Navigating uncertainty in environmental DNA detection of a nuisance marine macroalga

Early detection of nuisance species is crucial for managing threatened ecosystems and preventing widespread establishment. Environmental DNA (eDNA) data can increase the sensitivity of biomonitoring programs, often at minimal cost and effort. However, eDNA analyses are prone to errors that can complicate their use in management frameworks. To address this, eDNA studies must consider imperfect detections and estimate error rates. Detecting nuisance species at low abundances with minimal uncertainty is vital for successful containment and eradication. We developed a novel eDNA assay to detect a nuisance marine macroalga across its colonization front using surface seawater samples from Papahānaumokuākea Marine National Monument (PMNM), one of the world's largest marine reserves. Chondria tumulosa is a cryptogenic red alga with invasive traits, forming dense mats that overgrow coral reefs and smother native flora and fauna in PMNM. We verified the eDNA assay using site-occupancy detection modeling from quantitative polymerase chain reaction (qPCR) data, calibrated with visual estimates of benthic cover of C. tumulosa that ranged from < 1% to 95%. Results were subsequently validated with high-throughput sequencing of amplified eDNA and negative control samples. Overall, the probability of detecting C. tumulosa at occupied sites was at least 92% when multiple qPCR replicates were positive. False-positive rates were 3% or less and false-negative errors were 11% or less. The assay proved effective for routine monitoring at shallow sites (less than 10 m), even when C. tumulosa abundance was below 1%. Successful implementation of eDNA tools in conservation decision-making requires balancing uncertainties in both visual and molecular detection methods. Our results and modeling demonstrated the assay's high sensitivity to C. tumulosa, and we outline steps to infer ecological presence-absence from molecular data. This reliable, cost-effective tool enhances the detection of low-abundance species, and supports timely management interventions.

A spatially explicit assessment of factors shaping attitudes toward African elephant conservation

Conservation plans that explicitly account for the social landscape where people and wildlife co-occur can yield more effective and equitable conservation practices and outcomes. Yet, social data remain underutilized, often because social data are treated as aspatial or are analyzed with approaches that do not quantify uncertainty or address bias in self-reported data. We conducted a survey (questionnaires) of 177 households in a multiuse landscape in the Kenya-Tanzania borderlands. In a mixed-methods approach, we used Bayesian hierarchical models to quantify and map local attitudes toward African elephant (Loxodonta africana) conservation while accounting for response bias and then combined inference from attitude models with thematic analysis of open-ended responses and cointerpretation of results with local communities to gain deeper understanding of what explains attitudes of people living with wildlife. Model estimates showed that believing elephants have sociocultural value increased the probability of respondents holding positive attitudes toward elephant conservation in general (mean increase = 0.31 [95% credible interval, CrI, 0.02-0.67]), but experiencing negative impacts from any wildlife species lowered the probability of respondents holding a positive attitude toward local elephant conservation (mean decrease = -0.20 [95% CrI -0.42 to 0.03]). Qualitative data revealed that safety and well-being concerns related to the perceived threats that elephants pose to human lives and livelihoods, and limited incentives to support conservation on community and private lands lowered positive local attitude probabilities and contributed to negative perceptions of human-elephant coexistence. Our spatially explicit modeling approach revealed fine-scale variation in drivers of conservation attitudes that can inform targeted conservation planning. Our results suggest that approaches focused on sustaining existing sociocultural values and relationships with wildlife, investing in well-being, and implementing species-agnostic approaches to wildlife impact mitigation could improve conservation outcomes in shared landscapes.

Combining ranger records and biogeographical models to identify the current and potential distribution of an expanding mesocarnivore in southern Europe

Human-wildlife conflicts (HWC) are increasing and are potentially harmful to both people and wildlife. Understanding the current and potential distribution of wildlife species involved in HWC, such as carnivores, is essential for implementing management and conservation measures for such species. In this study, we assessed both the current distribution and potential distribution (forecast) of the Egyptian mongoose (Herpestes ichneumon) in the central part of the Iberian Peninsula. We acquired data concerning mongoose occurrences through an online questionnaire sent to environmental rangers. We used the municipality level as the sampling unit because all municipalities within the study area were covered at least by one ranger. Using the information provided by rangers (i.e. occurrences in their municipalities), we constructed environmental favourability distribution models to assess current and potential mongoose distribution through current distribution models (CDM) and ecological models (EM), respectively. >300 rangers participated in the survey and mongooses were reported in a total of 181 of 921 municipalities studied. The CDM model showed a current distribution mainly concentrated on the western part of the study area, where intermediate-high favourability values predominated. The EM model revealed a wider potential distribution, including the south-east part of the study area, which was also characterised by intermediate-high favourability values. Our predictions were verified using independent data, including confirmation of mongoose reproduction by rangers, reports by other experts, and field sampling in some areas. Our innovative approach based on an online survey to rangers coupled with environmental favourability models is shown to be a useful methodology for assessing the current distribution of cryptic but expanding wildlife species, while also enabling estimations of future steps in their expansion. The approach proposed may help policy decision-makers seeking to ensure the conservation of expanding wildlife species, for example, by designing awareness campaigns in areas where the target species is expected to arrive.

Quantitative assessment of reef foraminifera community from metabarcoding data

Describing living community compositions is essential to monitor ecosystems in a rapidly changing world, but it is challenging to produce fast and accurate depiction of ecosystems due to methodological limitations. Morphological methods provide absolute abundances with limited throughput, whereas metabarcoding provides relative abundances of genes that may not correctly represent living communities from environmental DNA assessed with morphological methods. However, it has the potential to deliver fast descriptions of living communities provided that it is interpreted with validated species-specific calibrations and reference databases. Here, we developed a quantitative approach to retrieve from metabarcoding data the assemblages of living large benthic foraminifera (LBF), photosymbiotic calcifying protists, from Indonesian coral reefs that are under increasing anthropogenic pressure. To depict the diversity, we calculated taxon-specific correction factors to reduce biological biases by comparing surface area, biovolume and calcite volume, and the number of mitochondrial gene copies in seven common LBF species. To validate the approach, we compared calibrated datasets of morphological communities from mock samples with bulk reef sediment; both sample types were metabarcoded. The calibration of the data significantly improved the estimations of genus relative abundance, with a difference of ±5% on average, allowing for comparison of past morphological datasets with future molecular ones. Our results also highlight the application of our quantitative approach to support reef monitoring operations by capturing fine-scale processes, such as seasonal and pollution-driven dynamics, that require high-throughput sampling treatment.

Towards a standardized framework for AI-assisted, image-based monitoring of nocturnal insects

Automated sensors have potential to standardize and expand the monitoring of insects across the globe. As one of the most scalable and fastest developing sensor technologies, we describe a framework for automated, image-based monitoring of nocturnal insects-from sensor development and field deployment to workflows for data processing and publishing. Sensors comprise a light to attract insects, a camera for collecting images and a computer for scheduling, data storage and processing. Metadata is important to describe sampling schedules that balance the capture of relevant ecological information against power and data storage limitations. Large data volumes of images from automated systems necessitate scalable and effective data processing. We describe computer vision approaches for the detection, tracking and classification of insects, including models built from existing aggregations of labelled insect images. Data from automated camera systems necessitate approaches that account for inherent biases. We advocate models that explicitly correct for bias in species occurrence or abundance estimates resulting from the imperfect detection of species or individuals present during sampling occasions. We propose ten priorities towards a step-change in automated monitoring of nocturnal insects, a vital task in the face of rapid biodiversity loss from global threats. This article is part of the theme issue 'Towards a toolkit for global insect biodiversity monitoring'.

Implications of the scale of detection for inferring co-occurrence patterns from paired camera traps and acoustic recorders

Multifunctional landscapes that support economic activities and conservation of biological diversity (e.g., cattle ranches with native forest) are becoming increasingly important because small remnants of native forest may comprise the only habitat left for some wildlife species. Understanding the co-occurrence between wildlife and disturbance factors, such as poaching activity and domesticated ungulates, is key to successful management of multifunctional landscapes. Tools to measure co-occurrence between wildlife and disturbance factors include camera traps and autonomous acoustic recording units. We paired 52 camera-trap stations with acoustic recorders to investigate the association between 2 measures of disturbance (poaching and cattle) and wild ungulates present in multifunctional landscapes of the Colombian Orinoquía. We used joint species distribution models to investigate species-habitat associations and species-disturbance correlations. One model was fitted using camera-trap data to detect wild ungulates and disturbance factors, and a second model was fitted after replacing camera-trap detections of disturbance factors with their corresponding acoustic detections. The direction, significance, and precision of the effect of covariates depended on the sampling method used for disturbance factors. Acoustic monitoring typically resulted in more precise estimates of the effects of covariates and of species-disturbance correlations. Association patterns between wildlife and disturbance factors were found only when disturbance was detected by acoustic recorders. Camera traps allowed us to detect nonvocalizing species, whereas audio recording devices increased detection of disturbance factors leading to more precise estimates of co-occurrence patterns. The collared peccary (Pecari tajacu), lowland tapir (Tapirus terrestris), and white-tailed deer (Odocoileus virginianus) co-occurred with disturbance factors and are conservation priorities due to the greater risk of poaching or disease transmission from cattle.

Estimating distribution and abundance of wide-ranging species with integrated spatial models: Opportunities revealed by the first wolf assessment in south-central Italy

Estimating demographic parameters for wide-ranging and elusive species living at low density is challenging, especially at the scale of an entire country. To produce wolf distribution and abundance estimates for the whole south-central portion of the Italian wolf population, we developed an integrated spatial model, based on the data collected during a 7-month sampling campaign in 2020-2021. Data collection comprised an extensive survey of wolf presence signs, and an intensive survey in 13 sampling areas, aimed at collecting non-invasive genetic samples (NGS). The model comprised (i) a single-season, multiple data-source, multi-event occupancy model and (ii) a spatially explicit capture-recapture model. The information about species' absence was used to inform local density estimates. We also performed a simulation-based assessment, to estimate the best conditions for optimizing sub-sampling and population modelling in the future. The integrated spatial model estimated that 74.2% of the study area in south-central Italy (95% CIs = 70.5% to 77.9%) was occupied by wolves, for a total extent of the wolf distribution of 108,534 km2 (95% CIs = 103,200 to 114,000). The estimate of total population size for the Apennine wolf population was of 2557 individuals (SD = 171.5; 95% CIs = 2127 to 2844). Simulations suggested that the integrated spatial model was associated with an average tendency to slightly underestimate population size. Also, the main contribution of the integrated approach was to increase precision in the abundance estimates, whereas it did not affect accuracy significantly. In the future, the area subject to NGS should be increased to at least 30%, while at least a similar proportion should be sampled for presence-absence data, to further improve the accuracy of population size estimates and avoid the risk of underestimation. This approach could be applied to other wide-ranging species and in other geographical areas, but specific a priori evaluations of model requirements and expected performance should be made.

Modelling heterogeneity in the classification process in multi-species distribution models can improve predictive performance

Species distribution models and maps from large-scale biodiversity data are necessary for conservation management. One current issue is that biodiversity data are prone to taxonomic misclassifications. Methods to account for these misclassifications in multi-species distribution models have assumed that the classification probabilities are constant throughout the study. In reality, classification probabilities are likely to vary with several covariates. Failure to account for such heterogeneity can lead to biased prediction of species distributions. Here, we present a general multi-species distribution model that accounts for heterogeneity in the classification process. The proposed model assumes a multinomial generalised linear model for the classification confusion matrix. We compare the performance of the heterogeneous classification model to that of the homogeneous classification model by assessing how well they estimate the parameters in the model and their predictive performance on hold-out samples. We applied the model to gull data from Norway, Denmark and Finland, obtained from the Global Biodiversity Information Facility. Our simulation study showed that accounting for heterogeneity in the classification process increased the precision of true species' identity predictions by 30% and accuracy and recall by 6%. Since all the models in this study accounted for misclassification of some sort, there was no significant effect of accounting for heterogeneity in the classification process on the inference about the ecological process. Applying the model framework to the gull dataset did not improve the predictive performance between the homogeneous and heterogeneous models (with parametric distributions) due to the smaller misclassified sample sizes. However, when machine learning predictive scores were used as weights to inform the species distribution models about the classification process, the precision increased by 70%. We recommend multiple multinomial regression to be used to model the variation in the classification process when the data contains relatively larger misclassified samples. Machine learning prediction scores should be used when the data contains relatively smaller misclassified samples.

Factors influencing the detection and occupancy of little brown bats (Myotis lucifugus)

Using acoustics to survey for bats has increased as the need for data on increasingly rare species has also increased. We set out to better understand the difference between mist netting and acoustic detection probabilities between these two methods for the little brown bat (Myotis lucifugus), a species highly impacted by white-nose syndrome and currently considered for federal listing in the United States. We also analyzed occupancy relationships with local and landcover variables. We surveyed 15 sites using mist netting paired with an acoustic recorder for multiple nights to estimate detection probability of this species. We also deployed acoustic recorders at another 73 sites. We found that detection rates for mist netting were very low but increased with day of year and decreased from proximity to water. Acoustic surveys had higher detection rates, but there was an approximately 30% likelihood of false-positive detections. At the mean distance to water and day of year, acoustic surveys had a detection rate 55 times higher than mist netting. There were not significant factors influencing occupancy of little brown bats, only a slight positive relationship between forested largest patch, landscape patch richness and forest basal area. Given the declines in little brown bat populations since white-nose syndrome, it is even more critical that we consider the very low detection rate of mist netting compared to acoustic surveys.

Development and validation of an environmental DNA assay to detect federally threatened groundwater salamanders in central Texas

The molecular detection of DNA fragments that are shed into the environment (eDNA) has become an increasingly applied tool used to inventory biological communities and to perform targeted species surveys. This method is particularly useful in habitats where it is difficult or not practical to visually detect or trap the target organisms. Central Texas Eurycea salamanders inhabit both surface and subterranean aquatic environments. Subterranean surveys are challenging or infeasible, and the detection of salamander eDNA in water samples is an appealing survey technique for these situations. Here, we develop and validate an eDNA assay using quantitative PCR for E. chisholmensis, E. naufragia, and E. tonkawae. These three species are federally threatened and constitute the Septentriomolge clade that occurs in the northern segment of the Edwards Aquifer. First, we validated the specificity of the assay in silico and with DNA extracted from tissue samples of both target Septentriomolge and non-target amphibians that overlap in distribution. Then, we evaluated the sensitivity of the assay in two controls, one with salamander-positive water and one at field sites known to be occupied by Septentriomolge. For the salamander-positive control, the estimated probability of eDNA occurrence (ψ) was 0.981 (SE = 0.019), and the estimated probability of detecting eDNA in a qPCR replicate (p) was 0.981 (SE = 0.011). For the field control, the estimated probability of eDNA occurring at a site (ψ) was 0.938 (95% CRI: 0.714-0.998). The estimated probability of collecting eDNA in a water sample (θ) was positively correlated with salamander relative density and ranged from 0.371 (95% CRI: 0.201-0.561) to 0.999 (95% CRI: 0.850- > 0.999) among sampled sites. Therefore, sites with low salamander density require more water samples for eDNA evaluation, and we determined that our site with the lowest estimated θ would require seven water samples for the cumulative collection probability to exceed 0.95. The estimated probability of detecting eDNA in a qPCR replicate (p) was 0.882 (95% CRI: 0.807-0.936), and our assay required two qPCR replicates for the cumulative detection probability to exceed 0.95. In complementary visual encounter surveys, the estimated probability of salamanders occurring at a known-occupied site was 0.905 (SE = 0.096), and the estimated probability of detecting salamanders in a visual encounter survey was 0.925 (SE = 0.052). We additionally discuss future research needed to refine this method and understand its limitations before practical application and incorporation into formal survey protocols for these taxa.

Sampling methodology influences habitat suitability modeling for chiropteran species

Technological advances increase opportunities for novel wildlife survey methods. With increased detection methods, many organizations and agencies are creating habitat suitability models (HSMs) to identify critical habitats and prioritize conservation measures. However, multiple occurrence data types are used independently to create these HSMs with little understanding of how biases inherent to those data might impact HSM efficacy. We sought to understand how different data types can influence HSMs using three bat species (Lasiurus borealis, Lasiurus cinereus, and Perimyotis subflavus). We compared the overlap of models created from passive-only (acoustics), active-only (mist-netting and wind turbine mortalities), and combined occurrences to identify the effect of multiple data types and detection bias. For each species, the active-only models had the highest discriminatory ability to tell occurrence from background points and for two of the three species, active-only models preformed best at maximizing the discrimination between presence and absence values. By comparing the niche overlaps of HSMs between data types, we found a high amount of variation with no species having over 45% overlap between the models. Passive models showed more suitable habitat in agricultural lands, while active models showed higher suitability in forested land, reflecting sampling bias. Overall, our results emphasize the need to carefully consider the influences of detection and survey biases on modeling, especially when combining multiple data types or using single data types to inform management interventions. Biases from sampling, behavior at the time of detection, false positive rates, and species life history intertwine to create striking differences among models. The final model output should consider biases of each detection type, particularly when the goal is to inform management decisions, as one data type may support very different management strategies than another.

DNA High-Throughput Sequencing for Arthropod Gut Content Analysis to Evaluate Effectiveness and Safety of Biological Control Agents

The search for effective biological control agents without harmful non-target effects has been constrained by the use of impractical (field direct observation) or imprecise (cage experiments) methods. While advances in the DNA sequencing methods, more specifically the development of high-throughput sequencing (HTS), have been quickly incorporated in biodiversity surveys, they have been slow to be adopted to determine arthropod prey range, predation rate and food web structure, and critical information to evaluate the effectiveness and safety of a biological control agent candidate. The lack of knowledge on how HTS methods could be applied by ecological entomologists constitutes part of the problem, although the lack of expertise and the high cost of the analysis also are important limiting factors. In this review, we describe how the latest HTS methods of metabarcoding and Lazaro, a method to identify prey by mapping unassembled shotgun reads, can serve biological control research, showing both their power and limitations. We explain how they work to determine prey range and also how their data can be used to estimate predation rates and subsequently be translated into food webs of natural enemy and prey populations helping to elucidate their role in the community. We present a brief history of prey detection through molecular gut content analysis and also the attempts to develop a more precise formula to estimate predation rates, a problem that still remains. We focused on arthropods in agricultural ecosystems, but most of what is covered here can be applied to natural systems and non-arthropod biological control candidates as well.

Linking frass and insect phenology to optimize annual forest defoliation estimation

It is often logistically impractical to measure forest defoliation events in the field due to seasonal variability in larval feeding phenology (e.g., start, peak, and end) in any given year. As such, field data collections are either incomplete or at coarse temporal resolutions, both of which result in inaccurate estimation of annual defoliation (frass or foliage loss). Using Choristoneura pinus F. and Lymantria dispar dispar L., we present a novel approach that leverages a weather-driven insect simulation model (BioSIM) and defoliation field data. Our approach includes optimization of a weighting parameter (w) for each instar and imputation of defoliation. Results show a negative skew in this weighting parameter, where the second to last instar in a season exhibits the maximum consumption and provides better estimates of annual frass and foliage biomass loss where sampling data gaps exist. Respective cross-validation RMSE (and normalized RMSE) results for C. pinus and L. dispar dispar are 77.53 kg·ha^-1 (0.16) and 38.24 kg·ha^-1 (0.02) for frass and 74.85 kg·ha^-1 (0.10) and 47.77 kg·ha^-1 (0.02) for foliage biomass loss imputation. Our method provides better estimates for ecosystem studies that leverage remote sensing data to scale defoliation rates from the field to broader landscapes and regions.•Utilize fine temporal resolution insect life cycle data derived from weather-driven insect simulation model (BioSIM) to bridge critical gaps in coarse temporal resolution defoliation field data.•Fitting distributions to optimize the instar weighting parameter (w) and impute frass and foliage biomass loss based on a cumulative density function (CDF).•Enables accurate estimation of annual defoliation impacts on ecosystems across multiple insect taxa that exhibit distinct but seasonally variable feeding phenology.

Detecting Leishmania in dogs: A hierarchical-modeling approach to investigate the performance of parasitological and qPCR-based diagnostic procedures

BACKGROUND

Domestic dogs are primary reservoir hosts of Leishmania infantum, the agent of visceral leishmaniasis. Detecting dog infections is central to epidemiological inference, disease prevention, and veterinary practice. Error-free diagnostic procedures, however, are lacking, and the performance of those available is difficult to measure in the absence of fail-safe "reference standards". Here, we illustrate how a hierarchical-modeling approach can be used to formally account for false-negative and false-positive results when investigating the process of Leishmania detection in dogs.

METHODS/FINDINGS

We studied 294 field-sampled dogs of unknown infection status from a Leishmania-endemic region. We ran 350 parasitological tests (bone-marrow microscopy and culture) and 1,016 qPCR assays (blood, bone-marrow, and eye-swab samples with amplifiable DNA). Using replicate test results and site-occupancy models, we estimated (a) clinical sensitivity for each diagnostic procedure and (b) clinical specificity for qPCRs; parasitological tests were assumed 100% specific. Initial modeling revealed qPCR specificity < 94%; we tracked the source of this unexpected result to some qPCR plates having subtle signs of possible contamination. Using multi-model inference, we formally accounted for suspected plate contamination and estimated qPCR sensitivity at 49-53% across sample types and dog clinical conditions; qPCR specificity was high (95-96%), but fell to 81-82% for assays run in plates with suspected contamination. The sensitivity of parasitological procedures was low (~12-13%), but increased to ~33% (with substantial uncertainty) for bone-marrow culture in seriously-diseased dogs. Leishmania-infection frequency estimates (~49-50% across clinical conditions) were lower than observed (~60%).

CONCLUSIONS

We provide statistical estimates of key performance parameters for five diagnostic procedures used to detect Leishmania in dogs. Low clinical sensitivies likely reflect the absence of Leishmania parasites/DNA in perhaps ~50-70% of samples drawn from infected dogs. Although qPCR performance was similar across sample types, non-invasive eye-swabs were overall less likely to contain amplifiable DNA. Finally, modeling was instrumental to discovering (and formally accounting for) possible qPCR-plate contamination; even with stringent negative/blank-control scoring, ~4-5% of positive qPCRs were most likely false-positives. This work shows, in sum, how hierarchical site-occupancy models can sharpen our understanding of the problem of diagnosing host infections with hard-to-detect pathogens including Leishmania.

Integrating basic and applied research to estimate carnivore abundance

A clear connection between basic research and applied management is often missing or difficult to discern. We present a case study of integration of basic research with applied management for estimating abundance of gray wolves (Canis lupus) in Montana, USA. Estimating wolf abundance is a key component of wolf management but is costly and time intensive as wolf populations continue to grow. We developed a multimodel approach using an occupancy model, mechanistic territory model, and empirical group size model to improve abundance estimates while reducing monitoring effort. Whereas field-based wolf counts generally rely on costly, difficult-to-collect monitoring data, especially for larger areas or population sizes, our approach efficiently uses readily available wolf observation data and introduces models focused on biological mechanisms underlying territorial and social behavior. In a three-part process, the occupancy model first estimates the extent of wolf distribution in Montana, based on environmental covariates and wolf observations. The spatially explicit mechanistic territory model predicts territory sizes using simple behavioral rules and data on prey resources, terrain ruggedness, and human density. Together, these models predict the number of packs. An empirical pack size model based on 14 years of data demonstrates that pack sizes are positively related to local densities of packs, and negatively related to terrain ruggedness, local mortalities, and intensity of harvest management. Total abundance estimates for given areas are derived by combining estimated numbers of packs and pack sizes. We estimated the Montana wolf population to be smallest in the first year of our study, with 91 packs and 654 wolves in 2007, followed by a population peak in 2011 with 1252 wolves. The population declined ~6% thereafter, coincident with implementation of legal harvest in Montana. Recent numbers have largely stabilized at an average of 191 packs and 1141 wolves from 2016 to 2020. This new approach accounts for biologically based, spatially explicit predictions of behavior to provide more accurate estimates of carnivore abundance at finer spatial scales. By integrating basic and applied research, our approach can therefore better inform decision-making and meet management needs.

Designing a surveillance program for early detection of alien plants and insects in Norway

Naturalized species of alien plants and animals comprise < 3% of biodiversity recorded in Norway but have had major impacts on natural ecosystems through displacement of native species. Encroachment of alien species has been especially problematic for coastal sites close to transport facilities and urban areas with high density housing. The goal of our field project was to design and test a surveillance program for early detection of alien species of vascular plants and terrestrial insects at the first phase of establishment in natural areas. In our 3-year project (2018–2020), we sampled 60 study plots in three counties in the Oslofjord region of southern Norway. Study plots (6.25 ha) were selected by two criteria: manual selection based on expert opinion (27 plots) or by random selection based on weights from a hotspot analysis of occurrence of alien species (33 plots). Vascular plants were surveyed by two experienced botanists who found a total of 239 alien species of vascular plants in 95 rounds of surveys. Insects and other invertebrates were captured with a single Malaise trap per site, with 3–4 rounds of repeated sampling. We used DNA-metabarcoding to identify invertebrates based on DNA extractions from crushed insects or from the preservative media. Over 3500 invertebrate taxa were detected in 255 rounds of sampling. We recorded 20 alien species of known risk, and 115 species that were new to Norway, including several ‘doorknocker’ species identified by previous risk assessments. We modeled the probabilities of occupancy (ψ) and detection (p) with occupancy models with repeated visits by multiple observers (vascular plants) or multiple rounds of sampling (insects). The two probabilities covaried with risk category for alien organisms and both were low for species categorized as no known or low risk (range = 0.052–0.326) but were higher for species categorized as severe risk (range = 0.318–0.651). Selecting sites at random or manually did not improve the probability of finding novel alien species, but occupancy had a weak positive relationship with housing density for some categories of alien plants and insects. We used our empirical estimates to test alternative sampling designs that would minimize the combined variance of occupancy and detection (A-optimality criterion). Sampling designs with 8–10 visits per site were best for surveillance of new alien species if the probabilities of occupancy and detection were both low, and provided low conditional probabilities of site occupancy ($$\hat{\psi }_{condl}$$ ψ ^ condl ≤ 0.032) and a high probabilities of cumulative detection ($$\hat{p}*$$ p ^ ∗  ≥ 0.943). Our field results demonstrate that early detection is feasible as a key component of a national surveillance program based on early detection and rapid response.

Forest disturbance and occupancy patterns of American ermine (Mustela richardsonii) and long-tailed weasel (Neogale frenata): results from a large-scale natural experiment in Maine, United States

Weasels are small mustelid carnivores that play an important role as predators of small mammals in a wide array of ecosystems. However, their response to land use, such as forest harvest for timber products, is seldom the subject of focused research and management projects. Both the American ermine, also known as the short-tailed weasel (Mustela richardsonii), and the long-tailed weasel (Neogale frenata) are native to Maine, United States, where commercial timber harvesting is widespread. The effects of this forest disturbance on weasels are poorly understood, so to contribute toward filling this knowledge gap, we conducted a 4-year, large-scale field study: specifically, our objective was to assess the effects of forest disturbance caused by timber harvest on occupancy patterns of ermines and long-tailed weasels occupancy patterns in Maine. We collected data from 197 survey sites (three camera traps each) over 4 years and analyzed over 7,000 images of weasels using dynamic false-positive occupancy models. We found that American ermines were widely distributed across the state (naïve occupancy at 54% of sites), while long-tailed weasels were rarer (naïve occupancy at 16% of sites). Both species responded positively to forest disturbance, with higher occupancy probabilities as disturbance increased, especially at the larger scales. American ermines were more likely to occupy stands with a higher percentage of conifer trees, while no such relationship was found for long-tailed weasels. We conclude that current forest harvest practices in Maine are not detrimental to weasel populations, but that the two species warrant continued monitoring.

Tracking an invasion front with environmental DNA

Data from environmental DNA (eDNA) may revolutionize environmental monitoring and management, providing increased detection sensitivity at reduced cost and survey effort. However, eDNA data are rarely used in decision-making contexts, mainly due to uncertainty around (1) data interpretation and (2) whether and how molecular tools dovetail with existing management efforts. We address these challenges by jointly modeling eDNA detection via qPCR and traditional trap data to estimate the density of invasive European green crab (Carcinus maenas), a species for which, historically, baited traps have been used for both detection and control. Our analytical framework simultaneously quantifies uncertainty in both detection methods and provides a robust way of integrating different data streams into management processes. Moreover, the joint model makes clear the marginal information benefit of adding eDNA (or any other) additional data type to an existing monitoring program, offering a path to optimizing sampling efforts for species of management interest. Here, we document green crab eDNA beyond the previously known invasion front and find that the value of eDNA data dramatically increases with low population densities and low traditional sampling effort, as is often the case at leading-edge locations. We also highlight the detection limits of the molecular assay used in this study, as well as scenarios under which eDNA sampling is unlikely to improve existing management efforts.

Using interview surveys and multispecies occupancy models to inform vertebrate conservation

Species distribution data are an essential biodiversity variable requiring robust monitoring to inform wildlife conservation. Yet, such data remain inherently sparse because of the logistical challenges of monitoring biodiversity across broad geographic extents. Surveys of people knowledgeable about the occurrence of wildlife provide an opportunity to evaluate species distributions and the ecology of wildlife communities across large spatial scales. We analyzed detection histories of 30 vertebrate species across the Western Ghats biodiversity hotspot in India, obtained from a large-scale interview survey of 2318 people who live and work in the forests of this region. We developed a multispecies occupancy model that simultaneously corrected for false-negative (non-detection) and false-positive (misidentification) errors that interview surveys can be prone to. Using this model, we integrated data across species in composite analyses of the responses of functional species groups (based on disturbance tolerance, diet, and body mass traits) to spatial variation in environmental variables, protection, and anthropogenic pressures. We observed a positive association between forest cover and the occurrence of species with low tolerance of human disturbance. Protected areas were associated with higher occurrence for species across different functional groups compared with unprotected lands. We also observed the occurrence of species with low disturbance tolerance, herbivores, and large-bodied species was negatively associated with developmental pressures, such as human settlements, energy production and mining, and demographic pressures, such as biological resource extraction. For the conservation of threatened vertebrates, our work underscores the importance of maintaining forest cover and reducing deforestation within and outside protected areas, respectively. In addition, mitigating a suite of pervasive human pressures is also crucial for wildlife conservation in one of the world's most densely populated biodiversity hotspots.

Accounting for detection probability with overestimation by integrating double monitoring programs over 40 years

In the context of wildlife population declines, increasing computer power over the last 20 years allowed wildlife managers to apply advanced statistical techniques that has improved population size estimates. However, respecting the assumptions of the models that consider the probability of detection, such as N-mixture models, requires the implementation of a rigorous monitoring protocol with several replicate survey occasions and no double counting that are hardly adaptable to field conditions. When the logistical, economic and ecological constraints are too strong to meet model assumptions, it may be possible to combine data from independent surveys into the modelling framework in order to understand population dynamics more reliably. Here, we present a state-space model with an error process modelled on the log scale to evaluate wintering waterfowl numbers in the Camargue, southern France, while taking a conditional probability of detection into consideration. Conditional probability of detection corresponds to estimation of a detection probability index, which is not a true probability of detection, but rather conditional on the difference to a particular baseline. The large number of sites (wetlands within the Camargue delta) and years monitored (44) provide significant information to combine both terrestrial and aerial surveys (which constituted spatially and temporally replicated counts) to estimate a conditional probability of detection, while accounting for false-positive counting errors and changes in observers over the study period. The model estimates abundance indices of wintering Common Teal, Mallard and Common Coot, all species abundant in the area. We found that raw counts were underestimated compared to the predicted population size. The model-based data integration approach as described here seems like a promising solution that takes advantage of as much as possible of the data collected from several methods when the logistic constraints do not allow the implementation of a permanent monitoring and analysis protocol that takes into account the detectability of individuals.

Measuring protected-area effectiveness using vertebrate distributions from leech iDNA

Protected areas are key to meeting biodiversity conservation goals, but direct measures of effectiveness have proven difficult to obtain. We address this challenge by using environmental DNA from leech-ingested bloodmeals to estimate spatially-resolved vertebrate occupancies across the 677 km² Ailaoshan reserve in Yunnan, China. From 30,468 leeches collected by 163 park rangers across 172 patrol areas, we identify 86 vertebrate species, including amphibians, mammals, birds and squamates. Multi-species occupancy modelling shows that species richness increases with elevation and distance to reserve edge. Most large mammals (e.g. sambar, black bear, serow, tufted deer) follow this pattern; the exceptions are the three domestic mammal species (cows, sheep, goats) and muntjak deer, which are more common at lower elevations. Vertebrate occupancies are a direct measure of conservation outcomes that can help guide protected-area management and improve the contributions that protected areas make towards global biodiversity goals. Here, we show the feasibility of using invertebrate-derived DNA to estimate spatially-resolved vertebrate occupancies across entire protected areas.

Invertebrate-derived eDNA (iDNA) is an emerging tool for taxonomic and spatial biodiversity monitoring. Here, the authors use metabarcoding of leech-derived iDNA to estimate vertebrate occupancy over an entire protected area, the Ailaoshan Nature Reserve, China.

Fine-scale ecological and anthropogenic variables predict the habitat use and detectability of sloth bears in the Churia habitat of east Nepal

Once widespread throughout the tropical forests of the Indian Subcontinent, the sloth bears have suffered a rapid range collapse and local extirpations in the recent decades. A significant portion of their current distribution range is situated outside of the protected areas (PAs). These unprotected sloth bear populations are under tremendous human pressures, but little is known about the patterns and determinants of their occurrence in most of these regions. The situation is more prevalent in Nepal where virtually no systematic information is available for sloth bears living outside of the PAs. We undertook a spatially replicated sign survey-based single-season occupancy study intending to overcome this information gap for the sloth bear populations residing in the Trijuga forest of southeast Nepal. Sloth bear sign detection histories and field-based covariates data were collected between 2 October and 3 December 2020 at the 74 randomly chosen 4-km2 grid cells. From our results, the model-averaged site use probability (ψ ± SE) was estimated to be 0.432 ± 0.039, which is a 13% increase from the naïve estimate (0.297) not accounting for imperfect detections of sloth bear signs. The presence of termite mound and the distance to the nearest water source were the most important variables affecting the habitat use probability of sloth bears. The average site-level detectability (p ± SE) of sloth bear signs was estimated to be 0.195 ± 0.003 and was significantly determined by the index of human disturbances. We recommend considering the importance of fine-scale ecological and anthropogenic factors in predicting the sloth bear-habitat relationships across their range in the Churia habitat of Nepal, and more specifically in the unprotected areas.

Why Are Species’ Traits Weak Predictors of Range Shifts?

We examine the evidence linking species’ traits to contemporary range shifts and find they are poor predictors of range shifts that have occurred over decades to a century. We then discuss reasons for the poor performance of traits for describing interspecific variation in range shifts from two perspectives: ( a) factors associated with species’ traits that degrade range-shift signals stemming from the measures used for species’ traits, traits that are typically not analyzed, and the influence of phylogeny on range-shift potential and ( b) issues in quantifying range shifts and relating them to species’ traits due to imperfect detection of species, differences in the responses of altitudinal and latitudinal ranges, and emphasis on testing linear relationships between traits and range shifts instead of nonlinear responses. Improving trait-based approaches requires a recognition that traits within individuals interact in unexpected ways and that different combinations of traits may be functionally equivalent.

Patch utilization and flower visitations by wild bees in a honey bee-dominated, grassland landscape

Understanding habitat needs and patch utilization of wild and managed bees has been identified as a national research priority in the United States. We used occupancy models to investigate patterns of bee use across 1030 transects spanning a gradient of floral resource abundance and richness and distance from apiaries in the Prairie Pothole Region (PPR) of the United States. Estimates of transect use by honey bees were nearly 1.0 during our 3.5-month sampling period, suggesting honey bees were nearly ubiquitous across transects. Wild bees more frequently used transects with higher flower richness and more abundant flowers; however, the effect size of the native flower abundance covariate (β ^ native  = 3.90 ± 0.65 [1SE]) was four times greater than the non-native flower covariate (β ^ n o n - n a t i v e  = 0.99 ± 0.17). We found some evidence that wild bee use was lower at transects near commercial apiaries, but the effect size was imprecise (β ^ distance  = 1.4 ± 0.81). Honey bees were more frequently detected during sampling events with more non-native flowers and higher species richness but showed an uncertain relationship with native flower abundance. Of the 4039 honey bee and flower interactions, 85% occurred on non-native flowers, while only 43% of the 738 wild bee observations occurred on non-native flowers. Our study suggests wild bees and honey bees routinely use the same resource patches in the PPR but often visit different flowering plants. The greatest potential for resource overlap between honey bees and wild bees appears to be for non-native flowers in the PPR. Our results are valuable to natural resource managers tasked with supporting habitat for managed and wild pollinators in agroecosystems.

Addressing context dependence in ecology

Context dependence is widely invoked to explain disparate results in ecology. It arises when the magnitude or sign of a relationship varies due to the conditions under which it is observed. Such variation, especially when unexplained, can lead to spurious or seemingly contradictory conclusions, which can limit understanding and our ability to transfer findings across studies, space, and time. Using examples from biological invasions, we identify two types of context dependence resulting from four sources: mechanistic context dependence arises from interaction effects; and apparent context dependence can arise from the presence of confounding factors, problems of statistical inference, and methodological differences among studies. Addressing context dependence is a critical challenge in ecology, essential for increased understanding and prediction.

The critical role of natural forest as refugium for generalist species in oil palm-dominated landscapes

In Borneo, oil palm plantations have replaced much of natural resources, where generalist species tend to be the principal beneficiaries, due to the abundant food provided by oil palm plantations. Here, we analyse the distribution of the Asian water monitor lizard (Varanus salvator) population within an oil palm-dominated landscape in the Kinabatangan floodplain, Malaysian Borneo. By using mark-recapture methods we estimated its population size, survival, and growth in forest and plantation habitats. We compared body measurements (i.e. body weight and body length) of individuals living in forest and oil palm habitats as proxy for the population's health status, and used general least squares estimation models to evaluate its response to highly fragmented landscapes in the absence of intensive hunting pressures. Contrary to previous studies, the abundance of lizards was higher in the forest than in oil palm plantations. Recruitment rates were also higher in the forest, suggesting that these areas may function as a source of new individuals into the landscape. While there were no morphometric differences among plantation sites, we found significant differences among forested areas, where larger lizards were found inhabiting forest adjacent to oil palm plantations. Although abundant in food resources, the limited availability of refugia in oil palm plantations may intensify intra-specific encounters and competition, altering the body size distribution in plantation populations, contrary to what happens in the forest. We conclude that large patches of forest, around and within oil palm plantations, are essential for the dynamics of the monitor lizard population in the Kinabatangan floodplain, as well as a potential source of individuals to the landscape. We recommend assessing this effect in other generalist species, as well as the impact on the prey communities, especially to reinforce the establishment of buffer zones and corridors as a conservation strategy within plantations.

Detection criteria and post-field sample processing influence results and cost efficiency of occupancy-based monitoring

Optimization of occupancy-based monitoring has focused on balancing the number of sites and surveys to minimize field efforts and costs. When survey techniques require post-field processing of samples to confirm species detections, there may be opportunities to further improve efficiency. We used scat-based noninvasive genetic sampling for kit foxes (Vulpes macrotis) in Utah, USA, as a model system to assess post-field data processing strategies, evaluate the impacts of these strategies on estimates of occupancy and associations between parameters and predictors, and identify the most cost-effective approach. We identified scats with three criteria that varied in costs and reliability: (1) field-based identification (expert opinion), (2) statistical-based morphological identification, and (3) genetic-based identification (mitochondrial DNA). We also considered four novel post-field sample processing strategies that integrated statistical and genetic identifications to reduce costly genetic procedures, including (4) a combined statistical-genetic identification, (5) a genetic removal design, (6) a within-survey conditional-replicate design, and (7) a single-genetic-replicate with false-positive modeling design. We considered results based on genetic identification as the best approximation of truth and used this to evaluate the performance of alternatives. Field-based and statistical-based criteria prone to misidentification produced estimates of occupancy that were biased high (˜1.8 and 2.1 times higher than estimates without misidentifications, respectively). These criteria failed to recover associations between parameters and predictors consistent with genetic identification. The genetic removal design performed poorly, with limited detections leading to estimates that were biased high with poor precision and patterns inconsistent with genetic identification. Both statistical-genetic identification and the conditional-replicate design produced occupancy estimates comparable to genetic identification, while recovering the same model structure and associations at cost reductions of 67% and 74%, respectively. The false-positive design had the lowest cost (88% reduction) and recovered patterns consistent with genetic identification but had occupancy estimates that were ˜32% lower than estimated occupancy based on genetic identification. Our results demonstrate that careful consideration of detection criteria and post-field data processing can reduce costs without significantly altering resulting inferences. Combined with earlier guidance on sampling designs for occupancy modeling, these findings can aid managers in optimizing occupancy-based monitoring.

Conditional natal dispersal provides a mechanism for populations tracking resource pulses after fire

Animals that persist in spatially structured populations face the challenge of tracking the rise and fall of resources across space and time. To combat these challenges, theory predicts that species should use conditional dispersal strategies that allow them to emigrate from patches with declining resources and colonize new resource patches as they appear. We studied natal dispersal movements in the black-backed woodpecker (Picoides arcticus), a species known for its strong association with recent post-fire forests in western North America. We radio-tracked juveniles originating from seven burned areas and tested hypotheses that environmental and individual factors influence dispersal distance and emigration rates—investigating emigration while additionally accounting for imperfect detection with a novel Bayesian model. We found that juveniles were more likely to leave natal areas and disperse longer distances if they were heavier or hatched in older burned areas where resources are increasingly scarce. Juveniles were also more likely to leave their natal burn if they hatched in a nest closer to the fire perimeter. While dispersing across the landscape, black-backed woodpeckers selected for burned forest relative to unburned available habitat. Together, these results strongly support the hypothesis that black-backed woodpecker populations track resource pulses across fire-prone landscapes, with conditional natal dispersal acting as a mechanism for locating and colonizing newly burned areas. Lending empirical support to theoretical predictions, our findings suggest that changes in resource distribution may shape dispersal patterns and, consequently, the distribution and persistence of spatially structured populations.

Effects of Training on Side-Scan Sonar Use as a Fish Survey Tool: A Case Study of Alligator Gar

Consumer-grade side-scan sonar has become a versatile fisheries management tool. First applied to assess habitat, its use has expanded to surveying fishes in recent years. However, an important consideration is the skill and experience of users, which can affect both the accuracy and comparability of surveys. To this end, we characterized the ability of a small sample of novice users (N = 8) to identify Alligator Gar Atractosteus spatula in imagery, as well as the effect of a 2-h training exercise on user performance. Prior to training, mean accuracy (expressed as the difference between observed and expected counts) among participants ranged from −2.6 to 1.3 fish and precision ranged from ±1.2 to ±2.4 fish, with the majority of participants underestimating the number of Alligator Gar present in the imagery. False positives (i.e., identifying Alligator Gar in imagery when none were present) were common among participants. Posttraining mean accuracy ranged from −3.1 to 0 among participants and precision ranged from ±1.6 to ±3.2 fish. The frequency of false positives was significantly reduced following training, and participants reported significant increases in confidence associated with image interpretation. The relatively high accuracy and precision we observed prior to training indicated that side-scan sonar can be easily incorporated into large-scale fishery monitoring efforts for Alligator Gar. However, our results also suggested that a rather minimal investment in training can further improve consistency and reduce uncertainty among novice users.

A Framework for Estimating Human-Wildlife Conflict Probabilities Conditional on Species Occupancy

Managing human-wildlife conflicts (HWCs) is an important conservation objective for the many terrestrial landscapes dominated by humans. Forecasting where future conflicts are likely to occur and assessing risks to lives and livelihoods posed by wildlife are central to informing HWC management strategies. Existing assessments of the spatial occurrence patterns of HWC are based on either understanding spatial patterns of past conflicts or patterns of species distribution. In the former case, the absence of conflicts at a site cannot be attributed to the absence of the species. In the latter case, the presence of a species may not be an accurate measure of the probability of conflict occurrence. We present a Bayesian hierarchical modeling framework that integrates conflict reporting data and species distribution data, thus allowing the estimation of the probability that conflicts with a species are reported from a site, conditional on the species being present. In doing so, our model corrects for both false-positive and false-negative conflict reporting errors. We provide study design recommendations using simulations that explore the performance of the model under a range of conflict reporting probabilities. We applied the model to data on wild boar (Sus scrofa) space use and conflicts collected from the Central Terai Landscape (CTL), an important tiger conservation landscape in India. We found that tolerance for wildlife was a predictor of the probability with which farmers report conflict with wild boars from sites not used by the species. We also discuss useful extensions of the model when conflict data are verified for potential false-positive errors and when landscapes are monitored over multiple seasons.

Effects of landscape composition on wetland occupancy by Blanding’s Turtles (Emydoidea blandingii) as determined by environmental DNA and visual surveys

Habitat loss and degradation have led to the extinction of many species worldwide. The endangered Blanding’s Turtle (Emydoidea blandingii (Holbrook, 1838)), a semi-aquatic freshwater turtle, occupies a wide range of wetlands and landscapes primarily in southeastern Canada and the Great Lakes region of the United States. We explored whether the probability of wetland occupancy by Blanding’s Turtles is affected by the surrounding landscape. We used visual surveys, environmental DNA, and Atlas data to document the presence of Blanding’s Turtles in wetlands in Ottawa, Ontario, Canada. We tabulated landscape composition at multiple scales surrounding the wetlands to determine whether landscape composition can predict wetland occupancy. Generally, wetlands surrounded by forest and other undisturbed lands were most likely to harbour Blanding’s Turtles, whereas those surrounded by more human-disturbed lands were least likely to harbour Blanding’s Turtles. Larger wetlands and a high proportion of wetlands in the surrounding landscape also increased the probability of occupancy by Blanding’s Turtles. Finally, older wetlands were more likely to be occupied by Blanding’s Turtles. The ability to estimate a species’ probability of occupancy can aid in conservation efforts, such as critical habitat delineation.

Linking camera-trap data to taxonomy: Identifying photographs of morphologically similar chipmunks

Remote cameras are a common method for surveying wildlife and recently have been promoted for implementing large-scale regional biodiversity monitoring programs. The use of camera-trap data depends on the correct identification of animals captured in the photographs, yet misidentification rates can be high, especially when morphologically similar species co-occur, and this can lead to faulty inferences and hinder conservation efforts. Correct identification is dependent on diagnosable taxonomic characters, photograph quality, and the experience and training of the observer. However, keys rooted in taxonomy are rarely used for the identification of camera-trap images and error rates are rarely assessed, even when morphologically similar species are present in the study area. We tested a method for ensuring high identification accuracy using two sympatric and morphologically similar chipmunk (Neotamias) species as a case study. We hypothesized that the identification accuracy would improve with use of the identification key and with observer training, resulting in higher levels of observer confidence and higher levels of agreement among observers. We developed an identification key and tested identification accuracy based on photographs of verified museum specimens. Our results supported predictions for each of these hypotheses. In addition, we validated the method in the field by comparing remote-camera data with live-trapping data. We recommend use of these methods to evaluate error rates and to exclude ambiguous records in camera-trap datasets. We urge that ensuring correct and scientifically defensible species identifications is incumbent on researchers and should be incorporated into the camera-trap workflow.

Improving the reliability of eDNA data interpretation

Global declines in biodiversity highlight the need to effectively monitor the density and distribution of threatened species. In recent years, molecular survey methods detecting DNA released by target-species into their environment (eDNA) have been rapidly on the rise. Despite providing new, cost-effective tools for conservation, eDNA-based methods are prone to errors. Best field and laboratory practices can mitigate some, but the risks of errors cannot be eliminated and need to be accounted for. Here, we synthesize recent advances in data processing tools that increase the reliability of interpretations drawn from eDNA data. We review advances in occupancy models to consider spatial data-structures and simultaneously assess rates of false positive and negative results. Further, we introduce process-based models and the integration of metabarcoding data as complementing approaches to increase the reliability of target-species assessments. These tools will be most effective when capitalizing on multi-source data sets collating eDNA with classical survey and citizen-science approaches, paving the way for more robust decision-making processes in conservation planning.

Occupancy modeling and resampling overcomes low test sensitivity to produce accurate SARS-CoV-2 prevalence estimates

Background

We evaluated whether occupancy modeling, an approach developed for detecting rare wildlife species, could overcome inherent accuracy limitations associated with rapid disease tests to generate fast, accurate, and affordable SARS-CoV-2 prevalence estimates. Occupancy modeling uses repeated sampling to estimate probability of false negative results, like those linked to rapid tests, for generating unbiased prevalence estimates.

Methods

We developed a simulation study to estimate SARS-CoV-2 prevalence using rapid, low-sensitivity, low-cost tests and slower, high-sensitivity, higher cost tests across a range of disease prevalence and sampling strategies.

Results

Occupancy modeling overcame the low sensitivity of rapid tests to generate prevalence estimates comparable to more accurate, slower tests. Moreover, minimal repeated sampling was required to offset low test sensitivity at low disease prevalence (0.1%), when rapid testing is most critical for informing disease management.

Conclusions

Occupancy modeling enables the use of rapid tests to provide accurate, affordable, real-time estimates of the prevalence of emerging infectious diseases like SARS-CoV-2.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-021-10609-y.

Crowdsourcing snake identification with online communities of professional herpetologists and avocational snake enthusiasts

Species identification can be challenging for biologists, healthcare practitioners and members of the general public. Snakes are no exception, and the potential medical consequences of venomous snake misidentification can be significant. Here, we collected data on identification of 100 snake species by building a week-long online citizen science challenge which attracted more than 1000 participants from around the world. We show that a large community including both professional herpetologists and skilled avocational snake enthusiasts with the potential to quickly (less than 2 min) and accurately (69-90%; see text) identify snakes is active online around the clock, but that only a small fraction of community members are proficient at identifying snakes to the species level, even when provided with the snake's geographical origin. Nevertheless, participants showed great enthusiasm and engagement, and our study provides evidence that innovative citizen science/crowdsourcing approaches can play significant roles in training and building capacity. Although identification by an expert familiar with the local snake fauna will always be the gold standard, we suggest that healthcare workers, clinicians, epidemiologists and other parties interested in snakebite could become more connected to these communities, and that professional herpetologists and skilled avocational snake enthusiasts could organize ways to help connect medical professionals to crowdsourcing platforms. Involving skilled avocational snake enthusiasts in decision making could build the capacity of healthcare workers to identify snakes more quickly, specifically and accurately, and ultimately improve snakebite treatment data and outcomes.