This book provides a much-needed guide to the sound use of camera trapping for the most common ecological applications to wildlife research. Each phase involved in the use of camera trapping is covered:
- Selecting the right camera type
- Set-up and field deployment of your camera trap
- Defining the sampling design: presence/absence, species inventory, abundance; occupancy at species level; capture-mark-recapture for density estimation; behavioural studies; community-level analysis
- Data storage, management and analysis for your research topic, with illustrative examples for using R and Excel
- Using camera trapping for monitoring, conservation and public engagement.
Each chapter in this edited volume is essential reading for students, scientists, ecologists, educators and professionals involved in wildlife research or management.
|Product dimensions:||6.60(w) x 9.50(h) x 0.70(d)|
About the Author
Fridolin Zimmermann is a carnivore conservation scientist with a PhD on Eurasian lynx conservation and ecology. He is currently coordinator of the large carnivore monitoring in Switzerland at Carnivore Ecology and Wildlife Management (KORA).
Collectively they have nearly 30 years of professional experience in the use of camera trapping for wildlife research, and have worked on a range of species, habitat and study types.
Read an Excerpt
Francesco Rovero and Fridolin Zimmermann
Camera trapping is the use of remotely triggered cameras that automatically take images and/or videos of animals or other subjects passing in front of them. This technology is changing rapidly, largely driven by market demands in the northern hemisphere, with a large proportion of the buyers being recreational hunters. The majority of commercially available camera trap models are passive infrared digital cameras triggered by an infrared sensor detecting a differential in heat and motion between the background temperature and a moving subject, such as animals, people, or even a vehicle, passing in front of them (see Chapter 2). Camera trapping as a scientific tool is widely used across the globe especially to study medium-to-large terrestrial mammals and birds, but is increasingly being also applied to other faunal targets, for example arboreal mammals (e.g. Goldingay et al. 2011), semi-aquatic mammals (e.g. river otter Lontra canadensis; Stevens et al. 2004), small mammals (e.g. Oliveira-Santos et al. 2008) and herpetofauna (e.g. Pagnucco et al. 2011). Moreover, a new type of underwater camera trap was recently designed (Williams et al. 2014) using stereo-cameras, which greatly increase the amount of quantitative data that can be extracted from images (i.e. fish size, position and orientation). The first underwater tests have successfully illustrated the potential of this technology to reveal new insights into marine organisms.
Over the last 15 years, and in particular since 2006, there has been an exponential increase in the number of published scientific studies that used camera trapping. The number of publications per year that used camera trapping increased from less than 50 during 1993–2003 to more than 200 during 2004–2014 with a relative peak of 234 in 2012 (Figure 1.1). This vast and impressive increment in the use of this tool has been accompanied by the widening of wildlife research applications, from basic faunal inventories to focal species studies, from behavioural studies to advanced, inferential studies in numerical ecology (Rovero et al. 2010; O'Connell et al. 2011; Meek et al. 2012; McCallum 2013; reviews in Rovero et al. 2013; Royle et al. 2013a).
1.1 A brief history of camera trapping
We briefly review the key steps in the advent of camera trapping since its first applications; more detailed accounts of the history of camera trapping can be found elsewhere (Sanderson and Trolle 2005; Kucera and Barrett 2011).
Camera trapping was invented in the late 1890s by George Shiras III, a lawyer and passionate naturalist who perfected a way of photographing wildlife at night with a large-format camera and a hand-operated flash. Shiras soon gained considerable acclaim for his stunning night photographs of deer and other animals (Sanderson and Trolle 2005). The first camera trap photos were taken when Shiras set up his camera so that he could take a picture remotely by pulling on a long trip-wire. Eventually, he arranged the trip-wire so that an animal triggered the camera, hence taking its own pictures. His articles in the National Geographic Magazine from 1906 to 1921 created considerable interest in wildlife photography (Shiras 1913). Subsequently, in the late 1920s, Shiras taught Frank M. Chapman (then Curator of Ornithology at the American Museum of Natural History in New York) how to use camera traps for his research work in the tropical rainforest of Barro Colorado Island in Panama. Chapman used Shiras's camera traps to capture images of the diverse and, at that time, poorly known fauna, including tapirs (Tapirus bairdii), ocelots (Leopardus pardalis) and pumas (Puma concolor). For many years, Chapman was one of the few researchers to use camera traps.
Several decades passed before researchers rediscovered camera traps as a tool, and Seydack (1984) was probably the first to use automatic camera traps to study rainforest mammals. He collected data for inventorying species as well as to estimate bushbuck abundance and identify individual leopards in Africa. Griffiths and van Schaik (1993) used camera traps to study rainforest mammals in Indonesia, and realised the potential of this method to detect species' presence and to study the behaviour, activity patterns and abundance of elusive mammals (Griffiths and van Schaik 1993; van Schaik and Griffiths 1996). Meanwhile, Karanth begun to use camera traps to identify individual tigers in Nagarahole National Park, India (Karanth 1995). His success with applying capture–recapture models to estimate population density from camera trap data (Karanth and Nichols 1998) led the way for camera trapping coupled with inferential statistics to become a powerful methodology for wildlife research.
Hunters, especially in the USA, began using camera traps in the late 1980s to search for trophy deer and other big-game species. This created a small industry resulting in an increasing range of camera trap models spanning a range of prices. At the same time, technology advanced quickly and modern camera traps soon became relatively small, waterproof plastic enclosures integrating all essential parts into one system (infrared sensor, digital camera, and flash).
1.2 Efficiency of camera trapping and advantages over other wildlife detection methods
Camera trapping is considered a non-invasive method that generally causes a minimum of disturbance to the study animals. While the presence of camera traps, the noise in the ultra-infrasonic range emitted by some models (Rovero et al. 2013; Meek et al. 2014), the smell signature of humans on the unit (Muñoz et al. 2014) and the flash (see below and Chapter 2 for details) could potentially modify the behaviour of passing animals, these potential sources of disturbance are clearly not comparable to those from faunal detection methods that require trapping and handling of animals. The majority of camera models and study types deploy LED flashes, which produce a red glow that is more or less visible to animals depending on the camera model (but see Chapter 2 for details); xenon white flashes, in contrast, produce an instantaneous and intense white light. The potential disturbance to animals of these types of flashes is discussed in the specific study designs (Chapters 5–9).
Camera traps work day and night and can be left unattended in the field for several weeks and even months. Such automatism not only allows for intensive and prolonged data collection over large and potentially remote areas, but makes the traps suited to study animals that are rare, elusive, and only live in remote areas. Camera trapping has also proved more efficient at detecting diurnal species compared to line transects as sighting rates with the latter may be too low, making robust assessments difficult and/or not cost efficient (e.g. Rovero and Marshall 2009).
The vital advantage of camera trapping in comparison to indirect methods used to record the presence of medium-sized to large terrestrial mammals (e.g. dung and track counts) is that photographs provide objective records ('hard fact'), or evidence, of an animal's presence and enable identification of the species. In addition, camera trapping provides information on activity pattern (from the day and time imprinted in the image) and species coexistence (Monterroso et al. 2014), on behaviour, and on the pelage characteristics that in turn can enable individual identification (see Chapter 7).
Taken together, these aspects make camera trapping a cost-efficient method for faunal detection in spite of the initial capital investment needed to purchase the equipment (e.g. Silveira et al. 2003; Rovero and Marshall 2009; De Bondi et al. 2010). Importantly, moreover, they make camera trapping relatively easy to deploy, and hence highly suitable to standardisation, as shown, for example, by the Tropical Ecology, Assessment and Monitoring (TEAM) network (http://www.teamnetwork.org), which implements a protocol of intensive sampling of terrestrial vertebrates simultaneously in (currently) 17 sites across the tropics.
The high efficiency of camera trapping to inventory communities of medium to large, predominantly terrestrial mammals and birds has been shown by a number of studies. For example, camera trapping involving a survey effort of 1,035–3,400 camera trap days detected 57–86% of the total number of species known to exist in the respective community of tropical forest mammals (review in Rovero et al. 2010; Rovero et al. 2014). An assessment of carnivores in the Udzungwa Mountains of Tanzania detected 15 species through camera trapping while only 9 were detected through other methods (observations, road killings, scats, and other signs; De Luca and Mpunga 2005). In a grassland area in central Brazil, Silveira et al. (2003) found camera trapping to be the most appropriate method for mammal inventory in all environmental conditions, allowing for rapid assessment of the community and its conservation status.
Camera trapping has also been instrumental in discovering new species, such as the giant sengi, or elephant-shrew Rhynchocyon udzungwensis discovered in Tanzania in 2005 (Rovero and Rathbun 2006) or the hairy-nosed otter Lutra sumatrana from Sabah, that had been deemed extinct. Similarly, camera trapping continues to reveal new range records of elusive species (e.g. Jackson's mongoose Bdeogale jacksoni and Abbott's duiker Cephalophus spadix in montane forests of Tanzania: Rovero et al. 2005; De Luca and Rovero 2006) and documents the expansion of species into new areas (e.g. Amur leopard Panthera pardus orientalis in China, golden jackal Canis aureus in Switzerland; Figure 1.2).
Several recent studies have compared the efficiency of capture–recapture studies to estimate the density of naturally marked animals (see Chapter 7) based on camera trapping versus genetic identification of scats. Hence, Janecka et al. (2011) conducted snow leopard (Panthera uncia) surveys in the Gobi Desert of Mongolia and showed that the abundance estimated from noninvasive genetic surveys was inflated because of methodological (i.e. inability to age and thus exclude juveniles, inadequate sampling strategy, the persistence of scat in cold, dry environments making it difficult to differentiate older from recent scats) and biological reasons (i.e. deposition of scats by multiple snow leopards on common sites). Similarly Anile et al. (2014) studied the wildcat (Felis silvestris silvestris) in Italy and found that spatially explicit capture–recapture analyses based on scat collection gave the highest and also less precise density estimates because of the lower number of captures and recaptures compared to camera trapping.
The efficiency of camera trapping for estimating the density and abundance of naturally marked and territorial animals through capture–recapture analysis is such that this approach replaces telemetry, one of the traditional methods used to estimate density through determining home range size (e.g. Karanth and Nichols 1998; Zimmermann et al. 2013; Anile et al. 2014). Telemetry requires trapping of animals for collaring, and following a representative proportion of the population is highly time consuming; therefore telemetry, unlike camera trapping-based density estimation, does not enable a rapid assessment of the population size, which might in fact change over the course of capture campaign.
Even though attempts have been made to use camera trapping to estimate home range size (e.g. Gil-Sánchez et al. 2011), bio-logging (i.e. the use of miniaturised animal-attached tags for logging and/or relaying of data about an animal's movements, behaviour, physiology and/or environment; Rutz and Hays 2009), which includes telemetry, is the method of choice for studying an animal's movements, space usage and resource selection, and can potentially provide a range of additional information about an individual's biology (e.g. activity pattern).
There is increasing consensus and early evidence that the integration of camera trapping with methods such as bio-logging (Royle et al. 2013b; Sollmann et al. 2013a,b), genetic sampling (Sollmann et al. 2013c) or presence signs (Blanc et al. 2014) can greatly enhance the informative power of studies. For example Blanc et al. (2013) used data from a Eurasian lynx (Lynx lynx) population in France monitored by means of camera trapping and presence signs and showed that abundance estimates were more precise when capture–recapture and occupancy analyses were combined.
Camera features related to specific ecological applications
Francesco Rovero and Fridolin Zimmermann
Camera trap functioning is complex and has changed vastly from its early origins (Shiras 1906; Shiras 1913; Guiler 1985) to current-day models. The first commercially available camera traps in the 1980s were xenon white flash systems connected via circuitry to a separate camera, which was often an off-the-shelf camera wired to respond to a break in an infrared beam (active infrared (AIR), see below). Over the last 20 years, technological advances have led to sophisticated units comprising a self-contained package that includes sensors and camera.
The range of camera trap brands and models currently on the market is vast, with new functions being introduced each year. Camera brands and models can vary greatly in features and specifications (Cutler and Swann 1999; Swann et al. 2011), however they have consistent features and components to function, the main ones being shown in Figure 2.1. In this chapter, we (1) briefly describe the camera trap components and camera systems, (2) describe the key technological features to be evaluated when choosing a camera trap system, and (3) outline the optimal interactions of these features in relation to the study designs described in section 2.4 and Chapters 5–9.
2.2 Camera trap systems
The majority of modern-day camera traps rely on a passive infrared sensor (PIR, also called a 'pyroelectric sensor') to detect a differential in heat and motion between a subject and the background temperature, and on an infrared/LED flash array to illuminate the target area. All animals have a heat signature in the infrared spectrum and the PIR detects this difference and triggers the camera (Meek et al. 2012). For a comprehensive glossary of technical terms we refer the reader to Meek et al. (2014a).
One limitation of PIR sensors is the way they detect differences between the target animal and the background temperature. The optimum condition for camera trapping is where the temperature differential between the target and the background is greater than 2.7°C (see Meek et al. 2012). PIR camera traps can therefore prove unreliable when the ambient temperature falls within the body temperature range of most mammals (31.5–36.5°C, with recorded peaks of 42.5°C). A second limitation of PIR sensors is that they can be triggered by the movement of pockets of hot air or by the motion of vegetation in the detection zone. This problem can be limited by avoiding pointing the camera directly at a background that is directly exposed to solar radiation. Another limitation of the PIR systems is that the field of view of the camera lens is rarely, with the exception of Reconyx cameras, equal to the detection zone of the PIR as will be explained in section 2.3.
Based on both sensor and flash technology, there are three main categories of camera traps, the first two being the most common ones:
1. PIR with infrared flash: the majority of camera traps on the market today use PIR sensors, despite the aforementioned shortcomings, coupled to an infrared LED flash, and take monochrome images at night.
2. PIR with white flash: two types of white flash cameras are available on the market: xenon and white LED. Xenon gas based flash systems were amongst some of the earliest of camera trap designs but faded into the background once infrared technology started to be used. In recent times, however, the demand for white flash camera traps has been reignited and, in 2012, Reconyx and Scoutguard models (see Appendix 2.1 for websites of camera trap producers) equipped with white LED flashes were released onto the market. It should be noted that the performance of white LED flashes does not match that of xenon flashes when optimal picture clarity is required (see section 2.2).
Excerpted from "Camera Trapping for Wildlife Research"
Copyright © 2016 Francesco Rovero and Fridolin Zimmermann.
Excerpted by permission of Pelagic Publishing.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.
Table of Contents1. Introduction
2. Camera features related to specific ecological applications
3. Field deployment of camera traps
4. Camera trap data management and interoperability
5. Presence/absence and species inventory
6. Species-level occupancy analysis
7. Capture–recapture methods for density estimation
8. Behavioural studies
9. Community-level occupancy analysis
10. Camera trapping as a monitoring tool at national and global levels
11. Camera traps and public engagement