Mammalian responses to Pleistocene climate change in southeastern Australia

Gavin J. Prideaux^{1, *}, Richard G. Roberts², Dirk Megirian³, Kira E. Westaway⁴, John C. Hellstrom⁵, and Jon M. Olley⁶

1. School of Biological Sciences, Flinders University, Bedford Park, South Australia 5042, Australia, 2. GeoQuEST Research Centre, School of Earth and Environmental Sciences, University of Wollongong, Wollongong, New South Wales 2522, Australia, 3. Museum of Central Australia, Alice Springs, Northern Territory 0871, Australia, 4. GeoQuEST Research Centre, School of Earth and Environmental Sciences, University of Wollongong, Wollongong, New South Wales 2522, Australia, 5. School of Earth Sciences, University of Melbourne, Melbourne, Victoria 3010, Australia, 6. Commonwealth Scientific and Industrial Research Organisation, Land and Water, Canberra, Australian Capital Territory 2601, Australia

INTRODUCTION Return to TOC

Few topics in Quaternary science match the late Pleistocene megafauna extinction debate for the intensity of polemic it has generated, with most authorities championing either a human or a climatic cause. Nearly 170 yr after Charles Darwin first pondered the demise of giant mammals (Darwin, 1839), the debate over what killed the megafauna on the continents, and when, remains hampered by a lack of basic data on faunal change through time and on the ecologies of extinct species (Barnosky et al., 2004a). Australia underwent the worst extinctions of all the continents, losing 90% of its megafauna by ca. 45 ka (Roberts et al., 2001; Miller et al., 2005). Opinions remain strongly divided between the relative importance of climatic changes preceding and concurrent with human arrival (e.g., Trueman et al., 2005; Wroe and Field, 2006), and the activities of humans via overhunting or habitat disturbance (e.g., Johnson, 2005; Miller et al., 2005). The roles of these agents will be contended until we establish how communities responded to climate changes prior to human arrival. Until now, no Australian studies have detailed how regional fauna responded to climate change between glacial-interglacial cycles or across an individual cycle. This is partly because few well-stratified faunal successions have been discovered, but more so because of the rarity of field-based studies and the historical difficulties involved with dating deposits beyond the limit of radiocarbon dating (ca. 45 ka).

SETTING Return to TOC

The Naracoorte Caves World Heritage Area (NCWHA) in southeastern South Australia (Fig. 1) is one of very few localities on Earth known to preserve a composite record of mammals and other vertebrates spanning the past 500 k.y. (Moriarty et al., 2000). Limestone caves in the region contain sediment deposits that preserve rich, diverse mammal assemblages (Wells et al., 1984; Reed and Bourne, 2000), most of which accumulated via pitfall and/or as regurgitated pellets from roosting owls. Given the duration and richness of the mammal record, it is fortuitous that one of the best records of Australian terrestrial climate over the past 500 k.y. has been derived from NCWHA speleothems (Ayliffe et al., 1998). This indicates marked temporal fluctuations in effective moisture induced by glacial-interglacial cycling.

Two prior studies found no evidence of faunal change in the NCWHA record (Moriarty et al., 2000; Grün et al., 2001), but these were based on comparisons of composite species lists for sites spanning multiple climatic intervals, so any potential changes are far less likely to have been revealed than via detailed biostratigraphic analyses. Here we examine changes in species richness, assemblage composition, species abundances, and sediment attributes through five fossil-rich middle Pleistocene units in Cathedral Cave, and compare these with local late Pleistocene faunas. We relate the observed shifts to the speleothem paleoclimate record, and then consider the role of climate change in the extinction of the regional megafauna.

CHRONOLOGY AND SEDIMENTOLOGY Return to TOC

Our study entailed the excavation of a 6.25 m² × 1.1-m-deep pit on the distal extremity of the sediment fan in the Fossil Chamber of Cathedral Cave (Fig. 1). The morphology of the chamber and general nature of the deposit were described by Brown and Wells (2000). Excavation revealed a stratigraphy composed of five units, numbered sequentially from top to base. Optical dating of quartz sediment in unit 1 gives an age of 206 ± 16 ka (Fig. 2), indicating that the sequence-capping flowstone (²³⁰Th/²³⁴U age of 159 ± 2 ka, Brown and Wells, 2000) was deposited well after the entrance pipe was blocked, terminating sediment and faunal accumulation. Optical dating gives depositional ages of 231 ± 21 ka, 257 ± 21 ka, and 292 ± 19 ka for units 2a, 2b, and 3, respectively (Fig. 2). Two samples from unit 4 returned statistically indistinguishable optical ages with a weighted mean of 528 ± 41 ka, highlighting the existence of an ∼230 k.y. hiatus between the deposition of units 4 and 3. Hiatuses in sediment deposition reflect extended intervals of entrance pipe blockage, typical of NCWHA deposits that have accumulated via narrow solution pipes (Moriarty et al., 2000). The ²³⁰Th/²³⁴U dating on the 20–30-cm-thick flowstone separating units 4 and 3 indicates that calcite deposition commenced at 396 ± 19 ka. Our excavations reveal that the stalactite tip dated as 399 ka by Ayliffe et al. (1998) does not penetrate unit 4 (contra Moriarty et al., 2000). The upper portion of the flowstone complex, which includes lenses interbedded with the base of unit 3, is bracketed by ²³⁰Th/²³⁴U ages of 297 ± 9 ka and 289 ± 10 ka (Fig. 2).

Cave sediments are often highly reliable climate proxies (e.g., Barnosky et al., 2004b). Our petrographic analysis (see GSA Data Repository^{footnote 1}) reveals that unit 4 is composed of clean quartz sand with a low frequency of well-sorted, well-rounded clay aggregates, and a paucity of carbonate and charcoal clasts. This denotes a texturally and compositionally mature sediment deposited during a period of high effective precipitation, with a low frequency of wildfires. Unit 3 contains more clay aggregates than unit 4, probably reflecting a slight change in sedimentary provenance, but the units are otherwise very similar in texture and composition, so we conclude that unit 3 also accumulated during a relatively wetter interval. This is supported by the thick flowstone interbedded with the lower part of unit 3. While a distinct contact between units 2b and 2a is evident in stratigraphic section, the two reddish units are similar in their sedimentology. Clay aggregates, ranging markedly in size, shape, and composition, are dominant within both units. Carbonate lithoclasts and charcoal-rich horizons are common. These imply deposition during intervals of lower effective precipitation, with a high frequency of wildfires. Unit 1 is composed of beige-colored quartz sand, similar in composition to that of unit 4, interbedded with sediments similar to those of unit 2a. This implies greater fluctuations in effective precipitation during the accumulation of unit 1.

The petrographic analysis clearly indicates that several different climatic intervals are sampled by the Cathedral Cave sequence. While the commencement of the NCWHA speleothem paleoclimate record (Ayliffe et al., 1998) postdates the deposition of unit 4, the sedimentological evidence suggests that a broadly similar climate prevailed when units 4 and 3 accumulated. In light of the optical age for unit 3, the ²³⁰Th/²³⁴U ages on the upper portion of the flow-stone, and the fact that speleothem formation in the NCWHA ceased for ∼50 k.y. following the ca. 270 ka glacial maximum (Fig. 2; Ayliffe et al., 1998), unit 3 evidently accumulated during the 300–270 ka wetter interval. Units 2b and 2a were deposited during the drier interval that followed (270–220 ka). The optical age for unit 1 indicates that it was deposited during a period of variable speleothem deposition (Fig. 2; Ayliffe et al., 1998), which agrees with the sedimentological data. Overall, the general pattern in sediment attributes corresponds closely to the speleothem paleoclimate record (Fig. 2).

PALEONTOLOGY Return to TOC

Like most sites in the NCWHA, the small mammal (<5 kg body mass) component of the Cathedral Cave assemblage is dominated by murine rodents, while the most diverse and abundant large mammals (>5 kg) are kangaroos. A total of 62 nonflying species (see ^{footnote 1}) exceeds that collected from any other Australian Pleistocene site, except for the Fossil Chamber in Victoria Fossil Cave (also in NCWHA), from which 65 species have been retrieved (Reed and Bourne, 2000). Initial analysis of the Cathedral Cave deposit showed that sediments and animals entered through a single solution pipe (Brown and Wells, 2000). Because the mechanisms of faunal accumulation (large mammals fell in; small mammals fell in and/or their remains were regurgitated by owls roosting within the cave) were apparently constant through time, sampling biases must also have remained essentially the same. Therefore, variations in relative abundances of species between units are likely to accurately reflect changes in the mammal community prompted by climate-driven changes in local vegetation.

Sample sizes for large mammals are much lower for Cathedral Cave units 2a and 1 than for units 4–2b. We consider this an accurate reflection of their rarity in the local community, as both sediment volumes and small mammal densities are relatively high. It is also unlikely that a body-size filter (e.g., partly blocked solution pipe) was in place, because three of the four largest species (Zygomaturus trilobus, Procoptodon goliah, Thylacoleo carnifex) are present in both units. Minimal reworking within the site is confirmed by the articulated and associated skeletal remains of small and large mammals encountered in all units.

A plot showing species richness through the sequence reveals a peak for small mammals in unit 2b, which is directly responsible for the peak in overall species richness (Fig. 3A). To examine the influence of sample size variation between units on this pattern, we conducted a rarefaction analysis for each unit using Analytic Rarefaction 1.3 (developed by S. Holland; see ^{footnote 1}). This shows that the unit 2b peak is due to a larger sample size relative to the other units. Overall, small mammal species richness was markedly similar ca. 530 ka and 280 ka, and from ca. 280 to 200 ka. Large mammal sample sizes are lower than for small mammals, although there is substantial variation between units. The trend in density of identifiable large mammal specimens (number of specimens identified to species level per cubic meter of sediment) reveals a peak in unit 3 followed by a decline to consecutively lower levels in units 2b and 2a (Fig. 3B). Among the large mammals, both megafauna (extinct species with body mass estimates of >30 kg or attaining estimates of ≥30% greater body mass than their closest living relatives) and nonmegafauna follow generally similar trajectories (Fig. 3B).

FAUNAL RESPONSES TO CLIMATE CHANGE Return to TOC

Plots for a broad range of taxa (Fig. 4) reveal three general faunal trends across the glacial-interglacial cycle during which units 3–1 accumulated (ca. 280–200 ka; Fig. 2). Four small and several large mammals show trend 1, high relative abundances in unit 3 followed by marked subsequent decline (Figs. 4A, 4B). Two small and two large mammals exhibit trend 2, where relative abundances are low in unit 3, rise to a peak in unit 2b, subsequently decrease in unit 2a, and then remain stable in unit 1 (Figs. 4C, 4D). Trend 3 species are united by successive increases in relative abundance from unit 2b to unit 1, but vary in abundance changes from unit 3 to unit 2b (Figs. 4E, 4F). All other species are represented by insufficient specimens to infer any meaningful trends.

Habitat preferences of extant species that reach their highest relative abundances in unit 3 (Figs. 4A, 4B) along with the extinct browser guild (excluding Procoptodon goliah and ‘P.’ gilli) indicate that unit 3 was deposited during a wetter interval. This is supported by the flowstone interbedded with unit 3, the steep erosional contact between units 4 and 3, and the sedimentology. Today, Pseudomys fumeus, Potorous tridactylus, Macropus rufogriseus, and Wallabia bicolor occupy wooded habitats with well-developed shrub understoreys (Strahan, 1995). The presence of minor heathland and grassland habitats is intimated by moderate abundances of Pseudomys apodemoides, P. australis, and Mastacomys fuscus. All three faunal trends reflect substantial changes in the proximal vegetation by the time unit 2b accumulated. Most large browsers and thicket-dwelling mammals decline, but a range of species shows the reverse trend. The sharpest increases are shown by Pseudomys apodemoides (Fig. 4C), which lives in Banksia heathland and open mallee woodland in the region today (Strahan, 1995), and ‘Procoptodon’ gilli, a likely heathland specialist (Prideaux, 2004). Cercartetus lepidus and Macropus greyi show similar but less marked increases in unit 2b, and also prefer heathland (Strahan, 1995). The unit 2b fauna clearly reflects a drier, more heath-dominated environment than the unit 3 fauna, a shift strongly supported by the sedimentological and effective moisture records.

Species exhibiting trends 1 and 2 decline between units 2b and 2a. Concomitant increases in Pseudomys australis, Mastacomys fuscus, Perameles gunnii, Macropus giganteus, and Procoptodon goliah (trend 3; Figs. 4E, 4F) reflect the spread of grassland and open woodland (Strahan, 1995; Prideaux, 2004). Little is known of the ecology of the recently extinct Pseudomys auritus, but its occurrence in late Holocene deposits in grassland areas just east of the NCWHA (Strahan, 1995) corroborates the paleoenvironmental inferences drawn from the other species.

The NCWHA speleothem record (Ayliffe et al., 1998) shows that the region was characterized by low effective moisture between 270 and 220 ka; unit 2a (231 ± 21 ka) accumulated toward the end of this drier interval, at or soon after peak interglacial conditions at 240 ka (Fig. 2). Unit 1 (206 ± 16 ka) was deposited soon after the reinitiation of speleothem deposition in the NCWHA; although the megafauna show no sign yet of recovery, increases in relative abundances of three extant trend 1 mammals (Figs. 4A, 4B) highlight an incipient reversal of the regional climatic trend and foreshadow the onset of cooler glacial conditions (Ayliffe et al., 1998).

DISCUSSION Return to TOC

Faunal trends from units 3–1, including the marked decline in large mammal densities in the upper units, provide the first evidence from prehuman Australia to show that regional faunas responded to local environmental changes induced by shifts in effective moisture within glacial-interglacial cycles. Our data also reveal that unit 4 (ca. 530 ka) is nearly identical in species richness to unit 3 (ca. 280 ka), and 40 of the 44 species recorded in unit 4 are present higher in the sequence. In light of these observations, it is significant that 54 of the 62 Cathedral Cave species also have local late Pleistocene records (see ^{footnote 1}) (contra Wroe and Field, 2006) and that 4 of the 8 missing species are still extant. This implies that, while the 500 k.y. preceding the megafaunal extinction interval were typified by climate-induced fluctuations in local populations, the ranges of some larger species contracting away from Naracoorte during drier times, the mammal fauna was resilient in the long-term. For example, eight species of megafauna are apparently lost from the record in the 270–220 ka drier interval, but at least six of these species are known from local late Pleistocene sites (see ^{footnote 1}). Although no NCWHA succession has yet been identified that is within the 155–115 ka drier interval (Fig. 2; Ayliffe et al., 1998), the fauna clearly survived it. The Naracoorte evidence does not support long-term, staggered, climate-mediated species extinctions.

Large mammals were abundant during the 300–270 ka wetter interval, so the high diversity of large mammals in the NCWHA (Reed and Bourne, 2000; Roberts et al., 2001) within the climatically similar 90–75 ka interval (Fig. 2) is not surprising. Effective moisture at this time was high both locally (Ayliffe et al., 1998) and further inland (e.g., Magee et al., 2004). The period was also marked by sustained forest development to the southeast (Harle et al., 2002). Reduction of the large mammal fauna in the Cathedral Cave sequence likely initiated ca. 270 ka (a glacial maximum), and although it is conceivable that a similar climatic interval within the last glacial cycle (75–60 ka; Fig. 2) might have had some impact on populations, effective moisture actually remained steady locally until the last glacial maximum (Fig. 2; Ayliffe et al., 1998).

CONCLUSIONS Return to TOC

On the basis of the earlier part of the NCWHA record, the persistence of relatively cool, moist conditions for most of the last glacial should have favored the megafauna, a conclusion comparable to that drawn from inland localities by Miller et al. (2005). Therefore, extinction of the NCWHA megafauna by ca. 45 ka (Pate et al., 2002) cannot have been caused solely or primarily by climate change, especially given the persistence of all other mammals into the Holocene (McDowell, 2001). The disparity between pre-human and posthuman changes at Naracoorte resembles that described for an analogous North American mammalian succession in Porcupine Cave, Colorado (Barnosky et al., 2004b). These records show that mammal faunas on both continents were well adapted to Quaternary climatic variations prior to the arrival of humans.

Acknowledgments

We thank S. Bourne (South Australian Department for Environment and Heritage), GreenCorp teams led by M. Crewe and G. Bradford, members of the Friends of the Naracoorte Caves and the Cave Exploration Group of South Australia, the cave guides, and G. Gully for help with excavation and hauling sediment bags. Funding for Prideaux was provided, in part, by a grant to R. Wells from the Commonwealth Natural Heritage Trust Extension Bushcare Program. The Australian Research Council supported Roberts through a Senior Research Fellowship, and Westaway held a Postgraduate Award and a Tuition Fee-Waiver Scholarship from the University of Wollongong. We also thank J. Abrantes for assistance with optically stimulated luminescence dating sample preparation and measurement, M. McDowell for guidance with rodent identifications, L. Ayliffe, A. Baynes, and C. Johnson for commenting on drafts, and A. Barnosky, J. Mead, and G. Miller for constructive reviews.

REFERENCES CITED Return to TOC

1. Ayliffe, L.K., Marianelli, P.C., Moriarty, K.C., Wells, R.T., McCulloch, M.T., Mortimer, G.E., Hellstrom, J.C., 1998, 500 ka precipitation record from southeastern Australia: Evidence for interglacial relative aridity: Geology, v. 26, p. 147–150, 10.1130/0091–7613(1998)026<0147:KPRFSA>2.3.CO;2. Find this article online
2. Barnosky, A.D., Koch, P.L., Feranec, R.S., Wing, S.L., Shabel, A.B., 2004a, Assessing the causes of late Pleistocene extinctions on the continents: Science, v. 306, p. 70–75, 10.1126/science.1101476. Find this article online
3. Barnosky, A.D., Bell, C.J., Emslie, S.D., Goodwin, H.T., Mead, J.I., Repenning, C.A., Scott, E., Shabel, A.B., 2004b, Exceptional record of mid-Pleistocene vertebrates helps differentiate climatic from anthropogenic ecosystem perturbations: National Academy of Sciences Proceedings, v. 101, p. 9297–9302, 10.1073/pnas.0402592101. Find this article online
4. Brown, S.P., Wells, R.T., 2000, A middle Pleistocene vertebrate fossil assemblage from Cathedral Cave, Naracoorte, South Australia: Royal Society of South Australia Transactions, v. 124, p. 91–104.
5. Darwin, C.R., 1839, Journal of researches into the geology and natural history of the various countries visited by H.M.S. Beagle, under the command of Captain Fitzroy, R.N., from 1832 to 1836: London, Henry Colburn629.
6. Grün, R., Moriarty, K.C., Wells, R.T., 2001, Electron spin resonance dating of the fossil deposits in the Naracoorte Caves, South Australia: Journal of Quaternary Science, v. 16, p. 49–59, 10.1002/1099–1417(200101)16:1<49::AID-JQS570>3.0.CO;2-#. Find this article online
7. Harle, K.J., Heijnis, H., Chisari, R., Kershaw, A.P., Zoppi, U., Jacobsen, G., 2002, A chronology for the long pollen record from Lake Wangoom, western Victoria (Australia) as derived from uranium/thorium disequilibrium dating: Journal of Quaternary Science, v. 17, p. 707–720, 10.1002/jqs.684. Find this article online
8. Johnson, C.N., 2005, What can the data on late survival of Australian megafauna tell us about the cause of their extinction?Quaternary Science Reviews, v. 24, p. 2167–2172, 10.1016/j.quascirev.2005.06.006. Find this article online
9. Magee, J.W., Miller, G.H., Spooner, N.A., Questiaux, D., 2004, Continuous 150 k.y. monsoon record from Lake Eyre, Australia: Insolation-forcing implications and unexpected Holocene failure: Geology, v. 32, p. 885–888, 10.1130/G20672.1. Find this article online
10. McDowell, M.C., 2001, The analysis of Late Quaternary fossil mammal faunas from Robertson Cave (5U17, 18, 19) and Wet Cave (5U10, 11) in the Naracoorte Caves World Heritage Area, South Australia [M.S. thesis]: Adelaide, Flinders University378.
11. Miller, G.H., Fogel, M.L., Magee, J.W., Gagan, M.K., Clarke, S.J., Johnson, B.J., 2005, Ecosystem collapse in Pleistocene Australia and a human role in megafaunal extinction: Science, v. 309, p. 287–290, 10.1126/science.1111288. Find this article online
12. Moriarty, K.C., McCulloch, M.T., Wells, R.T., McDowell, M.C., 2000, Mid-Pleistocene cave fills, megafaunal remains and climate change at Naracoorte, South Australia: Towards a predictive model using U-Th dating of speleothems: Palaeogeography, Palaeoclimatology, Palaeo-ecology, v. 159, p. 113–143, 10.1016/S0031–0182(00)00036–5. Find this article online
13. Pahnke, K., Zahn, R., Elderfield, H., Schulz, M., 2003, 340,000-year centennial-scale marine record of Southern Hemisphere climatic oscillation: Science, v. 301, p. 948–952, 10.1126/science.1084451. Find this article online
14. Pate, F.D., McDowell, M.C., Wells, R.T., Smith, A.M., 2002, Last recorded evidence for megafauna at Wet Cave, Naracoorte, South Australia 45,000 years ago: Australian Archaeology, v. 54, p. 53–55.
15. Prideaux, G.J., 2004, Systematics and evolution of the sthenurine kangaroos: University of California Publications in Geological Sciences, v. 146, p. i–xvii, 1–623. Find this article online
16. Reed, E.H., Bourne, S.J., 2000, Pleistocene fossil vertebrate sites of the South East region of South Australia: Royal Society of South Australia Transactions, v. 124, p. 61–90.
17. Roberts, R.G., Flannery, T.F., Ayliffe, L.K., Yoshida, H., Olley, J.M., Prideaux, G.J., Laslett, G.M., Baynes, A., Smith, M.A., Jones, R., Smith, B.L., 2001, New ages for the last Australian megafauna: Continent-wide extinction about 46,000 years ago: Science, v. 292, p. 1888–1892, 10.1126/science.1060264. Find this article online
18. Strahan, R., 1995, The mammals of Australia: Sydney, Australia, Reed New Holland756.
19. Trueman, C.N.G., Field, J.H., Dortch, J., Charles, B., Wroe, S., 2005, Prolonged coexistence of humans and megafauna in Pleistocene Australia: National Academy of Sciences Proceedings, v. 102, p. 8381–8385, 10.1073/pnas.0408975102. Find this article online
20. Wells, R.T., Moriarty, K.C., Williams, D.L.G., 1984, The fossil vertebrate deposits of Victoria Fossil Cave, Naracoorte: An introduction to the geology and fauna: Australian Zoologist, v. 21, p. 305–333. Find this article online
21. Wroe, S., Field, J.H., 2006, A review of the evidence for a human role in the extinction of Australian megafauna and an alternative interpretation: Quaternary Science Reviews, v. 25(in press), doi: 10.1016/j.quascirev.2006.03.005.

Figures Return to TOC

Figure 1. Map of Australia showing location of Naracoorte Caves.

Figure 2. A: δ18O record from core MD97–2120, southwest Pacific (Pahnke et al., 2003), long paleoclimate record from mid-latitude Southern Hemisphere. B: Histogram of 230Th/234U ages for Naracoorte Caves World Heritage Area speleothems (from Ayliffe et al., 1998), composite of Gaussian probability curves for each date (normalized to unit area). Higher peaks indicate more sharply defined ages (higher relative probability of accuracy). Depositional hiatuses correspond to drier periods. Megafaunal extinction interval is from Roberts et al. (2001). Filled circles indicate mean optical ages for Cathedral Cave units 3–1 (errors at 1σ); open circles indicate 230Th/234U ages on flowstones (errors at 2σ).

Figure 3. Species richness and large mammal density trends through Cathedral Cave sequence. A: Species richness for all mammals, and for small (<5 kg) and large species (>5 kg). Expected small mammal species richness (E) is derived from rarefaction analysis (see footnote 1). B: Density trends for large mammals. NISP is number of specimens identified to species level.

Figure 4. Species relative abundance trends across one glacial-interglacial cycle: Cathedral Cave units 3–1 (ca. 280–200 ka). A: Trend 1 small mammals. B: Trend 1 large mammals. C: Trend 2 small mammals. D: Trend 2 large mammals. E: Trend 3 small mammals. F: Trend 3 large mammals. Relative abundance is MNIspecies/MNItotal% (MNI—minimum number of individuals).

^{footnote 1} GSA Data Repository item 2007016, paleontological and dating methods, data tables, species lists, tables for faunal relative abundance data, and petrographic descriptions and interpretations, is available online at www.geosociety.org/pubs/ft2007.htm, or on request from E-mail: editing@geosociety.org or Documents Secretary, GSA, P.O. Box 9140, Boulder, CO 80301, USA.

*Current address: Department of Earth and Planetary Sciences, Western Australian Museum, Perth, Western Australia 6000, Australia; E-mail: gavin.prideaux@museum.wa.gov.au