Australia: The Land Where Time Began

A biography of the Australian continent 

Archaeological Sites Near the Dampier Archipelago – the Potential for the Discovery of New Sites

A probable 30,000 year archaeological record is preserved by the islands of the Dampier Archipelago that reflects the change from a continental to an island environment following sea level rise after the last glacial. In the Dampier Islands region the geomorphological history, combined with the preliminary hydrodynamic modelling of past tidal regimes provides the basis for a new model of the way in which the shelf landscape may have developed between the Last Glacial Maximum (LGM) (c.20,000 BP), through the marine transgression of the Holocene and up to the present. An assessment was made of the key Late Pleistocene and Holocene archaeological deposits by the use of first-order geomorphological principles.

It is shown by Ward et al. that archaeology is most likely to be present in deposits that are associated with the early phase of inundation of the Dampier Archipelago; from c.9-7 ka BP. Relative sea levels were about -30 m to -15 m at the time, which was when coastal configuration was complex and the variety and scale of intertidal and shallow sub-tidal environments wide. Ward et al. anticipate that in contrast coastal archaeology older than ~12 ka BP, a time when the post-glacial sea levels were below ~50 m, will have been exposed to a phase of more rapid coastal currents on the continental shelf, and therefore eroded or poorly preserved. The study of Ward et al. was aimed at improving prospection for the later management of submerged elements, which are as yet unknown, of the west Australian rich archaeological heritage.

Sea levels around the Earth have fluctuated by about 120 m over the past million years, corresponding with the cyclical contraction and expansion of the ice caps (Lambeck et al., 2002). These changing sea levels have had dramatic effects on the palaeogeography of the Australian continental shelf, repeatedly exposing and drowning vast areas since the earliest occupation of Australia. Islands off the Western Australian coastline that were formerly part of the continental shelf landmass became separated from the mainland finally during the eustatic sea level rise following the end of the last glacial period (Chappell et al., 1996). Possibly not surprisingly, several of these have archaeological evidence of occupation dating back to the Pleistocene (e.g. Dortch & Morse, 1984; McDonald & Veth, 2009; Veth et al., 2007; see also; Bowdler, 2003).

Also, evidence of Pleistocene age archaeological material is preserved on the islets and islands that comprise the Dampier Archipelago and the adjacent mainland (Dortch, 2002; McDonald & Veth, 2009; Mulvaney, 2011, 2013). Also, it is indicated by archaeological evidence on the islands of the possible presence of a transition from terrestrial to a mixed terrestrial and marine economy dating to the Early Holocene (post-10,000 BP), that has been asserted, reasonably, to have been associated with the rising of the sea level (McDonald & Veth, 2009; Mulvaney, 2010, 2011, 2013; McDonald, 2011). This transition is exemplified, however, by the change in style and fauna that is depicted in the rock art (petroglyphs) of the islands. A dominating feature of the Dampier Archipelago is these petroglyphs, with estimates of more the 1 million images present in more than 1,800 sites that have been recorded (Bird & Hallam. 2005). The particular geological and geomorphic processes that have formed the rock slopes of the Dampier Archipelago which provided the specific conditions for the production of rock art and its preservation is equally remarkable.

An initial investigation was undertaken by the West Australian Museum to locate and identify submerged rock art during the Dampier Archaeological Project in 1981, was unsuccessful (see DAS, 1984). A conclusion that was reached by a subsequent investigation that potential engraving sites would be limited to granophyre rocks on the submarine slopes of islands and reefs which make up less than 1% of the area of the Archipelago (Dortch, 2002). It was further argued by Dortch (2002) that other kinds of archaeological sites would have been obliterated or buried by marine conditions, concluding that future investigations must work to identify those shelf floor areas where the preservation of sites (or relevant non-archaeological features) will have been the greatest (see Dortch, 1997). In the Dampier Archipelago region there is, e.g., a concentration of rock art that is more recent around key economic resources (e.g. freshwater, fishing areas) and art that is older and is likely to have been similarly located near to palaeoeconomic resources (McDonald & Veth, 2000). The past (changes in) location of these palaeoeconomic resources and distribution of sites in general, therefore need to be defined better from geomorphological principles (see also Veth et al., 1993). Also, the nature and deposition of sedimentary deposits dating to the Holocene, and the coastal environmental changes that are inferred, are believed likely to form critical evidence with which to elucidate the causes of any archaeological transition (cf. Ward & Larcombe, 2008), and to increase understanding of the modern marine environment in general.

It was argued in the paper (Ward & Larcombe, 2008) that the archaeological potential is best considered as the combined probability of contemporary occurrence and long-term preservation, that latter having taken into account the destructive the destructive agents that stem from both natural and/or anthropogenic processes. There is relatively little understanding of the impacts of the natural dynamics of the marine environment, nor the later impacts associated with modern developmental activities (e.g. ports and harbours), upon the occurrence of archaeological records in the marine environment. In this paper Ward et al. (a) review the post-glacial geomorphological and stratigraphic development of the Dampier Archipelago, (b) present results of new palaeotidal modelling, and (c) use these to help describe the likely nature of the main landform elements and potential for associated archaeology. They aimed their research at gaining a better understanding of the areas geoarchaeological [potential?], to aid future management of any submerged archaeological sites that may occur in the Dampier region.

Aburara and Mardudhunera descendants and Wonggoottoo (or Ngarluma) people held the custodianship. According to Ward et al. it is not clear if the people who traditionally occupied the islands were a subgroup of the neighbouring Ngarluma people or a distinct group in their own right. Unfortunately events of the Flying Foam killings decimated these people (Mulvaney, 2011). Murujuga, meaning ‘hip bone sticking out’, is regarded by White (DAS, 1979) as referring to the rugged, rocky structure of the archipelago. Anecdotal evidence from traditional elders refers to submerged fish traps and drowned rivers. Ward et al. suggest that an improved understanding of the past landscape of the Dampier Archipelago would likely be of great interest to the Ngarluma and neighbouring Aboriginal groups in the area.

Regional setting

Regional geology and geography

The location of the Dampier Archipelago is on the inner part of the northwest marine region boundary (Baker et al., 2008) on the inner continental shelf. At this location the continental shelf is broad (~100 km wide) and comprises a shallow basin, with fields of mobile sandwaves and sand banks, large marginal plateaux that are separated from the surrounding shelf, that is wider, by saddles and troughs (Heap & Harris, 2008). The geomorphic development of the Archipelago was described (Semeniuk et al., 1982; Semeniuk & Würm, 1987), the understanding of which underpins the assessment of the geoarchaeological potential.

The size of the 42 islands comprising the Dampier Archipelago range from <20 m² to more than 33 m² and occupy an area of approximately 4,000 km². Legendre Island and some of the small outer islands are the remnants of consolidated limestone ridges which delineate a previous coastline (Semeniuk et al., 1982). The outer islands and the Burrup Peninsula (formerly Dampier Island, 118 m² in area are composed predominantly of granophyres, gabbro and basalt (Donaldson, 2011), that are cut by dolerite dykes which extend offshore (Semeniuk et al., 1982; Dortch, 2002). Precambrian rock, which is overlain by limestone from the Pleistocene, comprises the base of Mermaid Sound (Geofund, Pty Ltd. 1994). Immediately beyond the Dampier Archipelago there are older formations from the Pleistocene which include silty sand, minor gravel and coastal limestone deposits representing lithified dune, beach and offshore bar deposits (Kojan, 1994).

The limestone is overlain by alluvial and colluvial sand and gravel deposits that date to the Pleistocene and Holocene (Geofund Pty Ltd, 1994: 2). Unconsolidated calcareous sands, also known as limesand, are the youngest, regionally, deposits from the Holocene, which form a surficial blanket across most of the inner shelf, with localised accumulations of carbonate gravel  (Baker et al., 2008) and passing onshore to coastal dunes and old beach deposits (Hickman & Strong, 2003). Around the Dampier Archipelago the sediments of the seabed are influenced by the submarine topography as well as by the islands themselves, with localised depocentres of sand and mud that are thicker, especially between some of the islands. The geological surface beneath this upper unit from the Holocene deposits essentially defines the former Dampier landscape.

The northwest and southeast sides of the Burrup Peninsula have rocky coastlines that are interspersed by small bays with mud or muddy sand in the lower tidal zone, and are fringed with mangroves in the mud- to upper tidal zone. There were natural mudflats between the Burrup Peninsula and the mainland, which have now been developed into salt crystallisation ponds. In the 1960s industrial expansion effectively linked what was Dampier Island to the mainland, formal gazettal of the Burrup Peninsula in 1979. Dampier is now a major industrial port for the export of iron ore and the major processing of gas derived from the North-West Shelf. With the continued push for the commercial development of the Dampier Archipelago, as well as other areas of the adjacent continental shelf, there is a strong and increasing need to develop the archaeological knowledge base for management purposes (Bednarik, 2002; Mulvaney, 2011). Indeed, much of the understanding of cultural heritage of the Dampier region has resulted from commercial surveys prior to the construction of industrial facilities (Vinnicombe, 1987; Bednarik, 2007; Mulvaney, 2011).

The present oceanic regime

The Archipelago, which is situated on the inner part of the continental shelf, is subject to long period swell waves that originate in the southern Indian Ocean. Generally, the waves approach the Dampier depths of about 50 m or so, mean significant wave heights are about 0.4-0.5 with maxima about 2.3 m. Along the outer continental shelf the Leeuwin Current flows southwestwards, therefore having little effect on waters inside the Archipelago. At less than 1.3 m in height waves that are generated by the wind are small and are generated by local winds from the west in summer and from the southeast in winter (Semeniuk et al., 1982). Large swell waves, with significant wave heights of up to 10 m, may be generated by cyclonic disturbances, with wind gusts of 70 m/second, increased wave action and alongshore current surges exceeding 40 cm/second (Pearce et al., 2003).

There is an unusually large ocean tide in the northeast corner of the Indian Ocean, with a range of about 3 m (Mazzega, 1993). The result of the passage of this large tide across the northwest continental shelf is that the tidal currents are a dominant factor that controls the dynamics of the marine and coastal landscape with a semidiurnal tide that attains a maximum tidal range of 6.3 m, mean spring range of 5.6 m and mean neap range of 1 m at Dampier (Pearce et al., 2003). Seawards of the Archipelago tidal current speeds are up to 30 cm/second on spring tides and may reach up to 50 m/s at the offshore seawards entrances of the Archipelago, though they are generally much weaker (20 cm/s) in Mermaid Sound (Pearce et al., 2003). As well as the strong tides, currents in the Dampier Archipelago are also associated with local winds, large scale ocean circulation and continental shelf waves, and all are influenced by the complex local topography (Pearce et al., 2003). High waves, fast currents and the transport of sediment, and associated storm surges that may increase the height of high tide at the coast by up to 2.0 m, and important causes of these is cyclones (Jones et al., 2005). Off the coast of Western Australian tsunamis are rare, though they have been reported to have reached heights of 4-6 m on the northwest coast following the 1994 Java and 1977 Sumbawa earthquakes, respectively (Cummins, 2005), though the nature of the supporting evidence is not clear.

The glacial and post-glacial oceanographic regime – tidal modelling

According to Ward it is likely that the general regional oceanographic conditions varied through the postglacial transgression and the Holocene, though not necessarily in a manner that is easily characterised.  

1.     First, there is the nature of the tidal regime in the northeast Indian Ocean in the past that is relatively unknown. There are clear indications that in the open ocean away from the coast, the tidal range at lowstand about 20 ka BP was almost twice as great as at present (Rudnick & Ferrari, 1999; Egbert et al., 2004). Then, overall the ‘ocean tide’ will probably have increased in tidal range through the period of post-glacial rise of sea level, though the precise change pattern is as yet not known and it cannot be assumed that there was a steady monotonic decrease in ocean tidal range.

2.     There is the issue of the complex interaction of the tide with the width of the shelf, bathymetry of the shelf and coastal geography which are changing, as the post-glacial transgression occurred. In particular, as the flatter areas of the shelf were inundated by the rising post-glacial sea, tidal ranges and the speeds of tidal currents at a point on the shelf is in water that is increasingly deep are likely to have changed significantly over time as a result of seabed friction with the tidal wave and interaction of the tidal wave with the width of the shelf that is changing (e.g. Dyer, 1986). Therefore, not only were sea levels of the past variable, but also past tidal ranges and tidal dynamics.

Ward et al. used the Sparse Hydrodynamic Ocean Code (SHOC) model to assess tides of the past, while recognising the above uncertainties. Rather than trying to generate actual hindcasts of the tides (which, in their opinion, is not possible at present), however, they used to model to constrain the general nature of past tides. SHOC is a state-of-the-art hydrodynamic model that was developed by CSIRO Marine and Atmospherics research over the last 3 decades, that began with the pioneering work of Fandry (1981, 1982, 1983), (Fandry et al., 1985; Fandry & Steadman (1989, 1994) and improved later by Herzfeld & Waring (2008) and (Herzfeld & Waring, 2008; Herzfeld et al., 2008). SHOC was configured on a 5 km x 5 km horizontal grid and 26 layers in the vertical, ranging in thickness from 5 m near the surface of the ocean to 100 m depth that exceed 100 m. An area of 750 x 750 km off the Dampier region was covered by this model. Bathymetry data was obtained from Geoscience Australia, which has a horizontal resolution of ~200 m.

The model was forced along the open boundaries by tides only using:

1)    The modern ocean tide, and

2)    The likely tide at the Last Ocean Maximum (LGM).

The model was run for both, and there runs for 7 scenarios each corresponding to a decrease in water depth uniformly from 0, 10, 20, 30, 40, 50. 80-120 m. Each run of the model corresponded to a period of 32 days, which allowed for a full spring-neap tidal range. Ward et al. chose to use the modern bathymetry, and merely change the level of the water to simulate conditions in the past, though without detailed information on the changing sedimentary bodies and/or areas of net erosion through the period of post glacial sea level rise. Also they modelled the tides using only 4 principal tidal components and did not include any hydrodynamic influences of winds or waves, as their geoarchaeological focus was principally on the tidal ranges as a control on the past configurations of the coasty and the preservation of ancient archaeological sites. According to Ward et al. these simplifications are sufficient for their present purposes.

Regarding the geoarchaeology, the key result is that the coastal tidal ranges were of sufficient magnitude throughout the period of the transgression, so from a purely geomorphological perspective, there was the potential for the occurrence of significant areas of intertidal habitats that were potentially exploitable. Also, during the middle part of the Holocene transgression tidal currents were likely to have been at their strongest, with implications for the preservation of coastal deposits that formed prior to this period.

Wave incident at the coastline during the glacial sea level lowstand will probably have been very strong, at a time when the coastline was located about 150 km northwest of the present coastline and the seabed to seawards was relatively deep.  Ward et al. suggest that as the shelf was inundated and the waves were moderated by the shallow bathymetry, the incident waves at the transgressing shoreline will have become weaker. There may have been a slight increase in cyclonic activity through the Holocene as a result of the increasing mass of shallow warm shelf waters (see Larcombe & Carter, 2004 for the essential argument) though there may have been little significant change in cyclonic activity as the relative sea level stabilised about 6,500 BP. The detailed effects of cyclones on the transport of sediment on the shelf and at the coastline are hard to determine at the present, but it is likely they were significant throughout the highstand of sea level, ~6.5-4.0 ka BP, as a result of the potential for generation of strong along-shore currents to the southwest that were driven by the wind.

Reconstructing past coastal environments

Records of past sea level

The reconstruction of landscape that has been drowned relies critically on an understanding of sea level change in the past. As well as with sediment and sediment processes, sea level controls the likely location of archaeological sites (Ward et al., 2006). The need for chronological control in relation to the last marine transgression has been identified for the Dampier Archipelago (McDonald & Veth, 2009: 52) and for the adjacent Burrup Peninsula (Veth et al., 1993). There are, however, few sea level data or relative sea level curves for the offshore coast of Western Australia, and no marine sea level data for the Dampier itself. This is due partly to the erosive or at least non-depositional environment for much of the transgressive and highstand continental shelf (Dix, 1989; Dix et al., 2005; Baker et al., 2008). The relative lack of substantial coral reefs, the dominance of calcareous sediments (Baker et al., 2008) and the large tidal ranges, the latter particularly along the northern parts of the Western Australian coast are linked with the overarching sedimentary control. Therefore, for this study Ward et al. used sea level data from the wider region, as well as from the sea level curves that were isotopically inferred (Chappell et al., 1996; Yokoyama et al., 2006). Radiocarbon ages are given as calibrated values (cal ka BP) as calculated anew using the ShCal04 curve (McCormac et al., 2004) in 0xCal v.4.1 (Ramsay, 2011) or as provided by the original publications.

The considerable tidal range associated with the extensive continental shelf (average 100 km wide) through the post glacial Holocene further complicates the estimating of mean sea levels of the past. Also, there is considerable tidal range along the coastline of northwestern Australia of the present (Mazzega, 1983). Changing tidal ranges are an important confounding factor on regional reconstructions of past sea levels, as there are few, if any, geological indicators of mean sea level (see review of Larcombe et al., 1995). As a consequence it is very difficult to use observations of submerged morphology to infer sea levels of the past, or, because of the changed nature and location on the shelf and coastal sediment bodies, to infer the relief of former landscapes in some areas (Wywroll et al., 1995). However, Ward et al. combined the chronological data that were available with modern bathymetric data to develop a series of maps which illustrate the inundation and formation of the Dampier Archipelago, as a first-order approach. As well as geochronology, these maps were used to review the post-glacial geomorphological and stratigraphic development of the Archipelago, as well as assessing the landform elements for their archaeological potential.

Marine regression, Late Pleistocene (>45 ka BP, sea level ~50 m lower than present)

The area westward of the Dampier Archipelago prior to the last marine transgression was a coastal sand plain that was scattered with knolls and hillocks that were dissected by river courses, with a shore line at a distance of 110 km to the west (Semeniuk et al., 1982; Semeniuk & Würm, 1987). Barrow Island, 100 km to the west (with a maximum elevation of (+62 m) and Montebello Island (+27 m) formed dissected limestone hills on the lowland landscape (Bird & Hallam, 2006; Veth et al., 2007). The area of the Dampier Archipelago provided a terrain that was more rugged and geologically complex, with Mount Burrup (129 m) and Dolphin Island (120 m) forming the highest points.

Regional fracture trends or the location of weathered dolerite dykes, controlled the configuration of rives that dissected the coastal plain (Semeniuk et al., 1982: 103). It was confirmed by surveys for the port of Dampier that a natural channel (<12 m) existed between West Lewis and Malus Islands and also between Rosemary and Enderby Islands (Geofund, 1984). The northwards running waters of the Nickols and Maitland Rivers are located to the east and west of the Burrup Peninsula. As is the case now, these rivers probably flowed for short periods following rainfall events (Semeniuk et al., 1982). According to Ward et al. the Fortescue River may have formed a principal dividing rive system between the Montebello Islands and the Dampier Archipelago in the Late Pleistocene, though at present it debouches about 75 km further southwest of Dampier. These various rivers together formed a series of overlapping or adjacent alluvial fans, deltas and other coastal deposits that in their relative dominance as accretionary systems were staggered in time (Semeniuk et al., 1993: 238). A priority for determining archaeological palaeo-resources in this region is tracing these river systems off shore. Palaeoriver systems can also be useful in recording the response of alluvial and estuarine depositional environments to sea level change (e.g. Ryan et al., 2007). On the shelf of northeast Australia lowstand palaeochannels have been identified (e.g. Fielding et al., 2003, 2005; Ryan et al., 2007), though there has been little investigation as yet of similar palaeochannels off northwest Australia.

Aeolian shoreline calcareous sand ridges that have been built up along successive transgressive shorelines (Semeniuk et al., 1993), associated with the accretion and abandonment of low angle alluvial fans. Ward et al. suggest particular features might correspond to 1 or more periods of dune building in Australia centred around 70 ka BP, 42 ka BP and/or 35 ka BP (Rhodes et al., 2004: 301). Consolidated dune limestone on the outer islets and island in the north of the Archipelago are formed by abandoned dune barrier systems dating to the Pleistocene (Jones, 2004). Cementation of sandy calcareous sediments forms such limestone in the zone where groundwater emerges onto the intertidal surface. The presence of such indurated deposits provides, accordingly, an indication of groundwater levels in the past and, if related, something about sea level. Archaeological evidence of past use or occupation may also be preserved by cemented Pleistocene dune deposits, or capped by indurated sediment. Indurated deposits occur commonly in the Dampier Archipelago, and, as exemplified on Enderby Island and watering Cove, embedded in many of them are stone artefacts and marine mollusc shells of species that are often by eaten humans (Dortch, 2002: 39). As noted (O’Connell & Allen, 2007: 401), however, evidence of littoral and marine resource use in pre-LGM times is scarce, probably due to most of the relevant archaeological deposits being under water.

In this region the first evidence of Aboriginal occupation dates to ~35 ka BP associated with hearth features and artefacts at a site in the inland Pilbara (Law et al., 2010), to ~34 ka BP at Cape Range (Morse, 1993) and ~27 ka BP on the Montebello Islands (Veth, 1994; Veth et al., 2007). In the Dampier Archipelago there are no early dates for occupation, the only Pleistocene date coming from a fragment of trumpet shell (syrinx aruanus) that has been dated to ~21 ka BP (Lourblanchet, 1992). There is no question that at times pf lower relative sea level the northwest Australian shelf was occupied by Aboriginal groups that subsisted on marine as well as land based resources (McDonald, 2011: 12).

Sea level lowstand, LGM (~18 ka BP, sea level ~120 m below present)

The oldest marine radiocarbon date available for the northwest shelf region is ~25 ka BP, and Ward et al. suggest that even this date, which was obtained from a calcareous tube, is probably older than the sediments in which it occurs (James et al., 2004). Therefore, interpretation of Late Pleistocene sea level change needs to make use of regional data, such as the sea level curve (Yokoyama et al., 2006) that was isotopically inferred. The maximum land exposure of the Dampier Archipelago region, based on this curve, would have occurred during the LGM, at about 18 ka BP, with a eustatic sea level stillstand at about 130 m depth. This stillstand is regionally evident from submerged strandlines (or shorelines) at -120 m, and extending along the entire continental shelf from the North West Cape to the Bonaparte Gulf (Jones et al., 1973, his Fig. 16). Submerged shorelines at depths less than 120 m are less well preserved, occurring at various levels between -105 and -60 m (Jones et al., 1973: 37), which indicates an absence of lengthy stillstands during the Holocene part of the post-glacial transgression. Strong tidal currents may, however, have limited the development and preservation of some shoreline sequences.

Therefore the contemporary shoreline during the LGM was around 160 km offshore from the Burrup Peninsula and 50 km offshore from Barrow Island (Hook et al., 2004). The configuration of the lowstand and low-lying coastline would appear to have been relatively simple, based on the modern bathymetry and sedimentary history (e.g. James et al., 2004). The rocky hills of the Dampier Archipelago, as well as Barrow Island, the Montebello Islands and the Cape Range areas would have been part of the hinterland backing the extensive coastal plain at this time (Morse, 1993; V eth, 1994; Veth et al., 2007). It is not likely that the major rivers in the region, the Nickol, Maitland and Fortescue Rivers were very active during this period. Freshwater seepage can occur, however, via buried, discrete aquifers, which might exist in embayments or in the connective tidal lands between islands (Semeniuk, 1993) seepages have provided freshwater for thousands of years to coastal Aboriginal groups in northwest Australia (Mathews et al., 2011). It has in fact been argued (Mulvaney, 2013: 9) that the presence of potable water in the Dampier Archipelago, was an important factor bringing people to this area during the more arid phase. Subterranean freshwater seepage from the hinterland occurs well into the dry season even at present (Semeniuk & Würm, 1987).

There is an apparent decline in occupation of sites at this time (Shulmeister, 1992; however see Smith et al., 2008), with the exception of the arid shoreline of northwest Australia, where it is argued that small, mobile groups of people utilised hinterland and coastal resources during the Late Pleistocene (Veth, 1999; Veth et al., 2007; see also Hallam, 1987). The development of tidal inlets and lagoons, together with large areas of intervening intertidal flats, could have provided abundant food for humans, would have been favoured by the large tidal range of 4-5 m. Therefore, there may have been significant areas of potentially exploitable intertidal habitats along the northwest coastline during the LGM, rather than being diminished in terms of resources (O’Connell & Allen, 2012). Archaeological data that dates from this period supports this (e.g. O’Connor & Veth, 2000; Veth et al., 2007).

Mid-transgression (~18-11.7 ka BP, sea level rising towards ~50 m below present)

Global warming and consequent melting of the polar ice sheets from about 18 ka BP lead to a rapid rise in relative sea levels and later, the re-initiation of the summer monsoon (Wyrwoll & Miller, 2001). It is indicated by the tidal modelling results that tidal currents were likely to have been at their strongest during this phase of marine transgression when relative sea level was 50-30 m below present levels, and tidal range was at least as large as at present. Ward et al. suggest the coastline would have been within 30 km of its present location at that time, which make the macrotidal coastline accessible to small, mobile groups of people that occupied the area of the Dampier archipelago.

Yet another period of dune building occurred at this time (Rhodes et al., 2004), with a large seasonal supply of sand being provided by monsoonal winds to the coastline. In the area of the Fitzroy River further to the north, some circumstantial evidence was found that longitudinal dunes dating to this age may be preserved below the sea level on the shelf (Fairbridge, 1964 in Wyrwoll, 1979: 134), and evidence is provided by geophysical records for similar drowned dune systems dating to the Pleistocene within the Dampier Archipelago (Geofund Pty Ltd, 1994). There is strong evidence of past shoreline features being preserved as bathymetric features on the sea floor elsewhere in Western Australia, which includes off Port Hedland at water depths of -13.5 to -12.5 m (BHP, 2011), off James Price Point, near Broome, at water depths of -15 to -8 m (eureka, 2012) and in deeper waters (>50 m) off the Carnarvon Shelf (Nichol et al., 2012) and at Ningaloo Reef (Collins et al., 2003). According to Ward et al. the cemented limestone features drawn by Kojan (1994) can be interpreted similarly as submerged shoreline features. However, it is not yet clear which of these palaeoshoreline features are from the Pleistocene and which are from the Holocene, though they should be amenable to dating by luminescence. It is indicated by the existence of such shoreline features that have been submerged, that regressive or transgressive coastal dune (rather than lowstand coastal plain dunes as have been described (Hearty & O’Leary, 2008)) are preserved on parts of the upper continental shelf areas of Western Australia.

The mid to late transgression would have been associated with a much more complex coastal configuration as the (now inner shelf) topographic highs became inundated, to form estuarine and then marine embayments between major ridges, rocky promontories, headlands and offshore islands, as has been argued previously (Larcombe & Carter, 2004) for the central Great Barrier Reef shelf. A greater diversity of coastal sedimentary environments available to be exploited by humans would have been produced by increased coastal complexity, which is consistent with the increased depiction of marine fauna in the rock art record. O’Connor & Chappell (2003) did indeed argue that the opportunities for coastal exploitation would have been greatest at times of rising sea levels when reefs, mangrove belts and estuaries would have been richest and in closest proximity with each other. There is certainly pollen evidence just off Cape Range Peninsula for the establishment of mangrove ecosystems from about 14 ka BP (Van der Kaars & DeDeckker, 2002; see also; Van der Kaars, 1991).

Late transgression – Early Holocene (~11.7-6.5 ka BP, sea level -50 to +1.5 m)

It is indicated by coral records from the Abrolhos Islands (~1,100 km to the south) that between 9.8 and 8.0 ka BP, there was a generally rapid rise in sea level (~10 m/ka by U/Th age determination), reaching ~10 m below the present level about 8.2 ka BP, at which time the rate of rise decreased to about 3.3 m/ka (Collins et al., 1993). According to Ward et al. the earliest reliable sea level dates relate to the terminal Pleistocene, the time when sea level began to encroach on the outer limits of the Dampier Archipelago, with a U-series date of 9.8 ± 0.9 ka BP at 24 m below present levels (Eisenhauer et al., 1993). The islands formed when the land between many of the larger topographic points of the Dampier Archipelago became inundated. With continuing sea level rise over the following 1,000 years completed separation of most of the high points from the mainland to form the Mermaid Strait. Coastal occupation in the archipelago began at ~8 uncal. Ka BP (McDonald & Veth, 2009: 56), as soon as the sea reached what are the outer islands of the present.

The then shoreline would be between East Lewis Island and East Intercourse Island, with an arm or spit reaching out towards Withnell Bay, if the interpreted palaeoshorelines sequences of Kojan (1994) are valid. Evidence which supports this came from midden sites that dated to about this period on the Burrup Peninsula (Vinnicombe, 1987; Bradshaw, 1995) and the outer islands of the Dampier Archipelago (Bradshaw, 1995; McDonald & Veth, 2009). Dates which are consistent with coastal proximity are provided by excavation of a stratified shell deposit within a sand dune Wadjuru Rockpool) on Rosemary Island. The lower layer is comprised predominantly of mangrove shellfish and at 120 cm depth is dated to 9.5 ka BP (Bradshaw, 1995: 37). The upper layer was dated to 7.3 ka BP and is dominated by species of a rocky shore (Bradshaw, 1995, 37). A date of 9.9 ka BP, which has not been published, (8.9 ± 0.42 uncal, ka BP, S-ANU 18636) has now been obtained (by KM) for a surface midden that comprises Terebralia species shell (mangrove habitat gastropod), which was located adjacent to a creek in the western portion of Enderby Island. If palaeoshorelines sequences are actually represented by the limestone sequences of Kojan (1994), these midden dates from the early Holocene may well approximate a seawards position of the shoreline at or immediately beyond these islands.

It is indicated by the records of the local midden from this period a predominant use of resource from mangrove systems (Lorblanchet & Jones 1979: Bird & Hallam, 2006), which is consistent with the development of macrotidal conditions, intertidal areas that were large and more humid conditions which allowed the development of extensive mangrove systems (Semeniuk et al., 1982). A wide-range of mangrove-associated fauna on nearby Montebello Island, which is ~100 km to the west, that included gastropods (Terebralia sp., Telescopium sp.), mud crabs (Scylla serrate) and mud lobster (Thalassemia anomala) have been found in midden deposits that were dated to between 10 and 75 ka BP (Veth, 1999). At Cape Range, which is further to the south, where sea level may have reached near to the coastline of the present earlier than at the Archipelago, excavations have shown that Terebralia was being consumed from ~11 ka BP (Przywolnik, 2005: 190). At this time rivers that drained the mainland will have formed into larger deltaic systems, though they still had limited flow. Petroglyphs which indicate the growing importance of these estuarine and/or marine resources are thought to date from this time (Mulvaney, 2013).

Closely spaced hills of bedrock would develop into separate island systems, such as Enderby Island, Rosemary Island, Malus Island, and West Intercourse Island. Useful sea level indicators within their stratigraphy may have been provided by sedimentary infill of lowlands between these bedrock hills, now islands. A land bridge between the mainland proper and Burrup Island would have presented similar connective tidal deposits, but preservation of archaeological remains may have been compromised by the tidal creeks and channels that dissect these sediments at present. The local topography would have influenced circulation of seawater then, as now (Pearce et al., 2003: 30). Many of the inundated river valleys became straits, channels and embayments which can be traced into terrestrial drainage systems of the present (Semeniuk et al., 1982).

Highstand – mid-Late Holocene (c. 6.5 – 4 ka BP to present

A sea level highstand of +1-2 m above present level from 6.4 ka BP (Eisenhauer et al., 1993; Wyrwoll et al., 1995; Collins et al., 2006) until 4.0 ka BP, at least, or possibly as late as 2.5 ka BP (Lessa & Masselink, 2006). Emergent subtidal beds dated to 3.4 ka BP in the area of Port Smith (Hearty et al., 2006) and supratidal mudflats near Broome that dated to 2.7 ka BP (Baker et al., 2001; Lessa & Masselink, 2006), provided supporting evidence for this sea level highstand. Further evidence of this highstand is (or was) apparent at Dampier Point (Point Parker) (Bednarik, 2007) as well as Port Hedland (Mulvaney, 2011). A macrotidal regime that was similar to that of the present would have enhanced this higher sea level, and during cyclones storm surges raising sea level further above the astronomical tidal height. This equates to between 2,400 and 4,000 years of higher sea levels, which will have influenced occupation in the past and the use of the area.

Some lowland coastal barriers would have been breached by a sea level oscillation of 2 m above present levels, and would have inundated nearby lowland areas, and also probably would have divided some of the larger islands (e.g. Burrup; Enderby Island) into smaller island units. At the present low lying islands such as Cohen Island, Brigadier Island, Lady Nora Island, Conzinc Island, Sandy Island, Boler rock and Bare Rock would have been submerged during cyclonic storms much as they are during cyclonic storm surges of the present. Ward et al. suggest that it is possible that fringing intertidal platforms and rocky reefs, such as those around Kendrew, Brigadier, Legendre, Haüy and Delambre Islands may have formed about this time.

A change in midden composition is indicated by dated shell middens after ~4 ka BP, from being dominated by mangrove species, which included Anadara granosa from rocky shoreline, sandy beaches or mudflats (McDonald & Veth, 2009 and references therein). Changing shoreline ecology (see Clune & Harrison, 2009: 71) as both mangroves  (Creswell & Semeniuk, 2011) and shellfish (Morrison, 1983) are sensitive in terms of the habitat they require, may have been linked to the change in shellfish diet.

Following the highstand of the mid-Late Holocene, there was a progressive decline in sea level with the present height being reached at about 1.5 ka BP (Wyrwoll et al., 1995). Low lying islands and hinterland areas by this fall, with the possible result that progradation of salt flats occurred, with narrow bays and inlets being infilled, and erosion in older morphological features. An interplay between the accumulation of sediment (of alluvial and tidal deposits) and the tidal erosion and degradation of the sand plain (Semeniuk et al., 1982) associated with changing sea levels, are reflected in offshore lowland delta systems that are associated with the Maitland and Nickol Rivers. As a result of this there are a variety of tidal and marine habitat types that are unique to the Maitland delta complex (Semeniuk et al., 1982: 101). Ward et al. suggests it may be possible that these nearshore and shoreline environments preserve a Holocene archaeological record.

Towards a predictive model of archaeological potential

In the marine areas of the Dampier Archipelago the relative lack of geomorphological and chronological detail limits interpretation of archaeological and palaeoenvironmental patterns. As discussed in this paper, however, there is sufficient information available to make a preliminary assessment which includes potential for site preservation within the preserved landscape.

It was concluded by Ward et al. that protected embayments (tidal flats and shallow subtidal bays) and large intertidal areas have archaeological potential that is relatively high. Some dune systems, particularly cemented dunes, may also offer a high potential for archaeology (e.g. Eureka, 2012), as well as possibly helping to locate palaeoshorelines. They wide elevation that mangrove forest and coral reef cover limits their usefulness as sea level indicators (Lewis et al., 2008), though their distribution may have been altered in association with nearshore sedimentary and saline conditions. ‘Fixed biological indicators’ such as oysters, tubeworms and barnacles associated with rocky shores can allow the determination of more accurate evidence of sea level changes (Baker et al., 2001; Lewis et al., 2008).

The tidal modelling of Ward et al. indicates, at the same time, that there is uncertainty in regard to tidal ranges during the post-glacial transgression, and therefore significant uncertainty of the elevations (and archaeological significance) of mean sea level through that period. Therefore, it is more important to assess the way in which shoreline environments will have been used by past occupants, instead of trying to determine the precise mean sea level per se, from a geoarchaeological perspective (see also McDonald & Veth, 2009). Understanding the past may best be achieved by concentrating on those sedimentary units which have the greatest likelihood of preserving archaeological sites (Ward & Larcombe, 2008; see also Dortch, 2002). Investigations might in this way first focus on the uppermost intertidal and low supratidal zones (i.e. high tide), as there is a high potential for the presence there of occupation and midden deposits. Similarly, low tide deposits are probably important, as they would be associated with a likely foraging area that would be relatively rich (though with low potential for archaeological remains).

The sensitivity of the shoreline ecology to sedimentary and environmental conditions means that it may be possible to obtain useful information from ecological and geological studies of past and present mangrove, salt flat and estuarine environments. In the Dampier Archipelago the current environmental situation has a profound influence on the ecology of the area (Pearce et al., 2003), and will have influenced similarly ecology and the use of resources in the past. It is likely that resource-rich environments have been the focus for resource use or occupation (including the depiction of rock art). Some insight into sources of freshwater and mangrove diversity, and therefore available palaeoeconomic resources, may be offered by possible links between subterranean ostracods and development of host aquifers (Reeves et al., 2007).

It is also important to define the extent of offshore river and channel systems from both an archaeological and palaeoenvironmental perspective. The optimum locations for preserved prehistoric features during transgression are areas that are relatively protected, such as landwards-facing channel sides of meandering estuarine channels. In the case of Dampier Archipelago, this means mangrove flats. Delineation of all of these features might be helped by surveys with high resolution sub-bottom geophysical profilers.

According to Ward et al. future research first needs:

1.     To identify the spatial distribution of sedimentary units that are submerged offshore and date them, in order to distinguish between Holocene and modern sediment bodies.

2.     More work is required to fully record and date the lowstand dune systems that occur on the mid-to-upper continental shelf areas of Western Australia.

In the Dampier Archipelago and nearby initial documentation of probably submerged dune sequences indicate that these submerged shoreline systems have not been destroyed by transgressing seas (cf. Hearty & O’Leary, 2008) and they have the potential to provide important insights into understanding sea level change in the Pleistocene and Holocene. It is also apparent that the critical nature of establishing a local and regional chronology that is applicable to both terrestrial and marine deposits, and which extends beyond the end of the Pleistocene.

Conclusion

The natural features and the archaeological heritage of the area remain at risk from existing and future development, though the national heritage significance of the Dampier Archipelago is recognised. The potential for discovery of submerged archaeological sites is part of this significance. By the use of the available chronological and geological data, combined with new tidal modelling, Ward et al. provide an assessment of the key sediment bodies from the Late Pleistocene and Holocene that may preserve submerged archaeological deposits.

They suggest that it is most likely that archaeology will be present in deposits associated with the early phase of inundation of the Dampier Archipelago itself, dated around 9-7 ka BP (relative local sea levels between about -30 m and -15 m). The coastal configuration at this time was most complex and the variety and scale of intertidal and shallow subtidal environments was increased. Contrasting with this, coastal configuration formed when the post-glacial sea levels were below ~50 m (i.e. older than ~12 ka BP) and has been subsequently to a mid-transgressive phase of faster tidal currents on the continental shelf, and may now be eroded or poorly preserved. The need to improve definition of sea level oscillations and sedimentary regimes of the past, particularly over the Holocene period, where it is possible to accurately map the onshore archaeological record is highlighted by this preliminary assessment. One important means of resolving past sea level change and shoreline morphology is analysis of the nature and age of coastal dune sequences that are submerged which accumulated during periods of low sea level (and which may have been differentially preserved).

The accuracy of predictive modelling of land use in the past and archaeological site distribution will be developed by further geochronological and geoarchaeological research along the lines suggested in this paper. This will in turn improve prospecting techniques and the future management of the submerged element of Western Australia’s rich archaeological heritage that is as yet unknown.

Sources & Further reading 

Ward, I., et al. "The potential for discovery of new submerged archaeological sites near the Dampier Archipelago, Western Australia." Quaternary International(0).