Article

Improving the random encounter model method to estimate carnivore densities using data generated by conventional camera-trap design

Authors:
  • IBiCo - Instituto de Biología de la Conservación, Madrid, España
  • Instituto de Biología de la Conservación - IBiCo
Article

Improving the random encounter model method to estimate carnivore densities using data generated by conventional camera-trap design

If you want to read the PDF, try requesting it from the authors.

Abstract

The random encounter model, a method for estimating animal density using camera traps without the need for individual recognition, has been developed over the past decade. A key assumption of this model is that cameras are placed randomly in relation to animal movements, requiring that cameras are not set only at sites thought to have high animal traffic. The aim of this study was to define a correction factor that allows the random encounter model to be applied in photo-trapping surveys in which cameras are placed along tracks to maximize capture probability. Our hypothesis was that applying such a correction factor would compensate for the different rates at which lynxes use tracks and the surrounding area, and should thus improve the estimates obtained with the random encounter model. We tested this using data from a well-known Iberian lynx Lynx pardinus population. Firstly, we estimated Iberian lynx densities using a traditional camera-trapping design followed by spatially explicit capture-recapture analyses. We estimated the differential use rate for tracks vs the surrounding area using data from a lynx equipped with a GPS collar, and subsequently calculated the correction factor. As expected, the random encounter model overestimated densities by %. However, the application of the correction factor improved the estimate and reduced the error to %. Although there are limitations to the application of the correction factor, the corrected random encounter model shows potential for density estimation of species for which individual identification is not possible.

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the authors.

... whether or not they accounted for variance in all parameters) or reported explicitly that they had not considered precision in some of the measured variables (e.g. Balestrieri et al., 2016;Garrote et al., 2021;Pfeffer et al., 2018). ...
... Looking into the bibliography (Appendix S1), we observed that most of the deficient procedures to estimate day range are those in which tagged animals with GPS collars were used to estimate day range without accounting for tortuosity (e.g. Caravaggi et al., 2016;Garrote et al., 2021;Massei et al., 2018;Rovero & Marshall, 2009;Zero et al., 2013). It is well described that estimate day range assuming straight-line distances between consecutive fixes notably underestimate day range, and some studies concluded that more than 5 fixesÁmin 1 would be required to get tolerably accurate estimates (Marcus Sennhenn-Reulen et al., 2017). ...
... Looking into literature (Appendix S1), we observed that habitual practice is to determine the dimensions of the detection zone by a series of trials in which the camera was approached by a person from varying directions (e.g. Cusack et al., 2015;Garrote et al., 2021;Loonam et al., 2021;Massei et al., 2018;Rowcliffe et al., 2008). In this respect, some studies have evidenced that detection zone is determined by different factors such as environmental conditions and camera trap settings Rowcliffe et al., 2011). ...
Article
Full-text available
Population density estimates are important for wildlife conservation and management. Several camera trapping‐based methods for estimating densities have been developed, one of which, the random encounter model (REM), has been widely applied due to its practical advantages such as no need for species‐specific study design. Nevertheless, most of the studies in which REM has been assessed against referenced methods have sampled one population, precluding evaluation of the circumstances under which REM does or does not perform well. At this point, a review of all REM assessments could be useful to provide an overview of method reliability and highlight the main factors determining REM performance. Here we used a combination of literature review and empirical study to compare the performance of REM with independent methods. We reviewed 34 studies where REM was applied to 45 species, reporting 77 REM‐reference density comparisons; and we also sampled 13 populations (ungulates and lagomorphs) in which we assessed REM performance against independent densities. The results suggested that appropriate procedures to estimate REM parameters (namely day range, detection zone and encounter rate) are mandatory to obtain unbiased densities. Deficient estimates of day range and encounter rate lead to an overestimation of density, while deficient estimates of detection zone conducted to underestimations. Finally, the precision achieved by REM was lower than reference methods, mainly because of the high levels of spatial aggregation observed in natural populations. In this situation, simulation‐based results suggest that c. 60 camera placements should be sampled to achieve acceptable precision (i.e. coefficient of variation below 0.20). The wide range of situations and scenarios included in this study allow us to conclude that REM is a reliable method for estimating wildlife population density when using appropriate estimates of REM parameters and sampling designs. Overall, these results pave the way to wider application of REM for monitoring terrestrial mammals. Several camera trapping‐based methods for estimating wildlife population densities have been developed, one of which, the random encounter model (REM), has been widely applied because of its practical advantages such as no need for species‐specific study design. Here we used a combination of literature review and empirical study to compare the performance of REM with “gold standard” reference methods. We reviewed 77 REM‐reference density comparisons, and we also sampled 13 mammal populations. The results suggested that appropriate procedures to estimate REM parameters (namely day range, detection zone and encounter rate) are mandatory to obtain unbiased densities. Deficient estimates of day range and encounter rate led to overestimation of density, while deficient estimates of detection zone resulted in underestimations. In conclusion, the REM is a reliable method for estimating wildlife population densities when using appropriate estimates of REM parameters and sampling designs. Overall, these results pave the way for wider application of REM for monitoring terrestrial mammals.
... Random encounter model is a reliable method of estimating population density of multiple species using camera traps 228 not they accounted for variance in all parameters) or reported explicitly that they had not considered precision as one of the measured variables (e.g. Balestrieri et al., 2016Garrote et al., 2021). ...
... Many studies did not estimate REM parameters for the target population but considered values from published studies, which could lead to over-or underestimation of parameters and thus densities (e.g. Anile et al., 2014;Caravaggi et al., 2016;Garrote et al., 2021). Similarly, studies that estimated REM parameters based on incomplete data or estimated detection zones by testing on humans would probably result in biased density estimates (e.g. ...
Thesis
Full-text available
A better understanding of population density (i.e. the number of individuals per unit area) is essential for wildlife conservation and management. Despite the fact that a wide variety of methods with which to estimate population density have already been described and broadly used, there are still relevant gaps. In the last few decades, the use of remotely activated cameras (camera traps) has been established as an effective sampling tool when compared with alternative methods. Camera trapping could, therefore, be considered a reliable tool with which to monitor those situations in which classical methods have relevant limitations. It could, for example, be used with species whose behaviour is elusive and which have low detectability (as is the case of most mammals), or populations in which the animals can be identified individually by the spot patterns on their bodies. However, there is lack of information regarding those species for which it is not possible to identify individual animals (i.e. unmarked species). Some authors that have applied camera trapping originally considered relative abundance indexes in order to monitor unmarked populations. These indices were based on encounter rates (i.e. the number of animals detected per sampling unit) observed in camera trapping studies. Methods with which to estimate the population density of unmarked populations were later described, the first of which was the random encounter model (REM). The REM models the random encounters between moving animals and static cameras in order to estimate population density. The REM does this by employing three basic parameters: i) encounter rate, ii) detection zone (area in which the cameras effectively detect animals), and iii) day range (average daily distance travelled by each individual in the population). When this thesis was first started, it was broadly discussed that the application of the REM was limited by the difficulties involved in estimating the parameters required, especially the day range. In this context, the aim of this thesis was to develop and harmonise camera trapping methodologies so as to estimate the population density and movement parameters of unmarked populations, working principally in the REM framework. The first research carried out for this thesis comprised a review of published studies concerning REM, which found that i) wrong practices in the estimation of REM parameters were frequent, and ii) the REM has rarely been compared with reference densities in empirical studies. We, therefore, then went on to evaluate the main factors that affect the probability of detection and the trigger speed of camera traps, which are relevant for encounter rate and detection zone estimation. This is shown in Chapter 1. We subsequently evaluated and described new methodologies that use camera traps to estimate the movement parameters of unmarked populations. We also evaluated the seasonal and spatial variation in these parameters. The information regarding this is provided in Chapter 2. Finally, we assessed the performance of the REM in a wide range of scenarios, and we compared it with other recently described camera trapping methods used to estimate the population density of unmarked species, as detailed in Chapter 3. The results reported in Chapter 1 show that camera trap performance as regards trigger speed and detection probability are highly influenced by different factors, such as the period of the day, the camera trap model, deployment height or sensitivity, among others. We monitored the community of birds and mammals in the study area, and we discovered that a relevant proportion of the animals that entered the theoretical detection zone were not usually recorded. These missed detections introduce bias into the encounter rate, and consequently into density. However, several camera trapping methods with which to estimate effective detection zone have been described, and they should be applied to all the populations monitored. With regard to the day range, we considered the wild boar as a model species and showed that assuming straight-line distances between consecutive locations obtained by telemetry devices underestimates this parameter, while movement behaviours should be accounted when using camera traps to estimate day range, as shown in Chapter 2.1. We then explored the use of camera traps to monitor movement parameters in greater depth, and showed that they are a reliable method. We described a new procedure with which to estimate the day range that accounts for movement behaviour, and for the ratio between fast and slow speeds. The new procedure performed well in the wide range of scenarios that we simulated, and was also tested with populations of mammals around the world. In this respect, we also described a machine learning protocol with which to identify movement behaviour obtained from camera trap records. All of this is described in Chapter 2.2. We subsequently showed that geographical (e.g. altitude), environmental (e.g. habitat fragmentation), biological (e.g. species) and management (e.g. hunting) factors affect the day range, and we reported variable day ranges in ungulates and carnivores across Europe, as shown in Chapter 2.3. We use the combination of a literature review and an empirical study to compare REM densities with those obtained using reference methods. The results showed a strong correspondence between the REM and reference densities, especially when REM parameters are estimated accurately for the target population. We also showed that the precision of the REM is lower than that of the reference methods, and provided further insights into the survey design in order to increase precision. This information is provided in Chapter 3.1. Finally, and as shown in Chapter 3.2, we used ungulates and carnivores as a target in order to compare the REM, random encounter and staying time (REST), and camera trap distance sampling (CT-DS). The REST and CTDS are two recently described methods with which to estimate the population density of unmarked species using camera traps. The results showed that the performance of the three methods is similar in terms of accuracy and precision. We recommend a survey design that will make it possible to apply all the methods, as the final selection of one of them will be mediated by the number of animals recorded and the camera trap performance. In conclusion, the results of this thesis show the usefulness of camera trapping to monitor the movement parameters and population density of wildlife and contributes with a methodological practical step forwards. In summary, the REM approach, which was tuned in this thesis, proved to be a reliable method in a wide range of environmental scenarios. The REM can be firmly established as a reference method to be implemented in multispecies monitoring programmes in the coming years, considering that it does not need to identify individual animals or spatial autocorrelation in captures. However, future developments of the REM in particular, and camera trapping unmarked methods in general, should be focused on optimising surveys designs in order to increase precision. Before this thesis was begun, the main limitations of applying the REM were the estimation of REM parameters, along with its reliability. This has, however, already been dealt with, and the main gap now concerns the low precisions obtained.
... The size of the territory varies depending on the abundance of its main prey, the wild rabbit, with the territories of the males being greater than those of the females (Female: 300-800 ha; male: 600-1200ha) The Iberian lynx plays the role of apex predator of the terrestrial vertebrate community in the Mediterranean ecosystem. The presence of the species n lynx affects the spatial distribution of other mesocarnivores as red fox (Vulpes vulpes), Egyptian mongoose (Herpestes ichneumon), beech marten (Martes foina), wildcat (Felis sylvestris), and common genet (Genetta genetta) (Garrote et al. 2019) Unverifiable observations, a type of anecdotal occurrence data, or tracks and scats detection and species assignment based on morphology are often used to assess the ranges or abundances of carnivores (Al-Johany 2007; Din & Nawaz 2010). However, the use of such data has been widely criticized since misidentification is likely to occur (Garrote & Ayala 2015, Monterroso et al. 2013). ...
... Therefore, REM method could not be applied to study design, commonly used in carnivores, as Iberian lynx, where cameras are placed on tracks to maximize capture probability. Nevertheless, Garrote et al. (2019) developed a correction factor (CF) for expected deviations from REM density estimates using data generated by conventional camera-trap design. The correction factor corrects for the differential use-rate between tracks and the rest of the area made by lynx: ...
... This is because practitioners are required to devote more time and effort to reliably identify individuals which lack distinct markings (Mattioli et al., 2018) or by choosing a survey method that can detect individuals based on other features. Failure to adopt appropriate survey methods increases the chances of individual misidentification (Soller et al., 2020), may inflate or deflate capture rates Garrote et al., 2021), bias sex/age-class structure of the population (Balme et al., 2012), and adversely impact population management and conservation decisions (Balme et al., 2010). ...
Article
Full-text available
Large felids represent some of the most threatened large mammals on Earth, critical for both tourism economies and ecosystem function. Most populations are in a state of decline, and their monitoring and enumeration is therefore critical for conservation. This typically rests on the accurate identification of individuals within their populations. We review the most common and current survey methods used in individual identification studies of large felid ecology (body mass > 25 kg). Remote camera trap photography is the most extensively used method to identify leopards, snow leopards, jaguars, tigers, and cheetahs which feature conspicuous and easily identifiable coat patterning. Direct photographic surveys and genetic sampling are commonly used for species that do not feature easily identifiable coat patterning such as lions. We also discuss the accompanying challenges encountered in several field studies, best practices that can help increase the precision and accuracy of identification and provide generalised ratings for the common survey methods used for individual identification.
Article
Full-text available
Reported strike events between wildlife and aircraft are hazardous to aircraft and airfield operations and are increasing globally. To develop effective mitigation strategies, the relative hazard a species poses to aircraft, as well as information relating to its life history, are key to the development of effective mitigation strategies in Wildlife Hazard Management Plans. However, given the complex nature of airfield environments with access restrictions and the presence of sensitive equipment, the collection of high‐quality ecological data can be difficult. Here we use motion‐activated camera traps to collect activity data on a population of Irish hares (Lepus timidus hibernicus) inhabiting the airfield at Dublin International Airport, to investigate the link between hare activity and aircraft activity in relation to hare strikes. Camera traps revealed that the hare population at the airfield largely displayed a bimodal crepuscular activity pattern, with activity peaking at sunrise and at sunset. Recorded hare strike times at the airfield were closely associated with hare activity times with a high temporal overlap between these datasets. In comparison, hare activity and aircraft movement activity had a moderate overlap across all seasons, with strikes peaking at times with low aircraft movements. We demonstrate the importance of understanding the circadian and seasonal activity patterns of hazardous species at airfields for targeted strike mitigation. The collection of high‐quality ecological data in airfield environments is difficult. Here, we use camera traps to collect activity data on a population of Irish hares inhabiting the airfield at Dublin International Airport to investigate the link between hare activity and aircraft activity in relation to hare‐aircraft collisions (strikes). Using predator‐prey data analytical methods, we investigate the circadian (24‐hours) and seasonal activity patterns of the hares in relation to aircraft movements to identify periods of increased risk and found that strike patterns were more closely associated with hare activity at the airfield rather than aircraft activity. This highlights the importance of understanding the ecology and life histories of the fauna using an airfield environment to inform the development of suitable strike mitigation strategies.
Article
Full-text available
The use of camera traps to estimate population size when animals are not individually recognizable is gaining traction in the ecological literature, because of its applicability in population conservation and management. We estimated population size of synthetic animals with four camera trap sampling-based statistical models that do not rely on individual recognition. Using a realistic model of animal movement to generate synthetic data, we compared the random encounter model, the random encounter and staying time model, the association model and the time-to-event-model and we investigated the impact of violation of assumptions on the population size estimates. While under ideal conditions none of the models provided reliable population estimates, when synthetic animal movements were characterised by differences in speed (due to diverse behaviours such as locomotion, grazing and resting) none of the model provided both unbiased and precise density estimates. The random encounter model and the time-to-event-model provided precise results but tended to overestimate population size, while the random encounter and staying time model was less precise and tended to underestimate population size. Lastly, the association model was unable to provide precise results. We found that each tested model was very sensitive to the method used to estimate the range of the field-of-view of camera traps. Density estimates from both random encounter model and time-to-event-model were also very sensitive to biases in the estimate of animals’ speed. We provide guidelines on how to use these statistical models to get population size estimates that could be useful to wildlife managers and practitioners.
Article
1.Camera trapping plays an important role in wildlife surveys, and provides valuable information for estimation of population density. While mark-recapture techniques can estimate population density for species that can be individually recognized or marked, there are no robust methods to estimate density of species that cannot be individually identified. 2.We developed a new approach to estimate population density based on the simulation of individual movement within the camera grid. Simulated animals followed a correlated random walk with the movement parameters of segment length, angular deflection, movement distance and home-range size derived from empirical movement paths. Movement was simulated under a series of population densities. We used the Random Forest algorithm to determine the population density with the highest likelihood of matching the camera trap data. We developed an R package, cameratrapR, to conduct simulations and estimate population density. 3.Compared with line transect surveys and the random encounter model, cameratrapR provides more reliable estimates of wildlife density with narrower confidence intervals. Functions are provided to visualize movement paths, derive movement parameters, and plot camera trapping results. 4.The package allows researchers to estimate population sizes/densities of animals that cannot be individually identified and cameras are deployed in a grid pattern.
Article
Full-text available
In marginal habitats, populations should reach lower densities, as a consequence of both lower overall abundance and increasing home range size. To test if recently colonized riparian woods crossing intensively cultivated lowlands of NW Italy represent marginal habitats for the forest-specialist pine marten Martes martes, we assessed its population density by a recently developed, camera-trap-based non-invasive method, the Random Encounter Model (REM). As the central assumption of the REM is that animals move randomly with respect to cameras, we suspected that this method may be unsuitable for species with a strong tendency to use linear elements of the territory as usual paths and select habitats, such as woodland, which are likely to be under-represented in fragmented landscapes. To test for the efficacy of the REM, we also applied a faecal DNA-based genetic census to obtain an independent estimate of the minimum number of individuals occurring in the study area.Camera-trapping used 10 camera-traps, deployed for 10 days within a 2 km2 large unit, for a total of 6 units and 12 km2. Pine marten density was estimated at 0.48 (0.36-0.60) ind/km2. All the faecal samples identified by a mDNA-based PCR-RFLP method as pine marten were genotyped at 15 microsatellite loci using a multiplex protocol. We identified 15 different individuals, corresponding to a density ranging between 0.8 and 2.0 in/km2. Using the most conservative genetic estimate, the REM underestimated population density of about 60% proving to be unreliable for estimating pine marten population size. We suggest that this may be the case for elusive species for which the assessment of average daily movements cannot be achieved without the use of invasive methods.
Article
Full-text available
Large carnivores such as jaguars (Panthera onca) are species of conservation concern because they are suffering population declines and are keystone species in their ecosystems. Their large area requirements imply that unprotected and ever-increasing agricultural regions can be important habitats as they allow connectivity and dispersal among core protected areas. Yet information on jaguar densities across unprotected landscapes it is still scarce and crucially needed to assist management and range-wide conservation strategies. Our study provides the first jaguar density estimates of Colombia in agricultural regions which included cattle ranching, the main land use in the country, and oil palm cultivation, an increasing land use across the Neotropics. We used camera trapping across two agricultural landscapes located in the Magdalena River valley and in the Colombian llanos (47–53 stations respectively; >2000 trap nights at both sites) and classic and spatially explicit capture-recapture models with the sex of individuals as a covariate. Density estimates were 2.52±0.46–3.15±1.08 adults/100 km2 in the Magdalena valley, whereas 1.12±0.13–2.19±0.99 adults/100 km2 in the Colombian llanos, depending on analysis used. We suggest that jaguars are able to live across unprotected human-use areas and co-exist with agricultural landscapes including oil-palm plantations if natural areas and riparian habitats persist in the landscape and hunting of both jaguar and prey is limited. In the face of an expanding agriculture across the tropics we recommend land-use planning, adequate incentives, regulations, and good agricultural practices for range-wide jaguar connectivity and survival.
Article
Full-text available
Camera traps are used to estimate densities or abundances using capture-recap- ture and, more recently, random encounter models (REMs). We deploy REMs to describe an invasive-native species replacement process, and to demonstrate their wider application beyond abundance estimation. The Irish hare Lepus timidus hibernicus is a high priority endemic of conservation concern. It is threatened by an expanding population of nonnative, European hares L. euro- paeus, an invasive species of global importance. Camera traps were deployed in thirteen 1 km squares, wherein the ratio of invader to native densities were cor- roborated by night-driven line transect distance sampling throughout the study area of 1652 km2. Spatial patterns of invasive and native densities between the invader’s core and peripheral ranges, and native allopatry, were comparable between methods. Native densities in the peripheral range were comparable to those in native allopatry using REM, or marginally depressed using Distance Sampling. Numbers of the invader were substantially higher than the native in the core range, irrespective of method, with a 5:1 invader-to-native ratio indi- cating species replacement. We also describe a post hoc optimization protocol for REM which will inform subsequent (re-)surveys, allowing survey effort (camera hours) to be reduced by up to 57% without compromising the width of confidence intervals associated with density estimates. This approach will form the basis of a more cost-effective means of surveillance and monitoring for both the endemic and invasive species. The European hare undoubtedly represents a significant threat to the endemic Irish hare.
Article
Full-text available
Camera traps are increasingly important in studies of mammals throughout the world. Typically, cameras are placed along trails or other travel routes. Yet, species composition and photographic rate may differ between cameras set along trails and those set off trails. We tested this idea in eastern Ecuador. Pairs of cameras were placed at 10 locations along narrow (< 1 m wide) trails in lowland forest and approximately 50 m away from trail sites in adjacent forest. Excluding people, there was little difference in total number of records between trail (333 photographs, 506 trap days) and off-trail sites (306 photographs, 509 trap days). Capture rates varied among locations (11 to 148/100 trap days on trails; 19 to 217/100 trap days off trails) and were not correlated between pairs of cameras on and off trails (r = 0.37, P = 0.29). People were only photographed along trails but capture rates of other species on trails were not correlated with numbers of people photographed at the same site (r = -0.10, P > 0.75). Twenty-three species were photographed, including 21 on trails and 22 off trails; Panthera onca was only photographed along trails whereas Tinamus major and Priodontes maximus were only photographed off trails. Species accumulation curves were similar for both sets of cameras; in both cases, curves approached an asymptote after about 200 records. Latency to first detection (LTD) varied from < 1 day (e.g., Mazama americana in trail and off-trail cameras) to 294 days (Procyon cancrivorus in on-trail cameras). Overall, LTD values were correlated between pairs of cameras on and off trails (rs = 0.66, P < 0.01); means did not differ between cameras on and off trails. Species composition varied among trap locations but trail and off-trail cameras did not form distinct groups based on species composition.
Article
Full-text available
Estimating population densities of small mammals (< 100g) has typically been carried out by intensive livetrapping, but this technique may be stressful to animals and the effort required is considerable. Here, we used camera traps to detect small mammal presence and assessed if this provided a feasible alternative to livetrapping for density estimation. During 2010–2012, we used camera trapping in conjunction with mark–recapture livetrapping to estimate the density of northern red-backed voles (Myodes rutilus) and deer mice (Peromyscus maniculatus) in the boreal forest of Yukon, Canada. Densities for these 2 species ranged from 0.29 to 9.21 animals/ha and 0 to 5.90 animals/ha, respectively, over the course of this investigation. We determined if hit window—the length of time used to group consecutive videos together as single detections or “hits”—has an effect on the correlation between hit rate and population density. The relationship between hit rate and density was sensitive to hit window duration for Myodes with R 2 values ranging from 0.45 to 0.59, with a 90-min hit window generating the highest value. This relationship was not sensitive to hit window duration for Peromyscus, with R 2 values for the tested hit windows ranging from 0.81 to 0.84. Our results indicate that camera trapping may be a robust method for estimating density of small rodents in the boreal forest when the appropriate hit window duration is selected and that camera traps may be a useful tool for the study of small mammals in boreal forest habitat.
Article
Full-text available
Camera trapping with capture-recapture analyses has provided estimates of the abundances of elusive species over the last two decades. Closed capture-recapture models (CR) based on the recognition of individuals and incorporating natural heterogeneity in capture probabilities are considered robust tools; however, closure assumption is often questionable and the use of an Mh jackknife estimator may fail in estimations of real abundance when the heterogeneity is high and data is sparse. A novel, spatially explicit capture-recapture (SECR) approach based on the location-specific capture histories of individuals overcomes the limitations of closed models. We applied both methods on a closed population of 16 critically endangered Western Derby elands in the fenced 1,060-ha Fathala reserve, Senegal. We analyzed the data from 30 cameras operating during a 66-day sampling period deployed in two densities in grid and line arrays. We captured and identified all 16 individuals in 962 trap-days. Abundances were estimated in the programs CAPTURE (models M0, Mh and Mh Chao) and R, package secr (basic Null and Finite mixture models), and compared with the true population size. We specified 66 days as a threshold in which SECR provides an accurate estimate in all trapping designs within the 7-times divergent density from 0.004 to 0.028 camera trap/ha. Both SECR models showed uniform tendency to overestimate abundance when sampling lasted shorter with no major differences between their outputs. Unlike the closed models, SECR performed well in the line patterns, which indicates promising potential for linear sampling of properly defined habitats of non-territorial and identifiable herbivores in dense wooded savanna conditions. The CR models provided reliable estimates in the grid and we confirmed the advantage of Mh Chao estimator over Mh jackknife when data appeared sparse. We also demonstrated the pooling of trapping occasions with an increase in the capture probabilities, avoiding violation of results.
Article
Full-text available
The proportion of animals in a population that breeds is an important determinant of population growth rate. Usual estimates of this quantity from field sampling data assume that the probability of appearing in the capture or count statistic is the same for animals that do and do not breed. A similar assumption is required by most existing methods used to test ecologically interesting hypotheses about reproductive costs using field sampling data. However, in many field sampling situations breeding and nonbreeding animals are likely to exhibit different probabilities of being seen or caught. In this paper, we propose the use of multistate capture-recapture models for these estimation and testing problems. This methodology permits a formal test of the hypothesis of equal capture/sighting probabilities for breeding and nonbreeding individuals. Two estimators of breeding proportion (and associated standard errors) are presented, one for the case of equal capture probabilities and one for the case of unequal capture probabilities. The multistate modeling framework also yields formal tests of hypotheses about reproductive costs to future reproduction or survival or both fitness components. The general methodology is illustrated using capture-recapture data on female meadow voles, Microtus pennsylvanicus. Resulting estimates of the proportion of reproductively active females showed strong seasonal variation, as expected, with low breeding proportions in midwinter. We found no evidence of reproductive costs extracted in subsequent survival or reproduction. We believe that this methodological framework has wide application to problems in animal ecology concerning breeding proportions and phenotypic reproductive costs.
Article
Full-text available
We present the results of a camera trap survey conducted in 2008 in the Atlantic Forest of Iguazú National Park, Misiones, Argentina, testing whether placing camera traps on dirt roads/trails or in off-road locations produce important biases in the recorded species. Seven pairs of camera trap stations were active for 26.6 ± 8.9 days; for each pair, one station was located on a narrow unpaved road and the other about 50 m from the road. We used the first order Jackknife estimator to compare species richness between on-road vs. off-road locations. We used records from another camera trap survey conducted at Iguazú National Park in 2006-2007 to assess whether species with a high Road-use Index (ratio of photographs of animals walking along roads to photographs of animals crossing the roads) had a higher ratio of records on roads/off road stations in the 2008 survey. Multivariate ANOVA based on dissimilarities (ADONIS) was used to compare mammal assemblages recorded at stations located on roads vs. off roads. We obtained 228 independent records of 15 species of medium-large sized terrestrial mammals. Stations located on roads had a higher recording rate (1.06, SD = 0.57 vs. 0.24, SD = 0.13 records per day) and recorded more species than off-road stations (15 vs. 10 recorded species; 19.3, SE = 2.8 vs. 14.3, SE = 2.8 species estimated with the 1st order Jackknife model). Species differ in their relative probabilities of being recorded on roads vs. off roads, something that can be predicted with the Road-use Index. The ADONIS indicated that the mammal assemblage surveyed on roads was statistically dissimilar to that surveyed off roads, a result that can be explained by the differential tendency of the species to use roads and trails.
Article
Full-text available
Camera traps are an increasingly popular tool for monitoring medium to large mammals, but the influence of camera trap placement on the detection probabilities of different species has seldom been investigated. In this study we explore the influence of roads on the detection probability of medium to large mammals in three vegetation types in the Little Karoo, an arid biodiversity hotspot. We placed cameras in nine 100 m-long transects, running perpendicular from roads within a conservation area. The camera traps were spaced at*25 m intervals, and were active for an average of 88 days each. Detection probabilities relative to distance from roads showed extensive variation between species and habitat types. There was no clear relationship between distance from the road and the detection probability of most species and guilds, although carnivore detection probability declined significantly as distance from roads increased in all vegetation types. Our results suggest that there is considerable inter-specific variation in detection probability that is significantly influenced by camera trap location relative to roads. Therefore studies that seek to maximise the detection rates of particular species or guilds (e.g. carnivores) by placing cameras on prominent roads and trails are unlikely to provide reliable estimates of the relative abundance of the broader range of sympatric species; a trend observed elsewhere but hitherto untested in arid environments. We recommend that future studies employ a mixed design of cameras located on- and off-roads to provide better estimates of biodiversity in general and predators specifically.
Article
Full-text available
The number of animals in a population is conventionally estimated by capture-recapture without modelling the spatial relationships between animals and detectors. Problems arise with non-spatial estimators when individuals differ in their exposure to traps or the target population is poorly defined. Spatially explicit capture—recapture (SECR) methods devised recently to estimate population density largely avoid these problems. Some applications require estimates of population size rather than density, and population size in a defined area may be obtained as a derived parameter from SECR models. While this use of SECR has potential benefits over conventional capture—recapture, including reduced bias, it is unfamiliar to field biologists and no study has examined the precision and robustness of the estimates. We used simulation to compare the performance of SECR and conventional estimators of population size with respect to bias and confidence interval coverage for several spatial scenarios. Three possible estimators for the sampling variance of realised population size all performed well. The precision of SECR estimates was nearly the same as that of the null-model conventional population estimator. SECR estimates of population size were nearly unbiased (relative bias 0—10%) in all scenarios, including surveys in randomly generated patchy landscapes. Confidence interval coverage was near the nominal level. We used SECR to estimate the population of a species of skink Oligosoma infrapunctatum from pitfall trapping. The estimated number in the area bounded by the outermost traps differed little between a homogeneous density model and models with a quadratic trend in density or a habitat effect on density, despite evidence that the latter models fitted better. Extrapolation of trend models to a larger plot may be misleading. To avoid extrapolation, a large region of interest should be sampled throughout, either with one continuous trapping grid or with clusters of traps dispersed widely according to a probability-based and spatially representative sampling design.
Article
Full-text available
The European wildcat is an elusive felid that is declining across its range. Sicily hosts a distinctive insular wildcat population, the conservation of which requires much better ecological knowledge than is currently available, particularly population density. We simultaneously used two noninvasive methods (camera-trapping and scat-collection) to estimate the population density of wildcats on the Etna volcano. We conducted genetic analyses to identify individuals and to detect potential hybridization with the domestic cat. We analyzed individual capture-histories from camera-trapping and scat-collection using the spatially explicit capture-recapture (SECR) model. Furthermore, we applied the random encounter model (REM), which does not require individual identification, to the camera-trapping data. We identified 14 wildcats from 70 photographic detections (6.48 detections/100 trap-days) obtained from 1080 camera-trapping days over 4 months, and we estimated to have identified all the individuals living in the study area (10.9 km−2). On the contrary, we identified 10 wildcats from 14 out of 39 scats collected from 391 km of transects walked. The estimated densities (individuals km−2 ± se) were 0.32 ± 0.1 (SECR camera-trapping), 1.36 ± 0.73 (SECR scat-collection) and 0.39 ± 0.03 (REM). The population density estimates obtained from SECR camera-trapping and REM overlapped, although we recommend care when applying the latter. The SECR scat-collection gave the highest population density (and less precise) estimates because of the low number of capture and recaptures; however, the population size estimated with this method matched the number of individuals photographed. The population density of the wildcat in Etna falls in the medium-high range of those reported in literature, highlighting the role of this ecosystem for the long-term conservation of the wildcat in Sicily. Camera-trapping is confirmed as a useful tool to assess the wildcat population density and, in this case, was complemented by the genetic analysis that confirmed individual identity.
Article
Full-text available
Methods that accurately estimate animal abundance or density are crucial for wildlife management. Although numerous techniques are available, there have been few comparisons of the precision and cost-effectiveness of different approaches. We assess the precision and cost of three methods for estimating densities of the Endangered Grevy's zebra Equus grevyi. We compare distance sampling and photographic capture–recapture, and a new technique, the random encounter model (REM) that uses camera-trap encounter rates to estimate density. All three methods provide comparable density estimates for Grevy's zebra and are preferable to the common practice of raw counts. Photographic capture–recapture is the most precise and line-transect distance sampling the least precise. Line transects and photographic capture–recapture surveys are cost-effective in the first year and REM is most cost-effective in the long-term. The methods used here for Grevy's zebra may be applied to other rangeland ungulates. We suggest that for single species monitoring programmes in which individuals can be identified, photographic capture–recapture surveys may be the preferred method for estimating wildlife abundances. When encounter rates are low, distance sampling lacks the precision of the other methods but its cost advantage may make it appropriate for long-term or multi-species monitoring programmes. The REM is an efficient and precise method of estimating densities but has high initial equipment costs. We believe REM has the potential to work well for many species but it requires independent estimates of animal movements and group size.
Article
Full-text available
The proliferation of camera-trapping studies has led to a spate of extensions in the known distributions of many wild cat species, not least in Borneo. However, we still do not have a clear picture of the spatial patterns of felid abundance in Southeast Asia, particularly with respect to the large areas of highly-disturbed habitat. An important obstacle to increasing the usefulness of camera trap data is the widespread practice of setting cameras at non-random locations. Non-random deployment interacts with non-random space-use by animals, causing biases in our inferences about relative abundance from detection frequencies alone. This may be a particular problem if surveys do not adequately sample the full range of habitat features present in a study region. Using camera-trapping records and incidental sightings from the Kalabakan Forest Reserve, Sabah, Malaysian Borneo, we aimed to assess the relative abundance of felid species in highly-disturbed forest, as well as investigate felid space-use and the potential for biases resulting from non-random sampling. Although the area has been intensively logged over three decades, it was found to still retain the full complement of Bornean felids, including the bay cat Pardofelis badia, a poorly known Bornean endemic. Camera-trapping using strictly random locations detected four of the five Bornean felid species and revealed inter- and intra-specific differences in space-use. We compare our results with an extensive dataset of >1,200 felid records from previous camera-trapping studies and show that the relative abundance of the bay cat, in particular, may have previously been underestimated due to the use of non-random survey locations. Further surveys for this species using random locations will be crucial in determining its conservation status. We advocate the more wide-spread use of random survey locations in future camera-trapping surveys in order to increase the robustness and generality of inferences that can be made.
Article
Full-text available
Capture–recapture analysis of camera trap data is a conventional method to estimate the abundance of free-ranging wild felids. Due to notorious low detection rates of felids, it is important to increase the detection probability during sampling. In this study, we report the effectiveness of attractants as a tool for improving the efficiency of camera trap sampling in abundance estimation of Iberian lynx. We developed a grid system of camera stations in which stations with and without attractant lures were spatially alternated across known Iberian lynx habitat. Of the ten individuals identified, five were detected at stations with no attractant (blind sets), and nine, at the lured stations. Thirty-eight percent of blind set station’s independent captures and 10 % of lured station’s independent captures resulted in photographs unsuitable for correct individual identification. The total capture probability at lured stations was higher than that obtained at blind set stations. The estimates obtained with blind set cameras underestimated the number of lynxes compared to lured cameras. In our study, it appears that the use of lures increased the efficiency of trail camera captures and, therefore, the accuracy of capture–recapture analysis. The observed failure to detect known individuals at blind set camera stations may violate capture–recapture assumptions and bias abundance estimates.
Article
Full-text available
Automatically triggered cameras taking photographs or videos of passing animals (camera traps) have emerged over the last decade as one of the most powerful tool for wildlife research. In parallel, a wealth of camera trap systems and models has become commercially available, a phenomenon mainly driven by the increased use of camera traps by sport hunters. This has raised the need for developing criteria to choose the suitable camera trap model in relation to a range of factors, primarily the study aim, but also target species, habitat, trapping site, climate and any other aspect that affects camera performance. There is also fragmented information on the fundamentals of sampling designs that deploy camera trapping, such as number of sampling sites, spatial arrangement and sampling duration. In this review, we describe the relevant technological features of camera traps and propose a set of the key ones to be evaluated when choosing camera models. These Features are camera specifications such as trigger speed, sensor sensitivity, detection zone, flash type and flash intensity, power autonomy, and related specifications. We then outline sampling design and camera features for the implementation of major camera trapping applications, specifically: (1) faunal inventories, (2) occupancy studies, (3) density estimation through Capture-Mark-Recapture and (4) density estimation through the Random Encounter Model. We also review a range of currently available models and stress the need for standardized testing of camera models that should be frequently updated and widely distributed. Finally we summarize the “ultimate camera trap”, as desired by wildlife biologists, and the current technological limitations of camera traps in relation to their potential for a number of emerging applications.
Article
Full-text available
Statistical models for estimating absolute densities of field populations of animals have been widely used over the last century in both scientific studies and wildlife management programs. To date, two general classes of density estimation models have been developed: models that use data sets from capture-recapture or removal sampling techniques (often derived from trapping grids) from which separate estimates of population size (N ˆ) and effective sampling area (A ˆ) are used to calculate density (D ˆ 5 N ˆ/A ˆ); and models applicable to sampling regimes using distance-sampling theory (typically transect lines or trapping webs) to estimate detection functions and densities directly from the distance data. However, few studies have evaluated these respective models for accuracy, precision, and bias on known field populations, and no studies have been conducted that compare the two approaches under controlled field conditions. In this study, we evaluated both classes of density estimators on known densities of enclosed rodent populations. Test data sets ( n 5 11) were developed using nine rodent species from capture-recapture live-trapping on both trapping grids and trapping webs in four replicate 4.2-ha enclosures on the Sevilleta National Wildlife Refuge in central New Mexico, USA. Additional ''saturation'' trapping efforts resulted in an enumeration of the rodent populations in each enclosure, allowing the com- putation of true densities. Density estimates ( D ˆ) were calculated using program CAPTURE for the grid data sets and program DISTANCE for the web data sets, and these results were compared to the known true densities (D) to evaluate each model's relative mean square error, accuracy, precision, and bias. In addition, we evaluated a variety of approaches to each data set's analysis by having a group of independent expert analysts calculate their best density estimates without a priori knowledge of the true densities; this ''blind'' test allowed us to evaluate the influence of expertise and experience in calculating density estimates in comparison to simply using default values in programs CAPTURE and DIS- TANCE. While the rodent sample sizes were considerably smaller than the recommended minimum for good model results, we found that several models performed well empirically, including the web-based uniform and half-normal models in program DISTANCE, and the grid-based models Mb and Mbh in program CAPTURE (with Aadjusted by species-specific full mean maximum distance moved (MMDM) values). These models produced accurate Dvalues (with 95% confidence intervals that included the true D values) and exhibited acceptable bias but poor precision. However, in linear regression analyses comparing each model's Dvalues to the true D values over the range of observed test densities, only the web-based uniform model exhibited a regression slope near 1.0; all other models showed substantial slope deviations, indicating biased estimates at higher or lower density values. In addition, the grid-based Danalyses using full MMDM values for Warea adjustments required a number of theoretical assumptions of uncertain validity, and we therefore viewed their empirical successes with caution. Finally, density estimates from the independent analysts were highly variable, but estimates from web-based approaches had smaller mean square errors and better achieved confidence-interval coverage of D than did grid-based
Article
Full-text available
Densities of elusive terrestrial mammals are commonly estimated from camera-trap data. Typically, this is a 2-step process involving 1) fitting conventional closed population capture–recapture models to estimate abundance, and 2) using ad hoc methods to determine the effective trapping area. The methodology needs to be accurate, robust, and reliable when results are used to guide wildlife management. We critically review 47 published studies and discuss the problems associated with contemporary population estimates of elusive species from camera-trap data. In particular we discuss 1) individual identification, 2) sample size and capture probability, 3) camera location and spacing, 4) the size of the study area, and 5) ad hoc density estimation from the calculation of an effective trapping area. We also discuss the recently developed spatially explicit capture–recapture (SECR) models as an alternative approach that does not require the intermediate step of estimating an effective trapping area. We recommend 1) greater transparency in study design and quality of the data, 2) greater rigor when reviewing manuscripts, and 3) that more attention is given to the survey design to ensure data are of sufficient quality for analysis.
Article
Full-text available
The monitoring and management of species depends on reliable population estimates, and this can be both difficult and very costly for cryptic large vertebrates that live in forested habitats. Recently developed camera trapping techniques have already been shown to be an effective means of making mark-recapture estimates of individually identifiable animals (e.g. tigers). Camera traps also provide a new method for surveying animal abundance. Through computer simulations, and an analysis of the rates of camera trap capture from 19 studies of tigers across the species' range, we show that the number of camera days/tiger photograph correlates with independent estimates of tiger density. This statistic does not rely on individual identity and is particularly useful for estimating the population density of species that are not individually identifiable. Finally, we used the comparison between observed trapping rates and the computer simulations to estimate the minimum effort required to determine that tigers, or other species, do not exist in an area, a measure that is critical for conservation planning.
Article
Full-text available
a b s t r a c t Competition for mates, cub rearing, and other behaviors differ between males and females in large car-nivores. Although these differences can be reflected in patterns of habitat use, gender has rarely been incorporated into habitat models. We evaluated differences in habitat use between male and female jag-uars in the Mayan Forest of the Yucatan Peninsula by modeling occupancy as a function of land cover type, distance to roads, and sex. Nested models were fitted to high-spatiotemporal resolution satellite (GPS) telemetry, controlled for temporal autocorrelation, and eliminated selection bias of pseudo-absences using a semi-non-parametric bootstrap. Although both male and female jaguars prefer tall for-est, short forest was also preferred by females but avoided by males. Whereas females significantly avoided roads, males didn't and ventured into low-intensity cattle ranching and agriculture. Females' preference for intact forests and against roads led to their habitat being fragmented to a greater degree than that of males. Models that ignored sexual differences failed to capture the effect of roads and agri-culture on jaguar habitat use, blurred the distinction of use between short and tall forest, and underes-timated fragmentation of female jaguar habitat; but incorporating these differences increased precision of habitat maps and allowed the identification of potential jaguar-human conflict areas associated with male's use of cattle and agricultural lands. Specifying sex differences increases the power of habitat mod-els to understand landscape occupancy by large carnivores, and so greater attention should be paid to these differences in their modeling and conservation.
Article
Full-text available
We used extensive camera-trap surveys to study interindividual interactions among individually recognizable jaguars (Panthera onca) and plain-colored pumas (Puma concolor). Timed location data from a network of 119 trap stations in the Cockscomb Basin of Belize provide the 1st evidence of interspecific avoidance calibrated against intraspecific interactions among jaguars. Camera trapping has advantages over radiotelemetry in its potential to provide data on the complete array of individuals within the study area. The 23 individually identified male jaguars showed high levels of overlap in ranges, with up to 5 different males captured at the same location in the same month. Low levels of avoidance between individuals and a high flux of individuals contributed to low consistency in home-range ownership over the long term (3 months to 2 years). Jaguars and pumas had similar nocturnal activity schedules. Both species used similar habitats within the Cockscomb Basin, indicated by a high correlation in capture rates per location between species. Apart from their overall spatial similarities, jaguars and pumas avoided using the same location at the same time. This interspecific segregation was detectable over and above the spatial and temporal segregation of individual jaguars.
Article
Full-text available
Remote cameras are an increasingly important tool in management and wildlife studies. However, we often do not know if they provide an unbiased sample of populations. Using a marked, radiocollared population of coyotes (Canis latrans) of known social status, we evaluated the influence of temporal (daily and seasonal) and spatial (distance between units, habitat, and proximity to human structures) factors on vulnerability to photo-captures. During 8 unbaited camera sessions of 6 weeks each, we obtained 158 coyote photographs at a photo-capture success rate of 1.6%. We were able to identify not only marked individuals, but also a number of uncollared adults through variation in their pelage. Photo-capture of adults peaked 2 weeks after we established camera stations. Annual success for photographing adult coyotes was greatest during March and April, which corresponded with the dispersal season. The majority of photo-captures occurred at night, and adult photo-captures peaked around midnight, with smaller peaks at dawn and dusk. Rather than reflecting a circadian activity pattern, nighttime captures seemed to reflect when adult coyotes were most vulnerable to photo-capture. Characteristics of camera locations, such as amount of human activity, being on roads versus trails, and habitat type, also influenced the number of photo-captures. We conclude that remote cameras do not always provide an unbiased sample of populations and that animal behavior is important to consider when using these systems. Researchers using camera techniques need to carefully consider when, where, and how cameras are placed to reduce this bias.
Article
Full-text available
1. Calibrating indices of animal abundance to true densities is critical in wildlife studies especially when direct density estimations are precluded by high costs, lack of required data or model parameters, elusiveness and rarity of target species. For studies deploying camera traps, the use of photographic rate (photographs per sampling time) as an index of abundance potentially applies to the majority of terrestrial mammals where individual recognition, and hence capture–recapture analysis, are unfeasible. The very few studies addressing this method have either been limited by lack of independence between trapping rates and density estimations, or because they combined different species, thus introducing potential bias in camera trap detection rates. This study uses a single model species from several sites to analyse calibration of trapping rates to independently derived estimations of density. The study also makes the first field test of the method by Rowcliffe et al. (2008) for density derivation from camera trapping rates based on modelling animal-camera contacts.
Article
Full-text available
We examine the abundance and distribution of Sumatran tigers (Panthera tigris sumatrae) and nine prey species in Bukit Barisan Selatan National Park on Sumatra, Indonesia. Our study is the first to demonstrate that the relative abundance of tigers and their prey, as measured by camera traps, is directly related to independently derived estimates of densities for these species. The tiger population in the park is estimated at 40–43 individuals. Results indicate that illegal hunting of prey and tigers, measured as a function of human density within 10 km of the park, is primarily responsible for observed patterns of abundance, and that habitat loss is an increasingly serious problem. Abundance of tigers, two mouse deer (Tragulus spp.), pigs (Sus scrofa) and Sambar deer (Cervus unicolor) was more than four times higher in areas with low human population density, while densities of red muntjac (Muntiacus muntjac) and pigtail macaques (Macaca nemestrina) were twice as high. Malay tapir (Tapirus indicus) and argus pheasant (Argusianus argus), species infrequently hunted, had higher indices of relative abundance in areas with high human density. Edge effects associated with park boundaries were not a significant factor in abundance of tigers or prey once human density was considered. Tigers in Bukit Barisan Selatan National Park, and probably other protected areas throughout Sumatra, are in imminent danger of extinction unless trends in hunting and deforestation are reversed.
Article
Full-text available
The use of non-invasive long-term monitoring data to estimate home ranges of the critically endangered Iberian lynx has been evaluated. This programme began in 2002 and consisting of both annual latrine and camera-trap surveys, with the aims of detecting and individually identifying the maximum number of individuals and delineating female home range boundaries. Radio-tracking data were used to evaluate the accuracy of home range estimates constructed with camera-trapping data. There was little overlap of camera-trapping home ranges (7.0% ± 1.47), which suggests the existence of real territories consistent with the land tenure system expected for the species. Camera trapping home range estimates were half the size of radio-tracking data (54.1% ± 6.0 of overlapping). When comparing core areas, only the radio-tracking data did not yield improved results (36.7 ± 5.4 of overlapping). Estimation of territories, which escaped detection each year, ranged from 0.0% to 5.7%. The results produced by camera-trapping data in this non-intrusive monitoring programme could be considered precise, and are therefore well suited to provide the knowledge required for appropriate conservation of this endangered species. KeywordsCamera-trapping–Home range–Iberian lynx– Lynx pardinus –Monitoring–Sierra Morena
Article
Full-text available
The Iberian lynx (Lynx pardinus) has a highly restricted geographic distribution, limited even within the Iberian Peninsula. The last national survey reported less than 200 remaining individuals, distributed in two isolated areas—Andújar-Cardeña and Doñana—and in consequence, the Iberian lynx was listed by the International Union for Conservation of Nature as “Critically Endangered”. In this study, we estimate the Iberian lynx population size in the Doñana area using capture–recapture analysis of camera-trapping data. A model with different capture probability for each individual (Mh) yielded an estimate of 26 Iberian lynxes (SE = 5.26) more than 1year old. It is considered that a small slant in the estimation of the number of individuals could exist due to the presence of dispersers inside the study area that were not detected. Our study shows: (1) a reduction in number since the 1980s (45 individuals), and falling below the theoretical threshold of genetic viability, (2) changes in the species’ spatial distribution in this area, and (3) as for other carnivore species, photographic capture–recapture methods are applicable for estimating the size of Iberian lynx populations Keywords Lynx pardinus –Iberian lynx–Camera trapping–Capture–recapture–Population estimates–Doñana
Article
Full-text available
Density estimation is of fundamental importance in wildlife management. The use of camera traps to estimate animal density has so far been restricted to capture2013recapture analysis of species with individually identifiable markings. This study developed a method that eliminates the requirement for individual recognition of animals by modelling the underlying process of contact between animals and cameras. The model provides a factor that linearly scales trapping rate with density, depending on two key biological variables (average animal group size and day range) and two characteristics of the camera sensor (distance and angle within which it detects animals). We tested the approach in an enclosed animal park with known abundances of four species, obtaining accurate estimates in three out of four cases. Inaccuracy in the fourth species was because of biased placement of cameras with respect to the distribution of this species. Synthesis and applications. Subject to unbiased camera placement and accurate measurement of model parameters, this method opens the possibility of reduced labour costs for estimating wildlife density and may make estimation possible where it has not been previously. We provide guidelines on the trapping effort required to obtain reasonably precise estimates.
Article
Full-text available
The density of a closed population of animals occupying stable home ranges may be estimated from detections of individuals on an array of detectors, using newly developed methods for spatially explicit capture-recapture. Likelihood-based methods provide estimates for data from multi-catch traps or from devices that record presence without restricting animal movement ("proximity" detectors such as camera traps and hair snags). As originally proposed, these methods require multiple sampling intervals. We show that equally precise and unbiased estimates may be obtained from a single sampling interval, using only the spatial pattern of detections. This considerably extends the range of possible applications, and we illustrate the potential by estimating density from simulated detections of bird vocalizations on a microphone array. Acoustic detection can be defined as occurring when received signal strength exceeds a threshold. We suggest detection models for binary acoustic data, and for continuous data comprising measurements of all signals above the threshold. While binary data are often sufficient for density estimation, modeling signal strength improves precision when the microphone array is small.
Article
The tiger (Panthera tigris) is an endangered, large felid whose demographic status is poorly known across its distributional range in Asia. Previously applied methods for estimating tiger abundance, using total counts based on tracks, have proved unreliable. Lack of reliable data on tiger densities not only has constrained our ability to understand the ecological factors shaping communities of large, solitary felids, but also has undermined the effective conservation of these animals. In this paper, we describe the use of a field method proposed by Karanth (1995), which combines camera-trap photography, to identify individual tigers, with theoretically well-founded capture–recapture models. We developed a sampling design for camera-trapping and used the approach to estimate tiger population size and density in four representative tiger habitats in different parts of India. The field method worked well and provided data suitable for analysis using closed capture–recapture models. The results suggest the potential for applying this methodology to rigorously estimate abundances, survival rates, and other population parameters for tigers and other low-density, secretive animal species in which individuals can be identified based on natural markings. Estimated probabilities of photo-capturing tigers present in the study sites ranged from 0.75 to 1.00. Estimated densities of tigers >1 yr old ranged from 4.1 ± 1.31 to 16.8 ± 2.96 tigers/100 km2 (mean ± 1 se). Simultaneously, we used line-transect sampling to determine that mean densities of principal tiger prey at these sites ranged from 56.1 to 63.8 ungulates/km2. Tiger densities appear to be positively associated with prey densities, except at one site influenced by tiger poaching. Our results generally support the prediction that relative abundances of large felid species may be governed primarily by the abundance and structure of their prey communities.
Book
Remote photography and infrared sensors are widely used in the sampling of wildlife populations worldwide, especially for cryptic or elusive species. Guiding the practitioner through the entire process of using camera traps, this book is the first to compile state-of-the-art sampling techniques for the purpose of conducting high-quality science or effective management. Chapters on the evaluation of equipment, field sampling designs, and data analysis methods provide a coherent framework for making inferences about the abundance, species richness, and occupancy of sampled animals. The volume introduces new models that will revolutionize use of camera data to estimate population density, such as the newly developed spatial capture-recapture models. It also includes richly detailed case studies of camera trap work on some of the world's most charismatic, elusive, and endangered wildlife species. Indispensible to wildlife conservationists, ecologists, biologists, and conservation agencies around the world, the text provides a thorough review of the subject as well as a forecast for the use of remote photography in natural resource conservation over the next few decades.
Article
With the increasing popularity of remote photography in wildlife research, a large variety of equipment and methods is available to researchers. To evaluate advantages and disadvantages of using various types of equipment for different study objectives, we reviewed 107 papers that used either time-lapse or animal-triggered photography to study vertebrates in the field. Remote photography was used primarily to study avian nest predation, feeding ecology, and nesting behavior; additional applications included determining activity patterns, presence-absence monitoring, and estimating population parameters. Using time-lapse equipment is most appropriate when animals occur frequently in the photographic frame, the activity of interest occurs repeatedly, or no distinct event occurs to trigger a camera. In contrast, animal-triggered (light or mechanically triggered) systems are appropriate when events occur infrequently or unpredictably and there is a great likelihood that a trigger will be activated. Remote photography can be less time consuming, costly, and invasive than traditional research methods for many applications. However, researchers should be prepared to invest time and money troubleshooting problems with remote camera equipment, be aware of potential effects of equipment on animal behavior, and recognize the limitations of data collected with remote photography equipment.
Article
The random encounter model (REM) is a novel method for estimating animal density from camera trap data without the need for individual recognition. It has never been used to estimate the density of large carnivore species, despite these being the focus of most camera trap studies worldwide. In this context, we applied the REM to estimate the density of female lions (Panthera leo) from camera traps implemented in Serengeti National Park, Tanzania, comparing estimates to reference values derived from pride census data. More specifically, we attempted to account for bias resulting from non-random camera placement at lion resting sites under isolated trees by comparing estimates derived from night versus day photographs, between dry and wet seasons, and between habitats that differ in their amount of tree cover. Overall, we recorded 169 and 163 independent photographic events of female lions from 7,608 and 12,137 camera trap days carried out in the dry season of 2010 and the wet season of 2011, respectively. Although all REM models considered over-estimated female lion density, models that considered only night-time events resulted in estimates that were much less biased relative to those based on all photographic events. We conclude that restricting REM estimation to periods and habitats in which animal movement is more likely to be random with respect to cameras can help reduce bias in estimates of density for female Serengeti lions. We highlight that accurate REM estimates will nonetheless be dependent on reliable measures of average speed of animal movement and camera detection zone dimensions. © 2015 The Wildlife Society.
Article
Camera trap surveys exclusively targeting features of the landscape that increase the probability of photographing one or several focal species are commonly used to draw inferences on the richness, composition and structure of entire mammal communities. However, these studies ignore expected biases in species detection arising from sampling only a limited set of potential habitat features. In this study, we test the influence of camera trap placement strategy on community-level inferences by carrying out two spatially and temporally concurrent surveys of medium to large terrestrial mammal species within Tanzania's Ruaha National Park, employing either strictly game trail-based or strictly random camera placements. We compared the richness, composition and structure of the two observed communities, and evaluated what makes a species significantly more likely to be caught at trail placements. Observed communities differed marginally in their richness and composition , although differences were more noticeable during the wet season and for low levels of sampling effort. Lognormal models provided the best fit to rank abundance distributions describing the structure of all observed communities, regardless of survey type or season. Despite this, carnivore species were more likely to be detected at trail placements relative to random ones during the dry season, as were larger bodied species during the wet season. Our findings suggest that, given adequate sampling effort (> 1400 camera trap nights), placement strategy is unlikely to affect inferences made at the community level. However, surveys should consider more carefully their choice of placement strategy when targeting specific taxonomic or trophic groups.
Article
The search for easy-to-use indices that substitute for direct estimation of animal density is a common theme in wildlife and conservation science, but one fraught with well-known perils (Nichols & Conroy, 1996; Yoccoz, Nichols & Boulinier, 2001; Pollock et al., 2002). To establish the utility of an index as a substitute for an estimate of density, one must: (1) demonstrate a functional relationship between the index and density that is invariant over the desired scope of inference; (2) calibrate the functional relationship by obtaining independent measures of the index and the animal density; (3) evaluate the precision of the calibration (Diefenbach et al., 1994). Carbone et al. (2001) argue that the number of camera-days per photograph is a useful index of density for large, cryptic, forest-dwelling animals, and proceed to calibrate this index for tigers (Panthera tigris). We agree that a properly calibrated index may be useful for rapid assessments in conservation planning. However, Carbone et al. (2001), who desire to use their index as a substitute for density, do not adequately address the three elements noted above. Thus, we are concerned that others may view their methods as justification for not attempting directly to estimate animal densities, without due regard for the shortcomings of their approach.
Article
4017 (USA); present address: Swiss Ornithological Institute, CH – 6204 Sempach (Switzerland) Trolle M. & Kery M. 2005. — Camera-trap study of ocelot and other secretive mammals in the northern Pantanal. Mammalia 69 (3-4) : 405-412. ABSTRACT Reliable information on abundance of the ocelot (Leopardus pardalis) is scarce. We conducted the first camera-trap study in the northern part of the Pantanal wetlands of Brazil, one of the wildlife hotspots of South America. Using capture-recapture analysis, we estimated a density of 0.112 independ-ent individuals per km 2 (SE 0.069). We list other mammals recorded with camera traps and show that camera-trap placement on roads or on trails has striking effects on camera-trapping rates. RÉSUMÉ Étude par piège photographique d'ocelots et autres mammifères cachés dans le Pantanal nord .Bien que l'ocelot (Leopardus pardalis) soit relativement com-mun dans les plaines tropicales d'Amérique du Sud, peu de données sont dis-ponibles sur l'abondance de ce félidé. Nous présentons ici les résultats de la première étude par piège photographique publiée pour la partie nord des zones humides du Pantanal au Brésil, l'un des sites les plus importants d'Amérique du sud pour la faune sauvage. En combinant l'utilisation de pièges photographiques avec des analyses de capture-recapture, nous avons estimé une densité d'ocelots de 0.11 individus indépendants par km 2 (SE 0.069). Nous donnons aussi une liste d'autres mammifères enregistrés par piège photographique. Finalement, nous montrons que le choix des emplace-ments des pièges, sur des routes ou sur des sentiers, a un effet marqué sur le taux de capture photographique.
Article
Spatial ecology and diet of Iberian lynx (Lynx paradinus) and the abundance of its main prey, the European rabbit (Oryctolagus cuniculus), were studied in southwestern Spain from December 1992 to December 1996 when a decline in rabbit populations occurred. Our objectives were to relate spatial ecology of lynx to rabbit abundance, water availability, and protection from human disturbance. Rabbits were almost the bole prey of lynx; rabbit remains were present in 99.2% (n = 1,171) of feces analyzed. Rabbit abundance and density in 6 different habitats were estimated by line transect sampling. Rabbits were more abundant in Mediterranean scrubland, closely followed by ash stands. Pastureland and lentiscus in plantations had 4.5 times fewer rabbits, and pine plantations 15-20 times fewer rabbits, than the Mediterranean scrubland. Abundance of rabbits in the Mediterranean scrubland was not spatially uniform, as density for 1994 and 1995 varied from 42-55/ha in the area close to the edge of the marsh (locally called the V ra) to 2-7/ha far from the Vera. Rabbit density was 3.5 times higher during spring thane during autumn. A decline in rabbit density occurred in 1996 when numbers were 72-91% lower than the previous years. We mapped warren density in Mediterranean scrubland and pastureland to determine intra-habitat (differences in spatial distribution of rabbits. Warren density, entrance density, and mean number of entrances per warren declined significantly with distance from the Vera in the Mediterranean scrubland, following a negative exponential function (adj. r(2) ranging between 83 and 97%). However, none of these variables showed any trend in relation to the Vera in the pastureland. The lynx population contained three stable pairs of adults plus young raised each year, some of which remained on the study area as subadults (older than one year), The seasonal number of different lynx in the study area ranged between 7 and 17. Average adult and young/subadult seasonal density was 0.77 (range 0.72-0.88) and 0.46 individuals/km(2) (range = 0.07-1.12),respectively We estimated home range and daily movements of lynx to determine changes that might be due to sex, season, or changing prey density. On average, total lynx home range size was 7.3 km(2) for young, 9.5 km(2) for yl adult females, and 18.2 km(2) fo adult males. Mean bore areas (60% isopleth using the kernel approximation) were on average 15%, 10%, and 34% of total home ranges of young, adult females, and adult males, respectively. Significant differences were found for home range and core area sizes among sex-age classes, but neither season nor year affected home range size or core area size. Daily movements averaged 8.0 kill. Daily distance traveled was not affected by sex-age class or season, but was different among years, with lynx traveling shorter distances in 1993 and 1996. Daily home range size averaged 1.46 km(2), mid again only varied by year. Lynx daily movements were associated With permanent, artificial water sites. Habitat use by lynx was remarkably constant, with no differences detected among sex-age classes, active or inactive locations, seasons, or years. The habitat most frequently used was Mediterranean scrubland (53% of locations), and both it and ash stands were the only habitats preferred by lynx; pine and eucalyptus plantations were avoided, and marsh, pastureland and lentiscus in plantations were neither preferred or avoided.). When lynx were found in the non-preferred habitats, on most occasions (78%) animals were closer than 300 in from the edge of one of the two preferred habitats, whereas on only 4% of occasions were animals further than 1 km. This behavior was particularly accentuated when lynx moved through open habitats. Lynx appeared to respond to high human presence, as they were mainly located inside the National Park (82% of occasions), and when outside the Park they more frequently used the areas that were farther from a tourist village. We used a Geographical Information System (GIS) to estimate average rabbit density and number of rabbits within lynx home ranges. On average, rabbit density within home ranges was 5.6/ha. The habitats that sustained more rabbits were Mediterranean scrubland (74.2% of total number of rabbits within home ranges) and ash stands (32.2% for 6 home ranges where these habitats were available). Rabbit densities within core areas were similar to those found in home ranges. Throughout the study period the lowest rabbit density estimated within home ranges was about I/ha in autumn 1996. The number of rabbits per Lynx ranged between 1,367 in spring 1994 and 73 in autumn 1996. Rabbit density within home ranges, core areas, and the trapping area were not correlated with range size (P's > 0.2). Lynx and rabbits preferred the least human-transformed scrubland habitats therefore natural habitats must be favored over plantations and pastureland areas. The edges of the preferred habitats were also heavily used by rabbits and consequently by lynx. The vegetation structure of the preferred habitats was characterized by intermediate understory cover (25-35%), low tree cover, and large and frequent grasslands. The rabbit decline observed during the study did not affect lynx spatial behavior or reproduction. Therefore, rabbit densities as low as 1 and 4.6/ha for the times of the lowest and highest rabbit density (i.e., autumn and spring, respectively) were enough to sustain tire lynx population. Other factors such as the presence of permanent water sites and relatively low human presence are important components of quality lynx habitat.
Article
Conservation practices are supposed to get refined by advancing scientific knowledge. We study this phenomenon in the context of monitoring tiger populations in India, by evaluating the ‘pugmark census method’ employed by wildlife managers for three decades. We use an analytical framework of modern animal population sampling to test the efficacy of the pugmark censuses using scientific data on tigers and our field observations. We identify three critical goals for monitoring tiger populations, in order of increasing sophistication: (1) distribution mapping, (2) tracking relative abundance, (3) estimation of absolute abundance. We demonstrate that the present census-based paradigm does not work because it ignores the first two simpler goals, and targets, but fails to achieve, the most difficult third goal. We point out the utility and ready availability of alternative monitoring paradigms that deal with the central problems of spatial sampling and observability. We propose an alternative sampling-based approach that can be tailored to meet practical needs of tiger monitoring at different levels of refinement.
Article
Live-trapping capture-recapture studies of animal populations with fixed trap locations inevitably have a spatial component: animals close to traps are more likely to be caught than those far away. This is not addressed in conventional closed-population estimates of abundance and without the spatial component, rigorous estimates of density cannot be obtained. We propose new, flexible capture-recapture models that use the capture locations to estimate animal locations and spatially referenced capture probability. The models are likelihood-based and hence allow use of Akaike's information criterion or other likelihood-based methods of model selection. Density is an explicit parameter, and the evaluation of its dependence on spatial or temporal covariates is therefore straightforward. Additional (nonspatial) variation in capture probability may be modeled as in conventional capture-recapture. The method is tested by simulation, using a model in which capture probability depends only on location relative to traps. Point estimators are found to be unbiased and standard error estimators almost unbiased. The method is used to estimate the density of Red-eyed Vireos (Vireo olivaceus) from mist-netting data from the Patuxent Research Refuge, Maryland, U.S.A. Estimates agree well with those from an existing spatially explicit method based on inverse prediction. A variety of additional spatially explicit models are fitted; these include models with temporal stratification, behavioral response, and heterogeneous animal home ranges.
The R Project for Statistical Computing. www.r-project.org
  • Development Core
R DEVELOPMENT CORE TEAM () The R Project for Statistical Computing. www.r-project.org [accessed  December ].
) Monitoring an Endangered savannah ungulate, Grevy's zebra Equus grevyi: choosing a method for estimating population densities
  • V H Zero
  • S R Sundaresan
  • T G Brien
  • M F Kinnaird
ZERO, V.H., SUNDARESAN, S.R., O'BRIEN, T.G. & KINNAIRD, M.F. () Monitoring an Endangered savannah ungulate, Grevy's zebra Equus grevyi: choosing a method for estimating population densities. Oryx, , -.