Is the presence of a threatened arboreal mammal in residential areas related to remnant habitats?

Introduction

The expansion of urban areas is rapidly fragmenting natural ecosystems globally, replacing them with a mosaic of remnant vegetation and modified habitats. Within these landscapes, large remnants of natural habitat are considered critical to the sustainability of animal wildlife because they offer suitable resources and refuge from surrounding anthropogenic disturbances (Kowarik 2011). Although urbanisation has undoubtedly contributed to the decline of biodiversity worldwide (McKinney 2002), there is a growing body of evidence that some native species can occur not only in remnant habitat, but also in smaller and highly modified ‘green’ spaces (e.g. parklands, street vegetation and gardens) outside the reserve network and inside the city limits (McMaclagan et al. 2018; Soanes & Lentini 2019; Van Helden, Close, Stewart et al. 2020).

Cities can present wildlife with a mosaic of structurally complex and diverse habitats that contain novel resources (Aronson et al. 2017). Species with generalist habitat requirements and behaviourally flexible traits are most likely to utilise and prosper in highly modified urban habitats (Lowry et al. 2013) whereas threatened species, often with specific habitat requirements (Kani 2011), respond poorly to the disturbances associated with human-dominated areas (i.e. ‘the Human Disturbance Hypothesis’; Lawson et al. 2008). Despite this, threatened species also occur within highly modified urban habitats (Soanes & Lentini 2019; Van Helden, Close, Steven 2020; Van Helden, Close, Stewart et al. 2020), yet conservation strategies for urban wildlife continue to focus on the role remnant vegetation plays in maintaining their presence (see for example Soanes et al. 2019 and references therein). Understanding how wildlife can occur in these highly modified urban habitats and whether remnant vegetation plays a role in their persistence in these areas may contribute unexpected conservation opportunities for wildlife.

In this study, we explored whether the presence of a threatened arboreal marsupial within residential areas was influenced by distance to and size of nearby remnant vegetation. We chose the western ringtail possum (Pseudocheirus occidentalis) as a model of a Critically Endangered (Western Australian Biodiversity Conservation Act 2016) habitat specialist for which a dependence on remnant habitats is assumed (DPaW 2017), but never tested. We predicted that (i) possum presence was more likely in residential areas if remnant vegetation was accessible (i.e. within one home range of modified habitat), and (ii) in residential areas without accessible remnant vegetation, the presence of possums would be positively correlated with proximity to remnant vegetation. We explore the implications of our results for understanding species ecology in highly modified urban landscapes and to the development of conservation and urban planning strategies more broadly.

Methods

This study was conducted within the urban zone of the City of Albany, which supports a population of around 33 145 people (ABS 2016) in an area of approximately 90 km² (Fig. 1b). Over 20 large bushland remnants (~1 to ~600 ha) and multiple small patches of vegetation (<1 ha) scattered throughout the urban landscape are known to support western ringtail possums (Mathieson et al. 2020; Van Helden, Close, Stewart et al. 2020). The bushland remnants are typically dominated by peppermint (Agonis flexuosa), marri (Corymbia calophylla), eucalypt (jarrah (Eucalyptus marginanta), Albany blackbutt (Eucalyptus staerii), karri (Eucalyptus diversicolor) and Bullich (Eucalyptus megacarpa)), sheoak (Allocasuarina fraseriana) or coastal heath (Sandiford & Barret 2010; Van Helden et al. 2018; Bader et al. 2019; Mathieson et al. 2020) whereas the smaller vegetation patches comprise these species as well as a variety of non-native species.

The western ringtail possum, an arboreal marsupial weighing approximately 1 kg at maturity and once distributed over much of the southwest corner of Australia, is now largely restricted to three disjunct populations (Fig. 1a), two of which overlap with the urban centres of Albany (−35.02, 117.88) and Bunbury (−33.28, 115.73; see Van Helden, Close, Steven 2020). This possum occurs both in remnant natural habitats and highly modified habitats (road verges and residential gardens; Thompson & Thompson 2009; Van Helden, Close, Steven 2020; Van Helden, Close, Stewart et al. 2020) within urbanised landscapes.

Possum presence within the urban zone of Albany (Fig. 1b) was determined by spotlight surveying a total of 195, 100 m long transects, randomly positioned on the road network of urban Albany using the Quantum Geographic Information System (QGIS; version 3.4.4 with GRASS 7.4.4; QGIS Development Team 2009). We defined the ‘urban zone’ as any combination of landuse categories defined by the City of Albany as general industry, highway commercial, hotel/motel, light industry, local and neighbourhood centre, port industry, rail, regional centre (including mixed business and mixed use) and residential (special or tourist). Transects were separated by at least 100 m (approximately one home range width; see Van Helden et al. 2018) to maintain independence of neighbouring transects. Road-side and adjacent garden vegetation on both sides of each transect were spotlight-searched between June and July 2019 by two observers using 350 lumen head torches. Up to 15 transects were randomly selected to survey on any given night to reduce spatial, temporal or weather influence on the detection of possums across the entirety of the sampled area. No spotlight surveys were undertaken during rainfall, and all were conducted between 30 min after last light and midnight to maintain detection consistency among transects.

We used satellite imagery to identify and classify all remnant vegetation within urban Albany in QGIS as either large (>60 ha) or small (<20 ha, not including gardens) (i.e. no vegetation was between 20 and 60 ha in size; Fig. 1b). The area within 100 m from any point in all directions of each transect was defined (i.e. a 100 m buffer) in QGIS to represent the likely maximum area traversed by a possum (Van Helden et al. 2018). We identified whether remnant vegetation was present in the ‘buffer’ of each transect. For each transect with no remnant vegetation in the buffer, the shortest distance (m) to the closest large and small remnants was measured in RStudio (version 3.3.1 RStudio Team 2015).

To assess the potential influence of remnant vegetation on the presence of the possum, the proportion of transects on which possums were detected was compared for transects with and without remnant vegetation (of any size) inside the buffer using a proportion test in RStudio (version 3.3.1). A logistic regression in RStudio (version 3.3.1) was used to test whether the distance to small or large areas of remnant vegetation, or an interaction of these, influenced the presence of western ringtail possums on transects that contained no remnant vegetation within the 100 m buffer. We chose to separate large and small remnant vegetation in this analysis because of the differing roles they may play in sustaining wildlife in urban areas.

Results

Five transects (2.5%) included only large areas of remnant vegetation within the 100 m buffer, 109 transects (56%) included small areas of remnant vegetation, and 80 transects (41.5%) contained no remnant vegetation within the 100 m buffer. For the latter, the mean distance from each transect to the nearest large remnant was 470.8 ± 35.3 m SE, small remnant was 200.8 ± 12.0 m, and to remnant vegetation regardless of size was 183.9 ± 10.3 m.

A total of 48 possums were sighted on 33 transects (16.9% of sampled transects). Thirty-one of these animals were sighted on transects that contained remnant vegetation in the buffer, and 17 possums were detected on transects without remnant vegetation in the buffer. No possums were detected on the five transects that included large areas of remnant vegetation in the buffer. The proportion of transects on which possums were observed did not differ between transects with (21%, n = 24) and without (11%, n = 9) remnant vegetation within their buffer (χ = 2.543; P = 0.11). Neither the distance to small areas of remnant vegetation (P = 0.175), the distance to large areas of remnant vegetation (P = 0.118), nor an interaction of both these terms (P = 0.132) influenced the presence of western ringtail possums on transects with no vegetation within the 100 m buffer (Fig. 2).

Discussion

Although remnant patches of natural habitat are assumed critical for supporting threatened species with specific habitat requirements in anthropogenically modified landscapes, there is a growing body of evidence that some of these species can occur, persist and even prosper in these areas (e.g. McMaclagan et al. 2018; Soanes & Lentini 2019; Van Helden, Close, Steven 2020; Van Helden, Close, Stewart et al. 2020). Some of these species can achieve this with seemingly little dependence on remnant vegetation suggesting that other highly modified habitats may contribute to their presence in urban environments (e.g. McMaclagan et al. 2018). In this study, we tested the assumption that the presence of a highly threatened species in an urban environment would be dependent on remnant vegetation using a model species that was expected to exhibit some dependence on remnant habitat, but for which recent evidence of flexible resource use (Van Helden et al. 2018; Bader et al. 2019; Mathieson et al. 2020), and its widespread distribution in modified residential habitats (Van Helden, Close, Steven 2020; Van Helden, Close, Stewart et al. 2020) suggests otherwise. Our finding that the presence of the western ringtail possum was unrelated to remnant vegetation within the immediate surrounds, and the distance from remnant habitats (either large or small in size) promotes the opportunities highly modified habitats may present for wildlife conservation in urban landscapes (Soanes et al. 2019). Our findings contribute to a growing body of evidence that shows a diversity of animal taxa, including invertebrates, birds and reptiles, are capable of utilising ‘green spaces’, beyond the boundary of larger remnant reserves (see Soanes et al. 2019 for review) and exemplifies the unexpected and unrealised potential for novel urban habitats to support threatened species.

The finding that proximity to remnant vegetation (regardless of size) had no association with possum presence supports previous evidence of the same effect in residential gardens (Van Helden, Close, Steven 2020). Our results also support the contention that local mechanisms (e.g. vegetation cover, fruiting trees and presence of domestic pets) may influence the contemporary distribution of this species (Van Helden, Close, Steven 2020) and other taxa (e.g. Oprea et al. 2009) in modified habitats. This species’ flexibility in resource use (Van Helden et al. 2018; Bader et al. 2019) such as a diverse diet (Mathieson et al. 2020) and the ability to make use of artificial structures for resting sites (i.e. roof cavities) and movement pathways (i.e. fences; Thompson & Thompson 2009; Van Helden et al. 2018, in press) likely contributes to its apparent ability to access and utilise an abundance of novel resources that are available in modified urban habitats.

While this study provides preliminary evidence that the contemporary presence of possums in modified habitats is unrelated to the proximity of remnant vegetation, our sample size (i.e. a few transects with possums present) and approach (no transects in remnant natural habitat) constrain a direct comparison of the relative importance of natural and modified habitats across the urban landscape. It is possible that the distribution of this species may be influenced by processes operating at longer time scales (e.g. recruitment, dispersal and colonisation) and that the presence of remnant habitats may be critical for these processes. We also acknowledge that the presence of animals in modified habitats is not indicative of long-term viability of populations and may actually impart some conservation risk, such as acting as ecological traps (McMaclagan et al. 2018). Investigation of metrics such as residency, survival, reproduction and body condition of animals within modified environments will further our understanding of the potential value these novel habitats represent for wildlife conservation.

The results of this study contribute to a growing body of evidence that some taxa, in some circumstances, are capable of occurring in highly modified habitats outside of conservation estates and reserve networks, but within the city limits with seemingly little dependence on remnant vegetation. Our results promote the suggestion that highly modified habitats in urban landscapes can contribute conservation outcomes (Aronson et al. 2017; Soanes et al. 2019). Consideration of these highly modified spaces and their potential role in wildlife conservation warrants serious consideration, and they should not be overlooked in the development of wildlife management strategies in urban landscapes.

Author contributions

Marie B Busschots: Data curation (lead); formal analysis (supporting); investigation (lead); methodology (equal); validation (lead); visualization (equal); writing-original draft (equal); writing-review & editing (supporting). Paul Close: Conceptualization (equal); investigation (supporting); methodology (equal); project administration (lead); supervision (equal); validation (supporting); visualization (supporting); writing-original draft (lead); writing-review & editing (lead). Bronte Van Helden: Conceptualization (equal); data curation (supporting); formal analysis (equal); investigation (supporting); methodology (equal); supervision (equal); validation (equal); visualization (equal); writing-original draft (supporting); writing-review & editing (equal). Peter Speldewinde: Conceptualization (supporting); data curation (supporting); formal analysis (supporting); methodology (supporting); supervision (equal); validation (supporting); writing-review & editing (supporting).