, in particular sandy or gravelly sediments preserved within lower energy depositional environments, has aided in the recognition of tsunami events where written and instrumental records are unavailable (reviewed most recently by Bourgeois, 2009). Absolute dating of tsunami deposits allows an estimate of tsunami recurrence intervals, which can form the basis for tsunami hazard assessments.

Phra Thong Island is situated off the Andaman coast of Thailand, within 600 km of the Sunda Trench, the main source of tsunamis for the region (Fig. 1a). The western (Indian Ocean) coast has a series of sparsely vegetated, low, sandy Holocene beach ridges and marshy, intermittently inundated swales, which prograde from east to west. The east of the island comprises mangrove-fringed tidal inlets (Fig. 1b). Tides in the region are semi-diurnal and microtidal with mean spring tidal ranges of less than 2 m. In some settings, tsunami deposits can be difficult to distinguish from storm deposits. No known Indian Ocean cyclonic storm surges have inundated the western Thai coast (Murty and Flather, 1994). In this region, within 10° of the equator, cyclonic winds are limited due to the Coriolis minimum. Therefore the beach ridges of Phra Thong provide an ideal setting to study tsunami deposits without interference from storm activity.

The 2004 Indian Ocean tsunami crested higher on Phra Thong Island than almost anywhere else east of Sumatra, penetrating up to 2 km inland with run-up heights of up to 19 m (Tsuji et al., 2006). An eyewitness reported tsunami flow depths of around 10 m at the main study site (Jankaew et al., 2008). The tsunami caused local erosion of beach face sediments (Fig. 1c) and deposited a sand sheet 5–20 cm thick over most of the western half of the island. The sand sheet is composed predominantly of medium to very fine sand, grading up to coarse silt. Horizontal bedding and locally derived rip-up clasts are evident at some locations (Fujino et al., 2008, Fujino et al., 2009, Jankaew et al., 2008).

Jankaew et al. (2008) describe up to four tsunami sand sheets that predate the 2004 event preserved within the peaty soils of two marshy swales across the western half of Phra Thong Island. Sediments in these deposits range from coarse sand to coarse silt and generally have one or more fining up sequences. The more landward swale studied (Y) preserves four sand sheets whilst the seaward swale (X) preserves three sand sheets (Fig. 2). In swale X, close limiting maximum radiocarbon ages for sand sheet deposition were obtained from the uppermost centimetre of soil beneath sheet B, the penultimate sand sheet (550–700 cal BP), and sheet C, the lowermost event (2150–2350 cal BP). However, within the sand sheets themselves, horizontal laminae defined by leaf fragments yielded a conflicting range of ages (560–570 and 1070–1270 cal BP in sheet B of swale Y; between 6780 and 4260 cal BP in sheet C of swale X), indicating that the organic material in these sand sheets was reworked from older deposits. Therefore, these ages only provide distant limiting maxima for sand sheet deposition. Jankaew et al. (2008) could only tentatively correlate sand sheets associated with the penultimate tsunami between swales, based on the stratigraphic context of the deposits and the fact that detrital organics in swale X yielded ages too young to correlate with the lower sand sheets. They did not attempt to correlate the lowermost sand sheets across the swales due to a lack of material suitable for radiocarbon dating below sheet B in swale Y. However, radiocarbon ages on non-abraded shells (2482–2769 cal BP) underlying swale Y provide a limiting maximum for the formation of the swales and constrain the deposition of all sand sheets to the past 3000 years.

Radiocarbon is the most common method used to date tsunami deposits. Growth position plant fossils (where plant roots are fixed within the underlying soil whilst the stem protrudes in to the tsunami-deposited sediment after having been overrun by the advancing tsunami) provide excellent material for dating (Cisternas et al., 2005) however, these cannot always be found. Dating detrital plant and shell is less reliable as the material could have been reworked for hundreds to thousands of years before entrainment by the tsunami (Donato et al., 2008). This phenomenon is evident from the wide range of radiocarbon ages of detrital leaves preserved in sand sheets B and C at Phra Thong (Jankaew et al. 2008). These detrital organics cannot give reliable estimates of the time of tsunami deposition. Furthermore, the lowermost sand sheet at Phra Thong does not contain suitable material for radiocarbon dating in the underlying soil or within the sand sheet. Therefore, other dating methodologies are needed to correlate the sand sheets between the swales, to refine and extend the chronology at this site, and to provide a more robust recurrence rate for large tsunamis in the Indian Ocean.

OSL has the advantage of directly dating sediments. It measures the luminescence signal from the release of trapped energy that has accumulated in sediment grains as a result of environmental radiation since burial. Exposure to sunlight releases energy, effectively zeroing the signal. Therefore, OSL determines the elapsed time since sediments were last exposed to sunlight (Huntley et al., 1985, Aitken, 1998). Clean quartz grains exposed to direct sunlight can be bleached within a few seconds (Aitken, 1998). Incomplete exposure of sediments to sunlight during transport is common in many depositional environments and results in a heterogeneous distribution of trapped charge in the quartz grains (Murray and Olley, 2002). The process of sediment transport by tsunami is often rapid; at most tens of minutes for each half wave with total sedimentation lasting several hours (Murari et al., 2007, Jaffe et al., 2008). The flow may be too turbid to expose all sediments to light and the tsunami may occur at night. Therefore, whilst some bleaching may occur during transport, it is unlikely that it is sufficient to completely reset the luminescence signal (Huntley and Clague, 1996).

The assumption when using OSL to date tsunami-deposited sediments is that the source sediments are well bleached prior to tsunami entrainment and not exposed to daylight after deposition (Huntley and Clague, 1996). Tsunamis entrain sediments from the terrain over which they flow; usually shallow intertidal and beach face facies. Intertidal sediments from various locations have been shown to have a near-zero residual luminescence signal as they are regularly exposed to bioturbation of at least a few tens of centimetres by burrowing molluscs and worms and shallow-water reworking by tidal flows (Huntley and Clague, 1996, Banerjee et al., 2001, Madsen et al., 2005). Beach face sediments deposited by a combination of wind and wave action have also been shown to have near-zero ages (Lopez and Rink, 2008).

Asymmetric De populations in tsunami luminescence samples may result if the tsunami entrains incompletely zeroed sediment from deeper in the sedimentary profile (from > 50 cm sediment depth or > 10 m water depth for sub-tidal sediments: Murari et al., 2007). If these sediments are not exposed to sufficient sunlight during tsunami transport, for example if the tsunami occurs at night, the resulting luminescence age will be older than the timing of deposition. One method for detecting incomplete bleaching is to examine the distributions of equivalent doses. Positively skewed distributions (> 20%) suggest partial bleaching (Olley et al., 2004a). Such distributions have been described in paleotsunami sediments from North America (Huntley and Clague, 1996, Ollerhead et al., 2001). In preliminary studies of the 2004 tsunami sand sheet on Indian Ocean shores, Murari et al. (2007) found a narrow range of De measurements, which were interpreted as evidence for effective bleaching. They inferred that despite the size of the 2004 tsunami, it only eroded and entrained well-bleached sediment from the near shore zone. Bishop et al. (2005) found residual ages of less than 10 years in most tsunami samples from Thailand however; a small proportion of samples had residual ages of several thousand years. Robinson et al. (2008) found variations in the bleaching signal between different grain sizes within the 2004 sand sheet in Thailand.

Luminescence dating has been employed with various degrees of success to date tsunami events in Tunisia (Wood, 1994), Lisbon (Banerjee et al. 2001), north America (Huntley and Clague, 1996, Ollerhead et al., 2001), Israel (Reinhardt et al. 2006), Australia (Switzer and Jones, 2008) and the Andaman Islands (Kunz et al., 2010). The aims of this paper are twofold, Firstly, we test the applicability of using OSL to date tsunami deposited sediments in tropical beach ridge plain settings. As the exact time of the 2004 Indian Ocean tsunami is known, sand sheets deposited by this tsunami in Koh Phra Thong offer an excellent opportunity to study bleaching characteristics which allows a assessment of the effectiveness of the OSL technique to date tsunami deposited sediments in this depositional setting. Secondly, we report the first OSL results from paleotsunami-deposited sand sheets in western Thailand. This is significant as the available material for radiocarbon dating in the swales meant that Jankaew et al. (2008) were not able to provide well constrained ages for the two older tsunami events in the Phra Thong sequence. By directly dating the tsunami-deposited sediments, we provide ages for all sand sheets described in Jankaew et al. (2008). This provides greater constraint on the recurrence intervals of large Holocene tsunamis in the Indian Ocean. Such results are important for the assessment of tsunami hazard and the implementation of future tsunami mitigation strategies (Sieh 2006).

Section snippets

Methods

A total of twelve OSL samples were collected in metal tubes from freshly cleaned vertical sediment faces. Seven samples were taken from the tsunami-deposited sand sheets in swales X and Y, one sample from intertidal flat sand which underlies the swales, and four samples from the beach ridges surrounding the swales. The 180–212 μm quartz fraction was extracted under subdued, red light using standard procedures (e.g. Galbraith et al., 1999). Equivalent doses (De) were determined for multiple

Results

OSL ages are presented in Fig. 2 and Table 1. All samples contained quartz with luminescence signals dominated by the fast OSL component, displayed acceptable recycling ratios, and showed good internal consistency.

The OSL age from the underlying tidal flat deposits represents the time before the freshwater peat began to accumulate in the swales and shows that all four sand sheets were deposited after 2540 ± 340 years ago. This age overlaps with radiocarbon ages from growth-position shells in the

Incomplete bleaching and the suitability of OSL for dating tsunami deposits

As the exact time of the 2004 Indian Ocean tsunami is known, sand sheets deposited by the tsunami offer an excellent opportunity to study bleaching characteristics. The sand sheets at Phra Thong provide ideal candidates: based on their sedimentology and diatom assemblages, the tsunami sourced its sediment from beach and subtidal facies (Sawai et al., 2009), which should have a near-zero age. There is little chance of post-depositional disturbance to reset the OSL signal because in a little over

Conclusion

Examination of the bleaching characteristics of the 2004 Indian Ocean tsunami sand sheet on Phra Thong Island suggests that the majority of sediment entrained and deposited by tsunamis in this environment derives from well-bleached shallow sub-tidal and onshore settings. The direct dating of tsunami deposited sediment by OSL can circumvent problems associated with radiocarbon dating of older reworked material within the tsunami deposit. This paper shows that with the application of a minimum

Acknowledgments

The Authors thank Brian Atwater, Scott Nichol and Andrew McPherson for helpful pre-submittal discussions and reviews and the Australian Tsunami Warning System for providing funding for fieldwork and OSL dating. The manuscript was further improved by the comments of two anonymous reviewers.

References (48)

  • M. Oritz et al.

    Source area and rupture parameters of the 31 December 1881 Mw=7.9 Car Nicobar earthquake estimated from tsunamis recorded in the Bay of Bengal

    Journal of Geophysical Research

    (2003)
  • M.C. Quigley et al.

    Quaternary faults of south-central Australia: palaeoseismicity, slip rates and origin

    Australian Journal of Earth Sciences

    (2006)
  • A. Reid

    An Indonesian Frontier: Achenese and Other Histories of Sumatra

    (2005)
  • K. Sieh

    Sumatran megathrust earthquakes: from science to saving lives

    Philisophical Transactions of the Royal Society of America

    (2006)
  • M.J. Aitken

    An Introduction to Optical Dating: The Dating of Quaternary Sediments by the Use of Photon-Stimulated Luminescencerd

    (1998)
  • R. Bilham et al.

    Partial and complete rupture of the Indo-Andaman Plate boundary 1847–2004

    Seismological Research Letters

    (2005)
  • P. Bishop et al.

    Age-dating of tsunami deposits: lessons from the 26 December 2004 tsunami in Thailand

    The Geographical Journal

    (2005)
    J. Bourgeois

    Chapter 3. Geologic effects and records of tsunamis

  • J. Bourgeois

    Chapter 3. Geologic effects and records of tsunamis

  • M. Chlieh et al.

    Coseismic slip and afterslip of the great Mw 9.15 Sumatra-Andaman earthquake of 2004

    Bulletin of the Seismological Society of America

    (2007)
  • M. Cisternas et al.

    Predecessors to the giant 1960 Chile earthquake

    Nature

    (2005)
  • M.L. Cupper

    Luminescence and radiocarbon chronologies of playa sedimentation in the Murray Basin, southeastern Australia

    Quaternary Science Reviews

    (2006)
  • S.V. Donato et al.

    Identifying tsunami deposits using bivalve shell taphonomy

    Geology

    (2008)
  • S. Fujino et al.

    Thickness and grain-size distribution of Indian Ocean tsunami deposits at Khao Lak and Phra Thong Island, south-western Thailand: in tsunamites: features and implications

    (2008)
  • S. Fujino et al.

    Stratigraphic evidence for pre-2004 tsunamis in southwestern Thailand

    Marine Geology

    (2009)
  • Cited by (0)

    View full text