Medieval Forewarning of the 2004 Indian Ocean Tsunami in Thailand

Introduction

The 2004 Indian Ocean tsunami lacked precedent in recent centuries, both in devastated coasts and its source area.
The tsunami claimed victims on shores without a tsunami disaster for 200+ years.
The earthquake's magnitude of 9.2 surpassed the Sumatra-Andaman catalog, which lacked a 19th or 20th-century earthquake larger than magnitude 7.9.
The tsunami and earthquake resulted from a fault rupture of 1,500 km.
The study identifies probable precedents for the 2004 tsunami using sedimentary evidence at a beach-ridge plain 125 km north of Phuket.
The 2004 tsunami ran 2 km across the plain, coating ridges and swales with a 5–20 cm thick sand sheet.
Peaty soils in marshy swales preserve remains of earlier sand sheets less than 2,800 years old.
The most recent full-size predecessor to the 2004 tsunami occurred about 550–700 years ago if responsible for the youngest pre-2004 sand sheets.

2004 Tsunami Impact on Phra Thong Island

The 2004 Indian Ocean tsunami crested highest in Thailand, reaching 20 m above sea level on Phra Thong Island.
The main wave formed a relentless flood, rising stepwise above treetops.
It ran more than 2 km inland across a Holocene plain of grassy beach ridges and tree-lined swales.
Flooding caused local erosion and widespread deposition.
The tsunami reamed out drainages cut across beach ridges up to 300 m inland.
It coated the island's western half with a sand sheet, lineated with sand streaks from tin mining spoil piles.
Mean particle size ranged from medium to very fine, with horizontal bedding and upward fining to coarse silt.

Investigation of Pre-2004 Sand Sheets

Researchers sought pre-2004 sand sheets at Phra Thong by digging pits and augering holes at over 150 sites.
Pre-2004 sand was found interbedded with peaty soils of swales that hold standing water at 20 sites.
No pre-2004 sand beds were found in the quartz-sand soils of ridges or slightly organic soils of damp swales.
Pre-2004 beds were traced across two marshy swales near a location where the 2004 tsunami reportedly flowed about 10 m deep.
These swales formed about 2,500 years ago, when the area's relative sea level was within 1–2 m of its present position.
The westerly swale (X) postdates its neighbor (Y).
Stratigraphic cross-sections were assembled from pits, auger borings, and a 35 m long trench.
Particle size was estimated, and a preliminary chronology was inferred from radiocarbon dating of plant remains and shells.
Diatom analyses were also conducted.

Characteristics of Sand Sheets in Swale X

Peaty soil in swale X contains two sand sheets (B and C) resembling the 2004 deposit.
Sheet C (earlier) is commonly 10 cm thick, with coarse to very coarse sand forming a discontinuous basal layer filling pockets in the underlying soil.
The rest of sheet C consists of very fine sand and coarse silt with horizontal laminae defined by leaf fragments.
Sheet C formed after 2,200–2,400 sidereal years ago, based on bark age from the uppermost 1 cm of the underlying buried soil.
Leaf fragments within sheet C gave conflicting ages, exceeding the bark age by thousands of years.
Sheet B is commonly 5 cm thick, fining upwards from fine sand to sandy silt, conformably overlying peaty soil with bark fragments in its uppermost 1 cm horizon.
Three bark fragments yielded ages between 530 $\pm$ 40 and 570 $\pm$ 40 radiocarbon years before AD 1950 (14C yr BP).
If scarcely younger than these fragments, sheet B was deposited about 550–700 sidereal years ago.

Sand Sheets in Swale Y

Three pre-2004 sand sheets alternate with peaty soil in swale Y.
All three sheets are similar in thickness to the overlying 2004 sand sheet and extend preferentially up the swale’s seaward side.
All formed after the swale ceased holding an intertidal flat marked by non-abraided molluscan shells 2,500–2,800 sidereal years old.
The lowest two sheets are undated and uncorrelated with swale X, consisting mainly of very fine to fine sand.
They lack sedimentary structures due to bioturbation blurring contacts with underlying soils.
The highest pre-2004 sheet (B) retains a sharp base and tabular shape extending the trench's full length.
This sheet typically fines upwards from basal fine or medium sand to parallel-laminated very fine sand abounding in leaf fragments.
It probably correlates with sheet B of swale X because each is the youngest pre-2004 sand sheet in its swale, and leaf fragments in swale Y yielded ages too young for correlation with sheet C.

Diatom Analysis and Preservation

The 2004 sand sheet abounds in brackish and marine diatoms, but earlier sand sheets in swales X and Y lack diatoms of any kind.
Marine and brackish-water diatoms aid in identifying tsunami deposits on temperate shores.
The dissolution of diatoms increases with temperature, potentially affecting preservation in tropical warmth.
Pre-2004 sheets are distinct and sharply bounded where swale soil is peaty but blurred by gradational contacts where the soil is just slightly organic and totally absent in the sandy soils of beach-ridge crests.
The 2004 tsunami deposit is already headed towards this fate: in wet swales, it has a protective cap of organic matter up to 5 cm thick, whereas on ridge crests, it lacks cover other than ejecta from burrows.

Geographic Setting and Alternative Explanations

Sheet B, if truly correlative between swales X and Y, initially spanned the intervening beach ridge for a total shore-normal length of no less than 100 m.
The geographic setting limits Phra Thong’s exposure to intense storms.
Thailand fringes the belt where the Coriolis minimum limits cyclonic winds, being less than 10° from the Equator.
Twentieth-century cyclones originated in Indian Ocean waters to its west moved towards India, Bangladesh, or Myanmar without producing a known storm surge in Thailand.
Tropical cyclones strike Thailand from its Pacific side but lose strength during their overland crossing to the Indian Ocean.
Phra Thong’s setting also disfavors sand-sheet deposition by river or wind.
Tidal inlets separate the island from the nearest rivers.
Aeolian dunes obscure little of the island’s striping by beach ridges and swales.

Chronological Evidence and Correlations

The middle Holocene ages of leaf fragments from sheet C imply scour into long-buried deposits beneath tidal inlets.
The 2004 tsunami demonstrated capacity for such scour by knocking down mangroves along an inner part of the inlet that bounds Phra Thong Island on the south.
The sand sheets represent infrequent events.
The soil between sheets C and B spans 1,500–1,850 years, although it may contain the bioturbated remains of an intervening sheet.
The interval between sheet B and the 2004 tsunami lasted nearly 550–700 years.
These time intervals are in the broad range of deductive estimates for the recurrence of giant earthquakes in the Sumatra–Andaman source region of the 2004 tsunami.
Sheet B, if little younger than AD 1300–1450, may correlate with tsunami and earthquake evidence elsewhere.
The youngest widespread pre-2004 sand sheet on a beach-ridge plain at Meulaboh, Sumatra, overlies plant detritus dated to AD 1290–1400.
Two coral fragments on a marine terrace in the Andaman Islands gave ages in the range AD 1200–1650.
There is no written evidence for a sheet-B tsunami on Sumatran and Sri Lankan shores that the 2004 tsunami would have overrun, based on accounts from Ibn Battuta (journey, AD 1325–1354) and the great Ming armadas (voyages, AD 1405–1433).

Tsunami Source and Implications

The pre-2004 sand sheets probably required ruptures larger than that of 1881; no sand sheet from the 1881 tsunami, which crested less than 1 m high on Indian tide gauges, is evident at Phra Thong Island.
The pre-2004 sheets may also require Sunda Trench earthquakes larger than magnitude 8.5 if such earthquakes would spawn Thai tsunamis only a few meters high.
Sand sheets of Phra Thong Island thus forewarn of infrequent catastrophe.
They provide public officials and coastal residents tangible evidence that the 2004 tsunami was not the first of its kind.
It remains to be determined whether centuries dependably separate such outsize tsunamis of Sumatra–Andaman source and whether these recur often enough to dominate Thailand’s probabilistic tsunami hazard.
Tsunamis without precedent in written history may threaten Indian Ocean shores that face other parts of the Sunda Trench and the Makran subduction zone.
Natural warnings from recent geological history will hopefully avert surprises from these additional tsunami sources.