Australia: The Land Where Time Began

A biography of the Australian continent 

Aboriginal Settlement in the LGM at Brockman, Pilbara, Western Australia

The authors say this paper describes the results and implications of recent excavations on the Hamersley Iron Brockman 4 tenement, close to Tom Price in Western Australia. The results were from 2 rock shelters in which evidence of Aboriginal occupation was found that began at least 32,000 BP, continuing throughout the Last Glacial Maximum (LGM). The authors1 propose the nature of Aboriginal foraging patterns that are displayed, based on the records of flaked stone and faunal remains, for the Brockman region.

Since excavations began in the Pilbara about 30 years ago there are 2 research questions that are still to be answered the age and continuity of the occupation sites. In regards to the settlement of the continent the age of the sites continues to be important for determining the timing and the direction from which the settlement took place. At the time the first synthesis and review of the initial archaeological excavations for the Hamersley Plateau (Brown, 1987) the oldest date that had been determined for an archaeological site was 26,300 + 500 BP (SUA1510) that had been obtained for the Newman Rock Shelter (P2055.2) (Brown, 1987:22, Troilett, 1982). It is not until recently that the antiquity of this site has been surpassed by estimate of the age from excavations at Djadjiling, Hope Downs in the eastern Hamersley Range, in spite of intensive archaeological work over 20 years, that was mostly carried out by consulting projects that included more than 50 excavations and more than 100 radiocarbon dates (see Slack, 2008). The date from Djadjiling indicates that occupation dates from at least 35,159 ± 537 BP (Morse in the same issue of Archaeology in Oceania).

According to Slack et al. (2009) there is a question concerning whether following the occupation of the Pilbara by the Aboriginal people the occupation continued throughout period of increased aridity that occurred during the last glacial period during OIS2, between about 29,000-15,000 cal. yr. BP (see Burroughs, 2005: 30, 93), and the aridity of the LGM in particular, a time when the sea levels were at their lowest between 22,000-19,000 cal. yr. BP (Yokoyama et al., 2000). Slack et al., (2009) say that over the past few decades there has been a consistent focus of research on the occupation of the Pilbara throughout the LGM, with the Hamersley and Chichester Rages being proposed as likely to have been refuges (Hiscock, 1988; Smith, 1987, 1989; Veth, 1989, 1993). Over these same few decades it has been shown by research that the impact on climate of the LGM has been more severe, and to have occurred at an earlier time than had previously been believed (and peaking in the Greenland ice-core isotope stratigraphy at 21,200 cal. yr. BP see Turney et al., 2006; Barrows & Juggins, 2005).

According to Slack et al., (2009) the nature of regional patterns of occupation during the LGM that has been revised and extended has been summarised on the basis of 7 specific rock shelter sites in the Pilbara Uplands that have been argued to exhibit refuge occupation during the LGM (see O’Connor & Veth, 2006: 33-39). Yirra (Veitch et al., 2005) and Milly’s Cave (Marwick, 2002) have been said to be the only sites to exhibit persuasive evidence of occupation during the LGM. The suggestion that there is no unequivocal evidence of occupation during the LGM at the remaining 5 sites (Marwick, 2002) is agreed with by (O’Connor & Veth, 2006). According to the analysis by Marwick the first 2 of these sites, Newman Rockshelter (Troilett, 1982) and Newman Orebody XXIX Rockshelter (Maynard, 1980), have stratigraphic records and radiocarbon chronologies that suggest, though don’t confirm, there is evidence of human occupation17,000-13,000 BP [i.e. 20,000-15,000 cal. yr. BP] (OxCal v4.0.5 was used in radiocarbon calibration in this article) (Marwick, 2002: 23; see also Comtesse, 2003). Similarly, evidence of human occupation at this period is regarded as ambiguous. Uncertainties have been a problem for the interpretation of artefacts and their relationship to carbon dates at Mesa J J24 (Hughes & Quartermaine, 1992), Malea Rockshelter (McDonald, Hale & Associate, 1997) and Manganese Gorge 8 (Veth, 1995:736).

The only sites in the interior of the Pilbara for which there is good evidence of occupation during the LGM   are Yirra and Milly’s Cave. It has been said that at Yirra and Milly’s Cave artefacts are found between conventional radiocarbon ages of 19,270 ± 140 BP (Wk-8954) (23,440 – 22480 cal. yr. BP) and (16,950 ± 90 BP (Wk-9148) (20,300 – 19,889 cal. yr. BP) which are consistent with refuge occupation during the LGM (Veitch et al., 2005:58). It is, however, not certain whether Yirra was occupied more intensively at the height of the LGM of immediately following it, as there are acknowledged unresolved problems with bioturbation, with critical dates at the peak, and with little additional information about the frequency of artefacts, the site and climatic history of the locality of the site.

According to Marwick’s 2002 paper the only clear indication of human occupation during the LGM has been found at the Milly’s Cave site. Slack et al. (2009) agree but it is suggested by re-evaluation of the timing of the LGM (Yokoyama et al., 2000; Lambeck & Chappell, 2001) that the site may have been only sporadically occupied before the close of the LGM. More intense occupation immediately following the peak of the LGM is suggested by Slack et al. (2009), to plausibly be indicated by the lowest radiocarbon determinations and frequency of artefacts at Milly’s Cave, and beneath this level, between about 21,000 and 30,000 cal. yr. BP, rates of discard are very low (see Marwick, 2002:25). Slack et al. (2009) have also noted that at this site the lower 2 radiocarbon dates of 14,150 ± 320 BP (18,024 – 16,022 cal. yr. BP) and 18,750 ± 460 BP (23,686-21,075 cal. ye. BP) are separated by as little as 5 cm of deposit. Slack et al. (2009) suggest that, as such, the data from Milly’s Cave is more compelling evidence for increased occupation towards the end of the LGM, rather than an increased level of occupation throughout the LGM.

If it is accepted that the Hamersley Plateau was a refuge area for humans during periods when aridity was extreme, the question arises what was the nature of this occupation? Referring to Milly’s Cave, Marwick suggests the ranges of territories were of reduced area. It is considered more generally by O’Connor and Veth that retraction to and within the ranges occurred, though it would be evident that there would be differences in reference to local catchments that range from being abandoned completely through to increased use (O’Connor & Veth, 2006:41).

The shortage of evidence for subsistence is a significant barrier to understanding the utilisation of refuge areas before, during and following the LGM. The main evidence that is needed is organic remains that are systematically analysed in conjunction with flaked stone. The move towards broad-spectrum diets at the terminal Pleistocene that has been discussed (Edwards & O’Connell, 1995), but a true understanding of the phenomenon has not yet been achieved, mainly as a result of the very few sites containing evidence of occupation that includes faunal and floral remains, and not just flaked stone.

Solid evidence has been found showing that the Pilbara region was occupied prior and during the LGM, though cultural remains have been found in a few sites that have been excavated and even fewer sites for which the work has been published. Sites such as Newman Rockshelter, Newman Orebody XXIX Shelter, Malea Rockshelter and Milly’s Cave have contained little faunal material. Faunal remains were found at Malea (Edwards & Murphy, 2003), but mostly is still to be published, little more than a species list of fauna and the fact that it is highly fragmented had been published at this time this article was published (Edwards & Murphy, 2003:45). Faunal material was recovered at Malea in only some of the excavation units, being confined to the upper 16 units. It is argued by Edwards & Murphy that this distribution of faunal remains is likely due to preservation factors, not the actual absence of the remains. Further work has been carried out at Malea and the analysis is now in progress and it is hoped that it will supply information that is needed to increase understanding of the subsistence and settlement of the area. At Marillana A, faunal remains were preserved, though discussion is limited to a quantitative analysis of the density per stratigraphic unit (Marwick, 2005:1362-4).

In Newman Orebody XXIX faunal remains are limited to 1 macropod molar found in the top excavation unit (Maynard, 1980:5), and data is missing from Newman Rockshelter and Milly’s Cave (Marwick, 2003). A significant problem is caused in the understanding of refuge areas by the absence of faunal data, as well as to knowledge of subsistence of early Aboriginal settlers as a whole.

In this paper (Slack et al. 2009) have reported new sites in the region that have the potential to provide subsistence data that is important and frameworks that are more robust concerning Aboriginal settlement on the Hamersley plateau during the LGM.

Brockman 4, Hamersley Plateau – Excavations

Excavations at 2 particular sites of a series of excavations in the Pilbara, about 60 km west of the of the town of Tom Price, Juunkan-1 and Juunkan-2 have provided further substantiation for the antiquity of the occupation in this region of more than 30,000 years. They have also provided compelling evidence that occupation continued even at the height of the LGM, 22,000-19,000 cal. yr. BP. As the location of Brockman 4 mining tenement, in which the sites are located, is well within the central Hamersley Plateau and is more than 75 km north of the nearest substantial watercourse, the Ashburton River, though it is ephemeral, and this location of the sites makes the finds interesting and to some extent unexpected.

Juunkan-1 and Juunkan-2 are both located within a small ironstone gorge not far from a small ephemeral watercourse, the Purlykunti Creek. There are 3 other Rockshelter sites in this gorge, though the occupation sequence at all 3 is very recent. There is also a very large scatter of artefacts on an extensive floodplain below the gorge. It is believed that all the dominant raw materials, ironstone, chert, quartz and siltstone, are available from the creek at and near to open scatter.


Slack et al. say new information concerning the prehistory of the Pilbara is provided by the results of their excavations at Brockman. Early occupation of beyond 35,000 BP is further supported by the data they provided. A continual, though infrequent, occupation of the Brockman region during OIS 2, and even at the height of the LGM, is indicated by the cultural sequence at Juunkan-2.

Analysis of the data in terms of landscape use by hunter gatherers was limited by the size of the sample, though Slack et al. (2009) made a number of observations and suggested hypotheses. It is indicated by the evidence from these 2 rockshelters that people have been living in this area of the Hamersley during the LGM. The local population may have been more residentially mobile at times of more severe aridity than might be expected, given the dominant refuge models and their previous application to the Pilbara (see Veth, 2005: 101). According to Slack et al. (2009) it is clear that people were retreating into gorges on the margins of the ranges near main river courses, as well as making a more complex use of the landscape. It is also suggested that they were ‘possibly following local weather patterns and allowing access to the less drained areas occurred’ (Slack et al. 2009). In this paper it is also considered likely residential mobility decreased following the LGM as rainfall increased. If this suggestion is correct it would explain the high discard rate at Brockman; the increase in faunal remains density at Juunkan-2 in the later phases of occupation, as well as those trends in the flaked stone that were observed at Milly’s Cave (Marwick, 2002: 29). Slack et al. (2009) also suggest that though there was a decrease in residential mobility there was an increase in logistical mobility, at least on the local level. The greater range of the raw materials and the larger sizes of the flaked material over the last few thousand years of the Pleistocene, which continued to the Middle Holocene, is the basis for this suggestion. The intensity of the apparent reduction and frequency of artefact discard increased slightly in the Middle and Late Holocene. Behavioural implications that are suggested by the faunal remains provide additional support for the occupation increase that occurred in later periods, though it is also shown by the fauna that both Juunkan-1 and Juunkan-2 were occupied continuously in all periods. This trend is considered likely to be related to increased levels of population, as has been suggested (Marwick, 2002), given the results of other excavations within this area all dating to this period.


It is suggested that the results of this ongoing project further emphasise that the archaeology of the Pilbara region will continue to play an important role in developing an understanding of the timing of arid settlement, and the nature of hunter gatherer subsistence during periods of uncertainty such as the LGM.

Sources & Further reading

  1. Slack, M., M. Fillios and R. Fullagar (2009). "Aboriginal Settlement during the LGM at Brockman, Pilbara Region, Western Australia." Archaeology in Oceania 44(S1): 32-39.


Author: M. H. Monroe
Last updated: 25/11/2014
Journey Back Through Time
Experience Australia
Aboriginal Australia
National Parks
Photo Galleries
Site Map
                                                                                           Author: M.H.Monroe  Email:     Sources & Further reading