Multiple lines of evidence infer centurial-scale habitat change and resilience in a threatened plant species at Mount Dangar, Hunter Valley, New South Wales

Keywords: Acacia, centurial-scale habitat change, climate, dendrochronology, fire, radiocarbon dating, range-restricted endemic, resilience, threatened plants.

Management of threatened species occasionally requires the initiation of actions and interventions before a complete understanding of life history is gained, and this precautionary approach may avert species extinction (Scheele et al. 2018). In some cases, prompt intervention is warranted (e.g. Mackenzie and Keith 2009; Cuneo et al. 2018), whereas in others, the magnitude of the ecological knowledge gap may be too great to justify action. For many taxa, there is little knowledge available to inform how, for example, intervention in germination triggers or pollinator networks may benefit the species (Williams 2006; Silcock et al. 2015; Scheele et al. 2018), or whether modification to habitat is required to improve above-ground presence and dispersal mechanisms (Török et al. 2020). But for species where available life-history data span only a fraction of their expected life span (which may be 100 years or more), there is a need for caution before implementing management practices. Many plant species operate on long cycles of boom or bust, relying on rare environmental events (such as fire, floods or high rainfall years) to trigger episodic germination (Kenny et al. 2017; Zimmer et al. 2017; Elliott et al. 2019; Miller et al. 2019). Long periods of dormancy are typical in these species, often involving persistent seed banks held in the canopy or the soil for decades or centuries (Long et al. 2015). Species occurring in landscapes that rarely experience appropriate germination triggers can subsequently be the subject of inappropriate management decisions (Scheele et al. 2018), which may be unnecessary and compromise the allocation of scant conservation resources (Burgman 2002; Bateman et al. 2017).

Improved understanding of how major environmental disturbance events interact with a species’ life history will ultimately benefit its conservation (Scheele et al. 2018; Elliott et al. 2019). Environmental drivers for ecological change, such as the incidence and severity of fire and drought and the variable time scales over which they may operate, can dictate extinction and new recruitment in plant populations (Keith 1996; Zhang et al. 2019). Fire is a well researched driver of ecosystem regulation and rejuvenation (McLauchlan et al. 2020), and when infrequent fire events are coupled with stresses imposed under drought conditions, plant species may disappear from the above-ground flora and habitats may be substantially modified (e.g. Tozer et al. 2021). Conversely, with the appropriate environmental triggers, new generations of these species may reappear after seeds are released from dormancy, often in areas from which they were previously unknown (e.g. Bell and Holzinger 2015), and there are numerous examples in the literature where species have re-appeared after long absences following a major fire event (e.g. Miles and Cameron 2007; Cheal 2010; Binns 2013). Some authors also stress that persistence of rare species in certain habitats is entirely dependent on the occurrence of these rare disturbance events (e.g. Elliott et al. 2019; Miller et al. 2019); yet, such a phenomenon is rarely acknowledged in the context of threatened species management.

For one site in the Hunter Valley region of New South Wales, we suspect that the interplay of centurial-scale fire, drought and rainfall is a major driver of ecological change in the above-ground presence of common and rare plant species. Mount Dangar, lying near the confluence of the Hunter and Goulburn rivers, is a dominant feature in the regional landscape. Like other peaks of unusual geology supporting rare species (e.g. Mount Buffalo in Victoria: Calder and Calder 1998; Walsh 1998; banded iron formations in Western Australia: Miller et al. 2019), several range-restricted plant species occur on or in the immediate vicinity of Mount Dangar. In recent years, the management of each of these has become a focus of the New South Wales (NSW) Government through their Saving our Species program, including the critically endangered and range-restricted endemic tree Acacia dangarensis. Although contemporary survey data have shown little changes in the distribution and abundance of this species over the short term (S. Bell and P. Lamrock, unpubl. data), gaining a more comprehensive understanding of the ecology of this long-lived species requires examination of data from multiple sources across longer time frames to guide future management.

This paper uses written (explorer and surveyor journals) and oral (local residents) evidence, supported by herbarium, radiocarbon (¹⁴C) and dendrochronological analyses, to assess population fluctuations of A. dangarensis because they relate to historical fire disturbance and climatic variability affecting Mount Dangar over the past 200 years. We acknowledge at the outset that combining conventional and unconventional data sources such as these may result in intuitive rather than analytical findings; however, this is often necessary to further understanding of plant species and ecosystem changes operating over century-scale life cycles (Lunt 2002, and other papers in this volume; Caseldine and Turney 2010; Hédl et al. 2021). On the basis of our findings, the current-day population of A. dangarensis is placed into ecological context, and we provide guidance for future management of this species on Mount Dangar, and potentially also for other long-lived species that occur in similar landforms or in areas with similar disturbance regimes.

Mount Dangar is a unique natural feature within Goulburn River National Park, located in the upper Hunter Valley of New South Wales, 150 km north-west of the major regional town of Newcastle (Fig. 1). It comprises a basalt peak rising to 673-m elevation above sea level, 340 m above the surrounding landscape otherwise dominated by sandstone hills and ridges of the Triassic Narrabeen series. Present-day vegetation on Mount Dangar is determined primarily by aspect and soil depth, and fire history. It is dominated on deeper basalt soils by A. dangarensis on southern and eastern slopes, together with less abundant Eucalyptus ‘albemol’ (a purported hybrid between E. albens and E. moluccana: McRae and Cooper 1985), Callitris glaucophylla, Ficus rubiginosa and Brachychiton populneus. The mid-storey is dominated by Notelaea microcarpa, but few other shrub species are evident. Ground-layer vegetation is herbaceous and grassy in nature, varying widely in response to local rainfall conditions. Common taxa include Nyssanthes diffusa, Solanum brownii, Rytidosperma racemosum var. racemosum, Paspalidium criniforme, Einadia hastata, Einadia nutans, Clematicissus opaca, Dysphania carinata and Euphorbia planiticola. During times of drought, such as occurred in 2017–2019, almost all of the ground-layer vegetation dies back or is consumed by herbivores. The dense stands of A. dangarensis on sheltered slopes are even aged (S. Bell and P. Lamrock, unpubl. data) and often occur to the exclusion of all other tree species, consistent with major episodic germination events of several other Acacia species (e.g. Hunter 2005; Prober et al. 2007; T. Tame, unpubl. data, 1981).

Fig. 1.  Location of Mount Dangar (black dot) at the eastern end of Goulburn River National Park (GRNP), showing surrounding NPWS estate (grey shading).

The much drier and rockier northern basalt slope supports vine thicket vegetation dominated by N. microcarpa with Alectryon oleifolius, Pandorea pandorana, Clematicissus opaca, Cayratia clematidea, Cynanchum viminale subsp. australe and Adiantum aethiopicum (floristic group 1 of Curran et al. 2008). Steep basalt scree slopes associated with this habitat occasionally support stands of Senecio linearifolius var. dangarensis and Abutilon tubulosum. Surrounding sandstone landscapes carry a range of species typical of such habitats (L. Hill, unpubl. data, 1999), most of which are rarely, if ever, present on the basalt. These include the trees Eucalyptus caleyi subsp. caleyi, Eucalyptus nubila, Corymbia trachyphloia subsp. amphistomatica, Eucalyptus dawsonii, Eucalyptus punctata, Eucalyptus crebra, Acacia doratoxylon, Acacia crassa subsp. crassa and Eucalyptus sparsifolia. The fire sensitive Callitris endlicheri is a common component of the slopes and ridges across Goulburn River National Park, alluding to the scarcity of major high-intensity fire in these landscapes.

The low elevation of the Great Dividing Range in the Hunter Valley facilitates the presence of more western elements in the vegetation (Di Virgilio et al. 2012), which in combination with a dry climate infers fire patterns more typical of western lands (i.e. rare high-severity fires and long inter-fire intervals). Fire ignitions leading to major fire events are, consequently, rare in Goulburn River National Park, and since records began in 1983, over 66% of the reserve has remained unburnt (Lamrock 2017). Lightning strikes are the main ignition sources, although relative to the nearby Wollemi National Park, these occur at a lower frequency. Low-rainfall conditions contribute to relative slow plant growth, and, hence, considerable time is required to develop sufficient fuel structure to promote large wildfires. On average, large wildfires in the reserve occur at intervals of ∼20 years (L. Menke, pers. comm.).

Complete post-settlement histories of fire events in any one region are rarely available, and, hence, for this study multiple sources were examined. Fire records held by the NSW National Parks and Wildlife Service (NPWS) for Mount Dangar (1979–1980 to the present) were reviewed, and Landsat imagery (https://landsatlook.usgs.gov/viewer.html; July 1972–October 1992) was examined to search for fire scarring of the vegetation, as has been done elsewhere (e.g. Milne 1986; Callister et al. 2016). A review of historical reports in local newspapers and other written sources (https://trove.nla.gov.au/newspaper/; search terms ‘fire’, ‘wildfire’, ‘Mt Dangar’, ‘Mount Dangar’ and nearby local settlements ‘Gungal’, ‘Denman’, ‘Sandy Hollow’), and oral recollections from long-time residents of the area were also undertaken.

To provide a physical on-ground indication of the last fire-induced recruitment event of A. dangarensis that might corroborate written and oral evidence, dating of senesced individuals was undertaken using radiocarbon (¹⁴C) and dendrochronological techniques. Two individuals, one from each of the southern and northern faces of Mount Dangar and representative of the size class of the majority of the population, were analysed for radiocarbon at the Chronos ¹⁴Carbon-Cycle facility at UNSW (four samples taken from each tree). These eight samples were chemically pre-treated following the base–acid–base–acid-bleaching (BABAB) methods outlined in Turney et al. (2021), modified to include three initial acetone washes of 30 min each at 55°C and the first base step performed as 30 min base washes at 75°C, rather than one single base wash left at room temperature overnight. These samples were heavily resinous; so, seven base washes were performed during this first base step to ensure all mobile resin material had been removed from the wood. The eight radiocarbon samples were then graphited, measured and analysed as per Turney et al. (2021). Additionally, dendrochronological analysis was undertaken in conjunction with radiocarbon dating, following the multi-technique approach to tree dating discussed in Haines et al. (2018). The two trees were sampled at multiple heights along the trunk with cross-section discs collected. For each cross-section, ring counting was conducted and ring structures analysed and compared across discs within each tree.

Bureau of Meteorology (2022a, 2022b) climate data were examined to provide an outline of recurrent drought and rainfall patterns from the late 1800s until the present, to identify climatic cycles that may promote fire events. Few weather stations were recording data prior to 1900; hence, long-term patterns in rainfall variability and drought can be identified only by reference to documented trends. The Denman (Palace Street) weather station (Station number 61016; 105-m elevation; ∼20 km east-south-east of Mount Dangar) provided a useful proxy in the absence of more reliable data on rainfall at Mount Dangar; it has a near-continuous rainfall record of 131 years between January 1883 and September 2014 (October 2003 the only data gap). For the period October 2014 to December 2019, the Baerami (Old Dairy) weather station (Station number 61423; 205-m elevation; ∼5 km south of Mount Dangar) has been added into this dataset to provide a continuous 136-year record. To examine the interplay of climate and fire, ‘dry’ periods were defined as ≥3 consecutive years of below-average rainfall, and ‘wet’ periods were explored through analysis of moving average annual rainfall measured over consecutive intervals of 5–20 years. Identified wet periods were considered representative of the likely accumulation of plant growth material that might be sufficient to carry major fire.

Writings of early botanical explorers (pre-1900) are increasingly being called on to inform contemporary land management and re-create understandings of the vegetation prior to European settlement (e.g. Bowman and Panton 1993; Franco and Morgan 2007; Silcock et al. 2013; Green et al. 2020). To this end, a search of the literature was undertaken to review the ecological observations of explorers who traversed Mount Dangar. Post-1900, visitation by other more contemporary workers increased as European settlement expanded, although, unlike for the earlier explorers, few recorded their observations in any systematic way. As a proxy, examination of the plant collections contained within the Australia’s Virtual Herbarium database (AVH, see https://avh.chah.org.au/) was undertaken to glean information on post-1900 explorers, their collected specimens and habitat notes.

A review of NPWS fire records showed that no fires have occurred on Mount Dangar over the past 42 years, and although not all images were of a sufficient clarity to be certain, we could find no evidence of wildfire over the 1972–1992 period in Landsat imagery. Combined, data from these two modern-day sources imply that no fires have occurred across Mount Dangar over at least the past 49 years, since 1972. Historical written sources returned very few hints of wildfire prior to 1972. The Maitland Daily Mercury briefly reported on Tuesday 23 January 1906 (p. 3) that ‘a large bush fire in the vicinity of Mount Dangar, near Denman, was raging on Saturday night, and it is to be hoped did little damage to the holdings of settlers in the locality’. Meyer (1987) refers to a major bushfire that occurred in the Worondi and Gungal districts during the ‘New Year of 1905’, which, despite inconsistencies in year, may or may not represent this same blaze. These accounts represent the only documented evidence that Mount Dangar had potentially burnt. T. Tame (unpubl. data, 1981) briefly described a fire that began ‘in the vicinity of Sandy Hollow’ on the 2 December 1979, following 7 months of below-average rainfall, but there is no suggestion that it extended to Mount Dangar.

Fifth-generation resident of the settlement of Gungal, Clarise Judge, recalled a major fire event in November–December 1957 (while she was heavily pregnant) that burnt ‘everything’ between the Golden Highway at Gungal south to the Goulburn River. Mount Dangar lies within this area, and, although it cannot be definitively proven that Mount Dangar was burnt during this event, the implication is that it was. Corroborative evidence for this timeline was found in the results of ¹⁴C and dendrochronological analyses. The radiocarbon dates from all eight samples of A. dangarensis analysed are shown in Table 1, with samples collected from the lowest cross-sections (A1 and B1) as well from a cross-section mid-way up each tree (A4 and B4). For all four cross-sections, trees were sampled from the innermost ring, with additional samples taken from the lowest cross-sections from the outermost ring, as well as a ring midway between these points (Table 1, Fig. 2). Numerous false rings were noted; however, these rings did not appear around the entire circumference of the cross-sections. Additionally, true rings showed sharp boundaries throughout when viewed under magnification, but the false ring boundaries became blurred. This allowed for easy delineation between true and false rings, with the ring counts confirmed by the radiocarbon dates, and final ages assigned by combining the two methods. Comparing ring structures between the A1 and A4 samples as well the B1 and B4 samples also corroborated the ages assigned to the two trees. It was determined that Tree A presented rings from 1960 to 1996 and Tree B from 1962 to 1997. The innermost ring on Tree A and the outermost ring on Tree B are slightly different from the respective inner and outer dates given in Table 1 because of the poor quality in some rings (as seen in Fig. 2), which meant the ring sampled for radiocarbon analysis was selected a few rings away from the inner or outer ring. As the lowest cross-sections were sampled slightly off the ground and it takes time for a tree to commence growth after a fire, these results indicate that the fire of 1957–1958 was likely to be the source of seed germination for these specimens and the remainder of the population, given sampled trees were representative of the dominant size class.

Table 1.  Radiocarbon F¹⁴C values for the eight Acacia dangarensis samples from two trees analysed at the Chronos ¹⁴Carbon-Cycle Facility.

Fig. 2.  Cross-sections of the two Acacia dangarensis trees with the sample locations for radiocarbon dating noted as red stars.

The Bureau of Meteorology (2022a, 2022b) reported that three major droughts have affected regional New South Wales over the past 125 years, namely, the ‘Federation Drought’ (1895–1903), the ‘World War II Drought’ (1937–1945) and the ‘Millennium Drought’ (2001–2010). These drought events have persisted for 8–9 years, and additionally several shorter intense droughts have also occurred across this period (e.g. 1914–1915, 1965–1967, 1982–1983). The most recent drought began in mid-2017 and it is equivalent to a major drought event on the long-term historical record (100 years).

For Mount Dangar, the combined long-term rainfall record at Denman (n = 131 years, 1998–2014) and Baerami (n = 5 years, 2014–2019) shows nine dry periods over the course of 136 years (Fig. 3). These periods occurred at more-or-less regular intervals but many included occasional good years (above-average rainfall). Dry periods occurred from the commencement of records in 1883–1888 (6 years, possibly longer), 1899–1907 (9 years), 1912–1923 (12 years), 1925–1929 (5 years), 1935–1944 (10 years), 1964–1967 (4 years), 1979–1983 (4 years), 2002–2006 (5 years) and 2017–2019 (3 years). Four of these dry periods coincided with the three major droughts identified as the Federation Drought, World War II Drought and Millennium Drought, but there have been several other very dry periods in the area around Mount Dangar, spanning between 3 and 12 years of duration. The very dry year of 1957 (291 mm, half of the long-term annual average) corresponds to the period recalled by Clarise Judge, and is supported by ¹⁴C and dendrochronology results, for the major fire that was likely to be burning across Mount Dangar.

Fig. 3.  Annual rainfall for Denman (Station 61016; January 1883–September 2014) and Baerami (Station 61423; October 2014–December 2019), highlighting very dry years (darker histograms; <300 mm year⁻¹, or 50% of long-term annual average) and 8-year moving average rainfall (solid black line, 1955–1956 circled).

Prior to ignition of fire in November 1957, the total rainfall for January to October 1957 was just 213 mm (36% of the annual average) and this was the driest January to October period for any year over 136 years in this locality (the next-lowest was 227 mm in 1918). For the 10 years preceding the very dry 1957, seven of those years were wet years with well above-average rainfall. This suggests that plant growth over the 1947–1956 period would have been prolific, leading to a build-up of fuels sufficient to carry a fire over Mount Dangar. Contemporary observations show that fuel build-up and decay happens rapidly on Mount Dangar in response to rainfall (S. Bell and P. Lamrock, unpubl. data), differing from surrounding sandstone landscapes where fuel loads are more constant. High fuel accumulation followed by rapid drying prior to decay, therefore, provides conditions more conducive to fire and detecting previous rainfall patterns that replicate this scenario may identify other potential fire opportunities.

Consequently, examination of average annual rainfall over periods of 5–20 years between 1884 and 1957 found the 8-year moving average to be the greatest immediately prior to this fire (1956), at 774 mm and 186 mm above the long-term average (588 mm). The next-highest were the 7-year (765 mm), 9-year (758 mm) and 10-year (740 mm) averages, suggesting that between 7 and 10 years of high rainfall was sufficient to accumulate the necessary fuel for the 1957 wildfire event to carry across the basalt. Plotting the 8-year moving annual average rainfall for all years to 2019 showed the highest peak in 1956, immediately preceding the very dry 1957 and fire igniting in November of that year. Fig. 3 shows that the combination of very low rainfall (<300 mm year⁻¹) following a very wet preceding 8 years (>760 mm) has occurred only twice (1955, 1956) between 1890 and 2019, leading up to the 1957 fire event. Other wet periods have occurred, evidenced in the many peaks shown in Fig. 3, but these have not been followed by a very dry year where rainfall is less than half of the long-term annual average.

Allan Cunningham, botanical explorer and collector, undertook four expeditions through the Hunter Valley between 1823 and 1827 (Lee 1925; Whitehead 2013, 2017a, 2017b). On 22 April 1825, Cunningham climbed Mount Dangar, likely becoming the first European to reach the summit, a feat possibly not repeated until R. T. Baker and W. M. Carne collected there in 1903. Cunningham’s observations and notes have provided a significant reference point on the vegetation of Mount Dangar at that time, and, with a viewing of the paths where he walked (recreated in Whitehead 2017a, 2017b), are even more valuable. Cunningham traversed north-east from his campsite on the Goulburn River to Mount Dangar summit, and then descended down its south-easterly flank back to the Goulburn River. Fig. 4 presents an interpretation of this likely route, on the basis of Whitehead (2017a) but augmented by our own experiences traversing Mount Dangar and surrounding hills, and including the current-day distribution of A. dangarensis (S. Bell and P. Lamrock, unpubl. data). On this trek, Cunningham made note of over 30 plant taxa and contrasted those occurring at lower and upper elevations (see Supplementary Table S1). He made the specific observation that there was little difference in vegetation between that occurring near the summit (on basalt) and that lower down (presumably on the sandstone ranges), stating that ‘In our ascent to the Summit of Mt. D. the vegetation differs nothing from that of the Callitris ranges below it…’ (unpublished ‘Journal of Allan Cunningham’s excursions, 1822–1832’, p. 35). Cunningham recorded only a single wattle species on that day (‘an Acacia with large falcate leaves and cylindrical spikes’), most likely to be Acacia crassa subsp. crassa (Fig. 5), which today is common on the sandstone hills; no Acacia was recorded from the summit. This is of particular interest, because current-day knowledge of the distribution of A. dangarensis on Mount Dangar shows that Cunningham walked through these areas on both his ascent and descent that day (see Fig. 4), and in 2022 this species dominates the slopes here in almost mono-specific stands (Fig. 5). Had A. dangarensis been present above-ground in 1825, it might be expected that Cunningham would have collected, or at the very least written about the presence of, this species.

Fig. 4.  Allan Cunningham’s probable route on 22 April 1825. Both his ascent and descent passed through present-day vegetation dominated by Acacia dangarensis (see Fig. 5). Route modified from that shown in Whitehead (2017a) on the basis of recent field reconnaissance.

Fig. 5.  Acacia crassa subsp. crassa (top left) growing on the sandstone hills around Mount Dangar, the likely ‘Acacia with large falcate leaves and cylindrical spikes’ of Cunningham. Acacia linearifolia (top right), the current-day name for a specimen collected by Cunningham in 1827 near Mount Dangar (his ‘Acacia dangieri’). Acacia dangarensis (middle row) dominating on the south-western (left; 2018, during drought) and south-eastern (right; 2020, post-drought) spurs of Mount Dangar, the likely route of Cunningham’s trek in 1825. Senecio linearifolius var. dangarensis (lower left), and circular leaflets of Kennedia retrorsa (lower right), both resident on Mount Dangar.

Of interest among Cunningham’s other observations is a ‘Senecio sp. glaucous lanceolate leaves’ from the upper slopes, which can refer only to Senecio linearifolius var. dangarensis, a taxon that has been until recently known solely from Mount Dangar (Fig. 5). Cunningham also noted ‘Kennedia rubicunda or a new sp. rounded leaves’, which is almost certainly the vulnerable Kennedia retrorsa (Fig. 5) occurring on the sandstone slopes and gullies in the vicinity of Mount Dangar (Kennedia rubicunda is a more coastal species, not known from the mountain nor Goulburn River National Park; L. Hill, unpubl. data, 1999). The fact that Cunningham observed and noted these rare species is testament to his observational skills and ability to identify regional novelties.

Other pre-1900 explorers mention Mount Dangar in their writings but few with a botanical lean appeared to have climbed it. Ludwig Leichhardt explored much of the Hunter Valley in the 1840s, including the area around Wybong, Denman and Mount Dangar. He wrote on 18 April 1843 ‘I had found the Dodonaea pinnatifida [perhaps Dodonaea multijuga or Dodonaea boroniifolia] on similarly dry ground near Mt Dangar’ implying that he knew the country reasonably well (Darragh and Fensham 2013), but apparently did not climb it.

From 1900, there has been increased interest in botany and botanical exploration in the Hunter region, as in other parts of the country. An extract of all plant specimens from AVH undertaken in January 2022 for Mount Dangar and immediate surrounds shows that 28 botanists have collected there over a period of 115 years between 1903 and 2018 (see Supplementary Table S2). Several of these collectors were eminent botanists of their time (Supplementary Table S3), working in a period when there were many poorly explored regions and lodging many thousands of plant specimens in different herbaria. W. M. Carne and R. T. Baker were the first to collect in 1903, submitting just three taxa. Other collectors, such as J. L. Boorman in 1904, were fastidious observers and collected nearly 60 different taxa from the area. Boorman collected seven Acacia species (but not A. dangarensis) as well as other (now) threatened taxa, including Kennedia retrorsa, Pomaderris queenslandica and Senecio linearifolius var. dangarensis. Subsequent expeditions were undertaken by many other well-regarded botanists between 1908 and 2018 (see Table S2).

Acacia dangarensis was collected on Mount Dangar only from 1979 at its initial discovery, and it remains extant in 2022. No earlier observation records exist in the NSW Bionet Wildlife Atlas database (searched January 2022) for this species. Notes associated with the 1979 collections refer to dominant straight-stemmed trees 8–10 m in height, but trees of up to 18 m high now occur in some parts of the mountain.

Allan Cunningham described the vegetation on Mount Dangar in 1825, with little specific observations of the dominant canopy species, yet such knowledge would have been particularly informative for threatened-species management today. However, he did observe that there was no obvious change from the Callitris-dominated sandstone ranges at lower elevations to the basalt slopes higher up. This was despite the distinction made between the two underlying geologies; ‘Mount Dangar is of a sandstone whereas its summit is a whin [=hard dark rock, basalt] of coarse fragments’. This suggests that at the time Cunningham was on Mount Dangar, much of it was characterised by Callitris. The upper slopes today support some individuals of very old C. glaucophylla (Fig. 6), and these predominantly occur on the exposed western end of the mountain but are scattered across other areas including within low forest of A. dangarensis. Drought in eastern NSW during 1823–1824 (Nicholls 1988; Fenby and Gergis 2013) and general dry conditions on the NSW central slopes between 1820 and 1828 (Freund et al. 2017) suggest that ground-layer vegetation may have been scant in 1825, which may explain the lack of distinction made by Cunningham between sandstone and basalt habitats. Our own observations on the effect of the recent 2017–2019 drought on ground vegetation at Mount Dangar would support this.

Fig. 6.  Callitris glaucophylla on the western end of Mount Dangar today, perhaps a remnant of what may have once dominated the whole mountain.

Low-stature box (Eucalyptus ‘albemol’) dominates the upper summit of Mount Dangar today, and scattered individuals occur across most slopes, yet Cunningham made little mention of these in his notes from 1825. Box eucalypts were one of the earliest recognisable trees in the first days of the colony, given their presence on the Cumberland Plain (Eucalyptus moluccana: Benson and Howell 2002) and dominance on the western slopes of NSW (e.g. numerous mentions in Cunningham’s 1823 journey from Bathurst to the Liverpool Plains; Field 1825). Even after leaving Mount Dangar and on the same journey northward in 1825, Cunningham made mention of box trees near the Liverpool Plains (Lee 1925). For the summit of Mount Dangar, Cunningham recorded only ‘a small eucalypti’ among a number of other shrubs and herbs, suggesting that eucalypts may have been insignificant there in 1825 (or perhaps he was unable to recall their identity when writing his notes that evening). The fact that he did not think it necessary to contrast the vegetation occurring on the basalt with that on the lower sandstone areas, as he did with the geology, provides strong circumstantial evidence that, in 1825, the two habitats appeared similar and were characterised by Callitris. This makes the apparent absence of A. dangarensis at that time, when it is so dominant today, particularly intriguing. We suspect that, in 1825, there was no stand-dominating A. dangarensis present on Mount Dangar, but that it lay in dormancy as seeds in the soil awaiting a fire event of sufficient intensity.

However, there remains the possibility that A. dangarensis was present above ground as scattered individuals in 1825 and was missed or overlooked by Cunningham, and that it persisted for subsequent decades in low abundance. Given our observations and analyses that the even-aged stands of trees now present on Mount Dangar germinated c. 1960, evidence for A. dangarensis existing prior to this should be found in the observations of earlier workers, but this is lacking; four separate collecting events (1903–1953; see Table S2) by eminent botanists did not report on or collect the species during this period. Prior to Cunningham’s visit, historical weather patterns suggest that conditions suitable for a fire-induced mass germination event last occurred between 1730 and 1760 (discussed later), although sporadic, intermittent germination from the seed bank may have occurred in the inter-fire interval after this period.

The interplay of fire and drought has shaped much of the vegetation in Australia (e.g. Sakaguchi et al. 2013; Zhang et al. 2017a, 2017b), including parts of the upper Hunter Valley (Tierney and Watson 2009; T. Tame, unpubl. data, 1981). This region receives between 500 and 800 mm of rainfall annually, and, as a consequence, most areas support an open vegetation with few areas of rainforest or wet sclerophyll forest (McRae and Cooper 1985; Peake 2006; L. Hill, unpubl. data, 1999). This suggests that wildfires are rarely contained by structural features of the vegetation, except where low-fuel basalt habitats limit fire spread under most conditions. Low rainfall and high temperatures throughout the year are unconducive to rapid plant growth, meaning that fuel build-up in these landscapes is a slow process. Irregular, but extreme, fire events, therefore present opportunities for major stand-altering disturbances to the vegetation, particularly for habitats dominated by slow growing and fire-sensitive Callitris species.

Stand-altering fire events might explain the differing habitat observed on Mount Dangar by Cunningham in 1825, compared with that present today. Denham et al. (2016), for example, documented changes to the vegetation following a major fire event in the Warrumbungle Ranges, in areas that had remained unburnt for at least 50 years. This region has a climate comparable to that in the upper Hunter Valley (rainfall 750 mm year⁻¹, warm to hot summers and mild winters), and shares with it many plant species. Following extreme wildfire there in 2013, habitats dominated by both Eucalyptus and Callitris underwent considerable structural change, with severely burnt areas suffering nearly 100% Callitris death but eucalypts displaying considerably better survival. Any recovery of Callitris to pre-fire densities in these areas now relies on post-fire seedling emergence; however, in the interim, major habitat changes have occurred in areas where Callitris once dominated. Mature stands of C. glaucophylla are known to reduce fire potential in highly flammable environments, and fire-free intervals of at least 50 years are required to enable sufficient recovery to supress all but extreme fire events (Cohn et al. 2011). This implies considerable modification of the landscape post-fire, and recovering habitats will be shaped by subsequent climate and fire recurrence.

Fires in the Goulburn River valley are generally limited by slow plant growth and fuel accumulation, but when they do occur they tend to be major events (L. Menke, pers. comm.). The preponderance of C. endlicheri (another fire-sensitive species) across sandstone habitats of Goulburn River National Park supports the notion that major fire is rare, and historical and contemporary observations of fire events provide further evidence. For example, on two occasions Meyer (1987) discussed a major bushfire that occurred in 1905 at the foot of Mount Dangar and associated hills, but there is no indication that the mountain was affected by this fire, and observations by NPWS fire practitioners conducting hazard-reduction burns at the sandstone–basalt interface support this fire behaviour (Lamrock 2017). As evident on Fig. 3, 1905 was one of five very dry years where annual rainfall was less than half of the long-term average, but in this case the preceding 7–10-year rainfall amounts were insufficient to accumulate the fuel necessary for wildfire.

More recently, a major fire event beginning to the west of Goulburn River National Park in 2017 (the ‘Sir Ivan Doherty Fire’) ignited near Leadville and rapidly spread east. This blaze ultimately burnt through 55 324 ha of forest and grassland and was eventually brought under control only with a favourable weather change. Without this change in weather, it is feasible that this fire may have continued east right through the national park to Mount Dangar. This very dry 2017 (338 mm at Denman) was preceded by 10 years of mostly above-average rainfall approaching that seen in 1957, conducive to a major fire event. Hindsight now suggests that the Sir Ivan Doherty Fire may have been of the appropriate scale required to burn Mount Dangar, consuming the current cohort of A. dangarensis but breaking dormancy in the well stocked soil seed bank underneath.

Previous studies on the incidence of historical droughts and wet periods in Australia have provided strong corroborative evidence for the role of extreme fire and climate conditions on the ecology of Mount Dangar. The drought atlas for eastern Australia (Palmer et al. 2015) shows two major periods of excessive wet conditions over the past 500 years, one of these (1950s−1970s) corresponding to the period leading up to the postulated 1957–1958 fire event at Mount Dangar, and the second for the 1730–1760 period. A third less obvious wetter period also occurred between 1560 and 1580. These three periods in history, where several successive very wet years (in which to accumulate fuels) were followed by drought (to promulgate extreme wildfire) recur approximately over a 200-year cycle. As postulated here, high levels of plant growth promoted through the successive wet years of 1949–1956, followed by rapid curing of vegetation during the very dry 1957 created ideal conditions for a major wildfire. The devastating effects of severe short-term reductions in rainfall leading to catastrophic fire events have been noted elsewhere (e.g. the Black Thursday bushfires of 1851, Freund et al. 2017), and repetition of this pattern during earlier centuries may have similarly led to major fire events driven by wet–dry climate phases. Key canopy species including C. glaucophylla and A. dangarensis may, consequently, be operating in alternating cycles of dominance across the mountain. Under this scenario, continuing A. dangarensis senescence and fire absence may permit C. glaucophylla to re-emerge as the dominant species over the next century, with A. dangarensis dominance potentially returning during the 22nd century.

Links between wet–dry climatic cycles and major fire events have been identified in other habitats elsewhere in Australia. For vegetation dominated by Eucalyptus camaldulensis in the Murray River valley, Zhang et al. (2017b) found that fire risk was best explained by rainfall conditions occurring prior to the event. Similarly, Silcock et al. (2016) established that fire events in the mulga (Acacia aneura) lands of semi-arid Queensland were infrequent and occurred only after at least 2 years of above-average summer rainfall, arguing that fire is, consequently, feasible only a few times per century. Both of these scenarios are similar to our proposed hypothesis for Mount Dangar. In this case, average or above rainfall over a 7–10-year period is first required to enable sufficient growth and accumulation of plant material, followed by a rapid change to severe drought, to prepare fuel for burning and promulgate fire.

Rare episodic events shaping major vegetation change can have important implications on the management of threatened plant species (Lunt 2002; Elliott et al. 2019; Miller et al. 2019). For Mount Dangar, it appears from fire history and rainfall records and the writings of Cunningham that the nature of the mountain has changed dramatically within the past 200 years, with a major shift identified owing to wildfire in 1957–1958. We hypothesise that this fire event removed most of the existing C. glaucophylla-dominated woodland and replaced it with dense forests of Acacia dangarensis. Now listed as a critically endangered species because of inappropriate fire regimes, observed senescence and little recruitment (NSW Threatened Species Scientific Committee 2018), understanding the ecology of this species is paramount for the future management of Mount Dangar, particularly in the face of declining above-ground populations (Bell and Elliott 2013). Complicating this understanding is the presence of three other threatened species on or below the mountain (and several others in adjacent habitats). However, knowledge gained from historical collections for two of these suggests strong resilience to environmental stochasticity; both Senecio linearifolius var. dangarensis and Kennedia retrorsa, although not named as such at the time, were anecdotally noted from Mount Dangar by Cunningham in 1825, and have been recorded intermittently there ever since, despite fluctuating wet and dry periods, and the postulated 1957–1958 fire event. The third species, Lasiopetalum longistamineum, occurs in sheltered locations below the mountain and was first collected in 1904 and has also shown resilience through multiple drought periods.

For the structure-changing small tree A. dangarensis, it is telling that the first recorded collections of this taxon were not until 1979 (see Supplementary Table S4), despite the visits of Cunningham in 1825 and eight further botanists over nine visits between 1903 and 1973. Most of these botanists were keen observers and collectors and it seems unlikely that all of them overlooked the presence of what is now a community dominant tree on the southern and eastern flanks of Mount Dangar. In 1904, J. L. Boorman undertook a detailed survey of the mountain, collecting 57 different taxa, which included 7 other Acacia species, and the threatened Kennedia retrorsa, Pomaderris queenslandica and Senecio linearifolius var. dangarensis. The postulated fire event in 1957 falls between L. A. S. Johnson’s visit there in 1953 where no A. dangarensis was collected, and initial collections of the species by R. Coveny in 1979 (when plants would have been ∼17 years of age). Other collectors in the intervening period (H. Dorman in 1966, A. N. Rodd in 1966 and 1968, J. R. H. Phillips in 1968, G. L. Webster in 1973) collected predominantly from sandstone-based habitats at lower elevations and may not have fully explored A. dangarensis habitat (see Table S2).

If, as postulated, germination of the current cohort of A. dangarensis was triggered by the 1957–1958 wildfire event, trees would now be ∼60 years of age. Dendrochronology and ¹⁴C dating of two sample trees, representative of the dominant size class of the population, support this germination date. This implies the following two possible scenarios for germination of the current cohort of trees: either scarification of seed following the fire event may have triggered some germination between 1958 and 1961, these seedlings succumbing to water stress, desiccation or the intense grazing pressure that would have inevitably followed; or, no germination occurred until the very wet 1963 (>320 mm above average), at which point sufficient ground-layer vegetation had developed to ease both desiccation and grazing impacts. Several other long-lived Acacia species display similar episodic germination events in response to rare fire events. For example, populations of Acacia silvestris (Clayton-Greene and Wimbush 1988), Acacia binervia (Benson and McDougall 1996), Acacia blakei (Hunter 2005), Acacia chrysotricha (P. Richards, unpubl. data, 2011), Acacia wollarensis (Bell and Driscoll 2017) and Acacia covenyi (Tozer et al. 2021) are all thought or known to germinate post-fire and develop dominant long-lived stands maintained by fire on long fire-free cycles.

Long-term viability of soil-stored A. dangarensis seed, perhaps for up to 150 years, is central to our hypothesis of centurial habitat change at Mount Dangar. Resistance to desiccation of Acacia seed in habitats that, like Mount Dangar, rarely experience major disturbance events such as fire is evident in many species through their hard, impervious outer layers. Leino and Edqvist (2010) suggested that most Acacia following this strategy occur in dry climates, the seeds surviving in a dried state and thereby reducing the rate of desiccation. Long et al. (2015) noted that some hard-seeded species including those in the Fabaceae have reportedly persisted in soil for up to 1300 years, although these are likely to be the exception rather than the rule. They argued that persistence of any seed in the soil seed bank is dependent on its resistance to germination (absence of suitable growth conditions) or death (natural attrition), and on the prevailing conditions that are conducive to those fates. Ewart (1908) made the astute observation that hard-coated seeds such as those of Acacia distribute themselves in time rather than space. The impermeable coats of his ‘macrobiotic’ seeds (those remaining viable for 15 to >100 years), best exhibited in the Fabaceae, show no special means of dispersal, and that ‘the probable extreme duration of vitality for any known seed [of the Fabaceae] may be set between 150 and 250 years’. Research by Gilbert (1959) on community succession in Tasmanian wet sclerophyll forests extended this maximum longevity for Acacia seeds, documenting how large numbers of Acacia dealbata seedlings emerged across extensive areas following logging of mixed eucalypt forest that had not been disturbed for an estimated 350–500 years. Such observations imply that it is perfectly feasible for large numbers of A. dangarensis seed to remain viable in the dry environment of Mount Dangar for periods of 100–150 years.

However, we recognise that it is also plausible for occasional intermittent germination and persistence of A. dangarensis to have occurred without fire disturbance, leading to localised replenishment of the seed bank in some areas. For this to have played a major role in current-day distribution, it must have regularly occurred at high densities (100s of individuals) and across a widespread area (∼80–100 ha) to explain the large dominant stands of the species that are present today, and such a scenario is not reflected in historical collections from the area (no observations or collections between 1825 and the wildfire in 1957). Subsequent to release, Acacia seed is primarily dispersed by ants (Auld 1996), and as a group these vectors move seed over only short distances from the parent plant (<3 m: Gómez and Espadaler 2013). Seed therefore accumulates in the seed bank close to its parent, and for a seed bank covering 80–100 ha to develop and source the current cohort of trees, the species would have either (1) been highly prevalent and easily observed by botanists consistently over the past 200 years, or (2) undergone intermittent germination and seed-bank accumulation progressively across 80–100 ha and over several centuries to account for current day distribution. The first of these scenarios seems improbable, and the second, at the very least, confirms centurial seed longevity in this species; therefore, mass germination triggered by major fire is the most likely life strategy operating in this species.

Multiple lines of evidence (review of fire and climate records, dendrochronology and ¹⁴C dating, historical and contemporary writings, oral accounts, botanical collections and observations) have outlined the possible ecological history of Mount Dangar over the past 200 years. We postulate that the only major fire occurring on the mountain during this time frame was in the Austral summer of 1957–1958, which resulted in a stand-altering change to the vegetation by removing the likely dominant C. glaucophylla canopy and stimulating mass germination of A. dangarensis c. 1960. This fire occurred during an exceptionally dry year following 7–10 years of very wet conditions, a scenario that appears to repeat approximately every 200 years. Prior to this germination event, it is likely that A. dangarensis was not present in the above-ground flora in any great numbers and may have persisted only in the soil seed bank for ∼150 years (to c. 1800). The current cohort of Acacia individuals, at 60 years of age, are in a senescent phase and may persist only for another decade. However, provided that occasional flowering and fruiting of individuals continues to replenish the seed bank, there is little urgency to instigate recovery actions until perhaps 50 years after such replenishment has ceased (c. 2080). Acacia dangarensis assumes the model of a species compatible with centurial-scale disturbance events, typically in the form of extreme fire with high longevity of soil stored seed. Given the current age of individuals, the next wildfire event to affect Mount Dangar should be managed for promulgation rather than suppression, to release seed from dormancy and refresh above-ground populations. On the basis of history, such an event will occur only under extreme environmental conditions and may not eventuate for a considerable period of time.

Supplementary material is available online.

The data that support this study will be shared upon reasonable request to the corresponding author.

The authors declare that they have no conflicts of interest.

Data collected for this project was completed under contract to the first two authors as part of a wider project funded through the NSW Government Saving our Species initiative, supported in-kind by remaining authors.