Evaluating and integrating spatial capture-recapture models with data of variable individual identifiability

Spatial capture-recapture (SCR) models have become the preferred tool for estimating densities of carnivores. Within this family of models are variants requiring identification of all individuals in each encounter (SCR), a subset of individuals only (generalized spatial mark-resight, gSMR), or no individual identification (spatial count or spatial presence-absence). Although each technique has been shown through simulation to yield unbiased results, the consistency and relative precision of estimates across methods in real-world settings are seldom considered. We tested a suite of models ranging from those only requiring detections of unmarked individuals to others that integrate remote camera, physical capture, genetic, and global positioning system (GPS) data into a hybrid model, to estimate population densities of black bears, bobcats, cougars, and coyotes. For each species, we genotyped fecal DNA collected with detection dogs during a 20-d period. A subset of individuals from each species was affixed with GPS collars bearing unique markings and resighted by remote cameras over 140 d contemporaneous with scat collection. Camera-based gSMR models produced density estimates that differed by <10% from genetic SCR for bears, cougars, and coyotes once important sources of variation (sex or behavioral status) were controlled for. For bobcats, SCR estimates were 33% higher than gSMR. The cause of the discrepancies in estimates was likely attributable to challenges designing a study compatible for species with disparate home range sizes and the difficulty of collecting sufficient data in a timeframe in which demographic closure could be assumed. Unmarked models estimated densities that varied greatly from SCR, but estimates became more consistent in models wherein more individuals were identifiable. Hybrid models containing all data sources exhibited the most precise estimates for all species. For studies in which only sparse data can be obtained and the strictest model assumptions are unlikely to be met, we suggest researchers use caution making inference from models lacking individual identity. For best results, we further recommend the use of methods requiring at least a subset of the population is marked and that multiple data sets are incorporated when possible.

Evaluation of camera trap-based abundance estimators for unmarked populations

Estimates of species abundance are critical to understand population processes and to assess and select management actions. However, capturing and marking individuals for abundance estimation, while providing robust information, can be economically and logistically prohibitive, particularly for species with cryptic behavior. Camera traps can be used to collect data at temporal and spatial scales necessary for estimating abundance, but the use of camera traps comes with limitations when target species are not uniquely identifiable (i.e., "unmarked"). Abundance estimation is particularly useful in the management of invasive species, with herpetofauna being recognized as some of the most pervasive and detrimental invasive vertebrate species. However, the use of camera traps for these taxa presents additional challenges with relevancy across multiple taxa. It is often necessary to use lures to attract animals in order to obtain sufficient observations, yet lure attraction can influence species' landscape use and potentially induce bias in abundance estimators. We investigated these challenges and assessed the feasibility of obtaining reliable abundance estimates using camera-trapping data on a population of invasive brown treesnakes (Boiga irregularis) in Guam. Data were collected using camera traps in an enclosed area where snakes were subject to high-intensity capture-recapture effort, resulting in presumed abundance of 116 snakes (density = 23/ha). We then applied spatial count, random encounter and staying time, space to event, and instantaneous sampling estimators to photo-capture data to estimate abundance and compared estimates to our presumed abundance. We found that all estimators for unmarked populations performed poorly, with inaccurate or imprecise abundance estimates that limit their usefulness for management in this system. We further investigated the sensitivity of these estimators to the use of lures (i.e., violating the assumption that animal behavior is unchanged by sampling) and camera density in a simulation study. Increasing the effective distances of a lure (i.e., lure attraction) and camera density both resulted in biased abundance estimates. Each estimator rarely recovered truth or suffered from convergence issues. Our results indicate that, when limited to unmarked estimators and the use of lures, camera traps alone are unlikely to produce abundance estimates with utility for brown treesnake management.

Estimating preharvest density, adult sex ratio, and fecundity of white-tailed deer using noninvasive sampling techniques

Adult sex ratio and fecundity (juveniles per female) are key population parameters in sustainable wildlife management, but inferring these requires abundance estimates of at least three age/sex classes of the population (male and female adults and juveniles). Prior to harvest, we used an array of 36 wildlife camera traps during 2 and 3 weeks in the early autumn of 2016 and 2017, respectively. We recorded white-tailed deer adult males, adult females, and fawns from the pictures. Simultaneously, we collected fecal DNA (fDNA) from 92 20 m × 20 m plots placed in 23 clusters of four plots between the camera traps. We identified individuals from fDNA samples with microsatellite markers and estimated the total sex ratio and population density using spatial capture-recapture (SCR). The fDNA-SCR analysis concluded equal sex ratio in the first year and female bias in the second year, and no difference in space use between sexes (fawns and adults combined). Camera information was analyzed in a spatial capture (SC) framework assuming an informative prior for animals' space use, either (a) as estimated by fDNA-SCR (same for all age/sex classes), (b) as assumed from the literature (space use of adult males larger than adult females and fawns), or (c) by inferring adult male space use from individually identified males from the camera pictures. These various SC approaches produced plausible inferences on fecundity, but also inferred total density to be lower than the estimate provided by fDNA-SCR in one of the study years. SC approaches where adult male and female were allowed to differ in their space use suggested the population had a female-biased adult sex ratio. In conclusion, SC approaches allowed estimating the preharvest population parameters of interest and provided conservative density estimates.

Tapirs in trouble: estimating Baird's tapir densities in the Sierra Madre de Chiapas, Mexico

Spatial capture–recapture models have been widely used to estimate densities of species where individuals can be uniquely identified, but alternatives have been developed for estimation of densities for unmarked populations. In this study we used camera-trap records from 2018 to estimate densities of a species that does not always have individually identifiable marks, Baird's tapir Tapirus bairdii, in the Sierra Madre de Chiapas, southern Mexico. We compared the performance of the spatial capture–recapture model with spatial mark–resight and random encounter models. The density of Baird's tapir did not differ significantly between the three models. The estimate of density was highest using the random encounter model (26/100 km2, 95% CI 12–41) and lowest using the capture–recapture model (8/100 km2, 95% CI 4–16). The estimate from the spatial mark–resight model was 10/100 km2 (95% CI 8–14), which had the lowest coefficient of variation, indicating a higher precision than with the other models. Using a second set of camera-trap data, collected in 2015–2016, we created occupancy models and extrapolated density to areas with potential occupancy of Baird's tapir, to generate a population estimate for the whole Sierra Madre de Chiapas. Our findings indicate the need to strengthen, and possibly expand, the protected areas of southern Mexico and to develop an action plan to ensure the conservation of Baird's tapir.

Assessing the robustness of time-to-event models for estimating unmarked wildlife abundance using remote cameras

Recently developed methods, including time-to-event and space-to-event models, estimate the abundance of unmarked populations from encounter rates with camera trap arrays, addressing a gap in noninvasive wildlife monitoring. However, estimating abundance from encounter rates relies on assumptions that can be difficult to meet in the field, including random movement, population closure, and an accurate estimate of movement speed. Understanding how these models respond to violation of these assumptions will assist in making them more applicable in real-world settings. We used simulated walk models to test the effects of violating the assumptions of the time-to-event model under four scenarios: (1) incorrectly estimating movement speed, (2) violating closure, (3) individuals moving within simplified territories (i.e., movement restricted to partially overlapping circles), (4) and individuals clustering in preferred habitat. The time-to-event model was robust to closure violations, territoriality, and clustering when cameras were placed randomly. However, the model failed to estimate abundance accurately when movement speed was incorrectly estimated or cameras were placed nonrandomly with respect to habitat. We show that the time-to-event model can provide unbiased estimates of abundance when some assumptions that are commonly violated in wildlife studies are not met. Having a robust method for estimating the abundance of unmarked populations with remote cameras will allow practitioners to monitor a more diverse array of populations noninvasively. With the time-to-event model, placing cameras randomly with respect to animal movement and accurately estimating movement speed allows unbiased estimation of abundance. The model is robust to violating the other assumptions we tested.

Building a better baseline to estimate 160 years of avian population change and create historically informed conservation targets

Globally, anthropogenic land-cover change has been dramatic over the last few centuries and is frequently invoked as a major cause of wildlife population declines. Baseline data currently used to assess population trends, however, began well after major changes to the landscape. In the United States and Canada, breeding bird population trends are assessed by the North American Breeding Bird Survey, which began in the 1960s. Estimates of distribution and abundance prior to major habitat alteration would add historical perspective to contemporary trends and allow for historically based conservation targets. We used a hindcasting framework to estimate change in distribution and abundance of 7 bird species in the Willamette Valley, Oregon (United States). After reconciling classification schemes of current and 1850s reconstructed land cover, we used multiscale species distribution models and hierarchical distance sampling models to predict spatially explicit densities in the modern and historical landscapes. We estimated that since the 1850s, White-breasted Nuthatch (Sitta carolinensis) and Western Meadowlark (Sturnella neglecta) populations, 2 species sensitive to fragmentation of oak woodlands and grasslands, declined by 93% and 97%, respectively. Five other species we estimated nearly stable or increasing populations, despite steep regional declines since the 1960s. Based on these estimates, we developed historically based conservation targets for amount of habitat, population, and density for each species. Hindcasted reconstructions provide historical perspective for assessing contemporary trends and allow for historically based conservation targets that can inform current management.

The Rapid Rise of Next-Generation Natural History

Many ecologists have lamented the demise of natural history and have attributed this decline to a misguided view that natural history is outdated and unscientific. Although there is a perception that the focus in ecology and conservation have shifted away from descriptive natural history research and training toward hypothetico-deductive research, we argue that natural history has entered a new phase that we call “next-generation natural history.” This renaissance of natural history is characterized by technological and statistical advances that aid in collecting detailed observations systematically over broad spatial and temporal extents. The technological advances that have increased exponentially in the last decade include electronic sensors such as camera-traps and acoustic recorders, aircraft- and satellite-based remote sensing, animal-borne biologgers, genetics and genomics methods, and community science programs. Advances in statistics and computation have aided in analyzing a growing quantity of observations to reveal patterns in nature. These robust next-generation natural history datasets have transformed the anecdotal perception of natural history observations into systematically collected observations that collectively constitute the foundation for hypothetico-deductive research and can be leveraged and applied to conservation and management. These advances are encouraging scientists to conduct and embrace detailed descriptions of nature that remain a critically important component of the scientific endeavor. Finally, these next-generation natural history observations are engaging scientists and non-scientists alike with new documentations of the wonders of nature. Thus, we celebrate next-generation natural history for encouraging people to experience nature directly.

Investigation of REFINED CNN ensemble learning for anti-cancer drug sensitivity prediction

Motivation

Anti-cancer drug sensitivity prediction using deep learning models for individual cell line is a significant challenge in personalized medicine. Recently developed REFINED (REpresentation of Features as Images with NEighborhood Dependencies) CNN (Convolutional Neural Network)-based models have shown promising results in improving drug sensitivity prediction. The primary idea behind REFINED-CNN is representing high dimensional vectors as compact images with spatial correlations that can benefit from CNN architectures. However, the mapping from a high dimensional vector to a compact 2D image depends on the a priori choice of the distance metric and projection scheme with limited empirical procedures guiding these choices.

Results

In this article, we consider an ensemble of REFINED-CNN built under different choices of distance metrics and/or projection schemes that can improve upon a single projection based REFINED-CNN model. Results, illustrated using NCI60 and NCI-ALMANAC databases, demonstrate that the ensemble approaches can provide significant improvement in prediction performance as compared to individual models. We also develop the theoretical framework for combining different distance metrics to arrive at a single 2D mapping. Results demonstrated that distance-averaged REFINED-CNN produced comparable performance as obtained from stacking REFINED-CNN ensemble but with significantly lower computational cost.

Availability and implementation

The source code, scripts, and data used in the paper have been deposited in GitHub (https://github.com/omidbazgirTTU/IntegratedREFINED).

Supplementary information

Supplementary data are available at Bioinformatics online.

Spatially clustered count data provide more efficient search strategies in invasion biology and disease control

Geographic profiling, a mathematical model originally developed in criminology, is increasingly being used in ecology and epidemiology. Geographic profiling boasts a wide range of applications, such as finding source populations of invasive species or breeding sites of vectors of infectious disease. The model provides a cost-effective approach for prioritizing search strategies for source locations and does so via simple data in the form of the positions of each observation, such as individual sightings of invasive species or cases of a disease. In doing so, however, classic geographic profiling approaches fail to make the distinction between those areas containing observed absences and those areas where no data were recorded. Absence data are generated via spatial sampling protocols but are often discarded during the inference process. Here we construct a geographic profiling model that resolves these issues by making inferences via count data, analyzing a set of discrete sentinel locations at which the number of encounters has been recorded. Crucially, in our model this number can be zero. We verify the ability of this new model to estimate source locations and other parameters of practical interest via a Bayesian power analysis. We also measure model performance via real-world data in which the model infers breeding locations of mosquitoes in bromeliads in Miami-Dade County, Florida, USA. In both cases, our novel model produces more efficient search strategies by shifting focus from those areas containing observed absences to those with no data, an improvement over existing models that treat these areas equally. Our model makes important improvements upon classic geographic profiling methods, which will significantly enhance real-world efforts to develop conservation management plans and targeted interventions.

Review: COVID-19 highlights the importance of camera traps for wildlife conservation research and management

COVID-19 has altered many aspects of everyday life. For the scientific community, the pandemic has called upon investigators to continue work in novel ways, curtailing field and lab research. However, this unprecedented situation also offers an opportunity for researchers to optimize and further develop available field methods. Camera traps are one example of a tool used in science to answer questions about wildlife ecology, conservation, and management. Camera traps have long battery lives, lasting more than a year in certain cases, and photo storage capacity, with some models capable of wirelessly transmitting images from the field. This allows researchers to deploy cameras without having to check them for up to a year or more, making them an ideal field research tool during restrictions on in-person research activities such as COVID-19 lockdowns. As technological advances allow cameras to collect increasingly greater numbers of photos and videos, the analysis techniques for large amounts of data are evolving. Here, we describe the most common research questions suitable for camera trap studies and their importance for biodiversity conservation. As COVID-19 continues to affect how people interact with the natural environment, we discuss novel questions for which camera traps can provide insights on. We conclude by summarizing the results of a systematic review of camera trap studies, providing data on target taxa, geographic distribution, publication rate, and publication venues to help researchers planning to use camera traps in response to the current changes in human activity.

Next-Generation Camera Trapping: Systematic Review of Historic Trends Suggests Keys to Expanded Research Applications in Ecology and Conservation

Camera trapping is an effective non-invasive method for collecting data on wildlife species to address questions of ecological and conservation interest. We reviewed 2,167 camera trap (CT) articles from 1994 to 2020. Through the lens of technological diffusion, we assessed trends in: (1) CT adoption measured by published research output, (2) topic, taxonomic, and geographic diversification and composition of CT applications, and (3) sampling effort, spatial extent, and temporal duration of CT studies. Annual publications of CT articles have grown 81-fold since 1994, increasing at a rate of 1.26 (SE = 0.068) per year since 2005, but with decelerating growth since 2017. Topic, taxonomic, and geographic richness of CT studies increased to encompass 100% of topics, 59.4% of ecoregions, and 6.4% of terrestrial vertebrates. However, declines in per article rates of accretion and plateaus in Shannon's H for topics and major taxa studied suggest upper limits to further diversification of CT research as currently practiced. Notable compositional changes of topics included a decrease in capture-recapture, recent decrease in spatial-capture-recapture, and increases in occupancy, interspecific interactions, and automated image classification. Mammals were the dominant taxon studied; within mammalian orders carnivores exhibited a unimodal peak whereas primates, rodents and lagomorphs steadily increased. Among biogeographic realms we observed decreases in Oceania and Nearctic, increases in Afrotropic and Palearctic, and unimodal peaks for Indomalayan and Neotropic. Camera days, temporal extent, and area sampled increased, with much greater rates for the 0.90 quantile of CT studies compared to the median. Next-generation CT studies are poised to expand knowledge valuable to wildlife ecology and conservation by posing previously infeasible questions at unprecedented spatiotemporal scales, on a greater array of species, and in a wider variety of environments. Converting potential into broad-based application will require transferable models of automated image classification, and data sharing among users across multiple platforms in a coordinated manner. Further taxonomic diversification likely will require technological modifications that permit more efficient sampling of smaller species and adoption of recent improvements in modeling of unmarked populations. Environmental diversification can benefit from engineering solutions that expand ease of CT sampling in traditionally challenging sites.

Community-level modelling of boreal forest mammal distribution in an oil sands landscape

Anthropogenic landscape disturbances are known to alter, destroy, and fragment habitat, which typically leads to biodiversity loss. The effects of landscape disturbance generally vary among species and depend on the nature of the disturbances, which may interact and result in synergistic effects. Western Canada's oil sands region experiences disturbances from forestry and energy sector activities as well as municipal and transportation infrastructure. The effects of those disturbances on single species have been studied and have been implicated in declines of the boreal woodland caribou (Rangifer tarandus caribou). Yet, the specific responses of the mammal community, and of functional groups such as prey and predators, to those interacting disturbances are still poorly known. We investigated the responses of black bear, grey wolf, coyote, fisher, lynx, red fox, American red squirrel, white-tailed deer, moose, caribou, and snowshoe hare to both natural habitat and disturbance associated with anthropogenic features within Alberta's northeast boreal forest. We used a novel community-level modelling framework on three years of camera-trap data collected in an oil sands landscape. This framework allowed us to identify the natural and anthropogenic features which explained the most variation in occurrence frequency among functional groups, as well as compare responses to linear and non-linear anthropogenic disturbance. Occurrence frequency by predators was better explained by anthropogenic features than by natural habitat. Both linear and non-linear anthropogenic features helped explain occurrence frequency by prey and predators, although the effects differed in magnitude and spatial scale. To better conserve boreal biodiversity, management actions should extend beyond a focus on caribou and wolves and aim to restore habitat across a diversity of anthropogenic disturbances and monitor the dynamics of the entire mammal community.

Augustine B, Linden DW, B. Chandler R, Royle JA. Spatial capture-recapture with random thinning for unidentified encounters

Spatial capture-recapture (SCR) models have increasingly been used as a basis for combining capture-recapture data types with variable levels of individual identity information to estimate population density and other demographic parameters. Recent examples are the unmarked SCR (or spatial count model), where no individual identities are available and spatial mark-resight (SMR) where individual identities are available for only a marked subset of the population. Currently lacking, though, is a model that allows unidentified samples to be combined with identified samples when there are no separate classes of "marked" and "unmarked" individuals and when the two sample types cannot be considered as arising from two independent observation models. This is a common scenario when using noninvasive sampling methods, for example, when analyzing data on identified and unidentified photographs or scats from the same sites.Here we describe a "random thinning" SCR model that utilizes encounters of both known and unknown identity samples using a natural mechanistic dependence between samples arising from a single observation model. Our model was fitted in a Bayesian framework using NIMBLE.We investigate the improvement in parameter estimates by including the unknown identity samples, which was notable (up to 79% more precise) in low-density populations with a low rate of identified encounters. We then applied the random thinning SCR model to a noninvasive genetic sampling study of brown bear (Ursus arctos) density in Oriental Cantabrian Mountains (North Spain).Our model can improve density estimation for noninvasive sampling studies for low-density populations with low rates of individual identification, by making use of available data that might otherwise be discarded.

Abundance estimation of unmarked animals based on camera-trap data

The rapid improvement of camera traps in recent decades has revolutionized biodiversity monitoring. Despite clear applications in conservation science, camera traps have seldom been used to model the abundance of unmarked animal populations. We sought to summarize the challenges facing abundance estimation of unmarked animals, compile an overview of existing analytical frameworks, and provide guidance for practitioners seeking a suitable method. When a camera records multiple detections of an unmarked animal, one cannot determine whether the images represent multiple mobile individuals or a single individual repeatedly entering the camera viewshed. Furthermore, animal movement obfuscates a clear definition of the sampling area and, as a result, the area to which an abundance estimate corresponds. Recognizing these challenges, we identified 6 analytical approaches and reviewed 927 camera-trap studies published from 2014 to 2019 to assess the use and prevalence of each method. Only about 5% of the studies used any of the abundance-estimation methods we identified. Most of these studies estimated local abundance or covariate relationships rather than predicting abundance or density over broader areas. Next, for each analytical approach, we compiled the data requirements, assumptions, advantages, and disadvantages to help practitioners navigate the landscape of abundance estimation methods. When seeking an appropriate method, practitioners should evaluate the life history of the focal taxa, carefully define the area of the sampling frame, and consider what types of data collection are possible. The challenge of estimating abundance of unmarked animal populations persists; although multiple methods exist, no one method is optimal for camera-trap data under all circumstances. As analytical frameworks continue to evolve and abundance estimation of unmarked animals becomes increasingly common, camera traps will become even more important for informing conservation decision-making.

A novel modelling framework to explicitly simulate predator interaction with poison baits

Abstract ContextManagement of human–wildlife conflicts is of critical importance for both wildlife conservation and agricultural production. Population models are commonly used to simulate population dynamics and their responses to management actions. However, it is essential that this class of models captures the drivers and mechanisms necessary to reliably forecast future system dynamics. AimsWe aimed to develop a flexible modelling framework with the capacity to explicitly simulate individual interactions with baits (with or without the presence of other management tools), for which parameter estimates from field data are available. We also intended for the model to potentially accommodate multi-species interaction and avoidance behaviours. MethodsWe expanded an existing spatially explicit, individual-based model to directly simulate bait deployment, animal movements and bait consumption. We demonstrated the utility of this model using a case study from Western Australia where we considered two possible exclusion-fence scenarios, namely, the completion of a landscape-scale and smaller-scale fences. Within each of these proposed cells, using data obtained from a camera-trap study, we evaluated the performance of two levels of baiting to control wild dogs (Canis familiaris), in contrast with the option of no control. ResultsThe present study represents a substantial step forward in accurately modelling predator dynamics. When applying our model to the case study, for example, it was straightforward to investigate whether outcomes were sensitive to the bait-encounter probability. We could further explore interactions between baiting regimes and different fence designs and demonstrate how wild dog eradication could be achieved in the smaller cell under the more intense control scenarios. In contrast, the landscape-scale fence had only minor effects unless it was implemented as a preventive measure in an area where wild dogs were not already established. ConclusionsThe new component of the model presented here provides fine-scale control of single components of individual–bait interactions. ImplicationsThe effect of management actions (e.g. lures) that affect this process can be easily investigated. Multi-species modelling and avoidance behaviours can readily be implemented, making the present study widely relevant for a range of contexts such as multi-species competition or non-target bait uptake.

Population Estimate, Habitat-Use and Activity Patterns of the Honey Badger in a Dry-Deciduous Forest of Central India

Studies on carnivores are skewed toward larger species in India, limiting ecological information of the smaller ones. Basic ecological understanding like population density, distribution, habitat-use patterns of small carnivores is lacking. This inadequate knowledge has led to disagreement between conservation approaches in different landscapes. Honey badgers (Mellivora capensis) are cryptic carnivores distributed across large areas of Africa and Asia; however, fundamental ecological knowledge is scarce. The species is thought to exist at low population densities throughout its range. We used a large camera trap dataset from a tiger reserve in Maharashtra State, India to understand the population density, habitat preference, and diel activity pattern of the species. We applied an extension of the spatial count model for the estimation of population. Habitat preference analyses were carried out using generalized linear models and activity patterns were analyzed using kernel-density functions. The population density was estimated as 14.09 (95% CI 10–22.25) individuals per 100 km2. Habitat use revealed a positive association with forest cover and negative association with elevation. This may expose the species to other large carnivores in the habitat but honey badger activity pattern peaked at midnight retaining minimum temporal overlap with other large carnivores (e.g., tiger Panthera tigris, leopard Panthera pardus, and dhole Cuon alpinus) and moderate overlap with small carnivores (e.g., jungle cat Felis chaus, rusty-spotted cat Prionailurus rubiginosus). These behaviors, in turn, may facilitate the coexistence of species at such high density even with high carnivore density. We hope the findings of this study will fill the existing knowledge gap of this species and aid in guiding the conservation of the species in other landscapes and reserves.

Representation of features as images with neighborhood dependencies for compatibility with convolutional neural networks

Deep learning with Convolutional Neural Networks has shown great promise in image-based classification and enhancement but is often unsuitable for predictive modeling using features without spatial correlations. We present a feature representation approach termed REFINED (REpresentation of Features as Images with NEighborhood Dependencies) to arrange high-dimensional vectors in a compact image form conducible for CNN-based deep learning. We consider the similarities between features to generate a concise feature map in the form of a two-dimensional image by minimizing the pairwise distance values following a Bayesian Metric Multidimensional Scaling Approach. We hypothesize that this approach enables embedded feature extraction and, integrated with CNN-based deep learning, can boost the predictive accuracy. We illustrate the superior predictive capabilities of the proposed framework as compared to state-of-the-art methodologies in drug sensitivity prediction scenarios using synthetic datasets, drug chemical descriptors as predictors from NCI60, and both transcriptomic information and drug descriptors as predictors from GDSC.

Convolutional Neural Networks (CNN) are often unsuitable for predictive modeling involving nonimage based biological features. Here, the authors present a mapping termed REFINED to represent high dimensional vectors as compact images with spatial correlation that makes it compatible with CNN based learning.

Methods for Monitoring Large Terrestrial Animals in the Wild

Reliable information about wildlife is absolutely important for making informed management decisions. The issues with the effectiveness of the control and monitoring of both large and small wild animals are relevant to assess and protect the world’s biodiversity. Monitoring becomes part of the methods in wildlife ecology for observation, assessment, and forecasting of the human environment. World practice reveals the potential of the joint application of both proven traditional and modern technologies using specialized equipment to organize environmental control and management processes. Monitoring large terrestrial animals require an individual approach due to their low density and larger habitat. Elk/moose are such animals. This work aims to evaluate the methods for monitoring large wild animals, suitable for controlling the number of elk/moose in the framework of nature conservation activities. Using different models allows determining the population size without affecting the animals and without significant financial costs. Although, the accuracy of each model is determined by its postulates implementation and initial conditions that need statistical data. Depending on the geographical, climatic, and economic conditions in each territory, it is possible to use different tools and equipment (e.g., cameras, GPS sensors, and unmanned aerial vehicles), a flexible variation of which will allow reaching the golden mean between the desires and capabilities of researchers.

Spatial proximity moderates genotype uncertainty in genetic tagging studies

Accelerating declines of an increasing number of animal populations worldwide necessitate methods to reliably and efficiently estimate demographic parameters such as population density and trajectory. Standard methods for estimating demographic parameters from noninvasive genetic samples are inefficient because lower-quality samples cannot be used, and they assume individuals are identified without error. We introduce the genotype spatial partial identity model (gSPIM), which integrates a genetic classification model with a spatial population model to combine both spatial and genetic information, thus reducing genotype uncertainty and increasing the precision of demographic parameter estimates. We apply this model to data from a study of fishers (Pekania pennanti) in which 37% of hair samples were originally discarded because of uncertainty in individual identity. The gSPIM density estimate using all collected samples was 25% more precise than the original density estimate, and the model identified and corrected three errors in the original individual identity assignments. A simulation study demonstrated that our model increased the accuracy and precision of density estimates 63 and 42%, respectively, using three replicated assignments (e.g., PCRs for microsatellites) per genetic sample. Further, the simulations showed that the gSPIM model parameters are identifiable with only one replicated assignment per sample and that accuracy and precision are relatively insensitive to the number of replicated assignments for high-quality samples. Current genotyping protocols devote the majority of resources to replicating and confirming high-quality samples, but when using the gSPIM, genotyping protocols could be more efficient by devoting more resources to low-quality samples.

Zooming in on mechanistic predator-prey ecology: Integrating camera traps with experimental methods to reveal the drivers of ecological interactions

Camera trap technology has galvanized the study of predator-prey ecology in wild animal communities by expanding the scale and diversity of predator-prey interactions that can be analysed. While observational data from systematic camera arrays have informed inferences on the spatiotemporal outcomes of predator-prey interactions, the capacity for observational studies to identify mechanistic drivers of species interactions is limited. Experimental study designs that utilize camera traps uniquely allow for testing hypothesized mechanisms that drive predator and prey behaviour, incorporating environmental realism not possible in the laboratory while benefiting from the distinct capacity of camera traps to generate large datasets from multiple species with minimal observer interference. However, such pairings of camera traps with experimental methods remain underutilized. We review recent advances in the experimental application of camera traps to investigate fundamental mechanisms underlying predator-prey ecology and present a conceptual guide for designing experimental camera trap studies. Only 9% of camera trap studies on predator-prey ecology in our review use experimental methods, but the application of experimental approaches is increasing. To illustrate the utility of camera trap-based experiments using a case study, we propose a study design that integrates observational and experimental techniques to test a perennial question in predator-prey ecology: how prey balance foraging and safety, as formalized by the risk allocation hypothesis. We discuss applications of camera trap-based experiments to evaluate the diversity of anthropogenic influences on wildlife communities globally. Finally, we review challenges to conducting experimental camera trap studies. Experimental camera trap studies have already begun to play an important role in understanding the predator-prey ecology of free-living animals, and such methods will become increasingly critical to quantifying drivers of community interactions in a rapidly changing world. We recommend increased application of experimental methods in the study of predator and prey responses to humans, synanthropic and invasive species, and other anthropogenic disturbances.

Population density and habitat use of two sympatric small cats in a central Indian reserve

Despite appreciable advances in carnivore ecology, studies on small cats remain limited with carnivore research in India being skewed towards large cats. Small cats are more specialized than their larger cousins in terms of resource selection. Studies on small cat population and habitat preference are critical to evaluate their status to ensure better management and conservation. We estimated abundance of two widespread small cats, the jungle cat, and the rusty-spotted cat, and investigated their habitat associations based on camera trap captures from a central Indian tiger reserve. We predicted fine-scale habitat segregation between these sympatric species as a driver of coexistence. We used an extension of the spatial count model in a Bayesian framework approach to estimate the population density of jungle cat and rusty-spotted cat and used generalized linear models to explore their habitat associations. Densities of rusty-spotted cat and jungle cat were estimated as 6.67 (95% CI 4.07–10.74) and 4.01 (95% CI 2.65–6.12) individuals/100 km² respectively. Forest cover and evapotranspiration were positively associated with rusty-spotted cat occurrence whereas both factors had a significant negative relation with jungle cat occurrence. The results directed habitat segregation between these small cats with affinities of rusty-spotted cat and jungle cat towards well-forested and open scrubland areas respectively. Our estimates highlight the widespread applicability of this model for density estimation of species with no individual identification. Moreover, the study outcomes can aid in targeted management decisions and serve as the baseline for species conservation as these models allow robust population estimation of elusive species along with predicting their habitat preferences.

Estimation of free-roaming domestic dog population size: Investigation of three methods including an Unmanned Aerial Vehicle (UAV) based approach

Population size estimation is performed for several reasons including disease surveillance and control, for example to design adequate control strategies such as vaccination programs or to estimate a vaccination campaign coverage. In this study, we aimed at investigating the possibility of using Unmanned Aerial Vehicles (UAV) to estimate the size of free-roaming domestic dog (FRDD) populations and compare the results with two regularly used methods for population estimations: foot-patrol transect survey and the human: dog ratio estimation. Three studies sites of one square kilometer were selected in Petén department, Guatemala. A door-to-door survey was conducted in which all available dogs were marked with a collar and owner were interviewed. The day after, UAV flight were performed twice during two consecutive days per study site. The UAV's camera was set to regularly take pictures and cover the entire surface of the selected areas. Simultaneously to the UAV's flight, a foot-patrol transect survey was performed and the number of collared and non-collared dogs were recorded. Data collected during the interviews and the number of dogs counted during the foot-patrol transects informed a capture-recapture (CR) model fit into a Bayesian inferential framework to estimate the dog population size, which was found to be 78, 259, and 413 in the three study sites. The difference of the CR model estimates compared to previously available dog census count (110 and 289) can be explained by the fact that the study population addressed by the different methods differs. The human: dog ratio covered the same study population as the dog census and tended to underestimate the FRDD population size (97 and 161). Under the conditions within this study, the total number of dogs identified on the UAV pictures was 11, 96, and 71 for the three regions (compared to the total number of dogs counted during the foot-patrol transects of 112, 354 and 211). In addition, the quality of the UAV pictures was not sufficient to assess the presence of a mark on the spotted dogs. Therefore, no CR model could be implemented to estimate the size of the FRDD using UAV. We discussed ways for improving the use of UAV for this purpose, such as flying at a lower altitude in study area wisely chosen. We also suggest to investigate the possibility of using infrared camera and automatic detection of the dogs to increase visibility of the dogs in the pictures and limit workload of finding them. Finally, we discuss the need of using models, such as spatial capture-recapture models to obtain reliable estimates of the FRDD population. This publication may provide helpful directions to design dog population size estimation methods using UAV.

Multiple observation processes in spatial capture-recapture models: How much do we gain?

Population monitoring data may originate from multiple methods and are often sparse and fraught with incomplete information due to practical and economic constraints. Models that can integrate multiple survey methods and are able to cope with incomplete data may help investigators exploit available information more thoroughly. Here, we developed an integrated spatial capture-recapture (SCR) model to incorporate multiple data sources with imperfect individual identification. We contrast inferences drawn from this model with alternate models incorporating only subsets of the data available. Using extensive simulations and an empirical example of multi-method brown bear (Ursus arctos) monitoring data from northern Pakistan, we quantified the benefits of including multiple sources of information in SCR models in terms of parameter precision and bias. Our multiple observation processes SCR model (MOP) yielded a more complete picture of the underlying processes, reduced bias, and led to more precise parameter estimates. Our results suggest that the greatest gains from integrated SCR models can be expected in situations where detection probability is low, a large proportion of detections is not attributable to individuals, and the degree of overlap between individual home ranges is low.

Density-dependent space use affects interpretation of camera trap detection rates

Camera traps (CTs) are an increasingly popular tool for wildlife survey and monitoring. Estimating relative abundance in unmarked species is often done using detection rate as an index of relative abundance, which assumes that detection rate has a positive linear relationship with true abundance. This assumption may be violated if movement behavior varies with density, but the degree to which movement behavior is density-dependent across taxa is unclear. The potential confounding of population-level relative abundance indices by movement would depend on how regularly, and by what magnitude, movement rate and home-range size vary with density. We conducted a systematic review and meta-analysis to quantify relationships between movement rate, home-range size, and density, across terrestrial mammalian taxa. We then simulated animal movements and CT sampling to test the effect of contrasting movement scenarios on CT detection rate indices. Overall, movement rate and home-range size were negatively correlated with density and positively correlated with one another. The strength of the relationships varied significantly between taxa and populations. In simulations, detection rates were related to true abundance but underestimated change, particularly for slower moving species with small home ranges. In situations where animal space use changes markedly with density, we estimate that up to thirty percent of a true change in relative abundance may be missed due to the confounding effect of movement, making trend estimation more difficult. The common assumption that movement remains constant across densities is therefore violated across a wide range of mammal species. When studying unmarked species using CT detection rates, researchers and managers should explicitly consider that such indices of relative abundance reflect both density and movement. Practitioners interpreting changes in camera detection rates should be aware that observed differences may be biased low relative to true changes in abundance. Further information on animal movement, or methods that do not depend on assumptions of density-independent movement, may be required to make robust inferences on population trends.

Avoiding misleading messages: Population assessment using camera trapping is not a simple task

Population assessment is indispensable for appropriate and socially acceptable conservation and management of wildlife populations. This article critiques the paper by Campos-Candela et al. 2018 and highlights issues that could lead to inappropriate management and conservation policies.