Identification and characteristics of refuges for the threatened swamp antechinus (Antechinus minimus maritimus) under climate change; targeted surveys across the Otway Ranges, south-east Australia

Study area

This study was undertaken in the Otway Ranges, approximately 100 km south-west of Melbourne, Victoria (Fig. 1). The area (186,377 ha) is predominantly public land including the Great Otway National and Otway Forest Park (140,000 ha) and the Carlisle State Park (5600 ha), comprising heathland, woodland and forest (Parks Victoria 1998). The vegetation of the study area includes a diverse mosaic of Ecological Vegetation Communities (EVCs; the standard unit for classifying vegetation types in Victoria, based on floristics, lifeforms and ecological characteristics) predominantly of Wet Forest (42,357 ha) and Shrubby Wet Forest (32,858 ha), distributed inland from Lorne to Cape Otway, (Wilson, B. A., Burns, P., and Garkaklis, M. J. unpubl. data). Dry forest consisting of Lowland Forest (41,301 ha) and Shrubby Foothill Forest (29,910 ha) extends from the eastern Otways on a plateau, and coastally from Anglesea to Lorne. Heathy Woodland covers 23,198 ha and is distributed in an arc from the east (Anglesea) to central Otways north of Forrest, and southwest of Carlisle. Within the woodlands are areas of Riparian Scrub or Swampy Scrub/Woodlands (14,595 ha) and limited areas of Wet Heathlands (2158 ha).

Fig. 1.

Study area across the Otways including eastern (Anglesea–Lorne), central (Lorne–Cape Otway) and western (Lorne–Port Campbell) areas, showing location in Victoria, and historical records for swamp antechinus (1950s to 2020).

The most extensive recent fire in the area was the ‘Ash Wednesday’ wildfire (1983), which burnt 40,000 ha and left few unburnt small patches of vegetation (Wilson et al.1990, 2001). Fuel reduction burns have been undertaken since 1986 and an increase in burning has been implemented following recommendations arising from the 2009 Victorian Bushfires Royal Commission (Coates, F., Loyn, R., Ibbett, M., and Brown, G. unpubl. data; Parliament of Victoria 2010; Department of Environment and Primary Industry 2014). The total area of heathlands burnt by management between 1987 and 2008 (21 years) was 2266 ha (mean = 107.9 ha per annum), and between 2009 and 2012 (3 years) 3095 ha (mean = 1031.67 ha per annum), an annual increase of 900% (Wilson et al. 2017). In 2013, 42% of the Anglesea heathland was below its ‘minimum tolerable fire interval’, and 37% was at an early growth stage (Department of Environment and Primary Industry 2014). Regular fox baiting has occurred since 2005 (Parks Victoria and Department of Sustainability and Environment 2009).

Historical occurrence of the swamp antechinus across the Otways (1950s–2021)

To assess the historical distribution of the swamp antechinus across the Otways, all available records for the species were obtained for the period 1957–2021 (Wilson, B. A., Burns, P., and Garkaklis, M. J. unpubl. data). This included records from the Victorian Biodiversity Atlas (VBA) (as of 27 January 2021) together with all other known databases containing historical small mammal occurrence records from research institutes, universities, State Government authorities and individual researchers. Data were carefully audited to ensure spatial accuracy and to remove duplicate or unreliable records (Wilson, B. A., Burns, P., and Garkaklis, M. J. unpubl. data).

A Geographic Information System (GIS) approach was employed to map and calculate areas of all EVCs across the study area (Department of Environment, Land, Water and Planning 2019a). Swamp antechinus observations were mapped for each EVC, and calculations of swamp antechinus records per ha of each EVC was recorded as an indication of species’ preferred habitats. Accessibility to locations may have influenced location of the swamp antechinus records.

Assessment of current distribution and abundance of the swamp antechinus (2018–2023)

Surveys for the swamp antechinus targeted sites of notable healthy populations at long-term monitoring sites in the eastern Otways (1975–2007), historic records and suitable habitat identified with suitable EVCs and Habitat Distribution Models (NatureKit; Gibson et al. 2004). Furthermore, the surveys targeted sites downslope in gullies (Wilson, B. A., Garkaklis, M. J., Arnall, S., and Doherty, T. unpubl. data; Wilson, B. A., Burns, P., and Garkaklis, M. J. unpubl. data) and sites with higher Normalised Difference Vegetation Index (NDVI) values (Garkaklis and Wilson 2021) representing higher moisture and vegetation productivity, respectively, and indicative of more drought resistant refuges. Due to the targeted surveys and restrictions on time and resources, not all areas of potentially suitable habitat were surveyed.

Surveys were conducted between 2018 and 2023 with camera data obtained from 121 sites (518 cameras, 17,454 trap nights) and live trapping from 99 sites (5888 trap nights) (Supplementary Table S1). Nineteen of the sites had both camera and live trapping, and sampling sites were considered independent, being from 2 to 109 km apart. Sampling was conducted across three main areas, broadly delineated by differing EVCs: (1) the eastern Otways from Anglesea to Lorne coastally and inland to the Coastal Forest Park; (2) the central Otways from Lorne to Cape Otway coastally and inland to Forrest (dominated by wetter EVCs including Wet and Shrubby Wet Forest), and Carlisle Heath (dominated by Heathy Woodland and Wet Heathland); and (3) the western Otways from Cape Otway to Port Campbell Heath (dominated by Coastal Heathland Scrub and Damp Heath Scrub) (Fig. 1). Trapping was undertaken using Elliott traps (325 × 90 × 100 mm; Elliott Scientific Equipment, Upwey, Victoria) baited with a mixture of rolled oats, honey/golden syrup and peanut butter. At most sites, 30 Elliott traps were set in a transect formation in lines of 10 at 10–15-m intervals between lines and traps (0.3–0.5 ha), except at sites where the habitat patch was very small (<0.5 ha) necessitating only 4–10 traps being set (<0.5 ha). Traps were set for three to four nights. Species, sex, reproductive condition and body weight (g) were recorded for each animal and these data were used for multiple studies. Captured swamp antechinus individuals were identified or marked with ear notch before release at their capture point.

At each camera trapping site, predominantly one to four (and up to 10 at some sites) remote infrared cameras (Reconyx Hyperfire HC600 Passive Infrared, HP2W Hyperfire 2) were positioned approximately 100 m apart, for 3–6 weeks. Cameras were set to high sensitivity with a 0–15-s interval between triggers and three to five images taken per trigger. Bait lures of peanut butter and rolled oats were employed to attract animals. Cameras were secured vertically 0.2–1 m from ground level, at 0.5–0.9 m from the bait station, and dense vegetation was cleared from the field of view where required. Photographs of individuals were identified to species level using reference photographs from the area (B.A. Wilson pers. obs.) and species descriptions (Van Dyck and Strahan 2008; Van Dyck et al. 2013). Although identifying small mammals to species level from infrared (black and white) images can be challenging, we identified only five dusky antechinuses (Antechinus mimetes), recognised by their very contiguous dark fur, thickset body (especially the rump), long slender rostrum, very long claws, and tail which is 80% of body length. In contrast, the swamp antechinus has a shorter nose and ears, a lighter grizzled brown fur on upperparts (particularly on the rump), with very pale underparts, shorter claws, and a tail which is 70% of body length. Such differences provide confidence of correct identifications.

Trapping and camera data analyses

Although mark–recapture and other analyses have been used to derive estimates of populations (number of individuals known to be alive; density) for swamp antechinuses at particular sites (Wilson et al. 1986, 1990, 2001; Annett 2008; Magnusdottir et al. 2008; Sale et al. 2008), this precise population information was not available at the survey sites. To ensure consistent comparison, we used trapping rate as a proxy measure of abundance: Trap success = (total captures (including recaptures)/total number of nights) × 100. For camera trapping data, the total number of independent captures (visits) was used to calculate trapping success at each site by: Trap success = (total captures/total number of nights) × 100. An independent capture (visit) was defined as a 30-min interval between images of the same species (DeBondi et al. 2010; Claridge et al. 2010).

Identification and assessment of refuges for swamp antechinus

The results of recent surveys (2018–2023) were used to identify refuges based on several criteria: (1) survival of species (>20 years) during long-term monitoring; or (2) live capture success of >2.2 (2018–2023); or (3) the presence of swamp antechinus (live or camera) at least two trap sites (2018–2023); together with presence of suitable habitat (>30 ha). The results were used to map the location and extent of refuges supporting the species across the Otways, including refuges such as coastal dunes, gullies and waterways connecting coastal dunes and inland habitat (Wilson et al. 2017; Wilson and Garkaklis 2020). The refuges were delineated by incorporating the location records and mapping these in suitable habitat to topographic features (gully, road, beach) of unsuitable habitat. The location, shape, topographic features (creek, gully, dune), area (ha) and connectivity of refuges was estimated in GIS, together with the EVC.

In the initial assessment of potential refuges, three distinctive characteristics of refuges emerged: size, shape and connectivity, the latter being areas connecting small or large refuges, where suitable habitat for the swamp antechinus was present and the species was recorded at both ends indicating likely population connectivity. Three categories of refuges were then adopted: (1) micro-refuges consisting of smaller refuges (<500 ha) which were frequently linear; (2) mid-connected refuges which were medium-sized refuges (500–1000 ha) providing habitat connection features (e.g. creeks, gullies) which can connect populations; and (3) macro-refuges which were large refuges (>1000 ha) consisting of interconnected key habitats likely to support self-sustaining long-term populations of the species.

An assessment of threats was undertaken at each refuge. An estimation of extent of fragmentation of sites was based on the habitat patch size delineated by roads or tracks and scored as high (<340 ha, linear), medium (60–300 ha, not linear) and low (700–4800 ha, not linear). The presence of weeds and the plant pathogen, P. cinnamomi, were based on site observations, and scored as high (>30%), medium (20–30%) and low (<20%), and for P. cinnamomi classed as not applicable where vegetation present was non-susceptible such as coastal scrub. The presence of people, domestic dogs (Canis familiaris) and introduced herbivores including European rabbits (Oryctolagus cuniculus) and predators (fox, cat), were based on camera results and site observations. The erosion vulnerability of refuges was assessed based on the Victorian Coastal Hazard Assessment as Very High, High or Medium (CoastKit, www.marineandcoasts.vic.gov.au). The threat of wildfire was assessed as High for all refuges based on the Otway bushfire risk assessment (The State of Victoria Department of Environment, Land, Water and Planning 2020), and for fuel reduction burning as Low, if habitat was not included in spatial and temporal fuel management and ecological fire regime burn plans (e.g. coastal dunes), Medium if included in burn plans, and High if included in burn plans to be implemented in the next 3 years (The State of Victoria Department of Environment, Land, Water and Planning 2020, 2024).

Assessment of population abundance and habitat in the Carlisle Heath macro-refuge (2021)

Following preliminary survey work that identified the swamp antechinus in the Carlisle Heath (2019–2020), intensive live trapping was undertaken (April–June 2021) at 40 sites selected within suitable habitat. Trapping was conducted for four consecutive nights along two parallel 90-m transects (20 m apart) per site with 10 Elliott traps placed along each transect at 10-m spacings (total of 3744 trap nights). General trapping procedures described above were followed. The burn history at the study sites ranged from 1 to 82 years (obtained from Department of Environment, Land, Water and Planning 2019b). The abundance of swamp antechinuses in time since fire intervals of early (0–9 years), medium (10–20 years) and long (≥21 years) were calculated. In 2022, subsequent live trapping was conducted at eight of the 40 sites trapped in 2021, of which four were in recently burnt habitat (2022) and four were in long unburnt habitat (25 to ≥82 years post-fire). In 2023, repeat live trapping was conducted across four unburnt sites.

Comparison of capture success and density of swamp antechinus to previous studies

Estimates of trap success were compared to previous studies in the Otway region (Wilson et al. 2001, 2017). Density estimates for the trapping at Carlisle Heath, the number of individuals ‘known to be alive’ (Krebs 1989) divided by the trapping area, were compared to previous studies of the species including in the Otway Ranges.

Historical distribution of the swamp antechinus across the Otways (1950s–2021)

A total of 684 records for swamp antechinus were obtained between 1957 and 2021, primarily clustered in the eastern Otways, coastal Cape Otway and inland Carlisle areas (Fig. 1). The high numbers of records in the eastern Otways is related to intensive trapping studies on the species between the 1980s and 2007 (Gibson et al. 2004; Wilson et al. 1986, 2001, 2017; Magnusdottir et al. 2008; Wilson and Garkaklis 2020). The species has been recorded in the eastern Otways since the 1950s, Cape Otway since the 1970s, Carlisle heathlands since the 1980s, and Port Campbell since the 1990s.

Swamp antechinus was recorded predominantly in Heathy Woodlands, Coastal Scrub Grasslands and Woodlands, Heathland, Sand Heathland, Wet or Damp Forests, Lowland Forests, Coastal Headland Scrub, and Riparian Scrub/Swampy Riparian Woodland complex (Fig. 2). The highest percentage of swamp antechinus observations per ha EVC occurred in Sand Heathland (55%) followed by Coastal Tussock Grassland (10%) (Fig. 2).

Fig. 2.

Historic observation records of Swamp antechinus (1950s to 2020) in Ecological Vegetation Classes (EVCs) across the Otway Ranges.

The distribution records of swamp antechinus in Heathy Woodland and Heathland were predominantly in the east Otways, particularly Anglesea heath, and the west Otways in Carlisle heath. The species also occurred in Coastal Headland Scrub, Coastal Dune Scrub/Dune Grass mosaic and Sand Heathland along the coast in the eastern Otways and in Riparian Scrub/Swampy Riparian Woodland complex in Carlisle heath. The records in Wet Forest and Shrubby Wet Forest were largely across the central south Otways, north of Kennett River to Lavers Hill and central north Otways south of Forrest. The species was predominantly recorded in Coastal Dune Scrub, Coastal Tussock Grassland and Damp Sand Herb-rich Woodlands in the southwest near Princetown and Cape Otway. The swamp antechinus was also located in Riparian Scrub/Swampy Riparian Woodland complex.

Survey results for assessment of current distribution of the swamp antechinus (2018–2023)

Swamp antechinuses were recorded during camera surveys at 19 sites (15.1%), 10 sites in the eastern Otways, five in the central Otways and four in the western Otways (Table 1, Supplementary Table S1). Live survey trapping resulted in the recording of the species at 12 sites (17.1%) (Table 1), two in the eastern Otways (one inland and one at Painkalac Creek), and five in the central Otways (two inland at Carlisle Heath and three on coastal dunes at Cape Otway and Skenes Creek). In the western Otways, the species was trapped at five sites (four in coastal heath at Port Campbell National Park, and one in dunes at Princetown) (Table 1).

The species was located at 141 unique trap stations (individual live and camera traps) over the study period, including the survey results and population abundance assessment in the Carlisle heath refuge (Fig. 3). The largest percentage (36%) of locations was recorded inland in the central Otways in Heathy Woodland and Heathland. In the far eastern Otways, there were records at only three inland locations: two were in unburnt gullies (Heathy Woodland and Heathland) following fuel management burns (Watchorn et al. 2024), and one was located in Estuarine Woodland (Painkalac Creek) (Fig. 3). Further to the west, the species was recorded inland in Shrubby Wet Forest and Shrubby Foothills Forest in the confluence of several creeks.

Fig. 3.

Unique trap station records for Swamp antechinus from camera and live trapping surveys (2018–2023).

The swamp antechinus occurred in Coastal Headland Scrub, Coastal Dune Scrub/Dune Grass mosaic and Sand Heathland along the coast in the eastern Otways between Jan Juc and Spout Creek, in the central Otways (Grey River, Skenes Creek), in the western Otways (Aire River, Johanna River), and Cape Otway where a high percentage of locations (15%) were recorded. In the western Otways (Port Campbell), the species was recorded in Coastal Headland Scrub and Damp Heath Scrub (Fig. 3). Our re-surveys (camera and trapping) of known sites occupied by the swamp antechinus in the eastern Otways resulted in detection in Estuarine Woodland at Painkalac Creek where the species was last recorded in 2002; at Hutt Gully coastal dunes and Hutt Gully, last recorded in 2015; at Urquharts Bluff coastal headlands, last recorded in 2007; and inland at Edwards Creek Track Gully, last recorded in 2018 (Table 2, Supplementary Table S3). Success for live trapping ranged from 1.5 to 35.0 (mean = 9.1). Maximum success was recorded at Painkalac (35.0), Carlisle (21.5), Cape Otway, Port Campbell National Park, Gibsons Steps (18.8) and Parker Inlet (12.5) (Table 2, Supplementary Table S3).

Characteristics of refuges across the Otways for the swamp antechinus

There were 11 micro-refuges of between 39 and 339 ha, three mid-connected refuges of between 718 and 942 ha, and three macro-refuges of 1170, 4395 and 4859 ha, respectively, identified, mapped and assessed (Fig. 4, Supplementary Table S4).

Fig. 4.

Refuges identified for swamp antechinus in Otways (2018–2023). Numbers are refuges micro (1, 2, 4–9, 17), mid connected (3, 10, 13) and macro (14–16).

Micro-refuges were predominantly located in, or partially in, coastal dunes from Anglesea in the east to Johanna in the west (e.g. 1, 2, 5, 7, 8, 9, 12) (Fig. 4, Supplementary Table S4). The vegetation is predominantly Coastal Scrub and Estuarine Woodland which is characteristically long unburnt (since 1983), with structurally complex vegetation, high moisture and nutrients (Wilson, B. A., Garkaklis, M. J., Arnall, S., and Doherty, T. unpubl. data). The refuge at Aire River, Hordern Valley, is fragmented and surrounded by agriculture. There were few inland micro-refuges, and these were predominantly in Heathy Woodlands, Heathy Lowland Forest with several located in unburnt gullies subjected to fuel management burns, and another in a deep dissected gully.

Mid-connected refuges included the inland headwaters of creeks (Moggs, Coalmine, Spout, Grassy and Reedy) that provide connectivity between the creek systems and the coastal dune systems (Fig. 4, Supplementary Table S4). The vegetation is predominantly Shrubby Wet Forest, Shrubby Foothills Forest and Heathy Lowland Forest. Another mid-connected refuge provides connectivity along the coast between Cape Otway and Marengo and is predominantly Coastal Scrub/Grassland Mosaic and Lowland Forest, another, the Porcupine Reference area, is in Heathy Woodland/Lowland Forest (Fig. 4, Supplementary Table S4).

Macro-refuges such as Carlisle Heath include areas of poorly drained, Wet Heathlands (micro-refuges), interconnected by draining gullies and surrounded by Heathy Woodland (Fig. 4, Supplementary Table S4). The Port Campbell macro-refuge occurs on Coastal Headland interspersed with Damp Heath Scrub in deeper gullies that flow coastally from inland (Fig. 4, Supplementary Table S4).

Significant threats to refuges

An assessment of disturbances and threats occurring at each swamp antechinus refuge recognised erosion vulnerability due to sea level rise and tidal surges as of medium to very high risk in coastal micro-refuges (Supplementary Table S5). Further threats include habitat fragmentation, fire, weeds (coast wattle, Acacia sophorae, coastal tea-tree, Leptospermum laevigatum and sweet hakea, Hakea drupacea), foxes, cats, rabbits and recreation, including threats from dogs (especially off-leash dogs) and mountain bikes. The assessment found that inland micro-refuges are threatened predominantly by inappropriate fire regimes, the impacts of Phytophthora dieback, and the presence of foxes (Supplementary Table S5), and inland mid-connected refuges are threatened by wildfire, inappropriate fuel reduction burning and, where susceptible Heathy Woodland habitat is present, the impacts of Phytophthora dieback (Supplementary Table S5). Threatening processes in the macro-refuges are predominantly inappropriate fire regimes and the impacts of Phytophthora dieback, which contribute to severe habitat loss, while the presence of foxes under these conditions are considered likely to severely impact mammal prey (Supplementary Table S5).

Trapping results for the swamp antechinus at the Carlisle Heath macro-refuge (2021–2023)

In 2021, the swamp antechinus was the most frequently trapped small mammal species in the Carlisle Heath (152 captures) and was detected at 28 of the 40 trap sites with total trap success ranging from 1.2 to 33.4 and averaging 5.2% (Fig. 5). The total number of swamp antechinus individuals captured was 74 and ranged from one to eight animals at surveyed sites. The trap success was 10-times greater in long post-fire aged habitat (>20 years) and six-times greater in medium post-fire aged habitat (10–20 years) compared with early post-fire age (0–9 years) (Fig. 6). When live trapping was repeated (2022–2023) at a selection of the 40 sites trapped in 2021, no swamp antechinuses were recorded.

Fig. 5.

Trap success for swamp antechinus at sites across the Carlisle Heath (2021).

Fig. 6.

Trap success of swamp Antechinus in post fire age across the Carlisle Heath (2021). Trap night survey effort 0–9 years (864), 10–20 years (960) and >20 years (1028).

Refuge characteristics and ability to sustain populations

The findings of this investigation have highlighted the variability in refuges supporting the swamp antechinus, including size, shape, connectivity and habitat. Some native mammal species rely upon shifting refuges that are variable in space and move across the landscape through stochastic processes, such as rainfall or fire (Newsome et al. 1975; Pavey et al. 2017). In contrast, the swamp antechinus refuges identified here can be classified as fixed refuges as they occur in predictable locations that are used over time (Reside et al. 2014). Fixed refuges can arise through patchiness and be products of topographic complexity, including mountain ranges, rocky gorges, boulder piles, gullies or slopes (Reside et al. 2014; McDonald et al. 2015). They may be regions of reliable water (e.g. riparian zones, persistent waterholes or drainage lines), that have high vegetation productivity and support more stable animal assemblages during drought (Nimmo et al. 2015; Selwood et al. 2015, 2019; Selwood and Zimmer 2020; White et al. 2022). The swamp antechinus refuges identified across the Otways include similar mesic locations, including riparian zones, gullies and dune areas that have high structural vegetation complexity, moisture and nutrient levels (Wilson, B. A., Garkaklis, M. J., Arnall, S., and Doherty, T. unpubl. data). They range from linear micro-refuges of coastal dunes to interconnecting mid-connected refuges providing habitat connection features and larger macro-refuges of interconnected key habitats that are very likely to support large, self-sustaining populations of the species. The vegetation communities identified within the refuges are characteristically long unburnt and structurally complex, offering protection from predators by increasing shelter for swamp antechinus and decreasing predator activity and hunting efficiency (Doherty et al. 2022; von Takach et al. 2022).

The results provide evidence of survival of swamp antechinuses in dune systems over a 30-year period in the eastern Otways (Wilson et al. 1990, 2017; Wilson and Garkaklis 2020) and Cape Otway (Wilson 1983; J. White and A. Rendall pers. comm.), highlighting the resilience of these locations as refuge sites. Despite the presence of foxes and cats in these refuges, there is evidence that in areas where habitat structural diversity and complexity is maintained, small mammal communities, including swamp antechinus, are resilient (Wilson et al. 2017; Wilson and Garkaklis 2020). Allochthonous inputs such as marine wrack and seabird guano have been linked to increased nutrients and high food availability (e.g. invertebrates) in other similar coastal sites. Such inputs support dense vegetation cover, increased reproductive success and higher abundance including swamp antechinuses on coastal islands (Wolfe et al. 2004; Bancroft et al. 2005; Sale et al. 2006, 2008; Sale and Arnould 2012).

Many of the coastal dune micro-refuges have been identified as of very high erosion vulnerability and are highly unlikely to support mammals in the near future. Thus, there is a need to plan for rescue of threatened mammals, such as the swamp antechinus, prior to habitat loss and population extinctions, and undertake translocations to more suitable and sustainable habitats. The identification of Painkalac Creek as a micro-refuge for the swamp antechinus was an important result as monitoring had failed to locate the species here since 2002 (Wilson et al. 2001, 2017). Further, more recent surveys have recorded swamp antechinus 3 km upstream, confirming the significance of this site as a ‘source population’.

The swamp antechinus was recorded in a number of micro-refuges located inland in gullies, that connect with refuge sites within the coastal dunes. The results are significant because they provide evidence that the species continues to use the habitat on both sides of the Great Ocean Road. While the road is considered a significant barrier to movements of the species, many do have water management culverts and drains under the road. The utilisation of these by the swamp antechinus needs to be assessed to determine their use, and the potential to improve the connectivity of populations by provision of cover and refuge structures within the underpasses (Soanes et al. 2015; Goldingay et al. 2019).

The population of swamp antechinus in the Carlisle Heath no longer appears to be at high density, at least at the sites trapped in 2022–2023, indicating the need for caution on relying on this macro-refuge as a constant source of the species. The dramatic decline of the species in such a short timeframe and in a period of high rainfall is concerning. Such declines likely impact the genetic diversity of the populations, making them prone to inbreeding effects in the future, and less resilient to a changing climate. Threatening processes acting on the species in this location include incompatible fire regimes, introduced predators, changes in vegetation structure due to Phytophthora, and changes in rainfall patterns with climate change. Further investigation into the direct and indirect effects of fuel reduction burns and associated changes is required as limited data are available.

Management implications for swamp antechinus refuges in a drying climate

A key question in conservation management remains: once refuges for key species are identified, how do we use that knowledge to benefit their survival? The key threats identified for the Otways refuges were coastal erosion vulnerability, wildfire, inappropriate fuel reduction burning, the impacts of Phytophthora dieback, weeds, foxes and cats. Appropriate management to mitigate these threats will increase the survival and resilience of the species. However, such threats need to be consistently reassessed in the different refuges as they will likely change with altered conditions. For example, in La Nina years, the impacts of Phytophthora dieback and growth of weeds are much greater, and in El Nino years, wildfire risk is much higher.

There is an important need to refine the risk assessment of swamp antechinus in vulnerable coastal dunes and consider management intervention including timely translocations of swamp antechinuses to more stable refuges to prevent their extirpation and aid genetic management. Previous trial translocations of the swamp antechinus have been conducted and, recently, an evaluation has been completed of the feasibility of translocating the species within the Otway Ranges, to improve population viability based on genetic diversity (Aberton 1996; Wilson et al. 2017; Cripps and Parrott 2023). This work provides strong guidance for translocation of the species to avoid imminent loss of animals and important genetic diversity remaining in coastal dune refuges.

The need to implement appropriate fire management for wildfire and fuel reduction burning for this late successional species has long been identified (Wilson, B. A., Burns, P., and Garkaklis, M. J. unpubl. data; Threatened Species Scientific Committee 2016; Woinarski et al. 2014; Wilson et al. 2017). Strong protection of refuges from wildfire and fuel reduction burning is imperative. Detailed spatial fire data is required to inform optimum burning regimes, ensure burning is undertaken at a suitable time (e.g. outside times when young antechinus are in nests), and to maintain suitable successional aged habitat for survival, dispersal and recolonisation of the species (Bradstock et al. 2005; Clarke 2008; MacHunter et al. 2009; Driscoll et al. 2010). The fuel and ecological fire management strategy for the Otways region identifies the level of activity to reduce fuel hazard levels and/or optimise ecological outcomes and broadly prioritises minimising impacts to human settlements over the environment and wildlife (Department of Environment and Primary Industry 2014; Gazzard et al. 2020; The State of Victoria Department of Environment, Land, Water and Planning 2020). Mitigation of impacts on the swamp antechinus includes operational controls that reduce planned burn coverage to maintain refuge areas (The State of Victoria Department of Environment, Land, Water and Planning 2020). This has recently been implemented in practice with burning conducted to protect some of the refuges we have identified.

The plant pathogen Phytophthora cinnamomi has been identified as a major threat to mammals due to long-term vegetation degradation resulting in loss of susceptible plant species and simplification of vegetation structure, particularly in damp heathlands, and heathy woodlands (Laidlaw and Wilson 2006; Annett 2008; Wilson et al 2024). The swamp antechinus is at significant risk due to loss of susceptible keystone plant species (e.g. Xanthorrhoea australis) that contribute to complex vegetation structure and associated food resources (invertebrates, fungi) Woinarski et al. 2014; Threatened Species Scientific Committee 2016; Wilson et al 2024). Recent work has identified priority areas to protect high biodiversity values, including the presence of swamp antechinus, from Phytophthora dieback across the Otways and has developed management strategies for dealing with the threat and plans to contain the pathogen (Walshe, T., Wilson, B. A., Bridges, D., and Garkaklis, M. J. unpubl. data ; Wilson and Garkaklis (2023). These include increased hygiene procedures (e.g. disinfecting equipment, shoes, vehicles), and the application of phosphite (hand and aerial spraying) to reduce spread of the pathogen and enhance vegetation survival (Shearer and Tippett 1989; Hardy et al. 2001; Barrett and Rathbone 2018; Commonwealth of Australia 2018).

While there is little direct evidence of the impacts of introduced predators on the swamp antechinus, there is evidence of fox and cat predation on other mammals in the eastern Otways (Hutchings 1996, 2003; Wilson and Wolrige 2000; Wilson et al. 2001; Hradsky et al. 2017; Rees et al. 2023). Although fox baiting is regularly undertaken by land managers (Parks Victoria 2015), the effectiveness of the current program is likely low depending on the duration and intensity of baiting (Robley et al. 2019; Rees et al. 2023). There is little landscape-targeted cat control in the region. Recent management of foxes and cats has, however, been implemented in and around the coastal and estuary swamp antechinus refuges we have identified (Great Ocean Road Coast and Parks Authority 2022; Parks Victoria pers. comm., Surf Coast Shire pers. comm., M. Loughnan pers. comm.).

Further management is required to control invasive weeds, such as coast wattle and coastal tea-tree which outcompete all understorey plant species resulting in severely depauperate mammal habitats (Mitchell 2007). In coastal micro-refuges located close to human habitation and recreation, the presence of dogs, cats, human activities and littering are also threats to the swamp antechinus, requiring community education on responsible pet ownership, habitat protection areas and litter removal programs to assist the survival of the swamp antechinus.

The swamp antechinus is affected by multiple, likely interacting disturbances, that can accelerate population declines through additive or synergistic effects (Doherty et al. 2015). Fire can increase the activity of foxes and cats by removing vegetation which normally provides prey with refuge (McGregor et al. 2014, 2015; Leahy et al. 2015; Hradsky et al. 2017; Miritis et al.2023). There is evidence that increased predator activity is more likely when areas were burnt more recently (Doherty et al. 2023). For the swamp antechinus, there may be a critical period immediately post-fire when it is most vulnerable to an elevated risk of predation and within which management interventions are likely to be most impactful. The impacts of severe habitat loss and cover for swamp antechinus due to Phytophthora dieback is also likely to result in increased activity of foxes and cats. Integrated management of predators and fire is strongly recommended (Doherty et al. 2015, 2023) and should be paired with the application of phosphite to reduce spread of the pathogen and enhance the survival of vegetation.

It is recommended that refuges are monitored regularly (2–4 years) to confirm loss or recovery of populations and that the monitoring be linked to an assessment of management effectiveness, to aid adaptive management (Threatened Species Scientific Committee 2016; Wilson et al. 2017). Climate change will continue to have major impacts upon landscapes and species, and identifying critical refuges and protecting them to maintain and increase populations is essential.