Discussion

Cats and foxes

We predicted that cat and fox activity would be higher in areas burnt by the prescribed fire, however, we did not find any evidence in support of this. Whilst some studies have found strong evidence of cat or fox activity increasing after fire (McGregor et al. 2016a; Miritiset al. 2023), others have found a negative response (Bliege Birdet al. 2018; Lothian et al. 2022), no response (Hradskyet al. 2017c; Moore et al. 2018; Senior et al.2022), or a context-dependent response. For instance, in the USA, swift foxes (V. velox ), which are closely related to red foxes, only used burnt areas more frequently if their existing core home ranges were burnt (Thompson et al. 2008). Indeed, a recent analysis of existing evidence found that there was a high likelihood of neutral responses to fire being recorded for both cats and foxes in Australia (55% and 67%, respectively; Doherty et al. [2023]). One explanation for our result could be that the prescribed fire in this study was too mild to illicit an increase in cat or fox activity. A GPS study of feral cats in northern Australia found that, whilst cats strongly selected for areas recently burnt by severe fires, they did not select for areas recently burnt by mild fire—despite these areas having a high abundance of small mammal prey (McGregor et al.2014). The authors suggested that cats did not benefit from mildly burnt areas due to unburnt patches likely providing refuge for prey, making hunting less profitable compared to severely burnt areas (McGregoret al. 2014).
While cats did not increase their activity following the prescribed fire, they were more active in areas with a shorter time since fire and lower NDVI—both indicators of simpler vegetation structure (Haslemet al. 2016). Cats also favoured swampy riparian woodlands, which is typically the most structurally complex vegetation type in our study area. The seemingly conflicting habitat complexity preferences may be explained by resource availability. Feral cats are capable of exploiting a diverse range of habitats, and it is generally considered that they prefer dense habitats for shelter—such as riparian woodlands—and more open habitats for hunting (Lozano et al. 2003; Dohertyet al. 2015; Stobo-Wilson et al. 2020). Whilst foraging efficiency may theoretically be highest in recently burnt areas, these areas may be sub-optimal, and thus avoided, if prey availability is low (Pyke et al. 1977). The small mammal population in the eastern Otway Ranges has been declining for several decades and is now depauperate (Wayne et al., 2017; Wilson & Aberton, 2006; Wilson et al., 2001; Wilson et al., 2017; Wilson & Garkaklis, 2020). While we did not find a relationship between fire and small mammal activity (discussed below under Small and medium-sized mammals ), the most frequently detected small mammal species, the bush rat, almost completely avoids recently burnt areas (Lees et al. 2022). Further, during a separate, concurrent small mammal study in the eastern Otways, we found that detections of other small mammal species in burnt areas were very uncommon (Watchorn, unpublished data). These factors may partially explain why neither cats nor foxes increased their activity in burnt areas, and why cat activity was higher in relatively open areas that likely supported higher prey availability.
We did, however, find that fox activity increased across the study area after the fire, although due to the proximity of our control and treatment sites, we cannot be sure if the prescribed fire drove this change. Juvenile foxes typically disperse in winter (April to June in this region; Baker et al. 2001), and Hradsky et al. (2017) also observed an increase in fox activity at their control sites following a prescribed fire, which was thought to be driven by dispersing foxes. Spatially independent control sites would have better isolated the impact of the fire from other factors like seasonal fox dispersal, however, equipment limitations made such an approach unfeasible for this study.
We predicted that cat and fox activity would be higher near farms and towns due to increased resource availability (Hradsky et al.2017c). We found some support for this prediction, with foxes more active closer to farmland. Across their global range, red foxes, as well as other generalist mesopredators such as coyotes (Canis latrans ) and golden jackals (C. aureus ), use farmlands due to the availability of denning habitat, water and food (e.g., livestock, rodents, human food waste), as well as edge habitats which can support relatively diverse fauna communities and hunting opportunities (Gosselink et al. 2003; Šálek et al. 2014; Laux et al. 2022; Aikawa and Saito 2023). Indeed, we found that the eastern grey kangaroo, a common prey item of the red fox (Stobo-Wilson et al. 2022), was also more active near farmlands—likely due to the availability of pasture for grazing (Arnold et al. 1992; Maguireet al. 2006). In a GPS tracking study, Hradsky et al.(2017c) found that foxes in the Otway Ranges selected for farmland and forest-farmland interfaces at night, possibly due to livestock carcasses or prey availability. Collectively, these findings indicate that farm peripheries could provide effective fox baiting targets for land managers (Engeman and Linnell 1998; Carter and Luck 2013). Further high-resolution telemetry and resource mapping studies may provide further insight into the specific features foxes use (e.g., water dams, livestock carcasses etc.,), which may further improve bait uptake in these areas.

Macropods

Fire stimulates a short-term increase in plant nutrients, such as nitrogen (N), increasing both the nutritional availability and palatability for herbivores (Eby et al. 2014). This, in turn, drives temporary increases in the abundance of large herbivores after fire—a phenomenon observed around the world (Klop et al. 2007; Raynor et al. 2015; Reid et al. 2023). As such, we expected eastern grey kangaroos and swamp wallabies to increase their activity in recently burnt areas. Swamp wallaby activity was higher in areas with a shorter time since fire, a result consistent with studies showing this species’ preference for early successional vegetation and minimal short-term impacts of fire (Styger et al. 2011; Swanet al. 2021; Chard et al. 2022). However, we found no evidence that macropod activity increased at sites burnt by the prescribed fire or with Fire extent. Rather, swamp wallaby activity decreased following the prescribed burn. One explanation for this may be that the willingness of macropods to forage in recently burnt areas can vary with predator abundance. For example, Banks (2001) found that eastern grey kangaroos were more active in open areas were foxes had been removed, but spent more time near forest edges when foxes were present. Di Stefano et al. (2009) also found that swamp wallabies favour more structurally complex vegetation to increase obscurity from predators, whilst other large herbivores, such as the white-tailed deer and caribou (Rangifer tarandus ) in North America, also reduce their use of open areas—despite an abundance of food in those areas—to reduce the risk of predation (Joly et al. 2003; Cherryet al. 2017).

Small and medium-sized mammals

The small and medium-sized mammals in our study area typically prefer dense vegetation, where food and denning resources are greater and the risk of predation is likely lower (Catling et al. 2001; Dexteret al. 2011). We therefore predicted that the activity of these groups will be positively associated with NDVI and decrease after the prescribed fire (White et al. 2022). In support of this prediction, medium-sized mammal activity decreased with Fire extent, a trend also seen with medium-sized prey in Europe (Sokos et al.2016), North America (Zwolak and Foresman 2007), and elsewhere in Australia (Robley et al. 2023). Hradsky et al. (2017a) observed that fox consumption of medium-sized mammals, such as bandicoots, doubled after a patchy prescribed fire in the Otways, highlighting the increased vulnerability of these species to predation following fire, notwithstanding possible scavenging of carcasses.
Medium-sized mammals also showed a complex relationship with NDVI; while the main effect of NDVI was generally positive, activity was negatively associated with NDVI before the fire and slightly increased afterwards. This may suggest that these species increased their selection of microhabitat features associated with higher NDVI (e.g., gullies, sedges) following the fire (e.g., Swan et al. 2016; Fordyceet al. 2016; Lees et al. 2022). However, the high level of uncertainty with this interaction limits our ability to draw strong conclusions. Notably, while higher NDVI is typically associated with higher species richness for the mammals comprising this species group (Youngentob et al. 2015b; Dorph et al. 2021; Whiteet al. 2022), individually, these species have exhibited positive (Miritis et al. 2020; White et al. 2022), negative (Ralph 2021), and neutral (Youngentob et al. 2015b; Hale et al.2016) responses to NDVI. These varied relationships are likely due to factors such as drought, fire history, and habitat type (Youngentobet al. 2015a; White et al. 2022). Nonetheless, our findings suggest that NDVI may be a useful means of identifying and conserving productive and structurally complex areas which may facilitate mammal occurrence and diversity in this landscape (Sukmaet al. 2019; Rivarola 2022), especially following disturbances such as drought (White et al. 2022), plant pathogen spread (Casey 2022), or fire (Dorph et al. 2021).
None of the variables predicted the activity of small mammals, a finding at odds with previous studies on these species that identified relationships with NDVI (Hale et al. 2016; Chadwick et al.2022) and changes in habitat use following fire (Swan et al.2016; Fordyce et al. 2016; Lees et al. 2022). One reason for this discrepancy may be that our camera arrangement, in terms of both density and spatial extent, was insufficient to reliably detect discrete patterns of microhabitat use for these species, especially considering the depauperate state of the small mammal community (Wilson and Garkaklis 2020). The incorporation of fine-scale, site-level habitat information—such as understorey vegetation structure (e.g., Hradskyet al. 2017a; Lees et al. 2022)—may have further improved our ability to detect any potential response to fire or vegetation structure.

Conclusions

We found little evidence to support our predictions that cat, fox, and macropod activity would increase following the prescribed burn, or that small mammal (<800 g) activity would decrease. Medium-sized mammals were negatively associated with Fire extent, and other studies in the region have identified similar trends following patchy prescribed burns, along with increased consumption of these species by foxes (Hradsky et al. 2017a). This suggests that these species may be particularly vulnerable to fire and invasive predators in this region, and may benefit from preserving large unburnt refuges and from targeted post-fire interventions to reduce predation, such as the provision of artificial refuges (Watchorn et al. 2022).
Looking forward, there are several key research questions which should be prioritised to conserve fauna inside protected areas. The majority of evidence demonstrating increases in cat activity post-fire comes from Australia’s tropical north (Doherty et al. 2023). It remains unclear whether cats or foxes temporarily adjust their core home range or move long distances to hunt in recently burnt areas in the temperate forests, woodlands, and other ecosystems of southern Australia. Whilst camera traps are useful, their appropriateness for answering this question is limited by their inherently patchy spatial coverage across the landscape, as well as the delay associated with deploying them immediately after prescribed fires, especially in temperate forests. To this end, we recommend future studies employ the BACI experimental framework with animal-borne GPS loggers (Le Pla et al. 2023). This will improve our understanding of the fine-scale movements and habitat use of cats, foxes, and native prey species immediately before and after prescribed fire (e.g., McGregor et al. 2014). Animal-born video cameras could also provide valuable insight into cat and fox hunting success and prey selection (McGregor et al.2015). These approaches, whilst challenging to implement, will allow for a more nuanced understanding of fine-scale responses to prescribed fire, thereby informing more effective management strategies within protected areas.