CAUGHT ON CAMERA

Citizen science in the State Forest

JANUARY 2019

A REPORT ON THE FIRST FIVE YEARS OF MONITORING THE EFFECTS OF FIRE ON WILDLIFE Prepared by Christine Connelly, Richard Loyn, Caitlin Griffith, Dr Sera Blair Victorian National Parks Association The Victorian National Parks Association is a community-led nature conservation organisation. Our vision is to ensure is a place with a diverse and healthy natural environment that is protected, respected and enjoyed by all. We work with all levels of government, the scientific community and the general community to achieve long term, best practice environmental outcomes and help shape the agenda for creating and managing national parks, conservation reserves and other important natural areas across land and sea in Victoria. We are also Victoria’s largest bushwalking club and provide a range of education, citizen science and activity programs that encourage Victorians to get active for nature. NatureWatch The Victorian National Parks Association’s NatureWatch program is a citizen science program which gets community involved in collecting scientific data on Victorian native plants and . The program builds links between community members, scientists and land managers and develops scientifically based, practical projects which contribute to a better understanding of species and ecosystems, and the management of natural areas. Project Partners Wombat Forestcare Wombat Forestcare is a community group dedicated to protecting and enhancing the natural ecosystems of the Wombat State Forest and surrounding areas.

Eco Insights Eco Insights is a consultancy established by Richard Loyn that conducts strategic projects in Australia and Papua New Guinea. Richard is an ecologist with special interests in forests, fire, wetlands, threatened species, bird migration and conservation.

PAGE 2 | Caught on Camera | Citizen science in the Wombat State Forest Acknowledgements NatureWatch and the Caught on Camera project would not exist without the tremendous amount of work of many volunteers who have been involved in helping establish the project, setting up cameras, maintaining equipment, team leading, identifying wildlife in images and data management. Over the first five years of the project, each year around 45 volunteers contributed over 1,500 hours in the field, over 1,000 hours in the office identifying images and managing equipment and countless hours on other behind the scenes tasks. Wombat Forestcare input and support, Gayle Osborne, Eddie Schambre and several others. DELWP input and support, Matt Chick, Robin Holmes, Kerryn McTaggart and Sarah Bellhouse. ARI input into project development, Matt Bruce, Peter Menkhorst and Richard Loyn. EcoInsights project scientific support, Richard Loyn. Victorian National Parks Association volunteer team leaders, office volunteers and field volunteers. Project supported by: Since 2012, the project has received funding from the Foundation for National Parks and Wildlife, the ANZ Staff Foundation, Lord Mayor’s Charitable Foundation, Field Naturalists Club of Victoria Environment Fund, Perpetual Rowe Family Foundation. Caught on Camera is supported by the Victorian Government. This project is conducted with full ethics approvals and under DELWP Research Permit #10007466.

CONTENTS Summary 4 A project for the wider community 4 What have we found? 4 Introduction 6 Caught on Camera project 6 Project aims 6 This report 6 Background 7 The importance of understanding the fire response 7 Why use motion-sensing cameras? 7 Monitoring in Wombat State Forest 7 A 10-year endeavour 7 Methods 8 Study area 8 Sites 8 Survey methods 10 Training 10 Setting up monitoring sites 11 Species identification 12 Habitat assessments 12 Above: Christine Connelly. Photo by Deb Connelly. Analysis 13 Cover images: Eastern Grey and Results 14 Brush-tailed Phascogale (inset). Photos by VNPA. What did we observe? 14 Back cover images: Common Wombat and Associations between species 17 Crimson Rosella (inset). Photos by VNPA. Associations with habitat features 18 Relationships with fire history 19 Trends over time 20 Discussion 23 A community achievement 23 Understanding our results 23 Limitations and recommendations 24 What’s next 25 References 26

Caught on Camera | Citizen science in the Wombat State Forest | PAGE 3 Summary Between 2012 and 2016 The participation of the community more than 200 volunteers A project for the in this project demonstrates a great contributed over 2,500 wider community passion for science and conservation hours to monitor wildlife in Wombat State Forest. in Wombat State Forest Caught on Camera has been a as part of the Caught on major achievement for the local We’re incredibly grateful to everyone Camera project. Thanks to and wider community. that has contributed to the project the inspiring volunteer effort, to date and we look forward to Through the project, we’ve the community amassed five working with new and old friends built and strengthened positive years of data on 13 native to complete our ambitious and and long-lasting links between mammals and 15 native valuable 10-year study. the community, scientists and birds, including threatened the government. species. This is part of a 10 What have we found? year project and this report The community came together The community effort on this project marks our findings at the half to develop and deliver this has amassed five years of records that way mark. project over the past five years, document 13 native mammal species with volunteers from Wombat Back in 2012, Wombat and 15 native bird species, including Forestcare and VNPA being Forestcare and the VNPA involved in: were concerned that we did Mammals: not know enough about the Establishing the project science and methods Echidna effects of fire on wildlife. Reviewing project progress Brush-tailed Phascogale So, with the support of the Setting up and packing down cameras 100 Agile Antechinus Department of Environment, times at 44 different monitoring sites Dusky Antechinus Land Water and Planning’s Habitat assessments Common Brushtail Possum Arthur Rylah institute for Sending in recorded data Mountain Brushtail Possum Environmental Research Maintaining equipment Common Ringtail Possum (ARI), we set about devising Leading teams in the field a method to explore the Identifying wildlife in many thousands Common Wombat question; ‘what is the impact of images Black Wallaby Overview of final reports of fire on mammals?’ Eastern Grey through the VNPA’s Bush Rat NatureWatch program. Swamp Rat

Birds:

Southern Boobook Laughing Kookaburra Crimson Rosella Superb Fairy-wren White-browed Scrubwren Spotted Quail-thrush Grey Shrike-thrush Grey Currawong Pied Currawong Australian Magpie White-winged Chough Scarlet Robin Flame Robin Eastern Yellow Robin Bassian Thrush

PAGE 4 | Caught on Camera | Citizen science in the Wombat State Forest We have also recorded nine Other recent studies in similar We don’t have a lot of data on introduced mammal species ecosystems have shown that the less common species but to and one introduced bird species: longer-term effects of fire on date our results appear to show flora and fauna are generally that common ground-dwelling Common Blackbird more subtle than those in the mammal and bird species in Black Rat first three years. Our project the two forest types remained House Mouse appears to support this, at common over the study period. European Rabbit least for the common species. Brown Hare So far, there is no evidence Feral Goat of a dramatic response for Sambar Deer any species, and several Red Fox species have shown only Cat subtle responses. Domestic Dog (with its human!)

Brush-tailed Phascogales: an exciting find of a threatened species! Capturing images of a threatened Brush-tailed Phascogale (Phascogale tapoatafa) is an incredibly exciting outcome for this project. They are a small with a distinctive bottle-brush tail that live in dry leafy forests and nest in tree hollows and feed on insects and nectars. Being photographed at three sites across the Wombat State Forest demonstrates this is an important habitat area for this species which is listed as threatened under the Flora and Fauna Guarantee Act (1988). A few months before these images were taken, Wombat Forestcare picked up the species on some of their cameras – these were the first recordings of the species in this part of the Wombat State Forest since the 1970’s!

Brush-tailed Phascogale.

Caught on Camera | Citizen science in the Wombat State Forest | PAGE 5 INTRODUCTION

Caught on Camera Project aims project The Caught on Camera project This original aim remains, objectives in Wombat State although we have refined our Caught on Camera is an Forest are: project slightly since we started. ongoing project that involves We have extended the study to working with local community • To provide crucial data on look at fire response in a broader groups, land managers and the long-term impacts of range of species, including some scientists to establish long- prescribed burning on wildlife ground-foraging birds, and to term wildlife monitoring, using to land managers, community increase the number of sites in motion-sensing cameras. Field groups, scientists and each vegetation type and time deployment of the cameras and government. since fire category. management of equipment is • To create working partnerships undertaken by local community with government, researchers volunteers, with support of and community, to carry out This report volunteers from elsewhere in long-term monitoring. This report presents the Victoria, recruited through the • To demonstrate and promote results of the first five years’ VNPA. to government the need of monitoring, and provides In Wombat State Forest and for ongoing, strategic and recommendations for the Bunyip State Park, Caught on comprehensive monitoring in ongoing monitoring. Camera involves looking at response to fire. the long-term impact of fire We specifically designed the on wildlife. In the Hindmarsh/ project to explore the question: Wimmera region and at Bank Australia’s conservation ‘What mammal species are reserves, we explored the value located at ‘Recently Burnt’, of revegetation efforts for ‘Medium time since burnt’ and wildlife. ‘Long Unburnt’ sites in Foothills Forest and Forby Forest in In other locations we are Wombat State Forest?’ carrying out exploratory surveys, especially where contemporary survey data are lacking. Wombat

PAGE 6 | Caught on Camera | Citizen science in the Wombat State Forest BACKGROUND

The importance of Why use motion- Monitoring in understanding the sensing cameras? Wombat State fire response Monitoring of fauna using traditional Forest survey techniques (e.g. trapping, “Much monitoring of fauna is of such spotlighting) can be highly labour Since 2012, annual monitoring took a small scale and short duration that intensive. Motion-sensing cameras place over March to July at up to the statistical likelihood of detecting provide an alternative method, and the 20 sites each year. The monitoring a positive or negative effect of the opportunity to gather data on some season commences with a management regime is minute. Such animal groups (e.g. small mammals, community training session in March, shortcomings will only be overcome some arboreal mammals and ground- which includes deploying the cameras through broad-scale and/or long-term foraging birds) with much less effort. at the first four sites. studies of fauna.” (Clarke 2008) Cameras can be set-up at multiple Throughout the monitoring season, In Australia, we have a limited locations and left to automatically eight cameras are rotated around the understanding of the impacts of detect and record species throughout sites every three weeks, to ensure that fire on our biodiversity particularly the day and night. The method is also every site is monitored once. Each the impacts on fauna (Clarke 2008, much less stressful for the fauna than deployment involves two teams of MacHunter et al. 2009). We need trapping and can provide data on the around 4-5 volunteers. repeatable, ongoing monitoring or presence of some species that are there will continue to be gaps in our unlikely to be caught in traps. understanding of the impacts of A 10-year planned burning in the landscape. It’s important to note that motion- endeavour sensing cameras do not provide For land managers to improve accurate data on population sizes We established a one year trial, to test planned burning techniques, (except in rare cases when individual the study design and the community’s particularly to meet the needs of animals can be recognised) but capacity to carry out the study, our native fauna, it is essential to may have potential to provide with a view to carrying out annual understand the impact of fire on all data on relative abundance across monitoring over the long-term life forms, not just the impacts on the different habitats. (>10 years). plants. This is particularly relevant when working in a fragmented Motion-sensing cameras also offer At the end of the year, we reported landscape, such as Victoria. a terrific opportunity for community on our results and reviewed our engagement — the photographs of success and decided that the project “The ability of fire planners to animals ‘Caught on Camera’ in their was successful and should continue. meaningfully implement the dual natural environment can be shared We established a strategic plan with aspirations of protecting life and throughout the community as well planning and review actions to be property and achieving ecological as to a wider audience online. This undertaken at four yearly intervals. goals is dependent on the availability can serve as an educational tool and of science and evidence that informs inform the community about the In 2015, we conducted a detailed operational processes, and monitoring richness of their local environment, review of the project and determined that influences future management.” fostering a more meaningful that we are tracking well; the (MacHunter et al. 2009) relationship between the community data were of good quality and the and the place in which they live. project should continue with minor Through long term community- adjustments. driven projects that are developed with scientific rigour and linked to Specifically, we determined that we management techniques, such as should increase our sample size by planned burning, it is possible for the adding more sites across the site community to contribute to building categories, to improve our likelihood our knowledge base on the impact of of detecting responses. fire on fauna.

Caught on Camera | Citizen science in the Wombat State Forest | PAGE 7 Methods Study area Sites The study was conducted in From 2012 to 2015, 28 sites Parameters for selecting Wombat State Forest in central- were established, a subset monitoring sites: west Victoria, approximately of 30 sites established by • Include sites with planned 80 km north-west of central Wombat Forestcare and the burns in the subsequent 1-2 . The forest covers Department of Sustainability years (to collect pre-burn data and Environment as Community about 70,000 ha and extends and then monitor change). along both sides of the Great Research sites (Macak et al. • Sites in Forby Forest and Dividing Range, from Creswick to 2010). In 2016, a further 16 Foothills Forest EVDs. Mount Macedon. sites were added to increase the statistical power of the study, • Consider opportunities to Wombat forest has a long history providing greater capacity to address spatial distribution of of timber harvesting, and was detect trends in results. burns in the landscape, and the extensively logged during the patterns of burn intensity within gold rush era of the mid-1800s. The sites were selected to each burn. Patchiness can be Virtually all old-growth trees have represent sites that have been manipulated at both scales been removed from the forest. burnt at different time intervals, through the planning process Large-scale harvesting ceased in which were classified into three and decisions about when to 2002 (Macak et al. 2010), but the ‘time since fire’ categories; burn (weather on the day). Fine- forest continues to be managed scale patchiness can also be • recently burnt (RB: 0 - 10 years under the Forests Act, thus mining manipulated by decisions about since the last fire), and further logging activities ignition patterns. • medium burnt (IB: 11 - 41 cannot be ruled out. years since the last fire) and • All sites are easily accessible (by 2WD vehicle and short The forest comprises a • long unburnt (LU: 42+ years walk). mixed eucalypt foothill forest since the last fire). • Sites are at least 1km apart. with dominant Messmate Sites were located within (Eucalyptus obliqua) overstorey, • Design includes recently burnt two vegetation communities sites (0-3 years since burn) if in combination with Peppermints (ecological vegetation division (E. dives and E. radiata) and possible. categories for fire planning; Candlebark Gum (E. rubida) • If there is an interest or need EVDs); Foothills Forest and identified in the future, sites in (Leversha 1996). Sixty per cent of Forby Forest. the vegetation types that occur in new EVCs can be included at a the Wombat forest are classified The project planning workshop later stage. as threatened. provided a 10-year project plan So far, 44 sites have been to select 20 sites for monitoring selected and surveyed across The Wombat forest provides each year. This could include habitat for over 180 native the study area (Figure 1), with 28 some repeated sites and some monitored in the first four years fauna species, new sites and result in up to 100 including 18 rare or threatened and a further 16 surveyed in the or more monitoring sites over fifth year covered by this report. species (e.g. , the next 10 years. , Square-tailed Kite, White-throated Needletail and Lace Monitor). The forest is also inhabited by several key fire response species such as Mountain Brushtail Possum, Agile Antechinus and Black Wallaby (Macak et al. 2010; MacHunter et al. 2009).

PAGE 8 | Caught on Camera | Citizen science in the Wombat State Forest Figure 1. Map of the study sites within the Wombat State Forest and their time since fire category (green = long unburnt, blue= medium burnt, red= recently burnt).

Caught on Camera | Citizen science in the Wombat State Forest | PAGE 9 Survey methods Training This is an ongoing community project and on-ground activities are carried out by community volunteers. Each year, coordination and training is led by the VNPA. The monitoring season starts in March and begins with a community training activity, to launch the project and train participants in how to set-up cameras. Between 2012 and 2017, each annual training activity was attended by 25-50 eager participants , who were provided with project background, information about VNPA and Wombat Forestcare, visited several sites in Wombat forest and were trained in how to set-up bait stations and cameras.

PHOTOS: IAN KENINS

PAGE 10 | Caught on Camera | Citizen science in the Wombat State Forest Setting up monitoring sites

Following the training activity, site. Two motion-sensing ground, facing south to community volunteers from cameras (heat-in-motion type, avoid sun glare at sunrise VNPA and Wombat Forestcare Reconyx Hyperfire HC500) and sunset. volunteered every three weeks were installed at each site, to pack down, move and set-up approximately 100 m apart. The bait station comprised six cameras and bait stations at the stainless steel tea strainers, 20 sites monitored each year. Each camera was attached which were secured inside a to the nearest suitable tree at small rectangular cage and A GPS and location description the designated location and attached to a plastic garden were used to navigate to each positioned 50 cm above the stake, as shown in Figure 2.

Figure 2. Camera and bait station field set-up

Caught on Camera | Citizen science in the Wombat State Forest | PAGE 11 PHOTO: IAN KENINS

A bait station was placed The cameras were left to operate Habitat assessments 2-3 m in front of each camera for a minimum of 21 days. with the base of the cage Upon collection, cameras were At the first camera set-up at each approximately 30 cm from the checked to see if they were still site, habitat was assessed within ground. The tea strainers were operating. a one hectare area centred on the filled with a bait mixture of cameras. Several habitat variables rolled oats, peanut butter and Species identification were recorded, to describe upper-, golden syrup, which were used mid- and under-storey vegetation The species of wildlife in each to attract wildlife to the centre structure and to not any special photo taken by the cameras of the cameras’ field of view. features of the site. Key habitat were identified and counted by features such as leaf litter or fallen The cameras were aligned so volunteers under the supervision logs were scored on a scale from that the bait station appeared of the NatureWatch Coordinator. 0 (absent) to 5 (very prevalent). in the horizontal centre of The images of wildlife which Practice sessions helped ensure the frame and the bottom were difficult to identify were that this was estimated consistently of the bait cage was in the then sent to ARI or Eco Insights by all assessors. vertical centre of the frame. for further identification. Where The vegetation between each possible, animals have been Four photographs of the habitat camera and its bait station identified to species level. were taken from the locations where and one metre behind was In cases where some doubt cameras were installed, using hand- cleared (within the field of persisted they were assigned held digital cameras. The first is a view of the camera) to ensure to a more generic category, for photograph of the canopy, and the that this did not obscure any example “Unidentified Brushtail rest are horizontal photographs photographs of animals. Possum”. of the vegetation immediately surrounding the site.

PAGE 12 | Caught on Camera | Citizen science in the Wombat State Forest We had planned to examine sites Analysis in a third EVD (Grassy/Heathy Dry We calculated the number of Forest) but found too few sites for days on which each species this to be useful. Hence, we focused mainly on Foothills Forest and Forby was identified (using means for the two cameras) for each site Forest. Two sites in Grassy/Heathy and year. This measurement is Dry Forest were analysed along with sites in Forby Forest, as some called a ‘site-day’. We used this as a measure of the prevalence intermingling was found between of that species on the site at those EVDs. that time, or the intensity with Table 1 shows the spread of which the species was using sites across these ‘time since fire’ the site. We then assessed categories and the vegetation types. any relationships between Because there is less Forby Forest in this prevalence measure for the Wombat forest, we located two each species and the following sites within a similar EVD, ‘Grassy/ variables: Heathy Dry Forest’. • The prevalence of other Table 1. Number of sites in species. each fire category and • The habitat variables we ecological vegetation recorded. division. Time since fire • The vegetation type (Ecological categories include: recently Vegetation Division; EVD). burnt (3-14 years before • Fire history (number of the survey); medium-burnt recorded fires since 1970, and (15-50 years before) or ‘time since fire’ as a numerical long-unburnt (no recorded value of ‘years’). fires for 50 years or more).

We categorised the sites as Ecological Total Recently Medium Long recently burnt (3-14 years before Vegetation Sites Burnt Burnt Unburnt the survey); medium-burnt (15- Division

50 years before) or long-unburnt Foothills Forest 29 7 12 10 (no recorded fires for 50 years Forby or or more). This differs slightly Grassy/Healthy 15 4 5 6 from the original design (see Dry Forest pages 10-11), but it provided TOTAL 44 11 17 16 similar numbers of sites in each category. It happens that we did not survey any sites burnt Statistical analyses were run to less than three years before the identify significant correlations survey, and we will attempt to and other more complex include such very recently burnt relationships with EVD and time sites in the future because some since fire. Further statistical species are known to favour models will be developed in the those early stages after fire. future for selected species.

Caught on Camera | Citizen science in the Wombat State Forest | PAGE 13 Results

What did we observe? Altogether we recorded 23 most common mammal species The same seven species also topped mammal species and 16 in descending order were the list in terms of numbers of days bird species with the motion- Common Wombat (36 sites), detected, accounting for 74.2% of sensing cameras. Black Agile Antechinus (31), Red Fox mammal site-days. The two small Wallabies were detected on (28), Mountain Brushtail Possum species in the group (Agile Antechinus all 44 sites and were recorded (26), Eastern Grey Kangaroo and Bush Rat) were recorded on more on more days than any other (22), and Bush Rat (18). All were days at each site than the larger species, species. In terms of numbers native except for the Red Fox. reflecting their small home ranges and of sites recorded, the next Details in Table 2. consequently higher density.

Black Wallaby Common Wombat

Red Fox Bush Rat

Mountain Brushtail Possum Eastern Yellow Robin

PAGE 14 | Caught on Camera | Citizen science in the Wombat State Forest No other mammal species was recorded on more than 10 sites (Table 2). Black Rats appeared to be common on one of the three sites where they were recorded (WCR8, with Forby Forest last burnt in 2004) whereas other species were detected infrequently. The species recorded were mainly ground-foragers, although Agile Antechinus also forage extensively in trees and five species feed mainly in trees (Brush-tailed Phascogale, Koala and the three possum species). No gliders or were recorded with the camera traps.

Agile Antechinus Koala

Of the 22 mammal species, 13 were native and nine were introduced to Australia. Introduced mammal species contributed 7.1% of the mammal site-days, showing that introduced mammals form only a modest component of the mammal fauna detected. The most commonly detected bird species was the Superb Fairy-wren, and it was found on 20 sites, followed by the Grey Shrike-thrush at 18 sites (Table 2).

Feral cat Superb Fairy-wren

Caught on Camera | Citizen science in the Wombat State Forest | PAGE 15 Table 2: Species prevalence represented by the number of sites they were recorded visiting, the percentage of total sites they visited and the mean number of days (site days) they were recorded at each site.

Mean number Number of Percentage of of site days sites recorded sites visited (n=44) (n=44)

Native mammals species Black Wallaby 3.94 44 100.0 Common Wombat 1.49 36 81.8 Agile Antechinus 1.76 31 70.5 Any Brushtail 0.96 29 65.9 Mountain Brushtail Possum 0.77 26 59.1 Eastern Grey Kangaroo 0.45 22 50.0 Bush Rat 0.79 18 40.9 Common Ringtail Possum 0.09 10 22.7 Common Brushtail Possum 0.16 9* 20.5 Antechinus sp. 0.05 9 20.5 Echidna 0.06 9 20.5 Koala 0.08 7 15.9 Unknown Rattus sp. 0.07 6 13.6 Brushtail sp. 0.01 3 6.8 Brush-tailed Phascogale 0.04 3 6.8 Dusky Antechinus 0.01 2 4.5 Swamp Rat 0.01 1 2.3 Native bird species Superb Fairy-wren 0.56 20 45.5 Grey Shrike-thrush 0.49 18 40.9 White-browed Scrubwren 0.11 12 27.3 Grey Currawong 0.17 12 27.3 Australian Magpie 0.19 9 20.5 Laughing Kookaburra 0.16 7 15.9 White-winged Chough 0.13 7 15.9 Spotted Quail-thrush 0.03 4 9.1 Scarlet Robin 0.02 3 6.8 Currawong sp. 0.01 2 4.5 Southern Boobook 0.01 1 2.3 Crimson Rosella 0.01 1 2.3 Pied Currawong 0.00 1 2.3 Flame Robin 0.02 1 2.3 Eastern Yellow Robin 0.12 12 27.3 Bassian Thrush 0.06 8 18.2 Introduced species Red Fox 0.55 28 63.6 Cat 0.03 5 11.4 Black Rat 0.21 3 6.8 European Rabbit 0.05 3 6.8 Common Blackbird 0.02 3 6.8 Sambar or Deer sp. 0.01 2 4.5 Domestic Dog 0.01 2 4.5 House Mouse 0.01 2 4.5 Brown Hare 0.03 2 4.5 Feral Goat 0.04 1 2.3 * Identification of some Common Brushtail Possum records requires further confirmation. Local observers consider it to be unlikely to be as prevalent as these data suggest.

PAGE 16 | Caught on Camera | Citizen science in the Wombat State Forest All of the birds observed forage All bird species recorded, except extensively from the ground or Common Blackbird, were native low vegetation, and most were species. Introduced species classed as members of two contributed 0.7% of the bird site- guilds that take invertebrates days, showing that introduced from open ground or damp birds form a very small shady ground respectively. component of the bird fauna detected. Another common species (Grey Shrike-thrush) is classed as a Associations between general insectivore, which takes invertebrates and other food animal species from a wide range of situations. We found positive correlations between species that favour The only seed-eating bird similar habitats, including Black recorded (Crimson Rosella) feeds Wallaby and Common Wombat; mainly in the tree canopy but also Agile Antechinus and Echidna; from the ground or low plants. and Grey Shrike-thrush and Two carnivorous species Eastern Yellow Robin. These (Laughing Kookaburra and species all occupy a wide range Southern Boobook) feed mainly of habitats but the first pair by pouncing on ground-dwelling favours reasonably dense shrub prey though they may take prey cover and the other two pairs from a range of situations. become most common where shrub cover is interspersed with Laughing Kookaburra areas of open ground. Dusky Antechinus were identified at two sites where Bassian Thrush was detected (WCR13 & 28), and both species have a general preference for forest gullies, damp or wet forest. Birds that favour drier situations (e.g. White-winged Chough, Scarlet Robin and Spotted Quail- thrush) also showed positive correlations with each other, showing that they tended to favour similar sets of sites. Many bird species that feed from Few negative correlations tall shrubs, tree trunks or the were found, suggesting that eucalypt canopy are known to be competition between species common in the forest, but were was not a major force in shaping not detected with camera traps. the bird community in this forest.

Caught on Camera | Citizen science in the Wombat State Forest | PAGE 17 Associations with habitat features In terms of the designated Table 3: Mean number of site days of each species EVD, two native species by Ecological Vegetation Division. (Echidna and Spotted Quail- thrush) appeared to favour Site EVD Foothills Forby Forest or Grassy/Heathy Dry Forest or Forest Grassy/Heathy Forby Forest over Foothills Forest. Two introduced Number of sites 29 15 species (European Rabbit Native mammal species and Feral Cat) were also Echidna 0.04 0.10 found more often in Brush-tailed Phascogale 0.05 0.00 those EVDs. Agile Antechinus 1.97 1.33 Dusky Antechinus 0.01 0.00 Three of the uncommon Common Brushtail Possum 0.19 0.11 native mammal species Mountain Brushtail Possum 0.90 0.53 recorded were only found in Any Brushtail 1.11 0.65 Foothills Forest: these were Common Ringtail Possum 0.11 0.04 Brush-tailed Phascogale, Koala 0.08 0.09 Swamp Rat and Dusky Common Wombat 1.70 1.07 Antechinus. Fewer birds Black Wallaby 3.91 4.01 were recorded in the two Eastern Grey Kangaroo 0.48 0.41 sites containing Grassy/ Bush Rat 1.03 0.32 Heathy Dry Forest than in Swamp Rat 0.01 0.00 sites containing Forby Forest Native bird species or Foothills Forest. Southern Boobook 0.02 0.00 In general, the designated Crimson Rosella 0.02 0.00 EVD explained only a Laughing Kookaburra 0.16 0.15 small part of the variation Superb fairy wren 0.66 0.37 in mammal and bird White-browed Scrubwren 0.11 0.10 communities between sites. Spotted Quail-thrush 0.02 0.06 Mean numbers of site-days Grey Shrike-thrush 0.50 0.46 for each species in each EVD Australian Magpie 0.16 0.26 are shown in Table 3. Pied Currawong 0.00 0.00 Grey Currawong 0.17 0.17 Currawong sp. 0.00 0.01 White-winged Chough 0.17 0.04 Scarlet Robin 0.03 0.00 Flame Robin 0.00 0.05 Eastern Yellow Robin 0.15 0.05 Bassian Thrush 0.08 0.03 Introduced species Black Rat 0.08 0.47 House Mouse 0.02 0.00 European Rabbit 0.03 0.08 Brown Hare 0.02 0.07 Feral Goat 0.00 0.12 Sambar or Deer sp. 0.01 0.01 Red Fox 0.63 0.41 Cat 0.01 0.06 Domestic Dog 0.00 0.02 Common Blackbird 0.00 0.04

PAGE 18 | Caught on Camera | Citizen science in the Wombat State Forest Correlations were evident with stringybarks (mainly Messmate) Four native species showed some of the measured habitat was negatively correlated with weak positive relationships with characteristics. For example, Mountain Brushtail Possum but time since fire, suggesting a the abundance of tree-ferns was positively with Agile Antechinus. long process of recovery after positively associated with Dusky an initial negative impact of fire. Antechinus and Bassian Thrush. Numbers of Agile Antechinus These were Echidna, Bush Rat, Sedges and Blackwood wattles correlated positively with the Grey Shrike-thrush and Superb were positively associated number of trees with dead tops, Fairy-wren1. with Common Wombat and and negatively with the number insectivorous birds that feed of trees lacking that feature. Two uncommon species were from damp ground (including Tree size and number of hollows found mainly on sites that had White-browed Scrubwren and correlated positively with total not been burned for a long time: Eastern Yellow Robin). birds and several bird species. these were Dusky Antechinus and (more surprisingly) the The proportion of smooth- Relationships with introduced European Rabbit. barked eucalypts (gums, mainly fire history Introduced House Mouse was Candlebark) was positively only found in two sites (WCR4 correlated with Koala (which No species showed strong and 19), both of which were feeds preferentially from these relationships with fire history, long-unburnt. species) and Mountain Brushtail showing that most species can Possum (which favours the tolerate a range of fire regimes. The three records of Brush- damp habitats occupied by However, weak relationships tailed Phascogale were in gums). The proportion of were found for some species Foothills Forest in each of the (Table 4). three time since fire categories (recently burnt, medium burnt, long unburnt).

Short-beaked Echidna

1 Statistical significance for all tests: p < 0.1

Caught on Camera | Citizen science in the Wombat State Forest | PAGE 19 Table 4: Mean number of site days for each species Trends over time by fire history category. We surveyed 11 of the 44 sites every year from 2012 to 2015, Site Long Medium Recently which enables us to examine unburnt burnt burnt trends over time at these sites. Data from those sites Native mammals species are graphed in Figures 3-6 for Echidna 0.11 0.05 0.00 four common native mammal Brush-tailed Phascogale 0.02 0.06 0.02 species: along with most Agile Antechinus 2.00 2.13 0.82 species, these were expected to Dusky Antechinus 0.02 0.00 0.00 demonstrate a response to fire. Common Brushtail Possum 0.09 0.22 0.17 Mountain Brushtail Possum 0.96 0.69 0.61 Two of the 11 sites were classed Any Brushtail 1.06 0.95 0.81 as recently burnt: site WCR30 Common Ringtail Possum 0.14 0.08 0.03 was burnt in 2009 (three years Koala 0.12 0.04 0.09 before our study began), and Common Wombat 1.67 1.46 1.26 site WCR18 in 2005 (7 years). Black Wallaby 3.73 4.02 4.13 Three of the sites were classed Eastern Grey Kangaroo 0.45 0.35 0.63 as medium time since fire, Bush Rat 1.37 0.61 0.22 having been burnt 10-50 years Swamp Rat 0.02 0.00 0.00 previously (WCR5, 14 & 23), and Native bird species six were classed as long-unburnt Southern Boobook 0.00 0.00 0.05 (WCR3, 12, 19, 26, 28 & 29). It Crimson Rosella 0.00 0.00 0.05 was expected that the recently Laughing Kookaburra 0.02 0.23 0.25 burnt sites (especially WCR30) Superb fairy wren 1.00 0.36 0.23 might show more evidence of White-browed Scrubwren 0.13 0.07 0.13 successional change than the Spotted Quail-thrush 0.03 0.06 0.00 sites burnt longer ago. However, Grey Shrike-thrush 0.78 0.36 0.25 little support for that idea was Australian Magpie 0.21 0.16 0.22 found for these four species Pied Currawong 0.01 0.00 0.00 (Figures 3-6). Grey Currawong 0.13 0.21 0.16 Nevertheless, some differences Currawong sp 0.01 0.01 0.00 in response were evident2: White-winged Chough 0.09 0.22 0.05 Scarlet Robin 0.04 0.02 0.00 • Black Wallabies appeared to Flame Robin 0.05 0.00 0.00 increase over time on the most Eastern Yellow Robin 0.17 0.11 0.06 recently burnt site (WCR30) Bassian Thrush 0.08 0.07 0.02 but remained more stable on Introduced species the set of other sites (Figure 3). Black Rat 0.00 0.01 0.84 • Common were House Mouse 0.04 0.00 0.00 initially more common on European Rabbit 0.13 0.00 0.00 WCR30 than on other sets of Brown Hare 0.00 0.06 0.05 sites, but decreased to average Feral Goat 0.00 0.00 0.16 levels in the fourth year of the Sambar or Deer sp. 0.01 0.00 0.02 survey (Figure 4). Red Fox 0.48 0.60 0.60 • Bush Rats were not recorded Cat 0.05 0.02 0.01 on WCR30 except in the third Domestic Dog 0.01 0.00 0.01 year of the survey (2014), and Common Blackbird 0.02 0.00 0.05 numbers fluctuated greatly in the other sets of sites (Figure 5).

2 Further analysis is needed to assess the statistical significance of these results.

PAGE 20 | Caught on Camera | Citizen science in the Wombat State Forest Figure 3: Black Wallaby relationship with fire history.

9

8

7

6 Long unburnt

5 Medium burnt

WCR18 (2005) 4 WCR30 (2009)

3 Average number of site days per year number of site days Average

2

1

0 2012 2013 2014 2015

Figure 4. Common Wombat relationship with fire history.

6

5

4 Long unburnt

Medium burnt 3 WCR18 (2005)

WCR30 (2009)

2 Average number of site days per year number of site days Average

1

0 2012 2013 2014 2015

Caught on Camera | Citizen science in the Wombat State Forest | PAGE 21 Figure 5: Bush Rat relationship with fire history.

5

4.5

4

3.5

Long unburnt 3 Medium burnt 2.5 WCR18 (2005)

2 WCR30 (2009)

Average number of site days per year number of site days Average 1.5

1

0.5

0 2012 2013 2014 2015

Figure 6: Agile Antechinus relationship with fire history. • Agile Antechinus declined 7 dramatically after the first year of the survey on

66 both recently burnt sites, and on the set of 5 three medium- burnt sites, but Long unburnt not on the set of 4 six long-unburnt Medium burnt sites (Figure 6). A similar decline WCR18 (2005) occurred on the 3 WCR30 (2009) long-unburnt sites the following year, with a partial Average number of site days per year number of site days Average 2 recovery in 2015.

1

0 2012 2013 2014 2015

PAGE 22 | Caught on Camera | Citizen science in the Wombat State Forest Discussion

Some of our participants, “Looking at the photos there is A community volunteers and community something incredibly intimate achievement group leaders reflected on their and unguarded about them. It's learnings through an anonymous as if wildlife social-realism meets This important study on the survey: the monochromatic aesthetic of impacts of fire on wildlife in the night vision surveillance and we Wombat State Forest is carried “Participation has broadened my are becoming voyeurs of another out almost entirely by volunteers. knowledge and appreciation of intelligence at work -- which we More than 45 volunteers each local plants and animals and the would not have encountered year contributed over 2,500 habitats in which they live.” otherwise.” (Brüggemeier and hours of work to amass five “Community members have a Miranda 2012) years of highly valuable data desire for ecological knowledge on mammals in Wombat State and their participation allows for Forest. Understanding our the sharing of this knowledge.” results The project has opened a new We’ve also had some exciting window on the ground-dwelling Wombat State Forest has been and unexpected outcomes. mammal fauna in the Wombat subject to a long history of Caught on Camera in Wombat State Forest, and helps to gold-mining, logging, wildfire State Forest was featured in improve our understanding and planned burns. Most of French documentary, ‘Nature’s of how different species may the 44 sites surveyed had been Keepers’. Representatives from change over time, and respond subject to some or all of these VNPA and Wombat Forestcare to habitat and the way the disturbances. were interviewed and the field habitat is managed. deployment of the project was Fire causes substantial changes Without this study we would filmed to provide an international to fauna habitat, which are have no contemporary data audience with an understanding most marked in the first three about the ground-dwelling of the project and our aims. years. Influxes of native species mammal fauna in these forests associated with open habitats ‘Nature in the Dark’ was a in the decade following one are often seen after wildfire, and creative arts project that arose of the longest droughts in these may include introduced from the images we collected. species such as House Mouse recent history (1997-2009) and Artists were provided with our subsequent changes in policy (Friend 1993) and native species wildlife images and re-worked such as Superb Fairy-wren, relating to management of fire and interpreted the original and logging. Scarlet Robin and White-browed footage in their own way, with Woodswallow (Loyn 1997; Loyn several artists’ interpretations The project has seen passionate and McNabb 2015). community members, scientists curated into a single piece that and managers working together was projected at various venues. Smaller influxes may also occur to increase our understanding According to the curators; after planned burns, as shown in of mammals in Wombat various studies including some State Forest. The project in the Wombat State Forest has expanded the skills and (Humphries 1994; Loyn et al. knowledge of everyone involved. 2003; Loyn and McNabb 2015). The current study focused on longer-term effects, and we did not examine sites burnt less than three years before the surveys.

Caught on Camera | Citizen science in the Wombat State Forest | PAGE 23 Some of the species that As with any project, our prosper in the first three years Limitations and study does not provide a after fire (e.g. House Mouse, recommendations complete picture, and there Superb Fairy-wren and Scarlet are some limitations with the Robin) turned out to be more Camera trapping has become field methods and with the prevalent in long-unburnt forest a popular and valuable survey exploratory analysis presented than in younger age-classes. method, and refinements will here. But, despite the limitations, Other studies have shown that continue to be made to the this project has progressed some of the mammals and birds technology and the different our knowledge of fauna in that favour open stands in early ways of analysing data Wombat State Forest. More stages after fire or logging may (e.g. Swann et al. 2004; complex statistical analyses avoid the later stages where Nelson and Scroggie 2009; will be needed to unlock the shrub cover may be dense, but Meek et al. 2014). potential of this study to inform become more prevalent as those us about effects of habitat and At present, the number of stands open up again after many fire management. The analyses photographs does not provide years (Loyn 1997, 2004, 2012). reported here are exploratory, a robust measure of species and more sophisticated abundance, as individual Recent studies in Victorian approaches are possible (Robley animals may spend long and foothill forests have shown that et al. 2010; Gillespie et al. 2016) variable periods at the bait longer-term effects of fire on which we look forward to station on a single visit. Some flora and fauna are generally exploring in the future. subtler than those in the first ecologists have argued that three years (Muir et al. 2015; quantitative conclusions cannot With the present data, it is Leonard et al. 2016; Kelly et be made with camera-trap data clear that some species were al. 2017). The current project unless individual animals are detected much more often on supports this conclusion, identified, but we believe that some sites than others, in terms showing the response of the approach is too limited, and of numbers of photographs common ground-dwelling quantitative analyses of the taken and numbers of days on mammal and bird fauna in these sort we have undertaken can which they were photographed. forests. provide important information The number of days on which about the relative abundance the species is detected is highly The project has also shown of mammal species in different likely to reflect the relative that these forests continue to types of habitat. Analysis abundance of that species on support some less common based on presence-absence the site (for species with small species, including the Brush- would clearly be inadequate for home ranges), or the intensity tailed Phascogale, which was common species such as Black of use that site (for species with formerly only known from the Wallaby, which was recorded on large home ranges). drier northern parts of Wombat every site but exhibited marked State Forest. It is much harder variation in prevalence between to determine how they may be sites (different numbers of affected by current or future site-days). management of fire or anything else because of the low numbers we detected. However, it is reassuring that they can still be found in Wombat State Forest and continue to make use of widely distributed and less- threatened habitats including Foothills Forest.

PAGE 24 | Caught on Camera | Citizen science in the Wombat State Forest What’s next At a workshop to develop our The following reporting will be long-term project plan, attended undertaken: by the project partners, we 1. Regular reporting back to decided the best way to progress community via social media was to increase the number of and emails. sites by monitoring new sites 2. Reporting all feral and each year. The benefits of this threatened species records approach were identified as: to DELWP within a week of • Increase the statistical power downloading data. of the project and capacity to 3. Reports every five years answer the research question. prepared for a mixed • Easily make it possible to audience (scientists, land include several of the following managers, community and options: volunteers) and published by 1. Adding monitoring of birds VNPA. at existing sites. 4. Article for peer-reviewed 2. Adding monitoring in other science journal by around vegetation types 2022 (including up to 10 3. Setting up monitoring sites years of data). to monitor burn patterns at We’re committed to continue specific sites. this project for at least ten 4. Setting up monitoring at years. We’re looking forward sites before and after fire. to a continued and lasting 5. Gather more detailed relationship between the habitat data. community, scientists and 6. Include more ‘recently government representatives burnt’ sites (0-4 years involved in Caught on Camera in since fire). Wombat State Forest. • Make it possible to analyse the data against different variables. Black Wallaby

Caught on Camera | Citizen science in the Wombat State Forest | PAGE 25 References

Brüggemeier, J. and Miranda, Leonard, S., Bruce, M., Christie, Loyn, R.H. and McNabb, E.G. M. 2012. Curatorial Statement: F., Di Stefano, J., Haslem, A., 2015. Bird population responses Nature in the Dark 1. http:// Holland, G., Kelly, L., Loyn, R., to wildfire and planned burns in unlikely.net.au/nitd, accessed MacHunter, J., Rumpff, L., the foothill forests of Victoria, August 2018. Bennett, A., Clarke, M. and York, Australia. Journal of Ornithology. A. 2016. Foothills fire and biota. 156: 263-273. Clarke, M. F. 2008. Catering Fire and Adaptive Management for the needs of fauna in fire Report no. 96. Department of Macak, P.V., Chick, R.R. and Loyn, management: science or Environment, Land, Water & R.H., 2010. Arboreal mammal just wishful thinking? Wildlife Planning, Melbourne, Victoria. habitat associations in Wombat Research 35: 385-394. State Forest. Unpublished Loyn, R.H. 1997. Effects of client report to the Wombat Friend, G.R. 1993. Impact of fire an extensive wildfire on birds Biodiversity Working Group/ on small in mallee in far eastern Victoria. Pacific Wombat Forestcare Inc., Arthur woodlands and heathlands of Conservation Biology 3: 221-234. Rylah Institute for Environmental temperate Australia: a review. Research, Department of Biological Conservation 65: 99- Loyn, R.H. 2004. Research for Sustainability and Environment, 114. ecologically sustainable forest Heidelberg, Victoria. management in Victorian eucalypt Gillespie, G., Gentles, T., Hill, B., forests pp 783-806 in D. Lunney MacHunter, J., Menkhorst, P. Choy, J.L, Mahney, T., Stevens, (Ed) Conservation of Australia’s and Loyn, R.H. 2009. Towards a A. and Stokeld, D. 2016. A guide forest fauna (second edition). process for integrating vertebrate for the use of remote cameras for Royal Zoological Society of New fauna into fire management wildlife surveys and surveillance in South Wales, Mosman, NSW. planning. Technical Report Series Northern Australia. Department no. 192, Arthur Rylah Institute of Land Resource Management, Loyn, R.H. 2012. Vertebrate for Environmental Research, Palmerston, NT. fauna and fire regimes: a Department of Sustainability conceptual model to aid fire Humphries, R, K. 1994. The and Environment, Melbourne, research and management. Victoria. effects of single autumn and Proceedings of the Royal Society spring prescribed fires on small of Victoria 124: 20-29. Meek, P.D., Ballard, G., mammal and ecology in Claridge, A., Kays, R., Moseby, Wombat State Forest. Masters Loyn, R.H., Cunningham, R.B. K., O’Brien, T., O’Connell, A., Thesis. University of , and Donnelly, C. 2003. Effects Sanderson, J., Swann, D. E., Victoria, 163pp. of fuel reduction burning at Tobler, M. and Townsend, S. two frequencies and seasons 2014. Recommended guiding Kelly, L.T., Haslem, A., Holland, on bird abundance in eucalypt G.J., Leonard, S., MacHunter, principles for reporting on forests of central Victoria: a camera trapping research. J., Bassett, M., Bennett, A.F., replicated experiment. Arthur Bruce, M.J., Clarke, M., Chia, Biodiversity Conservation 23: Rylah Institute, Heidelberg, 2321-2343. E., Christie, F., Di Stefano, J., for Department of Natural Loyn, R., McCarthy, M., Pung, Resources and Environment, A., Robinson, N., Sitters, H., Melbourne, Victoria. Swan, M. and York, A. 2017. Fire regimes and environmental gradients shape vertebrate and plant distributions in temperate eucalypt forests. Ecosphere 8: e01781.

PAGE 26 | Caught on Camera | Citizen science in the Wombat State Forest References (cont.)

Muir, A., MacHunter, J., Bruce, Robley, A., Gormley, A., M., Moloney, P., Kyle, G., Woodford, L., Lindeman, M., Stamation, K., Bluff, L., Liu, C., Whitehead, B., Albert, R., Bowd, Sutter, G., Cheal, D., & Loyn, R. M. and Smith, A. 2010. Evaluation 2015. Effects of fire regimes on of camera trap sampling designs biodiversity in Gippsland, Victoria: used to determine change in a retrospective approach. Arthur occupancy rate and abundance Rylah Institute for Environmental of feral cats. Technical Report Research, Heidelberg, Victoria, Series no. 201, Arthur Rylah for Department of Environment, Institute for Environmental Land, Water and Planning, Research, Department of Melbourne, Victoria. Sustainability and Environment, Heidelberg, Victoria. Nelson, J. L. and Scroggie, M. P. 2009. Remote cameras Swann, D. E., Hass, C. C., Dalton, as a mammal survey tool. D. C. and Wolf, S.A. 2004. Survey design and practical Infrared-triggered cameras for considerations. Arthur Rylah detecting wildlife: an evaluation Institute for Environmental and review. Wildlife Society Research, Heidelberg, Victoria. Bulletin 32: 357-365.

Red Fox

Caught on Camera | Citizen science in the Wombat State Forest | PAGE 27 Victorian National Parks Association Level 3, 60 Leicester St, Carlton VIC 3053 PH: 03 9341 6500 | EMAIL: [email protected] | WEB: vnpa.org.au All donations over $2 are tax-deductible. ABN 34 217 717 593