The Western Ringtail Possum (Pseudocheirus Occidentalis)

A major road and an artificial waterway are barriers to the rapidly declining western ringtail possum, Pseudocheirus occidentalis

Kaori Yokochi BSc. (Hons.)

This thesis is presented for the degree of Doctor of Philosophy of The University of Western Australia

School of Animal Biology Faculty of Science

October 2015

Abstract

Roads are known to pose negative impacts on wildlife by causing direct mortality, habitat destruction and habitat fragmentation. Other kinds of artificial linear structures, such as railways, powerline corridors and artificial waterways, have the potential to cause similar negative impacts. However, their impacts have been rarely studied, especially on arboreal species even though these animals are thought to be highly vulnerable to the effects of habitat fragmentation due to their fidelity to canopies. In this thesis, I studied the effects of a major road and an artificial waterway on movements and genetics of an endangered arboreal species, the western ringtail possum (Pseudocheirus occidentalis). Despite their endangered status and recent dramatic decline, not a lot is known about this species mainly because of the difficulties in capturing them. Using a specially designed dart gun, I captured and radio tracked possums over three consecutive years to study their movement and survival along Caves Road and an artificial waterway near Busselton, Western Australia. I studied the home ranges, dispersal pattern, genetic diversity and survival, and performed population viability analyses on a population with one of the highest known densities of P. occidentalis. I also carried out simulations to investigate the consequences of removing the main causes of mortality in radio collared adults, fox predation and road mortality, in order to identify effective management options. A rope bridge was built to provide this species with a safe passage across Caves Road in July 2013, and I present the results from 270 days of monitoring of the rope bridge and factors influencing the numbers of crossings.

No radio collared possums crossed the road successfully during my study, while two were killed on the road. No collared possums crossed the waterway, except for one accidentally falling into the waterway during a severe storm. None of the home ranges included the road or waterway, suggesting that they both act as physical barriers for possums. Even a 5 m wide firebreak was enough to limit the movements of some possums where canopy connection was not available. Individuals in partially cleared campsites mostly remained within groups of trees with continuous canopy connections. Home ranges were small (males: 0.31 ± 0.044 ha, females: 0.16 ± 0.017 ha), and their sizes were affected by sex and proximity to the waterway. These results highlight the exceptionally sedentary and arboreal nature of this species.

I used 12 species-specific microsatellite DNA loci to investigate the fine-scale spatial genetic structure and the effects of a road and an artificial waterway on the population of P. occidentalis. Spatial autocorrelation analyses identified positive genetic structure over distances up to 600 m in continuous habitat. The artificial waterway was associated with significant genetic divergence, while no significant genetic divergence was detected across the road. However, this increasingly busy road may cause future divergence, and road mortality can still contribute to loss of genetic diversity. Therefore, providing safe passages to reconnect habitat is suggested to maximise genetic diversity and prevent isolation of subpopulations.

Predation by red foxes (Vulpes vulpes) was the most common cause of mortality in adult radio collared possums contributing to 70 % of all confirmed mortalities. Road mortality also contributed to about 10 % of mortalities. A population viability analysis revealed that the probability of this important population going extinct in 20 years was alarmingly high (P = 0.921 with 95 % lower confidence interval of 0.903 and upper confidence interval of 0.937 ). Removal of the effects of road mortality and fox predation on adult and pouch young survival rates dramatically reduced the extinction probability (P = 0.318 without road mortality and P = 0. 004 without fox predation), indicating that reducing both road mortality and fox predation is essential to ensure the survival of this important population.

We monitored the rope bridge using motion sensor cameras and microchip readers for 270 days. Western ringtail possums started crossing the bridge 36 days after its installation, which was remarkably sooner than expected or previously reported. It took other possums and glider species in the eastern states of Australia seven to 17 months to start crossing rope bridges across roads. After a period of habituation, multiple individuals were found crossing the bridge every night at a rate of 8.87  0.59 complete crossings per night, which was at least double of those reported on bridges built in eastern Australia. The number of crossings increased over time and decreased on windy or warm nights. Brightness of the moon also slightly reduced the crossings by the possums. Longer monitoring and genetic analyses to test whether crossings result in gene flow are necessary to assess the true conservation value of this bridge. However, these early monitoring results suggest that rope bridges have the potential to be safe crossing structures for this species.

This study provides an example of an artificial linear structure other than a road having similar or even greater impacts on wildlife than a road. It therefore highlights the need for more research into the impacts of artificial structures such as waterways. The population of P. occidentalis I studied has a high probability of extinction in the near future and more effective management strategies, especially against the effects of fox predation and road mortality, are urgently needed in order to ensure its survival.

iii

Table of contents Abstract ………………………………………………………………………...……… i

Acknowledgement ………………………………...... ………………………………. .. vi

Declaration and publications ……………………………………………..………... viii

1. General introduction……………………………….……………………………. .... 1 1.1. Impacts of artificial linear structures ………………………...…….……….… 2 1.1.1. Impacts of roads ……………………………………………………….. 2 1.1.1.1. Habitat alteration, degradation and destruction ...... 2 1.1.1.2. Direct mortality ...... 3 1.1.1.3. Habitat fragmentation ...... 5 1.1.2. Impacts of artificial linear structures other than roads ...... 7 1.2. Mitigation measures ...... 8 1.2.1. Common measures ...... 8 1.2.2. Wildlife crossing structures ...... 9 1.3. The western ringtail possum ...... 11 1.3.1. Biology and ecology ...... 11 1.3.2. Decline and management ...... 12 1.3.3. Locke Nature Reserve and surrounding campsites ...... 14 1.4. Gaps in the knowledge ...... 16 1.5. Research aims ...... 17 1.6. Structure of the thesis ...... 18 1.7. References ...... 19

2. An artificial waterway and a road restrict movements and alter home ranges of the western ringtail possum ...... 33 Abstract ...... 34 Introduction ...... 35 Materials and methods ...... 36 Results ...... 44 Discussion ...... 49 References ...... 55 Supplementary results ....………………………………………………………...... 62

3. The western ringtail possum shows fine-scale population structure and limited gene flow across an artificial waterway ...... 63 Abstract ...... 64 Introduction ...... 65 Materials and methods ...... 66 Results ...... 70 Discussion ...... 76 References ...... 82 Appendix 1 …………………………………………………………………...…… 88

4. A predicted sharp decline of a stronghold population of the western ringtail possum calls for urgent reduction in fox predation and road mortality ……... 97 Abstract ...... 98 Introduction ...... 99 Materials and methods ...... 100 Results ...... 110 Discussion ...... 114 References ...... 122 Appendix 1 ……………………………………………………………………….. 129

5. A remarkably quick habituation and high use of a rope bridge by the western ringtail possum ………………………………………………………………….. 139 Abstract ...... 140 Introduction ...... 141 Materials and methods ...... 143 Results ...... 147 Discussion ...... 150 References ...... 155

6. General discussion ...... 161 6.1. Key findings ...... 162 6.2. Limitations ...... 166 6.3. Future research ...... 168 6.4. Management applications ...... 170 6.5. Conclusion ...... 171 6.6. References ...... 172

Acknowledgement

I simply would not have been able to complete this thesis without the generous help and advice from many people. There were a number of frustrating and tough periods, but support from a group of incredible people made the journey a great experience.

First of all, I would like to thank my supervisors Roberta Bencini and Jason Kennington. I simply cannot thank you enough for your continuous guidance, encouragement, support and advice throughout this long journey. Even though you are both extremely busy with teaching, other research and many other students, you were always a short walk/ e-mail/ phone call away and patiently answered many of my questions.

I would also like to thank Brian Chambers for his helpful advice and support in so many different aspects of this project, including the fieldwork set up and data analyses. You have no idea how helpful your practical advice on fieldwork and analyses were to me! Robert Black kindly provided me with his expert guidance and patiently taught me the life cycle analyses and population viability analysis. Thank you so much, Bob. I enjoyed our hours-long discussions. Big thanks also go to Mike Johnson for sharing his valuable advice at panel meetings and providing the final check up of this thesis. I would also like to very gratefully acknowledge Paul de Tores and Judy Clark for their extensive support, valuable advice and training at the initial stage of this project. I had the best trainers to learn the possum catching and handling skills from. I would also like to thank my thesis examiners for their thorough review of the thesis and valuable comments.

This was a large project involving over three years of continuous fieldwork, laboratory analyses, and construction of a rope bridge, and many agencies and organisations generously provided financial, in-kind and/or technical support. This project would simply not exist without the extensive support from our main industry partner, Main Roads Western Australia. I would especially like to thank Gerry Zoetelief and Alan Grist for their continuous support. The School of Animal Biology at UWA and the Western Australian Department of Parks and Wildlife also provided financial and technical support throughout this study. We received financial support from Western Power, the Satterley Property Group and the Holsworth Wildlife Research Endowment.

My gratitude goes to the DPaW Busselton office (John Carter) and the City of Busselton (Will Oldfield) for generously providing me with local support during the

vi fieldwork. I would like to thank the South West Aboriginal Land and Sea Council for their understanding and support for this project. Owners and managers of the campsites within my study area were incredibly generous and helpful. I would like to thank the Peppermint Park Eco Village, Abundant Life Centre, Scripture Union Camp Geographe, Christian Brethren Camp, and Legacy Camp for generously letting me wander around the camp and catch possums in the middle of the night. Uta Wicke at the Possum Centre in Busselton provided expert advice on the western ringtail possums. Thank you, Uta.

I could not have done my fieldwork without the help of over 100 volunteers. I am sorry that I cannot name all of you here but I truly enjoyed meeting all of you and you made the fieldwork extra enjoyable for me. Many of you came out to the bush more than once, and special thanks go to my best friend Chihiro Hirota who braved the hot, cold, dry, wet and windy conditions for countless times. I would also like to thank Kaarissa Harring-Harris for helping me collect data from the bridge in 2014.

A big thank you also goes to my office mates and fellow postgraduates at the school. You motivated me and made me think being stuck in front of a computer in the office isn’t so bad. I knew I wasn’t alone on this PhD journey!

To my friends, thank you so much for bearing with me talking non-stop about possums at one second and then trying to avoid talking about my thesis at all cost at next second. Many of you got dragged out to the bush for a fieldwork too and I hope you enjoyed meeting the possums.

To my Japanese and Australian families, thank you so much for encouraging me and always being so supportive of this long journey. Knowing that you are always behind me helped me tremendously throughout the study.

To my partner, Michael Galton-Fenzi, no words are sufficient to thank you for your limitless support and confidence in me. You always cheered me up and pushed me through when I was going through stressful times. You believed in me when I doubted myself and made sure I focused on my project.

Last but not least, I would like to say a massive thank you to all the possums that were involved in this study. I know they cannot read this, but they truly were the main contributors to this thesis, and I sincerely hope that this thesis can help the future survival of this unique and incredible species.

vii

Declaration and publications

This thesis is presented as a combination of scientific papers and chapters. All parts of this thesis have been written by Kaori Yokochi with advice from the supervisors (Roberta Bencini and W. Jason Kennington from UWA) and co-authors of the scientific papers (Brian K. Chambers in Chapter 2 and 4, and Robert Black in Chapter 4, both from UWA). My contributions (%) for each chapter of the thesis are outlined below.

Chapter 2: Yokochi K, Chambers BK, Bencini R (2015) An artificial waterway and road restrict movements and alter home ranges of endangered arboreal marsupial. Journal of Mammalogy. doi:10.1093/jmammal/gyv137

Conception and study design 60 %, data collection 90 %, data analyses 100 %, interpretation of results 90 %

Chapter 3: Yokochi K, Kennington WJ, Bencini R. (in review) A narrow artificial waterway is a greater barrier to gene flow than a major road for an endangered arboreal specialist: the western ringtail possum (Pseudocheirus occidentalis). PLOS ONE (currently under a revision after comments from reviewers)

Conception and study design 50 %, data collection 100 %, data analyses 70 %, interpretation of results 80 %

Chapter 4: Yokochi K, Black R, Chambers BK, Bencini R. (unpublished*) A predicted sharp decline of a stronghold population of the western ringtail possum calls for urgent reduction in fox predation and roadkill.

Conception and study design 60 %, data collection 90 %, data analyses 60 %, interpretation of results 80 %

*This chapter will be separated into two scientific papers for publications.

viii

Chapter 5: Yokochi K, Bencini R. (2015) A remarkably quick habituation and high use of a rope bridge by an endangered marsupial, the western ringtail possum. Nature Conservation 11: 79-94. doi:10.3897/natureconservation.11.4385

Conception and study design 60 %, data collection 95 %, data analyses 100 %, interpretation of results 100 %

The contribution of the different co-authors in the papers/chapters is mainly associated with the initial research directions, advice on data analysis when required, and editorial input in the drafts on the papers and/or chapters. I have obtained permissions from all co-authors to include these chapters in this thesis.

All procedures for sample and data collection in this thesis were approved by the Animal Ethics Committee at The University of Western Australia (RA/3/100/539 and RA/3/100/1213).

------Kaori Yokochi

Roberta Bencini (coordinating supervisor)

23 October 2015

Chapter 1. General Introduction

The landscape we live in is covered with artificial linear structures such as roads, railways, artificial waterways, and powerline corridors. These structures are essential for our modern society, providing us with vital services such as transport, water supply, drainage and electricity. However, they are not without cost to the natural environment, and damage done by these structures can be so large that they can threaten many wildlife species (Laurance et al. 2014).

In the last two decades, there has been an dramatic increase in the number of studies investigating the impacts of roads on ecosystems, and a field dedicated to such studies was developed and named “road ecology” (Forman 1998, Coffin 2007). Although artificial linear structures other than roads can potentially pose similar impacts on ecosystems, these structures have not been studied as much as roads to this date (Benítez-López et al. 2010). Studies on arboreal species are especially rare, even though they are thought to be highly vulnerable to the impacts of habitat fragmentation due to their fidelity to canopies (Lancaster et al. 2011, Taylor et al. 2011).

The western ringtail possum (Pseudocheirus occidentalis, Thomas 1888) is a strictly arboreal marsupial that mainly occurs in rapidly urbanised parts of the southwest of Western Australia, the only mainland biodiversity hotspot in Australia (Myers et al. 2000). Despite its recent classification as an endangered species, many aspects of its ecology and biology still remain unknown, mainly due to the difficulties in capturing this elusive species (Department of Parks and Wildlife WA 2014, Woinarski et al. 2014).

This thesis focuses on assessing the negative impacts of a major road and an artificial waterway on a population of western ringtail possums near Busselton, Western Australia. The effect of a rope bridge as a mitigation measure is also investigated.

1.1. Impacts of artificial linear structures

Construction of artificial linear structures normally involves clearing of land and introducing new surface and infrastructure. For example, roads are often covered with bitumen or gravels, powerline corridors are cleared areas lined with pylons with interconnecting power cables, and artificial waterways are filled with water. These new surfaces and land clearing, together with the traffic that runs on the surface, significantly change habitat and availability of resources for wildlife. For example, some raptor species have been found to benefit from roads because they prefer hunting in cleared areas or scavenge on roadkills (Meunier et al. 2000). On the other hand, where predators avoid habitats near roads, small prey species, such as the white footed mouse (Peromyscus leucopus), may benefit from the presence of roads (Rytwinski and Fahrig 2007). However, these cases seem to be rare exceptions, and the number of documented negative impacts of roads on wildlife considerably outweighs the documented positive impacts (Fahrig and Rytwinski 2009).

Although artificial linear structures can pose negative impacts on wildlife populations in many different ways (Trombulak and Frissell 2000, Jaeger et al. 2005, Clevenger and Wierzchowski 2006, Laurance et al. 2009), most of these impacts fall within three main categories: habitat loss and degradation, direct mortality, and habitat fragmentation. The vast majority of the studies on the negative impacts of artificial linear structures have focussed on roads; therefore, below I discuss the impacts of roads first, followed by the impacts of other types of artificial linear structures.

1.1.1. Impacts of roads

1.1.1.1 Habitat alteration, degradation and destruction

Land clearing prior to road construction inevitably results in removal and/or reduction of resources required for the survival for wildlife inhabiting the area (Trombulak and Frissell 2000, Grilo et al. 2010). Once the road opens to traffic, changes to the chemical and physical properties of the habitat along the road can further impact wildlife populations. For example, the greater amount of light penetrating into the vegetation along a road may deter animals that avoid open areas, while increased light, pollution, and soil compaction along a road may alter the species composition of vegetation, resulting in altered animal abundances (Trombulak and Frissell 2000, Jaeger et al. 2005, Bignal et al. 2007, Goosem 2007). Similarly, the increased level of noise along roads

2 can influence animal behaviours such as foraging, predator avoidance and courtship, resulting in reduced reproduction and/or survival (Forman and Alexander 1998, Kociolek et al. 2011, Shannon et al. 2014). For example, McClure et al. (2013) experimentally demonstrated that noise of a road alone dramatically decreased the abundance of birds. Roads can also benefit undesirable species. For example, roads attract hunters and predators such as grey wolves (Canis lupus), resulting in a lowered abundance of herbivores such as caribous (Rangifer tarandus caribou, Bowman et al. 2010). Introduced pest species, such as the cane toad (Bufo marinu), European red fox (Vulpes vulpes), and feral cats (Felis catus) in Australia, are known to use roads as dispersal corridors (May and Norton 1996, Seabrook and Dettmann 1996, Brown et al. 2006 ). Traffic on roads also aids dispersal of weed species and pathogens, such as the dieback fungus (Phytophthora cinnamomi) that is threatening native tree species throughout Australia (Goosem 2007, Cahill et al. 2008).

1.1.1.2. Direct mortality

Even before habitat loss and degradation caused by road construction affect wildlife populations, the processes involved with road constructions, especially clearing of the vegetation, can cause injuries and mortalities in slow-moving sedentary species (Trombulak and Frissell 2000). Although the exact number of animals killed during road constructions is unknown, the surface area of roads in the U.S.A. alone was estimated to be approximately 4.8 million ha in 1996 (Trombulak and Frissell 2000), indicating the magnitude of the impact of road constructions on wildlife.

Once the road opens to traffic, collisions between cars and animals pose a direct impact on the survival of the wildlife populations by directly increasing the rate of mortality. Millions of animals die on roads every year following collisions with vehicles. For example, Hobday and Minstrel (2008) estimated the annual number of animals killed on state roads in Tasmania to be up to 1,500,000, after accounting for carcass removal by scavengers and possibility of animals dying away from roads after collisions. By reviewing and updating results from previous studies, Erickson et al. (2005) estimated the annual number of birds killed on roads in the United States of America to be 80,000,000. Additionally roadkills numbers might be greatly underestimated especially for small animals because their carcasses disappear quickly (Santos et al. 2011). These numbers indicate the enormity of the issue of road mortality throughout animal taxa and regions.

However, the negative impacts of road mortality on populations cannot be assessed only by the numbers because some species may be numerous or highly reproductive enough to overcome the impact of road mortality. The effect of road mortality is thought be devastating especially for already fragmented small populations (Forman and Alexander 1998, Laurance et al. 2009), and many studies have supported this notion. For example, Jones (2000) found that the increased level of road mortality following the widening and sealing of a road in the South-west Tasmania World Heritage Area caused the local extinction of a population of the vulnerable eastern quoll (Dasyurus viverrinus) and halved the local population of the endangered Tasmanian devil (Sarcophilus harrisii). In a nine-year study of a roadside population of the threatened Florida scrub jay (Aphelocoma coerulescens), Mumme et al. (2000) found that the mortality of the breeders caused by collisions with vehicles significantly exceeded the production of yearlings near a highway, resulting in a roadside demographic sink.

Population Viability Analysis (PVA) is becoming a popular tool to assess the impacts of road mortality on wildlife populations because researchers can incorporate environmental and demographic stochasticities in predicting the population trend and compare scenarios such as removal of road mortality (Akçakaya and Sjögren-Gulve 2000). Ramp and Ben-ami (2006) conducted a PVA on a population of swamp wallabies (Wallabia bicolor) in Royal National Park near Sydney and concluded that this population is likely to go extinct in the next 100 years based on current mortality rates. However, they found that a 20 % reduction in the road mortality of females is likely to reverse the decline. Similarly, a PVA on a population of the common wombat (Vombatus ursinus) in Kosciuszko National Park in New South Wales, Australia identified road mortality as a decisive parameter that determines the population trend (Roger et al. 2011). The same PVA also identified landscape connectivity, which can be negatively affected by the presence of roads, as a parameter that affects the population viability. These examples highlight the strong impact that road mortality can pose on threatened wildlife populations. In addition, road mortalities can further threaten wildlife populations by continuously reducing the number of individuals in a population and by preventing individuals from reaching other patches of habitats, which can lead to lowered genetic diversity and the accumulation of genetic differences between isolated patches (Jackson and Fahrig 2011).

1.1 .1.3. Habitat fragmentation

Although the total landmass cleared for roads may not be as extensive as other anthropogenic alteration of the land, such as clearing for agricultural land, roads can limit movements of wildlife and act as a barrier between two separated habitats that were previously connected (Forman and Alexander 1998, Corlatti et al. 2009). If severe, this barrier effect can split a larger population into a number of smaller isolated populations without inter-population migrations. Small isolated populations have a higher risk of extinction because of their higher vulnerability to stochastic demographic changes and catastrophic events such as severe weather, fire and diseases (Foley 1997). For example, Berger (1990) followed the fate of 122 populations of bighorn sheep (Ovis canadensis) and found that populations with less than 50 individuals went extinct within 50 years in all cases and that populations with more than 100 individuals lasted longer, for up to 70 years. In addition, habitat fragmentation caused by anthropogenic barrier effects can further increase the risk of local extinction by preventing immigration and dispersal of juveniles, which are important sources of new individuals for declining populations (Crooks and Sanjayan 2006, Stewart and Van der Ree 2006).

Once small isolated populations go extinct, the barrier effect can also prevent recoloniastion of the empty habitats, resulting in a smaller number of subpopulations contributing to a metapopulation (Foley 1997). A metapopulation follows a path of extinction when the extinction rate of subpopulations surpasses the recolonisation rate (Foley 1997); therefore, severed connectivity among subpopulations increases the risk of extinction across larger landscapes (i.e. metapopulation extinction) especially for already threatened species (Ovaskainen and Hanski 2003).

A limitation on movements is also problematic because it can prevent migration of animals and reduce gene flow between these groups. When gene flow is reduced for an extended time period, the isolated group of animals can experience an increased level of inbreeding and genetic drift, which lowers the genetic diversity within the population (Frankham et al. 2002). Low genetic diversity can result in the expression of deleterious recessive alleles that are normally suppressed in genetically healthy populations, resulting in lowered fitness of the population (Frankham et al. 2002, Corlatti et al. 2009). Low genetic diversity can also reduce a population’s ability to adapt to environmental changes, such as climate change (Frankham et al. 2002). These genetic consequences ultimately increase the risk of extinction.

Barrier effects caused by roads can be so severe that Cushman et al. (2010) concluded that the habitat fragmentation effects caused by roads were more problematic than the effects caused by agricultural or urban land use in western Massachusetts, U.S.A. In a review on the genetic effects of roads on wildlife populations, Holderegger and Di Giulio (2010) found that fragmentation of habitats by roads led to rapid declines in genetic diversity within populations and increased genetic divergence between populations in a wide range of species including invertebrates, amphibians and mammals. In a study of flightless ground beetles (Carabus violaceus), Keller and Largiadér (2013 ) concluded that habitat fragmentation caused by roads caused increased genetic divergence, loss of genetic diversity, and possibly local extinctions. Genetic divergence across roads have also been reported in the desert big horn sheep (Ovis canadensis nelsoni, Epps et al. 2005), coyote (Canis latrans, Riley et al. 2006), bobcat (Lynx rufus, Riley et al. 2006), bank vole (Clethrionomys glareolus, Gerlach and Musolf 2000), common lizard species (Uta stansburiana, Plestiodon skiltonianus and Sceloporus occidentalis, Delaney et al. 2010), timber rattlesnake (Crotalus horridus, Clark et al. 2010), and red-backed salamanders (Plethodon cinereus, Marsh et al. 2008).

Arboreal animals are thought to be especially vulnerable to the effects of habitat fragmentation because of their fidelity to canopies (Lancaster et al. 2011). However, studies assessing the barrier effects of roads on arboreal species are still limited and results are varying. For example, Van der Ree et al. (2010) found that roads did not restrict the movement of squirrel gliders (Petaurus norfolcensis) as long as median strips with tall trees were present. Wilson et al. (2007), Radespiel et al. (2008), Goldingay et al. (2013), and Munguia-Vega et al. (2013) found that the movements of lemuroid ringtail possums (Hemibelideus lemuroides), golden-brown mouse lemurs (Microcebus ravelobensis), squirrel gliders, and black -tailed brush lizards (Urosaurus nigricaudus) were restricted over roads but not to the point where they prevented dispersal and gene flow. On the other hand, Lee et al. (2010) found that koalas (Phascolarctos cinereus) that were separated by roads were genetically divergent.

1.1.2. Impacts of artificial linear structures other than roads

As presented in the sections above, roads can have multi-fold impacts on populations and metapopulations because they can drive a threatened population to extinction by processes like habitat loss and degradation that lower resource availability, vehicle collisions that increases mortality, and barrier effects that prevent immigration. Roads can then prevent recolonisation of the empty habitat, which affects the survival and dynamics of metapopulations. Similarly, artificial linear structures other than roads, such as railways, powerline corridors and artificial waterways can have significant negative impacts on wildlife populations by causing habitat destruction, direct mortalities and habitat fragmentation.

Habitat is often cleared and altered in order to construct these structures, just as with roads; therefore, they can compromise the abundance and/or health of local wildlife populations (Mahoney and Schaefer 2002, Clauzel et al. 2013). Collisions between animals and trains occur on railways (Andreassen et al. 2005, de Oliveira et al. 2014), birds die of electrocution when they collide with powerlines (Bevanger 1995, Jenkins et al. 2010), and waterways can cause drowning of wildlife (Rautenstrauch and Krausman 1989, Peris and Morales 2004, García 2009) . These artificial linear structures can also restrict movements of wildlife and cause habitat fragmentation. For example, Bhattacharya et al. (2003) found that bumblebees (Bombus and Xylocopa species) avoided crossing a railway even though they had the ability to do so. St Clair (2003) found that forest dependent birds were 50 % less likely to cross a natural river than a busy highway in Banff National Park, and Marsh et al. (2007, 2008) detected significant genetic divergences across narrow natural waterways in some terrestrial species. Natural river systems have also been found to cause genetic divergence in some arboreal species (Eriksson et al. 2004, Goossens et al. 2005, Jalil et al. 2008, Quéméré et al. 2010). The barrier effects of artificial waterways can be similar or even greater than those of natural waterways because cleared and reinforced banks that are often associated with artificial waterways can further deter animals or prevent animals from climbing out of the waterway (Coulon et al. 2006).

Given these potential threats, more research into the impacts of artificial linear structures on Australian wildlife has been called since as early as 1990 (Andrews 1990), and several studies have since revealed the negative impacts of powerline corridors. For example, Goosem and Marsh (1997) and Wilson et al. (2007) found that power line

7 corridors restricted the movements of rainforest small mammals and strictly arboreal lemuroid ringtail possums. However, research on the negative impacts of artificial waterways on Australian wildlife is completely lacking despite the potentially high risk these waterways can pose.

1.2. Mitigation measures

1.2.1. Common measures

To address these negative impacts, many different types of mitigation measures have been implemented on roads and other artificial linear structures worldwide. Common measures along roads include reducing speed limits, installing warning signs for drivers, installing warning systems that alert drivers when animals are near the road, and installing systems to deter animals from roadsides (Jones 2000, Huijser et al. 2008, Laurance et al. 2009, Bond and Jones 2014). Exclusion fences have also been used along roads, railways and canals to prevent the access of animals to the linear structures (Carmichael 1991, Clevenger et al. 2001a, Laurance et al. 2009). Although the low costs and the ease of implementation have attracted road authorities worldwide to the options of reducing the speed limit and installation of road signs, these methods have been found to be ineffective in actually lowering the vehicle speed or reducing the number of collisions between vehicles and animals (Bissonette and Kassar 2008, Huijser et al. 2008). Installation of driver warning systems has increased in recent years; however, it is expensive (e.g. US$ 40,000 – 96,000 per km, Huijser et al. 2008), and some systems have been found to be ineffective for the traffic travelling at over 100 km/h (Gordon et al. 2004). Fences have been found to reduce the number of roadkills if they are properly installed and maintained (Clevenger et al. 2001a); however, they are expensive (US$ 52,000 – 82,000 per km on both sides of a road, Huijser et al. 2008) and cannot be used where vehicle need to access to properties along the road. Fences also do not address the issue of severed connectivity and restricted gene flow. In fact, for the species that avoid roads or for areas where road mortality is low, fences have been found to increase the extinction risk of wildlife populations by increasing the effect of habitat fragmentation (Jaeger and Fahrig 2004).

1.2.2. Wildlife crossing structures

Given the limitations of these common mitigation measures, wildlife crossing structures have become popular in recent years because they can in theory reduce road mortalities and maintain habitat connectivity by providing animals with safe passages across roads (Foster and Humphrey 1995, Clevenger and Huijser 2011). Many underpasses and overpasses have been constructed under and over roads worldwide, and a wide variety of terrestrial species have been found to use them (Clevenger and Huijser 2011) . The most famous and well documented case would be the underpasses and overpasses constructed across the Trans-Canada Highway in Banff National Park in Canada, which have been monitored for over 13 years. Multiple carnivorous and herbivorous mammalian species use these crossing structures (Clevenger and Waltho 2005), and crossing of these underpasses by grizzly (Ursus arctos) and black bears (Ursus americanus) has restored gene flow to a level that prevents genetic isolation (Sawaya et al. 2014). On a Spanish motorway, over 17 mammalian and reptilian species used overpasses and three different types of underpasses (Mata et al. 2008). In southeast Queensland, Australia, at least five native species of mammals were found to cross Compton Road using an overpass, and over 20 species of birds that avoid crossing the road were found either flying above or utilising the vegetation on the overpass (Bond and Jones 2008, Jones and Pickvance 2013). In southwest Western Australia, 15 mammalian, reptilian and bird species have been recorded to use underpasses to cross roads and highways (Chambers and Bencini 2015).

Although these crossing structures can benefit many species of varying taxa, not a single type of wildlife crossing structure serves all species, and some species show strong preference towards particular types of crossing structure. For example, in Banff National Park, grizzly bears, wolves (Canis lupus), deer (Odocoileus species), and elk (Cervus elaphus) preferred large open underpasses, while black bears and cougars (Puma concolor) preferred small confined underpasses possibly due to their affinity to cover (Clevenger and Waltho 2005). In northeast New South Wales, Australia, larger animals such as macropods, canids, and large lizards used overpasses more than underpasses to cross roads, while small animals such as bandicoots, rodents and frogs preferred underpasses (Hayes and Goldingay 2011). These preferences shown by different species highlight the importance of installing appropriate types of crossing structures based on the target species’ known biological and ecological characteristics.

Having an appropriate type of crossing structures is particularly important for strictly arboreal species like the lemuroid ringtail possum and the Herbert River ringtail possum (Pseudochirulus herbertensis) in tropical Queensland, Australia. These possum species rarely descend to the ground and would not use underpasses; however, Goosem et al. (2005) recorded the use of rope bridges by these possums. In other parts of the world, rope bridges and gliding poles have been built to facilitate movements of arboreal animals such as lemurs, opossums, monkeys, dormice, squirrels, and flying squirrels (Norwood 1999, Mass et al. 2011, Minato et al. 2012, Kelly et al. 2013, Teixeira et al. 2013, Sonoda 2014). Within Australia, these crossing structures have been built for glider and possum species as well as koalas. However, monitoring of the use of these structures by the target species is still limited to a handful of cases (Weston et al. 2011, Goldingay et al. 2013, Russell et al 2013, Taylor and Goldingay 2013, Soanes et al. 2013), and an assessment of factors influencing the use of these structures is lacking.

As well as benefitting native wildlife populations, wildlife crossing structures can also benefit introduced or pest species. For example, in Australia, introduced species such as red foxes, feral cats, rabbits (Oryctolagus cuniculus), black rats (Rattus rattus) and house mice (Mus musculus) have been recorded to use underpasses and overpasses (Bond and Jones 2008, Chambers and Bencini 2015). The use of crossing structures by these introduced species may have negative effects on native wildlife as easier and safer dispersal of introduced species across roads can result in increased competition and predation pressure on native wildlife populations that may already be under increased pressure from the presence of artificial linear structures. Several researchers have also suggested that predators, including introduced predators, may learn to use wildlife crossing structures as prey traps because the concentrated abundance of prey species near and within a confined crossing structure could act as an easy hunting ground for predators (Hunt et al. 1987, Little et al. 2002). However, this hypothesis is not supported by available data. In their review, Little et al. (2002) found no evidence of increased frequencies of predation in or near crossing structures. Dickson et al. (2005) also found that cougars in Southern California did not use underpasses to “trap” prey species. Similarly, Ford and Clevenger (2010) analysed predation data collected over 31 years and found no evidence of increased predation on ungulates near underpasses and overpasses in Banff National Park. Although predators do not seem to use crossing structures as prey traps, avoidance of crossing structures by prey species may occur if predators frequently use them (Little et al. 2002). Some prey species, such as many

10 mammalian species (reviewed by Apfelbach et al. 2005), toads (Bufo species, Flowers and Graves 1997), and blue tits (Cyanistes caeruleus, Amo et al. 2008), are known to identify chemical cues (i.e. odours) of their predators and avoid areas with the cues. This predator avoidance behaviour may be causing prey species to avoid crossing structures used by predators, as observed in North America by Foster and Humprey (1995) and Clevenger et al. (2001b). In Australia, whether predator avoidance behaviour is displayed by native prey species at crossing structures is unclear. For example, foxes, cats and multiple individuals of their prey species used an underpass at Slaty Creek in Victoria (Abson and Lawrence 2003) and underpasses near Mandurah, Western Australia (Chambers and Bencini 2015). This apparent lack of predator avoidance by Australian species could be because these introduced predators and native prey species did not co-evolve (Little et al. 2002, Mata et al. 2015). In support of this notion, several native prey species, such as the southern brown bandicoot (Isoodon obesulus), common brushtail possum (Trichosurus vulpecula), and Australian bush rats (Rattus fuscipes) have been found not to recognise or react to scent of foxes and cats (Banks 1998, Mella et al. 2011).

Wildlife crossing structures can also be costly to install (Can$1,750,000 for an underpass in Banff National Park and €5,600,000 for an overpass in Netherlands, Huijser et al. 2008), which can deter road construction agencies and relevant government authorities from implementing them. However, the benefit they can provide to wildlife and motorists is currently thought to outweigh the large cost (Mata et al. 2008, Polak et al. 2014).

1.3. The western ringtail possum

1.3.1. Biology and ecology

The western ringtail possum, also called ngwayir by the Noongar indigenous people of southwest Australia, is a medium-sized nocturnal marsupial weighing up to 1.3 kg (Wayne et al. 2005). These possums are strongly arboreal and strictly folivorous, mostly feeding on leaves and some flowers of selected tree species (Jones et al. 1994a). They prefer tree hollows as rest sites, but they can also construct nests called “dreys” from vegetative materials especially where tree hollows are not available (Ellis and Jones 1992). They have also been observed to rest in understorey vegetations and abandoned

11 rabbit burrows where tree hollows are sparse (Jones et al. 1994a, Clarke 2011). In coastal regions of southwest Western Australia, these possums are strongly associated with peppermint trees (Agonis flexuosa), with up to 99.6 % of their diet comprising of A. flexuosa leaves and the majority of their dreys made of materials gathered from A. flexuosa trees (Ellis and Jones 1992, Jones et al. 1994a).

They can reproduce all year around if the environment is favourable, but normally have two peak breeding periods: April to June and October to December in the A. flexuosa dominated coastal region, and May to June and October to November in the jarrah (Eucalyptus marginata) dominated inland region (Jones et al. 1994b, Wayne at al. 2005). The birth of a single young is most common, but mothers can give birth to twins and raise them to maturity on rare occasions (Jones et al. 1994b). Juveniles can reach sexual maturity within a year after birth, and the possums are thought to live for 4 to 5 years on average in the wild (Ellis and Jones 1992, Wayne et al. 2005).

1.3.2. Decline and management

Fossil records indicate that P. occidentalis historically occupied most of the southwest corner of Western Australia, ranging from southeast of Geraldton (400 km North of Perth) to the southern edge of the Nullarbor Plain (Figure 1; Woinarski et al. 2014). After European settlement in the 1830s, local extinctions of these possums were recorded as early as in the 1920s, and by the 1980s, they were seen only in patchy areas in the southern coastal strip between Bunbury and Albany and in inland riparian habitat in the Upper Warren and Perup area (Jones et al. 2004, Clarke 2011, Woinarski et al. 2014). Wilson (2009) found that populations of P. occidentalis in the coastal Bunbury region (“Gelorup region”) and coastal Busselton region were genetically distinct from each other, even though these regions were only 30 km apart. The genetic distinctiveness of the Albany population was not assessed in Wilson’s study; hence it remains unknown to this date. The Upper Warren area once held a large genetically distinct population; however, regular spotlighting surveys in the area have shown a dramatic decline of the species between the 1990s and the early 2000s, and no P. occidentalis has been recorded during spotlight surveys since 2009 (Wilson 2009, Wayne et al. 2012). Currently P. occidentalis is seen in very fragmented patches between Bunbury and Augusta and in Albany. The total area it currently occupies is estimated to be less than 500 km2 (Woinarski et al. 2014).

Australia

Perth

Bunbury / Gelorup Upper Warren / Perup Busselton

Augusta Albany

Figure 1 A map of historical and current ranges of Pseudocheirus occidentalis (adapted from Woinarski et al. 2014). The dark shade represents the historical range without recent records and the lighter shades are historical ranges that experienced rapid declines. Red crosses and green circles are locations of records before 1993 and between 1993 and 2012, respectively. Blue circles are locations where the species has been translocated successfully.

Pseudocheirus occidentalis is classified as Vulnerable by the International Union for Conservation of Nature (Morris et al. 2008) and by the Australian Environment Protection and Biodiversity Conservation Act 1999 (Department of the Environment, Water, Heritage and the Arts 2013). However, these classifications are considered outdated, and the Action Plan for Australian Mammals 2012 calls for its classification to be changed to critically endangered (Woinarski et al. 2014). As a result, its classification was recently changed to Endangered by the Western Australian government (Department of Parks and Wildlife WA 2014). The main threatening processes are habitat destruction and fragmentation caused by urbanisation and logging, predation by introduced predators such as red foxes and feral cats, and altered fire regimes (Wayne et al. 2006, Woinarski et al. 2014). Road mortalities are also common for this species (Trimming et al. 2009).

Several management options, including translocations and fox control programs have been attempted since the early 1990s to stop the decline of P. occidentalis. Due to their highly specialised requirements and susceptibility to predation, translocations have had a low success rate (de Tores et al. 2004, Clarke 2011). The effectiveness of extensive fox control programs routinely conducted in Western Australia has not been identified for P. occidentalis to this date mainly due to the difficulty in capturing these possums and estimating their numbers (de Tores et al. 2004, Woinarski et al. 2014). The likely effectiveness of other management options such as reduction of road mortalities has also not been assessed.

The A. flexuosa dominated southern part of the Swan Coastal Plain now remains their core habitat; however, this coastal region is currently one of the fastest developing regions in Australia (Australian Bureau of Statistics 2014) and the increasing number of large-scale developments is further threatening the persistence of this species (Woinarski et al. 2014). Opportunistic surveys suggest a recent rapid decline of P. occidentalis on the Swan Coastal Plain, but the current status and future direction of the populations in this region has not been formally investigated with population viability analysis (Woinarski et al. 2014). The coastal habitat in the Busselton region has a population with the highest known density of P. occidentalis (Jones et al. 1994a, 2007). This area has also been identified as an area likely to be highly affected by new road developments (Laurance et al. 2014). Many artificial linear structures are already present in this area; however, their negative impacts on P. occidentalis have never been investigated. Long-term studies on home ranges, dispersal pattern and basic demographic rates, including reproductive and survival rates of wild western ringtail possums are lacking in this important region. The exception is a study conducted by Jones et al. (1994b) who estimated home ranges from 10 weeks of monitoring and observed dispersal of five individuals.

1.3.3. Locke Nature Reserve and surrounding campsites

Within the Busselton region, the highest densities of P. occidentalis have been recorded at Locke Nature Reserve and the surrounding campsites (Jones et al. 2007). Locke Nature Reserve is a 200 ha reserve that provides mostly continuous A. flexuosa canopy with swampy open areas in the southern part. The Western Australian Department of Parks and Wildlife manages the reserve, and no recreational activities are permitted within its boundaries. Monthly baiting with sodium monofluoroacetate is conducted

14 within the reserve as part of the Western Shield Program, which aims to control introduced predators, such as red foxes (de Tores et al. 2004).

The area to the north of the reserve is Crown land leased to religious and youth groups that use it as campsites. In these campsites, parts of the vegetation have been cleared for recreational purposes and the canopy connections are more limited than in the nature reserve, but the remaining vegetation is still dominated by A. flexuosa. The area to the east of the reserve is a privately owned caravan park, and canopy cover in this park is limited except for the southern end of the park and northern end along the road where a strip of A. flexuosa trees has been kept as a road reserve.

Caves Road runs between the nature reserve and campsites on its north (Figure 1). Its existence as a narrow dirt road was recorded as early as the 1930s, and it was sealed in the 1960s to become the current single carriageway road. This 15 m wide road connects popular tourist destinations in the region. The recorded daily traffic volume on this road was 6,000 cars in 2008, but it can be up to 15,000 cars during the peak tourist season in summer (Main Roads WA 2009, G. Zoetelief, Pers. Comm.). With the cleared verges and no overhanging trees across the road, the total canopy gap across the road is 25 m. The speed limit on this road is 90 km h-1, and d ead western ringtail possums have regularly been seen on this road. Trimming et al. (2009) identified this section of Caves Road as a roadkill hotspot for P. occidentalis. However, the true impact this road poses on this population has never been assessed.

In addition to Caves Road, an artificially reinforced, straightened and widened section of the Buayanyup River runs between Locke Nature Reserve and the caravan park on its east. This 30 m wide artificial waterway was built in the 1930s to act as drainage in this historically flood-prone area. Banks on both sides are kept clear to provide public access and for maintenance, resulting in a 45 m gap in the canopy. This waterway is not used for transportation, and it contains water all year. As P. occidentalis is not known to swim voluntarily, this linear structure could potentially act as a barrier; however, its impact on the P. occidentalis population is not known.

Figure 2 A map of the study area near Busselton, Western Australia. Red lines and blue lines indicate the edges of Caves Road and an artificial waterway, respectively. Areas enclosed with black rectangles are study blocks in which data and samples from Pseudocheirus occidentalis were collected.

1.4. Gaps in the knowledge

As discussed above, the impacts of artificial linear structures on wildlife, especially the impacts of artificial linear structures other than roads on arboreal species, are largely unknown. Our understanding of important aspects of the biology, ecology and current status of the endangered P. occidentalis is also lacking. Listed below are the key questions that need to be resolved: a) Do artificial linear structures, including those other than roads, act as barriers to strictly arboreal species such as P. occidentalis? b) If artificial linear structures, including those other than roads, act as barriers to strictly arboreal species, does the barrier effect result in genetic divergence?

16 c) Do rope bridges provide these strictly arboreal animals with safe crossings of artificial linear structures? d) What are the home ranges of P. occidentalis in its core habitat and what influences their size? e) What are the basic demographic parameters such as reproductive and survival rates of P. occidentalis in its core habitat? f) What are the dispersal patterns of P. occidentalis and do dispersals result in fine scale genetic structure? g) Is the stronghold P. occidentalis population suffering from lowered genetic health? h) What is the current state of the stronghold population of P. occidentalis in its core habitat? i) What is the likely future direction of the P. occidentalis population in its core habitat? j) What are the estimated impacts of known threats on P. occidentalis in its core habitat, including road mortality and fox predation?

1.5. Research aims

The first retro-fitted rope bridge in Western Australia was built across Caves Road in 2013 to provide arboreal animals with a safe passage across the road. It was constructed by Main Roads Western Australia as a part of this PhD project to investigate some of the research questions outlined above. I formulated six specific research objectives to be addressed in this thesis. They are: a) to assess the impact of a road and an artificial waterway on the movements of P. occidentalis; b) to investigate the genetic impacts of a road and an artificial waterway on P. occidentalis; c) to gain more information on home ranges of P. occidentalis in A. flexuosa dominated habitat;

17 d) to assess the general genetic health and fine-scale genetic structure within a population of P. occidentalis e) to investigate the current status and predict the future direction of a population of P. occidentalis in its core habitat using population viability analysis and assess the likely effectiveness of two potential management options: removal of fox predation and road mortality; and f) to monitor the use of a newly constructed rope bridge and assess whether it provides P. occidentalis with safe passage across Caves Road and to determine which factors affect the use of the bridge.

1.6. Structure of the thesis

This thesis is presented as a series of scientific papers in accordance with section 10. 28 – 35 of the Postgraduate and Research Scholarship Regulations at The University of Western Australia. It addresses each of above aims using a population of P. occidentalis at Locke Nature Reserve and surrounding campsites.

There are six chapters in this thesis, consisting of:

Chapter 1 (this chapter) that provides a general introduction,

Chapter 2 (data chapter) that addresses the above objectives (a) and (c),

Chapter 3 (data chapter) that addresses the above objectives (b) and (d),

Chapter 4 (data chapter) that addresses the above objective (e),

Chapter 5 (data chapter) that addresses the above objective (f), and

Chapter 6 that provides a general discussion, overall conclusions, future research directions and management implications in terms of the conservation of P. occidentalis and management of the negative impacts of roads and artificial waterways.

Some parts of the introduction and method sections overlap among four data chapters because each chapter is presented as a standalone paper. Chapters 2 to 5 are written as scientific papers with multiple authors; therefore plural pronouns are used. Chapters 2 and 5 have been published in peer-reviewed journals, and Chapter 3 has been submitted

18 and is currently under review for publication. Two scientific papers will be produced from Chapter 4 and submitted to peer reviewed journals.

1.7. References

Abson RN, Lawrence RE (2003) Monitoring the use of the Slaty Creek wildlife underpass, Calder Freeway, Black Forest, Macedon, Victoria, Australia. Road Ecology Center, University of California Davis.

Akçakaya HR, Sjögren-Gulve P (2000) Population viability analyses in conservation planning: an overview. Ecological Bulletins: 9-21

Amo L, Galván I, Tomás G, Sanz JJ (2008) Predator odour recognition and avoidance in a songbird. Functional Ecology 22: 289-293

Andreassen HP, Gundersen H, Storaas T, Hudson (2005) The effect of scent-marking, forest clearing, and supplemental feeding on moose-train collisions. Journal of Wildlife Management 69: 1125-1132

Andrews A (1990) Fragmentation of habitat by roads and utility corridors: a review. Australian Zoologist 26: 130-141

Apfelbach R, Blanchard CD, Blanchard RJ, Hayes RA, McGregor IS (2005) The effects of predator odors in mammalian prey species: A review of field and laboratory studies. Neuroscience & Biobehavioral Reviews 29: 1123-1144

Australian Bureau of Statistics (2014) 3218.0 - Regional population growth, Australia, 2012-13. Australian Bureau of Statistics. http://www.abs.gov.au/AUSSTATS/[email protected]/Lookup/3218.0Main+Features12012- 13 [accessed 23 January 2015]

Banks PB (1998) Responses of Australian bush rats, Rattus fuscipes, to the odor of introduced Vulpes vulpes. Journal of Mammalogy 79: 1260-1264

Benítez-López A, Alkemade R, Verweij PA (2010) The impacts of roads and other infrastructure on mammal and bird populations: A meta-analysis. Biological Conservation 143: 1307-1316

Berger J (1990) Persistence of different-sized populations: an empirical assessment of rapid extinctions in bighorn sheep. Conservation Biology 4: 91-98

Bevanger K (1995) Estimates and population consequences of tetraonid mortality caused by collisions with high tension power lines in Norway. Journal of Applied Ecology 32: 745-753

Bhattacharya M, Primack RB, Gerwein J (2003) Are roads and railroads barriers to bumblebee movement in a temperate suburban conservation area? Biological Conservation 109: 37-45

Bignal KL, Ashmore MR, Headley AD, Stewart K, Weigert K (2007) Ecological impacts of air pollution from road transport on local vegetation. Applied Geochemistry 22: 1265-1271

Bissonette JA, Kassar CA (2008) Locations of deer–vehicle collisions are unrelated to traffic volume or posted speed limit. Human–Wildlife Conflicts 2: 122-130

Bond AR, Jones DN (2008) Temporal trends in use of fauna-friendly underpasses and overpasses. Wildlife Research 35: 103-112.

Bond ARF, Jones DN (2014) Roads and macropods: interactions and implications. Australian Mammalogy 36: 1-14

Bowman J, Ray JC, Magoun AJ, Johnson DS, Dawson FN (2010) Roads, logging, and the large-mammal community of an eastern Canadian boreal forest. Canadian Journal of Zoology 88: 454-467

Brown GP, Phillips BL, Webb JK, Shine R (2006) Toad on the road: Use of roads as dispersal corridors by cane toads (Bufo marinus) at an invasion front in tropical Australia. Biological Conservation 133: 88-94.

Cahill DM, Rookes JE, Wilson BA, Gibson L, McDougall KL (2008) Phytophthora cinnamomi and Australia’s biodiversity: impacts, predictions and progress towards control. Australian Journal of Botany 56: 279-310

Carmichael GB (1991) Fences as barriers to desert mule deer along canals in central Arizona. Master’s thesis, The University of Arizona, Arizona U.S.A.

Chambers BK, Bencini R (2015) Factors affecting the use of fauna underpasses by bandicoots and bobtail lizards. Animal Conservation (Early view: published online) 20

Clark RW, Brown WS, Stechert R, Zamudio KR (2010) Roads interrupted dispersal and genetic diversity in timber rattlesnakes. Conservation Biology 24: 1059-1069

Clarke JR (2011) Translocation outcomes for the western ringtail possum (Pseudocheirus occidentalis) in the presence of the common brushtail possum (Trichosurus vulpecula): health, survivorship and habitat use. PhD thesis. Murdoch University (Perth, Australia).

Clauzel C, Girardet X, Foltête J-C (2013) Impact assessment of a high-speed railway line on species distribution: Application to the European tree frog (Hyla arborea) in Franche-Comté. Journal of Environmental Management 127: 125-134

Clevenger AP, Chruszcz B, Gunson KE (2001a) Highway mitigation fencing reduces wildlife-vehicle collisions. Wildlife Society Bulletin 29: 646-653

Clevenger AP, Chruszcz B, Gunson K (2001b) Drainage culverts as habitat linkages and factors affecting passage by mammals. Journal of Applied Ecology 38: 1340-1349

Clevenger AP, Huijser MP (2011) Wildlife crossing structure handbook: design and evaluation in North America. United States Department of Transportation

Clevenger AP, Waltho N (2005) Performance indices to identify attributes of highway crossing structures facilitating movement of large mammals. Biological Conservation 121: 453-464

Clevenger AP, Wierzchowski J (2006) Maintaining and restoring connectivity in landscapes fragmented by roads. In: Crooks KR, Sanjayan M (Eds) Conservation Biology 14: Connectivity Conservation. Cambridge University Press, Cambridge, U.K., 502-535

Coffin AW (2007) From roadkill to road ecology: A review of the ecological effects of roads. Journal of Transport Geography 15: 396-406

Corlatti L, HacklÄnder K, Frey-Roos F (2009) Ability of wildlife overpasses to provide connectivity and prevent genetic isolation. Conservation Biology 23: 548-556

Coulon A, Guillot G, Cosson JF, Angibault JMA, Aulagnier S, Cargnelutti B, Galan M, Hewison AJM (2006) Genetic structure is influenced by landscape features: empirical evidence from a roe deer population. Molecular Ecology 15: 1669-1679

Cushman S, Compton B, McGarigal K (2010) Habitat fragmentation effects depend on complex interactions between population size and dispersal ability: modeling influences of roads, agriculture and residential development across a range of life- history characteristics. In: Cushman S, Huettmann F (Eds) Spatial Complexity, Informatics, and Wildlife Conservation. Springer Japan, 369-385 de Oliveira SM, Murray PJ, de Villiers DL, Baxter GS (2014) Ecology and movement of urban koalas adjacent to linear infrastructure in coastal south-east Queensland. Australian Mammalogy 36: 45-54 de Tores PJ, Hayward MW, Rosier SM (2004) The western ringtail possum, Pseudocheirus occidentalis, and the quokka, Setonix brachyurus, case studies: Western Shield review, February 2003. Conservation science Western Australia 5: 235-257

Delaney KS, Riley SPD, Fisher RN (2010) A rapid, strong, and convergent genetic response to urban habitat fragmentation in four divergent and widespread vertebrates. PLoS ONE 5: e12767

Department of the Environment Water, Heritage and the Arts (2013) Commonwealth conservation advice for Pseudocheirus occidentalis (western ringtail possum). DEWHA. Canberra, Australia http://www.environment.gov.au/biodiversity/threatened/species/pubs/25911- conservation-advice.pdf [accessed 20th January 2015]

Department of Parks and Wildlife WA (2014) Summary of additions, deletions and changes to the wildlife - Conservation (specially protected fauna) notice as of 2 December 2014. DPaW. Perth, Australia http://www.dpaw.wa.gov.au/images/documents/plants-animals/threatened- species/Listings/Summary_of_Changes_to_Specially_Protected_Fauna_Notice.pdf [accessed 20th January 2015]

Dickson BG, Jenness JS, Beier P, Gehrt (2005) Influence of vegetation, topography, and roads on cougar movement in southern California. Journal of Wildlife Management 69: 264-276

Ellis M, Jones B (1992) Observations of captive and wild western ringtail possums Pseudocheirus occidentalis. The Western Australian Naturalist 19: 1-10

Epps CW, Palsbøll PJ, Wehausen JD, Roderick GK, Ramey II RR, McCullough DR (2005) Highways block gene flow and cause a rapid decline in genetic diversity of desert bighorn sheep. Ecology Letters 8: 1029-1038

Erickson WP, Johnson GD, Young Jr DP (2005) A summary and comparison of bird mortality from anthropogenic causes with an emphasis on collisions. USDA Forest Service General Technical Report PSW-GTR-191. United States Department of Agriculture, pp. 1029-1042

Eriksson J, Hohmann G, Boesch C, Vigilant L (2004) Rivers influence the population genetic structure of bonobos (Pan paniscus). Molecular Ecology 13: 3425-3435

Fahrig L, Rytwinski T (2009) Effects of roads on animal abundance: An empirical review and synthesis. Ecology and Society 14 (1): 21

Flowers MA, Graves BM (1997) Juvenile toads avoid chemical cues from snake predators. Animal Behaviour 53: 641-646

Foley P (1997) 10 - Extinction models for local populations. In: Gilpin IHE (Ed) Metapopulation Biology. Academic Press, San Diego, 215-246

Ford AT, Clevenger AP (2010) Validity of the p rey-trap hypothesis for carnivore- ungulate interactions at wildlife-crossing structures. Conservation Biology 24: 1679- 1685

Forman RTT (1998) Road ecology: A solution for the giant embracing us. Landscape Ecology 13: III-V

Forman RTT, Alexander LE (1998) Roads and their major ecological effects. Annual Review of Ecology and Systematics 29: 207-231

Foster ML, Humphrey SR (1995) Use of highway underpasses by Florida panthers and other wildlife. Wildlife Society Bulletin 23: 95-100

Frankham R, Ballou JD, Briscoe DA (2002) Introduction to Conservation Genetics. Cambridge University Press, Cambridge, U.K.

García P (2009) Mortality of vertebrates in irrigation canals in an area of west-central Spain. Animal Biodiversity and Conservation 32: 123-126

Gerlach G, Musolf K (2000) Fragmentation of landscape as a cause for genetic subdivision in bank voles. Conservation Biology 14: 1066-1074

Goldingay RL, Harrisson KA, Taylor AC, Ball TM, Sharpe DJ, Taylor BD (2013) Fine- scale genetic response to landscape change in a gliding mammal. PLoS ONE 8: e80383

Goosem M (2007) Fragmentation impacts caused by roads through rainforests. Current Science 93: 1587-1595

Goosem M, Marsh H (1997) Fragmentation of a small-mammal community by a powerline corridor through tropical rainforest. Wildlife Research 24: 613-629

Goosem M, Weston N, Bushnell S (2005) Effectiveness of rope bridge arboreal overpasses and faunal underpasses in providing connectivity for rainforest fauna. In: Irwin CL, Garrett P, McDermott KP, Raleigh NC (Eds) International Conference on Ecology and Transportation. Center for Transportation and the Environment, North Carolina State University, San Diego, 304-316

Goossens B, Chikhi L, Jalil MF, Ancrenaz M, Lackman-Ancrenaz I, Mohamed M, Andau P, Bruford MW (2005) Patterns of genetic diversity and migration in increasingly fragmented and declining orang-utan (Pongo pygmaeus) populations from Sabah, Malaysia. Molecular Ecology 14: 441-456

Gordon KM, McKinstry MC, Anderson SH (2004) Motorist response to a deer-sensing warning system. Wildlife Society Bulletin 32: 565-573

Grilo C, Bissonette JA, Cramer PC (2010) Mitigation measures to reduce impacts on biodiversity. In: Jones SR (Ed) Highways: Construction, Management, and Maintenance. Nova Science Publishers, New York, U.S.A., 73-114

Hayes IF, Goldingay RL (2009) Use of fauna road-crossing structures in north-eastern New South Wales. Australian Mammalogy 31: 89-95

Hobday AJ, Minstrell ML (2008) Distribution and abundance of roadkill on Tasmanian highways: human management options. Wildlife Research 35: 712-726

Holderegger R, Di Giulio M (2010) The genetic effects of roads: A review of empirical evidence. Basic and Applied Ecology 11: 522-531

Huijser MP, McGowen PT, Fuller J, Hardy A, Kociolek A (2007) Wildlife-vehicle collision reduction study: Report to congress. United States Department of Transportation

Hunt A, Dickens H, Whelan R (1987) Movement of mammals through tunnels under railway lines. Australian Zoologist 24: 89-93

Jackson ND, Fahrig L (2011) Relative effects of road mortality and decreased connectivity on population genetic diversity. Biological Conservation 144: 3143- 3148

Jaeger JAG, Bowman J, Brennan J, Fahrig L, Bert D, Bouchard J, Charbonneau N, Frank K, Gruber B, von Toschanowitz KT (2005) Predicting when animal populations are at risk from roads: an interactive model of road avoidance behavior. Ecological Modelling 185: 329-348

Jaeger JAG, Fahrig L (2004) Effects of road fencing on population persistence. Conservation Biology 18: 1651-1657

Jalil MF, Cable J, Sinyor J, Lackman-Ancrenaz I, Ancrenaz M, Bruford MW, Goossens B (2008) Riverine effects on mitochondrial structure of Bornean orang-utans (Pongo pygmaeus) at two spatial scales. Molecular Ecology 17: 2898-2909

Jenkins AR, Smallie JJ, Diamond M (2010) Avian collisions with power lines: a global review of causes and mitigation with a South African perspective. Bird Conservation International 20: 263-278

Jones ME (2000) Road upgrade, road mortality and remedial measures: impacts on a population of eastern quolls and Tasmanian devils. Wildlife Research 27: 289-296

Jones DN, Pickvance J (2013) Forest birds use vegetated fauna overpass to cross multi- lane road. Oecologia Australis 17: 147-156

Jones BA, Henry J, Francesconi M (2007) An important local population of the western ringtail possum, Pseudocheirus occidentalis: a 2006 survey study of the population 25

and habitat in the Busselton localities of Siesta Park and Kealy. Unpublished report for GeoCatch, Busselton, Australia

Jones BA, How RA, Kitchener DJ (1994a) A field study of Pseudocheirus occidentalis (Marsupialia: Petauridae) I. Distribution and Habitat. Wildlife Research 21: 175-187

Jones BA, How RA, Kitchener DJ (1994b) A field study of Pseudocheirus occidentalis (Marsupialia: Petauridae) II. Population studies. Wildlife Research 21: 189-201

Jones B, Meathrel C, Calver M (2004) Hypotheses arising from a population recovery of the western ringtail possum Pseudocheirus occidentalis in fire regrowth patches in a stand of Agonis flexuosa trees in south-western Australia. In: Lunney D (Ed) Conservation of Australia's Forest Fauna. Royal Zoological Society of New South Wales, Mosman NSW, Australia, 656-662

Keller I, Largiadèr CR (2003) Recent habitat fragmentation caused by major roads leads to reduction of gene flow and loss of genetic variability in ground beetles. Proceedings of the Royal Society of London Series B: Biological Sciences 270: 417- 423

Kelly CA, Diggins CA, Lawrence AJ (2013) Crossing structures reconnect federally endangered flying squirrel populations divided for 20 years by road barrier. Wildlife Society Bulletin 37: 375-379

Kociolek AV, Clevenger AP, St. Clair CC, Proppe DS (2011) Effects of road networks on bird populations. Conservation Biology 25: 241-249

Lancaster ML, Taylor AC, Cooper SJB, Carthew SM (2011) Limited ecological connectivity of an arboreal marsupial across a forest/plantation landscape despite apparent resilience to fragmentation. Molecular Ecology 20: 2258-2271

Laurance WF, Clements GR, Sloan S, O'Connell CS, Mueller ND, Goosem M, Venter O, Edwards DP, Phalan B, Balmford A, Van Der Ree R, Arrea IB (2014) A global strategy for road building. Nature 513: 229-232

Laurance WF, Goosem M, Laurance SGW (2009) Impacts of roads and linear clearings on tropical forests. Trends in Ecology & Evolution 24: 659-669

Lee K, Seddon J, Corley S, Ellis WH, Johnston S, de Villiers D, Preece H, Carrick F (2010) Genetic variation and structuring in the threatened koala populations of Southeast Queensland. Conservation Genetics 11: 2091-2103

Little SJ, Harcourt RG, Clevenger AP (2002) Do wildlife passages act as prey-traps? Biological Conservation 107: 135-145

Mahoney SP, Schaefer JA (2002) Hydroelectric development and the disruption of migration in caribou. Biological Conservation 107: 147-153

Main Roads WA (2009) Statewide Traffic Digest 2003/04 - 2008/09. http://www.mainroads.wa.gov.au/OurRoads/Facts/TrafficData/Pages/default.aspx#.U wMF2mKSySp [accessed 5th February 2014]

Marsh DM, Page RB, Hanlon TJ, Bareke H, Corritone R, Jetter N, Beckman NG, Gardner K, Seifert DE, Cabe PR (2007) Ecological and genetic evidence that low- order streams inhibit dispersal by red-backed salamanders (Plethodon cinereus). Canadian Journal of Zoology 85: 319-327

Marsh D, Page R, Hanlon T, Corritone R, Little E, Seifert D, Cabe P (2008) Effects of roads on patterns of genetic differentiation in red-backed salamanders, Plethodon cinereus. Conservation Genetics 9: 603-613

Mass V, Rakotomanga B, Rakotondratsimba G, Razafindramisa S, Andrianaivomahefa P, Dickinson S, Berner PO, Cooke A (2011) Lemur bridges provide crossing structures over roads within a forested mining concession near Moramanga, Toamasina Province, Madagascar. Conservation Evidence 8: 11-18

Mata C, Bencini R, Chambers BK, Malo JE (2015) Predator-prey interactions at wildlife passages: between myth and reality. In: Van der Ree R, Smith DJ, Grilo C (Eds) Handbook of Road Ecology. John Wiley & Sons, Oxford. 190-197

Mata C, Hervás I, Herranz J, Suárez F, Malo JE (2008) Are motorway wildlife passages worth building? Vertebrate use of road-crossing structures on a Spanish motorway. Journal of Environmental Management 88: 407-415

May SA, Norton TW (1996) Influence of fragmentation and disturbance on the potential impact of feral predators on native fauna in Australian forest ecosystems. Wildlife Research 23: 387-400

McClure CJW, Ware HE, Carlisle J, Kaltenecker G, Barber JR (2013) An experimental investigation into the effects of traffic noise on distributions of birds: avoiding the phantom road. Proceedings of the Royal Society of London B: Biological Sciences 280 : 20132290

Mella VSA, Cooper CE, Davies SJJF (2011) Predator odour does not influence trappability of southern brown bandicoots (Isoodon obesulus) and common brushtail possums (Trichosurus vulpecula). Australian Journal of Zoology 58: 267-272

Meunier FD, Verheyden C, Jouventin P (2000) Use of roadsides by diurnal raptors in agricultural landscapes. Biological Conservation 92: 291-298

Minato S, Iwabuchi M, Aiba H, Ohtake K, Morris P (2012) Helping (Japanese) dormice to cross the road. Oryx 46: 325-326

Morris K, Burbidge A, Friend T (2008) Pseudocheirus occidentalis. 2009 IUCN Red List of Threatened Species. The IUCN Red List of Threatened Species. Version 2014.3. http://www.iucnredlist.org/details/40585/0 [accessed 10th August 2014]

Mumme RL, Schoech SJ, Woolfenden GE, Fitzpatrick JW (2000) Life and death in the fast lane: demographic consequences of road mortality in the Florida scrub-jay. Conservation Biology 14: 501-512

Munguia-Vega A, Rodriguez-Estrella R, Shaw WW, Culver M (2013) Localized extinction of an arboreal desert lizard caused by habitat fragmentation. Biological Conservation 157: 11-20

Myers N, Mittermeier RA, Mittermeier CG, da Fonseca GAB, Kent J (2000) Biodiversity hotspots for conservation priorities. Nature 403: 853-858

Norwood C (1999) Linkages in the landscape: The role of corridors and connectivity in wildlife conservation [Book Review]. Pacific Conservation Biology 5: 158

Ovaskainen O, Hanski I (2003) How much does an individual habitat fragment contribute to metapopulation dynamics and persistence? Theoretical Population Biology 64: 481-495

Peris S, Morales J (2004) Use of passages across a canal by wild mammals and related mortality. European Journal of Wildlife Research 50: 67-72

Polak T, Rhodes JR, Jones D, Possingham HP (2014) Optimal planning for mitigating the impacts of roads on wildlife. Journal of Applied Ecology 51: 726-734

Quéméré E, Crouau-Roy B, Rabarivola C, Louis Jr EE, Chikhi L (2010) Landscape genetics of an endangered lemur (Propithecus tattersalli) within its entire fragmented range. Molecular Ecology 19: 1606-1621

Radespiel U, Rakotondravony R, Chikhi L (2008) Natural and anthropogenic determinants of genetic structure in the largest remaining population of the endangered golden-brown mouse lemur, Microcebus ravelobensis. American Journal of Primatology 70: 860-870

Ramp D, Ben-Ami D (2006) The effect of road-based fatalities on the viability of a peri- urban swamp wallaby population. The Journal of Wildlife Management 70: 1615- 1624

Rautenstrauch KR, Krausman PR (1989) Preventing mule deer drownings in the Mohawk Canal, Arizona. Wildlife Society Bulletin 17: 280-286

Riley SPD, Pollinger JP, Sauvajot RM, York EC, Bromley C, Fuller TK, Wayne RK (2006) A southern California freeway is a physical and social barrier to gene flow in carnivores. Molecular Ecology 15: 1733-1741

Roger E, Laffan S, Ramp D (2011) Road impacts a tipping point for wildlife populations in threatened landscapes. Population Ecology 53: 215-227

Russell TC, Herbert CA, Kohen JL, Cooper D (2013) The incidence of road-killed possums in the Ku-ring-gai area of Sydney. Australian Journal of Zoology 61: 87-94

Rytwinski T, Fahrig L (2007) Effect of road density on abundance of white-footed mice. Landscape Ecology 22: 1501-1512

Santos SM, Carvalho F, Mira A (2011) How long do the dead survive on the road? Carcass persistence probability and implications for road-kill monitoring surveys. PLoS ONE 6: e25383

Sawaya MA, Kalinowski ST, Clevenger AP (2014) Genetic connectivity for two bear species at wildlife crossing structures in Banff National Park. Proceedings of the Royal Society B: Biological Sciences 281

Seabrook WA, Dettmann EB (1996) Roads as Activity Corridors for Cane Toads in Australia. The Journal of Wildlife Management 60: 363-368

Shannon G, Angeloni LM, Wittemyer G, Fristrup KM, Crooks KR (2014) Road traffic noise modifies behaviour of a keystone species. Animal Behaviour 94: 135-141

Soanes K, Lobo MC, Vesk PA, McCarthy MA, Moore JL, van der Ree R (2013) Movement re-established but not restored: Inferring the effectiveness of road- crossing mitigation for a gliding mammal by monitoring use. Biological Conservation 159: 434-441

Sonoda Y (2014) Effectiveness of wildlife passage and new monitoring techniques on the road. In: Seiler A (Ed.) Infra Eco Network Europe 2014 International Conference on Ecology and Transportation. (Malmö, Sweden), September 2014. IENE: 160

St Clair CC (2003) Comparative permeability of roads, rivers, and meadows to songbirds in Banff National Park. Conservation Biology 17: 1151-1160

Stewart CW, Van der Ree R (2006) Quantifying and mitigating the negative effects of roads on Australian wildlife. Naturschutz und Landschaftsplanung 38: 338-340

Taylor BD, Goldingay RL (2013) Squirrel gliders use roadside glide poles to cross a road gap. Australian Mammalogy 35: 119-122

Taylor AC, Walker FM, Goldingay RL, Ball T, van der Ree R (2011) Degree of landscape fragmentation influences genetic isolation among populations of a gliding mammal. PLoS ONE 6: e26651

Teixeira FZ, Printes RC, Fagundes JCG, Alonso AC, Kindel A (2013) Canopy bridges as road overpasses for wildlife in urban fragmented landscapes. Biota Neotropica 13: 117-123

Trimming EM, Chambers BK, Grillo D, de Tores PJ and Bencini R (2009) Road kills of the western ringtail possum (Pseudocheirus occidentalis) occur at specific hotspots. Semi-centenary and 55th Meeting of the Australian Mammal Society. (Perth, Australia), July 2009. AMS: 24

Trombulak SC, Frissell CA (2000) Review of ecological effects of roads on terrestrial and aquatic communities. Conservation Biology 14: 18-30

Van der Ree R, Cesarini S, Sunnucks P, Moore JL, Taylor A (2010) Large gaps in canopy reduce road crossing by a gliding mammal. Ecology and Society 15: 35

Wayne AF, Cowling A, Lindenmayer DB, Ward CG, Vellios CV, Donnelly CF, Calver MC (2006) The abundance of a threatened arboreal marsupial in relation to anthropogenic disturbances at local and landscape scales in Mediterranean-type forests in south-western Australia. Biological Conservation 127: 463-476

Wayne AF, Rooney JF, Ward CG, Vellios CV, Lindenmayer DB (2005) The life history of Pseudocheirus occidentalis (Pseudocheiridae) in the jarrah forest of south-western Australia. Australian Journal of Zoology 53: 325-337

Wayne A, Ward C, Vellios C, Maxwell M, Wilson I, Wayne J, Ward B, Liddelow G, Renwick J, Orell P (2012) Ngwayir (Pseudocheirus occidentalis) declines in the Upper Warren, the issue in brief. WA Department of Environment and Conservation, Perth, Australia

Weston N, Goosem M, Marsh H, Cohen M, Wilson R (2011) Using canopy bridges to link habitat for arboreal mammals: successful trials in the Wet Tropics of Queensland. Australian Mammalogy 33: 93-105

Wilson RF, Marsh H, Winter J (2007) Importance of canopy connectivity for home range and movements of the rainforest arboreal ringtail possum (Hemibelideus lemuroides). Wildlife Research 34: 177-184

Wilson KJ (2009) Quantifying the genetic effects of habitat fragmentation on the western ringtail possums (Pseudocheirus occidentalis) in south-west Western Australia. Honours thesis, Murdoch University, Perth, Australia

Woinarski JC, Burbidge AA, Harrison PL (2014) The action plan for Australian mammals 2012. CSIRO Publishing, Collingwood, Australia

Chapter 2

An artificial waterway and a road restrict movements and alter home ranges of the western ringtail possum

This chapter has been published in the Journal of Mammalogy as: Yokochi K, Chambers B, Bencini R (2015) An artificial waterway and road restrict movements and alter home ranges of endangered arboreal marsupial. Journal of Mammalogy doi:10.1093/jmammal/gyv137

An artificial waterway and a road restrict movements and alter home ranges of an endangered arboreal marsupial

Kaori Yokochi, Brian K. Chambers, Roberta Bencini

Abstract

Artificial linear structures can cause habitat fragmentation by restricting movements of animals and altering home ranges. The negative impacts of these linear structures, especially of those other than roads, on arboreal species have been rarely studied even though these species can be greatly affected because of their fidelity to the canopy. We studied the home ranges of an endangered arboreal marsupial, the western ringtail possum (Pseudocheirus occidentalis), with a focus on the impacts of a road and an artificial waterway on their movement. We radiotracked18 females and 19 males for 3 years along a major road and an artificial waterway near Busselton, Western Australia, and estimated home ranges using the a-local convex hull (a-LoCoH) estimator. No possum crossed the road successfully during the monitoring period while one crossed the waterway. Males had a mean home range size of 0.31 ± 0.044 (SE) ha, almost double that of the females at 0.16 ± 0.017 ha. Possums near the waterway had larger home ranges (0.30 ± 0.048 ha) than those near the road (0.19 ± 0.027 ha), and the size increased with proximity to the waterway, probably due to the greater availability of nearby canopy connections and the lower availability of preferable foliage. These results demonstrate that both the road and waterway represent significant physical barriers to possums, and the artificial waterway influenced home ranges more severely than the road. This suggests that linear infrastructure other than roads can affect movements of strictly arboreal animals, and negative impacts of these structures need to be assessed and mitigated by reconnecting their habitat, just as those of roads.

Introduction

Artificial linear structures can pose negative effects on wildlife populations by causing habitat destruction, lowered habitat quality, and habitat fragmentation due to barrier effects (Forman and Alexander 1998). Severed or restricted movements and gene flow across artificial linear structures can further increase the risk of extinction in threatened wildlife species by lowering their fitness and adaptability (Forman and Alexander 1998, Frankham et al. 2002, Epps et al. 2005). Habitat fragmentation can also restrict dispersal, which further increases the possibility of local extinction (Forman and Alexander 1998). Studies of the impacts of roads on wildlife have become more common in recent years (Clevenger and Wierzchowski 2006), and Forman (2000) suggested that road networks have ecological impacts on 20% of the land of the United States. Other linear infrastructure, such as artificial waterways, can also cause habitat fragmentation and restrict the movement of animals; however, studies on negative impacts of these types of infrastructure are rare compared to those of roads. Such studies are lacking especially on arboreal species even though they are thought to be highly vulnerable to the impact of habitat fragmentation due to their fidelity to canopies (McAlpine et al. 2006). To investigate the impacts of different types of artificial linear structures on the movement of arboreal animals, we studied the movements and home ranges of western ringtail possums (Pseudocheirus occidentalis Thomas 1888) near an existing road and an artificial waterway that was built to ease flooding in the area.

Pseudocheirus occidentalis has experienced a dramatic decline in its numbers and range due to factors such as destruction and fragmentation of habitat and predation by introduced predators (de Tores et al. 2008). This species was recently classified as critically endangered in Australian national action plans for mammals and is expected to decline further in the future due to these continuing threats as well as climate change (Molloy et al. 2014, Woinarski et al. 2014). Despite its endangered status, many aspects of its ecology and biology still remain unknown mainly due to the difficulty of capturing these strongly arboreal animals (de Tores et al. 2004, Wayne et al. 2005a, Woinarski et al. 2014). These nocturnal folivores occupy a limited geographic range in the southwest of Western Australia, a major biodiversity hotspot on mainland Australia (Myers et al. 2000). The peppermint tree (Agonis flexuosa) dominated woodlands of the Swan Coastal Plain between Bunbury and Dunsborough are their core habitat because A. flexuosa is their preferred food source and they are known to build nests (“dreys”) using materials mainly from this tree species (Ellis and Jones 1992, Jones et al. 1994a, de 35

Tores et al. 2004). However, the southwest region currently has the highest road density among all the regions in Western Australia except for the Perth Metropolitan region (Main Roads Western Australia 2014), and it is one of the most rapidly growing regions in Australia with up to 5.3% annual increase in its human population (Australian Bureau of Statistics 2014). This rapid urbanization will contribute to further loss and fragmentation of P. occidentalis habitat. Wayne et al. (2006) found that the abundance of P. occidentalis in noncoastal jarrah (Eucalyptus marginata) dominated forests was negatively associated with habitat loss caused by forest fragmentation. The negative impacts of habitat fragmentation by roads on other possum species have also been documented (Wilson et al. 2007, Lancaster et al. 2011); however, little is known about the impacts of roads on P. occidentalis, let alone the impacts of artificial waterways.

Pseudocheirus occidentalis is known to be sedentary and strongly reluctant to traverse on the ground (Jones et al. 1994b), and Wilson et al. (2007) found that similarly arboreal lemuroid ringtail possums (Hemibelideus lemuroides) avoided crossing a 5 to 20 m wide forestry dirt road and a power-line corridor at ground level. Given this information, we predicted that a major busy road and a 30 m wide artificial waterway would prevent P. occidentalis from crossing or further expanding their home ranges. Although these linear structures may present barriers on one side of possums’ home ranges, the available vegetation along and away from these linear structures would enable possums to extend their home ranges into different directions; therefore, we also expected that the size of home ranges of P. occidentalis would not differ near and away from the road or waterway. In the process of testing these hypotheses, we also aimed to gain basic information on their home ranges, such as differences due to sex, reproductive season, and land use of their habitat.

Materials and methods

Study area

This study was conducted in Locke Nature Reserve and surrounding campsites, 9 km west of Busselton, Western Australia (33° 39′ S 115° 14′ E). This 200 ha reserve was managed by the Western Australian Department of Parks and Wildlife with no recreational activities by the public permitted. Its A. flexuosa-dominated habitat was known to support a high density of P. occidentalis (de Tores and Elscot 2010). The public used the campsites surrounding the reserve throughout the year with a peak in

36 summer. They saw P. occidentalis regularly but they did not directly interact with the animals due to the strongly nocturnal and arboreal nature of the species.

Caves Road, a 15 m wide single carriageway providing a 25 m gap between vegetation canopies, separated the nature reserve in the south from campsites in the north (Figure 1). The traffic volume on this road was highly seasonal and reached up to 15,000 vehicles per day during the peak holiday season (G. Zoetelief, Main Roads Western Australia, pers. comm.). On the eastern edge of the reserve, an artificially reinforced, straightened, and widened part of the Buayanyup River (“artificial waterway”) ran from south to north, separating the reserve from a campsite. This 30 m wide artificial waterway was built in the 1930s to act as drainage in this historically flood-prone area and contained water throughout the year. Banks on both sides of the waterway were kept clear of vegetation to provide public access and access for maintenance, which resulted in a 45 m wide gap between canopy-level vegetation. Neither of these artificial linear structures had canopy connections across them, meaning that there was no connection among branches of trees that would allow possums to cross the clearings without descending to the ground.

Four 200 × 200 m study blocks (1A, 1B, 2A, and 2B) were set up so that 1A and 2A fell in the nature reserve, and 1B and 2B fell in campsites. 1A and 1B were directly opposite each other and separated by Caves Road, and 2A and 2B were directly opposite each other and separated by the waterway (Figure 1). The canopy connection in the campsites was limited compared to the nature reserve, in which canopy cover was mostly continuous except for a few firebreaks and a swampy area in the southern part of 2A. In 1A, there was a 5 m wide firebreak running parallel to the road without any canopy connection and a 4 m firebreak running perpendicular to the road with at least two sections with canopy connection that would allow P. occidentalis to cross the firebreak without descending to the ground (Table 1). In 2A, there was a 4 m wide firebreak running parallel to the waterway and a 2 m wide track running perpendicular to the waterway, both with more than three sections with canopy connections.

Figure 1 A map of the study area near Busselton, Western Australia. Solid lines represent borders of Caves Road (west to east) and an artificial waterway (north to south). 1A, 1B, 2A, and 2B are 200 × 200 m blocks where data from radio-collared Pseudocheirus occidentalis were collected. 1A and 2A are inside Locke Nature Reserve, and 1B and 2B are within campsites.

Table 1 Characteristics of study blocks and firebreaks at the study site near Busselton, Western Australia. Barrier is the type of artificial linear structure adjacent to the block, and direction of a firebreak is its direction against the closest barrier. Canopy connection is the number of sections with canopy connections that would allow possums to cross the firebreak without descending to the ground. Firebreaks Block Barrier Land use Direction Width (m) Canopy connection 1A Road Nature reserve Parallel 5 0 Perpendicular 4 2 2A Waterway Nature reserve Parallel 4 > 3 Perpendicular 2 > 3 1B Road Campsite Canopy connection was limited throughout. 2B Waterway Campsite Canopy connection was limited throughout.

Data collection

Pseudocheirus occidentalis was captured near and away from the road and waterway in all 4 blocks using a specially modified tranquilizer dart gun with darts containing a dose of 11-12 mg/kg of Zoletil 100 (Virbac Australia, Milperra, Australia) following a method developed by P. de Tores and reported by Clarke (2011). Initially, three adult males and three adult females in each block (i.e., 24 animals in total) were fitted with VHF radio collars with a mortality function (AVM Instrument Company, Ltd., Colfax, California, or Biotrack, Wareham, United Kingdom). However, the number of monitored animals fluctuated throughout the monitoring period between March 2010 and March 2013 due to mortality and failure of some transmitters. When we failed to pick up signals from a radio collar, we expanded the search area to check whether the individual had moved out of the range. We also searched for the particular individual near its last known location with spotlights. If we failed to locate and recapture the animal after a collar stopped transmitting or if an animal died, another adult of the same sex was captured and collared in the same block. Fifty-two adult individuals were monitored in total during 3 years.

Collared animals were located during the day and/or night with an average time span between locations of 6.4 ± 0.23 days. Each animal’s location was determined by homing in using a three element Yagi antenna (Sirtrack, Havelock North, New Zealand) and an R-1000 telemetry receiver (Communications Specialists Inc., Orange, California). Pseudocheirus occidentalis is sedentary and homing on individuals did not cause them to move away from the researchers, so it was possible to record coordinates of each animal’s locations using a handheld GPS unit (Mobile Mapper Pro, Magellan Navigation, Inc., Santa Clara, California). We recorded the species and visually estimated the height of every tree in which a collared possum was observed and calculated the proportion of A. flexuosa among the trees utilized by the possums. We also calculated the average height of the trees in the nature reserve, campsites, along Caves Road, and along the artificial waterway. Using Wilcoxon rank sum tests, the average height of A. flexuosa was compared between those in the nature reserve and those in campsites, between those within the thin strip along the road in block 1A and those outside of the strip in 1A, and between those within the thin strip along the waterway in block 2A and those outside of the strip in 2A. All procedures for handling animals were approved by the Animal Ethics Committee at The University of Western

Australia (RA/3/100/539 and RA/3/100/1213). We conducted our fieldwork following the Australian code of practice for the care and use of animals for scientific purposes (National Health and Medical Research Council 2004), which complies with the American Society of Mammalogists guidelines (Sikes et al. 2011).

Estimation of home ranges

Out of 52 individuals, 37 had enough data to be used for robust home range analyses (block 1A: 5 females and 7 males, 1B: 4 females and 5 males, 2A: 5 females and 5 males, 2B: 4 females and 2 males) after the assessment of incremental plots of estimated home range area against the number of locations using Ranges 8 (Kenward et al. 2008). This lack of data was due to mortality events and an unexpectedly large number of AVM collars failing prematurely. Both day and night locations were pooled for each individual in order to include foraging and resting locations into the estimation of its overall home range size. The average number of locations recorded for each individual was 62.6 ± 5.78 (ranging from 26 to 156). Time to independence, which is the time span required between location records in order to achieve temporal independence, was estimated using Ranges 8. For 12 individuals, time to independence was estimated to be over 21 days, which was impractical especially if we were to have an adequate number of locations for robust analyses. To estimate home ranges accurately, having an adequate number of locations is more important than achieving independence between them (Reynolds and Laundre 1990, Rooney et al. 1998, Kernohan et al. 2001). Given that P. occidentalis tends to go back to the same rest sites at dawn on consecutive days and that only one location was recorded on any day or night, we considered the existing sampling interval of approximately six days to be acceptable for the estimation of home ranges for this study.

There are two methods that have been commonly used to estimate home ranges in recent studies, kernel density estimators (KDEs) and local convex hull (LoCoH) estimator. KDE, described by Worton (1989), is currently the most commonly used method; however, several authors reported that it tends to overestimate or fragment home ranges especially when their shapes are complex (Getz and Wilmers 2004, Wilson et al. 2007, Cumming and Cornélis 2012, Kie 2013). LoCoH, especially the a-LoCoH method, has been reported to represent home ranges more accurately when there are “sharp” features like barriers although the process of implementation is more complicated than KDEs (Getz et al. 2007, Cumming and Cornélis 2012). Our study area 40 included potential sharp barriers such as a major road and an artificial waterway; therefore, we employed the a-LoCoH method to estimate the home ranges of P. occidentalis.

We used the tlocoh package (Lyons et al. 2013) in R version 3.0.1 (R Development Core Team 2013) without the time component s to estimate home ranges. We estimated the a value for each individual as described by Getz et al. (2007) and Lyons et al. (2013) and calculated 25%, 50%, and 95% isopleths of individual home ranges. Estimated home ranges were then visualized on ArcGIS 10 (ESRI 2012). We used the 95% isopleth as an estimate of home range to match the published literature, although the appropriateness of this value as a representation of a home range has been questioned when using KDE (Börger et al. 2006, Fieberg and Börger 2012). The 25% isopleth was regarded as the core home range.

We estimated home range sizes during the breeding (April to July and September to November) and non-breeding seasons (rest of the year), as observed in this study and also by Jones et al. (1994b). Studied animals were found to stay in the same area over consecutive breeding seasons, so data over three years were pooled for both seasons to increase sample sizes. After an assessment of incremental plots, home ranges in the breeding season were estimated for 8 females and 6 males, and those in non-breeding season were estimated for 11 females and 8 males. All the home range sizes were log- transformed to fit normality.

Factors influencing home range size

We assessed the effects of the type of the closest barrier, distance from the closest barrier, sex, and habitat land use on the size of home ranges using a generalized linear model in JMP 10 (SAS Institute Inc. 2012; Table 2a). We separated possums into two groups according to the type of their closest barrier (road or waterway). We then calculated the distance from the closest barrier for each individual as the shortest distance between Caves Road or the waterway and the central point of the 25% isopleth home range. To assess whether the distance to the barrier influenced home ranges differently for two kinds of barriers, we added the interactive model between type of barrier and distance to the barrier. We categorized animals in blocks 1A and 2A as the “nature reserve group,” and animals in 1B and 2B as the “campsite group.” We also included number of locations and body weight in the modelling to ensure home range

41 estimates were not affected by the duration of monitoring or the size of the animals. We assessed whether reproductive season had an effect on the size of home range using linear mixed models with standard least square personality while setting individual identity as a random effect (Table 2b). We added an interactive model between sex and breeding season in the analysis to assess whether breeding seasons affected home range differently for males and females. For all the modelling analyses, a normal distribution and identity link function were used.

For each set of candidate models, we ranked models based on their corrected Akaike Information Criterion (AICc) values. We regarded models that ranked higher than the null model and with ΔAICc of less than 2.0 as having a strong support, and those with ΔAICc between 2.0 and 7.0 as having weak support from our data (Burnham and Anderson 2002). When a model was found to have support from our data, we assessed the directionality and significance of the effects of factors based on parameter estimates and their 95% confidence intervals.

Table 2 Candidate models and corresponding hypotheses on a) overall home range size and b) reproductive seasonal home range size of Pseudocheirus occidentalis in Busselton, Western Australia. For the linear mixed model (b), the identity of the animal (ID) was included in all models as a random factor and sex and reproductive season were fixed factors. Variables used in model Hypothesis tested a) Overall home range size (Generalised linear model) Sex Males would have larger home ranges than females. Barrier type The type of the closest barrier would not affect the home range size. Distance to barrier Proximity to the closest barrier would not affect home range size. Type x Distance Type of the closest barrier or the distance from it would not affect home range. Habitat Possums in campsites would have larger home ranges than those in the nature reserve. Body weight Weight of possums would not affect home range size. Number of locations Location record number would not affect home range size. Null Home range size would vary randomly.

b) Reproductive seasonal home range size (Linear mixed model) Sex Males would have larger home ranges than females, but there would be no effect of reproductive season. Season Home range would be larger during breeding season, but there would be no effect of sex. Sex + Season Males would have larger home ranges than females, and home ranges would be larger during breeding season for both sexes. Sex x Season Males’ home range would be larger than females’, and it would expand more than female’s during the breeding season. Null (ID factor only) Home range size would vary with individuals.

Results

Movements across the road and waterway

Eight radio-collared individuals were located within 20 m of road edges for over 98% of their location records; however, none of them was found on the other side of its usual side on Caves Road during three years of monitoring (Figure 2). One male and one female were killed on the road 115 and 311 days after they were collared, respectively. Neither of them had been located on the other side of the road before the mortality events. One male that had been living on trees adjacent to the artificial waterway crossed the waterway from 2B to 2A 35 days after being collared and never crossed the waterway again for the rest of 210 days of monitoring.

Home ranges of P. occidentalis did not included Caves Road or the artificial waterway even in the individuals that lived adjacent to them (Figures 3a and 3b), and the firebreak without canopy connection also seemed to restrict movements of some possums (Figures 2 and 3a). In 1A, all five individuals observed within the thin strip of trees between Caves Road and the west to east firebreak had narrow, elongated home ranges along the road compared to those on the other side of the road or firebreak (Figure 3a). By contrast, in 2A, where there were canopy connections, all four individuals observed within the thin strip of trees between the artificial waterway and the north to south firebreak had home ranges that included the firebreak and vegetation on the other side (Figure 3b). Visual inspections of the home ranges also revealed that where canopy connections were limited within campsites, the core of the home ranges of possums overlapped with groups of trees with continuous canopy (Figure 3b).

In all locations, A. flexuosa accounted for more than half of the trees on which radio- collared possums were observed but the proportion of possums observed on A. flexuosa within the nature reserve was higher away from the road or waterway than along the road or waterway (Table 3). On average, A. flexuosa trees were significantly taller in the campsites than in the nature reserve (P < 0.001), away from the road than along the road (P = 0.001), and away from the waterway than along the waterway (P < 0.001; Table 3).

Figure 2 Location records of Pseudocheirus occidentalis along Caves Road near Busselton, Western Australia. Solid lines are the borders of Caves Road, and dotted lines are firebreaks. Different colours represent different individuals. Locations from individuals in campsites north of the road (1B) are marked with triangles and those from individuals in the nature reserve on the south (1A) are marked in circles.

Figure 3 Examples of estimated home ranges of Pseudocheirus occidentalis along Caves Road and an artificial waterway near Busselton, Western Australia. Home ranges of a) two individuals in a nature reserve south of the road (1A), two individuals in campsites north of the road (1B), and b) three individuals in a nature reserve west of the artificial waterway (2A), and two individuals in a campsite east of the waterway (2B) are presented. Solid lines are approximate edges of canopy covers adjacent to the road or waterway, and dotted lines are firebreaks. Areas with different shades represent the 95% (light), 50% (medium), and 25% isopleths (dark) of home ranges.

Table 3 Proportion of Agonis flexuosa among trees on which radio-collared Pseudocheirus occidentalis was observed and the average height of A. flexuosa in different locations near Busselton, Western Australia. N is the number of total trees recorded. Values in parentheses are SE values. Differences in the average heights were significant in all three comparisons (P < 0.05).

Location N % A. flexuosa Height (m) Nature reserve 944 84.5 8.33 (0.10) Campsite 270 86.3 9.06 (0.16) 1A Away from road 332 87.6 8.90 (0.16) Along road 69 73.4 7.92 (0.32) 2A Away from waterway 349 87.0 8.00 (0.15) Along waterway 25 59.5 6.30 (0.24)

Factors influencing the size of the home range

When we examined individual factors, sex had the strongest influence on home range size followed by the type of the closest barrier, both of which ranked higher than the null model and had significant effects (Table 4a). A model for the interaction between the type of and distance to the closest barrier also had stronger support from our data than the null model. None of the other factors such as land use and body weight had support from our data. The number of location records also did not have an effect on home range size, indicating that the number of location records collected was enough to conduct a robust estimation of home ranges.

The average home range size of males was 0.31 ± 0.044 (s.e.) ha, almost twice the size of the females’ at 0.16 ± 0.017 ha. The average home range size near the waterway (0.30 ± 0.048 ha, n = 16) was about 1.5 times larger than that near Caves Road (0.19 ± 0.027 ha, n = 21). Given the strong influences of sex and the type of the closest barrier indicated by their AICc values and 95% confidence intervals, a model combining these two factors was added to the analysis. This model had the strongest support among all models investigated, suggesting that both sex and the type of closest barrier affected the home range sizes of possums (Table 4a). When we assessed the effect of reproductive season on home range size, all three models including reproductive season had weaker

46 support than the null model, indicating that reproductive season was not a significant predictor of home range size for either sex (Table 4b).

Table 4 Results of a) generalized linear model analysis on overall home range size and b) mixed linear model analysis on reproductive seasonal home range size of Pseudocheirus occidentalis in Busselton, Western Australia. Parameter estimates are presented only for models that ranked higher than the null models or had at least weak support (ΔAICc > 7.0). * denotes parameter estimates with 95% confidence intervals outside of zero. In the case of categorical variables, parameter estimates ( s.e.) are for the categories presented in parentheses (e.g. the parameter estimate of the sex model is for females). AICc: corrected Akaike Information Criterion.

Model AICc ∆AICc Parameter estimates a) Overall home range size (Generalized linear model) Sex + Barrier type 0.56 - Sex: -0.142  0.036 (female)* Barrier: 0.117 ± 0.036 (waterway)* Sex 7.31 6.76 -0.127 ± 0.040 (female)* Barrier type x Distance 7.48 6.93 Barrier: 0.117 ± 0.039 (waterway)* Distance: -0.001 ± 0.001 Barrier x Distance: -0.003 ± 0.001* Barrier type 11.26 10.70 0.098 ± 0.043 (waterway)* Null 13.80 13.24 Habitat 14.25 13.69 Body weight 15.18 14.62 Number of locations 15.85 15.29

Distance to barrier 16.00 15.44 b) Reproductive seasonal home range size (Linear mixed model) Sex 18.63 - -0.212 ± 0.065 (female)* ID (null) 21.45 2.81 Season + Sex 25.53 6.89 Sex: -0.211 ± 0.065 (female)* Season: 0.031 ± 0.028 (breeding)* Reproductive season 28.15 9.52 Season x Sex 33.71 15.08

To further investigate the interaction between the type of and distance to the closest barrier, the effects of distance to the barrier on home range size were assessed separately for each group. Only the sex and null model had strong support (ΔAICc < 2.0) for possums near Caves Road, suggesting that the distance to the road was not a significant predictor of the home range size of possums (Table 5a). By contrast, the home range sizes of possums near the artificial waterway showed a significant negative relationship with the distance to the waterway (Table 5b).

Table 5 Results of generalized linear model analysis on home range size of Pseudocheirus occidentalis near a) Caves Road and b) an artificial waterway in Busselton, Western Australia. Parameter estimates are presented only for models that ranked higher than the null models or had at least weak support (ΔAICc > 7.0). * denotes parameter estimates with 95% confidence intervals outside of zero. For the sex factor, parameter estimates ( s.e.) are for females. AICc: corrected Akaike Information Criterion.

Model AICc ∆AICc Parameter estimates a) Caves Road Sex 6.32 - -0.114  0.052* Null 7.92 1.60 Sex + Distance 8.69 2.37 Sex: -0.109 ± 0.051* Distance: 0.001 ± 0.002 Distance 9.68 3.37 0.002 ± 0.002 b) Artificial waterway Sex + Distance -4.51 - Sex: -0.128 ± 0.041* Distance: -0.003 ± 0.001* Sex -1.79 2.72 -0.179 ± 0.045* Distance -0.50 4.01 -0.005 ± 0.001* Null 6.17 10.68

Discussion

Effects of the road on movement and home ranges

None of the collared possums was observed to successfully cross Caves Road in three years of monitoring, and both of the individuals that tried to cross the road were killed by vehicles. This lack of crossings indicates that the road was acting as a physical barrier to P. occidentalis, as expected. Although the possibility of possums crossing the road and returning to the original side between monitoring dates cannot be eliminated from our data, it is unlikely given the strong arboreal nature of this species. A similar lack of road crossings has also been reported in other arboreal animals, such as eastern chipmunks (Tamias striatus, Ford and Fahrig 2008) and squirrel gliders (Van der Ree et al. 2010). Russell et al. (2009) found that a close relative of P. occidentalis, the common ringtail possum (Pseudocheirus peregrinus), was frequently killed on roads in Sydney, Australia. This arboreal folivorous species is naïve on the ground just like P. occidentalis, and this characteristic was thought to be one of the main contributing factors to its high frequency of road mortality. This suggests that P. occidentalis may also be experiencing a high level of road mortality on roads with no canopy connections, such as Caves Road, and highlights the multiple manners in which roads can negatively affect P. occidentalis.

The proximity to the road did not affect the size of home ranges overall, as expected. Many vertebrate species expand their home ranges when the population densities and availability of food, refuge, or mates are low (Maher and Lott 2000). Pseudocheirus occidentalis in our study area depends heavily on A. flexuosa for food and refuge, so it is likely that their home range sizes are influenced by the availability and quality of A. flexuosa foliage as well as population density. Harring-Harris (2014) studied the population density of P. occidentalis along Caves Road, along the artificial waterway, and within the non-edge habitat of Locke Nature Reserve and found that the density of the possums did not differ in these three areas. However, in the same study, Harring- Harris (2014) found that the water content of A. flexuosa leaves was higher along the road and waterway than in the non-edge habitat and that nitrogen content of the leaves was also slightly higher along the road. This and our results suggest that the higher water and nitrogen content of the A. flexuosa foliage along the road did not result in reduced home range size of P. occidentalis. One possible explanation for this lack of difference could be that the quality of foliage away from the road was already sufficient

49 to sustain P. occidentalis in small home ranges, and the additional levels of water and nitrogen in foliage along the road did not change the behaviour of possums. Our study area is located within the most pristine habitat for P. occidentalis (Jones et al. 1994a, de Tores et al. 2004), and the home ranges estimated in this study were much smaller than those in a less ideal inland habitat studied by Wayne et al. (2000): 5.01 ha for males and 1.26 ha for females. These large differences in home range sizes suggest that the quality of habitat does affect the home range sizes of P. occidentalis; however, the increase in water and nitrogen content closer to Caves Road was possibly not large enough to alter their home range size.

It is also possible that other factors that increase the home range size of possums may have masked the effects of the higher quality foliage near the road. While roads may provide higher levels of water and nutrients to nearby soil due to runoff, they can also pose adverse effects on nearby vegetation by processes such as air and soil pollution and soil compaction (Trombulak and Frissell 2000). For example, Bignal et al. (2007) found that air pollution near roads increased defoliation in oak trees (Quercus petraea) and Smith et al. (2001) found that soil compaction reduced the root growth of young A. flexuosa. Although Smith et al. (2001) did not study trees near roads, their results suggest that A. flexuosa is vulnerable to a known impact of roads. The proportion of A. flexuosa trees within the thin vegetation strip along the road in 1A was found to be slightly lower than that away from the road, and A. flexuosa trees in this strip were shorter than those outside of the strip (Table 2). Whether these characteristics were caused by the presence of the road and whether they actually resulted in smaller quantity of A. flexuosa foliage is uncertain. An assessment of the levels of pollutants and soil properties as well as the quantity of A. flexuosa foliage near the road would be required to provide us with more insight into the negative impacts of roads on A. flexuosa trees and, as a result, P. occidentalis.

During our monitoring period, possums were regularly seen on the vegetation at the edge of the road foraging, grooming, socializing, and resting, and this suggests that they avoided crossing the road not because of environmental cues such as noise or light but because there was no canopy connection. Wilson et al. (2007) observed a similar trend in the lemuroid ringtail possum on a roadside. This suggests that the roadside vegetation is an important habitat for P. occidentalis as seen in other arboreal species studied by Downes et al. (1997). Preservation of such habitat could prove beneficial to this

50 endangered species especially if mitigation measures against the impact of habitat fragmentation are provided. Therefore, as suggested by several authors (Goosem et al. 2005, Laurance et al. 2009, Goldingay et al. 2013), providing a safe passage for arboreal animals to cross artificial linear structures, such as a wildlife crossing structure, would be an important step toward reducing and reversing their negative impacts. In support of this, as soon as a rope bridge was installed across Caves Road after the end of this study, possums started using it almost immediately and with a high frequency (Yokochi and Bencini 2015).

Effects of the artificial waterway on movement and home ranges

One male crossed the waterway during three years of monitoring; however, this crossing seemed to be a rare event caused by severe weather. The study area experienced strong gusts up to 55.4 km/h during the week it crossed the waterway (Australian Bureau of Meteorology 2014), and it is likely that the male fell in the water during this storm and subsequently swam to the other side. This individual never returned to the original side of the waterway after the crossing event, suggesting that the artificial waterway was also acting as a barrier for the possums, as we expected.

Contrary to our expectation, possums living closer to the waterway had larger home ranges than those away from it. The density of the possums was found to be similar near and away from the waterway, and the water contents in A. flexuosa foliage were slightly higher near the waterway possibly because of the proximity to the permanent water source (Harring-Harris 2014). Therefore, it seems that factors other than population density or water content in foliage caused the increased home range size. One possible cause of the home range expansion in P. occidentalis near the waterway is the lower availability of A. flexuosa foliage for food and refuge. We found that A. flexuosa trees within the vegetation strip next to the waterway were about 2 m shorter than those outside of the strip in the nature reserve, and the proportion of A. flexuosa was 25% lower inside the strip than outside (Table 2). Given the high dependence of P. occidentalis on A. flexuosa, it is possible that possums along the waterway had to expand their home ranges in order to have access to a greater number of A. flexuosa. The availability of canopy connections across the firebreak along the waterway would also have allowed possums to cross the firebreak in search for more food and refuge. All four individuals that were recorded on the narrow strip of the trees along the waterway incorporated the vegetation on the other side of the firebreak into their home ranges 51

(e.g., Figure 3b). This availability of canopy connections may also have caused them to have larger home ranges because they included the 2 m wide firebreak in their home ranges in order to get to the other side. Further investigations into the quantity of A. flexuosa leaves available near the waterway and other possible factors influencing the home range sizes, such as levels of other nutrients in foliage, are needed to identify what is causing this expansion of home ranges.

Both the road and the artificial waterway acted as a barrier to the movement of P. occidentalis; however, possums near the waterway had larger home ranges than those near the road. Brearley et al. (2010) also showed that impacts of barriers and urban edges on the squirrel glider changed depending on the types of the urban edges present, such as road edges or residential edges. It is therefore important that we do not assume that all barriers have uniform impacts on home ranges of wildlife.

Effects of the firebreaks

Surprisingly, a 5 m wide firebreak in the nature reserve along Caves Road with no canopy connection restricted the movements of some possums, and possums observed in the thin strip of vegetation between Caves Road and this firebreak had home ranges that were elongated along the road. The home ranges of possums on the opposite side of the road were not as elongated as those in 1A, indicating that this distinct home range shape was not caused by the presence of the road alone. Possums living next to the waterway in 2A did not have elongated home ranges even though there was a 4 m wide firebreak along the waterway, probably because of the canopy connections across the firebreak. Both the high quality of habitat and lack of canopy connections seem to have prevented the possums on the roadside in 1A from expanding their home ranges across the firebreak. Similar distinctive long and thin home ranges have also been seen in other arboreal marsupials, such as bobucks (Trichosurus cunninghami) and squirrel gliders (Petaurus norfolcensis) living in linear roadside remnants (Van der Ree and Bennett 2003, Martin et al. 2007). Possums in campsites with limited canopy connection showed a similar behavior as they tended to stay within groups of trees with continuous canopy and traveled only occasionally to other trees across cleared patches. This confirms the strong unwillingness of P. occidentalis to traverse on the ground unless they are driven by the lack of resources and highlights their high susceptibility to the effects of habitat fragmentation.

Other factors influencing home range

As we expected, males had larger home ranges than females, confirming the trend shown in previous studies on P. occidentalis (Jones et al. 1994b, Wayne et al. 2000, Clarke 2011) and in other arboreal marsupials (White 1999, Martin et al. 2007, Cruz et al. 2012, Law et al. 2013). However, reproductive season did not influence home range sizes of P. occidentalis, indicating that their home ranges during the non-breeding season are already large enough for individuals to find mates or extra resources during the breeding season. The breeding season of the possums coincide with the season when peppermint foliage is more abundant or nutritious (Jones et al. 1994a, 1994b, Wayne et al. 2005b), which would supply enough resources during this energetically demanding period. Our study area is considered to be a pristine habitat for P. occidentalis and it supports one of their highest known densities (Jones et al. 2007); therefore, the results may differ in other areas where the habitat is not ideal or where P. occidentalis occurs in lower densities.

Whether a possum lived in the nature reserve or campsites did not affect its home range size overall, and this is probably because the high quality foliage, high density of possums, and their strong fidelity to canopies prevented them from expanding home ranges despite the low density of trees in campsites. Although we did not have data on the density of A. flexuosa in the study area, it was evident that the density was lower in the campsites due to the presence of irrigated grass areas and structures such as toilet blocks and kitchens. The lower density of trees in campsites could result in each tree having more access to water, nutrients, and sunlight. Harring-Harris (2014) found that the moisture content of A. flexuosa leaves was higher in the campsite. As P. occidentalis obtains most of their water from consuming leaves, this makes the leaves in campsites more favorable to the possums than to those in the reserve. This proposition is supported by the observation by the same author that the density of P. occidentalis in campsites north of Caves Road was about four times higher than that in Locke Nature Reserve. We found that possums in campsites remained within a group of trees with continuous canopy most of the time, again highlighting their unwillingness to descend to the ground. Wright et al. (2012) found that another arboreal marsupial species, the Virginia opossum (Didelphis virginiana), had smaller home ranges in an urban environment than rural habitat due to the greater availability of food and refuges in the urban areas. Due to the more specialized diet of P. occidentalis compared to D.

53 virginiana, the effects of higher quality foliage, higher population density, and unwillingness to leave the canopy might have been just enough to counterbalance the effect of low density of trees, thus resulting in the lack of difference in home range sizes between campsites and nature reserve.

Although we did not detect differences in home range size between the nature reserve and campsites within our study site, home ranges estimated in jarrah forests (Wayne et al. 2000) and those of translocated possums (Clarke 2011) were almost 10 times larger than our estimates, indicating that the home range of this species can vary greatly depending on habitats and circumstances. For the first time, we estimated the home ranges of wild P. occidentalis in its core habitat based on long-term monitoring of multiple individuals, and home ranges of P. occidentalis in this study were expected to be smaller than those in other habitats.

Conclusion

Results from this study update and add to the essential ecological information required for the management of endangered P. occidentalis in its core habitat. Both a major road and an artificial waterway were acting as physical barriers to possums in our study, and individuals closer to the waterway had larger home ranges probably due to nearby canopy connections and the lower availability of their preferred foliage. These results indicate that permanent artificial linear structures other than roads can have a similar or greater impact on the movement and home range of strictly arboreal mammals, and their impact needs to be assessed and mitigated similarly to those of roads. Even a seemingly harmless scale of linear clearing, such as a firebreak that offered no canopy connections across it, limited the movements and home ranges of P. occidentalis; therefore, providing passages at the canopy level is essential when clearing is likely to affect the movement of this endangered species. Vegetation adjacent to linear structures can have a high conservation value for arboreal animals like P. occidentalis, and these vegetation strips need to be reconnected by means such as wildlife crossing structures to mitigate negative impacts of habitat fragmentation.

Acknowledgements

We thank the School of Animal Biology at The University of Western Australia, Main Roads Western Australia, the Western Australian Department of Parks and Wildlife, Western Power, and the Satterley Property Group for providing funding and technical support for this project. We gratefully acknowledge P J. de Tores for providing valuable advice, support and training throughout the initial part of this study. We would also like to thank the City of Busselton, Peppermint Park Eco Village, Camp Geographe, Abundant Life Centre, Christian Brethren Campsites and Possum Centre Busselton Inc. for their support and over 100 volunteers who helped us in the field.

References

Australian Bureau of Meteorology. 2014. Climate data online. http://www.bom.gov.au/climate/data/. Accessed 18 February 2014.

Australian Bureau of Statistics. 2014. 3218.0 - regional population growth, Australia, 2013–14. http://www.abs.gov.au/AUSSTATS/[email protected]/Lookup/3218.0Main+Features12012- 13. Accessed 23 January 2015.

Bignal, K. L., M. R. Ashmore, A. D. Headley, K. Stewart, and K. Weigert. 2007. Ecological impacts of air pollution from road transport on local vegetation. Applied Geochemistry 22:1265–1271.

Börger, L., et al. 2006. Effects of sampling regime on the mean and variance of home range size estimates. Journal of Animal Ecology 75:1393–1405.

Brearley, G., A. Bradley, S. Bell, and C. McAlpine. 2010. Influence of contrasting urban edges on the abundance of arboreal mammals: a study of squirrel gliders (Petaurus norfolcensis) in southeast Queensland, Australia. Biological Conservation 143:60–71.

Burnham, K. P., and D. R. Anderson. 2002. Model selection and multimodel inference - a practical information-theoretic approach. 2nd ed. Springer-Verlag New York, Inc., New York.

Clarke, J. R. 2011. Translocation outcomes for the western ringtail possum (Pseudocheirus occidentalis) in the presence of the common brushtail possum 55

(Trichosurus vulpecula): health, survivorship and habitat use. Ph.D. dissertation, Murdoch University, Perth, Australia.

Clevenger, A. P., and J. Wierzchowski. 2006. Maintaining and restoring connectivity in landscapes fragmented by roads. Pp. 502–535 in Conservation biology 14: connectivity conservation (K. R. Crooks and M. Sanjayan, eds.). Cambridge University Press, Cambridge, United Kingdom.

Cruz, J., D. R. Sutherland, and L. K. P. Leung. 2012. Den use, home range and territoriality of the koomal (Trichosurus vulpecula hypoleucus) with implications for current forest management strategies. Australian Journal of Zoology 60:141–151.

Cumming, G. S., and D. Cornélis. 2012. Quantitative comparison and selection of home range metrics for telemetry data. Diversity and Distributions 18:1057–1065. de Tores, P. J., and S. Elscot. 2010. Estimating the population size of a threatened arboreal marsupial: use of distance sampling to dispense with ad hoc survey techniques. Wildlife Research 37:512–523. de Tores, P. J., M. W. Hayward, and S. M. Rosier. 2004. The western ringtail possum, Pseudocheirus occidentalis, and the quokka, Setonix brachyurus, case studies: Western Shield review, February 2003. Conservation Science Western Australia 5:235–257. de Tores, P. J., S. M. Rosier, J. Jackson, J. Clarke, and L. Aravidis. 2008. Working to conserve the western ringtail possum. Landscope 25:54–61.

Downes, S. J., K. A. Handasyde, and M. A. Elgar. 1997. The use of corridors by mammals in fragmented Australian Eucalypt Forests. Conservation Biology 11:718– 726.

Ellis, M., and B. Jones. 1992. Observations of captive and wild western ringtail possums Pseudocheirus occidentalis. The Western Australian Naturalist 19:1–10.

Epps, C. W., P. J. Palsbøll, J. D. Wehausen, G. K. Roderick, R. R. Ramey II , and D. R. McCullough. 2005. Highways block gene flow and cause a rapid decline in genetic diversity of desert bighorn sheep. Ecology Letters 8:1029–1038.

ESRI . 2012. ArcMap. Ver. 10. Environmental Systems Research Institute, Redlands, California.

Fieberg, J., and L. Börger. 2012. Could you please phrase “home range” as a question? Journal of Mammalogy 93:890–902.

Ford, A. T., and L. Fahrig. 2008. Movement patterns of eastern chipmunks (Tamias striatus) near roads. Journal of Mammalogy 89:895–903.

Forman, R. T. T. 2000. Estimate of the area affected ecologically by the road system in the United States. Conservation Biology 14:31–35.

Forman, R. T. T., and L. E. Alexander. 1998. Roads and their major ecological effects. Annual Review of Ecology and Systematics 29:207–231.

Frankham, R., J. D. Ballou, and D. A. Briscoe. 2002. Introduction to conservation genetics. Cambridge University Press, Cambridge, United Kingdom.

Getz, W. M., S. Fortmann-Roe, P. C. Cross, A. J. Lyons, S. J. Ryan, and C. C. Wilmers. 2007. LoCoH: nonparameteric kernel methods for constructing home ranges and utilization distributions. PLoS One 2:e207.

Getz, W. M., and C. C. Wilmers. 2004. A local nearest-neighbor convex-hull construction of home ranges and utilization distributions. Ecography 27:489–505.

Goldingay, R. L., D. Rohweder, and B. D. Taylor. 2013. Will arboreal mammals use rope-bridges across a highway in eastern Australia? Australian Mammalogy 35:30– 38.

Goosem, M., N. Weston, and S. Bushnell. 2005. Effectiveness of rope bridge arboreal overpasses and faunal underpasses in providing connectivity for rainforest fauna. Pp. 304–316 in International conference on ecology and transportation (C. L. Irwin, P. Garrett, K. P. McDermott, and N. C. Raleigh, eds.). Center for Transportation and the Environment, North Carolina State University, San Diego.

Harring-Harris, K. 2014. The importance of preserving peppermint trees (Agonis flexuosa) in developed sites and providing alternative rest sites for western ringtail possums (Pseudocheirus occidentalis). Honors thesis, The University of Western Australia, Perth, Australia.

Jones, B. A., J. Henry, and M. Feancesconi. 2007. An important local population of the western ringtail possum, Pseudocheirus occidentalis: a 2006 survey study of the

population and habitat in the Busselton localities of Siesta Park and Kealy. GeoCatch, Busselton, Western Australia.

Jones, B. A., R. A. How, and D. J. Kitchener. 1994a. A field study of Pseudocheirus occidentalis (Marsupialia: Petauridae) I. Distribution and habitat. Wildlife Research 21:175–187.

Jones, B. A., R. A. How, and D. J. Kitchener. 1994b. A field study of Pseudocheirus occidentalis (Marsupialia: Petauridae) II. Population studies. Wildlife Research 21:189–201.

Kenward, R. E., N. M. Casey, S. S. Walls, and A. B. South. 2008. Ranges8: for the analysis of tracking and location data. Online manual. http://ranges- support.anatrack.com/wiki. Accessed 13 December 2014.

Kernohan, B. J., R. Gitzen, and J. J. Millspaugh. 2001. Analysis of animal space use and movements. Pp. 125–166 in Radio tracking and animal populations (J. J. Millspaugh and J. M. Marzluff, eds.). Academic Press, San Diego, California.

Kie, J. 2013. A rule-based ad hoc method for selecting a bandwidth in kernel home- range analyses. Animal Biotelemetry 1:13–25.

Lancaster, M. L., A. C. Taylor, S. J. B. Cooper, and S. M. Carthew. 2011. Limited ecological connectivity of an arboreal marsupial across a forest/plantation landscape despite apparent resilience to fragmentation. Molecular Ecology 20:2258–2271.

Laurance, W. F., M. Goosem, and S. G. W. Laurance. 2009. Impacts of roads and linear clearings on tropical forests. Trends in Ecology & Evolution 24:659–669.

Law, B., M. Chidel, A. Britton, and T. Brassil. 2013. Response of eastern pygmy possums, Cercartetus nanus, to selective logging in New South Wales: home range, habitat selection and den use. Wildlife Research 40:470–481.

Lyons, A., W. Turner, and W. Getz. 2013. Home range plus: a space-time characterization of movement over real landscapes. Movement Ecology 1:2.

Maher, C. R., and D. F. Lott. 2000. A review of ecological determinants of territoriality within vertebrate species. American Midland Naturalist 143:1–29.

Main Roads Western Australia. 2014. Annual report 2014. https://www.mainroads.wa.gov.au/Documents/Annual%20Report%20PDF%20- 58

%20FINAL%20-%20Small%20version%20for%20web.RCN- D14%5E23609562.PDF. Accessed 11 January 2015.

Martin, J. K., K. A. Handasyde, and A. C. Taylor. 2007. Linear roadside remnants: their influence on den-use, home range and mating system in bobucks (Trichosurus cunninghami). Austral Ecology 32:686–696.

McAlpine, C. A., et al. 2006. The importance of forest area and configuration relative to local habitat factors for conserving forest mammals: a case study of koalas in Queensland, Australia. Biological Conservation 132:153–165.

Molloy, S. W., R. A. Davis, and E. J. B. Van Etten. 2014. Species distribution modelling using bioclimatic variables to determine the impacts of a changing climate on the western ringtail possum (Pseudocheirus occidentalis; Pseudocheiridae). Environmental Conservation 41:176–186.

Myers, N., R. A. Mittermeier, C. G. Mittermeier, G. A. B. Da Fonseca, and J. Kent. 2000. Biodiversity hotspots for conservation priorities. Nature 403:853–858.

National Health and Medical Research Council. 2004. Australian code of practice for the care and use of animals for scientific purposes. 7th ed. Australian Government, Canberra, Australia:1–84.

R Development Core Team. 2013. R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria.

Reynolds, T. D., and J. W. Laundre. 1990. Time intervals for estimating pronghorn and coyote home ranges and daily movements. Journal of Wildlife Management 54:316– 322.

Rooney, S. M., A. Wolf, and T. J. Hayden. 1998. Autocorrelated data in telemetry studies: time to independence and the problem of behavioral effects. Mammal Review 28:89–98.

Russell, T. C., C. A. Herbert, and J. L. Kohen. 2009. High possum mortality on urban roads: implications for the population viability of the common brushtail and the common ringtail possum. Australian Journal of Zoology 57:391–397.

SAS Institute Inc. 2012. Program JMP. SAS Institute Inc., Cary, North Carolina.

Sikes, R. S., W. L. Gannon. The Animal Care and Use Committee of the American Society of Mammalogists. 2011. Guidelines of the American Society of Mammalogists for the use of wild mammals in research. Journal of Mammalogy 92:235–253.

Smith, K. D., P. B. May, and G. M. Moore. 2001. The influence of compaction and soil strength on the establishment of four Australian landscape trees. Journal of Arboriculture 27:1–7.

Trombulak, S. C., and C. A. Frissell. 2000. Review of ecological effects of roads on terrestrial and aquatic communities. Conservation Biology 14:18–30.

Van der Ree, R., and A. F. Bennett. 2003. Home range of the squirrel glider (Petaurus norfolcensis) in a network of remnant linear habitats. Journal of Zoology (London) 259:327–336.

Van der Ree, R., S. Cesarini, P. Sunnucks, J. L. Moore, and A. Taylor. 2010. Large gaps in canopy reduce road crossing by a gliding mammal. Ecology and Society 15:35.

Wayne, A. F., et al. 2005a. A comparison of survey methods for arboreal possums in jarrah forest, Western Australia. Wildlife Research 2005:701–714.

Wayne, A. F., et al. 2006. The abundance of a threatened arboreal marsupial in relation to anthropogenic disturbances at local and landscape scales in Mediterranean-type forests in south-western Australia. Biological Conservation 127:463–476.

Wayne, A. F., J. F. Rooney, C. G. Ward, C. V. Vellios, and D. B. Lindenmayer. 2005b. The life history of Pseudocheirus occidentalis (Pseudocheiridae) in the jarrah forest of south-western Australia. Australian Journal of Zoology 53:325–337.

Wayne, A., C. Ward, J. Rooney, and I. Wheeler. 2000. The immediate impacts of timber harvesting and associated activities on the Ngwayir (Pseudocheirus occidentalis) in the Jarrah forest of Kingston State Forest Block. Western Australian Department of Conservation and Land Management, Perth, Australia.

White, N. A. 1999. Ecology of the koala (Phascolarctos cinereus) in rural south-east Queensland, Australia. Wildlife Research 26:731–744.

Wilson, R. F., H. Marsh, and J. Winter. 2007. Importance of canopy connectivity for home range and movements of the rainforest arboreal ringtail possum (Hemibelideus lemuroides). Wildlife Research 34:177–184.

Woinarski, J. C., A. A. Burbidge, and P. L. Harrison. 2014. The action plan for Australian mammals 2012. CSIRO Publishing, Collingwood, Victoria, Australia.

Worton, B. J. 1989. Kernel methods for estimating the utilization distribution in home- range studies. Ecology 70:164–168.

Wright, J. D., M. S. Burt, and V. L. Jackson. 2012. Influences of an urban environment on home range and body mass of Virginia opossums (Didelphis virginiana). Northeastern Naturalist 19:77–86.

Yokochi K., and R. Bencini. 2015. A remarkably quick habituation and high use of a rope bridge by an endangered marsupial, the western ringtail possum. Nature Conservation 11:79–94.

Supplementary results

As the collar weight was found to be a significant factor influencing the survival of radio collared possums (see Chapter 4), I tested whether it influenced the home range sizes of possums. When I incorporated the collar weight model into the generalised linear model (Table 4), it was clear that the collar weight had no influence on the home range sizes of the radio-collared possums (Table A1).

Table A1 Results of generalized linear model analysis on home range size of Pseudocheirus occidentalis in Busselton, Western Australia. Parameter estimates are presented only for models that ranked higher than the null models or had at least weak support (ΔAICc > 7.0). * denotes parameter estimates with 95% confidence intervals outside of zero. In the case of categorical variables, parameter estimates ( s.e.) are for the categories presented in parentheses (e.g. the parameter estimate of the sex model is for females).AICc: corrected Akaike Information Criterion.

Model AICc ∆AICc Parameter estimates Sex + Barrier type 0.56 - Sex: -0.142  0.036 (female)* Barrier: 0.117 ± 0.036 (waterway)* Sex 7.31 6.76 -0.127 ± 0.040 (female)* Barrier type x Distance 7.48 6.93 Barrier: 0.117 ± 0.039 (waterway)* Distance: -0.001 ± 0.001 Barrier x Distance: -0.003 ± 0.001* Barrier type 11.26 10.70 0.098 ± 0.043 (waterway)* Null 13.80 13.24 Habitat 14.25 13.69 Body weight 15.18 14.62 Number of locations 15.85 15.29 Collar weight 15.97 15.41

Distance to barrier 16.00 15.44

Chapter 3

A narrow artificial waterway is a greater barrier to gene flow than a major road for an endangered arboreal specialist: the western ringtail possum (Pseudocheirus occidentalis)

This chapter is currently being revised for a publication with PLoS ONE. 63

A narrow artificial waterway is a greater barrier to gene flow than a major road for an endangered arboreal specialist: the western ringtail possum (Pseudocheirus occidentalis)

Kaori Yokochi, W. Jason Kennington, Roberta Bencini

Abstract

The fragmentation of habitats by roads and other artificial linear structures can have a profound effect on the movement of arboreal species due to their strong fidelity to canopies. Here, we used 12 microsatellite DNA loci to investigate the fine-scale spatial genetic structure and the effects of a major road and a narrow artificial waterway on a population of the endangered western ringtail possum (Pseudocheirus occidentalis) in Busselton, Western Australia. Using spatial autocorrelation analysis, we found positive genetic structure in continuous habitat over distances up to 600 m. These patterns are consistent with the sedentary nature of P. occidentalis and highlight their vulnerability to the effects of habitat fragmentation. Pairwise relatedness values and Bayesian cluster analysis also revealed significant genetic divergences across an artificial waterway, suggesting that it was a barrier to gene flow. By contrast, no genetic divergences were detected across the major road. While studies often focus on roads when assessing the effects of artificial linear structures on wildlife, this study provides an example of an often overlooked artificial linear structure other than a road that has a significant impact on wildlife dispersal leading to genetic subdivision.

Introduction

Roads and other artificial linear structures, such as railways, powerline corridors, and artificial waterways are thought to inhibit movements of animals, leading to the fragmentation of populations, increased inbreeding, and loss of genetic diversity (Forman and Alexander 1998). In a review on the genetic effects of roads on wildlife populations, Holderegger and Di Giulio (2010) found that fragmentation of habitats by roads quickly decreased genetic diversity within populations and increased genetic divergence between populations in a wide range of species including invertebrates, amphibians and mammals. Clark et al. (2010) and Epps et al. (2005) also found that relatively recently built roads limited the dispersal and increased genetic divergence of timber rattle snakes (Crotalus horridus) and bighorn sheep (Ovis canadensis nelsoni). Since inbreeding and reductions in genetic diversity increase the risk of extinction of isolated populations (Frankham 2005), it is crucial to consider the impacts of artificial linear structures when developing management strategies for threatened species.

Strictly arboreal species are thought to be more vulnerable than the majority of terrestrial species to the impacts of artificial linear structures without canopy connections because many of them tend to avoid descending to the ground, (Lancaster et al. 2011, Taylor et al. 2011). The western ringtail possum (Pseudocheirus occidentalis Thomas 1888) is a medium-sized nocturnal marsupial endemic to the southwest of Western Australia, the only international biodiversity hotspot on mainland Australia (Myers et al. 2000). This species is likely to be susceptible to the negative impacts of artificial linear structures due to their known sedentary nature and strong fidelity to canopies (Jones et al. 1994b, Chapter 2: Yokochi et al. 2015). Studies on their movements suggest that their dispersal range is small (Jones et al. 1994b). They also have small home ranges (< 0.5 ha), and a road and an artificial waterway have been found to restrict their movements (Chapter 2: Yokochi et al. 2015).

Over the last few decades P. occidentalis has gone through a dramatic decline in numbers and range due to anthropogenic factors such as habitat destruction and fragmentation and the impact of introduced predators (Burbidge and McKenzie 1989, Woinarski et al. 2014). The Bunbury – Busselton region in the southwest of Western Australia is one of the last strongholds for this species. However, it is one of the fastest growing regions in Australia (Australian Bureau of Statistics 2014), and suitable habitat for the possums is disappearing due to the rapid urbanisation. Despite its endangered

65 conservation status, relatively little is known about the population structure of P. occidentalis (Woinarski et al. 2014). The only genetic studies done to date are a phylogenetic study using mitochondrial DNA, which supported their status as a single species, and a broad-scale population genetic study that identified three distinct populations within the current range of the species based on microsatellite markers (Wilson 2009).

In this study, we used microsatellite markers to investigate whether the previously reported small home ranges and limited dispersal in P. occidentalis are supported by the presence of fine-scale genetic structure. We also investigated whether a road and artificial waterway without canopy connection were associated with genetic divergences. Given the limited movements across artificial linear structures, a strong reluctance to traverse on the ground, and lack of evidence that the species voluntarily swims (Chapter 2: Yokochi et al. 2015), we predicted that there would be genetic differentiation across both the road and artificial waterway.

Materials and methods

Study site

This study was conducted in Locke Nature Reserve and surrounding campsites, 9 km west of Busselton, Western Australia (33° 39' 32'' S; 115° 14' 26'' E), where the habitat dominated by peppermint trees (Agonis flexuosa) is supporting a high density of P. occidentalis (de Tores and Elscot 2010, Jones et al. 1994a). We set up seven 200 m x 200 m study blocks, 1A, 1B, 1-2C, 1D, 2A, 2B and 2D, chosen so that they were small enough to fall within boundaries of campsites, large enough to contain a sufficient number of individuals for sampling, and far enough from each other within continuous habitat to prevent individuals from including multiple blocks in their home ranges (Figure 1). Caves Road, running from east to west, separated the nature reserve in the south from campsites in the north with no canopy connection (Figure 1). A record of this road as a narrow gravel road exists as early as the 1930s, and it became a sealed 15 m wide single carriageway approximately 50 years ago in the 1960s. With the cleared verges on both sides and no branches in contact across it, this unfenced road provided a 25 m canopy gap. The recorded traffic volume on Caves Road was about 6,000 vehicles per day in 2008 (Main Roads Western Australia 2009). However, the traffic volume on this road is highly seasonal and it can be up to 15,000 vehicle per day during the peak holiday season in summer (G. Zoetelief, Main Roads WA, Pers. Comm.). On the eastern 66 edge of the reserve, an artificially altered part of the Buayanyup River (“artificial waterway”) ran from north to south, separating the reserve in the west from a campsite in the east (Figure 1). This 30 m wide waterway was built approximately 80 years ago in the 1930s to prevent flooding in the area, and contained water all year around. With cleared paths on both banks and no branches overhanging, this artificial waterway provided a permanent 45 m gap in the canopy.

Figure 1 A map of a study area near Busselton, Western Australia. Red lines represent the borders of Caves Road (west to east) and blue lines represent the borders of an artificial waterway (north to south). 1A, 1B, 1-2C, 1D, 2A, 2B and 2D are 200 m × 200 m study blocks where samples from Pseudocheirus occidentalis were collected. 1A, 1-2C and 2A were inside Locke Nature Reserve, and 1B, 1D, 2B and 2D were within partially cleared campsites.

Sample collection and microsatellite genotyping

Between March 2010 and November 2012, a total of 145 adult possums were captured using a specially modified tranquiliser dart gun with darts containing a dose of 11 - 12 mg/kg of Zoletil 100® (Virbac Australia, Milperra, NSW Australia). Sample sizes within each block ranged from 10 to 32 (1-2C: n = 12, 1A: n = 26, 1B: n = 32, 1D: n = 10, 2A: n = 31, 2B: n = 22, 2D: n = 12). Sample collection followed the methods developed by P. de Tores and reported by (Clarke 2011). A thin slice of ear tissue was removed from each animal under anaesthesia using Isoflurane, and each tissue was stored in either dimethyl sulfoxide solution (Seutin et al. 1991) or 90 % ethanol until DNA was extracted. Coordinates of each capture location were recorded using a handheld GPS unit (Mobile Mapper Pro®, Magellan Navigation, Inc. California USA). All procedures for capturing and handling animals followed the Australian code of practice for the care and use of animals for scientific purposes (National Health and Medical Research Council 2004), and were approved by the Animal Ethics Committee at The University of Western Australia (RA/3/100/539 and RA/3/100/1213). Permits to access Locke Nature Reserve (CE003434) and to capture the possums for scientific purposes (SF008419) were obtained from Western Australian Department of Parks and Wildlife.

DNA was extracted from each sample using Qiagen DNeasy Blood and Tissue kit (Qiagen, Venlo, Netherlands) following the manufacturer’s instructions. Concentration and quality of each DNA sample were then determined using a NanoDrop ND-1000 spectrophotometer (Thermo Fisher Scientific Inc., Massachusetts, USA). We used 12 species-specific microsatellite markers (A1, A106, A119, A122, A127, A2, A6, B104, C111, D104, D113, and D114) following PCR conditions described by Wilson et al. (2009). Genotypes at each locus were determined using an ABI 3700 Genetic Analyzer with a GeneScan-500 LIZ dye size standard (Applied Biosystems Inc., California, USA).

Data analysis

The presence of null alleles was assessed for each locus using Microchecker v.2.2.3 (Van Oosterhout et al. 2004). Genetic diversity within each block was quantified by calculating allelic richness (AR) and Nei (1978)’s estimator of gene diversity (H) within Fstat v.2.9.3 software package (Goudet 2002). Deviations from random mating were assessed using randomization tests, with results characterized with the inbreeding coefficient (FIS) statistic. Significantly positive FIS values indicate a deficit of 68 heterozygotes relative to a random mating model, while negative results indicate an excess of heterozygotes. Genotypic disequilibrium was tested between each pair of loci within each study block. FIS values and tests for deficits in heterozygotes and genotypic equilibrium were calculated using the Fstat v.2.9.3. Tests for differences in genetic diversity and FIS among study blocks were performed using Friedman’s ANOVA and Wilcoxon rank tests with the R v.3.0.1 statistical package (R Depelopment Core Team 2013).

We performed a Spatial Autocorrelation (SA) analysis on the samples collected from the nature reserve only (blocks 1A, 1-2C and 2A, n = 69) to investigate how the genetic similarity of individuals changed over geographical distance within continuous vegetation (i.e. whether a fine-scale genetic structure was present without the presence of artificial linear structures). The result from this analysis would also tell us whether distances that were the same as the widths of the road and artificial waterway were large enough to cause genetic divergence in continuous vegetation. SA analyses were conducted using GenAlEx v.6.501 (Peakall and Smouse 2012) with the results presented in two different ways. Firstly, mean genetic correlation coefficients (r) were calculated and plotted over a range of distance classes increasing at 100 m intervals to obtain autocorrelograms. Secondly, because estimates of r are influenced by the size of distance classes (Peakall et al. 2003), we also performed Multiple Distance Class (MDC) analyses to calculate and plot r for a series of increasing distance class sizes. Decreasing r with increasing distance interval class indicates significant positive spatial structure, and the distance interval class at which r is no longer greater than zero represents the limit of detectable genetic structure (Peakall et al. 2003). Tests for statistical significance were carried out by random permutation and calculating the bootstrap 95% confidence limits of r using 1000 replicates. We also conducted 2- Dimensional Local Spatial Autocorrelation (2DLSA) analysis using GenAlEx. This analysis investigates uniformity of spatial autocorrelation over the study area by estimating local autocorrelation (lr) by comparing each individual with its nearest neighbours (Double et al. 2005). Calculations of lr were made using the nearest five individuals with statistical significance determined using permutation tests.

Population structure across the whole study area was assessed using pairwise FST values and Bayesian clustering analysis. Pairwise FST values and tests for genetic differentiation between blocks were calculated using Fstat. The Bayesian clustering analysis was performed using Structure v.2.3.4 (Pritchard et al. 2000). This method 69 identifies genetically distinct clusters (K) based on allele frequencies across all loci. Analyses were based on an ancestry model, which assumes admixture and correlated allele frequencies with study blocks used as prior information about the origin of the samples. A burn-in period of 100,000 and 1,000,000 Markov Chain Monte Carlo (MCMC) iterations were used for 10 replicate runs for each number of clusters (K) ranging from 1 to 10. We then determined the most likely value of K using the ∆K method of Evanno et al. (2005) implemented in Structure Harvester v.0.6.93 (Earl and von Holdt 2012). Each individual’s average membership to K clusters from 10 replicate runs was calculated and re-organised using Clumpp v.1.1.2 (Jakobsson and Rosenberg 2007) and visualised using Distruct v.1.1 (Rosenberg 2004).

The effects of artificial linear structures were then examined by comparing relatedness values between pairs of individuals on the same side of the road or waterway with relatedness values between pairs of individuals on opposite sides of the road or waterway. The effect of the road was examined without data from individuals on the other side of the waterway (2B and 2D) to remove the effect of the waterway, and the effect of the waterway was examined without data from those on the other side of the road (1B and 1D) to remove the effect of the road. Pairwise relatedness values were calculated using the method of Queller and Goodnight (1989) with GenAlEx. We tested for differences in pairwise relatedness values using Mann-Whitney U test in R.

Results

Genetic variation within blocks

Two loci were identified as having null alleles (locus A119 in block 1B and locus A1 in blocks 2A and 2D). Since there was no consistent pattern in the presence of null alleles (i.e. the loci with null alleles were not the same across study blocks), all loci were retained for further analysis. No FIS values were significantly different from zero indicating that the groups of possum in all blocks were in Hardy-Weinberg Equilibrium. There was no evidence of genotypic disequilibrium between any pair of loci within any study block after correcting for multiple comparisons. There were also no significant differences in allelic richness (χ2 = 8.73, P = 0.189, d.f. = 6), gene diversity (χ2 = 8.55, P 2 = 0.200, d.f. = 6) or FIS (χ = 4.12, P = 0.660, d.f. = 6) among study blocks (Table 1).

Table 1 Genetic variation among P. occidentalis within each study block. Standard errors are in parentheses. N is the mean sample size per locus, AR is the mean allelic richness based on a sample size of 10 individuals, H is the mean gene diversity, and FIS is the inbreeding coefficient. No FIS values were significantly different from zero.

Block N AR H FIS 1-2C 12 3.5 (0.3) 0.62 (0.03) – 0.02 1A 26 3.4 (0.3) 0.57 (0.04) – 0.02 1B 32 3.7 (0.3) 0.60 (0.03) 0.01 1D 10 3.3 (0.3) 0.56 (0.05) 0.00 2A 31 3.7 (0.3) 0.61 (0.04) 0.04 2B 22 3.6 (0.3) 0.60 (0.03) – 0.01 2D 12 3.9 (0.3) 0.64 (0.03) 0.06

Fine-scale genetic structure within continuous habitat

Fine-scale spatial genetic structure was detected within continuous habitat. In SA analysis, the genetic correlation coefficient (r) was significantly positive up to 100 m and it intercepted zero at 347 m (Figure 2a). The MDC analysis showed significantly positive r values over distances up to 600 m (Figure 2b). All size class bins were well represented, with a minimum of 95 and 515 pairwise comparisons per bin for the SA analyses and MDC analysis respectively. The 2DLSA analysis revealed clusters of positive genetic correlation in all three blocks, indicating that positive genetic structure was not confined to one particular area within the nature reserve (Figure 3).

Figure 2 (a) A correlogram plot and (b) a multiple distance class plot based on 69 Pseudocheirus occidentalis in continuous habitat in Locke Nature Reserve near Busselton, Western Australia. Dotted lines (a) and small blue markers (b) represent upper and lower 95 % confidence intervals around zero. Red circle markers on the solid line (a) and red markers (b) are the genetic correlation values (r) that differ significantly from zero based on bootstrap resampling (P values are shown in parentheses).

Figure 3 Plot of two-dimensional local spatial autocorrelation analyses of Pseudocheirus occidentalis sampled in Locke Nature Reserve near Busselton, Western Australia.

Red lines represent the borders of Caves Road (west to east) and blue lines represent the borders of an artificial waterway (north to south). Markers represent geographical locations of the local spatial autocorrelation analyses with significantly positive (solid symbols) or non-significant values (open symbols) based on five nearest neighbours. Coordinates are based on GDA 94 projection (zone 50).

Population structure across the whole study area and the effect of artificial linear structures

Significant genetic divergences were observed between 14 (67 %) pairs of blocks. These pairs of blocks occurred on the same and opposite sides of Caves Road, as well as the same and opposite sides of the artificial waterway (Table 2). Pairwise FST values ranged from 0.017 to 0.079, with the highest FST occurring between blocks 1D and 2D, which were separated by both the road and waterway (Table 2).

Population structure across the whole study area was also evident with the Bayesian clustering analysis. Analysis of the Structure results using the ΔK method clearly identified K = 3 as the most likely number of genetic clusters (Figure 4). A bar plot of 73 individuals’ memberships to each cluster showed that most individuals from the western side of the artificial waterway (blocks 1-2C, 1A, 1B, 1D and 2A) were predominantly assigned to cluster 1 (shown in pink in Figure 5), whereas those from the eastern side (blocks 2B and 2D) were predominantly assigned to clusters 2 (yellow) and 3 (blue) respectively (Figure 5).

Table 2 Pairwise FST values (below diagonal) and P-values from tests of differentiation (above diagonal) between blocks. Significant divergences are highlighted in bold text. The adjusted significance level for multiple comparisons is 0.0024. P-values were obtained after 2,100 permutations.

1-2C 1A 1B 1D 2A 2B 2D 1-2C - 0.0010 0.0019 0.0010 0.0081 0.0005 0.0095 1A 0.050 - 0.0010 0.0148 0.0033 0.0005 0.0005 1B 0.050 0.035 - 0.0081 0.0033 0.0005 0.0005 1D 0.056 0.016 0.033 - 0.0033 0.0005 0.0005 2A 0.033 0.012 0.017 0.023 - 0.0005 0.0005 2B 0.060 0.023 0.061 0.040 0.034 - 0.0019 2D 0.041 0.061 0.047 0.079 0.036 0.045 -

3 Delta K Delta 2

0 1 2 3 4 5 6 7 8 9 10 K

Figure 4 Summary of ∆K estimates (Evanno et al. 2005) for varying numbers of genetic clusters (K) derived from the STRUCTURE analysis of 145 adult Pseudocheirus occidentalis from Busselton, Western Australia.

Figure 5 Summary of the Bayesian clustering analysis assuming three admixed populations of Pseudocheirus occidentalis in Busselton, Western Australia.

Each column represents an individual’s estimated membership to three genetic clusters represented by different colours. Vertical dotted lines separate individuals sampled from different blocks. Vertical bold solid lines represent the presence of Caves Road and an artificial waterway.

The pairwise relatedness analysis also indicated that the individuals separated by the artificial waterway were significantly less related to each other than the individuals on the same side of the waterway (Figure 6, P = 0.002) indicating a significant genetic divergence across the waterway. However, the relatedness between individuals from different blocks on the same side of the road did not differ significantly from that between individuals from opposite sides of the road (Figure 6 , P = 0.219) indicating there was no detectable genetic divergence across the road. Pairwise relatedness (RS) values between individuals from the same blocks were significantly higher than RS values between individuals from different blocks on the same or opposite sides of the road or the waterway (Figure 6 , P < 0.001).

Figure 6 Mean pairwise relatedness values (Queller and Goodnight 1989) for Pseudocheirus occidentalis individuals sampled from the same block (“Within”), different blocks on the same side of an artificial barrier (“Same”) and opposite sides of an artificial barrier (“Opposite”).

Different letters below the bars indicate significant differences found between the mean pairwise relatedness values with the Mann-Whitney U test. Error bars represent the 95% confidence intervals determined by bootstrap resampling.

Discussion

The fine-scale population structure detected in this study confirms that dispersal in P. occidentalis is limited. The maximum distance at which a significantly positive genetic structure was detected in continuous habitat with SA and MDC analyses was 600 m. This finding is consistent with the sedentary nature of this species observed in previous telemetry studies (Clarke 2011, Jones et al. 1994b, Chapter 2: Yokochi et al. 2015). Fine-scale population structure was also evident with the pairwise relatedness values, which were significantly higher between individuals from the same blocks than between 76 individuals from different blocks. Furthermore, the fine-scale genetic structuring was observed equally in all study blocks within the nature reserve, indicating that this genetic structuring was not restricted to one part of the nature reserve. Similar patterns of positive fine-scale genetic structuring have been reported in other small to medium sized mammals, such as the Australian bush rat (Rattus fuscipes, Peakall et al. 2003), the brush-tailed rock-wallaby (Petrogale penicillata, Hazlitt et al. 2004), the Eurasian badger (Meles meles, Pope et al. 2006), the southern hairy-nosed wombat (Lasiorhinus latifrons, Walker et al. 2008), and the squirrel glider (Petaurus norfolcensis, Goldingay et al. 2013). All of these species are thought to have short dispersal distances due to their philopatry or small body size, both of which may have contributed to the positive genetic structuring found in this P. occidentalis population.

SA analysis on another species of possum, the common brushtail possum (Trichosurus vulpecula) found positive spatial structure up to distances of 896 m and 454 m for males and females, respectively (Stow et al. 2006). Male-biased dispersal of up to 10 km was also recorded for this species in field studies (Clout and Efford 1984). Another arboreal marsupial, the koala (Phascolarctos cinereus) has also been recorded to disperse up to 10.6 km (Dique et al. 2003) with a spatial autocorrelation analysis showing no evidence of limited dispersal (Lee et al. 2010). In this study, we could not investigate the fine- scale genetic structure of males and females separately due to small sample sizes. Such analyses are required to test the hypothesis that dispersal in P. occidentalis is driven by males, which is based on observations of only a few individuals (Jones et al. 1994b). Nevertheless, the presence of fine-scale population structure at short geographical distances (300 - 400 m) highlights the exceptionally high level of philopatry in P. occidentalis even compared to other arboreal marsupials, highlighting their high susceptibility to the impacts of habitat fragmentation.

In support of our prediction about the effect of the artificial waterway, both the pairwise relatedness and Bayesian cluster analyses indicated that possums separated by the artificial waterway were genetically distinct, indicating that it was acting as a barrier to gene flow. The waterway was much narrower than the spatial scale at which positive genetic structure was detected within continuous vegetation, and this rules out the possibility that the genetic divergence across the waterway was purely due to geographical distance. Thus, it appears that even a 30 m wide artificial waterway is sufficient to restrict gene flow in P. occidentalis, enhancing population structure over short distances. 77

On the other hand, Caves Road did not seem to have an apparent effect on the genetic connectivity of P. occidentalis, according to the results from pairwise relatedness and Bayesian cluster analyses, even though it was found to be restricting the movements of possums in a previous study (Chapter 2: Yokochi et al. 2015). Caves Road is a busy road without canopy connection. However, it is not fenced and possum roadkills have been recorded in the area (Trimming et al. 2009), indicating that some possums try to cross it. One to ten migrants per generation are considered to be enough to maintain genetic homogeneity (Mills and Allendorf 1996), and it is possible that the small number of possums that successfully crossed the road had been sufficient to maintain enough gene flow to prevent genetic divergence. The gap in the vegetation across the road was 20 m narrower than that across the artificial waterway, which may also have contributed to the apparently higher permeability of the road. Moreover, Caves Road was bituminised only 40 to 50 years ago, with significant increases in traffic volumes occurring only recently, so its barrier effect may not have had sufficient time to result in genetic divergences detectable with the microsatellite loci used in this study. What we detect in the genetic structure of animals today reflects historic rather than present events, and depending on the circumstances, it could take tens and hundreds of generations for the effects of habitat fragmentation to become detectable (Keyghobadi 2007). Therefore, we still should not disregard the risk of this road causing genetic divergence within this important population of P. occidentalis.

Another possible explanation for the difference in the genetic effect of the road and waterway is the difference in effective population size of possums along these linear structures. A population with a smaller effective population size suffers a greater effect of genetic drift, which results in a greater rate of genetic divergence (Marsh et al. 2008). While this remains a possibility, a previous study has shown that the density of P. occidentalis within Locke Nature Reserve does not differ near Caves Road or the artificial waterway (Harring-Harris 2014). Our data also indicate that the levels of genetic diversity and inbreeding were not different among blocks. Therefore, it is unlikely that the difference in the level of genetic divergence between these barriers was caused by the effective population size alone.

In a study investigating genetic divergence among red-backed salamanders (Plethodon cinereus) occurring on different sides of roads and streams, Marsh et al. (2008, 2007) found that 2–7 m wide streams caused small, but significant genetic divergence. Amongst the roads, only one 104 m wide highway caused apparent genetic divergence 78 out of six roads examined. The other five roads (13 – 47 m) were at least 30 years older than the highway, but much narrower and not as busy. Quéméré et al. (2010) also found that the primary causes of genetic structuring in the golden-crowned sifaka (Propithecus tattersalli) were a river and geographical distances rather than a road. On the other hand, Estes-Zumpf et al. (2010) found that none of the roads or creeks they examined caused genetic divergence in the pygmy rabbit (Brachylagus idahoensis) possibly due to the greater mobility of this species. These results, together with our own, indicate that different types of linear barriers can have different levels of permeability to the movement of different species, and the permeability of barriers also depends on the mobility of the species. Therefore, it is important to assess the negative impacts of a wider range of artificial linear structures especially in areas where endangered sedentary species occur.

FST values also suggested that the groups of possums on the east of the waterway were genetically different from those on the west. The samples from block 2B were significantly different from the all other blocks, probably due to the presence of the waterway and the mostly cleared land between blocks 2B and 2D. Block 2D was significantly different from all other blocks except 1-2C. There is a road bridge across the waterway north of these two blocks (see Figure 1), and it is possible that some P. occidentalis have used this bridge to cross the waterway in the past, resulting in the non- significant pairwise FST between blocks 2D and 1-2C. It should also be noted that the sample sizes within these blocks (N = 12 for both) were considerably smaller than the others, lowering the reliability of the FST value. The pairwise FST values showed inconclusive results across the road. For example, FST values between blocks 1A and 1B and blocks 1-2C and 1D were found to be significant; however, FST values between blocks 1B and 2A and blocks 1D and 2A were non-significant despite the presence of a road and greater distances between these blocks. This inconsistency may have been caused by the small sample sizes in some of the study blocks and/or the fine scale of the genetic divergence examined in this study. Individual-level analyses, such as pairwise relatedness analyses and Bayesian cluster analyses, have been suggested to be more appropriate than population-level analyses, such as estimated paiwise FST, for studies examining fine-scale genetic divergence because combining individuals within populations can lead to loss of fine scale information (Broquet et al. 2006). Given the small geographical scale of this study, the clear results shown by both of individual-

79 level analyses are likely to be more robust than the inconsistent results shown by a population-level analysis.

All blocks had high levels of genetic variation and did not deviate from Hardy- Weinberg equilibrium, suggesting that the possums in these blocks are not experiencing a high level of inbreeding despite the presence of two barriers. This is probably because blocks were still connected with other patches of habitat outside our study area and they were not completely isolated. The different cluster membership in the Bayesian cluster analysis for blocks 2B and 2D may also be explained by these connections to other patches. Blocks 2B and 2D were isolated by the waterway from the other study blocks, and were separated from each other by a mostly cleared campsite. Given the exceptionally highly sedentary and arboreal nature of this species (Chapter 2: Yokochi et al. 2015), the limited canopy connection between these two blocks (Figure 1) may have been enough to cause this differentiation between neighbouring blocks. However, 2D had continuous canopy covers to the habitat east of the study area via a road reserve (Figure 1), possibly altering the allele frequencies in this area compared to those on the western side of the waterway. Block 2B was the most isolated among all blocks because the area to the east and south of this block was mostly cleared farmland (Figure 1), and this isolation seems to be reflected in its memberships to mostly yellow and blue genetic clusters (Figure 5) rather than mostly red and blue like the other blocks. To confirm this, DNA samples from possums in the area east and south of the campsite would need to be analysed; however, this was outside of the scope of our study.

Even though possums in our study area did not seem to suffer from a low genetic diversity or a high level of inbreeding at the time of sample collection, the barrier effects of the waterway and road still need to be mitigated to maximise their ability to disperse, given the current increasing pressure of habitat clearing. Caves Road has not caused a detectable genetic divergence among possums yet, but it is restricting their movements and causing mortality (Chapter 2: Yokochi et al. 2015), which also reduces genetic diversity over generations (Jackson and Fahrig 2011). The number and range of P. occidentalis is expected to decline dramatically due to the impacts of climate change, even without these barrier effects (Molloy et al. 2014). To restore connectivity, effective mitigation measures that provide possums with a safe passage across the waterway and road should be considered. Yokochi and Bencini (2015) found that P. occidentalis quickly habituated to a rope bridge built across Caves Road after the end of this study and multiple individuals frequently crossed it, highlighting the potential of this wildlife 80 crossing structure as an effective mitigation measure. Therefore, research into the capability of these rope bridges of restoring gene flow as well as reducing road mortality in P. occidentalis is warranted.

Conclusion

Both telemetry data from a previous study (Chapter 2: Yokochi et al. 2015) and genetic data from this study show that P. occidentalis have limited dispersal making them highly susceptible to the impacts of habitat fragmentation. Indeed, we found evidence to suggest that an artificial waterway has been limiting the gene flow of P. occidentalis, resulting in genetic divergence within spatial scales of hundreds of metres. By contrast, a busy road does not seem to have had a detectable impact on population structure, though future impacts cannot be ruled out. While studies investigating negative effects of artificial linear structures on population structure tend to focus on roads, our study provides an example where an artificial waterway poses significant genetic impacts on an endangered arboreal species. We therefore urge for more research to be conducted on the impacts of artificial linear structures other than roads. Our study shows that mitigation measures should be considered for a wide range of artificial linear structures that divide wildlife habitats.

Acknowledgments

We would like to thank the School of Animal Biology at The University of Western Australia, Main Roads Western Australia, the Western Australian Department of Parks and Wildlife, Western Power, the Satterley Property Group and the Holsworth Wildlife Research Endowment for providing funding and support for this project. We gratefully acknowledge Mr. Paul J. de Tores for providing valuable advice, support and training for the fieldwork. Yvette Hitchen at Helix Molecular Solutions Pty Ltd provided technical assistance with DNA extractions and conducted PCR and scoring of genotypes for this project. We would also like to thank the City of Busselton, Peppermint Park Eco Village, Camp Geographe, Abundant Life Centre, Christian Brethren Campsites, Legacy Camp and Possum Centre Busselton Inc. for their support and over 100 volunteers who helped us in the field.

References

Australian Bureau of Statistics (2014) 3218.0 - Regional Population Growth, Australia, 2012-13. http://www.abs.gov.au/AUSSTATS/[email protected]/Lookup/3218.0Main+Features12012- 13 [accessed 23 rd January 2015]

Broquet T, Johnson CA, Petit E, Thompson I, Burel F, Fryxell JM (2006) Dispersal and genetic structure in the American marten, Martes americana. Molecular Ecology 15: 1689-1697. doi:10.1111/j.1365-294X.2006.02878.x

Burbidge AA, McKenzie NL (1989) Patterns in the modern decline of Western Australia's vertebrate fauna: Causes and conservation implications. Biological Conservation 50: 143-198. doi:http://dx.doi.org/10.1016/0006-3207(89)90009-8

Clarke JR (2011) Translocation outcomes for the western ringtail possum (Pseudocheirus occidentalis) in the presence of the common brushtail possum (Trichosurus vulpecula) : health, survivorship and habitat use. PhD dissertation, Perth: Murdoch University.

Clark RW, Brown WS, Stechert R, Zamudio KR (2010) Roads interrupted dispersal and genetic diversity in timber rattlesnakes. Conservation Biology 24: 1059-1069. doi:10.1111/j.1523-1739.2009.01439.x

Clout MN, Efford MG (1984) Sex differences in the dispersal and settlement of brushtail possums (Trichosurus vulpecula). Journal of Animal Ecology 53: 737-749. doi:10.2307/4656 de Tores PJ, Elscot S (2010) Estimating the population size of a threatened arboreal marsupial: use of distance sampling to dispense with ad hoc survey techniques. Wildlife Research 37: 512-523. doi:doi:10.1071/WR10090

Dique DS, Thompson J, Preece HJ, de Villiers DL, Carrick FN (2003) Dispersal patterns in a regional koala population in south-east Queensland. Wildlife Research 30: 281-290. doi:http://dx.doi.org/10.1071/WR02043

Double MC, Peakall R, Beck NR, Cockburn A (2005) Dispersal, philopatry, and infidelity: dissecting local genetic structure in superb fairy-wrens (Malurus cyaneus). Evolution 59: 625-635. doi:10.2307/3448979

Earl D, von Holdt B (2012) STRUCTURE HARVESTER: a website and program for visualizing STRUCTURE output and implementing the Evanno method. Conservation Genetics Resources 4: 359-361. doi:10.1007/s12686-011-9548-7

Estes-Zumpf WA, Rachlow JL, Waits LP, Warhett KI (2010) Dispersal, gene flow, and population genetic structure in the pygmy rabbit (Brachylagus idahoensis). Journal of Mammalogy 91: 208-219. doi:10.2307/27755188

Evanno G, Regnaut S, Goudet J (2005) Detecting the number of clusters of individuals using the software structure: a simulation study. Molecular Ecology 14: 2611-2620. doi:10.1111/j.1365-294X.2005.02553.x

Forman RTT, Alexander LE (1998) Roads and their major ecological effects. Annual Review of Ecology and Systematics 29: 207-231

Frankham R (2005) Genetics and extinction. Biological Conservation 126: 131-140. doi:http://dx.doi.org/10.1016/j.biocon.2005.05.002

Goldingay RL, Harrisson KA, Taylor AC, Ball TM, Sharpe DJ, Taylor BD (2013) Fine- scale genetic response to landscape change in a gliding mammal. PLoS ONE 8: e80383. doi:10.1371/journal.pone.0080383

Goudet J. (2002) FSTAT, a program to estimate and test gene diversities and fixation indices (version 2.9.3). Available: http://www2.unil.ch/popgen/softwares/fstat.htm. Updated from Goudet J. FSTAT (Version 1.2): A computer program to calculate F- Statistics. J Hered. 1995; 86: 485–6.

Harring-Harris K (2014) The importance of preserving peppermint trees (Agonis flexuosa) in developed sites and providing alternative rest sites for western ringtail possums (Pseudocheirus occidentalis). Honours thesis, Perth, Australia: The University of Western Australia.

Hazlitt SL, Eldridge MDB, Goldizen AW (2004) Fine-scale spatial genetic correlation analyses reveal strong female philopatry within a brush-tailed rock-wallaby colony in

southeast Queensland. Molecular Ecology 13: 3621-3632. doi:10.1111/j.1365- 294X.2004.02342.x

Holderegger R, Di Giulio M (2010) The genetic effects of roads: A review of empirical evidence. Basic and Applied Ecology 11: 522-531. doi:http://dx.doi.org/10.1016/j.baae.2010.06.006

Jackson ND, Fahrig L (2011) Relative effects of road mortality and decreased connectivity on population genetic diversity. Biological Conservation 144: 3143- 3148. doi:http://dx.doi.org/10.1016/j.biocon.2011.09.010

Jakobsson M, Rosenberg NA (2007) CLUMPP: a cluster matching and permutation program for dealing with label switching and multimodality in analysis of population structure. Bioinformatics 23: 1801-1806. doi:10.1093/bioinformatics/btm233

Jones BA, How RA, Kitchener DJ (1994a) A field study of Pseudocheirus occidentalis (Marsupialia: Petauridae) I. Distribution and habitat. Wildlife Research 21: 175-187. doi:doi:10.1071/WR9940175

Jones BA, How RA, Kitchener DJ (1994b) A field study of Pseudocheirus occidentalis (Marsupialia: Petauridae) II. Population studies. Wildlife Research 21: 189-201. doi:doi:10.1071/WR9940189

Keyghobadi N (2007) The genetic implications of habitat fragmentation for animals. Canadian Journal of Zoology 85: 1049-1064. doi:10.1139/z07-095

Lee T, Zenger KR, Close RL, Jones M, Phalen DN (2010) Defining spatial genetic structure and management units for vulnerable koala (Phascolarctos cinereus) populations in the Sydney region, Australia. Wildlife Research 37: 156-165. doi:http://dx.doi.org/10.1071/WR09134

Main Roads Western Australia (2004) Statewide Traffic Digest 2003/04 - 2008/09. http://www.mainroads.wa.gov.au/OurRoads/Facts/TrafficData/Pages/default.aspx#.U wMF2mKSySp [accessed 5th February 2014].

Mills LS, Allendorf FW (1996) The One-migrant-per-generation rule in conservation and management. Conservation Biology 10: 1509-1518. doi:10.2307/2387022

Molloy SW, Davis RA, Van Etten EJB (2014) Species distribution modelling using bioclimatic variables to determine the impacts of a changing climate on the western ringtail possum (Pseudocheirus occidentals; Pseudocheiridae). Environmental Conservation 41: 176-186. doi:doi:10.1017/S0376892913000337

Myers N, Mittermeier RA, Mittermeier CG, da Fonseca GAB, Kent J (2000) Biodiversity hotspots for conservation priorities. Nature 403: 853-858. doi:doi:10.1038/35002501

Nei M (1978) Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 89: 583-590

National Health and Medical Research Council (2004) Australian code of practice for the care and use of animals for scientific purposes. 7th ed. Australian Government, Canberra, Australia: 1–84.

Peakall R, Ruibal M, Lindenmayer DB (2003) Spatial autocorrelation analysis offers new insights into gene flow in the Australian bush rat, Rattus fuscipes. Evolution 57: 1182-1195. doi:10.2307/3448817

Peakall R, Smouse PE (2012) GenAlEx 6.5: genetic analysis in Excel. Population genetic software for teaching and research—an update. Bioinformatics 28: 2537- 2539. doi:10.1093/bioinformatics/bts460

Pope LC, Domingo-Roura X, Erven K, Burke T (2006) Isolation by distance and gene flow in the Eurasian badger (Meles meles) at both a local and broad scale. Molecular Ecology 15: 371-386. doi:10.1111/j.1365-294X.2005.02815.x

Pritchard JK, Stephens M, Donnelly P (2000) Inference of population structure using multilocus genotype data. Genetics 155: 945-959

Queller DC, Goodnight KF (1989) Estimating relatedness using genetic markers. Evolution 43: 258-275. doi:10.2307/2409206

R Depelopment Core Team (2013) R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria.

Rosenberg NA (2004) distruct: a program for the graphical display of population structure. Molecular Ecology Notes 4: 137-138. doi:10.1046/j.1471- 8286.2003.00566.x

Seutin G, White BN, Boag PT (1991) Preservation of avian blood and tissue samples for DNA analyses. Canadian Journal of Zoology 69: 82-90. doi:10.1139/z91-013

Stow AJ, Minarovic N, Eymann J, Cooper DW, Webley LS (2006) Genetic structure infers generally high philopatry and male-biased dispersal of brushtail possums (Trichosurus vulpecula) in urban Australia. Wildlife Research 33: 409-415. doi:http://dx.doi.org/10.1071/WR06019

Taylor AC, Walker FM, Goldingay RL, Ball T, van der Ree R (2011) Degree of landscape fragmentation influences genetic isolation among populations of a gliding mammal. PLoS ONE 6: e26651. doi:10.1371/journal.pone.0026651

Trimming E, Chambers B, Grillo D, de Tores P, Bencini R (2009) Road kills of the western ringtail possum (Pseudocheirus occidentalis) occur at specific hotspots. Semi-centenary and 55th Meeting of the Australian Mammal Society. Australian Mammal Society, Perth, Australia, 24.

Van Oosterhout C, Hutchinson WF, Wills DPM, Shipley P (2004) Micro-checker: software for identifying and correcting genotyping errors in microsatellite data. Molecular Ecology Notes 4: 535-538. doi:10.1111/j.1471-8286.2004.00684.x

Walker FM, Sunnucks P, Taylor AC (2008) Evidence for habitat fragmentation altering within-population processes in wombats. Molecular Ecology 17: 1674-1684. doi:10.1111/j.1365-294X.2008.03701.x

Wilson K, de Tores P, Spencer P (2009) Isolation and characterisation of polymorphic microsatellite markers in the western ringtail possum, Pseudocheirus occidentalis. Conservation Genetics Resources 1: 123-125. doi:doi:10.1007/s12686-009-9030-y

Woinarski JC, Burbidge AA, Harrison PL (2014) The action plan for Australian mammals 2012. CSIRO Publishing, Collingwood, Australia.

Yokochi K, Bencini R (2015) A remarkably quick habituation and high use of a rope bridge by an endangered marsupial, the western ringtail possum. Nature Conservation 11: 79-94. doi:10.3897/natureconservation.11.4385

Yokochi K, Chambers B, Bencini R (2015) An artificial waterway and road restrict movements and alter home ranges of endangered arboreal marsupial. Journal of Mammalogy: doi:10.1093/jmammal/gyv137

Appendix 1

Table A1 Multilocus genotypes of the 145 adult Pseudocheirus occidentalis used in the study. The microsatellite loci (A1, A127, A106, A6, A2, D113, A122, D104, C111, A119, D114 and B104) are described in Wilson et al. (2009).

Individual Block Microsatellite loci A1 A127 A106 A6 A2 D113 1 CR1-2C 162 162 249 249 211 211 232 232 162 172 231 231 2 CR1-2C 162 172 246 249 211 219 240 242 168 168 231 231 3 CR1-2C 162 175 249 249 211 237 232 242 166 168 231 231 4 CR1-2C 162 172 246 249 211 237 232 232 168 172 227 231 5 CR1-2C 162 172 249 249 229 237 232 242 168 168 227 231 6 CR1-2C 162 162 249 249 211 227 232 232 168 172 227 227 7 CR1-2C 172 175 246 246 211 229 232 232 168 172 227 227 8 CR1-2C 172 172 246 249 237 237 232 232 162 172 227 231 9 CR1-2C 162 172 246 249 211 237 232 242 168 172 227 231 10 CR1-2C 162 172 249 249 211 229 232 242 168 168 227 231 11 CR1-2C 162 172 246 246 219 229 240 240 162 168 222 231 12 CR1-2C 162 175 246 249 211 211 232 232 166 168 227 231 13 CR1A 172 172 249 249 211 211 232 249 168 168 227 231 14 CR1A 162 172 246 249 211 211 232 242 166 172 227 231 15 CR1A 162 175 246 249 211 237 232 232 168 172 231 231 16 CR1A 162 175 249 249 211 211 232 232 168 168 227 231 17 CR1A 162 170 246 249 211 237 232 232 168 172 227 231 18 CR1A 170 175 246 249 211 237 242 242 168 172 227 231 19 CR1A 162 172 246 249 211 211 232 240 168 172 227 231 20 CR1A 172 172 246 246 211 211 232 232 172 172 227 231 21 CR1A 162 175 246 249 211 229 232 232 168 172 227 231 22 CR1A 162 162 249 249 211 211 232 244 172 172 231 231 23 CR1A 162 175 249 249 211 211 232 247 168 172 231 231 24 CR1A 162 162 249 249 211 211 232 247 168 168 227 227 25 CR1A 162 162 249 249 237 237 232 232 172 172 227 227 26 CR1A 162 175 249 249 211 211 232 242 172 172 227 235 27 CR1A 162 175 249 249 211 229 240 242 168 168 227 227 28 CR1A 162 175 249 249 211 229 232 240 172 172 227 231 29 CR1A 162 172 249 249 211 211 232 247 168 172 227 227 30 CR1A 162 162 249 249 211 229 240 247 166 168 227 227 31 CR1A 172 175 249 249 211 211 232 242 172 172 227 231 32 CR1A 175 175 246 249 211 211 232 232 168 168 227 227 33 CR1A 172 175 246 249 211 211 232 242 168 172 227 227 34 CR1A 162 172 249 249 211 211 232 232 168 172 231 231

35 CR1A 172 172 249 249 211 237 232 232 168 168 227 231 36 CR1A 172 172 246 249 211 229 232 242 168 172 231 231 37 CR1A 162 162 246 249 211 211 232 232 168 168 231 231 38 CR1A 162 175 249 249 211 211 232 242 172 172 227 231 39 CR1B 175 175 246 249 211 229 232 242 166 172 231 235 40 CR1B 162 170 249 249 211 229 232 242 172 172 227 231 41 CR1B 162 175 249 249 211 221 232 232 168 172 227 231 42 CR1B 162 172 246 249 225 227 232 240 168 168 231 231 43 CR1B 162 172 249 249 227 229 242 242 168 172 227 227 44 CR1B 162 170 246 249 211 229 232 232 172 172 231 231 45 CR1B 172 172 246 246 211 229 232 240 168 172 227 231 46 CR1B 162 162 249 249 211 237 232 232 172 172 231 231 47 CR1B 162 170 246 249 229 229 232 244 172 172 231 235 48 CR1B 164 172 246 249 211 229 232 247 168 172 227 231 49 CR1B 170 172 249 249 211 229 232 232 168 172 227 231 50 CR1B 162 172 246 249 211 211 232 247 162 172 227 231 51 CR1B 172 175 249 249 211 221 232 240 168 172 222 231 52 CR1B 170 172 249 249 211 227 242 242 172 172 222 227 53 CR1B 162 175 246 249 211 229 232 232 162 172 227 227 54 CR1B 162 172 249 249 229 237 232 242 172 172 227 231 55 CR1B 162 162 246 249 211 211 232 240 172 172 231 231 56 CR1B 175 175 246 249 227 237 232 242 168 172 227 231 57 CR1B 175 175 249 249 229 229 232 242 166 172 227 235 58 CR1B 172 172 246 246 211 237 232 240 172 172 231 231 59 CR1B 162 172 246 249 229 229 232 240 168 172 227 227 60 CR1B 162 172 246 249 227 237 232 242 172 172 222 227 61 CR1B 162 172 249 249 211 229 232 244 172 172 231 231 62 CR1B 162 172 249 249 229 237 232 232 172 172 231 231 63 CR1B 162 162 246 249 229 229 232 242 172 172 227 227 64 CR1B 162 170 246 246 237 237 232 242 172 172 227 231 65 CR1B 162 162 246 246 237 237 232 232 172 172 227 231 66 CR1B 162 172 249 249 211 227 232 242 168 172 227 231 67 CR1B 162 172 249 249 211 227 232 242 162 172 231 231 68 CR1B 162 172 249 249 211 237 232 247 172 172 231 231 69 CR1B 170 175 246 249 229 229 232 244 162 172 227 227 70 CR1B 162 162 246 249 211 229 232 247 162 172 227 231 71 CR1D 172 175 249 249 211 237 242 247 168 172 231 231 72 CR1D 175 175 246 249 211 221 232 232 172 172 227 231 73 CR1D 162 172 249 249 211 237 232 242 172 172 227 227 74 CR1D 162 162 249 249 211 211 232 247 168 172 227 231 75 CR1D 172 175 249 249 211 211 232 244 162 168 227 231 76 CR1D 162 175 246 249 221 229 232 247 168 172 231 231 89

77 CR1D 162 172 246 249 211 211 232 242 168 172 227 231 78 CR1D 162 172 249 249 211 229 232 238 162 168 227 227 79 CR1D 170 170 249 249 211 211 232 238 168 172 227 227 80 CR1D 172 172 249 249 211 211 232 242 172 172 227 231 81 CR2A 170 172 249 249 211 211 232 242 166 168 227 231 82 CR2A 175 175 246 249 211 229 232 247 168 168 227 231 83 CR2A 175 175 249 249 211 211 232 242 166 168 227 231 84 CR2A 162 172 246 246 211 237 232 242 168 168 222 227 85 CR2A 162 162 249 249 211 227 232 232 168 168 227 231 86 CR2A 162 175 249 249 211 211 244 247 172 172 227 227 87 CR2A 162 162 249 249 211 229 240 247 166 168 227 227 88 CR2A 162 162 249 249 211 229 232 232 172 172 231 235 89 CR2A 162 162 249 249 211 211 232 247 172 172 231 231 90 CR2A 162 175 249 249 211 227 232 247 162 172 227 231 91 CR2A 162 162 249 249 211 211 232 240 166 172 227 227 92 CR2A 162 162 246 249 211 211 242 247 168 172 227 231 93 CR2A 162 162 246 249 211 237 232 247 172 172 227 231 94 CR2A 162 162 249 249 211 227 232 247 168 172 227 227 95 CR2A 175 175 249 249 211 229 232 232 168 172 227 227 96 CR2A 172 172 246 249 237 237 232 232 168 172 222 231 97 CR2A 162 175 249 249 229 237 232 247 168 172 227 231 98 CR2A 162 170 249 249 211 229 232 232 168 168 227 231 99 CR2A 162 175 249 249 227 237 232 247 162 172 227 227 100 CR2A 162 162 249 249 211 229 232 232 168 172 227 231 101 CR2A 162 162 249 249 211 229 232 232 172 172 231 231 102 CR2A 162 162 249 249 227 229 240 247 166 172 227 227 103 CR2A 162 162 249 249 211 227 232 247 162 172 227 231 104 CR2A 162 175 246 249 237 237 232 242 168 168 231 231 105 CR2A 175 175 246 249 211 237 238 247 166 168 231 231 106 CR2A 162 175 246 249 237 237 232 232 168 172 231 231 107 CR2A 162 170 249 249 211 227 232 232 162 168 227 227 108 CR2A 162 172 249 249 229 237 232 247 172 172 231 231 109 CR2A 162 162 249 249 229 237 232 242 162 168 231 231 110 CR2A 172 175 249 249 211 229 244 247 172 172 227 231 111 CR2A 172 175 246 249 211 211 242 247 168 172 227 231 112 CR2B 162 162 246 249 211 229 232 242 166 172 227 231 113 CR2B 172 175 246 249 211 211 232 249 166 168 231 231 114 CR2B 162 170 246 249 227 229 240 240 168 172 227 231 115 CR2B 162 175 249 249 211 229 232 249 168 172 227 231 116 CR2B 162 172 249 249 211 211 232 249 172 172 222 227 117 CR2B 162 162 246 249 211 211 242 249 168 172 222 227 118 CR2B 172 175 249 249 211 211 232 244 168 168 227 227 90

119 CR2B 162 172 246 249 211 227 232 232 168 168 227 231 120 CR2B 162 162 246 246 229 229 232 240 168 172 227 227 121 CR2B 175 175 246 249 211 211 232 232 166 168 231 231 122 CR2B 162 170 249 249 211 211 232 244 166 168 227 231 123 CR2B 162 175 246 249 211 211 232 232 166 172 222 227 124 CR2B 162 175 249 249 211 229 232 249 172 172 227 231 125 CR2B 162 172 246 246 211 211 232 232 168 168 222 231 126 CR2B 162 162 246 249 211 211 232 232 168 168 231 231 127 CR2B 162 172 249 249 211 227 232 242 166 168 231 231 128 CR2B 170 172 249 249 211 227 232 244 168 172 227 231 129 CR2B 172 172 246 249 211 211 232 249 166 172 227 231 130 CR2B 162 172 249 249 211 237 249 249 168 172 222 227 131 CR2B 172 175 249 249 227 229 232 232 168 172 227 231 132 CR2B 162 175 249 249 211 227 232 240 168 172 227 227 133 CR2B 162 162 246 246 229 229 240 240 168 172 227 227 134 CR2D 162 162 246 249 229 229 240 240 166 172 222 231 135 CR2D 162 162 246 249 211 229 232 240 168 168 222 231 136 CR2D 162 162 246 249 211 229 240 240 168 172 222 231 137 CR2D 162 162 246 246 211 237 232 247 166 168 222 231 138 CR2D 162 162 249 249 229 233 232 242 168 172 227 231 139 CR2D 162 162 246 246 211 239 240 240 168 168 222 231 140 CR2D 170 170 249 249 211 229 232 244 168 168 227 231 141 CR2D 162 172 246 249 227 229 232 240 172 172 227 227 142 CR2D 162 175 249 249 229 229 232 240 164 166 227 227 143 CR2D 162 162 246 246 229 229 240 242 168 168 227 227 144 CR2D 162 162 246 249 211 229 242 242 168 168 222 231 145 CR2D 164 175 249 249 211 211 232 240 172 172 227 227

Individual Block Microsatellite loci A122 D104 C111 A119 D114 B104 1 CR1-2C 209 212 205 209 207 211 283 283 269 273 277 277 2 CR1-2C 209 212 206 213 207 211 279 279 273 273 277 277 3 CR1-2C 209 212 205 209 207 211 279 289 250 273 266 273 4 CR1-2C 205 205 205 213 211 211 279 291 250 254 270 277 5 CR1-2C 205 209 209 213 207 207 279 279 250 273 266 277 6 CR1-2C 205 212 205 213 207 211 279 283 254 257 266 277 7 CR1-2C 205 209 213 213 211 211 279 279 250 257 273 273 8 CR1-2C 205 210 213 213 207 211 279 291 250 250 270 277 9 CR1-2C 205 205 205 213 207 211 279 279 254 257 270 273 10 CR1-2C 205 210 205 213 211 211 279 279 257 273 266 273 11 CR1-2C 209 212 213 213 211 211 279 279 273 273 266 277 12 CR1-2C 205 205 205 205 207 211 289 291 250 273 273 277 13 CR1A 205 209 205 205 211 211 279 291 250 257 270 273 14 CR1A 205 212 205 209 207 211 279 279 254 257 273 277 15 CR1A 212 212 205 213 207 211 291 291 269 273 273 273 16 CR1A 210 212 205 205 207 207 291 291 250 273 266 270 17 CR1A 212 212 205 205 207 211 291 291 254 269 273 273 18 CR1A 205 212 205 205 207 211 279 291 250 254 273 273 19 CR1A 205 212 205 213 207 211 291 291 269 273 270 273 20 CR1A 205 212 205 213 207 211 279 291 257 269 270 277 21 CR1A 205 212 205 205 207 211 279 279 250 254 273 273 22 CR1A 205 206 209 213 207 211 279 291 269 269 273 277 23 CR1A 205 212 209 213 207 211 291 291 254 273 273 273 24 CR1A 205 212 205 213 207 211 279 291 254 269 270 273 25 CR1A 205 212 205 205 207 207 283 291 269 269 273 277 26 CR1A 205 210 204 205 207 207 279 291 273 277 266 273 27 CR1A 210 212 205 213 211 211 279 291 250 273 270 273 28 CR1A 210 212 205 205 207 207 279 291 250 257 266 273 29 CR1A 205 210 205 213 207 207 279 291 250 257 266 273 30 CR1A 205 205 205 213 207 211 279 291 250 254 273 277 31 CR1A 205 210 205 205 207 207 279 291 250 257 273 273 32 CR1A 210 212 204 213 211 211 279 289 250 250 273 277 33 CR1A 210 212 205 213 211 211 279 279 250 257 270 277 34 CR1A 206 209 205 209 207 211 279 291 257 269 273 273 35 CR1A 210 212 205 205 207 211 279 291 257 269 273 277 36 CR1A 205 209 205 205 211 211 279 291 250 257 273 273 37 CR1A 206 212 206 206 211 211 291 291 269 269 273 273 38 CR1A 205 205 204 205 211 211 279 289 250 250 273 277 39 CR1A 205 209 205 205 207 211 279 283 250 269 270 270

40 CR1A 205 205 205 213 207 207 283 283 269 273 273 277 41 CR1B 209 212 213 213 207 207 289 291 250 273 273 273 42 CR1B 209 209 205 209 207 211 289 289 250 277 273 273 43 CR1B 205 205 205 213 207 211 279 283 250 250 273 277 44 CR1B 205 212 205 205 207 211 279 291 254 254 273 273 45 CR1B 205 210 205 213 207 211 283 291 250 257 273 273 46 CR1B 212 212 205 205 211 211 279 279 250 269 273 277 47 CR1B 205 212 205 209 207 211 279 291 254 257 273 277 48 CR1B 209 212 205 205 207 207 289 291 250 257 266 273 49 CR1B 205 210 205 205 207 211 279 279 254 257 273 273 50 CR1B 205 210 205 205 207 207 279 279 257 269 273 273 51 CR1B 209 212 209 213 207 207 289 291 250 257 273 277 52 CR1B 205 205 205 209 207 211 279 279 250 254 266 273 53 CR1B 205 212 205 205 211 211 279 279 250 273 270 273 54 CR1B 212 212 205 205 207 207 279 279 257 269 270 273 55 CR1B 209 212 205 205 207 219 279 279 269 273 270 273 56 CR1B 209 212 209 213 211 211 291 291 250 250 273 273 57 CR1B 209 212 205 205 207 211 279 291 250 269 270 273 58 CR1B 212 212 205 209 207 207 291 291 250 250 273 273 59 CR1B 205 209 205 209 211 211 279 279 250 257 270 277 60 CR1B 205 212 209 213 207 211 279 279 250 254 266 277 61 CR1B 205 209 205 213 207 211 279 279 257 273 270 277 62 CR1B 205 212 205 205 207 211 279 291 250 269 273 277 63 CR1B 205 212 205 205 207 211 279 279 250 254 273 277 64 CR1B 212 212 205 205 207 211 279 279 254 254 273 277 65 CR1B 205 212 205 205 207 211 279 279 254 254 273 273 66 CR1B 209 212 205 209 207 211 279 283 250 257 270 273 67 CR1B 209 210 205 205 207 207 279 279 257 273 273 273 68 CR1B 210 212 205 205 207 211 279 279 257 269 273 273 69 CR1B 205 212 201 205 211 211 279 279 250 254 270 273 70 CR1B 205 205 205 205 207 207 279 291 269 269 273 277 71 CR1B 205 212 205 205 211 211 291 291 250 250 273 277 72 CR1B 212 212 205 213 207 211 279 291 250 250 273 273 73 CR1B 206 210 205 205 211 211 279 279 257 269 266 277 74 CR1B 205 206 205 205 207 211 279 283 250 269 273 277 75 CR1D 205 212 205 205 207 211 279 291 250 257 266 273 76 CR1D 205 212 205 205 207 207 279 283 250 250 273 273 77 CR1D 205 205 205 205 211 211 279 283 257 269 277 277 78 CR1D 205 212 205 205 207 211 279 291 254 254 273 277 79 CR1D 205 209 205 209 211 211 283 283 250 254 273 277 80 CR1D 205 206 205 205 211 211 279 283 257 257 277 277 81 CR1D 205 210 205 213 207 211 279 279 254 257 266 273 93

82 CR1D 209 209 206 209 207 207 279 291 250 257 266 270 83 CR1D 209 209 205 205 211 211 279 291 250 250 273 277 84 CR1D 205 209 205 205 207 211 279 279 250 269 273 277 85 CR2A 209 212 205 213 207 211 289 291 257 273 270 273 86 CR2A 210 210 205 213 207 207 289 291 250 269 273 277 87 CR2A 205 205 205 205 211 211 279 291 250 273 273 273 88 CR2A 205 212 205 205 207 211 289 291 250 250 273 277 89 CR2A 205 209 206 213 207 207 279 279 250 254 273 277 90 CR2A 209 212 206 209 207 211 279 279 250 269 273 273 91 CR2A 205 210 205 209 211 211 279 291 254 273 266 270 92 CR2A 205 209 205 205 211 211 279 279 254 269 273 277 93 CR2A 205 212 205 209 207 211 279 283 273 277 273 277 94 CR2A 209 212 205 205 211 211 289 291 257 273 273 273 95 CR2A 205 212 205 209 211 211 279 291 250 269 273 273 96 CR2A 210 212 205 213 207 207 279 289 250 250 273 273 97 CR2A 205 210 205 205 207 211 279 291 250 254 273 277 98 CR2A 210 210 205 205 207 207 279 279 254 254 266 273 99 CR2A 205 212 205 209 207 211 279 289 269 273 273 273 100 CR2A 205 210 205 213 207 207 279 291 254 254 273 277 101 CR2A 205 205 205 205 207 211 289 291 250 254 273 273 102 CR2A 205 205 205 205 207 211 279 291 250 273 273 273 103 CR2A 205 212 205 206 207 207 279 279 250 273 273 273 104 CR2A 205 205 205 205 207 211 279 291 250 273 277 277 105 CR2A 205 205 205 209 211 211 291 291 250 250 266 277 106 CR2A 205 212 205 209 207 211 279 279 273 273 273 277 107 CR2A 205 209 205 213 207 211 283 291 261 269 270 273 108 CR2A 212 212 209 209 207 211 279 279 269 277 273 273 109 CR2A 209 212 205 213 207 207 279 283 250 273 270 273 110 CR2A 210 212 210 210 207 207 279 289 269 269 273 277 111 CR2A 205 209 205 206 207 211 283 291 257 257 270 273 112 CR2A 205 210 205 213 207 211 289 291 254 257 270 270 113 CR2A 205 212 205 209 207 211 291 291 250 257 273 273 114 CR2A 205 212 205 205 207 211 283 291 250 273 273 273 115 CR2A 205 212 205 205 211 211 283 291 250 250 266 273 116 CR2B 205 215 205 205 211 211 291 291 250 257 273 273 117 CR2B 205 212 205 205 211 211 283 291 250 273 266 273 118 CR2B 209 209 205 209 211 211 291 291 254 273 273 273 119 CR2B 205 205 209 209 207 211 279 283 250 273 270 273 120 CR2B 209 212 205 205 211 211 291 291 254 273 273 273 121 CR2B 205 205 205 209 207 211 279 291 250 250 273 277 122 CR2B 209 210 205 213 207 211 289 291 257 273 270 273 123 CR2B 205 209 213 213 211 211 289 291 254 257 273 273 94

124 CR2B 205 215 206 206 211 211 291 291 250 250 273 273 125 CR2B 205 210 205 209 211 211 279 291 254 273 273 277 126 CR2B 205 212 205 209 207 211 279 283 254 273 273 277 127 CR2B 205 209 205 205 207 211 289 291 250 254 273 277 128 CR2B 205 209 205 209 207 211 291 291 254 269 273 277 129 CR2B 205 215 205 209 211 211 279 291 257 257 270 273 130 CR2B 209 215 205 205 211 211 291 291 257 273 273 273 131 CR2B 205 212 205 205 207 211 291 291 250 250 273 273 132 CR2B 205 209 205 213 211 211 289 291 269 277 266 270 133 CR2B 205 209 205 205 207 211 279 283 254 254 270 281 134 CR2B 205 210 205 213 207 211 279 289 250 269 273 273 135 CR2B 209 212 205 213 207 211 279 283 254 273 273 273 136 CR2B 210 212 213 213 207 211 279 283 250 273 266 273 137 CR2B 210 212 205 205 207 211 283 291 273 273 270 273 138 CR2D 205 212 205 209 211 219 291 291 250 254 266 277 139 CR2D 209 212 213 213 207 211 283 291 250 254 273 273 140 CR2D 210 212 205 209 207 211 279 291 254 257 266 277 141 CR2D 205 210 205 205 211 211 279 289 250 254 273 277 142 CR2D 205 212 205 205 207 211 279 279 250 273 270 277 143 CR2D 205 205 205 205 207 207 279 289 254 273 273 281 144 CR2D 209 212 205 205 211 211 279 291 250 250 270 270 145 CR2D 210 210 205 205 207 207 279 289 257 273 266 273

Chapter 4

A predicted sharp decline of a stronghold population of the western ringtail possum calls for urgent reduction in fox predation and road mortality

This chapter will be separated into two scientific papers for publications: one on the negative effect of radio collars on the survival of possums and the other on the population viability analysis. 97

A predicted sharp decline of a stronghold population of the western ringtail possum calls for urgent reduction in fox predation and road mortality

Kaori Yokochi, Robert Black, Brian K. Chambers, Roberta Bencini

Abstract

The western ringtail possum (Pseudocheirus occidentalis) was recently classified as endangered; however, knowledge of its life cycle or population viability is very limited. For the first time, we projected the near future change in a stronghold population of this species in Busselton, Western Australia, using Population Viability Analyses (PVA) based on survival and fecundity data collected over three years of continuous monitoring. We found that this population has experienced a decline in recent years, and models of the population predicted a 92.1 % probability of extinction within 20 years. Fox predation was the most common cause of mortality in radio collared adult possums and contributed to 16 cases out of 23 confirmed mortalities. Road mortality accounted for the mortalities of two collared possums. Removal of the effects of fox predation on adult and pouch young survival rates dramatically reduced the probability of extinction to 0.4 %. Removal of the effects of road mortality also reduced the probability of extinction to 31.8 %. These simulated results indicate that these management options are likely to effectively slow down the decline of this population. However, this population was predicted to decline even in the fox removal scenario, suggesting that conservation efforts to increase other demographic parameters, such as fecundity rate, are also necessary to stop the decline. Our results emphasise the alarmingly poor outlook for this species and call for an urgent and more intense management of its threatening processes. In the process of PVA, we unexpectedly found that the weight of radio collars affected the survival of adult possums even though the collars used were only up to 2.7 % of the animal’s body weight, about half of the limit set by the current recommendation. Therefore, we call for more research on the impacts of collars on P. occidentalis and other strictly arboreal specialist, and review of the current recommended limit. In the mean time, researchers should use the lightest possible transmitters on P. occidentalis.

Introduction

Predicting the future trend of wildlife populations can provide researchers and managers with valuable information on how to ensure the survival of populations and/or species. Population viability analysis (PVA) has become a popular tool to make such predictions because it can provide quantitative estimates of population trends, such as probability of extinction, sensitivity of the population to a change in vital rates, and predicted effectiveness of management strategies, while incorporating uncertainties such as demographic and environmental stochasticity into the projections (Akçakaya and Sjögren-Gulve 2000). Detailed information on survival and fecundity of the target species is required for an effective PVA; however, this information is hard to obtain for some species especially if they are rare and/or elusive.

The western ringtail possum (Pseudocheirus occidentalis) is one such species because its capture rate by conventional trapping is extremely low due to their arboreal and folivorous nature (Wayne et al. 2005a). Because of this, studies on its population viability have never been conducted despite its recent classification as endangered (Department of Parks and Wildlife WA 2014a). Its numbers and range have declined dramatically in recent years. The causes of this decline have been postulated to be destruction and fragmentation of habitats, predation by introduced European foxes (Vulpes vulpes) and feral cats (Felis catus), and road mortality (Trimming et al. 2009, Clarke 2011, Woinarski et al. 2014). The species has a restricted geographic range in the southwest of Western Australia, and the Busselton region holds one of the few stronghold populations left for this species (Jones et al. 1994a). However, this coastal region is currently one of the fastest developing regions in Australia (Australian Bureau of Statistics 2009), and an increasing number of large-scale developments threaten the persistence of this species. Furthermore, the southwest of Western Australia has been experiencing drier and hotter conditions for the last decade, and these conditions are predicted to worsen in the next 50 years due to climate change (Indian Ocean Climate Initiative 2012). This change is likely to impact P. occidentalis adversely as this species is not well adapted to dry, hot climates. Yin (2006) found that P. occidentalis is unable to lose heat effectively at temperatures above 35 °C because evaporative heat loss through fur licking is the main strategy against overheating. In coastal regions, western ringtail possums are highly dependent on the foliage of peppermint trees (Agonis flexuosa), from which they obtain most of their water intake (Jones et al. 1994b). However, during extended period of hot weather, the possums have been observed to 99 descend to the ground to drink water and to seek refuge in understorey vegetation, which increases the risk of predation by ground dwelling predators such as foxes (Jones et al. 1994b, Department of Parks and Wildlife WA 2014b). In inland habitat dominated by jarrah (Eucalyptus marginata), long-term spotlight surveys have revealed up to 99 % decline in the detection rate of P. occidentalis between 1998 and 2009 (Wayne et al. 2012). A decline of this species in coastal regions has also been suggested but not confirmed due to a lack of systemised and continuous monitoring of populations (Woinarski et al. 2014).

By fitting multiple individuals from a population in Busselton with radio collars and monitoring them for over three years, we were able to estimate survival and fecundity rates of P. occidentalis and to construct its life cycle graph and transition matrices for the first time to build population viability models. These models allowed us to identify the causes of mortality and factors influencing their survival, and as a result, estimate the recent state and future direction of the population. We also simulated the effects of removal of foxes and road mortalities on the population. Based on available information on this species and region, we hypothesised that the future projection of a P. occidentalis population in Busselton would show a declining trend. Because fox predation and roads are known to cause mortality in the possums, we also expected that removing fox predation and road mortality would ease the predicted decline.

Materials and methods

Study area

We studied a population of P. occidentalis in Locke Nature Reserve and surrounding campsites, located about 9 km West of Busselton, Western Australia (33°39'S 115°14'E). This area supports one of the highest densities of P. occidentalis (Jones et al. 2007, de Tores and Elscot 2010). The 200 ha reserve provides mostly continuous canopy cover except for a few fire breaks and a swampy area in the southern part. The Western Australian Department of Parks and Wildlife manages the reserve, and no recreational activities are permitted inside. The reserve is baited monthly with kangaroo meat containing sodium monofluroacetate as part of the Western Shield program implemented to eradicate or reduce the fox population (de Tores et al. 2004). These baits are unlikely to directly affect P. occidentalis due to this species’ strictly arboreal nature and folivorous diet. Campsites are located to the north of the reserve across the 15 m wide Caves Road and to the west across an artificially widened and straightened 100 part of the Buayanyup River, which is a 30 m wide “artificial waterway” (Figure 1). Due to the presence of irrigated grass areas for recreational use, the canopy connection in the campsites is limited compared with the nature reserve. We set up seven 200 m x 200 m blocks (1A, 1B, 1-2C, 1D, 2A, 2B and 2D) in this area to study the P. occidentalis population (Figure 1).

Figure 1 A map of the study area near Busselton, Western Australia. Red lines represent the edges of Caves Road and blue lines represent the edges of a 30 m wide artificial waterway. Areas enclosed by black lines are 200 m x 200 m study blocks where Pseudocheirus occidentalis were caught and monitored.

Data collection

117 female and 97 male possums were captured in study blocks using a specially modified tranquiliser dart gun with darts containing a dose of 11-12 mg/kg of Zoletil 100® (Virbac Australia, Milperra, NSW Australia) following a method developed by P.

101 de Tores and reported by Clarke (2011). We assessed their reproductive status by measuring testes for males and checking the pouch for young and evidence of reproduction, such as elongated teats and enlarged mammary glands for females. When a pouch young was present, its body length including its head, presence of fur, and its ability to open its eyes were recorded in order to estimate its age and therefore its birth month.

We initially fitted VHF radio collars with a mortality function (AVM Instrument Company, Ltd., Colfax, California USA, or Biotrack, Wareham, UK) to three adult females and three adult males in each of four blocks: 1A, 1B, 2A and 2D (24 individuals in total). All collars used in our study had the same design: a brass loop band acting as an antenna and a transmitter and a battery coated with epoxy resin. We recorded survival of each collared animal once weekly or fortnightly at an average time span between monitoring of 9.33  0.37 (s.e.) days by homing in on the signal from the radio collars and by directly sighting each animal at night. We monitored the possums on a total of 130 occasions from March 2010 to July 2013. We also recorded the presence of any companions, such as mates and/or young at foot at every sighting. The number of monitored animals fluctuated throughout the study period due to mortality and failure of some transmitters. If an animal died or we failed to locate and recapture the animal after a transmitter failure, another animal of the same sex was captured and collared in the same block. Fifty-three individuals (22 females and 31 males) weighing between 795.2 and 1307.0 g were monitored in total. The weight of collars varied between 15.8 and 22.6 g (1.2 – 2.7 % of body weight), depending on the model of the collar and the size of the individual as larger individuals required longer bands.

When a collared animal was found dead, the likely cause of death was established from available evidence. For example, the cause of mortality was recorded as road mortality if a body was located on the road and it showed obvious signs of a motor vehicle collision (e.g. the carcass was flattened) or a sign of blunt trauma identified through post mortem examination by a veterinarian. The cause of mortality was recorded as a predation event if the body had obviously been torn apart, eaten or buried. In cases of suspected predation events, swab samples of saliva were taken from tooth marks on the collar or chewed parts of the body. We used a pool of DNA markers that are specific to the red fox, cat, dog and the western quoll (Dasyurus geoffroii) to genetically identify the predator species from the swab samples. We performed melt-curve analyses, which utilises a specific dye with melting temperature that varies depending on the length of 102 the DNA fragments. By assessing at which temperature the dye disappears, scientists can identify which species-specific marker amplified the particular DNA fragment (Berry and Sarre 2007). We cannot be completely certain whether the saliva was left on the body because the predator preyed on the possum or because it scavenged on an already deceased possum; however, foxes and cats are known to hunt and prey on possum species, including P. occidentalis (Kinnear et al. 2002, Clarke 2011, Yokochi, personal observation), so we assumed that the mortality was caused by direct predation. Post mortems were also performed on presumably predated carcasses to confirm the signs of predation if bodies were still present and intact.

Survivorship analysis

Using survival data from radio telemetry monitoring, we constructed candidate models with factors that potentially influenced the survival rate of adult possums. We conducted a survivorship analysis using the known fate model with logit link function in program MARK v.7.1 (White 1999). Factors incorporated in candidate models were sex, type of closest barrier (road or waterway), habitat type (nature reserve or campsite), year, life cycle season, average rainfall, body weight and weight of collars. We counted a year from May to April, and life cycle seasons were set as May to October (season 1) and November to April (season 2) so that two peaks in breeding cycles of P. occidentalis fell at the beginning of these seasons (Jones et al. 1994b, Chapter 2). We obtained rainfall data from a weather station at Busselton Regional Airport, which was located approximately 15 km from the study site (Australian Bureau of Meteorology 2014). We included the weight of collars in the modelling analysis to check the assumption that collars do not affect the survival of studied animals (Pollock et al. 1995).

Candidate models that represented our data the best were identified using corrected Akaike Information Criterion (AICc) values. We considered models with ∆AICc values of less than 2.0 to have strong support and those with ∆AICc values between 2.0 and 7.0 to have weak support (Burnham and Anderson 2002), compared with the best fitting model. In models with strong support, we identified the directionality of the effects from their parameter estimates and checked for their significant divergence from zero using 95 % confidence intervals.

103

As the collar weight was found to be the most influential factor for the survival rate (see Results), we re-ran the same survivorship modelling analysis with data only from possums wearing collars lighter than 19.9 g, to minimise the effect of collar weight on the estimated survival rates. If a heavy collar on an individual was replaced with a light collar during the study period, only the data after switching to a light collar for that particular individual were included in the analysis. Survival data from 15 females and 19 males over 105 weeks satisfied this criterion. From this survivorship analysis with light collars, we then estimated the survival rate of adult females in seasons 1 and 2 in 2010, 2011 and 2012 (six estimates in total) by model averaging sex, year and season models. 95 % confidence intervals for estimated survival rates were calculated using the profile likelihood method within program MARK.

Life cycle matrices

We investigated the life cycle of females only because pooling both sexes in matrix- based PVA can lead to an underestimation of the extinction risk by ignoring the demographic stochasticity in sex ratio (Brook et al. 2000a). We focused on females because they are the reproductively limiting sex and we could estimate their fecundity parameters from our data while this could not be done for males. In favourable conditions, female possums in the Busselton region can reproduce twice a year at around the beginning of each life cycle season (Jones et al. 1994b, Yokochi, personal observation). None of the radio collared females reproduced twice in a life cycle season. After each of two breeding peaks, the young stays in its mother’s pouch for approximately three months and become an independent juvenile at the approximate age of 6 to 7 months (Jones et al. 1994b, Yokochi, personal observation). In captivity, females have been observed to start breeding as early as 305 days old (Ellis and Jones 1992). The average life span of P. occidentalis in the wild is thought to be 4 to 5 years (Wayne et al. 2005b), and the oldest known age of collared females at the end of our monitoring period was 5 years old, at which stage, the animal was still reproductively active.

Based on these life history characteristics, we constructed a Leslie life cycle graph and transition matrix with six-monthly time steps up to five years of age, for each of two life cycle seasons in each of three monitoring years (i.e. six seasonal matrices in total). A

Leslie matrix can be constructed from observed demographic data using 1) fx = mx. P0, where fx is a fecundity parameter, mx is the number of female young produced by a

104 female in one time step (i.e. six months), and P0 is the probability of a female pouch young surviving to become an independent juvenile (Ebert 1999). This fx = mx. P0, model had one juvenile stage and 9 adult stages (Figure 2). An estimated rate of change in the population size during one time step (λasympt “asymptotic lambda”) can be calculated from this matrix, then the same population information can be restated by constructing a second matrix using 2) fx = mx. λasympt (Ebert 1999).

Figure 2 A Leslie life cycle graph for female Pseudocheirus occidentalis in Busselton, Western

Australia where fx = mx. P0. The time interval for each step is six months. fx is a fecundity parameter, mx is the number of female young produced by a female in one time step, P0 is the survival rate of female pouch young, PJ is the survival rate of female juveniles, and PA is the survival rate of female adults. J refers to the juvenile stage (0.5 to 1 year old), A1 (1 to 1.5 years old adult),

A2 (1.5 to 2 years old), A3 (2 to 2.5 years old), A4 (2.5 to 3 years old), A5 (3 to 3.5 years old), A6 (3.5 to 4 years old), A7 (4 to 4.5 years old), A8 (4.5 to 5 years old) and A9 (5 to 5.5 years old).

The basic survivorship and fecundity parameters required in our life cycle transition matrices were mx, P0, survival rate of a juvenile female (PJ), and survival rate of an adult female (PA). For each of six life cycle seasons, we calculated mx by averaging the numbers of female pouch young present in the pouches of each adult female caught in the season. We estimated P0 by dividing the number of surviving offspring of radio collared females at the point of independence by the number of newborns observed with adult females in either capture or radio tracking data in the season. Because of the lack of data on the sex ratio of newborns, we assumed that P0 was the same for males and females, as observed in a closely related species, common ringtail possums (P. peregrinus, How et al. 1984). No information on PJ of P. occidentalis was available because juveniles were too small to be collared and once independent, they could no longer be observed with their collared mothers. PJ is thought to be lower than P0 or PA

105 due to juveniles’ naivety and need to traverse over greater distances in unknown territories for dispersal, which increases the risk of road mortality and predation (A.

Wayne, J. Clarke, and U. Wicke, Pers. Comm.). How et al. (1984) also found that PJ of

P. peregrinus was lower than their P0 or PA. Although P. peregrinus is a much more common species than P. occidentalis, these two species share similar biological characteristics, including being arboreal folivores, having two breeding peaks per year, and having an average life span of 4 to 5 years (Thomson and Owen 1964, How et al.

1984, Woinarski et al. 2014); therefore, we used the PJ value of 0.367 reported by How et al. (1984) as an estimate of PJ for all six seasons in our analyses. PA for each of six seasons was estimated from the survivorship analysis described earlier.

We used the Leslie matrices with fx = mx. P0 to obtain the estimate of λasympt and the elasticity for each element in the matrix. Elasticity values for eight PA were added to obtain a single elasticity value for PA, and elasticity values for nine fx were added to obtain an elasticity value for P0 (Ebert 1999). Using the λasympt value, we then constructed a fx = mx. λasympt matrix and calculated the stable age distribution (Cx) for the season. This process was repeated for each of six seasons to obtain a set of six fx = mx. λasympt matrices and six lists of Cx values. We used the “lambda”, “elasticity”, and “stable.stage” functions in the popbio package v.2.4 (Stubben et al. 2012) within R v3.0.1 (R Foundation for Statistical Computing 2013, available at http://www.R- project.org) for these calculations.

Demographic and environmental stochasticity projection

The fx = mx. λasympt transition matrix, starting female population size, a vector of starting numbers of females in 11 life stages, and a matrix of the variances of fx (i.e. var(fx)) were necessary to conduct a population projection to estimate demographic stochasticity for each of six seasons, using the “multiresultm” function in the popbio package for R. We estimated the starting female population size in our PVA based on the most recent adult population estimate in Locke Nature Reserve, which was 353 (de Tores and Elscot 2010). Given that 117 out of 214 possums (54.7 %) we captured were females, we calculated that 193 out of 353 adults would be females. We averaged the six sets of Cx values derived from fx = mx. λasympt transition matrices to obtain one set of Cx values. We then estimated the proportion of female adults in the studied population to be 64.0 % by summing the Cx values of nine adult stages. Based on this percentage, we calculated the total starting female population size (i.e. number of adults, juveniles, and

106 pouch young) to be (193 / 0.64) = 302. In each of six life cycle seasons, the starting number of females in each life stage was calculated by multiplying 302 by the corresponding Cx value for the stage. Because the λasympt is constant, the var(fx), or the 2 variance of mx. λasympt, was calculated as λasympt * var(mx), where var(mx) is the variance of mx in each season. Using these parameters, we ran 1,000 simulations of population projection over 20 years. We applied the “multiresultm” function on the starting numbers of the population, and continued to apply it on the newly created set of numbers until it reached 20 years (i.e. 40 six-monthly time steps). In each of 1,000 simulations, a six-monthly rate of change in population size after adding demographic stochasticity (λdemog) was calculated as -1 (Dorai-Raj 2015). We performed all these calculations and projections in R v3.0.1. The R script for these steps is presented in Appendix 1.

Impacts of fox predation and road mortality

To assess how fox predation was affecting the population, we estimated new sets of PA values through the same survivorship analyses, while assuming individuals that were killed by foxes had survived for the remainder of the monitoring period. We are aware that individuals may die of other causes even if they escaped fox predations. However, given that the majority of mortalities of P. occidentalis were presumed to be caused by fox predation in the studied population, the manipulation of the mortality values was kept simple for the simulation purposes. As pouch young are completely reliant on their mothers for their survival in the wild, the rate of increase in PA following the removal of fox predation was also applied to P0. PJ for this assessment remained the same as the original analyses because causes of mortality were unknown for juveniles. We then conducted the same set of projection processes with the new sets of PA and P0 values. We also conducted the same analyses for the scenario of removal of road mortality by assuming all adult possums that were killed by vehicles had survived for the remainder of the monitoring period. For each scenario, r and P(ext) were calculated and compared with the estimates from the actual field data using profile likelihood 95 % confidence intervals.

The processes and assumptions involved with our projections and simulations are summarised in Figure 3 and Table 1.

107

Table 1 A list of assumptions and their justifications made while performing population viability analyses (PVA) of a stronghold population of Pseudocheirus occidentalis in Busselton, Western

Australia. PA, PJ, and P0 are survival rates of adult, juvenile and pouch young, respectively.

Assumption Source, justification and comments 1) Possums died of fox predation if fox Scavenging is possible but unable to saliva was present. distinguish in the field. Foxes are known to actively prey on P. occidentalis. 2) The trend of the female population Females are the reproductively limiting represents the trend of the whole sex. Brook et al. (2000a) population. 3) PA is unaffected by collar weight PA for 19.9 g collar is within 95 % C.I. when collars are < 20 g. of P for 15.8 g collar. A 4) PJ of P. occidentalis is the same as The two species are closely related and that of P. peregrinus. share many biological characteristics. How et al. (1984) 5) P0 is the same for males and How et al. (1984) females. 6) Population is in stable age Although unlikely, this was assumed, as distribution. we could not assess this with our data. 7) The population size of adults is 353. de Tores & Elscot (2010) 8) Population goes extinct if the Conservative value set for PVA. Likely number of females falls below two. to underestimate the extinction risk if the real threshold is higher than two. 9) No catastrophes (severe weather This was assumed as data are events, fires or diseases) for the next unavailable. Likely to underestimate the 20 years. extinction risk. 10) No migration for the next 20 years. Unlikely but assumed as data unavailable. Likely to overestimate the extinction risk. 11) In a fox predation removal scenario, 70 % of mortality was caused by fox if a possum does not die of fox predation. Likely to overestimate the predation, it survives till the end of effect of fox predation removal. study period. 12) In a road mortality removal Likely to overestimate the effect of road scenario, if a possum does not die of mortality removal. Also assumed by road mortality, it survives till the Ramp & Ben-Ami (2006) end of study period. 13) Removal of fox predation or road Juveniles are likely to be affected by fox mortality does not affect PJ. predation and road mortality, but mortality data on juveniles are unavailable.

108

Figure 3 A flow chart showing the process of the population viability analyses. PY and BoM stand for pouch young and Australian Bureau of Meteorology (2014), respectively. PA, PJ, and P0 are survival rates of adult, juvenile and pouch young, respectively. fx is fecundity parameter, Cx is stable age distribution value, λasympt is asymptotic lambda, λproj is projected change in population size, and P(ext) is probability of extinction. Numbers in grey diamonds are assumptions made in the step. Numbers in yellow diamonds are assumptions relevant only in simulated scenarios. Refer to Table 1 for corresponding assumptions. Assumption 2 is not shown as it applies to the whole process. 109

Results

Causes of mortality and survivorship analysis

We confirmed mortalities of 23 radio collared possums in three years, and the number of mortalities ranged between 2 and 6 in each six monthly season (Table 2). Fox predation was the most common cause of mortality, contributing to up to 100 % of mortalities in each season. Road mortality contributed to two mortalities in total. Other causes of mortality included one cat predation, one predation in which the predator species could not be genetically determined, one general condition loss and two unknown cause. Possums wearing heavy radio collars (> 19.9 g) contributed to 48 % of mortalities even though only 36% of monitored possums were wearing heavy collars.

Table 2 The number and causes of confirmed mortalities of radio collared Pseudocheirus occidentalis and in Busselton, Western Australia. Season 1 is from May to October and season 2 is from November to April. “Unknown Predator” is the number of mortalities caused by predation events where the predator species could not be identified. Numbers in parentheses are numbers of possums wearing collars that were heavier than 19.9 g. “Radio collared” is the total number of possums monitored during the whole study period. Unknown Season Fox Cat Predator Road Other Total 2010-1 3 (3) 0 1 (1) 1 (1) 0 5 2010-2 3 (3) 0 0 1 (0) 0 4 2011-1 5 (2) 0 0 0 1 (0) 6 2011-2 1 (0) 0 0 0 1 (0) 2 2012-1 2 (0) 0 0 0 0 2 2012-2 2 (0) 1 (0) 0 0 1 (1) 4 Total 16 (8) 1 (0) 1 (1) 2 (1) 3 (1) 23 (11) Radio collared 53 (19)

110

When we performed a survivorship analysis, the weight of the collars was found to be the most influential factor affecting the survival of adult possums, with an increase in the collar weight decreasing their survival (Table 3). The null model had the second strongest support from our data, but its support was a quarter of that for the collar weight. All the other candidate models ranked lower than the null model and none of their parameters had a significant effect on the survival of possums. Given the strong effect of the collar weight on the survival of possums, we projected the estimated six- monthly survival rate of possums over the range of collar weights used in this study (15.8 – 22.6 g), based on the candidate model with collar weight as a factor. The estimated six monthly survival rate of possums wearing the lightest collar was 0.897 (95 % confidence intervals: 0.783, 0.954) and it fell by 15 % to 0.763 (0.665, 0.838) when the collar weight was increased to 19.9 g. The survival rate fell by a further 15.5 % to 0.623 (0.406, 0.800) with the heaviest collar.

Table 3 Results of a known fate model analysis on the survival of adult Pseudocheirus occidentalis near Busselton, Western Australia. The variable “LC (Life Cycle) season” refers to May to October (season 1) and November to April (season 2), “Rainfall” is the average rainfall, “Habitat” is whether the individual was located in Locke Nature Reserve or a campsite, and “Barrier type” refers to the closest barrier (road or artificial waterway). In the case of categorical variables, parameter estimates ( s.e.) are for the categories presented in parentheses (e.g. the parameter estimate of the sex model is for females). * indicates parameters with 95 % confidence intervals outside of zero. Variable AICc ∆AICc AICc weight Parameter estimates Collar weight 265.0 0.48 -0.244  0.11* Null 267.8 2.7 0.12 Body weight 268.5 3.5 0.08 0.002 ± 0.00 Sex 268.7 3.6 0.08 0.480 ± 0.46 (female) LC season 268.7 3.6 0.08 -0.478 ± 0.46 (season I) Rainfall 269.2 4.1 0.06 -0.079 ± 0.10 Habitat 269.7 4.6 0.05 -0.172 ± 0.52 (reserve) Barrier type 269.7 4.7 0.05 0.080 ± 0.46 (road) Year 275.2 10.2 0.00 -0.233 ± 1.16 (2010) 0.009 ± 1.17 (2011) -0.193 ± 1.19 (2012)

111

In the survivorship analyses, after the removal of data from individuals wearing heavy collars, the null model had the strongest support from data; however, differences in AICc values were small and all factors except for the year had strong support from our data (∆AICc < 2.0, Table 4). However, in all models, 95 % confidence intervals of the parameter estimates included zero, indicating that their effects were not significant.

Six-monthly PA values for females estimated by model averaging sex, life cycle season and year models were very similar over three years varying between 0.845 and 0.863

(Table 5). Six-monthly P0 fluctuated the most (0.333 – 1.000) out of three parameters estimated, and mx varied between 0.234 and 0.447 over three years.

Table 4 Results of a known fate model analysis on the survival of adult Pseudocheirus occidentalis with collars less than 20 g near Busselton, Western Australia. The variable “LC (Life Cycle) season” refers to May to October (season 1) and November to April (season 2), “Rainfall” is the average rainfall, “Habitat” is whether the individual was located in Locke Nature Reserve or a campsite, and “Barrier type” refers to the closest barrier (road or artificial waterway). In the case of categorical variables, parameter estimates ( s.e.) are for the categories presented in parentheses (e.g. the parameter estimate of the sex model is for females). 95 % confidence intervals of all parameters included zero. Variable AICc ∆AICc AICc weight Parameter estimates Null 133.5 0.23 Rainfall 133.7 0.2 0.21 -0.177 ± 0.13 Sex 134.6 1.1 0.13 0.599 ± 0.64 (female) Barrier type 134.7 1.2 0.13 -0.617 ± 0.72 (road) Body weight 135.3 1.8 0.10 -0.001 ± 0.00 LC season 135.4 1.8 0.09 -0.266 ± 0.63 (season I) Habitat 135.5 2.0 0.09 0.088 ± 0.64 (reserve) Year 139.2 5.7 0.01 0.442 ± 1.53 (2010) 0.179 ± 1.20 (2011) -0.111 ± 1.21 (2012)

112

Table 5 Six-monthly demographic parameters and results of a Leslie transition matrix analysis of female Pseudocheirus occidentalis in Busselton, Western Australia. Season 1 is from May to

October, and season 2 is from November to April. PA is the survival rate of adults per six months, PJ is the survival rate of juveniles per six months, P0 is the survival rate of pouch young per six months, and mx is the number of female young produced by a female in six months. λasympt (the estimated rate of change in population size in six months) and elasticity values were calculated from Leslie transition matrices using these demographic parameters. Elasticity values for P0 and PJ are presented together because they were identical. Numbers in parentheses are 95 % confidence intervals or variances in case of mx. Subscripted numbers are the sample sizes or the ranges of sample sizes in case of PA.

Six monthly demographic parameters Matrix analysis results

Season PA P0 mx λasympt Elasticity

PJ (P0) PA 2010-1 0.849 0.500 0.447 0.852 0.167 0.665 (0.656, 0.943) 2-4 (0.205, 0.795) 12 (0.240) 41 2010-2 0.863 0.333 0.263 0.746 0.144 0.712 (0.607, 0.945) 6-9 (0.000, 0.747) 6 (0.185) 30 2011-1 0.847 0.778 0.417 0.908 0.180 0.639 (0.669, 0.938) 8-17 (0.490, 1.000) 9 (0.193) 18 2011-2 0.862 1.000 0.234 0.868 0.168 0.664 (0.711, 0.940) 14-18 (1.000, 1.000) 4 (0.067) 8 2012-1 0.845 0.500 0.250 0.774 0.152 0.696 (0.662, 0.938) 10-13 (0.062, 0.938) 6 (0.214) 8 2012-2 0.859 0.750 0.375 0.894 0.174 0.652 (0.702, 0.941) 12-15 (0.260, 1.000) 4 (0.143) 8

PVAs on observed and simulated data

The λasympt values for all six life cycle seasons were smaller than one and indicated that the number of female P. occidentalis in this population would decrease by between 9 and 25 % in each season if the population was in a stable age distribution (Table 5). The elasticity values for the survival rate of adults (PA) were about four times greater than those of pouch young (P0) or juveniles (PJ ) in all life cycle seasons, suggesting that changes in PA alter the rate of population changes in female P. occidentalis more than those in P0 or PJ do. According to our PVA, the studied population has over 92 % probability of going extinct in the next 20 years (Table 6).

113

In a scenario where the road mortality effect on survival rates of adults and pouch young was removed from our dataset, λproj was smaller than one, indicating that the population would still be in decline; however, the predicted rate of decline was smaller than the decline predicted from our original data, and P(ext) dropped to 32 % (Table 6). In a scenario where fox predation effect was removed, the population was also projected to decline, but the rate of decline was even smaller than the “no road mortality” scenario and P(ext) in the next 20 years was almost zero (0.4 %, Table 6).

Table 6 Results from population viability analyses of a Pseudocheirus occidentalis population in Busselton, Western Australia based on an observed dataset and two management scenarios. Analyses were run using data from females only. “No road” and “No fox” are scenarios where road mortality and fox predation were removed from the estimation of adult and pouch young survival rates. Six monthly λproj is the projected change in the number of females over six months, and P(ext) is the projected probability of the number of females dropping to below two in 20 years. Numbers presented in brackets are 95 % lower and upper confidence intervals.

Scenario λproj P(ext) Observed 0.827 (0.788, 0.882) 0.921 (0.903, 0.937) No road 0.882 (0.837, 0.882) 0.318 (0.290, 0.347) No fox 0.891 (0.882, 0.907) 0.004 (0.001, 0.009)

Discussion

Current and future decline of P. occidentalis

Our population projections suggest that the studied population of P. occidentalis is declining and has an alarmingly high risk of extinction within the next 20 years. These projections were built upon many assumptions (Table 1 and Figure 3), so the rate of decline should not be accepted as a precise prediction. However, data from other studies seem to support this recent steep decline. In early 2009, the density of P. occidentalis within continuous vegetation at Locke Nature Reserve was estimated to be 5.84 ha-1 from four months of monitoring (de Tores and Elscot 2010). Four and a half years later in a recent study, Harring-Harris (2014) estimated the density of the possums within continuous vegetation at this reserve to be 1.03 ha-1 from one month of monitoring. This reduction in the density gives us an estimated population decline of 0.825 per six

114 months, which is close to our λproj estimate of 0.827 per six months. Although these density estimates were based on the numbers of possums detected by different operators, both studies were conducted intensively in short time periods and employed a distance sampling method, which is thought to be a more accurate method than other conventional methods to estimate a density of this elusive species (de Tores and Elscot 2010, Finlayson et al. 2010). Therefore, the two estimates are comparable and provide strong support for our projections of an alarming decline in this population.

Although the projected decline of the studied population from our PVA was similar to the recent decline observed by Harring-Harris (2014), the future population projection in this study may be optimistic compared with the reality. One of the assumptions made in this analysis was that environmental conditions will remain the same for the next 20 years. Although 2010 - 2013 had a lower rainfall and greater number of days with a maximum temperature of 35 °C or above compared to previous years (Australian Bureau of Meteorology 2014, summarised in Figure 4), the Busselton region is expected to experience even less rain and more frequent hot days in the future due to climate change (Indian Ocean Climate Initiative 2012). This drier and hotter climate is likely to negatively impact the survival of P. occidentalis as these possums struggle to cope with dry conditions with ambient temperatures above 35 °C (Yin 2006). Other assumptions included no catastrophes, such as severe weather events, fires or diseases. These events could happen in the next 20 years, especially in view of the predicted changes in climate, and would dramatically accelerate the possums’ decline. The accuracy of the probability of extinction estimated from PVA is debatable (Brook et al. 2000b, Fieberg and Ellner 2000, Coulson et al. 2001, Ball et al. 2003, McCarthy et al. 2003, Schödelbauerová et al. 2010), so it should not be taken as a precise prediction; however, it does provide indications of the general direction in which the population is headed. In the case of this stronghold population of P. occidentalis, the future outlook is poor even when we take these uncertainties into account.

115

45 1200 40 1000 35 30 800 25 600 20 15 400

10 (mm) rainfall Annual

Number of hot days hot of Number 200 5 0 0 1998 2000 2002 2004 2006 2008 2010 2012 Year Figure 4 Number of days with a maximum temperature of or above 35°C and annual rainfall between 1998 and 2013 in Busselton, Western Australia. Bars represent the number of hot days and the line graph represents the change in annual rainfall. The monitoring period for this study (2010 to 2013) is highlighted by dark shaded bars and markers on the line graph. Upper and lower dotted lines represent the average annual rainfall and average number of hot days, respectively. Climate data were obtained from a weather station at Busselton Regional Airport (Australian Bureau of Meteorology 2014).

Our results have serious implications for the entire species because the studied population is thought to be one of the largest and densest populations left, and it is located in the most pristine habitat for the species (Jones et al. 1994a, 2007). Therefore, it is possible that some of other remaining populations of this endangered species are experiencing an even greater decline than that reported here. Based on the spotlight count data of P. occidentalis between 1997 and 2004 in the inland jarrah-dominated Upper Warren region (Wayne et al. 2012), the finite rate of their population change can be calculated as 0.453 per six months. This value is about a half of the rate estimated in this study, suggesting that the Upper Warren population declined at twice the rate of the stronghold population we studied. As a result, no possums have been sighted in most of the inland survey areas after 2006, even though the Upper Warren population used to be a large, genetically diverse population (Wilson 2009). This rapid decline supports the notion that P. occidentalis in other parts of its remaining range may be experiencing a greater decline and calls for more urgent and effective conservation efforts to be implemented to ensure this species’ survival. Long-term monitoring of P. occidentalis populations in other parts of its remaining range, such as in the Albany region, is also urgently needed to assess their population trends.

116

Two management scenarios

Presumed predation by foxes was the most common cause of mortality in radio collared adult P. occidentalis, and the simulated removal of fox predation in PVA resulted in a slower rate of decline and a dramatically lower probability of extinction in the next 20 years, as we expected. This result suggests that the removal or reduction of foxes in the area is likely to be effective in slowing down the decline of P. occidentalis. Locke Nature Reserve is baited monthly with baits containing sodium monofluroacetate toxin to control foxes (de Tores et al. 2004, C. Forward, pers. comm.); however, fox predation is still having a significant negative impact on the study population. Due to public safety concerns, baiting cannot be conducted in campsites or along roads, which makes it extremely hard to eradicate foxes from the study area because foxes are currently free to roam between the baited nature reserve and unbaited campsites. The farmland south of the reserve could also act as source of foxes without effective fox control measures. Reducing and maintaining a low density of foxes in the whole region with the aim of eventual eradication is essential to ensure the survival of P. occidentalis as a species, and more effective baits and/or methods to control foxes on a broader scale need to be explored. We also recorded one instance of predation by a cat in the area, so cats need to be controlled concurrently with foxes because their number can increase as the number of foxes decreases (Glen and Dickman 2005, de Tores and Marlow 2012).

Two radio collared adult possums were killed by vehicles on the road, and the simulated removal of road mortality in PVA resulted in a slower rate of population decline and a lower probability of extinction, as expected. Although the removal of road mortality did not seem to ease the decline of possums as much as the removal of fox predation, mortalities on roads need to be minimised to maximise the chance of this species persisting. One of the most common methods to reduce road mortality is installing fences along the road; however, it is impractical for this species in our study area because of the arboreal nature of the possums and the need for public accesses to the campsites along Caves Road. Road mortalities need to be reduced by providing possums with safe ways to cross roads by means such as rope bridges. In a previous study, Caves Road was found to limit the movements of P. occidentalis (Yokochi et al. 2015: Chapter 2). Isolated populations have a higher risk of extinction because of its higher vulnerability to stochastic demographic changes and catastrophic events such as severe weather, fire and diseases (Foley 1997). If no action was taken to increase the connectivity, the isolation caused by the road may further threaten the survival of this 117 population. A high level of connectivity between subpopulations also encourages gene flow, dispersal and immigration, which can prevent further decline of the population. Therefore, installing wildlife crossing structures across the road is likely to benefit this declining population in multiple ways.

We conducted PVAs on both of these scenarios by changing only the adult and pouch young survival rates because data for juvenile mortality were not available. Removal of fox predation and road mortality is also likely to increase the survival rate of juveniles because at this stage individuals are thought to be most vulnerable to these threats. Therefore, the actual positive effects of removal of foxes and road mortality may be greater than those predicted in this study. On the contrary, some assumptions made during the modelling processes may have caused overestimation of the positive impacts that removal of these threats would have on the studied population. For example, we simulated the removal of threats by assuming that individuals that were killed by the particular threat would survive for the remainder of the monitoring period; however, it is possible that these individuals would die of other causes during the monitoring period. This is especially the case for the simulation of the removal of road mortality because two road-killed individuals that were assumed to have survived could have instead died of the most common cause of mortality, fox predation. We are not alone in making this assumption in a simulation of reduction in road mortality (Ramp and Ben-Ami 2006); however, to what degree this uncertainty affected the outcome of our simulations cannot be known from our data, and caution should be paid when interpreting our results.

Nonetheless, increased adult and pouch young survival rates due to the removal of common threats still resulted in projected rates of population change (λproj) of less than one, suggesting that this population is likely to keep declining even without fox predation or road mortality. One possible explanation for this lack of impact of the simulated management of threats could be that the observed adult survival rates were already high (0.85 – 0.86 per six months). Because of the high base survival rates, removal of fox predations and removal of road mortalities only resulted in 9 % and 4 % increase in survival rates, respectively. The high elasticity value of the adult survival rates meant that these small changes were enough to significantly lower the population extinction risk in the near future (i.e. 20 years); however, they were still not enough to change the population trend. Therefore, conservation efforts to increase other demographic rates, such as fecundity rate, need to be undertaken concurrently with the reduction in fox predation and road mortality. By comparing reproductive trend in 118 different habitats, Jones et al. (1994b) suggested that improved habitat quality can increase the fecundity rate of P. occidentalis. However, which particular habitat components and attributes encourage survival and reproduction of P. occidentalis is still largely unknown and this population is already thought to be living in the most pristine environment in its range; therefore, more research is needed on preferred macro- and microhabitats of this species.

Impacts of collar weight

We unexpectedly found that the weight of the collars strongly affected the survival of the possums. While adult possums wearing 15.8 g radio collars had a 90% chance of survival for six months, those with 20 g collars only had a 76 % chance, and those wearing 22.6 g collars had a less than 63 % chance of survival. Individuals wearing collars over 20 g accounted for half of mortalities caused by predation by foxes even though only 36 % of the monitored individuals wore heavy collars. This suggests that heavier collars may have somehow increased the chance of possums being predated. It is possible that the extra weight of collars slowed movements of possums and/or forced them to come down to the ground, thus making them more vulnerable to predation. Several researchers have examined the effects of radio collars on the survival of small mammalian species, but their results are contradictory (Withey et al. 2001). For example, wearing radio collars reduced the body condition of European badgers (Meles meles; Tuyttens et al. 2002) and biased the sex-ratio of water voles (Arvicola terrestris; Moorhouse and Macdonald 2005); however, collars did not affect the short term survival of house mice (Mus domesticus; Pouliquen et al. 1990), yellow-necked mice (Apodemus flavicollis; de Mendonça 1999) or root voles (Microtus oeconomus; Johannesen et al. 1997). Studies into the effects of radio collars on arboreal mammals seem to be lacking, and this study is the first instance where the negative effect of collar weight on the survival of an arboreal marsupial species has been tested and demonstrated. This result is extremely alarming because the weight of the heaviest collar used in this study was about half of the recommended maximum collar weight set in guidelines that are commonly used by wildlife researchers (Animal Ethicks Infolink n.d., Sikes and Gannon 2011). These guidelines recommend researchers to fit radio collars of up to 5 % (10 % in case of Sikes and Gannon 2011) of the animals’ body weight, which would be approximately 50 g (or 100 g) for an average 1000 g P. occidentalis.

119

We reduced the effect of collar weights on our projections by omitting data from possums wearing collars heavier than 20 g (i.e. more than 2 % of their body weight); therefore, the results from our PVA should not have been affected by the effects of the collars. However, this means that putting heavy collars on P. occidentalis is likely to accelerate the decline of this species; therefore, researchers must be cautious when monitoring this species in the future. The trap success rate of these possums is extremely low due to their highly arboreal nature and folivorous diet (Wayne et al. 2005a); therefore, we will still have to rely on radio transmitters to monitor them. The weight and size of transmitters might decrease as lighter and smaller batteries become available with advances in technology, which, in turn, will decrease the burden of collars on animals. However, for now we must ensure that we use the lightest and smallest collars available on this species in order to minimise the negative impact of the collars. The current recommended maximum weight of radio collars appears to be excessive for western ringtail possums, and possibly for other small to medium sized arboreal marsupials. This study was not designed to investigate the effects of the collar weight on animals, so more well-designed studies are needed on the impacts of collars on arboreal mammals; however, until then, we strongly suggest changing the maximum collar weight to 2 % of the body weight for arboreal marsupials especially if long-term monitoring is expected.

Limitations and future research directions

As discussed above, many assumptions were made in the process of performing our PVA (Table 1). One of the assumptions made was that if fox saliva was found on the collar or carcass of a dead possum, the death was due to fox predation. We are aware that the fox could have scavenged on the carcass of a possum that had died of a different cause; therefore, the observed number of “fox predation” event may be an overestimate. One could try to distinguish between predation and scavenging events by the presence of subcutaneous haemorrhage on a carcass because trauma caused while the prey is alive would cause subcutaneous haemorrhaging (Miller et al. 1985). However, this observation is not possible if the carcass is not intact at discovery, which was the case for many recorded mortalities in this study. Given this lack of data and the knowledge that foxes do actively prey on P. occidentalis (Kinnear et al. 2002, Clarke 2011, Yokochi, personal observation), we assumed that foxes actively predated on radio collared P. occidentalis, and readers should be aware that the estimated impacts of removal of fox predations may have been overestimated as a result. 120

Another assumption in our PVA was that the studied population would experience no migration for the next 20 years. This assumption is very unlikely to be true because the study area was connected to other known P. occidentalis habitats (Chapter 3). However, there were no data available on the rate of migration from/ to this population, so migration was not incorporated in our models. Given the rapid rate of decline estimated in our PVA (λproj of 0.827 per six months) and the starting female population size of 302, at least 52 female immigrants per six months would be required to stop the population decline. Given the highly arboreal and sedentary nature of P. occidentalis in this population (Yokochi et al. 2015: Chapter 2), the actual migration rate is unlikely to be high enough to change the population trend. Nonetheless, immigration is an important source of new individuals for a declining population (Foley 1997); therefore, this assumption is likely to have resulted in an overestimation of extinction risk. Research on the migration rates of P. occidentalis is needed to eliminate the uncertainty of this assumption from future PVA.

The juvenile survival rate (PJ ) used in our analyses was based on data from a similar, but different ringtail possum species in a different habitat 40 years ago; therefore, it was only a general estimation. Having accurate parameter estimates is important in PVA because they are the basis of the model construction and they can alter the resulting projections significantly (Taylor 1995). This PJ value was the best estimation available to us at the time of this study, especially given that PJ is thought to be lower than PA or

P0 in P. occidentalis (Pers. Comm. A. Wayne, J. Clarke, and U. Wicke). Elasticities of

PJ were also lower than those of adult survival rate (PA), suggesting that having an accurate PA is more important in accurately projecting this population than PJ.

Nevertheless, studies into the survival of juveniles are needed so that a more accurate PJ can be incorporated into future analyses. Our limited observational data may have caused particularly low values of mx (the number of female young produced by a female in one time step). Our data were the best estimate available at the time of this study because the only other available data on the reproductive rate of wild female P. occidentalis were those estimated by Jones et al. (1994a) based on one-year-long monitoring of only three females. Demographic parameter estimates based on larger sample sizes and longer monitoring periods would increase the accuracy of the projections. However, the species is extremely difficult to capture and therefore obtaining larger samples size would represent a serious challenge. Data on demographic

121 parameters of males would also give us a more thorough picture of the status of this population by enabling us to incorporate both sexes into the projections.

Conclusion

Using PVA, we predicted that a stronghold population of the endangered western ringtail possum would decline with a high probability of extinction within the next 20 years if no action is taken. Fox predation was the most common cause of mortality in this population, followed by road mortality, and removal of these known threats dramatically reduced the projected probability of extinction. However, this population was predicted to decline even with the removal of these threats, suggesting that conservation efforts to increase other demographic parameters, such as fecundity rate, need to be implemented concurrently with mitigation of the impacts of fox predation and road mortality. Our results highlight the poor outlook for this species and call for the urgent implementation of conservation strategies. The current recommended maximum weight of radio collars is also excessive for this species, and we recommend using collars that are as light as possible, up to 2 % of the animals’ body weight, for this species and possibly also for other arboreal specialists.

References

Akçakaya HR, Sjögren-Gulve P (2000) Population viability analyses in conservation planning: an overview. Ecological Bulletins: 9-21. doi:10.2307/20113245

Animal Ethics Infolink (n.d.) Radio Tracking in Wildlife Research. http://www.animalethics.org.au/policies-and-guidelines/wildlife-research/radio- tracking [accessed 17th October 2015].

Australian Bureau of Meteorology (2014) Climate Data Online. http://www.bom.gov.au/climate/data/ [accessed 15th October 2014].

Australian Bureau of Statistics (2009) 3218.0 - Regional Population Growth, Australia, 2007-08. http://www.abs.gov.au/ausstats/[email protected]/Products/3218.0~2007- 08~Main+Features~Main+Features [accessed 10th August 2014].

Ball S, Lindenmayer D, Possingham H (2003) The predictive accuracy of population viability analysis: a test using data from two small mammal species in a fragmented

122

landscape. Biodiversity & Conservation 12: 2393-2413. doi:10.1023/a:1025821506931

Berry O, Sarre SD (2007) Gel-free species identification using melt-curve analysis. Molecular Ecology Notes 7: 1-4. doi:10.1111/j.1471-8286.2006.01541.x

Brook BW, Burgman MA, Frankham R (2000a) Differences and congruencies between PVA packages: the importance of sex ratio for predictions of extinction risk. Ecology and Society 4 (1):6. available at http://www.ecologyandsociety.org/vol4/iss1/art6/

Brook BW, O'Grady JJ, Chapman AP, Burgman MA, Akçakaya HR, Frankham R (2000b) Predictive accuracy of population viability analysis in conservation biology. Nature 404: 385-387. available at http://www.nature.com/nature/journal/v404/n6776/suppinfo/404385a0_S1.html

Burnham KP, Anderson DR (2002) Model Selection and Multimodel Inference - A Practical Information-Theoretic Approach. Springer-Verlag New York, Inc., New York, USA.

Coulson T, Mace GM, Hudson E, Possingham H (2001) The use and abuse of population viability analysis. Trends in Ecology & Evolution 16: 219-221. doi:10.1016/S0169-5347(01)02137-1 de Mendonça PG (1999) Impact of radio-collars on Yellow-necked Mice, Apodemus flavicollis (Mammalia, Rodentia). Mammal Review 29: 127-134. doi:10.1046/j.1365-2907.1999.00042.x de Tores PJ, Elscot S (2010) Estimating the population size of a threatened arboreal marsupial: use of distance sampling to dispense with ad hoc survey techniques. Wildlife Research 37: 512-523. doi:10.1071/WR10090 de Tores P, Marlow N (2012) The relative merits of predator-exclusion fencing and repeated fox baiting for protection of native fauna: five case studies from Western

123

Australia. In: Somers MJ, Hayward M (Eds) Fencing for Conservation. Springer New York, 21-42. doi:10.1007/978-1-4614-0902-1_3 de Tores PJ, Hayward MW, Rosier SM (2004) The western ringtail possum, Pseudocheirus occidentalis, and the quokka, Setonix brachyurus, case studies: Western Shield review, February 2003. Conservation science Western Australia 5: 235-257

Department of Parks and Wildlife WA (2014a) Summary of additions, deletions and changes to the wildlife - Conservation (specially protected fauna) notice as of 2 December 2014. DPaW. Perth, Australia. http://www.dpaw.wa.gov.au/images/documents/plants-animals/threatened- species/Listings/Summary_of_Changes_to_Specially_Protected_Fauna_Notice.pdf [accessed 20th January 2015]

Department of Parks and Wildlife WA (2014b) Western ringtail possum (Pseudocheirus occidentalis) recovery plan. Wildlife management program No. 58/ DPaW. Perth, Australia. http://www.dpaw.wa.gov.au/images/documents/plants-animals/threatened- species/recovery_plans/WRP_RECOVERY_PLAN_Final_24022014.pdf [accessed 2nd December 2014]

Dorai-Raj S (2012) binom, R package version 1.1-1. http://cran.r- project.org/web/packages/binom/

Ebert TA (1999) Plant and animal populations: methods in demography. Academic Press, San Diego, U.S.A.

Ellis M, Jones B (1992) Observations of captive and wild western ringtail possums Pseudocheirus occidentalis. Western Australian Naturalist 19: 1-10

Fieberg J, Ellner SP (2000) When is it meaningful to estimate an extinction probability? Ecology 81: 2040-2047. doi:10.1890/0012-9658(2000)081[2040:wiimte]2.0.co;2

Finlayson GR, Diment AN, Mitrovski P, Thompson GG, Thompson SA (2010) Estimating western ringtail possum (Pseudocheirus occidentalis) density using distance sampling. Australian Mammalogy 32: 197-200

Foley P (1997) 10 - Extinction models for local populations. In: Gilpin IHE (Ed) Metapopulation Biology. Academic Press, San Diego, 215-246

124

Glen AS, Dickman CR (2005) Complex interactions among mammalian carnivores in Australia, and their implications for wildlife management. Biological Reviews 80: 387-401. doi:10.1017/s1464793105006718

Harring-Harris K (2014) Selection of day-time shelter and night-time feeding trees by western ringtail possums in Locke Nature Reserve and neighbouring caravan park. Honours thesis, The University of Western Australia. Perth, Australia.

How R, Barnett J, Bradley A, Humphreys W, Martin R (1984) The population biology of Pseudocheirus peregrinus in a Leptospermum laevigatum thicket. Possums and gliders: 261-268

Indian Ocean Climate Initiative (2012) Western Australia's weather and climate: A synthesis of Indian Ocean Climate Initiative stage 3 research. The Commonwealth Scientific and Industrial Research Organisation and Bureau of Meteorology Australia

Johannesen E, Andreassen HP, Steen H (1997) Effect of radiocollars on survival of root voles. Journal of Mammalogy 78: 638-642. doi:10.2307/1382914

Jones BA, Henry J, Francesconi M (2007) An important local population of the Western Ringtail Possum, Pseudocheirus occidentalis: a 2006 survey study of the population and habitat in the Busselton localities of Siesta Park and Kealy. Unpublished report for GeoCatch, Busselton, Australia.

Jones BA, How RA, Kitchener DJ (1994a) A field study of Pseudocheirus occidentalis (Marsupialia: Petauridae) I. Distribution and Habitat. Wildlife Research 21: 175-187

Jones BA, How RA, Kitchener DJ (1994b) A field study of Pseudocheirus occidentalis (Marsupialia: Petauridae) II. Population studies. Wildlife Research 21: 189-201

Kinnear JE, Sumner NR, Onus ML (2002) The red fox in Australia—an exotic predator turned biocontrol agent. Biological Conservation 108: 335-359. doi:10.1016/S0006- 3207(02)00116-7

McCarthy MA, Andelman SJ, Possingham HP (2003) Reliability of relative predictions in population viability analysis. Conservation Biology 17: 982-989. doi:10.1046/j.1523-1739.2003.01570.x

125

Miller FL, Gunn A, Broughton E (1985) Surplus killing as exemplified by wolf predation on newborn caribou. Canadian Journal of Zoology 63: 295-300. doi:10.1139/z85-045

Moorhouse TP, Macdonald DW (2005) Indirect negative impacts of radio-collaring: sex ratio variation in water voles. Journal of Applied Ecology 42: 91-98. doi:10.1111/j.1365-2664.2005.00998.x

Pouliquen O, Leishman M, Redhead TD (1990) Effects of radio collars on wild mice, Mus domesticus. Canadian Journal of Zoology 68: 1607-1609. doi:10.1139/z90-239

Pollock KH, Bunck CM (1995) A capture-recapture survival analysis model for. Journal of Applied Statistics 22: 661

Ramp D, Ben-Ami D (2006) The effect of road-based fatalities on the viability of a peri- urban swamp wallaby population. The Journal of Wildlife Management 70: 1615- 1624

Schödelbauerová I, Tremblay RL, Kindlmann P (2010) Prediction vs. reality: Can a PVA model predict population persistence 13 years later? Biodiversity & Conservation 19: 637-650. doi:10.1007/s10531-009-9724-1

Sikes RS, Gannon WL (2011) Guidelines of the American Society of Mammalogists for the use of wild mammals in research. Journal of Mammalogy 92: 235-253. doi:10.1644/10-mamm-f-355.1

Stubben C, Milligan B, Nantel P (2012) popio, R package version 2.4. http://cran.r- project.org/web/packages/popbio/

Taylor BL (1995) The reliability of using population viability analysis for risk classification of species. Conservation Biology 9: 551-558. doi:10.2307/2386609

Thomson JA, Owen WH (1964) A field study of the Australian ringtail possum Pseudocheirus peregrinus (Marsupialia: Phalangeridae). Ecological Monographs 34: 27-52. doi:10.2307/1948462

126

Tuyttens FAM, Macdonald DW, Roddam AW (2002) Effects of radio-collars on European badgers (Meles meles). Journal of Zoology 257: 37-42. doi:10.1017/s0952836902000638

Wayne AF, Cowling A, Ward CG, Rooney JF, Vellios CV, Lindenmayer DB, Donnelly CF (2005a) A comparison of survey methods for arboreal possums in jarrah forest, Western Australia. Wildlife Research 2005: 701-714

Wayne AF, Rooney JF, Ward CG, Vellios CV, Lindenmayer DB (2005b) The life history of Pseudocheirus occidentalis (Pseudocheiridae) in the jarrah forest of south- western Australia. Australian Journal of Zoology 53: 325-337

White GC, Burnham KP (1999) Program MARK: Survival estimation from populations of marked animals. Bird Study 46 Supplement: 120-138. available at http://www.phidot.org/software/mark/

Withey JC, Bloxton TD, Marzluff JM (2001) Chapter 3 - Effects of tagging and location error in wildlife radiotelemetry studies. In: Joshua J. Marzluff, Millspaugh JM (Eds) Radio Tracking and Animal Populations. Academic Press, San Diego, 43- 75. doi:10.1016/B978-012497781-5/50004-9

Woinarski JC, Burbidge AA, Harrison PL (2014) The action plan for Australian mammals 2012. CSIRO Publishing, Collingwood, Australia

Yin HK (2006) The metabolic and hygric physiology of western ringtail possum (Pseudocheirus occidentalis). Honours thesis, Perth: Curtin University of Technology

127

Yokochi K, Chambers BK, Bencini R (2015) An artificial waterway and road restrict movements and alter home ranges of endangered arboreal marsupial. Journal of Mammalogy. doi:10.1093/jmammal/gyv137

128

Appendix 1

Below I present the R script used in the population viability analyses of Pseudocheirus occidentalis. This script was developed by Robert Black and Kaori Yokochi, based on Elbert (1999). An example set of transition life cycle matrices required is given first, followed by the script.

Table A1 “WRP-F-5yr 2010-1nf.txt”. A fx = mx. P0 transition life cycle matrix of female Pseudocheirus occidentalis in the first life cycle season of 2010.

J A1 A2 A3 A4 A5 A6 A7 A8 A9 J 0 0.088 0.088 0.088 0.088 0.088 0.088 0.088 0.088 0.088 A1 0.367 0 0 0 0 0 0 0 0 0 A2 0 0.863 0 0 0 0 0 0 0 0 A3 0 0 0.863 0 0 0 0 0 0 0 A4 0 0 0 0.863 0 0 0 0 0 0 A5 0 0 0 0 0.863 0 0 0 0 0 A6 0 0 0 0 0 0.863 0 0 0 0 A7 0 0 0 0 0 0 0.863 0 0 0 A8 0 0 0 0 0 0 0 0.863 0 0 A9 0 0 0 0 0 0 0 0 0.863 0

Table A2 “WRP-F-5yr 2010-1nf ml.txt “. A fx = mx. λ transition life cycle matrix of female Pseudocheirus occidentalis in the first life cycle season of 2010.

PY J A1 A2 A3 A4 A5 A6 A7 A8 A9 PY 0 0 0.196 0.196 0.196 0.196 0.196 0.196 0.196 0.196 0.196 J 0.333 0 0 0 0 0 0 0 0 0 0 A1 0 0.367 0 0 0 0 0 0 0 0 0 A2 0 0 0.863 0 0 0 0 0 0 0 0 A3 0 0 0 0.863 0 0 0 0 0 0 0 A4 0 0 0 0 0.863 0 0 0 0 0 0 A5 0 0 0 0 0 0.863 0 0 0 0 0 A6 0 0 0 0 0 0 0.863 0 0 0 0 A7 0 0 0 0 0 0 0 0.863 0 0 0 A8 0 0 0 0 0 0 0 0 0.863 0 0 A9 0 0 0 0 0 0 0 0 0 0.863 0

129

Table A3 “WRP-F-5yr 2010-1h ml v.txt”. A var(fx) matrix of female Pseudocheirus occidentalis in the first life cycle season of 2010.

PY J A1 A2 A3 A4 A5 A6 A7 A8 A9 PY 0 0 0.067 0.067 0.067 0.067 0.067 0.067 0.067 0.067 0.067 J 0 0 0 0 0 0 0 0 0 0 0 A1 0 0 0 0 0 0 0 0 0 0 0 A2 0 0 0 0 0 0 0 0 0 0 0 A3 0 0 0 0 0 0 0 0 0 0 0 A4 0 0 0 0 0 0 0 0 0 0 0 A5 0 0 0 0 0 0 0 0 0 0 0 A6 0 0 0 0 0 0 0 0 0 0 0 A7 0 0 0 0 0 0 0 0 0 0 0 A8 0 0 0 0 0 0 0 0 0 0 0 A9 0 0 0 0 0 0 0 0 0 0 0

R script for PVA library(popbio) library(binom)

###2010 ##Season1 #fx=mxp0 wrp2010.1nf <- read.table("WRP-F-5yr 2010-1nf.txt", header=T, row.names =1) wrp2010.1nf.mat <- as.matrix(wrp2010.1nf) adl <- c("J", "A1", "A2", "A3", "A4", "A5", "A6", "A7", "A8", "A9") lambda(wrp2010.1nf.mat) #Use lambda value to construct fx=mx.lambda matrix #fx=mx.lambda matrix saved as "WRP-F-5yr 2010-1nf ml.txt"

#fx=mx.lambda wrp2010.1nfml <- read.table("WRP-F-5yr 2010-1nf ml.txt", header=T, row.names =1) wrp2010.1nfml.mat <- as.matrix(wrp2010.1nfml) adl <- c("PY", "J", "A1", "A2", "A3", "A4", "A5", "A6", "A7", "A8", "A9") stable.stage(wrp2010.1nfml.mat) lambda(wrp2010.1nfml.mat) #Check if lambda value is the same as fx=mxp0. If not, check for errors in matrices.

##Season2 #fx=mxp0 wrp2010.2nf <- read.table("WRP-F-5yr 2010-2nf.txt", header=T, row.names =1) wrp2010.2nf.mat <- as.matrix(wrp2010.2nf) adl <- c("J1", "A1", "A2", "A3", "A4", "A5", "A6", "A7", "A8", "A9") lambda(wrp2010.2nf.mat) #Use lambda value to construct fx=mx.lambda matrix

#fx=mx.lambda wrp2010.2nfml <- read.table("WRP-F-5yr 2010-2nf ml.txt", header=T, row.names =1) wrp2010.2nfml.mat <- as.matrix(wrp2010.2nfml) adl <- c("PY", "J", "A1", "A2", "A3", "A4", "A5", "A6", "A7", "A8", "A9") stable.stage(wrp2010.2nfml.mat) lambda(wrp2010.2nfml.mat) #Check if lambda value is the same as fx=mxp0. If not, check for errors in matrices.

###2011 #Season1 130

#fx=mxp0 wrp2011.1nf <- read.table("WRP-F-5yr 2011-1nf.txt", header=T, row.names =1) wrp2011.1nf.mat <- as.matrix(wrp2011.1nf) adl <- c("J", "A1", "A2", "A3", "A4", "A5", "A6", "A7", "A8", "A9") lambda(wrp2011.1nf.mat) #Use lambda value to construct fx=mx.lambda matrix

#fx=mx.lambda wrp2011.1nfml <- read.table("WRP-F-5yr 2011-1nf ml.txt", header=T, row.names =1) wrp2011.1nfml.mat <- as.matrix(wrp2011.1nfml) adl <- c("PY", "J", "A1", "A2", "A3", "A4", "A5", "A6", "A7", "A8", "A9") stable.stage(wrp2011.1nfml.mat) lambda(wrp2011.1nfml.mat) #Check if lambda value is the same as fx=mxp0. If not, check for errors in matrices.

##Season2 #fx=mxp0 wrp2011.2nf <- read.table("WRP-F-5yr 2011-2nf.txt", header=T, row.names =1) wrp2011.2nf.mat <- as.matrix(wrp2011.2nf) adl <- c("J1", "A1", "A2", "A3", "A4", "A5", "A6", "A7", "A8", "A9") lambda(wrp2011.2nf.mat) #Use lambda value to construct fx=mx.lambda matrix

#fx=mx.lambda wrp2011.2nfml <- read.table("WRP-F-5yr 2011-2nf ml.txt", header=T, row.names =1) wrp2011.2nfml.mat <- as.matrix(wrp2011.2nfml) adl <- c("PY", "J", "A1", "A2", "A3", "A4", "A5", "A6", "A7", "A8", "A9") stable.stage(wrp2011.2nfml.mat) lambda(wrp2011.2nfml.mat) #Check if lambda value is the same as fx=mxp0. If not, check for errors in matrices.

###2012 #Season1 #fx=mxp0 wrp2012.1nf <- read.table("WRP-F-5yr 2012-1nf.txt", header=T, row.names =1) wrp2012.1nf.mat <- as.matrix(wrp2012.1nf) adl <- c("J", "A1", "A2", "A3", "A4", "A5", "A6", "A7", "A8", "A9") lambda(wrp2012.1nf.mat) #Use lambda value to construct fx=mx.lambda matrix

#fx=mx.lambda wrp2012.1nfml <- read.table("WRP-F-5yr 2012-1nf ml.txt", header=T, row.names =1) wrp2012.1nfml.mat <- as.matrix(wrp2012.1nfml) adl <- c("PY", "J", "A1", "A2", "A3", "A4", "A5", "A6", "A7", "A8", "A9") stable.stage(wrp2012.1nfml.mat) lambda(wrp2012.1nfml.mat) #Check if lambda value is the same as fx=mxp0. If not, check for errors in matrices.

##Season2 #fx=mxp0 wrp2012.2nf <- read.table("WRP-F-5yr 2012-2nf.txt", header=T, row.names =1) wrp2012.2nf.mat <- as.matrix(wrp2012.2nf) adl <- c("J1", "A1", "A2", "A3", "A4", "A5", "A6", "A7", "A8", "A9") lambda(wrp2012.2nf.mat) #Use lambda value to construct fx=mx.lambda matrix

#fx=mx.lambda wrp2012.2nfml <- read.table("WRP-F-5yr 2012-2nf ml.txt", header=T, row.names =1) wrp2012.2nfml.mat <- as.matrix(wrp2012.2nfml) adl <- c("PY", "J", "A1", "A2", "A3", "A4", "A5", "A6", "A7", "A8", "A9") stable.stage(wrp2012.2nfml.mat) lambda(wrp2012.2nfml.mat) #Check if lambda value is the same as fx=mxp0. If not, check for errors in matrices. 131

################################################################ ##############Demographic stochasticity#########################

#Load fx=mx.lambda matrices t11 <- read.table("WRP-F-5yr 2010-1nf ml.txt", header = TRUE, row.names = 1) t12 <- read.table("WRP-F-5yr 2010-2nf ml.txt", header = TRUE, row.names = 1) t21 <- read.table("WRP-F-5yr 2011-1nf ml.txt", header = TRUE, row.names = 1) t22 <- read.table("WRP-F-5yr 2011-2nf ml.txt", header = TRUE, row.names = 1) t31 <- read.table("WRP-F-5yr 2012-1nf ml.txt", header = TRUE, row.names = 1) t32 <- read.table("WRP-F-5yr 2012-2nf ml.txt", header = TRUE, row.names = 1) t11 <- as.matrix(t11) t12 <- as.matrix(t12) t21 <- as.matrix(t21) t22 <- as.matrix(t22) t31 <- as.matrix(t31) t32 <- as.matrix(t32)

#Load var(fx) matrices v11 <- read.table("WRP-F-5yr 2010-1h ml v.txt", header = TRUE, row.names = 1) v12 <- read.table("WRP-F-5yr 2010-2h ml v.txt", header = TRUE, row.names = 1) v21 <- read.table("WRP-F-5yr 2011-1h ml v.txt", header = TRUE, row.names = 1) v22 <- read.table("WRP-F-5yr 2011-2h ml v.txt", header = TRUE, row.names = 1) v31 <- read.table("WRP-F-5yr 2012-1h ml v.txt", header = TRUE, row.names = 1) v32 <- read.table("WRP-F-5yr 2012-2h ml v.txt", header = TRUE, row.names = 1) v11 <- as.matrix(v11) v12 <- as.matrix(v12) v21 <- as.matrix(v21) v22 <- as.matrix(v22) v31 <- as.matrix(v31) v32 <- as.matrix(v32)

##2010-1 xx <- splitA(t11, r=1, c= 3:11) t11T <- xx$T t11F <- xx$F reps <- 1000 # number of trajectories tmax <- 40 # length of the trajectories 40 = 20 years totalpop <- matrix(0,tmax,reps) # initializes totalpop matrix to store trajectories nzero <- c(79,47,19,19,19,19,20,20,20,20,20) # starting population size, example only, calculated by [Starting N*Cx] for each stage. #Cx was obtained from [stable.stage] in the previous section. for (j in 1:reps) { n <- nzero for (i in 1:tmax) { n <- multiresultm(n,t11T, t11F, varF= v11) totalpop[i,j] <- sum(n) } }