Late-Holocene environmental change on the Nullarbor Plain,southwest Australia,based on speleothem pollen records

Abstract

Fossil pollen from two stalagmites is examined to reconstruct a c. 2400-year history of vegetation change on the Nullarbor Plain. Environmental changes are reflected by variation in chenopod species abundance, and by a peak in woody taxa between 1000 and 800 years ago which is interpreted as evidence of increased moisture conditions associated with a positive phase of the Southern Annular Mode. While no strong palynological signal is observed at the time of European colonization of Australia, a significant change occurs in the past 40 years, which is interpreted as a vegetation response to a recorded fire event. As speleothems (secondary cave carbonates including stalagmites, stalactites and flowstones) rarely contain enough fossil pollen for analysis, the taphonomic biases of speleothem archives remain poorly understood. This study, as well as being a high-resolution record of environmental change, presents an opportunity to examine these taphonomic filters. The record is shown to be sensitive to episodic deposition of presumably insect-borne pollen, but overall appears to provide a faithful representation of local and regional vegetation change. There is a need for greater research into taphonomic processes, if speleothem palynology is to be developed as a viable alternative to lacustrine sediments in the investigation of past environmental change.

Keywords

arid ecology palaeoecology palynology speleothem

Introduction

Arid climates are generally unfavourable for the preservation of organic remains, and for this reason, the Australian arid and semi-arid regions are poorly represented in reconstructions of late Quaternary palaeovegetation and palaeoclimate (Dixon et al., 2017; Petherick et al., 2013). A small number of fossil pollen records have documented late Quaternary environmental changes in Australia’s semi-arid regions (e.g. Boyd, 1990; Cupper, 2005; Luly, 1993; Martin, 1973; Singh and Luly, 1991; Thomas et al., 2001; Table 1, Figure 1), but the records are few and their age models are for the most part dependent on scarce organic remains for radiocarbon dating. In light of these limitations, researchers have turned to novel archives for the reconstruction of the past environments of Australia’s arid regions, including isotopic analysis of stick-nest rat middens (Allen et al., 2000; McCarthy and Head, 2001; Pearson, 1999; Pearson and Dodson, 1993) and emu eggs (Johnson et al., 1999). Although the isotopic analysis of speleothems (secondary calcite deposits including stalagmites, stalactites and flowstones), has long been used in the field of palaeoclimatology (Fairchild and Baker, 2012), speleothems have only rarely been used for the examination of entrapped fossil material. Despite its limited use (McGarry and Caseldine, 2004), speleothem palynology is a promising technique, in part due to the potential for dating speleothems precisely and accurately using U series decay, and also because of the geographic distribution of caves in diverse ecosystems on all points of the globe, including regions where ‘conventional’ wetland deposits may be unavailable (McGarry and Caseldine, 2004; Sniderman et al., 2016, 2019).

Table 1.

Sites of Holocene palynological studies from arid and semi-arid Australia.

Study	Site
Present study	Webbs Cave, Nullarbor Plain
Boyd (1990)	Dalhousie Springs
Cupper (2005)	Warrananga Playa and Pine Camp Playa, Darling Anabranch
Luly (1993)	Lake Tyrrell
Luly (2001)	Lake Frome and Lake Eyre
Martin (1973)	Madura Cave, Norina Rock Shelter and N145 Rock Shelter, Nullarbor Plain
Singh (1981)	Lake Frome
Singh and Luly (1991)	Lake Frome
Thomas et al. (2001)	Salt Lake, Little Desert National Park

Figure 1.

Holocene palynological sites from the Australian arid and semi-arid zones. Map developed from GPS coordinates using QGIS version 3.4.2.

In this study, we document late-Holocene vegetation change on the Nullarbor Plain, a large (200,000 km²), semi-arid limestone plateau in south-central Australia, which provides few opportunities for wetland pollen preservation. We recovered fossil pollen from two stalagmites that grew near the entrance to Webbs Cave, a shallow cave typical of the hundreds developed in the Miocene Nullarbor Limestone (Webb and James, 2006). The aims of the study were twofold: to identify changes in vegetation occurring on the Nullarbor Plain during the late Holocene, and to explore the utility of speleothem palynology as an alternative to lacustrine pollen records in arid areas.

Study area and study site

Webbs Cave (127°49′4″E, 31°46′14″S) is located on Mundrabilla Station, in the Nullarbor region of southeastern Western Australia (Figure 1). The cave contains large numbers of calcite speleothems, which grew primarily during the Pliocene (Sniderman et al., 2016; Woodhead et al., 2012, 2019), and a small number of undated, but presumably late Quaternary halite, gypsum and evaporatively precipitated calcite stalagmites (Goede et al., 1992; Webb and James, 2006). Two evaporative calcite stalagmites, designated WC38 and WC39, were collected for this study.

The marine Nullarbor limestone was deposited during successive periods of inundation, beginning in the mid to late Eocene (Webb and James, 2006). Some time after c. 14 million years ago the plain was uplifted, and the regional climate more or less simultaneously underwent aridification (Crisp and Cook, 2007; Webb and James, 2006). This history of marine sedimentation, together with the effects of aridity, has led to the development of shallow red calcareous soils with an average depth of less than 1 metre (Gillieson et al., 1996; Webb and James, 2006). Episodic erosion has caused many parts of the plain to be stripped of soil, leaving bare, stony, pedogenic, calcrete pavements (Gillieson et al., 1996).

Webbs Cave is located at the ecotone between the coastal Hampton Bioregion, characterized by low woodlands dominated by species of mallee Eucalyptus, and the inland Nullarbor Bioregion, characterized by extensive treeless saltbush shrubland (Department of the Environment and Energy, 2017). Today, the region surrounding Webbs Cave is dominated by a community of saltbush (Atriplex spp.) and bluebush (Maireana spp.) in which low woodland accounts for less than 10% of foliage cover (Beard et al., 2013). In favourable conditions, saltbush is accompanied by the grasses, Austrostipa nitida, A. eremophila and Rytidosperma caespitosum (Poaceae), the annual herbs Zygophyllum ovatum (Zygophyllaceae), Lepidium oxytrichum, L. rotundum (Brassicaceae), and the white everlasting daisy, Rhodanthe floribunda (Asteraceae). Dominant tree-shrub associations include Acacia papyrocarpa (Mimosoideae: Fabaceae), Myoporum platycarpum (Myoporeae: Scrophylariaceae), Eucalyptus oleosa (Myrtaceae), Pittosporum angustifolium (Pittosporaceae) and Eremophila spp. (Myoporeae: Scrophulariaceae) (Beard et al., 2013). The predominant control on species distribution on the Nullarbor Plain appears to be soil chemistry. Many species are restricted to the siliceous sands on the eastern and western margins of the Nullarbor as they are unable to tolerate the alkalinity of the loam soils overlying limestone. Myrtaceous trees, however, appear to be less sensitive to soil chemistry than to water availability. The gaps in Eucalyptus distribution made apparent by Parsons’ (1970) survey coincide with aridity, and not with soil alkalinity.

There is evidence of human impact on the landscape both prior and subsequent to European colonization. Koonalda cave, on the eastern Nullarbor Plain, retains Aboriginal rock art as well as evidence of extensive mining of flint nodules (Wright, 1971). In the more recent Holocene, debate exists around the extent of human impact on the Nullarbor Plain (see contrasting theories presented by Martin et al., 1973; Williams et al., 2015). Written records provide accounts of extensive agricultural use of the region by European settlers. Mundrabilla Station has been grazed by sheep since 1876 (Waddell et al., 2010), and more broadly across the Nullarbor Plain, Santalum spp. were harvested extensively, both for timber and for incense (Brown, 1919, as cited in Parsons, 1970).

Methods

Two stalagmites were collected from the same chamber in Webbs Cave, within 20 metres of the entrance (Figure 2). WC38 was actively growing at the time of collection, but WC39 was collected lying on its side on the cave floor.

Figure 2.

Model of the cave site, constructed by J. Hellstrom, using Photoscan software. There are two entrances to the cave. The stalagmites are positioned 18.9 metres from the western entrance, 9.8 metres below the surface, and approximately 1 metre apart. WC38 was actively growing at the time of collection, so its position is certain. WC39, by contrast, was found on its side, so it is not known whether this was the position of the stalagmite as it grew. Site of collection is denoted by the circular marker.

U-Th dating

Uranium (U) and Thorium (Th) were extracted from small pieces (0.05–0.13 g) of stalagmite that had been cut from clean bands along the growth axis using a dental drill, and their isotopic ratios were determined using a Nu Instruments Multi-collector inductively coupled plasma mass spectrometer (MC-ICP-MS) at the University of Melbourne following Hellstrom (2003) and Drysdale et al. (2012). Age–depth models were developed using a finite positive growth rate model (Hendy et al., 2012; Scholz et al., 2012), corrected for detrital Th using the method described in Hellstrom (2006).

Stalagmite preparation and description

Each stalagmite was sampled by defining contiguous samples of roughly equal thickness (20 mm for WC39 and 5 mm for WC38) along the growth axis, which were cut using a circular saw for large samples, and a band saw for finer cuts. Once cut, the numbered samples were weighed and photographed, and rinsed with distilled water several times. Pollen was extracted from each sample using standard palynological methods (Bennett and Willis, 2001; Moore et al., 1991; Traverse, 2007), modified for speleothems (Sniderman et al., 2016). Individual samples were first cleaned in distilled water, and surface etched with dilute HCl followed by several water rinses. They were then spiked with one Lycopodium tablet each (containing 20,848 spores per tablet, sourced from Lund University), and treated with concentrated HCl, 10% KOH solution, acetolysis and concentrated HF (Supplemental Material, available online) to remove the majority of organic and mineral content. The acid- and alkali-resistant residues were mounted in glycerol on permanent glass slides.

Pollen analysis

Pollen was counted at ×400 magnification with the aid of PolyCounter version 3.1.6. (Nakagawa, 2013). A pollen sum exceeding 200 grains was achieved for all but six samples, in which 180 was the lowest number of grains counted. To aid in identification of pollen grains, a list of species currently recorded in the Nullarbor and Hampton bioregions was downloaded from the Atlas of Living Australia (ALA, 2016), and used to inform the development of a reference collection. Genera not represented well in the palynological literature and online pollen databases (e.g. the Australasian Pollen and Spore Atlas (Haberle et al., 2007), available at http://apsa.anu.edu.au) were targeted for collection from the University of Melbourne Herbarium (MELU). To prepare reference materials, pollen-laden anthers were removed from identified herbarium specimens under a dissecting microscope. Pollen was processed as above (excluding the application of HCl and HF), and mounted on slides for examination.

Data analysis

Pollen diagrams were generated using Psimpoll version 4.27 (Bennett, 2009), with counts expressed as percentages of the pollen sum. The pollen sum excludes unidentified pollen grains, as well as three insect-pollinated types, Nicotiana, cf. Halgania and cf. Fabaceae, that contribute to extreme, single-sample peaks (Faegri and Iversen, 1989). Zonation of pollen diagrams was carried out in Psimpoll, based on the agglomerative clustering method, CONISS. The number of statistically significant zones was evaluated using a broken stick model (Bennett, 2009). Species interactions were investigated using a stratigraphically unconstrained Principal Components Analysis (PCA). PCA was performed on Hellinger-transformed percentage data (Legendre and Gallagher, 2001), using R version 3.3.2 (R Foundation for Statistical Computing, 2016). Co-ordinates were extracted from R and plotted using Graphpad Prism version 6.0c. Charcoal particles, defined as opaque, angular objects exceeding 10 µm, were counted and were used to infer changing intensity and frequency of fires. Both charcoal flux and charcoal:pollen ratios were calculated.

Results

Chronology

As estimated by their U-Th age models (Figure 3), the younger stalagmite, WC38, began growing approximately 435 years ago (2σ uncertainty = 253–790) and was still actively growing at the time of collection. The older stalagmite, WC39, grew continuously from 2362 (2σ uncertainty = 1867–2830) to 571 years ago (2σ uncertainty = 0–867). The age model for WC39 is based on 20 U-Th determinations, and the age model for WC38 is based on 10 (two additional samples at 8.8 cm from the tip of WC38, and 6.5 cm from the tip of WC39 were discarded as outliers, after further sampling). The age models take into account both the depth uncertainties and 2σ age uncertainties of each individual age determination, and the constraint provided by superposition of the samples.

Figure 3.

Monte Carlo U-Th age-depth model for stalagmites: (a) WC38 and (b) WC39.

Model uncertainty is high, particularly at the tip of the older stalagmite. This uncertainty is largely a function of the high detrital component of the stalagmites. The U-Th dating technique rests on the assumption that all measured Th is produced in situ by decay of U, and that no initial Th was present at the time of crystallization of the calcite. Contamination of the stalagmite with mineral dust containing detrital Th, therefore, increases the age uncertainty of the model and requires a correction for detrital Th (Hellstrom, 2006). While a stalagmite that is rich in detrital material does present some challenges for U-Th dating, the trade-off would have been the selection of cleaner stalagmites with lower pollen counts. Ages are reported here as best estimates, with 2σ age uncertainties.

Fossil pollen and charcoal record

The combined pollen records for the two stalagmites can be divided into three statistically significant zones. The most significant zonal boundary (Figure 4) separates the two stalagmites, and the second most significant boundary separates the uppermost sample of the younger stalagmite, representing the past 30 years, from all earlier samples. Zones A through C are described below.

Figure 4.

Pollen percentage for both stalagmites combined. Varying widths of bars represent duration in years of each contiguous sample, according to the U-Th age model (see age scale on y axis). X axis shows percentage of pollen sum; major ticks = 10%, minor ticks = 5%. Pollen and charcoal fluxes are given, and since they strongly co-vary, charcoal: pollen ratio is also provided. The three zones, A, B and C, were delineated using the constrained incremental sum of squares (CONISS) agglomerative clustering method, with the most significant split occurring between the two stalagmites. The second most significant split separates the most recent sample from the remainder of the WC38 samples. Zonation excludes the insect-pollinated taxa, Fabaceae, cf. Nicotiana and cf. Halgania, so as not to obscure the environmental signal. Dashed lines indicate the boundaries of the five phases in the discussion. Fifty-eight pollen and spore types are recorded, which are separated into five categories, broadly reflecting plant functional types, for summary purposes.

Zone A (2362 (1870–2830) to 570 (0–860) years ago) encompasses the entirety of the older stalagmite. The record shows evidence of long-term environmental change (see the c. 1000-year long trend in undifferentiated Eucalyptus from 1900 to 800 years ago), but these trends are occasionally interrupted by episodic, stochastic deposition of idiosyncratic pollen assemblages (see for example, high percentages of cf. Fabaceae, cf. Halgania and Nicotiana grains in the sample corresponding to 1240 years ago).

The record begins at 2362 years ago with an anomalous sample, distinguished by its distinct composition of Chenopodiaceae pollen sub-types. Cf. Maireana is relatively low, while cf. Atriplex is more abundant, and Poaceae is entirely absent. The arid-adapted Eremophila reaches its greatest representation in this sample. The period from 2200 to 1400 years ago represents the period of greatest stability over the 2360-year record. Cf. Maireana (25% of all pollen) dominates the chenopod assemblage, and Myrtaceous trees begin their long-term trend of increasing abundance. At 1400 years ago, cf. Maireana abruptly declines, cf. Atriplex increases modestly, and there is a sustained increase in the overall diversity of non-Myrtaceous trees and shrubs. Myrtaceous pollen continues to rise, interrupted by a drop at 1000 years ago, which may partly be a statistical response to a large peak of monosulcate monocot pollen (although Asteraceae pollen similarly declines at around 1000 years ago but recovers more slowly). The drop is consistent through all Myrtaceous tree types for this single sample, but the upwards trend is resumed in the following two samples, culminating at around 800 years ago. After 800 years ago, until the cessation of speleothem deposition, Asteraceae and cf. Maireana recover at the expense of Myrtaceae types.

Zone B (435 (254–790) to 30 (0–198) years ago) includes the interval preceding European colonization of Australia, up until 30 years ago. It is characterized by rising values of cf. Maireana, declining values of cf. Atriplex, generally low but stable counts of Myrtaceous pollen and greater grass coverage. It includes an initial, short period of high pollen flux, and the largest charcoal peak of the record. The highest percentages of Cupressaceae (cf. Callitris) grains are found in the upper part of this zone, within the past ~150 years.

Zone C (30 (0–198) years ago to present) comprises a single sample, marked by an abundance of Chenopodiaceae grains, which make up more than 60% of all grains counted. The zone shows the most significant change in the chenopod pollen assemblage of any time in the preceding 2200 years. Although they make up only a small percentage of the total, grains of the fire-ephemeral Gyrostemonaceae (Baker et al., 2005) are seen in their highest abundance of the whole study period.

Principal components analysis

Broken stick modelling suggests there are six significant axes for PCA, but the first two, presented at Figure 5, account for 30% of the total variance in the data. PCA suggests that the species mix of chenopods, rather than major adjustments to vegetation structure or life form, drive the separation of zones. This is most readily demonstrated by the distance between cf. Maireana and cf. Atriplex on both axes, and their apparent influence on the separation of Zone C in particular. Zones A and B are similarly affected by the species mix of chenopods, but also show a weak separation between woody (Zone B) and herbaceous (Zone A) taxa.

Figure 5.

Principal components analysis for species (a) and sites (b), using Axis 1 (20.6% of all variation) and Axis 2 (9.7% of all variation). Zones correspond to those marked in Figure 4.

Pollen and charcoal concentration and accumulation

Pollen concentration for the younger stalagmite was much lower than for the older stalagmite, with an average of 1640 grains per gram of calcite in the latter and 784 grains per gram of calcite in the former. However, taking the rate of growth of each stalagmite into account, the pollen flux (pollen grains cm^-2 year^-1) was somewhat higher in the younger stalagmite (Figure 4). The pollen and charcoal flux to the younger stalagmite was also more variable, with two large injections at 40 and 440 years ago.

Discussion

Interpretation of speleothem-based palynological archives

Pollination syndromes influence both the spatial range of pollen dispersal, and the relative abundance of pollen types in a sedimentary record (Hunt and Rushworth, 2005). Anemophilous (wind-pollinated) taxa are typically both high producers and long-distance dispersers of pollen, and as such, they are often over-represented in the fossil record relative to their contribution to the vegetation (Faegri and Iversen, 1989; Jacobson and Bradshaw, 1981; Pickett et al., 2004), particularly in treeless environments where wind-borne pollen may be transported over large distances (Dimbleby, 1962). However, it is not clear that this over-representation of anemophilous pollen types occurs within cave and speleothem deposits that are to a greater or lesser degree protected from the aerially transported pollen rain. Studies of cave sites (Burney and Burney, 1993; Hunt and Rushworth, 2005) suggest that both the pollination syndrome of plants, and the visitation of the cave by animals are important predictors of pollen assemblages found within the cave environment. In caves in arid climates, pollen deposition by animals such as insects may be important, as arid areas often host a higher proportion of animal-pollinated taxa (Horowitz, 1992, though note the dominance of wind-pollinated Chenopodiaceae and Poaceae in the Australian arid region) and insects are likely to visit accessible stalagmites in search of water. Faegri and Iversen (1989) note that, for zoophilous taxa with specialized pollination syndromes, pollen deposition is often unpredictable. The anomalous sample at 1200 years ago, containing high percentages of cf. Fabaceae, cf. Halgania and Nicotiana may be an example of this phenomenon. In addition, slow rates of speleothem growth (at least relative to wetland sediment accumulation rates), together with the generally low rates of pollen delivery, can serve to amplify the impact of stochastic, insect-borne pollen deposition (Burney and Burney, 1993). In reconstructions based on speleothem data, the pollen sum is typically lower than for wetland records (Fairchild and Baker, 2012; Horowitz, 1992; McGarry and Caseldine, 2004). As a result, insect-borne pollen influxes, if included in the pollen sum, have the potential to obscure the environmental signal (Pickett et al., 2004). Thus, the peaks of cf. Fabaceae, cf. Halgania and Nicotiana at 1240 years ago are interpreted here as indicators of the local presence of those taxa, but their high pollen percentages are unlikely to reflect any change in their ecological importance on a regional scale.

Overlaid on this insect-borne pollen signal are the wind-pollinated and long-distance dispersed animal-pollinated taxa that represent a combination of both the local and regional vegetation. The stalagmites analysed here grew relatively close to the cave entrance, where they were likely exposed to the airborne pollen ‘rain’. The dominance of the pollen records by ecologically dominant, wind-pollinated taxa such as Chenopodiaceae and Poaceae suggests that, in overall terms, this record is consistent with the regional-scale pollen assemblage that would be captured in more conventional sediments such as lakes and wetlands.

Compared with the spectrum of wetland/lacustrine basins, speleothem pollen records are perhaps most analogous to the pollen profiles of small, ‘closed-canopy’ forest hollows and peat bogs, in that all have very small pollen catchments and undergo limited stratigraphic mixing of pollen once deposited (Bradshaw, 1988; Jacobson and Bradshaw, 1981). As the model of catchment size proposed by Jacobson and Bradshaw (1981) is influenced primarily by air flow, some consideration should be given to the potential differential sorting of pollen between the two stalagmites used in this study. The causes of the difference in pollen composition between the two stalagmites (as measured by the high level zone boundary placed between them) is unknown, but may reflect either a real change in vegetation outside the cave in the time between deposition of the older and younger stalagmite, or slightly different taphonomic biases between the stalagmites. The precise position of the two stalagmites relative to the cave interior morphometry is unknown, but differences in pollen composition between the two stalagmites may partly reflect differential pollen sorting, driven by localized patterns of airflow within the cave; if so, these sorting patterns must have remained stable over timescales of centuries.

Poor taxonomic resolution is a persistent challenge for palynological analysis, particularly in environments where ecologically important changes in vegetation composition occur between species that cannot be distinguished palynologically (Birks and Birks, 2000; Newsome, 1999). This is an important question in this study, because we distinguish between two Chenopodiaceae morphotypes, cf. Maireana and cf. Atriplex, within a family that, based on fossil pollen, is not usually identified below family level. The assignment of these two types at the generic level is not based on pollen morphology alone – other members of the Chenopodiaceae share these morphotypes – but modern chenopod communities on the southern Nullarbor Plain are dominated by Maireana and Atriplex, which, based on our arid-zone reference material, can be distinguished from each other on the basis of exine sculpture. Separation of the major morphological types allows a more nuanced discussion of environmental processes, given that the divergent responses of Maireana and Atriplex species to environmental conditions such as moisture availability and fire are known (Beadle, 1981; Gillieson et al., 1996). The depth of this discussion is limited, however, because ecological responses commonly occur at the species, rather than genus, level, and much of the literature in Australia is limited to the ecology of pastorally important species. Given their apparent sensitivity to environmental change, and their potential to be used as arid-zone palaeoecological indicators, further study of both the pollen morphology and ecology of Australian arid-zone chenopods is warranted.

Late-Holocene environments on the Nullarbor Plain

Chenopods have dominated the landscape of the Nullarbor Plain for the majority of the past 2360 years. Changes in both their abundance relative to other families (Zone A-B boundary) and between Chenopodiaceae pollen types (Zone B-C boundary) are largely responsible for the partitioning of the assemblage into three zones. Temporal trends in the pollen assemblage are compared with both instrumental and reconstructed climate and human influences, giving rise to five distinct phases for discussion.

Stable cf. Maireana phase (2360–1420 years ago)

This phase is characterized by relatively stable dominance of cf. Maireana and is interpreted as a continuation of the late-Holocene aridity that is thought to have begun at approximately 5500 years ago and reached a maximum at approximately 3000 years ago (Petherick et al., 2013). Since then, conditions have remained arid, though variability of rainfall has increased towards modern times (Petherick et al., 2013). Following the single anomalous sample at 2360 years ago, there is a c. 700-year period of relative stability of this cf. Maireana-dominated assemblage, throughout which Myrtaceae counts are consistently low. According to Beadle (1981), low annual rainfall in arid Australia has the effect of both reducing the total number of species, and eliminating woody species. Both of these conditions are satisfied during this phase; three of the five samples that make up this dry stage share the lowest species richness (S = 19) observed for any of the 50 samples counted.

Transition to grass and herb-dominance, coupled with increasing Myrtaceae (1420–960 years ago)

With the exception of the anomalous insect-pollinated sample at 1200 years ago, this appears to be a transitional phase, with a shift away from chenopod dominance towards a more diverse grass and herb-dominant state that, based on Beadle’s (1981) descriptions of arid Australian community dynamics, may be a response to disturbance or rain. The arid zone has the highest proportion of annuals of any climatic zone in Australia, many of which possess specific adaptations to arid environments, including seed longevity and the ability initiate seedling development upon receipt of multiple correlating environmental signals (Noy-Meir, 1973). These features allow them to readily colonize under favourable environmental conditions, with population doubling times much shorter than those of trees (Prentice, 1985). The beginning of this period is marked by a small peak in charcoal, so it may be that the initial expansion of forbs and grasses in this phase resulted from fire disturbance.

Myrtaceous peak: vegetation response to centennial-scale climate variability (960–570 years ago)

Myrtaceous trees increase in importance within a well-defined interval from 960 to 800 years ago, accompanied by a rise in other trees, Acacia and Casuarinaceae, and the woody shrub, Westringia. Given that low woodland communities occur only on the wettest, southern margin of the Nullarbor, this distinct vegetation signal implies increasing water availability, and is consistent with the observations of Cohen et al. (2012), who documented a water level highstand at Lake Callabonna, in the Lake Frome complex (Figure 1), which they interpreted as evidence of an exceptionally wet Medieval Climate Anomaly (MCA). In Western Tasmania, Fletcher et al. (2018) found a discrete episode of high fire activity at c. 870 years ago, followed by distinct changes in vegetation that they interpreted as evidence of dry conditions associated with the positive phase of the Southern Annular Mode (see also Abram et al., 2014; Moreno et al., 2014). The anti-correlation between an episode of relatively high rainfall on the Nullarbor, and of high fire activity in western Tasmania is consistent with spatial patterns of hydroclimatic response to the Southern Annular Mode. Positive phases of the Southern Annular Mode result in anomalously warm, dry conditions in southern South America, western Tasmania and the South Island of New Zealand, which coincide with anomalously wet conditions over mainland Australia (Gillett et al., 2006), including on the Nullarbor Plain (Figure 6).

Figure 6.

Southern Annular Mode reconstructions of (a) Abram et al. (2014), (b) Moreno et al. (2014), (c) SAM-related peaks in charcoal in Tasmania (Fletcher et al., 2018), are compared with (d) Webbs Cave Myrtaceous trees. Higher values of arboreal pollen are interpreted as an increase in precipitation on the southern Nullarbor.

Late-Holocene cooling, and European settlement (440–30 years ago)

The penultimate phase comprises all but the most recent sample of the younger stalagmite, WC38. The most significant change in vegetation during this phase is the shift in the composition of chenopods relative to those of the older stalagmite, WC39. It is unclear whether this shift simply reflects differential pollen delivery onto the two stalagmites, or is a true response to changed environmental conditions. A possible explanation of the change in composition between the two stalagmites – as reflected in the high level zonation boundary between them – could be the onset of the ‘Little Ice Age’ (715–150 years ago) (PAGES 2k Consortium, 2013). The timing of the ‘Little Ice Age’ spans much of the period that is missing from the present record and in the absence of overlapping records, little can be inferred about the response of Nullarbor Plain vegetation to this event.

This phase also spans the timing of European colonization of Australia. Elsewhere in the Australian arid zone, large shifts in vegetation in response to European colonization have been documented (Cupper, 2005; Luly, 1993; Thomas et al., 2001). If this record is indeed faithfully representing regional vegetation change, the Nullarbor Plain appears to have escaped these environmental effects, at least until approximately 50 years ago when Poaceae abruptly declines. Though other researchers have pointed to an increase in Poaceae as evidence of European colonization (Cupper, 2005; Dodson and Lu, 2000; Luly, 1993; Thomas et al., 2001), it is equally plausible that pastoral practices have caused a decline in Poaceae on the Nullarbor. While elsewhere in Australia introduced pasture grasses have been extensively planted, this practice was not adopted in the arid rangelands of the Nullarbor, where preferential grazing of native grasses (Beadle, 1981) may have caused a rapid decline in Poaceae. Associated increases in the representation of Stylidium and Brassicaceae-Westringieae pollen could tentatively be interpreted as responses, sometimes short lived, to altered management regimes, linked with loss of Indigenous management and/or changes in grazing intensity and dynamics.

Post-fire restructuring of vegetation (30 years ago to present)

Comparison of the charcoal record with historically documented fire records provides a way to assess whether or not our charcoal data accurately signal the timing of past fires in the region near Webbs Cave. Here, relatively large, ‘macrocharcoal’ fragments > 125 µm in diameter, which can provide insight into very local fires (Gavin et al., 2003; Whitlock and Larsen, 2001), were not analysed independently of smaller fragments, but the abrupt peak in the charcoal: pollen ratio that occurs approximately 40 years ago seems to correspond with a major fire that occurred on Mundrabilla Station in 1973 (Gillieson et al., 1996). Based on this correlation, earlier peaks, including that preceding the transition in vegetation structure at 1420 years ago, might also be interpreted as evidence of historical fires, even if their spatial extent is unknown (Patterson et al., 1987; Whitlock and Larsen, 2001).

Despite increasing rainfall in the late Holocene, fires on the Nullarbor may have been mitigated by the dominance of the fire-retarding Chenopodiaceae (Cupper, 2005; Gillieson et al., 1996). Cupper (2005) attributes the 1974 and 1975 fires in the Darling Anabranch, southeastern Australia, to enhanced fuel loads following high ENSO-driven rainfall. It would be tempting to apply the same argument here, were it not for the dramatic decline in Poaceae grains preceding the charcoal peak. Although it is unusual for chenopod shrublands to ignite without extensive grass cover, the vegetation response inferred from the pollen record is consistent with what is known of the ecology of the site: With the exception of the fire-sensitive A. nummularia (Gillieson et al., 1996), Maireana spp. recover from disturbance events more slowly than Atriplex spp., particularly where fire-affected areas are also subject to grazing (Beadle, 1981; Gillieson et al., 1996). This may explain the significant peak in non-Maireana grains in the interval following the most recent charcoal peak. Historically, the use of sedimentary charcoal peaks to infer palaeofires has been treated with some scepticism, partly because although fire events are discrete, the resulting charcoal peaks in the palaeo record are spread over a longer interval (Patterson et al., 1987). This record, with its high temporal resolution (~5 years, for the sample in question) reduces the age uncertainty around the timing of fire events, and in the case of the most recent 1973 fire, documents not only the timing of the fire event, but also the conditions preceding the event, and the vegetation response to disturbance.

Conclusion

The study demonstrates the potential of speleothem palynology to elucidate past environmental change in arid regions. The record contributes a high-resolution reconstruction of Holocene palaeovegetation in arid Australia, one of very few from Western Australia. The speleothem palynoflora appears to be susceptible to some stochastic pollen delivery mechanisms, and though this is not a problem that is unique to speleothems, fewer taphonomic models have been developed to aid in the interpretation of speleothem-based records. In future work, thorough floristic surveys of the local and regional area, as well as investigation of the use of caves and speleothems by potential animal pollen vectors, would assist the interpretation of the pollen record.

Over the past 2360 years, environmental changes have rarely been sufficient to cause large-scale change to the composition of vegetation on the Nullarbor Plain. Responses to disturbance are most commonly seen in changes in floristic composition of the dominant Chenopodiaceae shrubland, and ephemeral changes in the abundance of herbs. In contrast to other studies from arid Australia, no unambiguous evidence is observed of changes resulting from European colonization, but a single sample in the most recent period (the past 30 years) shows that a change may now be underway. The increase in pollen of Myrtaceae and other woody taxa in the record supports findings from elsewhere in arid Australia of an exceptionally wet period between c. 1000 and 800 years ago, and although a shift in composition is seen around the proposed time of the ‘Little Ice Age’, it is unclear whether this reflects true vegetation change over time, or taphonomic filters differentially affecting pollen deposition between stalagmites. In future studies of this kind, the selection of multiple stalagmites that overlap in time will resolve some of the uncertainty around this question.

Supplemental Material

S1_Simplified_summary_of_pollen_extraction_method – Supplemental material for Late-Holocene environmental change on the Nullarbor Plain, southwest Australia, based on speleothem pollen records

Supplemental material, S1_Simplified_summary_of_pollen_extraction_method for Late-Holocene environmental change on the Nullarbor Plain, southwest Australia, based on speleothem pollen records by Kia A Matley, JM Kale Sniderman, Andrew N Drinnan and John C Hellstrom in The Holocene

Supplemental Material

S2_WC38_Age_model – Supplemental material for Late-Holocene environmental change on the Nullarbor Plain, southwest Australia, based on speleothem pollen records

Supplemental material, S2_WC38_Age_model for Late-Holocene environmental change on the Nullarbor Plain, southwest Australia, based on speleothem pollen records by Kia A Matley, JM Kale Sniderman, Andrew N Drinnan and John C Hellstrom in The Holocene

Supplemental Material

S3_WC39_Age_model – Supplemental material for Late-Holocene environmental change on the Nullarbor Plain, southwest Australia, based on speleothem pollen records

Supplemental material, S3_WC39_Age_model for Late-Holocene environmental change on the Nullarbor Plain, southwest Australia, based on speleothem pollen records by Kia A Matley, JM Kale Sniderman, Andrew N Drinnan and John C Hellstrom in The Holocene

Supplemental Material

S4_Webbs_Cave_pollen_percentages – Supplemental material for Late-Holocene environmental change on the Nullarbor Plain, southwest Australia, based on speleothem pollen records

Supplemental material, S4_Webbs_Cave_pollen_percentages for Late-Holocene environmental change on the Nullarbor Plain, southwest Australia, based on speleothem pollen records by Kia A Matley, JM Kale Sniderman, Andrew N Drinnan and John C Hellstrom in The Holocene

Supplemental Material

S5_Chenopodiaceae_pollen_identification_and_classification – Supplemental material for Late-Holocene environmental change on the Nullarbor Plain, southwest Australia, based on speleothem pollen records

Supplemental material, S5_Chenopodiaceae_pollen_identification_and_classification for Late-Holocene environmental change on the Nullarbor Plain, southwest Australia, based on speleothem pollen records by Kia A Matley, JM Kale Sniderman, Andrew N Drinnan and John C Hellstrom in The Holocene

Footnotes

Acknowledgements

The author is thankful to Michael-Shawn Fletcher for providing lab space for pollen counting. Serene Paul and Claire MacGregor assisted with U-Th extraction and dating. Jo Birch of the University of Melbourne Herbarium, and Janelle Stevenson of the Australian National University provided access to plant and pollen collections for the development of reference material. Libby Rumpff and Matt Cupper provided invaluable comments on an earlier version of this manuscript.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Funding support was received from The Botany Foundation.

ORCID iDs

Kia A Matley

JM Kale Sniderman

Supplemental material

Supplemental material for this article is available online.

References

Abram

Mulvaney

Vimeux

et al. (2014) Evolution of the Southern Annular Mode during the past millennium. Nature Climate Change 4: 564–569.

Allen

Head

Medlin

et al. (2000) Palaeo-ecology of the gap and Coturaundee Ranges, western New South Wales, using stick-nest rat (Leporillus spp.) (Muridae) middens. Austral Ecology 25: 333–343.

Atlas of Living Australia (ALA) (2016) Occurrence download. Available at: http://www.ala.org.au (accessed 13 May 2016).

Baker

Steadman

Plummer

et al. (2005) Seed dormancy and germination responses of nine Australian fire ephemerals. Plant and Soil 277: 345–358.

Beadle

NCW

(1981) The Vegetation of Australia. Cambridge: Cambridge University Press.

Beard

Beeston

Harvey

et al. (2013) The vegetation of Western Australia at the 1:3,000,000 scale. Explanatory memoir. Second edition. Conservation Science Western Australia 9: 1–152.

Bennett

(2009) Psimpoll: C Program for Plotting Pollen Diagrams and Analyzing Pollen Data (Version 4.27). Belfast: Queen’s University of Belfast.

Bennett

Willis

(2001) Pollen. In: Smol

Birks

HJB

Last

(eds) Tracking Environmental Change Using Lake Sediments: Terrestrial, Algal, and Siliceous Indicators, vol. 3. Dordrecht: Kluwer Academic Publishers, pp. 5–32.

Birks

HJB

(2000) Future uses of pollen analysis must include plant macrofossils. Journal of Biogeography 27: 31–35.

10.

Boyd

(1990) Quaternary pollen analysis in the arid zone of Australia: Dalhousie Springs, Central Australia. Review of Palaeobotany and Palynology 64: 331–341.

11.

Bradshaw

RHW

(1988) Spatially-precise studies of forest dynamics. In: Huntley

Webb

(eds) Vegetation History. Dordrecht: Kluwer Academic Publishers, pp. 725–751.

12.

Burney

(1993) Modern pollen deposition in cave sites: Experimental results from New York State. New Phytologist 124: 523–535.

13.

Cohen

Nanson

Jansen

et al. (2012) A pluvial episode identified in arid Australia during the Medieval Climatic Anomaly. Quaternary Science Reviews 56: 167–171.

14.

Crisp

Cook

(2007) A congruent molecular signature of vicariance across multiple plant lineages. Molecular Phylogenetics and Evolution 43: 1106–1117.

15.

Cupper

(2005) Last glacial to Holocene evolution of semi-arid rangelands in southeastern Australia. The Holocene 15: 541–553.

16.

Department of the Environment and Energy (2017) Interim biogeographic regionalisation for Australia. Available at: http://www.environment.gov.au/land/nrs/science/ibra/australias-bioregions-maps (accessed 21 May 2017).

17.

Dimbleby

(1962) The Development of British Heathlands and Their Soils (Oxford Forestry Memoirs No. 23). Oxford: Oxford University Press.

18.

Dixon

Tyler

Lorrey

et al. (2017) Low-resolution Australasian palaeoclimate records of the last 2000 years. Climate of the Past 13: 1403–1433.

19.

Dodson

(2000) A late Holocene vegetation and environment record from Byenup Lagoon, south-western Australia. Australian Geographer 31: 41–54.

20.

Drysdale

Paul

Hellstrom

et al. (2012) Precise microsampling of poorly laminated speleothems for U-series dating. Quaternary Geochronology 14: 38–47.

21.

Faegri

Iversen

(1989) Textbook of Pollen Analysis. New York: John Wiley.

22.

Fairchild

Baker

(2012) Speleothem Science: From Process to Past Environments. West Sussex: Wiley-Blackwell.

23.

Fletcher

M-S

Benson

Bowman

DMJS

et al. (2018) Centennial-scale trends in the Southern Annular Mode revealed by hemisphere-wide fire and hydroclimatic trends over the past 2400 years. Geology 46(4): 363–366.

24.

Gavin

Brubaker

Lertzman

(2003) An 1800-year record of the spatial and temporal distribution of fire from the west coast of Vancouver Island, Canada. Canadian Journal of Forest Research 33: 573–586.

25.

Gillett

Kell

Jones

(2006) Regional climate impacts of the Southern Annular Mode. Geophysical Research Letters 33: L23704.

26.

Gillieson

Wallbrink

Cochrane

(1996) Vegetation change, erosion risk and land management on the Nullarbor Plain, Australia. Environmental Geology 28: 145–153.

27.

Goede

Atkinson

Rowe

(1992) A giant Late Pleistocene halite speleothem from Webbs cave, Nullarbor Plain, Southeastern Western Australia. Helictite 30: 3–7.

28.

Haberle

Rowe

Hungerford

et al. (2007) The Australasian Pollen and Spore Atlas V1.0. Australian National University, Canberra. Available at: http://apsa.anu.edu.au/.

29.

Hellstrom

(2003) Rapid and accurate U/Th dating using parallel ion-counting multi-collector ICP-MS. Journal of Analytical Atomic Spectrometry 18: 1346–1351.

30.

Hellstrom

(2006) U–Th dating of speleothems with high initial ²³⁰Th using stratigraphical constraint. Quaternary Geochronology 1: 289–295.

31.

Hendy

Tomiak

Collins

et al. (2012) Assessing amino acid racemization variability in coral intra-crystalline protein for geochronological applications. Geochimica et Cosmochimica Acta 86: 338–353.

32.

Horowitz

(1992) Palynology of Arid Lands. Amsterdam: Elsevier.

33.

Hunt

Rushworth

(2005) Pollen taphonomy and airfall sedimentation in a tropical cave: The West Mouth of The Great Cave of Niah in Sarawak, Malaysian Borneo. Journal of Archaeological Science 32: 465–473.

34.

Jacobson

Bradshaw

(1981) The selection of sites for paleovegetational studies. Quaternary Research 16: 80–96.

35.

Johnson

Miller

Fogel

et al. (1999) 65,000 years of vegetation change in central Australia and the Australian summer monsoon. Science 284: 1150–1152.

36.

Legendre

Gallagher

(2001) Ecologically meaningful transformations for ordination of species data. Oecologia 129: 271–280.

37.

Luly

(1993) Holocene palaeoenvironments near Lake Tyrrell, semi-arid Northwestern Victoria, Australia. Journal of Biogeography 20: 587–598.

38.

Luly

(2001) On the equivocal fate of Late Pleistocene Callitris Vent. (Cupressaceae) woodlands in arid South Australia. Quaternary International 83–85: 155–168.

39.

McCarthy

Head

(2001) Holocene variability in semi-arid vegetation: New evidence from Leporillus middens from the Flinders Ranges, South Australia. The Holocene 11: 681–689.

40.

McGarry

Caseldine

(2004) Speleothem palynology: An undervalued tool in Quaternary studies. Quaternary Science Reviews 23: 2389–2404.

41.

Martin

(1973) Palynology and historical ecology of some cave excavations in the Australian Nullarbor. Australian Journal of Botany 21: 283–316.

42.

Moore

Webb

Collinson

(1991) Pollen Analysis. Oxford: Blackwell Scientific Publications.

43.

Moreno

Vilanova

Villa-Martínez

et al. (2014) Southern Annular Mode-like changes in Southwestern Patagonia at centennial timescales over the last three millennia. Nature Communications 5: 4375.

44.

Nakagawa

(2013) PolyCounter (Version 3.1.6). Available at: http://polsystems.rits-palaeo.com/.

45.

Newsome

(1999) Pollen-vegetation relationships in semi-arid southwestern Australia. Review of Palaeobotany and Palynology 106: 103–119.

46.

Noy-Meir

(1973) Desert ecosystems: Environment and producers. Annual Review of Ecology and Systematics 4: 25–51.

47.

PAGES 2k Consortium (2013) Continental-scale temperature variability during the past two millennia. Nature Geoscience 6: 339–346.

48.

Parsons

(1970) Mallee vegetation of the southern Nullarbor and Roe Plains, Australia. Transactions of the Royal Society of South Australia 94: 227–242.

49.

Patterson

Edwards

Maguire

(1987) Microscopic charcoal as a fossil indicator of fire. Quaternary Science Reviews 6: 3–23.

50.

Pearson

(1999) Late Holocene biological records from the middens of stick-nest rats in the central Australian arid zone. Quaternary International 59: 39–46.

51.

Pearson

Dodson

(1993) Stick-nest rat middens as sources of paleoecological data in Australian deserts. Quaternary Research 39: 347–354.

52.

Petherick

Bostock

Cohen

et al. (2013) Climatic records over the past 30 ka from temperate Australia – a synthesis from the Oz-INTIMATE workgroup. Quaternary Science Reviews 74: 58–77.

53.

Pickett

Harrison

Hope

et al. (2004) Pollen-based reconstructions of biome distributions for Australia, Southeast Asia and the Pacific (SEAPAC region) at 0, 6000 and 18,000 14C yr BP. Journal of Biogeography 31: 1381–1444.

54.

Prentice

(1985) Pollen representation, source area, and basin size: Toward a unified theory of pollen analysis. Quaternary Research 23: 76–86.

55.

R Foundation for Statistical Computing (2016) R: A Language and Environment for Statistical Computing. Available at: https://www.R-project.org (accessed 17 February 2017).

56.

Scholz

Hoffmann

Hellstrom

et al. (2012) A comparison of different methods for speleothem age modelling. Quaternary Geochronology 14: 94–104.

57.

Singh

(1981) Late quaternary pollen records and seasonal palaeoclimates of Lake Frome, South Australia. Hydrobiologia 81–82: 419–430.

58.

Singh

Luly

(1991) Changes in vegetation and seasonal climate since the last full glacial at Lake Frome, South Australia. Palaeogeography, Palaeoclimatology, Palaeoecology 84: 75–86.

59.

Sniderman

JMK

Hellstrom

Woodhead

et al. (2019) Vegetation and climate change in southwestern Australia during the Last Glacial Maximum. Geophysical Research Letters 46: 1709–1720.

60.

Sniderman

JMK

Woodhead

Hellstrom

et al. (2016) Pliocene reversal of late Neogene aridification. Proceedings of the National Academy of Sciences of the United States of America 113: 1999–2004.

61.

Thomas

Enright

Kenyon

(2001) The Holocene history of mediterranean-type plant communities, Little Desert National Park, Victoria, Australia. The Holocene 1: 691–697.

62.

Traverse

(2007) Paleopalynology. Dordrecht: Springer.

63.

Waddell

Gardner

Hennig

(2010) An Inventory and Condition Survey of the Western Australian Part of the Nullarbor Region. Technical Bulletin No. 97. Kensington: Department of Agriculture and Food, Western Australia.

64.

Webb

James

(2006) Karst evolution of the Nullarbor Plain, Australia. In: Harmon

Wicks

(eds) Perspectives on Karst Geomorphology, Hydrology, and Geochemistry – A Tribute to Derek C. Ford and William B. White. Boulder, CO: The Geological Society of America (Special Paper 404), pp. 56–79.

65.

Whitlock

Larsen

(2001) Charcoal as a fire proxy. In: Smol

Birks

HJB

Last

(eds) Tracking Environmental Change Using Lake Sediments: Terrestrial, Algal, and Siliceous Indicators, vol. 3. Dordrecht: Kluwer Academic Publishers, pp. 75–98.

66.

Williams

Ulm

Turney

CSM

et al. (2015) Holocene demographic changes and the emergence of complex societies in prehistoric Australia. PLoS ONE 10(6): e0128661.

67.

Woodhead

Hellstrom

Pickering

et al. (2012) U and Pb variability in older speleothems and strategies for their chronology. Quaternary Geochronology 14: 105–113.

68.

Woodhead

Sniderman

JMK

Hellstrom

et al. (2019) The antiquity of Nullarbor speleothems and implications for karst palaeoclimate archives. Scientific Reports 9: 603.

69.

Wright

RVS

(1971) Archaeology of Koonalda cave. In: Mulvaney

Golson

(eds) Aboriginal Man and Environment in Australia. Canberra: Australian Institute of Aboriginal Studies, pp. 105–113.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.01 MB

0.03 MB

0.02 MB

0.27 MB