Camera Trapping: Wildlife Management and Research

By Peter Fleming, Peter Banks, Andrew Claridge and Jim Sanderson

Camera trapping in wildlife management and research is a growing global phenomenon. The technology is advancing very quickly, providing unique opportunities for collecting new biological knowledge. In order for fellow camera trap researchers and managers to share their knowledge and experience, the First International Camera Trapping Colloquium in Wildlife Management and Research was held in Sydney, Australia.

Camera Trapping brings together papers from a selection of the presentations at the colloquium and provides a benchmark of the international developments and uses of camera traps for monitoring wildlife for research and management. Four major themes are presented: case studies demonstrating camera trapping for monitoring; the constraints and pitfalls of camera technologies; design standards and protocols for camera trapping surveys; and the identification, management and analyses of the myriad images that derive from camera trapping studies. The final chapter provides future directions for research using camera traps.

Remarkable photographs are included, showing interesting, enlightening and entertaining images of animals 'doing their thing', making it an ideal reference for wildlife managers, conservation organisations, students and academics, pest animal researchers, private and public land managers, wildlife photographers and recreational hunters.

• Language: English
• Publisher: CSIRO PUBLISHING
• Release date: Nov 20, 2014
• ISBN: 9781486300419

Book preview

PART 1

Camera trapping for animal monitoring: case studies

1

Camera trapping for animal monitoring and management: a review of applications

Don E. Swann and Nic Perkins

Abstract

Camera traps are being used throughout the world to address a wide range of issues in wildlife management and to address both research and management questions that cannot be easily answered with other methods. In addition to detecting rare species and providing answers to practical management questions, camera traps have a potentially large role in assessing global changes in biodiversity of mammals. The quality of camera traps is continuing to improve, and field and analytical techniques are also moving rapidly forward. This paper reviews the current state of camera trapping in wildlife ecology with a focus on new and emerging applications in management and monitoring. Recent papers, including many in this volume, indicate that camera traps have the potential to be a powerful new tool in areas of animal ecology where they have not previously been widely used, such as estimation abundance, sampling of small animals, and establishing conservation priorities based on regional monitoring. In addition, the use of camera traps by citizen scientists and environmental educators continues to grow and become more integrated with more traditional scientific studies.

Introduction

Cameras that record images of wild animals when humans are not present have a long history in ecology, but their use dramatically increased with the introduction of commercial infrared-triggered cameras in the early 1990s. Today, the term ‘camera trap’ typically refers to cameras units that are triggered by the movement of an animal within a detection area, although the term also can describe cameras set to take photos at set time intervals. Nearly all camera traps used in current wildlife applications are small (the size of a shoebox or smaller), consist of only one piece, shoot digital still or video images, and are passively triggered using an infrared light source. Nevertheless, a dazzling array of commercial camera traps and optional features are available (Rovero et al. 2013; Swann et al. 2004, 2011; <http://www.trailcampro.com>), and these can be further modified by researchers.

Camera traps have been applied to nearly every aspect of vertebrate ecology, including to study nest ecology, research activity patterns and behaviour, document rare species or events, and estimate state variables such as species richness, occupancy, abundance (Cutler and Swann 1999; Kucera and Barrett 2011). Data recorded by camera traps typically consist of an image (e.g. Plate 1), series of images, or video of an individual or group of animals within the area of detection covered by the camera trap, as well as other information such as the date, time, and location of the photograph. Because most individuals in the image can be identified to species, the trap thus records the presence of that species at that place and time. Other information, such as behavioural data (e.g. Meek et al. 2012) or events such as predation or feeding (e.g. Zimmerman et al. 2011) can also be recorded. In some cases individual animals can be identified either through tags previously affixed by researchers or by unique natural marks. Some data gathered by physical trapping techniques, such as reproductive condition and genetic data, cannot usually be obtained by camera traps, but camera trap data are often combined with other techniques such as radio-telemetry (e.g. Larrucea et al. 2007) and genetics (e.g. Janečka et al. 2011).

Despite the limits of the data that can be gathered by camera traps, experience during the past two decades indicates why they are powerful tools for addressing conservation of populations of native species, especially mammals. First, camera traps provide basic knowledge of the distribution of mammals (their presence in a certain place), which is essential for conservation on both the local and regional scale, but often previously lacking for the many species that are nocturnal, avoid humans, and seek cover. Second, they are relatively inexpensive, which means they can be deployed very efficiently to increase sample sizes over wide areas (De Bondi et al. 2010). And third, camera traps are relatively non-invasive and safe for both humans and animals. From a practical point of view, it is obviously much easier to sample populations of very large mammals with camera traps than with live traps.

As a result, camera trapping has been truly significant for wildlife management and conservation throughout the world. Camera traps have documented species that are new to science, or occur in areas where they were thought to be locally extinct or not previously known to exist (e.g. Sangay et al., Chapter 10). Kucera and Barrett (2011) list several recent examples, including a new species of striped rabbit (Nesolagus timminsi) in South-east Asia (Surridge et al. 1999), a range extension for the Sulawesi palm civet (Macrogalidia musschenbroekii) (Lee et al. 2003) and the documentation of the wolverine (Gulo gulo) in California for the first time since 1922 (Moriarty et al. 2009). Camera traps also have been used to reliably estimate, for the first time, the abundance of the tiger (Panthera tigris; Karanth and Nichols 1998) and other species where individuals can be readily recognised based on their spot pattern. After many years of debate and poor information on the number of species of mammals present in natural areas, camera traps are now producing reliable estimates of species richness and other community measures (Tobler et al. 2008; O’Brien et al. 2011) that are not based on poor-quality observational data or generalised range maps (e.g. Newmark 1995; Parks and Harcourt 2002).

As demonstrated at the First International Camera Trapping Colloquium in Wildlife Management and Research, new developments in camera trapping are arriving at a staggering pace. Camera traps themselves are rapidly evolving, becoming faster, cheaper, more resilient, and more versatile (Meek 2012). Researchers are developing new techniques for deploying them, including using vertical mounts (Smith and Coulson 2012; Welbourne, Chapter 20). More researchers are aware that detectability is always < 1 in camera studies (Kéry 2011) and are becoming more adept and sophisticated at analysing camera trap data using occupancy and other methods (O’Connell and Bailey 2011). They are rapidly improving methods for managing data, including extracting data directly from photos, eliminating the need for data entry, and even developing methods for using pattern recognition to identify animals automatically (Falzon et al., Chapter 28). And of course, they are improving methods for sharing data through the internet and cell phone applications.

What is truly exciting about camera trapping in the modern era is that the explosive adoption of this technology, in synergy with improvements in camera trap quality, wildlife data analysis, and information management, is resulting in novel applications in wildlife conservation around the world. It is a testament to the rapid advances in camera trapping that several important developments in camera trapping have occurred since the very recent publication of a major book on the subject, Camera Traps in Animal Ecology (O’Connell and Bailey 2011). The goal of this paper is to provide a short summary of some of these applications and explore their potential for creating ground-breaking developments in the coming years.

Estimating animal abundance and density

Camera traps are often used to provide an index of abundance (also called relative abundance), such as the number of photos of a species per trap night. However, indices typically provide biased estimates of abundance (Anderson 2001), which in camera trapping is primarily due to spatial variability and detectability (O’Brien 2011). Many users of camera traps assume that the number of photographs per unit time is an accurate reflection of the number of individual animals present. However, this assumption may not be valid because many factors may influence the number of photographs, including attraction to the camera trap, trap shyness, use of or type of attractant, weather, ability of the camera trap to detect an animal when present, and others. Several recent studies (see Karanth et al. 2011) have followed Karanth (1995), who used camera traps and individual natural markings to estimate tiger abundance and density using capture–recapture models. However, this approach does not work for species without features that allow them to be individually recognised. The recent work by Marcus Rowcliffe and colleaques (Rowcliffe et al. 2008, 2011) is thus truly important for camera trap researchers, as it presents an opportunity to estimate density by modelling the underlying process of the encounter between the camera trap and the animal. Their random encounter model (REM) relies on characteristics of the camera trap (the distance and angle with which it detects animals) and the characteristics of animals that can be determined from videos taken at the camera sites, including its size and speed.

One of the complications of using the REM has to do with the difficulty of accurately estimating the area of the detection zone. Although this problem has been addressed using distance sampling techniques by Rowcliffe et al. (2011), papers presented in this volume (e.g. Welbourne, Chapter 20) suggest that another approach may be to mount cameras vertically (facing down at the ground), so as to more precisely control the area of detection. This approach may not be appropriate for larger animals in tropical areas where a large amount of vegetation is present, but might work in less vegetated areas and for smaller mammals in most habitats.

Studying small animals

Another interesting new direction for camera trapping is in studies of smaller animals, including mammals such as insectivores, rodents, and small marsupials, as well as birds and herpetofauna (e.g. Welbourne, Chapter 20). Camera traps have been used in small mammal research for some time, but typically to determine temporal patterns (e.g. Pearson 1959). Recent work on small mammals with more accurate new commercial traps, particularly in Australia with Reconyx™ brand traps that pick up smaller heat signatures (e.g. Meek et al. 2012; Weerakoon et al., Chapter 29), allow researchers to estimate variables such as abundance (where individual animals are marked or have identifiable patterns), species richness, and occupancy that previously required the use of live traps. The work by Rowcliffe et al. (2008, 2011) applies to estimating density of small mammal abundance, which has the potential to replace live trapping for many studies, especially if combined with vertical camera mounts on a network grid of small mammal sensors at random locations with no bait. A paper presented at the First International Camera Trapping Colloquium in Wildlife Management and Research by Falzon et al. (Chapter 28) presented a method for automated species identification that allows for removal of the consistent background image in photos and use of visual algorithms to identify individual species. Although this approach is still in early stages of development, it is expected to grow and have important implications for monitoring not only abundance and density using the REM, but also species richness and occupancy. Further, because infrared triggered cameras may always have imperfect detectability, it seems conceivable that in the future constant photo or video surveillance of a series of unbaited, random plots would produce the potential for higher detection rates and very accurate estimation of small mammal populations, as well as populations of herpetofauna and ground birds.

Finally, camera traps have typically been used in studies of mammals and birds, and only occasionally used in studies of reptiles and amphibians. Dustin Welbourne’s presentation at the Colloquium (Welbourne, Chapter 20) is interesting not only because it focuses on herpetofauna, but is potentially ground-breaking in its approach to inventory and monitoring of this taxonomic group. Even more than small mammals, reptiles and amphibians are very difficult to detect due to their small size and (as ectotherms) lack of a heat signature. Many species are rare, cryptic, and active only during specific windows of heat and humidity. Herpetologists have traditionally used a variety of methods to estimate population parameters, including observations and times searches (Visual Encounter Surveys, or VES), pitfall traps, hoop traps, and other methods. Welbourne’s work suggests that a more efficient and less invasive approach may be to use camera traps, mounted vertically over plates that can be naturally warmed to a different temperature than the surrounding substrate, in conjunction with drift fences. It remains to be seen whether this technique can be modified to sample the range of terrestrial reptiles and amphibians with different types of movement and habitat use (e.g. sand-dwelling, rock-dwelling, aquatic and aboreal). In addition, it may be possible to eliminate drift fences if capture rates are high enough. As with small mammals, this technique can further developed with the use of species recognition software and non-infrared triggered cameras that run continuously.

Applied management

While large conservation issues, such as preserving the biological diversity of our planet, require a large scale vision (Wilson and Peter 1988), most actual conservation happens locally. Every day, throughout the world, wildlife managers strive to reduce the impacts of introduced predators, develop structures to prevent threatened species from being killed by cars, create artificial water resources to replace lost natural sources, build gates on caves to protect bats, and engage in many other types of hands-on conservation efforts.

The basic type of data collected by camera traps – the image or images of an animal in a certain place at a certain time – has been very useful for addressing local wildlife management questions for some time, and continues to be very useful. Recent published papers around the world expand the uses of camera traps to topics as diverse as assessing the value of artificial water sources (Krausman et al. 2006); comparing different types of forestry practices on mammal communities (Samejima et al. 2012); assessing effects of human presence on large carnivore populations (Muhly et al. 2011); assessing effects of predator control on prey species populations (Salo et al. 2010); evaluating the potential for expansion of the wild boar population in Switzerland (Wu et al. 2011); and many others. Case studies at the colloquium that used camera traps for applied management applications included addressing the effectiveness of different types of baits for decreasing introduced carnivores while preserving native species (Antos and Yuen, Chapter 2; Moseby and Read, Chapter 14; Bengsen, Chapter 31); quantifying the use of water holes by introduced camels and their interactions with water availability and native animals (Ninti One Ltd 2013); whether and how native mammals use glide poles and other structures to successfully cross roadways (Taylor and Goldingay, Chapter 22); and identifying the predators of the New Zealand kea (Nestor notabilis, B. Barrett pers. comm.). In addition, camera traps are being used in some truly innovative management applications, such as to identify Tasmanian devils (Sarcophilus harrisii) with devil facial tumour disease and track the spread of this disease (Thalmann et al., Chapter 3).

The main trend in using camera traps in applied management studies is that camera traps are rapidly becoming more affordable and reliable. Many wildlife managers, especially in developing countries, have lacked effective methods for determining whether their conservation actions are indeed effective. The potential for obtaining larger numbers of camera traps make it easier not only to capture data that may provide insight into a particular management issue, but also to implement studies with a higher sample size, thus providing results that allow for greater inference.

Monitoring of animal communities at local, regional and international scales

Knowledge of the number of species occurring in a particular place, either locally as in a national park, or regionally or even globally, is essential for managing and conserving biodiversity, but the poor state of this knowledge remains one of the greatest hindrances to conservation (O’Brien et al. 2011). Because funding for conserving biodiversity are always less than what is needed, establishing conservation priorities is one of the most vexing problems in conservation biology (Isaac et al. 2007). Three-quarters of all species-based conservation projects are specifically aimed at charismatic megafauna (Leader-Williams and Dublin 2000). While part of the problem is the nature of the public’s attention, the lack of information on many species is also an issue.

There are many challenges to tracking of the status of different taxonomic groups, ranging from the sheer number of species to the complexities of their life cycles. The challenge of monitoring animals typically detected by camera traps, particularly medium and large mammals, is that they are often rare, nocturnal, and avoid humans. However, camera traps have proven to be very useful in over coming these issues. Thus the greatest current challenge may be to organise researchers to deploy cameras in a systematic way and share the data so the information can be most useful for monitoring trends over time and establishing conservation priorities. Because the focus is on entire communities, such an approach requires a randomised study design with unbaited camera traps. On a local level, an increasing number of such programs in national parks and other reserves are beginning to monitor communities in this way.

Even more exciting is the development of programs that monitor the status of communities on larger geographic scales. A primary example of this is the Terrestrial Ecology Assessment and Monitoring (TEAM) network (Ahumada et al. 2011; Jansen et al., Chapter 24), which, in an international partnership at 16 sites across the globe, collects data useful for conservation of tropical forest mammals. TEAM has developed a detailed, standard protocol and data storage and sharing methods that allow for comparisons of species richness, diversity, and evenness among sites. In addition, programs like the Wildlife Picture Index (WPI) (O’Brien et al. 2010; Townsend et al., Chapter 5) combine camera traps with occupancy analysis and generalised additive models to allow conservationists to monitor trends in biodiversity over time. The success of approaches such as TEAM and WPI suggest that similar standardised approaches can accomplish important conservation work across a variety of regional and national scales, and deserve to be implemented more widely.

Citizen science and education

Two final developments in applications of camera traps in monitoring and management are the growing use of camera traps in education and the establishment of citizen science networks that involve non-scientists in collecting important scientific data (e.g. Griffiths and Lewis, Chapter 9; Thomas, Chapter 8 volume). There is considerable overlap between these two areas. In a recent major event at Saguaro National Park in Arizona, the 2011 BioBlitz, students learned to use and deploy camera traps in an effort to both document mammals and learn more about their habits and conservation needs, and the park posted most of the photos gathered during and after the event (Fig. 1.1 and Plate 2; ). Photos from camera traps are being used the world over to excite people about the wildlife around them that usually remains unseen. These images are often quite powerful and have great potential for promoting conservation among citizens who might otherwise not be interested.

Fig. 1.1. Camera trapping as education: students at saguaro national Park, USA, set a camera trap as part of 2011 BioBlitz.

The primary user of camera traps in many countries, and particularly the United States, is by hunters. In the USA, the value of hunter-collected data in the documentation of rare and threatened cannot be underestimated. Where we live in Arizona, USA, for example, camera traps set by hunters have provided some of the best recent records of rare tropical cats such as jaguars and ocelots. Networks of camera traps maintained by volunteers are proving to be valuable for conservation in some areas (e.g. Erb et al. 2012). The effort to tap into the larger pool of existing camera trap data (which can be easily found by making a Google search of Flickr and camera traps, for example) includes eMammal (<http://www.facebook.com/eMammal>; ), which is being developed as an international archive of wildlife photos by the Smithsonian Institute. At the same time, some caution is in order. Camera traps are a neutral way to detect animals, and can be used just as easily as a tool to detect and kill wildlife as to detect and conserve them. As researchers share information, it is important to recognise that the potential downsides of sharing the secrets of wildlife.

Conclusion

What makes camera trapping unique among the many methods used by wildlife ecologists is that it is, in a sense, both a science and an art. It is a science in that camera traps can produce data that are useful for addressing questions about our world works. But the same process produces images that can be ‘appreciated primarily for their beauty or emotional power’, which is how the Oxford Dictionary (<http://oxforddictionaries.com/definition/english/art>) defines art. As Dustin Welbourne recently wrote, ‘the raw data collected with cameras and recorders can be experienced. It is visceral and tangible to our senses, and even seasoned researchers are often excited to ‘see’ what has been detected’ (Welbourne 2012b). It is important for scientists and managers that the art function of photos produced by camera traps does not become a substitute for the science function; that is, no matter how excellent and interesting a camera trap images are, they cannot replace the requirement of having clear objectives, strong study design, and efficient and accurate methods for managing photo data.

At the same time, the art value of camera traps has a place in conservation biology that is significant and not likely to go away soon. It is this dynamic between science and art that would seem to ensure that the value and creative uses of camera traps are likely to continue to grow, as today’s camera traps merge with newer technologies in ways that may rapidly change the definition of the term. The challenge for conservation biologists is to harness this creativity with sound science in ways that will help the Earth’s many magnificent creatures survive until the next generation.

References

Ahumada JA, Silva CEF, Gajapersad K, Hallam C, Hurtado J, Martin E, McWilliam A, Mugerwa B, O’Brien T, Rovero F, Sheil D, Spironello WR, Winarni N, Andelman SJ (2011) Community structure and diversity of tropical forest mammals: data from a global camera trap network. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences 366, 2703–2711.

Anderson DR (2001) The need to get the basics right in wildlife field studies. Wildlife Society Bulletin 29, 1294–1297.

Cutler TL, Swann DE (1999) Using remote photography in wildlife ecology: a review. Wildlife Society Bulletin 27, 571–581.

De Bondi N, White JG, Stevens M, Cooke R (2010) A comparison of the effectiveness of camera trapping and live trapping for sampling terrestrial small-mammal communities. Wildlife Research 37(6), 456–465.

Erb PL, McShea WJ, Guralnick RP (2012) Anthropogenic influences on macro-level mammal occupancy in the Appalachian Trail Corridor. PLoS ONE 7(8), e42574.

Isaac NJB, Turvey ST, Collen B, Waterman C, Baillie JEM (2007) Mammals on the EDGE: conservation priorities based on threat and phylogeny PLoS ONE 2(3), e296.

Janečka JE, Munkhtsog B, Jackson RM, Naranbaatar G, Mallon DP, Murphy WJ (2011) Comparison of noninvasive genetic and camera-trapping techniques for surveying snow leopards Journal of Mammalogy 92, 771–783.

Karanth KU (1995) Estimating tiger Panthera tigris populations from camera-trap data using capture-recapture models. Biological Conservation 71, 333–338.

Karanth KU, Nichols JD (1998) Estimation of tiger densities in India using photographic captures and recaptures. Ecology 79(8), 2852–2862.

Karanth KU, Nichols JD, Kumar NS (2011) Estimating tiger abundance from camera trap data: field surveys and analytical issues. In AF O’Connell, JD Nichols and KU Karanth (Eds) Camera Traps in Animal Ecology. Pp. 97–118. (Springer Publishing, Tokyo).

Kéry M (2011) Species richness and community dynamics: a conceptual framework. In AF O’Connell, JD Nichols and KU Karanth (Eds) Camera Traps in Animal Ecology. Pp. 207–232 (Springer Publishing, Tokyo).

Krausman PR, Rosenstock SS, Cain JW, III(2006) Developed waters for wildlife: science, perception values, and controversy. Wildlife Society Bulletin 34, 563–569.

Kucera TE, Barrett RH (2011) A history of camera trapping. In AF O’Connell, JD Nichols and KU Karanth (Eds) Camera Traps in Animal Ecology. Pp. 9–26. (Springer Publishing, Tokyo).

Larrucea ES, Brussard PF, Jaeger MM, Barrett RG (2007) Cameras, coyotes, and the assumption of equal detectability. The Journal of Wildlife Management 71, 1682–1689.

Leader-Williams N, Dublin HT (2000) Charismatic megafauna as ‘flagship species’. In A Entwistle and N Dunstone (Eds) Priorities for the Conservation of Mammalian Diversity: Has the Panda had its Day? Pp. 53–81. (Cambridge University Press Cambridge, England).

Lee RJ, Riley J, Hunowu I, Maneasa E (2003) The Sulawesi palm civet: expanded distribution of a little known endemic viverid. Oryx 37, 378–381.

Meek PD (2012) Refining and improving the use of camera trap technology for wildlife management and research in Australia and New Zealand. Unpublished report to the Winston Churchill Memorial Trust of Australia.

Meek PD, Zewe F, Falzon G (2012) Temporal activity patterns of the swamp rat (Rattus lutreolus) and other rodents in north-eastern New South Wales, Australia. Australian Mammalogy 34, 223–233.

Muhly TB, Semeniuk C, Massolo A, Hickman L, Musiani M (2011) Human activity helps prey win the predator-prey space race. PLoS ONE 6(3), e17050.

Newmark WD (1995) Extinction of mammal populations in western American national parks. Conservation Biology 9, 512–526.

Ninti One Ltd (2013) Managing the Impacts of Feral Camels across Remote Australia. Final Report of the Australian Feral Camel Management Project. (Ninti One Limited, Alice Springs).

O’Brien TG (2011) Abundance, density, and relative abundance: a conceptual framework. In AF O’Connell, JD Nichols and KU Karanth (Eds) Camera Traps in Animal Ecology. Pp. 71–96. (Springer Publishing, Tokyo).

O’Brien TG, Baille JE, Cuke M (2010) The wildlife picture index: monitoring top trophic levels. Animal Conservation 13, 335–343.

O’Brien TG, Kinnaird MF, Wibisono HT (2011) Estimation of species richness of large vertebrates using camera traps: an example from an Indonesian rainforest. In AF O’Connell, JD Nichols and KU Karanth (Eds) Camera Traps in Animal Ecology. Pp. 233–252. (Springer Publishing, Tokyo).

O’Connell AF, Bailey LL (2011) Inference for occupancy and occupancy dynamics. In AF O’Connell, JD Nichols and KU Karanth (Eds) Camera Traps in Animal Ecology. Pp. 191–206. (Springer Publishing, Tokyo).

Parks SA, Harcourt AH (2002) Reserve size, local human density, and mammalian extinctions in U.S. protected areas. Conservation Biology 16, 800–808.

Pearson OP (1959) A traffic survey of Microtus Reithrodontomys runways. Journal of Mammalogy 40, 169–180.

Rovero F, Zimmermann F, Berzi D, Meek P (2013) ‘Which camera trap type and how many do I need?’ A review of camera features and study designs for a range of wildlife research applications. Hystrix 24, 148–156 <http://www.italianjournal-of-mammalogy.it/article/view/6316/pdf>.

Rowcliffe JM, Field J, Turvey ST, Carbone C (2008) Estimating animal density using camera traps without the need for individual recognition. Journal of Applied Ecology 45, 1228–1236.

Rowcliffe JM, Carbone C, Jansen PA, Kays R, Kranstauber B (2011) Quantifying the sensitivity of camera traps: an adapted distance sampling approach. Methods in Ecology and Evolution 2, 464–476.

Salo P, Banks PR, Dickman CR, Korpimäki E (2010) Predator manipulation experiments: impacts on populations of terrestrial vertebrate prey. Ecological Monographs 80, 531–546.

Samejima H, Ong R, Lagan P, Kitayama K (2012) Camera-trapping rates of mammals and birds in a Bornean tropical rainforest under sustainable forest management. Forest Ecology and Management 270, 248–256.

Smith JK, Coulson G (2012) A comparison of vertical and horizontal camera trap orientations for detection of potoroos and bandicoots. Australian Mammalogy 34, 196–201.

Surridge AK, Timmins RJ, Hewitt GM, Bell DJ (1999) Striped rabbits in Southeast Asia. Nature 400, 726.

Swann DE, Hass CC, Dalton DC, Wolf SA (2004) Infrared-triggered cameras for detecting wildlife: an evaluation and review. Wildlife Society Bulletin 32, 357–365.

Swann DE, Kawanishi K, Palmer J (2011) Evaluating types and features of camera-traps in ecological studies: a guide for researchers. In AF O’Connell, JD Nichols and KU Karanth (Eds) Camera Traps in Animal Ecology. Pp. 27–43. (Springer Publishing, Tokyo).

Tobler MW, Carillo-Percastegui SE, Pitman RL, Mares R, Powell G (2008) An evaluation of camera traps for inventorying large and medium-sized terrestrial rainforest mammals. Animal Conservation 11, 169–178.

Welbourne D (2012b) Ready for my close-up: camera traps bring home the wild. The Conversation. <http://theconversation.edu.au/ready-for-my-closeup-camera-traps-bring-home-the-wild-9707>.

Wilson EO, Peter FM (Eds) (1988) Biodiversity. (National Academies Press, Washington).

Wu N, Abril C, Hinic V, Brodard I, Thur B, Fattebert J, Hussy D, Ryser-Degiorgis MP (2011) Freeranging wild boar: a disease threat to domestic pigs in Switzerland? Journal of Wildlife Diseases 47, 868–879.

Zimmerman F, Molinari-Jobin A, Ryser A, Breitenmoser-Wursten Ch, Presenti E, Breitenmoser U (2011) Status and distribution of the lynx (Lynx lynx) in the Swiss Alps 2005–2009. Acta Biologica Slovenica 54, 74–84.

2

Camera trap monitoring for inventory and management effectiveness in Victorian national parks: tailoring approaches to suit specific questions

Mark J. Antos and Kally Yuen

Abstract

This paper evaluates two approaches of camera trap use to answer different questions of interest to land managers by using two studies as examples. The first study aims to examine the effectiveness of fox baiting activities by comparing occupancy of foxes and their prey species over time across baited and unbaited areas. Each treatment had 20 camera trap sites that were activated annually for a 21-day monitoring period between 2009 and 2012. While there was no strong support for fox baiting having an effect on modelled occupancy rates for foxes and mammalian prey species, there was strong support for fox baiting having an effect on modelled occupancy of a ground-dwelling bird, the Bassian thrush.

The aim of the second study was to provide an inventory of ground-dwelling mammals across an entire national park with a lack of recent and comprehensive mammal surveys. Camera traps were deployed at 134 randomly selected sites for 1 to 4 weeks over a 6-month period during 2011–12. A total of 32 native species was detected, including 11 mammals. The survey revealed the presence of species of regional and state conservation concern including the long-nosed bandicoot and white-footed dunnart as well as cryptic species that have been only infrequently recorded in the past, such as the Lewin’s rail.

The implications of the findings for refinement of monitoring approaches and how they will be used to assist with land management are discussed. Specifically, the importance of appropriate experimental design, operator training, being prepared for unexpected but useful data and using a range of complementary survey methods are emphasised.

Introduction

Land management agencies are interested in monitoring fauna for a range of different reasons and to address multiple questions to help with future environmental management. Parks Victoria has developed the Signs of Healthy Parks Ecological Monitoring Program to guide this management (Parks Victoria 2010). Core to the program is the identification of the key natural values of a park, the threats to those values, and the development of monitoring questions and methods to provide information on the status of both. The program provides a prioritisation process that ensures that monitoring activities are well planned and designed, of management relevance and align with key legislative and management obligations. These principles are important to any successful longterm evaluation program (Hockings et al. 2000; Lindenmayer and Likens 2010).

The Signs of Healthy Parks Ecological Monitoring Program has identified ground-dwelling mammal assemblages as being key values in several Victorian national parks, including the Great Otway, Mornington Peninsula and Point Nepean National Parks. Camera traps were identified as being the most appropriate and efficient method for monitoring small mammals, including introduced predators, especially given their demonstrated usefulness in the survey of such species (e.g. Claridge et al. 2004; De Bondi et al. 2010; Nelson et al. 2010; Bengsen, Chapter 31; Robley et al., Chapter 27).

This paper examines two different approaches to the use of camera traps as a wildlife monitoring tool. In the Great Otway National Park, where good baseline data on mammal distribution exist, camera traps were used to address specific land management questions:

1 Is ongoing fox baiting successful in reducing red fox Vulpes vulpes occupancy?

2 Is there a positive response from potential prey species in areas subject to fox baiting?

The threat posed by foxes to a range of fauna, especially ground-dwelling mammals, has been well documented (e.g. Burbidge and McKenzie 1989; Saunders et al. 1995) and there are many large scale and resource intensive control projects underway (e.g. Murray et al. 2006). An understanding of the efficacy of control methods and approaches is necessary to evaluate and improve existing practices.

Within the Mornington Peninsula and Point Nepean National Parks information on the presence and distribution of mammal species is limited. Consequently, questions were posed around broad scale inventory:

1 Which mammal species are present in the park?

2 Which areas of the park are most important for conserving native mammal diversity?

Addressing these questions is important to park managers as it provides a baseline of what is currently present in the park and the results can be used to develop more focused conservation and management objectives.

Methods

Study sites

The Great Otway National Park is located on the coast and coastal ranges of south-west Victoria and protects over 103 000 ha (Fig. 2.1). It encompasses a relatively intact and highly diverse landscape which includes vegetation types such as cool temperate rainforest, wet forest, dry forest and woodland, heathland, coastal scrub and wetlands (Parks Victoria 2009). The park contains an altitudinal gradient ranging from sea level to montane forests over 600 m in elevation. It has a strong annual rainfall gradient ranging from over 2000 mm in the far south-west of the park to less than 800 mm in the north-east (VEAC 2003). Previous faunal surveys indicate that the park supports a diverse range of small and medium-sized ground-dwelling mammals, including dasyurids (Dasyuridae), bandicoots (Peramelidae) and long nosed potoroos (Potorous tridactylus)(Menkhorst 1996; VEAC 2003).

Point Nepean and Mornington Peninsula National Parks share a common boundary and will be hereafter referred to as the Mornington Peninsula National Park. The park is located on the southern end of the Mornington Peninsula on the central coast of Victoria (Fig. 2.1). It is a long linear park of over 3000 ha with a high edge to area ratio.

Fig. 2.1. Location of great Otway and Point nepean/mornington Peninsula national Parks (black) in victoria, Australia.

It primarily consists of coastal scrub and woodland although it also supports some dry forest further from the coast. It is the most highly visited national park in Victoria with ~2.5 million visits recorded per year (Parks Victoria 1998). Most of the park is located at sea level with the highest point rising to ~180 m and annual rainfall is 600–800 mm (Calder 1986). There have been few comprehensive fauna surveys of the park, but several older unpublished studies (e.g. Ehmke et al. 2008) indicate that bandicoots and white footed dunnarts (Sminthopsis leucopus) are present at a small number of sites, with Antechinus species being more widespread.

Experimental design

Great Otway National Park

Camera traps were deployed in two different locations in the park, separated by ~60 km, containing different habitats; the west Otways comprising wet forests and high rainfall and the east Otways comprising heathy woodlands and lower annual rainfall. Each location (approx. 5000 ha) was divided into an area subject to pulsed fox baiting (four times a year with 9 weeks per pulse interspersed with a 4-week resting period) with buried 1080 baits and one area used as a control with no baiting. Twenty camera traps were deployed at fixed sites in each area and the two areas were separated by a 2-km buffer with no baiting and no camera trap sites. Camera traps were spaced 1 km apart and were located within 100 m of existing track networks. Monitoring in the east Otways commenced in December 2008 and fox baiting commenced during April of 2009, thus providing pre-baiting data at both locations. In the west Otways, monitoring commenced during June 2009 and fox baiting commenced during April 2010.

In the west Otways meat lures (chicken with fish oil) were used to increase the possibility of detecting carnivorous marsupials while in the east Otways a mix of peanut butter, oat and golden syrup was used. Lures were placed in small cages and attached to fixed poles within 3 m of the camera traps. In order to test differences between different lures, the camera trap sites in the west Otways were also used during September and October 2009–11 but baited with a mix of peanut butter, oat and golden syrup and results were compared with the winter surveys when meat lures were used.

Camera trap models used were the PixController™ DSC-W35/55 set to high sensitivity with 10 s intervals between activation and one image taken per trigger. Some ScoutGuard® DTC530 camera traps were also used in conjunction with the PixControllers. These were continually rotated among individual sites and records made of which camera trap model was active at which location. Camera traps were active day and night for monitoring sessions consisting of a minimum 21-day period, but this occasionally fell short due to malfunction or operator error.

Mornington Peninsula National Park

In Mornington Peninsula National Park, 134 baited camera trap sites were established at randomly selected points proportionately stratified according to the main vegetation types present. Sites were located at least 300 m apart. ScoutGuard DTC530 camera traps were used on high sensitivity video mode set to film for 10 seconds with a 1-minute interval between recordings. Camera traps were mounted 1 m above the ground. From December 2011 to June 2012 camera traps were deployed at individual sites for one to 4 weeks and rolled out to new sites upon completion of each survey round. Lures consisting of a peanut butter, oat and golden syrup mix were placed in tea strainers and pegged into the ground ~1.5–2 m from camera traps.

Data analysis

All images were identified to species level where possible using experienced Parks Victoria staff and, where necessary, outside personnel. Some images were of insufficient quality to enable a positive identification and were excluded from all analyses. Data were sorted according to which species were detected at each camera trap site during each 24-hour period.

Great Otway National Park

The species of key interest for analysis were the red fox and a range of critical weight range mammals, namely bandicoots and potoroos. These species have been shown to be particularly sensitive to introduced predators (Burbidge and McKenzie 1989; Lunney 2001). Because detection rates for both species of bandicoot and the potoroo were low, they were combined into a single group for analysis. The effect of fox baiting on occupancy rate over time for red fox, the bandicoot/potoroo group and the Bassian thrush (Zoothera lunulata) was tested using a dynamic occupancy model based on MacKenzie et al. (2003). The Bassian thrush, a non-target species, was frequently detected by camera traps in the west Otways. Given that this species forages on the ground (Higgins et al. 2006) and is therefore likely to be susceptible to fox predation, we have analysed its occurrences over the years despite it not being an initial target species.

For each analysis, a total of 14 candidate models were fitted to the data, some allowing for the possible effect of fox baiting, some allowing for the possible effect of year and some allowing for possible effect of year for different fox baiting status on the occupancy probability or detection probability or both over time. Six of the candidate models have considered the possible effect of fox baiting on the occupancy rate over time. Of the six models, four allow for the possible effect of fox baiting only and the remaining two allow for possible effects of fox baiting and year. The Akaike Information Criterion (AIC) weights corresponding to these six candidate models were summed to obtain the level of support of the hypothesis that occupancy rate over time is dependent on fox baiting status. Classification of levels of support for hypotheses based on summed model weights follows MacKenzie et al. (2005). The top ranked candidate model, and those models for which the difference in AIC from the top ranked model was no more than two, were selected to examine the possible changes of occupancy rate over time (Burnham and Anderson 2002). In the case of more than one model being selected, a model average of the occupancy rate for each year was calculated. Standard errors (s.e.) for the occupancy rate for all years associated with each model were obtained based on 1000 bootstrap simulations. The analysis was carried out using function colext in the ‘unmarked’ package (Fiske and Chandler 2011) in R (R Development Core Team 2011). No detailed analysis of the data series from the east Otways is presented due to low detection of key species in this area over the years.

Mornington Peninsula National Park

A reporting rate for each species at each site was calculated as the number of days a species was detected per number of active camera trap days. For mapping purposes, the species richness score of each of the camera trap sites was converted into a continuous surface using the kriging interpolation algorithm. Kriging is a geostatistical estimator working under the assumption that an unknown point should be the average of the known points at its neighbours, weighted by the neighbours’ distance to the unknown point. The result is a grid of continuous cells, each with a predicted value of species richness.

Results

Great Otway National Park

During the course of the project, there was an improvement in the number of days that camera traps were operational in the west Otways ranging from a mean of 16.6 days in 2009 to 21.8 in 2012 (Table 2.1). Although the aim was to have camera traps active for at least a 3-week (21 days) period, various factors such as camera trap failure and operator error prevented this from occurring during the early stages of the project. With greater experience and training the duration of camera trap function days increased.

A total of 18 native species, including 11 mammals, were recorded over 4 years at 40 camera trap sites in the wet forests of the west Otways. The swamp wallaby (Wallabia bicolor), agile antechinus (Antechinus agilis), dusky antechinus (A. swainsonii) and the long-nosed bandicoot (Perameles nasuta) were all detected from > 20% of camera trap sites. The long-nosed potoroo and the southern brown bandicoot (Isoodon obesulus) were less frequently detected, typically not being recorded from more than three sites in any given year. Several bird species were also recorded with the most common being the Bassian thrush.

On balance, the meat and peanut butter, oats and golden syrup lures were both effective in attracting a wide range of wildlife in terms of the numbers of camera trap sites where species were detected. A closer examination reveals that some key species of interest such as the red fox, long-nosed bandicoot, southern brown bandicoot and long-nosed potoroo were recorded at more sites when the meat lures were used (Table 2.2), but the results for the last two species may have been by chance due to the low numbers detected.

For the west Otway monitoring sites using meat as lures, the sums of the AIC weight from the six candidate models which allow for the possible effect of fox baiting on the occupancy rate over time are 5% for red fox, 17% for bandicoots and potoroos and 90% for Bassian thrush. All candidate models are tabled in the Appendices (Appendix Tables A2.1–A2.3). Thus, there is little support for the hypothesis that there is an effect of fox baiting on the occupancy rate over time for foxes and bandicoots and potoroos whereas there is strong support for an effect of fox baiting on Bassian thrush occupancy. The graphed data (Figs 2.2–2.4) suggest that there has been an overall increase in the occupancy rate for bandicoots and potoroos and a possible reduction in the occupancy rate for red fox, although some of the error bars are large. Occupancy rates for Bassian thrush at baited sites remained steady at around 44% over the 4 years whereas at unbaited sites Bassian thrush occupancy rate decreases by 40% from 2009 to 2010, and after 2010, the rate remained between 25 and 29%.

Table 2.1. Summary of camera trap deployment times (based on 40 camera traps used per year) in the western area of the Great Otway National Park showing an improvement in the number of functioning camera trap days through time.

Table 2.2. A comparison of the number of sites (n = 40) in which a given species had been detected at least once for a particular year for different types of lure used in the western area of the Great Otway National Park.

Species of key interest for which further analysis was conducted are displayed in bold. Birds are excluded with the exception of the Bassian thrush. The trial ceased in 2011 after which meat lures only were used.

For the east Otway area, the detection of species of key interest during this study was much lower than the west Otways. Bandicoots were only detected in one or two sites in the first 3 years, although the number of sites had increased to six in the fourth year and the red fox was detected in three to six sites each year. There was no detection of long-nosed potoroos or Bassian thrush at all in this area. For these reasons, a detailed analysis of east Otway data is not presented here. The program needs to run longer to ascertain whether adequate data will be gathered and will be subject to review.

Fig. 2.2. Red fox Occupancy rates (± 1 s.e.) in the western study area of the Great Otway National Park, victoria. Fox baiting commenced between the 2009 and 2010 monitoring periods.

Fig. 2.3. Combined bandicoot (long-nosed and southern brown) and long-nosed potoroo occupancy rates (± 1 s.e.) in the western study area of the Great Otway National Park, Victoria. Fox baiting commenced between the 2009 and 2010 monitoring periods.

Mornington Peninsula National Park

A total of 32 native species was detected, including 11 mammals. The