Imperfect Detection Research Articles

Identifying factors linked with persistence of reintroduced populations: Lessons learned from 25 years of amphibian translocations

Conservation translocations are increasingly used to help recover imperiled species. However, success of establishing populations remains low, especially for amphibians. Identifying factors associated with translocation success can help increase efficiency and efficacy of recovery efforts. Since the 1990s, several captive and semi-captive facilities have produced Chiricahua Leopard Frogs ( Rana chiricahuensis ) to establish or augment wild populations in Arizona and New Mexico, USA. During this same time, personnel associated with several programs surveyed translocation and non-translocation sites for presence of amphibians. We used 25 years (1995–2019) of survey and translocation data for the federally threatened Chiricahua Leopard Frog to identify factors linked with population persistence. Our dataset included approximately 40,642 egg masses or animals translocated in 314 events to 115 distinct sites and > 5800 visual encounter surveys from 641 sites; 120 of these sites were also surveyed with environmental DNA methods in 2018. We used a hierarchical dynamic occupancy model that accounted for imperfect detection to identify patch- and landscape-level attributes associated with site occupancy, and then used predictions from that model to evaluate factors associated with population persistence at translocation sites. Across all sites, extinction probability for Chiricahua Leopard Frogs was higher in lotic (stream) than lentic (pond) habitats and when Western Tiger Salamanders ( Ambystoma mavortium ) were present. Restoration of sites specifically for frog conservation reduced extinction probability. Colonization of unoccupied sites increased moderately with increasing numbers of translocation sites within 2 km, indicating a benefit of translocation efforts beyond sites where frogs were stocked. At translocation sites, persistence was greater in lentic than lotic habitats and was negatively correlated with the proportion of years tiger salamanders were present. Increasing numbers of translocation events, especially of late-stage larvae, increased persistence. There was little difference in population persistence based on whether stock was from captive, semi-captive, or wild sources, but translocations during the dry season (January July) succeeded more than those after the typical arrival of summer rains (August December). Based on the number of years translocation sites were predicted to be occupied, 2 or more translocations produced, on average, a > 4-yr increase in predicted occupancy compared to sites without translocations. While translocations have increased the number of populations across the landscape, continued management of water availability and threats such as invasive predators and disease remain critical to recovery of the Chiricahua Leopard Frog. • Most translocation efforts are unsuccessful, often for unknown reasons. • We assessed factors linked with population persistence for 25 years of translocations. • Local features were paramount, including habitat, predators, and restoration history. • Timing and life stages stocked affected persistence, but rearing environment did not. • ≥ 2 translocations produced a > 4-yr increase in predicted population persistence.

Drivers of occupancy patterns for the red fox, Vulpes vulpes, in Mediterranean Eucalyptus plantations

The replacement of natural areas with forestry plantations is a worldwide expanding process with direct consequences for biodiversity and ecosystem functionality. In the Mediterranean region, Eucalyptus spp. plantations are widespread, forming monospecific landscapes that in Portugal dominate most of its forested areas. The reduction in the availability of native habitats induces important challenges to native wildlife, namely changes in habitat use patterns and behavior. In this study, we evaluated the influence of Eucalyptus globulus exotic plantations on the occupancy patterns of the red fox (Vulpes vulpes L.), a mesocarnivore widely distributed in the Palearctic, and we examined potential drivers shaping those patterns. We conducted camera trapping surveys in Central Portugal (Southwestern Europe), in eight 16 km2 grids, and analyzed the resulting occurrence data accounting for imperfect detection. Our analysis revealed a preference of the red fox for native vegetation over non-native plantations and avoidance of areas with higher human disturbance. Our data suggest that the current structure of exotic plantations can have a negative impact on species occurrence, even for generalist and resilient species such as the red fox. By gaining insight into landscape structures that promotes fox occupancy, our research can contribute to the development and implementation of more integrative management measures aiming to promote the presence and conservation of mesocarnivores in Eucalyptus dominated landscapes while ensuring sustainable exploitation of these plantations.

Using helicopter counts to estimate the abundance of Himalayan tahr in New Zealand's Southern Alps

AbstractEstimating the abundance and density of mountain ungulates is difficult because of rugged and remote terrain, high elevations, and rapidly changing weather. Helicopter surveys could overcome these problems, but researchers have seldom applied helicopter‐based survey methods at large spatial scales in mountain terrain. We used helicopters to count introduced Himalayan tahr (Hemitragus jemlahicus) at 117 plots, each of 4 km2, in New Zealand's Southern Alps during 2016–2019. The sampling frame was 7,844 km2 and we located the plots at the vertices of an 8‐km grid superimposed over the sampling frame (i.e., a systematic random sampling design). We conducted 3 repeat counts at each plot during summer–autumn. We used the repeat counts to estimate tahr abundance and density, corrected for imperfect detection, using a dynamic N‐mixture model for open populations. We estimated the population of tahr in the sampling frame using design‐based, finite sampling methods and model‐based inference procedures. The mean estimated density of tahr on each plot varied from zero to 31.7 tahr/km2. The mean densities of tahr varied among management units, ranging from 0.3 to 10.7 tahr/km2, and exceeded specified intervention densities in 6 of the 7 management units. The total design‐based estimate of tahr abundance in the sampling frame was 34,500 (95% CI = 27,750–42,900), with a coefficient of variation (CV) of 0.11. The corresponding model‐based estimate of total abundance was similar (34,550, 95% CI = 30,250–38,700) but was substantially more precise (CV = 0.06) than the design‐based estimate. The precision of the estimates for the individual management units was also better than that of the design‐based estimates, with CVs of <0.20 for all but 1 management unit. Our study provides a repeatable method for sampling mountain ungulates. More generally, robust estimation of abundance and density of mountain ungulates is possible by combining aerial surveys and open population models with an objective, probabilistic sampling design.

Estimating Road Mortality Hotspots While Accounting for Imperfect Detection: A Case Study with Amphibians and Reptiles

Wildlife road mortality tends to aggregate spatially at locations commonly referred to as road mortality hotspots. Predictive models can be used to identify locations appropriate for mitigation measures that reduce road mortality. However, the influence of imperfect detection (e.g., false absences) during road mortality surveys can lead to inaccurate or imprecise spatial patterns of road mortality hotspots and suboptimal implementation of mitigation measures. In this research, we used amphibians and reptiles as a case study to address imperfect detection issues when estimating the probability of road mortality hotspots using occupancy detection modeling. In addition, we determined the survey effort needed to achieve a high probability of detecting large roadkill events. We also assessed whether vehicle travel reductions associated with the COVID-19 pandemic travel restrictions led to reductions in road mortality. We conducted surveys at 48 sites throughout Rhode Island, USA, from 2019–2021. In total, we observed 657 carcasses representing 19 of Rhode Island’s 37 native species. Of the 19 native species, eight species of frogs, four species of salamanders, four species of snakes, and three species of turtles were observed. We documented a reduction in roadkill density and the proportion of dead versus live amphibians and reptiles in pandemic years (2020 and 2021), but we were unable to link reductions in roadkill density to reductions in traffic volume. Our model results indicated that large roadkill events were more likely to occur on roads near wetlands and with low traffic volume and were more likely to be detected as daily precipitation increased. We determined that there was a low probability of detecting large roadkill events, suggesting that imperfect detection influences detection of large roadkill events, and many were likely missed during our surveys. Therefore, we recommend using occupancy modeling to account for the influence of imperfect detection when estimating road mortality hotspots. This approach will more effectively guide the implementation of mitigation measures.

Inhomogeneous Poisson point process for species distribution modelling: relative performance of methods accounting for sampling bias and imperfect detection

Inhomogeneous Poisson point process for species distribution modelling: relative performance of methods accounting for sampling bias and imperfect detection

Integrating Public Health Surveillance and Environmental Data to Model Presence of Histoplasma in the United States.

Background:In the United States, the true geographic distribution of the environmental fungus Histoplasma capsulatum remains poorly understood but appears to have changed since it was first characterized. Histoplasmosis is caused by inhalation of the fungus and can range in severity from asymptomatic to life threatening. Due to limited public health surveillance and under detection of infections, it is challenging to directly use reported case data to characterize spatial risk.Methods:Using monthly and yearly county-level public health surveillance data and various environmental and socioeconomic characteristics, we use a spatio-temporal occupancy model to estimate latent, or unobserved, presence of H. capsulatum, accounting for imperfect detection of histoplasmosis cases.Results:We estimate areas with higher probabilities of the presence of H. capsulatum in the East North Central states around the Great Lakes, reflecting a shift of the endemic region to the north from previous estimates. The presence of H. capsulatum was strongly associated with higher soil nitrogen levels.Conclusions:In this investigation, we were able to mitigate challenges related to reporting and illustrate a shift in the endemic region from historical estimates. This work aims to help inform future surveillance needs, clinical awareness, and testing decisions for histoplasmosis.

Estimating species misclassification with occupancy dynamics and encounter rates: A semi‐supervised, individual‐level approach

Abstract Large‐scale, long‐term biodiversity monitoring is essential to conservation, land management and identifying threats to biodiversity. However, multispecies surveys are prone to various types of observation error, including false‐positive/false‐negative detection and misclassification, where a species is thought to have been encountered but not correctly identified. Previous methods assume an imperfect classifier produces species‐level classifications, but in practice, particularly with human observers, we may end up with extraspecific classifications including ‘unknown’, morphospecies designations and taxonomic identifications coarser than species. Disregarding these types of species misclassification in biodiversity monitoring datasets can bias estimates of ecologically important quantities such as demographic rates, occurrence and species richness. Here we present a joint classification‐occupancy model that accounts for species non‐detection and misclassification. Our framework accommodates extinction and colonization dynamics, allows for additional uncertain ‘morphospecies’ designations and makes use of individual specimens with known species identities in a semi‐supervised setting. We compare the performance of our model to a classification‐only model that discards information about occupancy and encounter rate. We illustrate our model with an empirical case study of the carabid beetle (Carabidae) community at the National Ecological Observatory Network Niwot Ridge Mountain Research Station, near Boulder, CO, USA. We also use simulations to evaluate model performance through validation metrics where varying fractions of the data are confirmed. The model supported imperfect classifier accuracy and favoured certain true species classifications strongly for some morphospecies. The model outperformed (e.g. precision) the reduced model that discarded occupancy information, and these differences were most pronounced for abundant species. Spatial and temporal dynamics from modelled occupancy and encounter rates may inform species misclassification probability, but this idea has not yet been tested. Our statistical framework explores this opportunity, and can be applied to datasets with imperfect species detection and classification, limited verification data and non‐species classifications.

Dynamic occupancy modelling to determine the status of a Critically Endangered lizard

AbstractMonitoring of cryptic or threatened species poses challenges for population assessment and conservation, as imperfect detection gives rise to misleading inferences about population status. We used a dynamic occupancy model that explicitly accounted for occupancy, colonization, local extinction and detectability to assess the status of the endemic Critically Endangered Bermuda skink Plestiodon longirostris. During 2015–2017, skinks were detected at 13 of 40 surveyed sites in Bermuda, two of which were new records. Ten observation-level and site-specific covariates were used to explore drivers of occupancy, colonization, extinction and detectability. Sites occupied by skinks tended to be islands with rocky coastal habitat and prickly pear cacti; the same variables were also associated with reduced risk of local extinction. The presence of seabirds appeared to encourage colonization, whereas the presence of cats had the opposite effect. The probability of detection was p = 0.45, and on average, five surveys were needed to reliably detect the presence of skinks with 95% certainty. However, skinks were unlikely to be detected on sites with cat and rat predators. Dynamic occupancy models can be used to elucidate drivers of occupancy dynamics, which in turn can inform species conservation management. The survey effort needed to determine population changes over time can be derived from estimates of detectability.

Quantifying the relationship between prey density, livestock and illegal killing of leopards

AbstractMany large mammalian carnivores are facing population declines due to illegal killing (e.g. shooting) and habitat modification (e.g. livestock farming). Illegal killing occurs cryptically and hence is difficult to detect. However, reducing illegal killing requires a solid understanding of its magnitude and underlying drivers, while accounting for the imperfect detection of illegal killing events. Despite the importance of illegal killing of large carnivores in comparison with other causes of mortality, its relationship with potential drivers such as livestock density and wild prey abundance is rarely described.Using ranger‐collected data (2007–2019) of leopard killing events and data on covariates (livestock density, wild prey abundance, road length, protected area size, elevation) across Iran, we applied a single‐visit N‐mixture model to jointly model variation in detection probability and expected annualized number of leopard killing events.Over the study period, we estimated 428 leopard mortalities (95% CI 184 to 1,014), which was 45% larger than the observed number. Expected intensity of leopard killing was positively related to protected area size, livestock density and wild prey abundance. Detection of leopard killing was higher in areas with more developed road networks.Synthesis and applications. Ranger‐based monitoring data on poaching of carnivores are cost effective, but traditional analysis does not take into account imperfect detection. We show that innovative statistics (single‐visit N‐mixture modelling) can reliably quantify poaching events and address their drivers, at large geographical scales. We used the example of the Persian leopard across Iran, but our approach is also applicable to understand killing dynamics of other species. Results suggest that a high frequency of leopard killing is likely to occur in areas with >100 livestock per km2and >450 individuals of wild prey per km2. This highlights the need for improved management of livestock grazing and effective measures around high‐risk protected areas to mitigate human–leopard conflict and reduce killing of leopards.

Metapopulation dynamics of multiple species in a heterogeneous landscape

AbstractCharacterizing the diversity of demographic strategies among species can inform research in topics such as trait syndromes, community stability, coexistence, and ecological succession. However, this diversity can depend on the spatial scale considered: at the landscape scale, species often form metapopulations, that is sets of local, sometimes short‐lived, populations, inhabiting discrete habitat patches. Metapopulation dynamics are most frequently analyzed in individual species or pairs of interacting species because of the large amount of data required for multiple species, and because species vary in their perceptions of what constitutes a favorable or unfavorable habitat. Here we evaluate, using a case study, whether a metapopulation model can be used to generate accurate estimates of demographic parameters and to describe the diversity of dynamics, responses to environment, and prospects of long‐term persistence in a guild of species inhabiting a common fragmented landscape. We applied this approach to a guild of 22 mollusk species that inhabit freshwater habitats on two islands of Guadeloupe, to compare metapopulation dynamics among species. We analyzed a 15‐year time series of occupancy records for 278 sites using a multistate occupancy model that estimated colonization and extinction rates as a function of site‐specific and year‐specific environmental covariates, then used model results to simulate future island metapopulation dynamics. Despite the diverse array of metapopulation trajectories—a mix of species with either stable, increasing, declining, or fluctuating metapopulations—and the inherent challenges associated with such data (e.g., imperfect detection, spatial and temporal heterogeneity), our model accurately captured among‐patch variation in suitability for many mollusk taxa. The dynamics of rare species or species with habitat preferences not fully captured by the retained set of covariates were less well described. For several species, we detected a negative correlation between extinction and colonization. This variation in habitat suitability created species‐specific extinction‐resistant pockets in the landscape. Our comparative analysis also revealed that species had distinct strategies for metapopulation dynamics, such as “fast‐turnover” species with both a high proportion of occupied sites and a high rate of site extinction in the landscape.

Stage‐specific abiotic and biotic associations of the imperilled Silver ShinerNotropis photogenisin an urban drainage

AbstractStage‐specific abiotic and biotic associations are often poorly described for imperilled freshwater fishes, including Silver Shiner (Notropis photogenis; Threatened), which may limit species protection and recovery efforts. Here, we used fish and habitat sampling data from an urban drainage in southern Ontario, Canada to quantify abiotic and biotic associations for juvenile and adult Silver Shiner, while considering the potential effects of imperfect detection on sampled fishes. Detection probability of Silver Shiner was high (>0.80) across life‐stages, but occupancy models for adults and juveniles included different covariates. The top‐selected occupancy model (lowest Akaike's Information Criterion [AICc]) for adults included a positive effect of site depth, whereas sampling year was included in the top‐selected occupancy model for juveniles. A redundancy analysis suggested unique stage‐specific biotic associations; a positive association was observed between the abundance of adult Silver Shiner and Rosyface Shiner (Notropis rubellus), whereas positive associations were observed between juvenile Silver Shiner and the abundance of Fathead Minnow (Pimephales promelas) and Smallmouth Bass (Micropterus dolomieu). Overall, the results of this study advance the knowledge of stage‐specific abiotic and biotic associations of Silver Shiner, and the probability of species detection, which will inform future monitoring and protection of the species in Canada. More broadly, continued efforts are needed to quantify early life‐stage habitat requirements for informing the recovery of imperilled freshwater species.

N-mixture models provide informative crocodile (Crocodylus moreletii) abundance estimates in dynamic environments.

Estimates of animal abundance provide essential information for population ecological studies. However, the recording of individuals in the field can be challenging, and accurate estimates require analytical techniques which account for imperfect detection. Here, we quantify local abundances and overall population size of Morelet’s crocodiles (Crocodylus moreletii) in the region of Calakmul (Campeche, Mexico), comparing traditional approaches for crocodylians (Minimum Population Size—MPS; King’s Visible Fraction Method—VFM) with binomial N-mixture models based on Poisson, zero-inflated Poisson (ZIP) and negative binomial (NB) distributions. A total of 191 nocturnal spotlight surveys were conducted across 40 representative locations (hydrologically highly dynamic aquatic sites locally known as aguadas) over a period of 3 years (2017–2019). Local abundance estimates revealed a median of 1 both through MPS (min–max: 0–89; first and third quartiles, Q1–Q3: 0–7) and VFM (0–112; Q1–Q3: 0–9) non-hatchling C. moreletii for each aguada, respectively. The ZIP based N-mixture approach shown overall superior confidence over Poisson and NB, and revealed a median of 6 ± 3 individuals (min = 0; max = 120 ± 18; Q1 = 0; Q3 = 18 ± 4) jointly with higher detectabilities in drying aguadas with low and intermediate vegetation cover. Extrapolating these inferences across all waterbodies in the study area yielded an estimated ~10,000 (7,000–11,000) C. moreletii present, highlighting Calakmul as an important region for this species. Because covariates enable insights into population responses to local environmental conditions, N-mixture models applied to spotlight count data result in particularly insightful estimates of crocodylian detection and abundance.

Imperfect detection of spills: A case study in U.S. onshore oil drilling

This paper specifies and estimates regression models to test several hypotheses about the operational and managerial drivers of pollution in a multi-rig onshore drilling organization. In addition to conventional models, detection-controlled models are also specified to explicitly control for the potential for imperfect reporting. The results suggest that continuity in operations and supervision act to reduce the likelihood of pollution. Additional variables such as site complexity are also significant. The results are largely consistent with related research on personal safety incidents. While the analysis was completed for one organization in one geographic area, the results may be applicable to similar regions and organizations. The results can be used to drive decisions regarding operating practices and managerial policies.

Perfect Manipulation of Open Stochastic Quantum Systems with Imperfect Detection

Perfect Manipulation of Open Stochastic Quantum Systems with Imperfect Detection

Habitat occupancy of sloth bear Melursus ursinus in Chitwan National Park, Nepal.

Mammals have experienced a massive decline in their populations and geographic ranges worldwide. The sloth bear, Melursus ursinus (Shaw, 1791), is one of many species facing conservation threats. Despite being endangered in Nepal, decades of inattention to the situation have hindered their conservation and management. We assessed the distribution and patterns of habitat use by sloth bears in Chitwan National Park (CNP), Nepal. We conducted sign surveys from March to June, 2020, in 4 × 4 km grids (n = 45). We collected detection/non‐detection data along a 4‐km trail that was divided into 20 continuous segments of 200 m each. We obtained environmental, ecological, and anthropogenic covariates to understand determinants of sloth bear habitat occupancy. The data were analyzed using the single‐species single‐season occupancy method, with a spatially correlated detection. Using repeated observations, these models accounted for the imperfect detectability of the species to provide robust estimates of habitat occupancy. The model‐averaged occupancy estimate for the sloth bear was 69% and the detection probability was 0.25. The probability of habitat occupancy by sloth bears increased with the presence of termites and fruits and in rugged, dry, open, undisturbed habitats. Our results indicate that the sloth bear is elusive, functionally unique, and widespread in CNP. Future conservation interventions and action plans aimed at sloth bear management must adequately consider their habitat requirements.

A new statistical approach for identifying rare species under imperfect detection

AbstractAimSpecies rarity is often used as a measure to assess the risk of extinction of species, and thus, different methods have been developed to describe the composition of rare species in biological communities. These methods usually depend on species attributes that are not always available and very often ignore imperfect species detection. In this work, we developed a new method to characterize species rarity in a community when species are detected imperfectly. Our modelling framework is based on Bayesian occupancy models to estimate species distributions under imperfect detection using presence‐nondetection data.InnovationWe propose a finite mixture occupancy model to identify rare species based on their occupancy and class‐membership probabilities. Here, we explored a two‐class finite mixture model to distinguish between rare and common species classes and presented the general modelling framework for a problem with more than two classes. By using simulations, we were able to compare our model results under different scenarios obtaining a high‐classification performance across all of them. Additionally, we applied our model to a data set of Odonata occurrence records that were partially observed due to imperfect detection and quantified the proportion of rare species on a national scale across waterbodies in the United Kingdom.Main conclusionsNowadays, biodiversity conservation involves monitoring programmes that target multiple species within a community where individual species responses may vary widely. This high variability makes the task of identifying the ecological processes that drive distributions of rare species difficult. Thus, our method represents a new approach to characterize the composition of a community in terms of species rarity while correcting for detectability bias. Our modelling framework also suggests lines of research and future developments for the understanding of how species rarity can be measured in a wide range of scenarios.

A novel application of hierarchical modelling to decouple sampling artifacts from socio-ecological effects on poaching intensity

Poaching is a global driver of wildlife population decline, including inside protected areas (PAs). Reducing poaching requires an understanding of its cryptic drivers and accurately quantifying poaching scales and intensity. There is little quantification of how poaching is affected by law enforcement intensity (e.g., ranger stations) versus economic factors (e.g., unemployment), while simultaneously accounting for imperfect detection. Using extensive data of poaching events (i.e., seizures) and censuses of nine ungulate species across the PAs and unprotected lands of Iran from 2010 to 2018, we developed a single-visit hierarchical (N-mixture) model to accurately estimate annual poaching of Iranian ungulates and to differentiate between social and ecological effects on annual poaching intensity. We found that poaching detectability increased with numbers of ranger stations. A recent surge in poaching (2013–2018) coincides with rising unemployment rate. We estimated that 19,727 ungulates (95% confidence interval 11,178–36,195) were poached across the country during 2010–2018. Poaching intensity was positively related to unemployment rate, road density, and ungulate abundance. Our simulations demonstrated that the Poisson and Negative binomial N-mixture models had adequate performance when the conditions of Sólymos et al. (2012) were satisfied, in particular, when at least one covariate is unique to both the detection and abundance parts of the model. Overall, we suggest that single-visit models offer unique insights into understanding the link between poaching intensity, economic conditions, and law enforcement in large-scale landscapes while accounting for imperfect detection of poaching events.

Methods Combining Genomic and Epidemiological Data in the Reconstruction of Transmission Trees: A Systematic Review.

In order to better understand transmission dynamics and appropriately target control and preventive measures, studies have aimed to identify who-infected-whom in actual outbreaks. Numerous reconstruction methods exist, each with their own assumptions, types of data, and inference strategy. Thus, selecting a method can be difficult. Following PRISMA guidelines, we systematically reviewed the literature for methods combing epidemiological and genomic data in transmission tree reconstruction. We identified 22 methods from the 41 selected articles. We defined three families according to how genomic data was handled: a non-phylogenetic family, a sequential phylogenetic family, and a simultaneous phylogenetic family. We discussed methods according to the data needed as well as the underlying sequence mutation, within-host evolution, transmission, and case observation. In the non-phylogenetic family consisting of eight methods, pairwise genetic distances were estimated. In the phylogenetic families, transmission trees were inferred from phylogenetic trees either simultaneously (nine methods) or sequentially (five methods). While a majority of methods (17/22) modeled the transmission process, few (8/22) took into account imperfect case detection. Within-host evolution was generally (7/8) modeled as a coalescent process. These practical and theoretical considerations were highlighted in order to help select the appropriate method for an outbreak.

Double-observer approach with camera traps can correct imperfect detection and improve the accuracy of density estimation of unmarked animal populations

Camera traps are a powerful tool for wildlife surveys. However, camera traps may not always detect animals passing in front. This constraint may create a substantial bias in estimating critical parameters such as the density of unmarked populations. We proposed the 'double-observer approach' with camera traps to counter the constraint, which involves setting up a paired camera trap at a station and correcting imperfect detection with a reformulated hierarchical capture-recapture model for stratified populations. We performed simulations to evaluate this approach's reliability and determine how to obtain desirable data for this approach. We then applied it to 12 mammals in Japan and Cameroon. The results showed that the model assuming a beta-binomial distribution as detection processes could correct imperfect detection as long as paired camera traps detect animals nearly independently (Correlation coefficient ≤ 0.2). Camera traps should be installed to monitor a predefined small focal area from different directions to satisfy this requirement. The field surveys showed that camera trap could miss animals by 3–40%, suggesting that current density estimation models relying on perfect detection may underestimate animal density by the same order of magnitude. We hope that our approach will be incorporated into existing density estimation models to improve their accuracy.

Accounting for uncertainty when estimating drivers of imperfect detection: An integrated approach illustrated with snorkel surveys for riverine fishes

Imperfect detection is a common issue affecting the accuracy of surveys that quantify animal abundance and distribution. To quantify detectability, counts are often calibrated to independent measures of abundance (e.g., via mark-recapture) but stochastic sampling variability in both data types is not typically accounted for. This practice may cause detectability to be quantified inaccurately and lead to overly confident predictions for out-of-sample applications. Our objective was to develop, apply, and simulation-test an integrated approach for quantifying detectability that better-accommodates uncertainty in the data. The method assumes mark-recapture and count surveys sample the same local abundance with error, allowing the construction of a joint likelihood function for both data sets. The model estimates coefficients that link detection probability to local covariates through a logit-linear model, which enables correcting counts for imperfect detection in locations where mark-recapture data are unavailable. We illustrate the application of the model with an empirical data set of over 100 paired snorkel count and mark-recapture electrofishing surveys for riverine salmonids in northeastern Oregon. Covariates that best explained heterogeneity in detectability included species, visibility, and channel unit type and depth, though substantial variability was attributed to site-level random effects. Estimated detection probability ranged from 0.02 to 0.92 among surveys and was higher for Chinook Salmon (Oncorhynchus tshawytscha) juveniles (mean: 0.38) than for steelhead/Rainbow Trout (O. mykiss; mean: 0.24). Simulation analyses revealed that our integrated model performed better (relative to a method that treated mark-recapture abundance estimates as known without error) with respect to (i) selection of covariates, (ii), credible interval coverage, (iii) accuracy of estimated random variability terms, and (iv) reduced sensitivity to violated mark-recapture assumptions surrounding behavioral effects. This model represents an improvement over simpler calibration methods, particularly for snorkel surveys, by applying more a rigorous statistical treatment of sources of variability while explicitly describing the mechanistic link between local conditions and detectability. The analytical methods we illustrate are general and could be broadly applied to quantify detectability in other biological surveys with paired abundance and count data.