Non-detection Error Research Articles

Errors in aerial survey count data: Identifying pitfalls and solutions.

Accurate estimates of animal abundance are essential for guiding effective management, and poor survey data can produce misleading inferences. Aerial surveys are an efficient survey platform, capable of collecting wildlife data across large spatial extents in short timeframes. However, these surveys can yield unreliable data if not carefully executed. Despite a long history of aerial survey use in ecological research, problems common to aerial surveys have not yet been adequately resolved. Through an extensive review of the aerial survey literature over the last 50 years, we evaluated how common problems encountered in the data (including nondetection, counting error, and species misidentification) can manifest, the potential difficulties conferred, and the history of how these challenges have been addressed. Additionally, we used a double‐observer case study focused on waterbird data collected via aerial surveys and an online group (flock) counting quiz to explore the potential extent of each challenge and possible resolutions. We found that nearly three quarters of the aerial survey methodology literature focused on accounting for nondetection errors, while issues of counting error and misidentification were less commonly addressed. Through our case study, we demonstrated how these challenges can prove problematic by detailing the extent and magnitude of potential errors. Using our online quiz, we showed that aerial observers typically undercount group size and that the magnitude of counting errors increases with group size. Our results illustrate how each issue can act to bias inferences, highlighting the importance of considering individual methods for mitigating potential problems separately during survey design and analysis. We synthesized the information gained from our analyses to evaluate strategies for overcoming the challenges of using aerial survey data to estimate wildlife abundance, such as digital data collection methods, pooling species records by family, and ordinal modeling using binned data. Recognizing conditions that can lead to data collection errors and having reasonable solutions for addressing errors can allow researchers to allocate resources effectively to mitigate the most significant challenges for obtaining reliable aerial survey data.

Evaluation of Portable Water Filtration Systems to Reduce the Transport Risk of Zebra Mussels and Asian Clams during Fish Spawning and Trap‐and‐Transfer Operations

AbstractInvasive species represent a substantial threat to aquatic systems, and aquaculture operations provide a vector for their transportation between waters. Moving large volumes of water during fish spawning or trap‐and‐transfer operations increases the risk of transporting invasive bivalve larvae, particularly from sources experiencing nondetection error. This study evaluated the effectiveness of two portable filter systems at reducing or eliminating zebra mussel Dreissena polymorpha and Asian clam Corbicula fluminea veligers from an infested water source. A small setup was assessed at intervals to 1,420 L using combinations of 20‐ and 5‐μm filters arranged in tandem, whereas a large setup was tested at intervals to 7,041 L with combinations of 30‐ and 5‐μm tandem filters operating at two pressure ranges. Zero veligers were observed in 36 small‐setup filtered water samples. Four zebra mussel veligers and zero Asian clam veligers were observed in 60 large‐setup samples. Total zebra mussel and Asian clam counts from 31 unfiltered control samples were 27,846 and 1,095, respectively. Overall, both filter systems were highly effective at removing zebra mussel and Asian clam veligers and represent simple, cost‐effective safeguards against nondetection error of aquatic invasive species in waters used for aquaculture practices.

A Bayesian semiparametric GLMM for historical and newly collected presence‐only data: An application to species richness of Ross Sea Mollusca

Historical data sets from vast and relatively inaccessible areas are sources of potentially unique information still valuable for biodiversity studies today. In many research fields, ranging from climate change to projection of species loss, great efforts have been made to integrate historical data sets with recent data to create databases that are as complete as possible. Unlocking the information contained in presence‐only data, largely prevalent in such databases, presents a challenge for statistical modeling because of insidious observational errors due to the opportunistic nature of the data‐gathering process. In this article, we propose an appropriate statistical method for the joint analysis of historical and newly collected presence‐only data, that is, a Bayesian semiparametric generalized linear mixed model with Dirichlet process random effects. The potential of the method is illustrated by considering the Ross Sea section of the SOMBASE, an international compilation of Southern Ocean Mollusc distributional records, from 1899 to 2004 and beyond. Despite the presence of sampling bias and nondetection errors, the proposed model draws latent information from the data, such that the resulting estimates of the parameters of interest not only are coherent with those obtained in indirectly related studies based on well‐structured data but also suggest interesting ideas for further research.

Monitoring the near-extinct European weather loach in Denmark based on environmental DNA from water samples

The European weather loach (Misgurnus fossilis) represents one of many European freshwater fishes in decline. Efficient monitoring is essential if conservation efforts are to be successful, but due to the species’ cryptic biology, traditional monitoring methods currently in use are inefficient, time consuming and likely prone to non-detection error. Here, we investigate the usefulness of environmental DNA (eDNA) monitoring as an alternative or supplementary method for surveying the Danish weather loach population, which is presumed to consist primarily of a single group of no more than 50 individuals. In 2008, the majority of historical Danish localities were surveyed, using traditional fishing techniques. We then applied eDNA methods to a number of these, as well as other potential localities. We successfully detected the weather loach at multiple sites in the single known remaining locality; a result that was later confirmed when local managers caught eight live specimens. Furthermore, the eDNA method indicated presence of the weather loach in another historical locality, where the species has not been observed since 1995. At the remaining localities, weather loach eDNA was not detected, providing further evidence for its absence. Importantly, the eDNA survey required less effort in person-hours and lower costs than the traditional fishing survey. This study confirms that eDNA monitoring is a valid supplement to traditional monitoring methods currently applied to monitor rare freshwater fishes. We propose that by providing reliable distribution data at lower cost and limited effort, the eDNA method can allow for increased management efficiency of endangered freshwater species such as the European weather loach.

Accounting for imperfect detection in ecology: a quantitative review.

Detection in studies of species abundance and distribution is often imperfect. Assuming perfect detection introduces bias into estimation that can weaken inference upon which understanding and policy are based. Despite availability of numerous methods designed to address this assumption, many refereed papers in ecology fail to account for non-detection error. We conducted a quantitative literature review of 537 ecological articles to measure the degree to which studies of different taxa, at various scales, and over time have accounted for imperfect detection. Overall, just 23% of articles accounted for imperfect detection. The probability that an article incorporated imperfect detection increased with time and varied among taxa studied; studies of vertebrates were more likely to incorporate imperfect detection. Among articles that reported detection probability, 70% contained per-survey estimates of detection that were less than 0.5. For articles in which constancy of detection was tested, 86% reported significant variation. We hope that our findings prompt more ecologists to consider carefully the detection process when designing studies and analyzing results, especially for sub-disciplines where incorporation of imperfect detection in study design and analysis so far has been lacking.

Classification of Endomicroscopic Images of the Lung Based on Random Subwindows and Extra-Trees

Recently, the in vivo imaging of pulmonary alveoli was made possible thanks to confocal microscopy. For these images, we wish to aid the clinician by developing a computer-aided diagnosis system, able to discriminate between healthy and pathological subjects. The lack of expertise currently available on these images has first led us to choose a generic approach, based on pixel-value description of randomly extracted subwindows and decision tree ensemble for classification (extra-trees). In order to deal with the great complexity of our images, we adapt this method by introducing a texture-based description of the subwindows, based on local binary patterns. We show through our experimental protocol that this adaptation is a promising way to classify fibered confocal fluorescence microscopy images. In addition, we introduce a rejection mechanism on the classifier output to prevent nondetection errors.

Observer effects and avian-call-count survey quality: Rare-species biases and overconfidence

ABSTRACT. Wildlife monitoring surveys are prone to nondetection errors and false positives. To determine factors that affect the incidence of these errors, we built an Internet-based survey that simulated avian point counts, and measured error rates among volunteer observers. Using similar-sounding vocalizations from paired rare and common bird species, we measured the effects of species rarity and observer skill, and the influence of a reward system that explicitly encouraged the detection of rare species. Higher self-reported skill levels and common species independently predicted fewer nondetections (probability range: 0.11 [experts, common species] to 0.54 [moderates, rare species]). Overall proportions of detections that were false positives increased significantly as skill level declined (range: 0.06 [experts, common species] to 0.22 [moderates, rare species]). Moderately skilled observers were significantly more likely to report false-positive records of common species than of rare species, whereas...

Non-detection errors in a survey of persistent, highly-detectable vegetation species

Rare, small or annual vegetation species are widely known to be imperfectly detected with single site surveys by most conventional vegetation survey methods. However, the detectability of common, persistent vegetation species is assumed to be high, but without supporting research. In this study, we evaluate the extent of false-negative errors of perennial vegetation species in a systematic vegetation survey in arid South Australia. Analysis was limited to the seven most easily detected persistent vegetation species and controlled for observer skill. By comparison of methodologies, we then predict the magnitude of non-detection error rates in a second survey. The analysis revealed that all but one highly detectable perennial vegetation species was imperfectly detected (detection probabilities ranged from 0.22 to 0.83). While focussed in the Australian rangelands, the implications of this study are far reaching. Inferences drawn from systematic vegetation surveys that fail to identify and account for non-detection errors should be considered potentially flawed. The identification of this problem in vegetation surveying is long overdue. By comparison, non-detection has been a widely acknowledged, and dealt with, problem in fauna surveying for decades. We recommend that, where necessary, vegetation survey methodology adopt the methods developed in fauna surveying to cope with non-detection errors.

Overwintering in nests on the ground in the harvest mouse

The harvest mouse Micromys minutus has adapted to inhabit the stalk zone of grassland vegetation and is identified by the characteristic nests this species builds on grasses above the ground. Such aerial nests have been considered as an almost exclusive sign of harvest mouse populations and have therefore also been used to determine habitat choice and population density of the harvest mouse. However, we found that nests built on the ground occur after farmers burn away grasslands dominated by native grass Miscanthus sinensis and Pleioblastus chino. The aim of this study was to determine whether the harvest mouse habitually builds this type of nest and, if so, when it occurs. Because it was difficult to locate nests on the ground when the soil was covered with grasses, we located all the nests in the native grasslands after burning in the spring in 2 consecutive years. We then attempted to estimate nesting time by comparing the mineral content of nest materials to mineral dynamics in M. sinensis leaves. Nests were distributed over all the grasslands investigated, and nesting times were calculated to occur between early September and middle March. Our findings demonstrate the harvest mouse repeatedly constructs this type of nest for overwintering. Considerable nondetection errors would have occurred with regard to this species by exclusively interpreting data on the presence or absence of aerial nests.

Survey of an otterLutra lutra population in Southern Italy: site occupancy and influence of sampling season on species detection

In order to assess the current status of the otterLutra lutra Linnaeus, 1758 in Southern Campania (Italy), we surveyed 141 sites for spraints (faeces) in 2001. Fifty-eight sites were surveyed during the winter and in the following summer in order to test, through an estimation-based approach, the influence of sampling season on species detection. Site occupation in the study area was high (69.5%) and possibly underestimated, because the survey was affected by non-detection errors. Our analyses showed that winter surveying markedly underestimated true otter occupancy at the 58 sites (51.7% vs 97.1%), whereas summer surveying was very reliable (91.4%). Rains and floods may have removed spraints during winter, thereby reducing the detection probability to 0.534. These results suggest that otter standard surveys in areas with Mediterranean-type climates should be conducted during summer or periods without prolonged precipitation. Comparing our results with those of the 1985 National survey, we found an occupancy increase from 65.8 to 100%. We could not establish whether this change reflected a population increase or was due to possible non-detection errors that occurred in the 1985 survey. However, the present occupancy substantiates the strategic relevance of the study area for planning the conservation of Italian otters.

A BAYESIAN STATE-SPACE FORMULATION OF DYNAMIC OCCUPANCY MODELS

Species occurrence and its dynamic components, extinction and colonization probabilities, are focal quantities in biogeography and metapopulation biology, and for species conservation assessments. It has been increasingly appreciated that these parameters must be estimated separately from detection probability to avoid the biases induced by non-detection error. Hence, there is now considerable theoretical and practical interest in dynamic occupancy models that contain explicit representations of metapopulation dynamics such as extinction, colonization, and turnover as well as growth rates. We describe a hierarchical parameterization of these models that is analogous to the state-space formulation of models in time series, where the model is represented by two components, one for the partially observable occupancy process and another for the observations conditional on that process. This parameterization naturally allows estimation of all parameters of the conventional approach to occupancy models, but in addition, yields great flexibility and extensibility, e.g., to modeling heterogeneity or latent structure in model parameters. We also highlight the important distinction between population and finite sample inference; the latter yields much more precise estimates for the particular sample at hand. Finite sample estimates can easily be obtained using the state-space representation of the model but are difficult to obtain under the conventional approach of likelihood-based estimation. We use R and WinBUGS to apply the model to two examples. In a standard analysis for the European Crossbill in a large Swiss monitoring program, we fit a model with year-specific parameters. Estimates of the dynamic parameters varied greatly among years, highlighting the irruptive population dynamics of that species. In the second example, we analyze route occupancy of Cerulean Warblers in the North American Breeding Bird Survey (BBS) using a model allowing for site-specific heterogeneity in model parameters. The results indicate relatively low turnover and a stable distribution of Cerulean Warblers which is in contrast to analyses of counts of individuals from the same survey that indicate important declines. This discrepancy illustrates the inertia in occupancy relative to actual abundance. Furthermore, the model reveals a declining patch survival probability, and increasing turnover, toward the edge of the range of the species, which is consistent with metapopulation perspectives on the genesis of range edges. Given detection/non-detection data, dynamic occupancy models as described here have considerable potential for the study of distributions and range dynamics.

Absent or undetected? Effects of non-detection of species occurrence on wildlife–habitat models

Presence–absence data are used widely in analysis of wildlife–habitat relationships. Failure to detect a species’ presence in an occupied habitat patch is a common sampling problem when the population size is small, individuals are difficult to sample, or sampling effort is limited. In this paper, the influence of non-detection of occurrence on parameter estimates of logistic regression models of wildlife–habitat relationships was assessed using analytical analysis and simulations. Two patterns of non-detection were investigated: (1) a random distribution of non-detection among occupied patches; and (2) a non-random distribution of non-detection in which the probability of detecting a species in an occupied patch covaried with measurable habitat variables. Our results showed that logistic regression models of wildlife–habitat relationships were sensitive to even low levels of non-detection in occupancy data. Both analytic and simulation studies show that non-detection yields bias in parameter estimation of logistic regression models. More importantly, the direction of bias was affected by the underlying pattern of non-detection and whether the habitat variable was positively or negatively related to occupancy. For a positive habitat coefficient, a random distribution of non-detection yielded negative bias in estimation, whereas linkage of the probability of non-detection to habitat covariates produced positive bias. For a negative habitat coefficient, the pattern was reversed, with a random distribution of non-detection leading to positive bias in estimation. A release–recapture livetrapping study of small mammals in central Indiana, USA, was used to illustrate the magnitude of non-detection in a typical field sampling protocol with varying levels of sampling intensity. Estimates of non-detection error ranged from 0 to 23% for seven species after 5 days of sampling. We suggest that for many sampling situations, relationships between probability of detection and habitat covariates need to be established to correctly interpret results of wildlife–habitat models.

Detector efficiency in the Greenberger-Horne-Zeilinger paradox: Independent errors

The Greenberger-Horne-Zeilinger (GHZ) paradox is subject to the detector-efficiency ``loophole'' in a similar manner as the Bell inequality. In a paper by J.-\AA{}. Larsson [Phys. Rev. A 57, R3145 (1998)], the issue is investigated for very general assumptions. Here, the assumptions of constant efficiency and independent errors will be imposed, and it will be shown that the necessary and sufficient efficiency bound is not lowered, but remains at $75%$. An explicit local-variable model is constructed in this paper to show the necessity of this bound. In other words, it is not possible to use the independence of experimental nondetection errors to rule out local realism in the GHZ paradox below $75%$ efficiency.

Multiple paternity in social Hymenoptera: estimating the effective mate number in single–double mating populations

Several hypotheses have been advanced to explain the widespread occurrence of multiple mating by queens in eusocial bees, wasps and ants. In order to test competing hypotheses, it is essential to obtain an accurate description of the distribution of mate number and the mates’ respective contributions to the brood in populations under study. In this study we examine current methods for estimating the effective mate number and other paternity frequency parameters in populations dominated by single and double mating of queens, based on genetical data from parent–offspring combinations. Estimation errors are due to limited sample size of offspring (nonsampling error) and limited variation of genetic markers (nondetection error). We demonstrate that limited sample size of offspring leads to an underestimation of the paternity contribution from the majority male of double‐mated queens (the paternity skew or sperm bias), thus overestimating the effectiveness of double matings and underestimating the population frequency of double‐mated queens. We propose a calculation procedure which avoids this bias. Furthermore, we propose a general method to correct simultaneously for errors from nondetection and nonsampling when estimating the population frequency of double‐mated queens. The calculation procedures presented are designed for male‐haplodiploid social Hymenoptera but some can be applied directly to studies of diplo‐diploid mating systems where single and double mating prevail. This has particular relevance to vertebrate studies where the number of offspring studied per mother is invariably low. The remaining calculation procedures are conceptually widely applicable, but details would have to be modified for other than haplodiploid mating systems.