camera trapping for the study
THE RAFFLES BULLETIN OF ZOOLOGY 2013
THE RAFFLES BULLETIN OF ZOOLOGY 2013 Supplement No. 28: 21–42
http://zoobank.org/urn:lsid:zoobank.org:pub:804A6DC9-A92A-41AE-A820-F3DA48614761
Date of Publication: 27 Nov.2013
© National University of Singapore
CAMERA TRAPPING FOR THE STUDY AND CONSERVATION
OF TROPICAL CARNIVORES
Sunarto
Department of Fish & Wildlife Conservation, Virginia Tech, 100 Cheatham Hall, Blacksburg, VA 24061-0321, USA
WWF-Indonesia, Graha Simatupang, Jl. TB. Simatupang No. 38, Tower 2 Unit C, Jakarta 12540, Indonesia
Email: [email protected] (Corresponding author)
Rahel Sollmann
Department of Forestry and Environmental Resources, North Carolina State University, Box 7646, Turner House, Raleigh, NC 27695-8003, USA
Azlan Mohamed
WWF-Malaysia, 49, Jalan SS 23/15, Taman SEA, 47400, Petaling Jaya Selangor, Malaysia
Institute for Tropical Biology and Conservation, Universiti Malaysia Sabah, 88999 Kota Kinabalu, Sabah, Malaysia
Marcella J. Kelly
Department of Fish & Wildlife Conservation, Virginia Tech, 100 Cheatham Hall, Blacksburg, VA 24061-0321, USA
ABSTRACT. — Past studies on tropical carnivores and other secretive animals relied on indirect evidence
of animal presence such as tracks, scats, or scrapes. While such evidence can be useful for basic studies,
using remotely-triggered camera traps offer researchers more reliable evidence of animal presence and, with
appropriate study design and analysis, provides an array of opportunities to investigate carnivore ecology.
We present an overview on camera trap uses for the study and conservation of wildlife, with a particular
focus on tropical carnivores. Our goals are to promote proper and effective application of camera trapping
and related analyses. We highlight major research avenues, give relevant examples and lessons learned from
published material and from our own experiences, and review available resources for implementation, from
review considers sampling design with respect to target species or groups of species, the state variable(s)
of interest, what constitutes a sample, sample size needed, collection of supporting data (independent
variables), reducing bias/minimising error, and data collection schedule. We also highlight some available
camera trap database management packages and available statistical packages to analyse camera trapping
management of species. Finally, we discuss future development of camera trapping technology and related
techniques for the study and conservation of carnivores in the tropics.
KEY WORDS. — camera trapping review, elusive carnivores, photographic sampling, predator, wildlife
research in the tropics
INTRODUCTION
of room for improvement in the study design and analysis,
allowing more thorough investigation of carnivore ecology.
Tropical carnivore ecologists often receive questions
about the number of times they have directly observed the
animals they study in the wild. Typically, the answer to
such questions is “hardly ever” or “never”. Past studies on
tropical carnivores and many other secretive animals relied on
indirect evidence of animal presence such as tracks, scats, or
scrapes, which can be useful for simple distribution mapping.
However, relatively recent techniques using camera traps
offer researchers more reliable evidence of animal presence.
Moreover, standardising effort and sampling protocol is
relatively easy to do in camera trapping; and there is plenty
Tropical rainforest carnivores have characteristics that make
camera traps an ideal study tool. These characteristics include
their body size (medium or large), morphology (natural marks
live (terrestrial—allowing for relatively simple placement
of equipment), behaviour (readily use trails), secretiveness/
Schaick, 1993; Karanth et al., 2004b), rarity (requiring large
21
Sunarto et al.: Camera trapping for the study and conservation of tropical carnivores
challenging for direct observation), and generally remote
locations, which make long-term studies, direct observations,
We limit our interpretation of camera traps to those “remotely
triggered cameras that automatically take images of whatever
walks in front of them” (Rovero et al., 2010: 102). There
are similar systems that are potentially useful to study wild
carnivores in the tropics but they are not the focus of this
review. Examples of such systems are non-triggered camera
traps programmed to periodically record images at certain
Compared to tracks or scat, pictures from camera traps
minimum ambiguity. Once they are set up, camera traps can
accumulate efforts quickly over large areas. Additionally,
camera traps can record information such as the date and
time of the photograph, temperature, and location, either
follow an animal (e.g., http://www.bbc.co.uk/news/scienceenvironment-12070732), surveillance (video) cameras that
are continuously recording events, or cameras attached to
animals to observe surroundings (e.g., http://boingboing.
net/2007/06/06/cat-has-camera-on-co.html). Although we
focus on the use of still images, we also consider video camera
traps that record motion pictures with or without sound.
properties, or noted on a related datasheet. Other supporting
information can also be collected including data related to
cover type), time (climatic parameters such as temperature,
rainfall, humidity, etc.), or survey effort (number of trap
nights, personnel involved, ad hoc/systematic, etc.).
trapping techniques, give examples of the equipment
used to study a variety of animal taxa, and explain some
technical aspects of the most commonly used camera trap
models, including set up. We also discuss preparation of
camera trapping studies, data management and analysis,
and presentation of results. To support the text, in an online
supplement we present some resources that can be useful to
help design and implement effective camera trapping studies
for tropical carnivores.
For studying certain taxa and for certain purposes, camera
trapping is often superior to other survey methods: For
example, in species inventories, camera trapping proved to
collection (Davison et al., 2002; Weckel et al., 2006), track
better data to investigate activity patterns than radio telemetry
(Bridges et al., 2004b) and has the additional advantage of
being non-invasive. The technique is especially popular for
THE EVOLUTION OF CAMERA
TRAP APPLICATIONS
Camera traps were initially developed mainly for aesthetic
Though advantageous in many aspects, camera trapping
cannot be considered a silver bullet to studying carnivore
use and activity patterns of small animals, including mice
and lizards. The wider uses of camera traps as a surveillance
and conditions, there are other techniques that work better.
For example, although camera traps might be used to
obtain estimates of minimum home ranges in individually
the main reason they became commercially available and
why technology rapidly developed. Use of camera traps
in ecological research has boomed since the last decade
(Rovero et al., 2010; O’Connell et al., 2011b) following
the successful combination of camera traps with rigorous
these limitations, camera traps offer many possibilities in
wildlife research.
OBJECTIVES AND SCOPE
traps have become an indispensable tool in many wildlife
studies worldwide ranging from simple documentation of
animal presence to rigorous investigation of animal ecology
based on quantitative, experimental and statistical inference.
Our goals are to promote proper applications of camera
trapping techniques and to increase the effectiveness of the
applications to achieve different objectives in the study and
conservation of tropical carnivores. With this both regional
of the history of camera trap development.
addition to existing texts and books dedicated to camera
trapping (e.g., Rovero et al., 2010; O’Connell et al., 2011a).
Against this background we present relevant examples
and lessons learned from published material and our own
experience, discuss major research avenues and data analysis
procedures, highlight study design and available resources
Today, camera traps are typically used to investigate medium
applied successfully in studies of other groups of animal
mammals (Oliveira-Santos et al., 2008), and predators of
to data management, analysis, and presentation of results.
22
THE RAFFLES BULLETIN OF ZOOLOGY 2013
is used in a variety of environmental conditions ranging
from cold temperate, higher altitude areas (Jackson et al.,
2006) to hot, humid tropical forests (Mohd-Azlan, 2009).
temperature is close to body temperature. On the other hand,
EQUIPMENT
The use of DSLR cameras overcomes several issues
still strongly associated with many digital camera traps
or rain, causing a higher rate of false triggering (Swann et
al., 2011), which is an issue particularly in the wet tropics.
There are a wide variety of camera trap models currently
available, from ready-to-use units to those that require
assemblage/development. As in general photography, camera
consequent higher shutter speed, and an overall higher image
quality (mainly due to the larger size of the image sensor),
among other features, make DSLR superior to point-and-
digital systems. Digital camera traps are superior in many
aspects including instant result viewing, much better data
storage capacity, more extensive metadata that comes with
images, the ability to shoot videos, and wider availability of
allows researchers to modify DSLR-based camera-traps such
as those from Nikon® or Canon® (check www.kenrockwell.
com, for a review) to their particular needs. However, a
DSLR camera is much more expensive, increasing equipment
integrated with communication networks such as cell or
satellite phone, allow researchers to receive images taken by
their camera units instantly on their phone or computer (e.g.,
weight and the extra work needed to weatherproof and to
assemble the necessary components.
). On the other hand,
analog camera systems, at least by the time this manuscript
was being written, have a higher picture quality/resolution,
is needed to run a camera trapping study. We present an
example list of camera trapping equipment in Appendix 1.
especially those exhibiting trap shyness (e.g.,Wegge et al.,
CAMERA TRAPS IN TROPICAL FORESTS
Every camera trapping study requires equipment that matches
the study objectives, conditions of the study area, and the
target animal(s). For tropical habitats, equipment must be
able to withstand high heat and particularly, high humidity.
Adding packets of desiccants inside the camera box helps
protect the unit from extreme moisture, but these need to
be replaced often. Frequent camera checking for cleaning
and maintenance is necessary.
susceptible to blurring. However, with recent technological
development, some infrared cameras (e.g., Reconxy®) can
take good pictures in the dark; and with proper camera
placement and setting, the picture quality can be enhanced,
increasing the prospect for individual identification of
Camera traps also vary in type of triggering mechanism.
Originally, camera traps relied on physical triggering
Habitats where tropical carnivores live are often remote and
boating to reach a village closest to the study area. From
Such mechanisms have some limitations with regard to the
physical characteristics of the animal (such as body weight)
that may cause the trigger to fail. Also, with physical triggers
of travel on foot. Therefore equipment weight is an important
issue but should not override equipment quality. Smaller
well-built cameras, although potentially pricey, may end
to trigger the camera. Currently available camera traps mostly
http://en.wikipedia.org/wiki/
http://www.
) infrared motion detectors.
logistics are extremely expensive. Using cheaper cameras
that perform poorly may also make the ultimate costs soar
and introduce bias into the data. Battery power must also
type with high durability. Rechargeable batteries can be
rechargeable batteries currently do not have nearly the same
components (e.g., the camera, transmitting, and receiving
units) providing more flexibility in camera positioning
or lithium batteries, which allow for less frequent camera
checks. Memory card size, which posed a dilemma until
recently, is no longer an issue as capacity has tremendously
increased while prices have declined. Swann et al. (2011)
provide further information on types and features of camera
traps and factors to consider in selecting the right equipment
for different study conditions.
usually easier, including the use of better quality cameras
only sensitive to objects with a different temperature from
the ambient (with warm-blooded animals being the target)
so that they can fail to record passing animals if ambient
23
Sunarto et al.: Camera trapping for the study and conservation of tropical carnivores
SURVEY DESIGN AND CONDUCT
anthropogenic factors. These variables also can be measured
on a sampling unit (i.e., camera station) or a study area scale.
Study objectives. — Two major factors usually motivate
As a call for caution, the use of these raw descriptive variables
such as a trapping rates as a surrogate for abundance does
not account for the fact that our ability to observe individuals
or species is imperfect, and that the probability of observing
a species (or an individual of a given species) is unlikely
management/conservation. Camera traps can give insight
into aspects of a species’ behavior, such as activity patterns
foraging/hunting, denning (Bridges et al., 2004), or species
Sauer, 1998; Pollock et al., 2002). Failure to account for
imperfect detection can lead to biased results. Analytical
approaches to account for imperfect detection are discussed
in the data analysis section.
2011; Sunarto, 2011; Sollmann et al., 2012).
conservation, camera traps are usually employed to
investigate one of two major issues: population parameters
and parameters related to species occurrence or distribution
Sample unit and size. — The sample unit can vary with the
study objectives. For example, in spatial terms a study using
occupancy models might consider a habitat unit, an area/
grid cell, or camera station as the sample unit, while on the
temporal scale the sampling occasion could be the sample
unit. Meanwhile, studies investigating animal activity might
consider the trap day, each animal record, or each individual
animal as their sample unit. When estimating abundance, the
individuals detected are the sampling unit, but the number
of times they are recaptured also determines whether the
sample size is large or small.
et al. 2002), inventories of carnivore and prey diversity
(O’Brien, 2008; Rovero et al., 2010), studying/mapping
geographic distribution (e.g., Moruzzi et al., 2002a), modeling
occupancy patterns and related habitat use/preferences (Linkie
et al., 2007), population estimation (absolute or relative
including variation of these parameters across geographic
locations or different habitat types (Kelly et al., 2008), and
the investigation of population dynamics (including survival,
Sample size needed depends on factors ranging from the
degree of precision one aims to achieve, to the complexity
of the analysis or the number of independent variables to
use in analytical models. Further, the amount of data an
investigator can actually collect will be limited by resources
available or the nature of the target animal or area. Burnham
al., 2010a]). Depending on the research objective, the study
design, and setup, the data collected will vary. Because
of the space limit, we do not discuss in great detail how
objective. Rather we present important general aspects to
consider when designing camera trap sampling and refer the
reader to more detailed literature.
discussion on this issue.
Data collection. — Measures such as abundance or presence
are also called state variables, because they describe the
state of the studied system. The most basic approach is to
use descriptive or summary variables to approximate state
variables of interest. These include the number of individuals
photographed, the number of (independent) pictures of
the target species, sampling effort, and the photographic
capture rates. These descriptive variables are very useful in
identifying hotspots of high animal activity or in comparing
effort and success across studies. Depending on the research
question, these measures can be determined for the entire
duration of a camera trap study, for sub-periods of time, across
all camera trap units or separately for each unit. The most
STUDY SETUP
a crucial part of the project setup and deserves time and
consideration on several levels:
Characteristics of target species.
to consider in designing camera trap sampling are the
characteristics of the target species. First, consider whether or
(and if so, what are the diagnostic characteristics). This
will not only determine whether investigators can focus
on analytical approaches that require individual-level data
(e.g., capture-recapture models) or species-level data (e.g.,
each camera station, as the data can then be analysed at the
camera’s lens. Even species with little or no apparent natural
study is divided into temporal sub-periods, data can also be
condensed to a binary detection/non-detection format (i.e.,
whether or not a species was detected in a given sub-period
or not). We may be interested in investigating correlations
between these measures with potential explanatory, or
independent, variables. Examples of independent variables
are those related to micro-habitat, macro-habitat, landscape
features, and environmental/climatic, socio-economic and
morphological details of the animals (e.g., bobcats [Lynx
rufus; Heilbrun et al., 2003], maned wolves, [Chrysocyon
brachyurus; Trolle et al., 2007], pumas, [Puma concolor;
Kelly et al., 2008], Javan rhino, [Rhinoceros sondaicus;
Hariyadi et al., 2011]).
24
THE RAFFLES BULLETIN OF ZOOLOGY 2013
Second, the movement range of target species will determine
how far apart traps should be spaced to achieve either
independence of sampling units (a prerequisite for occupancy
modeling) or to make sure all individuals in the sampled
area are exposed to traps (an assumption of capture-recapture
models).
indigenous people who do not want their pictures taken (D.
Priatna/ZSL, pers. comm.). Overall, the study should be
designed so that frequent camera checks are feasible with the
Finally, for species occurring at very low densities, targeting
Data collection schedule. — Factors to consider in the data
collection schedule include survey timing (time of year,
month, seasons), single versus multiple repeated surveys (e.g.,
investigation of trends over time), duration of the survey,
and how often cameras should/can be checked.
species may be the only means of collecting an adequate
number of records (e.g., Karanth, 2011). Under these
circumstances, any knowledge of preferences for certain
landscape structures such as roads or rivers will be helpful
in designing a camera trapping study.
When studying a single species, tailoring the study design
to its characteristics may be relatively straight forward.
However, when a study targets multiple species, for example
in assessing carnivore diversity or interaction among species,
it is important to ensure a study design that is balanced
regarding the different target species. When analytical
approaches do not account for differences in detection
probability among species (see Analytical approaches below),
it is imperative to evaluate or recognise the potential biases
aspects must be assessed on a site-by-site basis. Unfortunately,
Logistically, some areas might be impossible to survey during
season might be the only time areas can be accessed, for
example, if boat transportation requires adequate water
levels or if water supplies for drinking are only available
during the rainy season. Another logistical consideration is
whether local guides are needed and available during the
and extreme care should be taken in inference.
On a smaller scale, exactly where and how to set up camera
traps at the sampling sites must be evaluated at various
levels (for an example, see Fig. 1). Body size of the target
trap model will determine camera height and distance from
Camera trapping studies often have to be conducted when
there is little or no available information on particular
same species from a better studied region can be useful.
Logistics and operational considerations.
the study objectives, researchers must consider operational
conditions including security of cameras from various threats
such as harsh weather/climate, direct sunlight (for cameras
with passive sensors), human vandalism, and disturbance/
damage by organisms (including large and small animals,
insects, plants, and fungi). Different techniques can be applied
using desiccants to absorb moisture, armoring cameras with
extra protection to prevent damage by large animals and
talking directly to local leaders and/or adding persuasive notes
on each camera to prevent vandalism. Regular checking of
traps will help keep the study running smoothly in spite of
camera trap model (battery life, picture storage capacity)
of camera disturbance, etc.) checking cameras every other
week might be ideal as it balances between the need to ensure
cameras are operational and to minimise the disturbance to
be completely unacceptable, such as in areas inhabited by
Fig. 1. Careful selection of the site, height, and angle to set the
camera trap considering the characteristics of target animals is a
key factor in determining the success of a camera trapping study.
The present example shows a two-camera set-up for individual
sized carnivores; b) camera angle that is not directly facing the
other camera at the same station; c) pile of leaves to protect against
splashing mud in heavy rain; d) placement along old logging road
when applicable to increase the probability of detection; and e)
vegetation clearing to ensure clear picture of the animal and avoid
Sunarto et al.: Camera trapping for the study and conservation of tropical carnivores
survey. Anticipating holidays or farming seasons in villages
helps to avoid time loss. Ecologically, investigators should
anticipate that carnivores are likely to exhibit temporal/
seasonal variation in numbers or behavior (Barlow et al.,
Because of the wide range of conditions and dynamics of the
study systems, a technique that works well for one species
might not work for another. Even for the same species, the
ideal setup may vary by habitat type or geographic location.
variation is the parameter of interest. For other objectives,
there might be more flexible options such as sampling
during the season when probability of capturing carnivores
is the highest.
sampling techniques (Kelly, 2008).
Any state variable estimated is only meaningful for the time
frame in which it remains stable, i.e., in which the system
under study is ‘closed’. For example, if there is seasonal
variation in habitat cover, the distribution of a species can
change between seasons; or, in a situation where individuals
may be born or die during sampling (i.e., open population),
abundance would no longer be a meaningful measure.
Therefore, it is important that sampling takes place within a
time frame during which the parameter of interest is unlikely
to change. With rare and elusive species, investigators usually
need to compromise between sampling long enough to collect
enough data, but short enough so that the parameter under
investigation is biologically and ecologically meaningful.
The actual amount of time depends on the biology of target
species and study area. As an example, for big cats, a study
duration of 2–3 months is generally deemed adequate to
Camera trapping investigators likely will spend a large
amount of their time cataloging, managing, and analysing
their photographic data. Therefore, it is important to spend
adequate time thinking in advance about how camera trapping
data should be stored and managed, and to plan extensive
time for data entry.
MANAGING CAMERA TRAP DATA
Camera trap data contain a wide array of information, usually
only part of which is used by the investigator. However, we
advise entering photographic data on all non-target species
including humans as this information can be extremely
valuable to management or can serve as potential variables
to predict the target species’ presence or abundance. Not only
interactions or impacts of human use, but a complete database
will also make later analyses much easier, as researchers will
not have to slog back to the original photographs. Further,
a complete database enhances the ability to compare across
sites or share data and contribute information to other projects
interested in different species.
Silver, 2004).
Reducing bias and error.
the sampling technique across camera stations (and across
repeated samples in space and time), unless variation in
sampling technique is the factor of interest. When complete
standardisation is not possible, the impact of differential
sampling should be assessed and, if possible, minimised.
Other types of information that should be recorded with
a camera survey include: name of study area and its
management status and habitat type, survey name, time and
duration, geographic coordinates of each trap, and type of
cameras and settings used. We also advise recording species,
sex, individual (if possible), age category, number of animals,
time and date of record, camera number, and image number.
Raw data tables may form the base for derived, analysisspecific data tables, for example, individual detection/
non-detection matrices (X-matrix) for capture-recapture
models (see analysis section below). Relational databases,
the use of different camera trap brands/types/accessories,
shutter speed), the use of bait/lure versus non bait (Hegglin
roads (Sollmann et al., 2011), a response to trapping (i.e.,
trap-happiness and trap shyness [Sequin et al., 2003; Wegge et
al., 2004]), or change in detection over time and with season
allowing multiple tables to be queried at once to generate/
an analytical approach that adequately accounts for variable
detection, results of our analyses will exhibit complex bias.
Even when accounting for variation in detection, the more
sources of variation the more complex our model and the
more data we need.
manage camera trapping data.
database using a spreadsheet application such as Microsoft
Placement of camera traps at landscape features that are
preferentially visited by a certain species (e.g., forest roads
spread sheet for entering raw data), or use a database already
developed and made available by others. Examples of existing
to increase the probability of detection. Such a setup targets
a single species, but might not be optimal for other species
in the study area. Additionally one portion of the population
(males for example) may be more likely to be photographed
data include Camera Base, http://www.atrium-biodiversity.
org/tools/camerabase/ (Tobler, 2010), WWF-Malaysia
database by Harris et al. (2010) and Sundaresan et al. (2011).
errors into the data through careless selection of location
to inappropriate camera settings.
26
THE RAFFLES BULLETIN OF ZOOLOGY 2013
DATA ANALYSIS
probability of species occurrence at a site, and parameters
related to changes in occupancy over time, such as the
probability of local extinction or re-colonisation (MacKenzie
et al., 2006).
Due to space limitations we are unable to provide a
comprehensive overview of analytical approaches used with
camera trapping data, but refer readers to the recent book
by O’Connell et al. (2011) for a comprehensive review.
of an area, such as a distinct habitat unit (or fragment),
camera trapping data, organised according to the objective
of the study.
should be spatially independent, meaning they should be
Species inventory. — The most basic information camera
trapping can provide is a list of medium and large sized
mammal species in the study area—a species inventory
present at a single sample unit (for a detailed account on
Hines et al., 2010). The estimated probability of occupancy
refers to the entire sampling unit, and researchers may want
to deploy several camera traps in a single unit to achieve
better spatial coverage. This approach also can be used to
investigate aspects such as habitat associations (Sunarto et
al., 2013) or species interaction (Sunarto, 2011).
2006). Rovero et al. (2010) provide a detailed manual on the
use of camera traps for the inventory of terrestrial vertebrates
and several authors have evaluated and discussed camera
trapping for wildlife inventories (Silveira et al., 2003).
The failure to photograph a species should never simply be
mistaken for a proof of absence of the respective species
Favorable features of occupancy models include: 1) they
fully account for imperfect species detection and varying
detection probabilities among species, sites, time intervals
etc.; and 2) they tolerate missing sampling occasions
without affecting the parameter estimates (Hines, 2006). A
combination of these features and the rapid development of
the method and availability of related infrastructure (e.g.,
software to implement the analysis, guidebook, and expert
support) have facilitated the adoption of the approach in
recent camera trapping studies (e.g., Linkie et al., 2007;
aquatic habits, can be missed completely by camera traps
even with considerable sampling effort. To achieve a more
reliable “absence record”, a researcher can estimate the
amount of effort needed to detect the species at least once
based on a ‘guesstimated’ density (Carbone et al., 2001)
or the detection probability documented in other places or
for other similar species. This has been applied to clouded
leopards in Taiwan and to tigers (Panthera tigris) in South
China (Tilson et al., 2004; Chiang, 2007; Sanderson, 2009).
However, absence records should be interpreted with caution,
since the probability to detect a species likely varies with
models can be implemented in the R package “unmarked”
et al., 2006).
(White, 2009).
Using species accumulation curves (Colwell et al., 2004)
or capture-recapture based approaches (Boulinier et al.,
1998) can aid species inventories by accounting for species
not recorded in camera surveys. Free software tools, such
as Estimates (http://viceroy.eeb.uconn.edu/estimates), can
be used for camera trapping data to investigate issues of
species richness.
Other statistical procedures can be used with camera trap
1989), log-normal (Poisson) regressions, negative binomial
regression or other generalised linear models, to analyse
camera trapping data to reveal resource selection functions,
2004), distribution prediction (Linkie et al., 2006; Karanth
et al., 2009), and species interactions (Weckel et al., 2006;
Species distribution and occupancy. — With extensive
placement of the equipment across a geographic region of
interest, camera traps are very useful to investigate carnivore
distribution (Moruzzi et al., 2002b). For example, Pettorelli
et al. (2010) combined camera trap surveys across 11 sites in
Tanzania, East Africa with ecological niche factor analysis
(ENFA) to reveal distributional and habitat use patterns
for seven elusive carnivores simultaneously. However, this
approach does not explicitly model detectability.
However, some of these examples do not take imperfect
detection into account, which can potentially distort inference
Population abundance and density. — One of the most
common objectives of camera trapping is to estimate the
size and/or density of a population in a given area. Camera
trapping in combination with capture-recapture (CR)
models has proven useful not only for large carnivores with
conspicuous individual marks such as the tiger (Karanth,
With proper sampling design, data from camera trapping are
suitable for analyses using occupancy models (MacKenzie
et al., 2006). This approach has a wide array of applications
to estimate parameters related to species occurrence, such as
the percentage of an area occupied by a species (PAO), the
(Panthera onca ; Silver et al., 2004), leopard (Panthera
pardus ; Balme et al., 2010), and Sunda clouded leopard
(Neofelis diardi ; Wilting et al., 2012), but also for smaller
27
Sunarto et al.: Camera trapping for the study and conservation of tropical carnivores
individual heterogeneity in detection probability; 2) By
treating the trapping grid as embedded in a larger area, they
circumvent the problem of estimating an effective sampled
carnivores like ocelots (Leopardus pardalis
leopard cat (Prionailurus bengalensis; Mohamed et al., 2013)
without immediately conspicuous individual markings such
as pumas (Kelly et al., 2008; Negrões et al., 2010) or maned
wolves (Trolle et al., 2007).
This approach can be implemented using either maximum
Using CR models to estimate abundance from individual
detection/non-detection data has a long history (Otis et
(Efford, 2010) or the equivalent R package secr (Efford,
in combination with camera trapping data began in the
mid-1990s and has greatly enhanced our understanding of
carnivore population status over the last two decades (Karanth
SPACECAP (Singh et al., 2010). These approaches provide
et al., 2010b, Kery et al., 2010). Models can be expanded
to handle open populations, allowing the estimation of
demographic parameters such as survival and recruitment
models, SCR models often result in lower estimates of density
can be implemented in well-established software packages
al., 2012; Sunarto et al., 2013). This is probably a result of
these models more fully accounting for animal movement
off the trapping grid, which is most likely underestimated
by the ad hoc approach to animal movement (BondrupNielsen, 1983).
the size of so-called closed populations, these models can
also handle parameters describing the dynamics of open
populations, such as survival or recruitment rates (Amstrup
Abundance estimation when individuals cannot be
distinguished. — Any type of capture-recapture analysis
requires individual-level data and, thus, that individuals can
be distinguished based on camera trap pictures. Obviously,
this is not possible for a wide range of species, including
many carnivores. There are two alternative model-based
approaches towards this estimation problem: Rowcliffe et al.
(2008) formulated a model under which density is essentially
a function of the trap encounter rate and animal movement
speed and activity. Apart from the fact that reliable estimates
Although CR models provide a statistically sound means
of estimating abundance, deriving a density estimate is
problematic. Animal movement on and off the sampling grid
violates the assumption of geographic population closure
means that the abundance estimate refers to an area that is
larger than the polygon delineated by the outermost traps.
The standard approach is to buffer the trap polygon with half
the mean maximum linear distance moved by individuals
requires a camera trap setup that is random with regard to
animal movement (i.e., targeting landscape structures with
1998) and use this buffered area, the effective sampled
area, to estimate density by dividing abundance by this area.
Other approaches that have been used to estimate buffer
width include the full MMDM, or the radius of an average
for rare species, may result in prohibitively low amounts of
data. The second approach is essentially a variation of an
occupancy model, under which the probability of detecting a
species in a sampling unit is related to the species’ abundance
2006) or on information from the literature (Wallace et
al., 2003). Although some approaches performed well in
these point estimates of abundance into a meaningful estimate
of overall abundance for a study site or translating them into
the chosen buffer width, comparison of estimates from studies
determining the buffer width in different ways becomes
A third approach is the use of abundance indices, usually
some variation of the number of photographs of the focal
species per trap day (O’Brien, 2011). When used to estimate
absolute abundance or density, such an index requires an
independent estimate of density for calibration (Carbone et al.,
2001). The usefulness of this approach has been questioned
as the relation between the index and true density is unlikely
to be constant across sites, species or time (Jennelle et al.,
2002). More often, camera trapping data is used to derive
therefore, have focused on minimising such drawbacks.
A recent analytic development is spatial capture-recapture
(SCR) modeling. These models have two major advantages
over traditional CR models: 1) They make use of the spatial
information of individual captures to model individual
movement and account for differential exposure of individuals
to the trapping grid, thereby addressing a major source of
for example, to investigate relative abundance of prey species
28
THE RAFFLES BULLETIN OF ZOOLOGY 2013
in studies of carnivore feeding ecology (Weckel et al., 2006;
their comparison across time, space or species is extremely
problematic (O’Brien, 2011). Such comparisons are based
on the assumption that detection probability is constant
across these dimensions, which, as previously discussed,
PRESENTATIONS OF RESULTS
BEYOND RESEARCH
to erroneous conclusions about the abundance of species.
Such indices should only be used as a measure of trapping
success or activity rates, not as a measure of abundance,
unless there is strong evidence (which should be stated
explicitly) that the assumption about constant detection
probability is reasonable.
To have an impact and ultimately contribute to conservation
and management, research must be disseminated. Unlike
most of other research techniques that require some analysis
to reveal ecological processes and patterns before the data
become useful, camera trapping has the advantage that the
raw material, i.e., the pictures themselves, can generate
powerful information and are an invaluable tool for public
awareness/advocacy. Photographs provide the public with
Activity patterns and other aspects of behavior. — The time
of day a record was taken provides valuable information
on the activity patterns of species. Different approaches
can be implemented to analyse such data (e.g., van Schaik
any sophisticated statistics or graphs. Thus, presenting onpresentation can be an important contribution to raise
awareness for urgent management issues. From central
Sumatra, for example, awareness of tiger conservation
has been successfully generated through press releases of
camera trap photographs of a three-legged tiger, a victim of
illegal snares set by poachers (e.g., http://www.reuters.com/
), an adult tiger with cubs
(e.g.,
), a bulldozer passing on a
tiger trail (
), and more recently the
possible impact of rapid forest conversion to plantation on
the tiger population (
).
the simplest ways is to present the number or percentage
of pictures for certain time interval in a 24-h period (van
latest and more appropriate approach to analyse activity
data is by considering the time of day as ‘circular’ (Fisher,
1993). Using modeling techniques such as kernel density
estimation (KDR), activity patterns of different animals in
the same study area can be compared to investigate possible
Though not usually applicable for detailed behavioral studies
in carnivores, camera traps can document certain aspects of
Pictures and data from camera traps have also been
extensively used to support efforts to protect important
et al., 2004), handling of prey items, females with offspring,
scent-marking, use of water holes, and even some unusual
behaviors (e.g., http://www.bbcwildlifemagazine.com/
gallery/camera-trap-photo-year-2010-winners; Sanderson
North Sumatra (Sunarto et al., 2004) and Tesso Nilo National
Park in Central Sumatra (Departemen Kehutanan, 2009).
to advocate better protection of habitats harboring rare or
endangered species, as in the case of the rediscovery or
new species records from camera trapping. For example,
camera traps in the Malaysian state of Sabah on Borneo
rediscovered the world’s most threatened otter species—the
hairy-nosed otter (Lutra sumatrana) after over 100 years
Biodiversity monitoring. — With camera trapping it is
fairly easy to standardise sampling to a large degree by
using the same camera trap model, programming and setup
throughout the study area and in repeated samples. This
feature makes camera traps an ideal monitoring tool. Not
surprisingly, therefore, some developments have been made
toward the implementation of camera trapping for biodiversity
monitoring (particularly those employing repeated sampling
over relatively long time period) at various scales based
on certain indicators such as presence and occupancy of
mammals (Ahumada et al., 2011), the composition of the
Particularly for the case of tigers, camera trapping studies
have become the key source of information to determine
the conservation status (Chundawat et al., 2008; Linkie et
al., 2008) and to formulate the conservation strategy both at
(e.g., Soehartono et al., 2007).
et al., 2007), or top trophic level species (O’Brien et al., 2010).
actions such as removal of anthropogenic disturbance
FUTURE DEVELOPMENT
The last few years have seen massive progress in camera trap
development. Design-wise, cameras are becoming smaller in
29
Sunarto et al.: Camera trapping for the study and conservation of tropical carnivores
an effective camera trapping studies for tropical carnivores
including: 1) A list of equipment needed for a camera trapping
study (Appendix 1); 2) Examples of datasheets for camera
trap set-up (Appendix 2) and camera trap checking (Appendix
3); 3) An example of individual identification database
(Appendix 4); 4) An example of data entry spread sheet for
size and lighter in weight, while image quality is increasing;
settings are becoming more flexible, batteries more
makes camera trapping more environmentally friendly.
Simultaneously, camera traps are becoming cheaper. With
such developments, it is likely that camera traps will become
more integrated with other data loggers to record more
detailed biological, climatic, or environmental parameters.
A brief decision guide to study design and data analysis for
common purposes of camera trapping (Appendix 6).
The ability of digital camera traps to capture video sequences
or take sequential pictures will promote further development.
For example, video footage allows estimation of animal
movement speed, which would facilitate the use of the gas
model approach towards estimating density of animals where
individuals cannot be distinguished (Rowcliffe et al., 2008).
Another potential development is the use of three-dimensional
imaging with multi-lenses (Moynihan, 2010). Technically,
it should be possible to develop one camera with multiple
lenses, connected to the camera wirelessly, allowing a single
camera to capture images of an animal from different angles
ACKNOWLEDGEMENTS
We thank Andreas Wilting for inviting us to write this review
paper. Rini Sugiyanti read and provided useful comments
on the early draft of manuscript.
LITERATURE CITED
Hurtado, E. Martin, A. McWilliam, B. Mugerwa, T. O’Brien,
individuals, and/or physiological condition.
Andelman, 2011. Community structure and diversity of tropical
forest mammals: Data from a global camera trap network.
Philosophical Transactions of the Royal Society B: Biological
Sciences, 366: 2703–2711.
Technological advances related to the camera trap sensors
have been widely applied in the gaming world, allowing
the human body to be scanned, movements recorded, and
analysed (e.g., www.xbox.com). Similar technologies in
combination with the existing database and software might
enable future camera traps to, for example, automatically
identify species, individuals, gender; measure body mass,
describe general physiological condition, and characterise
movement. Eventually such new technologies will become
more accessible and economical.
Handbook
of Capture-Recapture Analysis. Princeton University Press,
Princeton, USA. 313 pp.
in a secondary forest in Peninsular Malaysia. Biodiversity and
Conservation, 15
patterns of wild felids in a secondary forest in Peninsular
Malaysia. Oryx, 40: 36–41.
The last few years have also seen the development in cameratrapping related software for data management (Harris et
the impact of non-protected areas in carnivore conservation:
Leopards in the Phinda-Mkhuze Complex, South Africa. Animal
Conservation, 13
(Kelly, 2001, Hiby et al., 2009), and analytical software
et al., 2010; Sundaresan et al., 2011). Simultaneously, new
analytical approaches are constantly being developed or
existing approaches extended—as the relatively recent
appearance of spatially explicit capture-recapture models
demonstrates.
Temporal variation in tiger (Panthera tigris) populations and
its implications for monitoring. Journal of Mammalogy, 90:
472–478.
However, all of these advances will not automatically make
Scavenging behavior in puma. Southwestern Naturalist, 50:
466–471.
Bondrup-Nielsen, S., 1983. Density estimation as a function of
live-trapping grid and home range size. Canadian Journal of
Zoology, 61
trapping will rest largely on the shoulders and in the hands
of the investigators and technicians who use them. We hope
Pollock, 1998. Estimating species richness: The importance of
heterogeneity in species detectability. Ecology, 79: 1018–1028.
in the future.
RESOURCES
2004. American black bear denning behavior: Observations
and applications using remote photography. Wildlife Society
Bulletin, 32: 188–193.
To support the text, in this supplement we present some
resources that can be useful to help design and implement
2002. First record in 68 years of Lowe’s servaline genet. Oryx,
36: 324–324.
30
THE RAFFLES BULLETIN OF ZOOLOGY 2013
kawasan Hutan Produksi Terbatas kelompok hutan Tesso Nilo.
Keputusan Menteri Kehutanan Nomor: Sk. 255/Menhut-Ii/2004.
inexpensive and automatic multiple-exposure photographic
system. Journal of Field Ornithology, 66: 37–43.
Ocelot Density and Home Range in Belize,
Central America: Camera-Trapping and Radio Telemetry.
Model Selection and
Multimodel Inference. 2nd Edition.
488 pp.
State University, Blacksburg, USA. 136 pp.
Leopardus pardalis in
Belize: The impact of trap spacing and distance moved on
density estimates. Oryx, 41: 469–477.
warblers. Journal of Wildlife Management, 67: 600–610.
density: Comparing radio telemetry with camera trapping.
Journal of Zoology, 275: 391–398.
traps. Oryx, 43: 427–430.
Carbone, C., T. Coulson, S. Christie, K. Conforti, J. Seidensticker,
Efford, M., 2004. Density estimation in live-trapping studies.
Oikos, 106
J. Holden, D. Martyr, K. Kawanishi, M. Sunquist, D. W.
Efford, M., 2010. Density 4.4: Spatially Explicit Capture-Recapture.
University of Otago, Dunedin, New Zealand.
The use of photographic rates to estimate densities of tigers
and other cryptic mammals. Animal Conservation, 4
models.
Edition
Austria.
of jaguars (Panthera onca) in the southern Pantanal, Brazil.
Journal of Mammalogy, 91: 722–736.
R Package. 2.0.0.
Statistical Analysis of Circular Data. Cambridge
University Press, Cambridge, UK. 296 pp.
F. J. Manly (eds.), Handbook of Capture-Recapture Analysis.
(ed.), R Package. 2.0.0. Edition. R Foundation for Statistical
Borneo. Oryx, 45: 119–124.
survey procedures for detection of forest carnivores. Journal
of Wildlife Management, 62: 1217–1226.
Chiang, P.-J., 2007. Ecology and Conservation of Formosan
Clouded Leopard, Its Prey, and Other Sympatric Carnivores
in Southern Taiwan
by sympatric jaguars and pumas across a gradient of human
disturbance in Belize. Biotropica, 42: 724–731
Chundawat, R. S., B. Habib, U. K. Karanth, K. Kawanishi, J.
Ahmad Khan, T. Lynam, D. Miquelle, P. Nyhus, Sunarto,
Panthera tigris
IUCN Red List of Threatened Species. Version 2013.1 http://
.
Franklin, N., Bastoni, Sriyanto, D. Siswomartono, J. Manansang
Riding the Tiger, Tiger Conservation in Human-Dominated
Landscapes. Cambridge University Press, Cambridge, UK.
Pp. 130–147.
extrapolating, and comparing incidence-based species
accumulation curves. Ecology, 85: 2717–2727.
Spatially explicit inference for open populations: Estimating
demographic parameters from camera-trap studies. Ecology,
91: 3376–3383.
Colwell, R. K., and J. A. Coddington, 1994. Estimating terrestrial
biodiversity through extrapolation. Philosophical Transactions
of the Royal Society of London Series B: Biological Sciences,
345(1311): 101–118.
Curtis, 2010b. Estimating black bear density using DNA data
from hair snares. Journal of Wildlife Management, 74
cats at the northern limit of their range: Activity patterns and
density estimates from camera trapping in Bolivian dry forests.
Studies on Neotropical Fauna and Environment, 41: 169–177.
Randrianantenaina, 2010. An assessment of carnivore relative
abundance and density in the eastern rainforests of Madagascar
using remotely-triggered camera traps. Oryx, 44: 219–222.
in wildlife ecology: A review. Wildlife Society Bulletin, 27:
Comparing Density Analyses and Carnivore
Ecology in Madagascar’s Southeastern Rainforest. Unpublished
existence and habitat use in the Mountain Pine Ridge Forest
Reserve, Belize. Animal Conservation, 2010: 1–10.
the potential biases in carnivore capture–recapture studies
associated with the use of lure and varying density estimation
techniques using photographic-sampling data of the Malagasy
civet. Population Ecology, 54
E. Messenger, 2002. On the origin of faeces: Morphological
versus molecular methods for surveying rare carnivores from
their scats. Journal of Zoology, 257: 141–143.
program for complex Bayesian modelling. Journal of the Royal
Statistical Society, Series D (The Statistician), 43: 169–177.
Departemen Kehutanan, 2009. Keputusan Menteri Kehutanan
31
Sunarto et al.: Camera trapping for the study and conservation of tropical carnivores
Tiger Summit 2010. St. Petersburg, Russia.
and monitoring techniques. WCS Global Carnivore Program.
Wildlife Conservation Society, Toronto, Canada. 49 pp.
Journal of Widlife Management,
K. U. Karanth, 2009. A tiger cannot change its stripes: Using a
three-dimensional model to match images of living tigers and
tiger skins. Biology Letters, 5: 383–386.
45: 273–276.
Genetta servalina archeri and
other endemic subspecies on the island of Unguja, Tanzania.
Small Carnivore Conservation, 29: 1–4.
Hines, J. E., 2006. Program PRESENCE Version 2.4, Patuxent
Widlife Research Center, Laurel, USA.
activity periods of some Amazonian mammals. Studies on
Neotropical Fauna and Environment, 40
trails: Occupancy modeling for cluster sampling. Ecological
Applications, 20
Applied Logistic Regression.
survey carnivore communities in Northeastern North America.
Wildlife Society Bulletin, 34
Hutajulu, M. B., Sunarto, S. Klenzendorf, J. Supriatna, A. Budiman
Field: Armoring the Camtrakker® cameratrap in a tropical
Asian forest. Wildlife Society Bulletin, 33
clouded leopards in Riau, Sumatra. SSC Cat Specialist Group
Conference Proceedings. Oxford University, Oxford, UK.
The Javan Rhino of Ujung Kulon: An
Investigation of its Population and Ecology through CameraTrapping
2006. Estimating snow leopard population abundance using
photography and capture-recapture techniques. Wildlife Society
Bulletin, 34: 772–781.
of photographic rates to estimate densities of tigers and other
cryptic mammals: A comment on misleading conclusions.
Animal Conservation, 5: 119–120.
on the abundance and activity periods of Sumatran rain forest
wildlife. Conservation Biology, 7: 623–626.
L. Christensen, 2009. Patterns and determinants of mammal
Journal of Applied Ecology, 46:
1189–1200.
new tool for the study of elusive rain forest animals. Tropical
Biodiversity, 1
Early Wildlife Photographers.
Panthera tigris populations
from camera-trap data using capture recapture models.
Biological Conservation, 71: 333–338.
of occupancy studies with imperfect detection. Methods in
Ecology and Evolution, 1: 131–139.
2004a. Estimation of tiger densities in the tropical dry forests
distribution modeling in ecology. Ecological Modelling, 135:
147–186.
sampling. Animal Conservation, 7
tiger abundance from camera trap data: Field surveys and
(Panthera tigris) and their prey to removal of anthropogenic
European Journal of
Wildlife Research, 55
Karanth (eds.), Camera Traps in Animal Ecology: Methods
and Analyses
Ecology,
Javan rhinos (Rhinoceros sondaicus) in Ujung Kulon National
Park using the mark recapture method based on video and
Pachyderm, 49: 90–99.
79
Thompson (ed.), Sampling rare or elusive species: Concepts,
designs, and techniques for estimating population parameters.
Automatic storage and analysis of camera trap data. Bulletin
of the Ecological Society of America, 91
wild boar (Sus scrofa) abundance and density using capture–
European Journal
of Wildlife Research, 54: 391–401.
Assessing tiger population dynamics using photographic
capture-recapture sampling. Ecology, 8
in a primary rainforest of Peninsular Malaysia. Biological
Conservation, 120: 329–344.
area: A camera trap study. Journal of Wildlife Management,
68: 1010–1017.
Noninvasive
Survey Methods for Carnivores
Pp. 110–140.
Using automatically triggered cameras to individually identify
bobcats. Wildlife Society Bulletin, 31
32
THE RAFFLES BULLETIN OF ZOOLOGY 2013
regression in habitat-selection studies. Journal of Wildlife
Management, 68: 774–789.
data for wildlife managers? Journal of Wildlife Management,
69: 849–860.
Trap success among camera types and across species, and habitat
Occupancy Estimation
and Modelling: Inferring Patterns and Dynamics of Species
Occurrence
Northeastern Naturalist, 15: 249–262.
Kelly, M., A. Noss, M. Di Bitetti, L. Maffei, R. Arispe, A. Paviolo,
species that are detected imperfectly. Ecology, 86: 1101–1113.
from camera trapping across three study sites: Bolivia,
Argentina, and Belize. Journal of Mammalogy, 89: 408–418.
Journal
Kelly, M. J., 2001. Computer-aided photograph matching in studies
of Applied Ecology, 42
Cheetahs. Journal of Mammalogy, 82: 440–449.
for elusive species. Animal Conservation, 11: 182–184.
(Panthera onca) in Bolivia’s Chaco? Camera trapping in the
Journal of Zoology, 262
capture-recapture samples to estimate temporary emigration and
breeding proportions. Journal of Applied Statistics, 22
survey areas and density estimates for ocelots in the Bolivian
Chaco. Biotropica, 40
temporary emigration using capture-recapture data with
Pollock’s robust design. Ecology, 78
W. P. Erickson, 2002. Resource Selection by Animals: Statistical
Design and Analysis for Field Studies. 2nd Edition. Springer
Netherlands, Netherlands. 221 pp.
Use of spatial capture-recapture modeling and DNA data to
estimate densities of elusive animals. Conservation Biology,
25
model for spatial capture-recapture data: Comment. Ecology,
92
Camera Traps in Animal Ecology: Methods and Analyses.
Conservation and conservation concerns. International Journal
of Environmental Studies, 63
Monitoring mammals in the Caxiuana National Forest, Brazil
– First results from the Tropical Ecology, Assessment and
Monitoring (TEAM) program. Biodiversity Conservation, 16:
Krebs, C., 1989. Ecological Methodology. Harper Collins, New
from count data: Application to the North American Breeding
Bird Survey. Ecological Applications, 8
Mohd-Azlan, J., 2009. The use of camera traps in Malaysian
rainforests. Journal of Tropical Biology and Conservation,
5: 81–86.
Williams, 2006. Assessing the viability of tiger subpopulations
in a fragmented landscape. Journal of Applied Ecology, 43:
Mohamed, A., R. Sollmann, H. Bernard, L. N. Ambu, P. Lagan,
S. Mannan, H. Hofer, A. Wilting, 2013. Density and habitat
use of the leopard cat (Prionailurus bengalensis) in three
commercial forest reserves in Sabah, Malaysian Borneo. Journal
of Mammalogy, 94, 82–89.
tropical rainforests: Sun bears in the Kerinci Seblat region,
Sumatra. Biological Conservation, 137: 20–27.
Panthera tigris corbetti, in a secondary forest
in Peninsular Malaysia.
, 51: 421–427.
in Sumatran rainforests. Journal of Zoology, 284: 224–229.
Li, 2002a. Assessing remotely triggered cameras for surveying
carnivore distribution. Wildlife Society Bulletin, 30: 380–386.
Panthera tigris ssp. sumatrae
IUCN Red List
of Threatened Species. Version 2013.1. http://www.iucnredlist.
.
Li, 2002b. Assessing remotely triggered cameras for surveying
carnivore distribution. Wildlife Society Bulletin, 30: 380–386.
patterns of medium and large-sized mammals of Sonora, Mexico.
Natural Areas Journal, 24
PC World
Negrões, N., P. Sarmento, J. Cruz, C. Eira, E. Revilla, C. Fonseca,
R. Sollmann, N. M. Tôrres, M. M. Furtado, A. T. A. Jácomo
2008. Comparing methods for sampling large- and medium-sized
mammals: Camera traps and track plots. European Journal of
Wildlife Research, 54: 739–744.
Journal of
Wildlife Management, 74
Estimating grizzly bear population size using camera sightings.
Wildlife Society Bulletin, 22: 74–82.
O’Brien, T. 2011. Abundance, density and relative abundance: A
33
Sunarto et al.: Camera trapping for the study and conservation of tropical carnivores
U. Karanth (eds.), Camera Traps in Animal Ecology: Methods
and Analyses
Estimating animal density using camera traps without the
need for individual recognition. Journal of Applied Ecology,
45: 1228–1236.
tigers, hidden prey: Sumatran tiger and prey populations in a
tropical forest landscape. Animal Conservation, 6: 131–139.
repeated presence-absence data or point counts. Ecology, 84:
777–790.
Camera
Traps in Animal Ecology
2009. A hierarchical model for estimating density in camera-trap
studies. Journal of Applied Ecology 46:118–127.
Camera Traps in Animal Ecology
Pp. 1–8.
Hierarchical Modeling and
Inference in Ecology. Academic Press, London, UK. 464 pp.
species richness for large- and medium-sized rainforest
mammals. Animal Conservation, 11: 179–181.
recapture models for estimating density from trapping arrays.
Animal
Camera Traps in Animal Ecology: Methods and Analyses.
Conservation, 13
capture-recapture data. Ecology, 89: 2281–2289.
words: The application of camera trapping to the study of birds.
Bird Conservation International, 18: S144–S162.
Sumatra. Cat News, 50: 6–9.
Camera Traps in Animal Ecology:
Methods and Analyses
Unattended cameras reveal secrets of some of the world’s
wildest places. American Scientist, 93
Activity pattern of Atlantic Forest small arboreal mammals as
revealed by camera traps. Journal of Tropical Ecology, 24:
species number determined from point-count data? A capturerecapture approach. Journal of Field Ornithology, 74
Wariness of coyotes to camera traps relative to social status
and territory boundaries. Canadian Journal of Zoology, 81:
1978. Statistical inference from capture data on closed animal
populations. Wildlife Monographs, 62
runways. Journal of Mammalogy, 40: 169–180.
K. Nayak, 2010. Evaluating capture–recapture population
and density estimation of tigers in a population with known
parameters. Animal Conservation, 13: 94–103.
Durant, 2010. Carnivore biodiversity in Tanzania: Revealing
the distribution patterns of secretive mammals using camera
traps. Animal Conservation, 13: 131–139.
trap, line transect census and track surveys: A comparative
evaluation. Biological Conservation, 114
Statistical inference for capture-recapture experiments. Wildlife
Monographs, 107: 1–97.
Silver, C. S., L. E. T. Ostro, L. K. Marsh, L. Maffei, A. J. Noss, M.
of camera traps for estimating jaguar Panthera onca abundance
and density using capture/recapture analysis. Oryx, 38
monitoring studies: Statistical methods for design and analysis.
Environmetrics, 13
Silver, S., 2004. Assessing Jaguar Abundance Using Remotely
Triggered Cameras
selectively logged forests for tiger Panthera tigris conservation:
A population density estimate in Peninsular Malaysia. Oryx,
43
Simpson, E. H. 1949. Measurement of diversity. Nature, 163: 688.
User’s Guide for Interactive
Program CAPTURE. Colorado State University, Fort Collins,
USA. 30 pp.
‘SPACECAP’: A program to estimate animal abundance and
density using spatially-explicit capture-recapture. Methods in
Ecology and Evolution, 3: 1067–1072.
Soehartono, T., H. T. Wibisono, Sunarto, D. Martyr, H. D. Susilo, T.
activity patterns from camera trap data. Journal of Agricultural,
Biological, and Environmental Statistics, 14: 322–337.
harimau Sumatera (Panthera tigris sumatrae
Kehutanan (ed.), Peraturan Menteri Kehutanan Nomor: P.42/
Menhut-Ii/2007
Manual On Field Recording Techniques
and Protocols for All Taxa Biodiversity Inventories. Belgium
of a jaguar population in the Brazilian Pantanal using cameraradio-telemetry. Biological Conservation, 129: 487–496.
34
THE RAFFLES BULLETIN OF ZOOLOGY 2013
demographics of black bears in coastal North Carolina and
Journal of Wildlife Management, 73: 1031–1039.
movements using spatial capture–recapture models for jaguars
in central Brazil. Biological Conservation, 144: 1017–1024.
Sollmann, R., M. M. Furtado, H. Hofer,A. T. Jácomo, N. M. Tôrres
pantanal using capture-recapture analysis of camera-trapping
data. Journal of Mammalogy, 84: 607–614.
space partitioning between two sympatric large predators,
the jaguar and puma in central Brazil. Mammalian BiologyZeitschrift für Säugetierkunde, 77: 41–46.
other secretive mammals in the northern Pantanal. Mammalia,
69: 409–416.
Sunarto, 2011. Ecology and Restoration of Sumatran Tigers in
Forest and Plantation Landscapes. Unpublished PhD Thesis,
Camera-trap studies of maned wolf density in the Cerrado
and the Pantanal of Brazil. Biodiversity and Conservation,
16: 1197–1204.
2004. Keanekaragaman mamalia dan burung di kawasan
Taman Nasional Batang Gadis, Sumatera Utara. Conservation
rain forest mammals. Biotropica, 28
trapping for jaguar (Panthera onca
Journal of Neotropical Mammal, 10: 133–139.
on the equator: Multi-point abundance estimates of the tiger
Panthera tigris in central Sumatra. Oryx, 47: 211–220.
Panthera onca)
feeding ecology: Distribution of predator and prey through time
and space. Journal of Zoology, 270
and analysis of camera trap data: Alternative approaches
(response to Harris et al. 2010). Bulletin of the Ecological
Society of America, 92
effort and trap shyness on estimates of tiger abundance from
camera trap studies. Animal Conservation, 7
and features of camera traps in ecological studies: A guide
Program Mark. Colorado State University,
Fort Collin, USA.
Karanth, (eds.), Camera Traps in Animal Ecology: Methods
and Analyses
News, 52: 9–10.
decline of wild South China tigers Panthera tigris amoyensis:
Field survey of priority tiger reserves. Oryx, 38: 40–47.
Analysis and
Management of Animal Populations: Modelling, Estimation, and
Decision Making
Tobler, M., 2010. Camera Base.
Conservation Research, San Diego, USA.
large- and medium-sized terrestrial rainforest mammals. Animal
Conservation, 11: 169–178.
species of ungulate in south-eastern Peru. Journal of Tropical
Ecology, 25: 261–270.
probability of detection compromise the use of indices in
aerial surveys of medium-sized mammals? Wildlife Research,
32
Cat
estimators of small mammal population size. Journal of
Mammalogy, 66: 13–21.
Wilting, A., A. Mohamed, L. N. Ambu, P. Lagan, S. Mannan, H.
clouded leopard Neofelis diardi in two commercial forest
reserves in Sabah, Malaysian Borneo. Oryx, 46: 423–426.
photography and tracking plates to monitor oral rabies vaccine
bait contact by raccoons in culverts. Wildlife Society Bulletin,
31: 387–391.
of fruit production cycles on Malayan sun bears and bearded
pigs in lowland tropical forest of Sabah, Malaysian Borneo.
Journal of Tropical Ecology, 21: 627–639.
Sunarto et al.: Camera trapping for the study and conservation of tropical carnivores
APPENDIX 1: EXAMPLE OF LIST OF CAMERA TRAPPING EQUIPMENT
Map
Compass
Radio and/or cell satellite phone
Data sheet
Keys for padlocks (if cameras are locked)
Laminated sheet (or dry erase board) or placard, for writing date, camera, and station number
Permanent marker
Ball point pen or pencil
Dry erase pen
Rag to wipe off dry erase pen
Extra bungee cords or nylon webbing
Extra sign (camera trapping “project sign” if needed)
Alcohol prep pads for cleaning debris from camera O-rings.
Umbrella –if raining
Tape measure (for taking trail measurements)
Machete or other vegetation cutting device (for clearing vegetation around camera)
Pocket-size multi-tool (always useful)
Weapons to protect yourself from dangerous animals (mace, gun, spray etc.).
Spare camera trap for malfunctions
36
37
MT14
MT12
MT08
MT01
MT15
MT04
Nanjak selatan
Nanjak utara
Baserah
169147.123
825503.123
169147.869
9960265.456
9972884.456
9960265.906
Unless the study site encompasses several zones, it
may be better to use UTM coordinates rather than
lat/long. Since UTMs are in meters, it is easy to tell
distances away from cameras or distances between
cameras stations.
23 Mar.2011
23 Mar.2011
23 Mar.2011
GPS Location
Easting (UTM X)
3.5
1
1.6
5.0
4.7
2.0
1.9
1.8
1.75
70
90
90
PA
PA
PA
PF
SF
PF
Notes
It is a good idea to collect categorical
data about each location. These codes can
be very specific to your study.
Habitat
Type
****
Trail type, width, and canopy cover have been shown to be
predictors of detectaiblity and/or habitat use. We would advise
to always collect this data. In addition, distance from camera to
the middle of the trail will make sure that you do not put your
camera so far away. This is to ensure that the nightime flash or
infrared will at least reach the target area. Camera makes all
have different flash/infra ranges.
R
G
T
Road (R),
Distance
Trail (T),
from
New Trail
Camera to Canopy
GPS Location
(NT),
Width of Middle of Cover (%) Land
Northing (UTM Game Trail Road or
Road or at Station Use
Y)
(G)
Trail (m) Trail (m) **
***
** Canopy cover: 0 = 0–10%, 10 = 10–20%, 20 = 20–30%, 30 = 30–40%, 40 = 40–50%, 50 = 50–60%, 60 = 60–70%, 70 = 70–80%, 80 = 80–90%, 90 = 90–100%.
*** Land use: PA protected area, OP oil palm plantation, AC Acacia , R roads, BA built up area.
**** Habitat: PF primary forest, SF secondary forest, BU bushland, GR grassland, FS fresh water swamp, RI riverine, PL plantation
This example uses two cameras
per station. In this case MT stands
for Moultrie brand camera.
TN11St03
TN11St02
TN11St01
Station
Date
(D/M/Y)
Nilo Forest Complex (TNFC)
Camera
#s
Physical Location
SITE: Tesso
APPENDIX 2: EXAMPLE OF A “SET-UP” CAMERA DATA SHEET FOR THE INITIAL DEPLOYMENT OF CAMERA TRAPS IN THE FIELD
THE RAFFLES BULLETIN OF ZOOLOGY 2013
38
SS, MJK, RS
Trigger camera with a placard that
reads at minimum: Station Number,
Camera Number and Date. Separate
treggering for each of the 2 cameras.
13 Apr.2011
ü
ü
MT12
MT14
ü
MT08
Open Camera. Press Off Button for DCs,
REs or MTs. DON'T TURN OFF BEC
DIGITALS
ü
ü
ü
ü
ü
ü
Battery Level % for Digitals or # of Green
Blinks for DCs (9-volts)
80%
25%
95%
75%
78%
85%
N
Y
N
N
N
N
Change Batteries? Yes (Y) NO (N)
Having a check sheet to follow in a
specific order prevents researchers from
forgetting key information or
maintenance procedure such as noting
number of photographs, battery levels,
whether memory cards were changed,
image quality, etc.
21
27
112
3
-
MT01
69
78
13 Apr.2011
13 Apr.2011
ü
Trigger with Station #, Camera #, and Date
on Display Card
ü
SS, MJK, RS
SS, MJK, RS
Today's Date
(d/m/yy)
# Pics Taken
MT15
MT04
Names or
Initials of
Camera
checkers
Noting researchers present is helpful to
address later questions that might arise.
TN11St03
TN11St02
TN11St01
Station Code
or Station #
Camera
Type &
Number
MT =
Moultrie
RE =
Reconyx
DC =
DeerCam
BEC =
Buckeye
-
Ds
-
-
Y
Y
Y
Y
Y
Digitals on Still Picture Mode (S) or Video
Mode (V)
V
S
S
S
S
S
Image Quality? High (H), Med (M), Low
(L)
-
M
H
M
L
M
Event Delay in Minutes
1
0.5
1
1
1
1
-
3
3
3
3
3
# Pictures per Event
ü
ü
ü
ü
ü
ü
Double checking all the settings at each
camera check is important because
malfunctions can cause the time
between events, # photos taken per
trigger, and especially date and time
stamp to become corrupted or incorrect.
Y
Which Batteries Changed? AAs, 9-Volts,
Cs, Ds, 6 Volts
-
Card or Film Swapped Out, or Images
Downloaded to Card? Yes (Y) or No (N)
-
Check Date/Time Stamp on Camera- is it
Correct?
Put a check mark for things that you have checked and a dash if nothing needed
ü
ü
ü
ü
ü
ü
Clean O-Rings (Camera Seal) with Cloth
Site: Tesso Nilo Forest Complex
Set, Lock, and Reposition
ü
ü
ü
ü
ü
ü
ü
ü
ü
ü
ü
ü
Make Sure Camera is On (AUTO for MTs
or Switch for REs)
ü
ü
ü
ü
ü
ü
Without a note to
specifically check the lense,
flash or sensor, dirt can
easily go unnoticed and
image quality or trigger
efficiency will be serously
compromised.
ü
ü
ü
ü
ü
ü
Clean Lens Cover, Flash Cover and Sensor
Cover
Survey Name: TNSyst2011
Trigger with Station #, Camera #, and Date
on Display card
Notes on specific camera issues can be very helpful.
Re-triggering cameras with the placard (Station#, Camera
#, and Date) is particulary important when you have
changed the memory card. This will alleviate the need to
keep track of which memory card is which since a
photographic record is taken.
Changed Ds, then camera read 99%
Camera did not trigger. Battery read good. Replaced with camera MT25
Camera out of position slightly and with some claw marks. Re-arranged and
repositioned to fix.
Very dirty from water splash up - cleaned lenses, o-rings well
Notes - include anything out of the ordinary, damage to cameras by animals,
suspected malfunctions, physical location if you change a camera location
etc.
APPENDIX 3: EXAMPLE OF A CAMERA CHECKING DATA SHEET USED WHEN MONITORING CAMERAS FOR PROPER FUNCTIONING IN THE FIELD
Sunarto et al.: Camera trapping for the study and conservation of tropical carnivores
39
Time
17:10
16:12
x-Location
825507.308
825503.123
y-Location
9972884.597
9972884.456
Place
AD17
MI01
Time should be recorded in 24-hour format and if the study area
is within only one zone, we suggest converting to UTM rather
than lat/long.
*Notes: some illustrative data have been added to clarify
the example illustration
Dates
5 Jun.2008
9 Jun.2008
Leopard Cats From Tesso Nilo, Central Sumatra*
Individual #1
Dates
1 May 2008
1 Jun.2008
1 Sept.2008
Individual #2
x-Location
173602.752
173602.752
173602.123
y-Location
32981.421
32981.421
32981.456
Place
AI20
AI20
MI02
Dates
25 Apr.2008
4 Jul.2008
5 Sept.2008
Time
16:18
2:14
3:15
x-Location
169147.869
169147.869
169147.123
y-Location
9960265.906
9960265.906
9960265.456
Place
AJ23
AJ23
MI03
While name of place or grid code can be used, we suggest using a code that incorporates the survey number
or date. For example these stations could be labelled as TN08St01 for Tesso Nilo survey 2008 camera
station 01.
Time
23:49
13:53
16:15
Individual #3
APPENDIX 4: EXAMPLE OF INDIVIDUAL IDENTIFICATION FOR LEOPARD CATS “CAPTURED” AT REMOTE CAMERA STATIONS
THE RAFFLES BULLETIN OF ZOOLOGY 2013
40
Homo sapiens
Sciurus carolinensis
Homo sapiens
Procyon lotor
Homo sapiens
Procyon lotor
Odocoileus virginianus
Homo sapiens
Ursus americanus
Ursus americanus
Ursus americanus
Tamias stratus
Homo sapiens
Odocoileus virginianus
Canis lupus familiaris
Odocoileus virginianus
Ursus americanus
Human
Gray Squirrel
Human
Raccoon
Human
Raccoon
White-tailed Deer
Unknown
Human
Black Bear
Black Bear
Black Bear
Eastern Chipmunk
Human
White-tailed deer
Dog
White-tailed deer
Black Bear
Use common and scientific name,
followed by a station code. In this
example form, the 7MLBS01 and
7MLBS02 are the stations 01 and 02 of
this survey. The 7 in front of MLBS
refers to the fact that this is the 7th
survey at the same site.
Scientific Name
Odocoileus virginianus
Ursus americanus
Odocoileus virginianus
Odocoileus virginianus
Procyon lotor
Homo sapiens
Homo sapiens
Odocoileus virginianus
Odocoileus virginianus
Homo sapiens
Common Name
White-tailed Deer
Black Bear
White-tailed Deer
White-tailed Deer
Raccoon
Human
Human
White-tailed Deer
White-tailed Deer
Human
1
4
1
1
1
4
1
1
1
3
1
1
1
1
1
1
1
1
14
33
3
1
3
3
1
9
3
42
2
3
3
2
58
3
3
11
# Animals # of
in Photo Photos
1
6
1
3
1
4
2
7
1
3
9
70
1
3
1
1
1
4
2
21
10 Sep.2010
12 Aug.2010
15 Aug.2010
16 Aug.2010
18 Aug.2010
21 Aug.2010
25 Aug.2010
25 Aug.2010
28 Aug.2010
29 Aug.2010
29 Aug.2010
30 Aug.2010
31 Aug.2010
1 Sep.2010
5 Sep.2010
9 Sep.2010
9 Sep.2010
9 Sep.2010
Date (M/D/Y)
11 Jan.2010
11 Jan.2010
11 Feb.2010
11 Mar.2010
11 Mar.2010
11 Mar.2010
11 Mar.2010
11 Jun.2010
11 Jun.2010
11 Jul.2010
9:29
14:02
16:47
8:32
2:18
11:17
4:15
13:29
13:53
13:44
19:21
9:02
19:07
11:16
10:30
8:00
9:40
6:28
Time
10:42
18:38
14:25
1:42
3:28
8:45
17:36
1:45
4:55
12:55
34–39; 28–33, 38–39
1–12;1–21
22–24
22
25–27
10–12
01
16–21;4–6
10–12
22–30,55–66;16–27,1–9
10–11
67–69
13–15
16–17
79–84,1–21; 19–30,1–18
22–24
25; 19–20
28–33; 22–27
Image #
121–126
127–129
632;133–135
633;136–141
142–144
634–641;145–192,1–15
642–643;16
645
646;22–24
647–655;25–40
Perhaps the trickiest part is determining the independent "events" or independent animal
captures. Most camera studies use 30 or 60 mins as the standard, albeit somewhat arbitrary, cutoff for considering a capture of a new event. Each row denotes an event, however there can be
more than one event in the photo such as when 2 deer are photographed in 1 picture (4th record
above). The total number of photos can be added together from both cameras to keep track of
total numbers of photos sifted through to obtain the data. Once in this type of spreadsheet, it is
relatively simple to use Pivot Table in Excel to sort the total numbers of events and photos by
animal or camera station.
7MLBS02
7MLBS02
7MLBS02
7MLBS02
7MLBS02
7MLBS02
7MLBS02
7MLBS02
7MLBS02
7MLBS02
7MLBS02
7MLBS02
7MLBS02
7MLBS02
7MLBS02
7MLBS02
7MLBS02
7MLBS02
Station #
7MLBS01
7MLBS01
7MLBS01
7MLBS01
7MLBS01
7MLBS01
7MLBS01
7MLBS01
7MLBS01
7MLBS01
Bear moved
cam
hunting dog
change card
Buck
change card
doe and fawn
Notes
foot
Hunter
foot
Tourist
foot
foot
Unknown
Researchers
foot
foot
Researcher
Researchers
foot
vehicle
Researcher
Unknown
Human Type Vehicle/Foot
It is good to note the camera numbers and especially
the image number to later relocate a specific photo
or group of photos. In this example, if two cameras
are noted, then both cameras fired. Whereas if one
camera is noted, only 1 of the 2 camera fired. All
images before the semi-colon came from the first
camera while all images after the semi-colon came
fro the second camera. BEC, MT, and RE refers to
Buckeye, Moultrie, and Reconyx brand cameras
MT215;RE07
MT215;RE07
RE07
MT215
RE07
MT215
RE07
MT215;RE07
R E 07
MT215;RE07
RE07
MT215
RE07
RE07
MT215;RE07
MT215
MT215;RE07
MT215;RE07
Cam #(s)
RX01
RX01
BEC17;RX01
BEC17;RX01
BEC17;RX01
BEC17;RX01
BEC17;RX01
BEC17
BEC17;RX01
BEC17;RX01
We suggest entering all data on all species including non-target species and humans as this information can be important in predicting target species presence or abundance.
APPENDIX 5: EXAMPLE OF DATA ENTRY SPREAD SHEET FOR RAW PHOTO DATA WITH TWO CAMERAS PER STATION
Sunarto et al.: Camera trapping for the study and conservation of tropical carnivores
THE RAFFLES BULLETIN OF ZOOLOGY 2013
APPENDIX 6: BRIEF GUIDE TO STUDY DESIGN AND DATA ANALYSIS FOR COMMON PURPOSES OF CAMERA TRAPPING
This brief guide will provide basic assistance in designing a camera trapping study and picking adequate analytical techniques depending
using camera traps, or approaches one might adopt for data analysis. Neither is this intended as a manual for camera trap study design
and data analysis. Rather, our aim was to cover the more frequently used approaches, touch upon some basic issues a researcher has to
Adequately planning a study is extremely important and this brief guide is intended to get one started.
STARTING POINT: SINGLE SPECIES (1) or MULTI SPECIES (2)
1. SINGLE SPECIES FOCUS
Question: What aspect of research are you interested in?
1.1 Presence – Question: Is my focal species present in the study area?
Option: Use method by Tilson et al. (2004) to estimate minimum effort to determine absence of your target species with more
reliability. No particular setup requirements except to increase the probability of detecting the target animal.
1.2 Distribution/habitat use/occupancy –
occurrence of my target species?
Otherwise go to 1.2.2
1.3 Abundance/density – Question: How abundant is my target species?
Otherwise go to 1.3.2
1.4 Activity pattern – Question: What time of the day is my target species active/inactive?
Option 1
Option 2
No particular setup requirements for either option.
1.5 Population dynamics – Question: How does my study population change over time; what are the demographic rates?
population model of population dynamics, see, for example, Pollock et al. (1990) and Karanth et al. (2006).
1.6 Occupancy dynamics – Question: How do patterns of occupancy change over time; what are the rates of local extinction
and colonisation?
et al. (2006), Chapter 7.
Option 1
trap point locations) use spatial capture-recapture models and make density a function of your habitat variable. For details,
see 1.3.1, Option 2.
All methods under 1.2.2 are also applicable.
1.2.2 Spatial distribution of species (no individual ID)
Option 1: Occupancy models (MacKenzie et al., 2006). Sample units need to be spatially independent, need to cover all
relevant environmental conditions, and refer to areas, NOT points.
Option 2: Regression on the number of photographs (log-linear, e.g., Foster et al., 2010) or detection/non-detection (logistic,
environmental conditions. However, regressions do not account for imperfect detection, and we suggest using occupancy
models.
Data requirements for both options: A fair number of sampling units; explanatory variables for all sampling units.
1.3.1 Abundance of individually distinguishable species
Option 1
produces abundance estimates, density must be estimated using an ad hoc approach to estimating animal movement.
Data requirements: Several individuals, several recaptures.
Option 2: Spatial capture-recapture (SCR) models (e.g., Royle et al., 2009, Efford, 2011). Camera trap array should cover
the extent of animal movement (but not necessarily several times, Marques et al., 2011); can contain some ‘holes’ but
overall spacing should be smaller than animal movement.
Data requirements: Several individuals, several spatially spread out recaptures, potentially individual and spatial explanatory
variables.
41
Sunarto et al.: Camera trapping for the study and conservation of tropical carnivores
1.3.2 Abundance of non-distinguishable species
Option 1
both in terms of logistics and in terms of the amount of data this will render) and animal movement speed needs to be
known or estimated (telemetry data, video traps).
Option 2
cannot easily be translated into a study area wide abundance estimate.
other methodologies.
2
MULTIPLE SPECIES FOCUS
Question: What aspect of research are you interested in?
2.1 Species inventory/richness – Question: What is the composition of the mammal assemblage at my study site? How rich is
the mammal assemblage?
Option 1: Make a list of detected species; sample study area with a balanced design (no particular species targeting, although
you can target certain focal groups such as terrestrial carnivores) and across different environmental conditions to increase the
chance for different species being detected; ideally, complement list with other methodologies (spotlighting); remember that
failure to detect is no proof of absence (see also 1.1).
Option 2: Estimate the number of species in the study area accounting for the fact that you likely missed some, i.e., with species
Option 3: Estimate species diversity; typical measures are Simpson’s diversity index (Simpson, 1949) and the Shannon’s diversity
index (Krebs, 1989); note that these require some measure of relative abundance and are therefore most likely not suitable for
camera trapping data, unless abundance can be determined with sound analytical methods (CR or SCR models, calibrated indices).
2.2 Spatial relationship between species – Question: Do species co-occur or avoid each other spatially?
Option 1: Multi-species occupancy models (e.g., Sunarto, 2011); for setup and data requirements see 1.2.1; see MacKenzie et
al. (2006) Chapter 8 and 9 for the formal extension of single species to multi species and community-level occupancy models.
Option 2: Regression analysis where data from one species is explanatory variable for the other species (Davis et al., 2010);
remember that regressions do not account for imperfect detection or differences in detection between species.
2.3 Temporal relationship between species – Question: How similar are the activity patterns of species?
Option
42
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement