Feral cat predation of the threatened Pilbara leaf-nosed bat – a key threatening process

Introduction

Feral cats (Felis catus) are recognised as a significant predation threat to wildlife and have contributed to severe declines and extinctions in native birds, mammals and reptiles across Australia (Lepczyk et al. 2004; Medina et al. 2011; Doherty et al. 2017; Legge et al. 2017; Woolley et al. 2019). Predation by feral cats is recognised as a key threatening process under the Australian Commonwealth Government’s Environment Protection and Biodiversity Conservation Act 1999 (EPBC Act) and threatens the survival of over 100 native species in Australia (Department of Climate Change, Energy, the Environment and Water 2023a). According to the Western Australian Department of Biodiversity, Conservation and Attractions (2023), six mammal, 11 reptile and 22 bird species are threatened by feral cat predation. Feral cats established alongside the colonisation of Australia by Europeans, and by 1900 they were present throughout most of Australia and many offshore islands (Abbott 2002). The exclusion and removal of cats is a major consideration for conservation of small to medium-sized vertebrates on a landscape scale, and especially for threatened species rescue programs that rely on protecting small, vulnerable populations in limited areas such as fenced properties and islands (e.g. Algar et al. 2007, 2010; Moseby et al. 2009; Campbell et al. 2011).

The Pilbara leaf-nosed bat (Rhinonicteris aurantia, Pilbara form; synonym Pilbara Diamond-faced Bat; Armstrong et al. 2016) is listed as Vulnerable under the EPBC Act. It is one of two population isolates of the orange leaf-nosed bat that have divergent morphology and echolocation call frequency characteristics (Armstrong 2002, 2005; Armstrong and Coles 2007) and is distributed throughout the Pilbara region of Western Australia. Its distribution is limited by its poor physiological capacity to manage water and heat balance (Kulzer et al. 1970; Baudinette et al. 2000) and the scarcity of underground roosts with suitably warm and humid roost microclimates (28–32°C and 85–100% relative humidity; Armstrong 2000, 2001) within natural caves and disused underground mines. Suitable natural caves within the ironstone terrain of the Hamersley Range and the margins of the Fortescue Valley coincide with a substantial focus for active mining activities. It also relies on disused gold and copper mines in the eastern Pilbara (around Bamboo Creek, Marble Bar and Nullagine), some of which have disappeared or are at risk of roost collapse (Armstrong 2001; Cramer et al. 2016). The species was listed under the EPBC Act in 2001 following the growth of iron ore mining in the Pilbara, partly because of concerns for encroachment on roost habitat. The Pilbara leaf-nosed bat has since become a significant consideration and research focus in the mining approval process in the Pilbara region (Cramer et al. 2016).

On a global level, studies quantifying or describing the impacts of cats (including domestic, stray and feral cats) on bats are less represented in the literature than other fauna groups and non-volant mammals (Ancillotto et al. 2013; unpubl. data., G. Vázquez-Domínguez., R. A Saldaña-Vázquez, M. V Orduña-Villaseñor and J. Schondube; Welch and Leppanen 2017). However, according to a review by Oedin et al. (2021), 86 bat species have been reported to be preyed upon or threatened by cats, a quarter of which are listed as Threatened or Near Threatened in IUCN Red List assessments. Within Australia, 24 species of bat, including a record of R. aurantia from the Northern Territory, were reported as cat prey (Woolley et al. 2019). The same study also recognised that this number could be an underestimate given recent taxonomic changes and that further research would likely reveal additional records of bat species preyed upon by cats.

Cats prey upon bats in both urban and regional habitats, with impacts on cave-, house- and tree-roosting bat species. Documented examples include the great fruit-eating bat (Artibeus lituratus), pale spear-nosed bat (Phyllostomus discolor) and vampire bat (Desmodus rotundus) in Argentina and Brazil (Delpietro et al. 1994; da Costa-Pinto 2020), central lesser short-tailed bat (Mystacina tuberculata rhyacobia) and long-tailed bat (Chalinolobus tuberculatus) in New Zealand (Scrimgeour et al. 2012; Borkin et al. 2022), as well as several species in Mexico (unpubl. data., G. Vázquez-Domínguez., R. A Saldaña-Vázquez, M. V Orduña-Villaseñor and J. Schondube), the United Kingdom (Khayat et al. 2020), Portugal (Rocha 2015) and Italy (Ancillotto et al. 2013). The ability to fly clearly does not always allow bats to evade cat predation, and bats are generally considered more at risk when in their roosts.

Evidence of cat predation on bats in Australia comes from observations on the eastern blossom bat (Syconycteris australis) and southern bent-winged bat (Miniopterus orianae bassanii) in south-eastern Australia (Phillips et al. 2001; E. van Harten, pers. comm., 2023). Recent publications have recognised feral cats as a threat to other mammal species in the same Pilbara cave habitats as those used by R. aurantia (Rothschild’s rock-wallaby (Petrogale rothschildi), Anderson et al. 2021; and ghost bat (Macroderma gigas), Cramer et al. 2022). The tendency of cats to hunt at cave roosts with low and narrow entrances that are used by M. gigas (Bat Call WA 2021a) also has implications for R. aurantia because the two bat species have been recorded in many of the same roosts. Despite recent (Bat Call WA 2021b) and historical reviews of threats to the species (Cramer et al. 2016; Threatened Species Scientific Committee 2016), none have identified feral cats as posing a potential threat to R. aurantia. Consultant biologists who recorded cat predation of R. aurantia at known bat roosts raised concerns about this threat at the Pilbara leaf-nosed bat conservation workshop in 2022 because of casual observations made throughout several mining project areas.

In this study we consolidate and quantify evidence of feral cat predation on R. aurantia. Our intention was to compile evidence as the basis for: (1) formal recognition of feral cats as a threat to R. aurantia; and (2) highlighting the need for targeted control of cats at and nearby known R. aurantia roosts to reduce the current documented level of impact and the future threat from cat predation. Understanding how feral cat predation impacts Pilbara bat colonies will provide further perspective on the range and level of key threats to their populations that is important within the broader context of other cumulative threats to bats, such as habitat loss and indirect disturbances associated from mining and other human activities.

Methods

To address the lack of published data on the issue of feral cat predation on R. aurantia and to document unpublished records, we conducted an online survey requesting records of cat predation on Australian bat species, particularly R. aurantia and other cave obligate species. The following recipients were included in the request for information: Australasian Bat Society, Inc.; Ecological Consultants Association (chapters in New South Wales, Victoria and Western Australia); various consulting companies in Western Australia (Biologic Environmental Survey Pty Ltd, Biota Environmental Sciences Pty Ltd, Eco Logical Australia Pty Ltd, Ecologia Environment Pty Ltd, Spectrum Ecology Pty Ltd, Stantec Pty Ltd and Terrestrial Ecosystems Pty Ltd); and several mining companies operating in the Pilbara region, including Fortescue Ltd. Of the four respondees, two noted interactions between feral cats and R. aurantia. We collated recorded instances of predation by cats on R. aurantia in the Pilbara, as well as anecdotal findings related to cat presence in the vicinity of bat roosts.

Study site description

Studies conducted by GHD Pty Ltd and Biologic Environmental Survey Pty Ltd in the Pilbara region of Western Australia have produced two empirical datasets of cat predation on R. aurantia at three recognised roost sites. The roosts are located 110 km (Roost 1), 141 km (Roost 2) and 152 km (Roost 3) south of Port Hedland (Fig. 1). All three roosts have been monitored for several years using a combination of visual (infrared camera traps), acoustic (full spectrum ultrasonic recorders) and for two roosts, collection of scat and wings for DNA testing. Thus, the significance of these caves to R. aurantia is well understood. Roost types are categorised according to the definitions in the Conservation Advice for the species (Threatened Species Scientific Committee 2016).

Fig. 1.

Location of the three focal roosts in the eastern Pilbara, Western Australia.

Roost 1

Roost 1 has been identified as a natural ‘permanent diurnal roost (Priority 1)’ and has been subject to an intensive monitoring study conducted by GHD from 2017 to 2023. The three-chambered cave sits near the top of a west-facing cliff of banded iron formation at an elevation of approximately 368 m. The roosting area is located in the rear chamber of the cave that is accessed via passages through chambers 1 and 2. The narrow entrance to the rear roosting chamber is ~50 × 70 cm and is at ground level. The cave is also occupied by the common sheath-tailed bat (Taphozous georgianus), Finlayson’s cave bat (Vespadelus finlaysoni) and, occasionally, by the ghost bat (M. gigas). Cat predation has been recorded at the entrance of the second chamber, which has an opening of ~1.5 m high and 1 m wide.

Roost 2

Roost 2 is also considered a ‘permanent diurnal roost (Priority 1)’. It is positioned on the eastern-facing slope of a hill ~300 m above the surrounding plain. The cave has two chambers. The outer chamber has a low (~0.5 m high, 2 m wide) entrance at ground level. Predation of bats has been recorded at the entrance of this first chamber. The inner chamber is connected by a long (~5 m) but small (50 cm wide × 30 cm high) passage at ground level. R. aurantia roost in the second chamber. The roost is also occupied by T. georgianus and V. finlaysoni.

Roost 3

Roost 3 is also considered to be a ‘permanent diurnal roost (Priority 1)’ and is located at the base of north-western-facing mesa-like escarpment, near the top of small ironstone range. It comprises a single main chamber with five pillars and three vertical ‘pipeholes’ in the ceiling. R. aurantia roost near the entrance of the cave within one of these vertical pipeholes (30 cm × 30 cm) that is ~1.2 m off the ground, and ~5 m from the cave entrance. M. gigas are regularly recorded roosting in another section of the cave, as are T. georgianus and V. finlaysoni.

Results

The total number of bat kills observed in video recordings and photographic images at all three focal roosts was 152–169 individuals (Table 1). The direction of bat flight was often difficult to determine, but it appeared bats both leaving and entering the roost were captured. Two individual cats were observed at the entrance to Roost 1 on two consecutive nights (one on 28 April 2021 and one on 29 April 2021). On the first of those nights, the cat (Table 1; Cat A) appeared to capture and feed on one small bat, which was identified as R. aurantia from various characteristics. In addition, at least five ‘active hunting without capture’ events were observed.

At Roost 2, 116 bats were killed by two cats (distinguishable from their coat patterns): 76 by ‘Cat B’ in March 2022 and a total of 40 by ‘Cat C’ between December and May 2023. In most images, the cat was observed to be either ‘stalking’ at the cave entrance (n = 1822 images; 39% of observations) or ‘passive hunting’ (n = 2178 images; 46% of observations), with a further 9% of observations of ‘feeding’, 3% of ‘active hunting without capture’ and 3% of ‘active hunting with capture’ (representations of each cat behaviour in Fig. 2; example images of Cat B and Cat C with R. aurantia in Supplementary S2). The colony size of Roost 2 was estimated at between 2456 and 2534 in December 2022 (unpubl. data, K. N. Armstrong). Therefore, based on observational data, cats removed ~4.7% of individuals within the colony over the period of observations.

Fig. 2.

Summary of cat behaviour observations from observed photographic images at Roost 2 between March 2022 and May 2023 and Roost 3 between June 2020 and September 2020.

A single adult cat (Roost 3 Cat D) visited Roost 3 on 35 nights between 24 June 2020 and 11 September 2020 (Supplementary S2). The cat successfully captured bats on eight nights, with a total of 33–51 predation events. On six nights between late June to mid-July 2020, three kittens were also seen accompanying Cat D. The kittens made attempts to capture bats on the nights they were present. However, they were not successful until 19 July, when one of the kittens captured one bat. The kittens were not observed again after 19 July. The adult cat observed at Roost 3 (Cat D) displayed very similar hunting behaviour to the two cats observed at Roost 2. It was observed stalking in 53 images (8%), passively hunting in 190 images (28%) or actively hunting – no capture in 365 images (53%), with the remaining time spent largely recorded as actively hunting with capture (44 images, 6%) and feeding (37 images, 5%) (Fig. 2). A roost census using IR video and ultrasonic recording in 2016 indicated that the colony size of R. aurantia likely comprised between 780 and 975 individuals. Therefore, the adult and her kittens potentially impacted 3.4–6.5% of individuals within the colony over 4 months.

In the analysis using the Agena MassARRAY genetic markers to determine the number of individuals among a collection of 31 wings collected at Roost 2 in February 2022, there were 31 unique genotypes, indicating that each wing was from a separate individual. Pairwise similarity values amongst the genotypes ranged from 0.57 to 0.76, much less than the ideal value of 1.0 (and well below the established threshold of 0.9) that would indicate a match of all 78 loci. Thus, there is evidence for a scenario whereby a single cat had captured and consumed at least 31 individual bats over a relatively short period. Given the direct observations of capture described from photographs and videos, the genetics-based determination of the number of individuals killed underestimated the level of predation by cats at least three-fold, given that the total number of bat kills over the observation periods was 152–169 individuals.

Discussion

A relatively small set of observations (comprising 28 days of cat visitation, three roost sites, four individual adult cats and a total kill of between 183 and 200 individual bats) provides a strong basis for concern that feral cats are having, and could continue to have, a significant impact on the Vulnerable-listed Pilbara leaf-nosed bat (R. aurantia). Adding to the significance of these observations, each of the three roosts is recognised as critical habitat for R. aurantia in its Pilbara distribution. The cats exhibited clear targeted hunting behaviour at Roost 2 and Roost 3, which prompts concern for other colonies across the region, highlighting the need for ongoing and suitably designed feral cat management at these and other significant roosts of the species.

R. aurantia are likely to be vulnerable to cat predation because of the accessibility to cats of roosts and roost entrances (e.g. Roost 2’s single, small entrance at ground level), the colonial roosting habit of the bat species and a reliance on specific microclimatic conditions (high temperature and high humidity) that leaves the species with a limited choice of suitable roost sites. As exemplified by Roost 3, R. aurantia often roost in a constricted chamber, which increases the risk of predation if there is only a single narrow opening for their egress. Cats observed in this study at both Roost 2 and Roost 3 were clearly aware of bats flying into and out of the cave entrances and exploited this behaviour, making them a relatively ‘low-effort’ prey item. Furthermore, at Roost 3, a cat also appeared to use the bat colony to train its young. Within Roost 1, R. aurantia roost and breed within a chamber at the rear of the cave that is accessible by a single narrow ground-level opening. Although a cat has not yet been observed to date at this internal entrance or within the roosting chamber, there is certainly potential for a significant number of bats to be captured and killed, should it be discovered by cats.

In addition to the physical characteristics of the roost that were considered an important factor in cat hunting success in the present study, other published studies have demonstrated that the impacts of feral cats can be amplified by increased hunting effectiveness in particular landscapes, such as fire scars (McGregor et al. 2015; Hohnen et al. 2023). Post-fire conditions may have exacerbated the impact of Cat B at Roost 2 (76 observed bat predation events) – a substantial fire had burned surrounding vegetation just 2 months prior to the commencement of monitoring at that roost. Such observations highlight that combinations of factors might multiply both the risk and impacts of a key threat.

The possibility that individual cats specialise on particular prey types contributes further to the level of threat that cats pose to R. aurantia in the Pilbara. The significant number of bats taken from the colonies in this study during the recording periods (total observations: 1 bat at Roost 1; 116 bats at Roost 2; and 35–52 bats at Roost 3) are likely to be a small subset of the actual impact cats have had at these roosts, and may be indicative of the possible scale of predation throughout the region. The effect of prey specialisation by an individual cat has been recorded previously by Scrimgeour et al. (2012), who observed hunting of the New Zealand short-tailed bat (Mystacina tuberculata rhyocobia) at a roost tree until the cat was captured 7 days after its detection. In Italy, Ancillotto et al. (2013) demonstrated that once a roost is found by a cat, targeted hunting can result in a substantial number of mortalities in a short time. They also noted that roosts were easily located because of the broad frequency range of hearing in cats, which encompasses most bat echolocation and social calls (in temperate bat species), as well as the strong smell associated with roosts. In addition, the magnitude of threat that cats pose to bat colonies can vary depending on the characteristics of individual cats, such as their size (e.g. large males) and previous level of experience. Moseby et al. (2015) found that individual cats can have a disproportionate and catastrophic impact that devastates the efforts of reintroduction programs. Therefore, it is imperative to investigate and monitor the relatively small number of roost locations where Pilbara leaf-nosed bats are particularly vulnerable so that cats exhibiting prey specialisation behaviour can be neutralised.

Supplementary material

Supplementary material is available online.