Achieving explanatory depth and spatial breadth in infectious disease modelling: Integrating active and passive case surveillance

Ideally, the data used for robust spatial prediction of disease distribution should be both high-resolution and spatially expansive. However, such in-depth and geographically broad data are rarely available in practice. Instead, researchers usually acquire either detailed epidemiological data with high resolution at a small number of active sampling sites, or more broad-ranging but less precise data from passive case surveillance. We propose a novel inferential framework, capable of simultaneously drawing insights from both passive and active data types. We developed a Bayesian latent point process approach, combining active data collection in a limited set of points, where in-depth covariates are measured, with passive case detection, where error-prone, large-scale disease data are accompanied only by coarse or remotely-sensed covariate layers. Using the example of malaria, we tested our method's efficiency under several hypothetical scenarios of reported incidence in different combinations of imperfect detection and spatial complexity of the environmental variables. We provide a simple solution to a widespread problem in spatial epidemiology, combining latent process modelling and spatially autoregressive modelling. By using active sampling and passive case detection in a complementary way, we achieved the best-of-both-worlds, in effect, a formal calibration of spatially extensive, error-prone data by localised, high-quality data.

Simultaneous-count models to estimate abundance from counts of unmarked individuals with imperfect detection

We developed a method to estimate population abundance from simultaneous counts of unmarked individuals over multiple sites. We considered that at each sampling occasion, individuals in a population could be detected at 1 of the survey sites or remain undetected and used either multinomial or binomial simultaneous-count models to estimate abundance, the latter being equivalent to an N-mixture model with one site. We tested model performance with simulations over a range of detection probabilities, population sizes, growth rates, number of years, sampling occasions, and sites. We then applied our method to 3 critically endangered vulture species in Cambodia to demonstrate the real-world applicability of the model and to provide the first abundance estimates for these species in Cambodia. Our new approach works best when existing methods are expected to perform poorly (i.e., few sites and large variation in abundance among sites) and if individuals may move among sites between sampling occasions. The approach performed better when there were >8 sampling occasions and net probability of detection was high (>0.5). We believe our approach will be useful in particular for simultaneous surveys at aggregation sites, such as roosts. The method complements existing approaches for estimating abundance of unmarked individuals and is the first method designed specifically for simultaneous counts.

The importance of considering multiple interacting species for conservation of species at risk

Conservation of species at risk of extinction is complex and multifaceted. However, mitigation strategies are typically narrow in scope, an artifact of conservation research that is often limited to a single species or stressor. Knowledge of an entire community of strongly interacting species would greatly enhance the comprehensiveness and effectiveness of conservation decisions. We investigated how camera trapping and spatial count models, an extension of spatial-recapture models for unmarked populations, can accomplish this through a case study of threatened boreal woodland caribou (Rangifer tarandus caribou). Population declines in caribou are precipitous and well documented, but recovery strategies focus heavily on control of wolves (Canis lupus) and pay less attention to other known predators and apparent competitors. Obtaining necessary data on multispecies densities has been difficult. We used spatial count models to concurrently estimate densities of caribou, their predators (wolf, black bear [Ursus americanus], and coyote [Canis latrans]), and alternative prey (moose [Alces alces] and white-tailed deer [Odocoileus virginianus]) from a camera-trap array in a highly disturbed landscape within northern Alberta's Oil Sands Region. Median densities were 0.22 caribous (95% Bayesian credible interval [BCI] = 0.08-0.65), 0.77 wolves (95% BCI = 0.26-2.67), 2.39 moose (95% BCI = 0.56-7.00), 2.64 coyotes (95% BCI = 0.45-6.68), and 3.63 black bears (95% BCI = 1.25-8.52) per 100 km² . (The white-tailed deer model did not converge.) Although wolf densities were higher than densities recommended for caribou conservation, we suggest the markedly higher black bear and coyote densities may be of greater concern, especially if government wolf control further releases these species. Caribou conservation with a singular focus on wolf control may leave caribou vulnerable to other predators. We recommend a broader focus on the interacting species within a community when conserving species.

Advances in population ecology and species interactions in mammals

The study of mammals has promoted the development and testing of many ideas in contemporary ecology. Here we address recent developments in foraging and habitat selection, source–sink dynamics, competition (both within and between species), population cycles, predation (including apparent competition), mutualism, and biological invasions. Because mammals are appealing to the public, ecological insight gleaned from the study of mammals has disproportionate potential in educating the public about ecological principles and their application to wise management. Mammals have been central to many computational and statistical developments in recent years, including refinements to traditional approaches and metrics (e.g., capture-recapture) as well as advancements of novel and developing fields (e.g., spatial capture-recapture, occupancy modeling, integrated population models). The study of mammals also poses challenges in terms of fully characterizing dynamics in natural conditions. Ongoing climate change threatens to affect global ecosystems, and mammals provide visible and charismatic subjects for research on local and regional effects of such change as well as predictive modeling of the long-term effects on ecosystem function and stability. Although much remains to be done, the population ecology of mammals continues to be a vibrant and rapidly developing field. We anticipate that the next quarter century will prove as exciting and productive for the study of mammals as has the recent one.

Generalized spatial mark-resight models with incomplete identification: An application to red fox density estimates

The estimation of abundance of wildlife populations is an essential part of ecological research and monitoring. Spatially explicit capture-recapture (SCR) models are widely used for abundance and density estimation, frequently through individual identification of target species using camera-trap sampling.Generalized spatial mark-resight (Gen-SMR) is a recently developed SCR extension that allows for abundance estimation when only a subset of the population is recognizable by artificial or natural marks. However, in many cases, it is not possible to read the marks in camera-trap pictures, even though individuals can be recognized as marked. We present a new extension of Gen-SMR that allows for this type of incomplete identification.We used simulation to assess how the number of marked individuals and the individual identification rate influenced bias and precision. We demonstrate the model's performance in estimating red fox (Vulpes vulpes) density with two empirical datasets characterized by contrasting densities and rates of identification of marked individuals. According to the simulations, accuracy increases with the number of marked individuals (m), but is less sensitive to changes in individual identification rate (δ). In our case studies of red fox density estimation, we obtained a posterior mean of 1.60 (standard deviation SD: 0.32) and 0.28 (SD: 0.06) individuals/km2, in high and low density, with an identification rate of 0.21 and 0.91, respectively.This extension of Gen-SMR is broadly applicable as it addresses the common problem of incomplete identification of marked individuals during resighting surveys.

Improving estimation of puma (Puma concolor) population density: clustered camera-trapping, telemetry data, and generalized spatial mark-resight models

Obtaining reliable population density estimates for pumas (Puma concolor) and other cryptic, wide-ranging large carnivores is challenging. Recent advancements in spatially explicit capture-recapture models have facilitated development of novel survey approaches, such as clustered sampling designs, which can provide reliable density estimation for expansive areas with reduced effort. We applied clustered sampling to camera-traps to detect marked (collared) and unmarked pumas, and used generalized spatial mark-resight (SMR) models to estimate puma population density across 15,314 km² in the southwestern USA. Generalized SMR models outperformed conventional SMR models. Integrating telemetry data from collars on marked pumas with detection data from camera-traps substantially improved density estimates by informing cryptic activity (home range) center transiency and improving estimation of the SMR home range parameter. Modeling sex of unmarked pumas as a partially identifying categorical covariate further improved estimates. Our density estimates (0.84–1.65 puma/100 km²) were generally more precise (CV = 0.24–0.31) than spatially explicit estimates produced from other puma sampling methods, including biopsy darting, scat detection dogs, and regular camera-trapping. This study provides an illustrative example of the effectiveness and flexibility of our combined sampling and analytical approach for reliably estimating density of pumas and other wildlife across geographically expansive areas.

Patterns of vector species richness and species composition as drivers of Chagas disease occurrence in Brazil

Chagas disease represents one of the major health issue in Latin America. Epidemiological control is focused on disease vectors, so studies on the ecology of triatomine vectors constitute a central strategy. Recently, research at large spatial scale has been produced, and authors commonly rely on the assumption that geographical regions presenting good environmental conditions for most vector species are also those with high risk of infection. In the present work, we provide an explicit evaluation for this assumption. Employing species distribution models and epidemiological data for Chagas disease in Brazilian territory, our results show that species richness is a poor predictor for the observed pattern of Chagas disease occurrence. Species composition proved to be a better predictor. We stress that research on macroecology of infectious diseases should go beyond the analysis of biodiversity patterns and consider human infections as a central part of the focal ecological systems.

Methods for assessing small-scale variation in the abundance of a generalist mesopredator

Estimating animal abundance is essential for research, management and conservation purposes. Although reliable methods exist to estimate absolute density for populations with individually marked animals, robust relative abundance indices (RAIs) may allow to track changes in population size when individual identification is not possible. Their performance, however, needs be thoroughly evaluated. We investigated the relative performance of several common faeces-based and camera-based RAIs for estimating small-scale variation in red fox abundance, a mesopredator of high relevance for management, in two different study areas. We compared precision, cost and performance of the methods in capturing relationships with covariates of local abundance. Random transect-based RAIs had a low mean, a comparatively high coefficient of variation and a high proportion of zeros, prohibiting or impeding analysis in relation to environmental predictors. Rectangular scat plots and transects along linear landscape features had an intermediate amount of zeros while retaining a high precision, but were less sensitive to local variation in abundance related to environmental predictors and required a large field effort. Camera trap-based RAIs yielded low to intermediate precision, but were more sensitive to small-scale variation in relative abundance than faeces-based methods. Camera traps were the most expensive methods for an initial monitoring session, but required the lowest field effort, were cheapest in the long run and were the least susceptible to observer bias and detection error under a robust sampling protocol. Generally, faeces count-based RAIs appear more suitable for studies that aim to compare local abundance between several study sites of equal landscape composition under constant detection probability. Camera traps provide more flexible data for studies that require accounting for influences of landscape composition on local abundance and are more cost-effective for long-term or continuous monitoring and more suitable to achieve high replication. Accordingly, the choice of the most suitable method and plot design is context-dependent.

Cluster capture-recapture to account for identification uncertainty on aerial surveys of animal populations

Capture-recapture methods for estimating wildlife population sizes almost always require their users to identify every detected animal. Many modern-day wildlife surveys detect animals without physical capture-visual detection by cameras is one such example. However, for every pair of detections, the surveyor faces a decision that is often fraught with uncertainty: are they linked to the same individual? An inability to resolve every such decision to a high degree of certainty prevents the use of standard capture-recapture methods, impeding the estimation of animal density. Here, we develop an estimator for aerial surveys, on which two planes or unmanned vehicles (drones) fly a transect over the survey region, detecting individuals via high-definition cameras. Identities remain unknown, so one cannot discern if two detections match to the same animal; however, detections in close proximity are more likely to match. By modeling detection locations as a clustered point process, we extend recently developed methodology and propose a precise and computationally efficient estimator of animal density that does not require individual identification. We illustrate the method with an aerial survey of porpoise, on which cameras detect individuals at the surface of the sea, and we need to take account of the fact that they are not always at the surface.

Spatial capture-recapture design and modelling for the study of small mammals

Spatial capture-recapture modelling (SCR) is a powerful analytical tool to estimate density and derive information on space use and behaviour of elusive animals. Yet, SCR has been seldom applied to the study of ecologically keystone small mammals. Here we highlight its potential and requirements with a case study on common voles (Microtus arvalis). First, we address mortality associated with live-trapping, which can be high in small mammals, and must be kept minimal. We designed and tested a nest box coupled with a classic Sherman trap and show that it allows a 5-fold reduction of mortality in traps. Second, we address the need to adjust the trapping grid to the individual home range to maximize spatial recaptures. In May-June 2016, we captured and tagged with transponders 227 voles in a 1.2-ha area during two monthly sessions. Using a Bayesian SCR with a multinomial approach, we estimated: (1) the baseline detection rate and investigated variation according to sex, time or behaviour (aversion/attraction after a previous capture); (2) the parameter sigma that describes how detection probability declines as a function of the distance to an individual´s activity centre, and investigated variation according to sex; and (3) density and population sex-ratio. We show that reducing the maximum distance between traps from 12 to 9.6m doubled spatial recaptures and improved model predictions. Baseline detection rate increased over time (after overcoming a likely aversion to entering new odourless traps) and was greater for females than males in June. The sigma parameter of males was twice that of females, indicating larger home ranges. Density estimates were of 142.92±38.50 and 168.25±15.79 voles/ha in May and June, respectively, with 2–3 times more females than males. We highlight the potential and broad applicability that SCR offers and provide specific recommendations for using it to study small mammals like voles.

Clair K, Fieberg JR. Time series sightability modeling of animal populations

Logistic regression models—or “sightability models”—fit to detection/non-detection data from marked individuals are often used to adjust for visibility bias in later detection-only surveys, with population abundance estimated using a modified Horvitz-Thompson (mHT) estimator. More recently, a model-based alternative for analyzing combined detection/non-detection and detection-only data was developed. This approach seemed promising, since it resulted in similar estimates as the mHT when applied to data from moose (Alces alces) surveys in Minnesota. More importantly, it provided a framework for developing flexible models for analyzing multiyear detection-only survey data in combination with detection/non-detection data. During initial attempts to extend the model-based approach to multiple years of detection-only data, we found that estimates of detection probabilities and population abundance were sensitive to the amount of detection-only data included in the combined (detection/non-detection and detection-only) analysis. Subsequently, we developed a robust hierarchical modeling approach where sightability model parameters are informed only by the detection/non-detection data, and we used this approach to fit a fixed-effects model (FE model) with year-specific parameters and a temporally-smoothed model (TS model) that shares information across years via random effects and a temporal spline. The abundance estimates from the TS model were more precise, with decreased interannual variability relative to the FE model and mHT abundance estimates, illustrating the potential benefits from model-based approaches that allow information to be shared across years.

A camera-based method for estimating absolute density in animals displaying home range behaviour

The measurement of animal density may take advantage of recent technological achievements in wildlife video recording. Fostering the theoretical links between the patterns depicted by cameras and absolute density is required to exploit this potential. We explore the applicability of the Hutchinson-Waser's postulate (i.e. when animal density is stationary at a given temporal and spatial scale, the absolute density is given by the average number of animals counted per frame), which is a counter-intuitive statement for most ecologists and managers who are concerned with counting the same individual more than once. We aimed to reconcile such scepticism for animals displaying home range behaviour. The specific objectives of this paper are to generalize the Hutchinson-Waser's postulate for animals displaying home range behaviour and to propose a Bayesian implementation to estimate density from counts per frame using video cameras. Accuracy and precision of the method was evaluated by means of computer simulation experiments. Specifically, six animal archetypes displaying well-contrasted movement features were considered. The simulation results demonstrate that density could be accurately estimated after an affordable sampling effort (i.e. number of cameras and deployment time) for a great number of animals across taxa. The proposed method may complement other conventional methods for estimating animal density. The major advantages are that identifying an animal at the individual level and precise knowledge on how animals move are not needed, and that density can be estimated in a single survey. The method can accommodate conventional camera trapping data. The major limitations are related to some implicit assumptions of the underlying model: the home range centres should be homogeneously distributed, the detection probability within the area surveyed by the camera should be known, and animals should move independently to one another. Further improvements for circumventing these limitations are discussed.

Response of pumas (Puma concolor) to migration of their primary prey in Patagonia

Large-scale ungulate migrations result in changes in prey availability for top predators and, as a consequence, can alter predator behavior. Migration may include entire populations of prey species, but often prey populations exhibit partial migration with some individuals remaining resident and others migrating. Interactions of migratory prey and predators have been documented in North America and some other parts of the world, but are poorly studied in South America. We examined the response of pumas (Puma concolor) to seasonal migration of guanacos (Lama guanicoe) in La Payunia Reserve in northern Patagonia Argentina, which is the site of the longest known ungulate migration in South America. More than 15,000 guanacos migrate seasonally in this landscape, and some guanacos also are resident year-round. We hypothesized that pumas would respond to the guanaco migration by consuming more alternative prey rather than migrating with guanacos because of the territoriality of pumas and availability of alternative prey throughout the year at this site. To determine whether pumas moved seasonally with the guanacos, we conducted camera trapping in the summer and winter range of guanacos across both seasons and estimated density of pumas with spatial mark-resight (SMR) models. Also, we analyzed puma scats to assess changes in prey consumption in response to guanaco migration. Density estimates of pumas did not change significantly in the winter and summer range of guanacos when guanacos migrated to and from these areas, indicating that pumas do not follow the migration of guanacos. Pumas also did not consume more alternative native prey or livestock when guanaco availability was lower, but rather fed primarily on guanacos and some alternative prey during all seasons. Alternative prey were most common in the diet during summer when guanacos also were abundant on the summer range. The response of pumas to the migration of guanacos differs from sites in the western North America where entire prey populations migrate and pumas migrate with their prey or switch to more abundant prey when their primary prey migrates.

Effects of scale of movement, detection probability, and true population density on common methods of estimating population density

Knowledge of population density is necessary for effective management and conservation of wildlife, yet rarely are estimators compared in their robustness to effects of ecological and observational processes, which can greatly influence accuracy and precision of density estimates. In this study, we simulate biological and observational processes using empirical data to assess effects of animal scale of movement, true population density, and probability of detection on common density estimators. We also apply common data collection and analytical techniques in the field and evaluate their ability to estimate density of a globally widespread species. We find that animal scale of movement had the greatest impact on accuracy of estimators, although all estimators suffered reduced performance when detection probability was low, and we provide recommendations as to when each field and analytical technique is most appropriately employed. The large influence of scale of movement on estimator accuracy emphasizes the importance of effective post-hoc calculation of area sampled or use of methods that implicitly account for spatial variation. In particular, scale of movement impacted estimators substantially, such that area covered and spacing of detectors (e.g. cameras, traps, etc.) must reflect movement characteristics of the focal species to reduce bias in estimates of movement and thus density.

Estimating carnivore community structures

Obtaining reliable estimates of the structure of carnivore communities is of paramount importance because of their ecological roles, ecosystem services and impact on biodiversity conservation, but they are still scarce. This information is key for carnivore management: to build support for and acceptance of management decisions and policies it is crucial that those decisions are based on robust and high quality information. Here, we combined camera and live-trapping surveys, as well as telemetry data, with spatially-explicit Bayesian models to show the usefulness of an integrated multi-method and multi-model approach to monitor carnivore community structures. Our methods account for imperfect detection and effectively deal with species with non-recognizable individuals. In our Mediterranean study system, the terrestrial carnivore community was dominated by red foxes (0.410 individuals/km²); Egyptian mongooses, feral cats and stone martens were similarly abundant (0.252, 0.249 and 0.240 individuals/km², respectively), whereas badgers and common genets were the least common (0.130 and 0.087 individuals/km², respectively). The precision of density estimates improved by incorporating multiple covariates, device operation, and accounting for the removal of individuals. The approach presented here has substantial implications for decision-making since it allows, for instance, the evaluation, in a standard and comparable way, of community responses to interventions.

Are we getting the full picture? Animal responses to camera traps and implications for predator studies

Camera trapping is widely used in ecological studies. It is often considered nonintrusive simply because animals are not captured or handled. However, the emission of light and sound from camera traps can be intrusive. We evaluated the daytime and nighttime behavioral responses of four mammalian predators to camera traps in road‐based, passive (no bait) surveys, in order to determine how this might affect ecological investigations. Wild dogs, European red foxes, feral cats, and spotted‐tailed quolls all exhibited behaviors indicating they noticed camera traps. Their recognition of camera traps was more likely when animals were approaching the device than if they were walking away from it. Some individuals of each species retreated from camera traps and some moved toward them, with negative behaviors slightly more common during the daytime. There was no consistent response to camera traps within species; both attraction and repulsion were observed. Camera trapping is clearly an intrusive sampling method for some individuals of some species. This may limit the utility of conclusions about animal behavior obtained from camera trapping. Similarly, it is possible that behavioral responses to camera traps could affect detection probabilities, introducing as yet unmeasured biases into camera trapping abundance surveys. These effects demand consideration when utilizing camera traps in ecological research and will ideally prompt further work to quantify associated biases in detection probabilities.

Density and activity patterns of pumas in hunted and non-hunted areas in central Argentina

Context Hunting has demographic effects on large and medium carnivores, causing population reductions and even extinctions worldwide. Yet, there is little information on carnivore demographic parameters and spatial and temporal land-use patterns in areas experiencing sport hunting, thus hindering effective conservation plans for such areas. Aims We estimated densities and determined activity patterns of pumas (Puma concolor) from camera-trapping surveys in a protected area and in a game reserve with sport hunting, in the Caldén forest of central Argentina. Methods We used both non-spatial and spatial mark–resight techniques to estimate and compare puma densities and we used kernel-density estimation (KDE) techniques to analyse and compare puma activity patterns between study sites. Key results Puma densities estimated from spatial models were lower than densities estimated from non-spatial mark–resight techniques. However, estimated density of pumas in the protected area was always higher (range = 4.89–9.32 per 100 km2) than in the game reserve (range = 0.52–1.98 per 100 km2), regardless of the estimation technique used. Trapping rates for large mammal prey were similar across sites. Pumas exhibited more nocturnal behaviour and high activity peaks at 0600 hours and 1100 hours in the hunted game reserve, whereas puma activity was spread more evenly around the clock in the protected area. Conclusions The higher puma densities in the protected area reflect the potential for such areas to function as refugia in a human-dominated landscape. However, the game reserve had a lower puma density than the protected area despite high trap rates of large prey, indicating that these areas may function as attractive sinks. Implications Our results could indicate that puma sport hunting in the Caldén forest should be managed at a metapopulation, regional level, and include both no-hunting areas (protected area, as potential sources) and hunting areas (game reserves, as potential sinks). Considering that our study areas were small and that this was an unreplicated study, we urge more research to be conducted, so as to determine whether sport hunting is compatible with puma conservation in the region.

Using occupancy-based surveys and multi-model inference to estimate abundance and distribution of crested gibbons (Nomascus spp.) in central Laos

Reliable assessments of species' status are prerequisites for monitoring the success of conservation programmes. However, survey conditions such as terrain and inaccessibility, compounded by the low densities of many species across Southeast Asia and other parts of the world are considerable barriers to obtaining robust populations estimates. We used an occupancy-based approach and multi-model inference to generate occupancy and abundance estimates for northern white-cheeked crested gibbons Nomascus leucogenys and southern white-cheeked crested gibbons N. siki in the Nam Kading National Protected Area (NKNPA) in central Lao Peoples' Democratic Republic (hereafter Laos). We present these estimates for gibbons within the context of a strategy designed to monitor multiple species and discuss the practical challenges to obtaining sufficient data for robust population estimates to detect change in gibbon status over time. We surveyed approximately 210 km² of habitat and estimate an abundance of 45 (SE = 17, CV = 37%) groups, giving an average site abundance of 0.21 (SE = 0.08, CV = 37%) groups per km² . We make recommendations for ongoing gibbon monitoring and discuss the wider implications for cost effective wildlife monitoring in Laos. Am. J. Primatol. 78:462-472, 2016. © 2015 Wiley Periodicals, Inc.