Biology subjects

Chamaille-Jammes, S.

Publications and source records attributed to Chamaille-Jammes, S..

At least 19 recordsLinked to original sources

bioRxiv2025

Evolutionary and ecological determinants of the phenology of births in wild large herbivores, a systematic review

1. IntroductionTiming of birth is a major determinant of individuals reproductive success and phenotypic performances, driven by multiple factors. However, there is currently no quantitative assessment of the support received by the drivers of the phenology of births in mammals. 2. Aimsour objective was to identify the drivers tested for their effect on the phenology of births in large herbivores and how much empirical support they received, as well as determine thematic, taxonomic and spatial gaps in the literature. 3. MethodsWe conducted a systematic review focusing on the drivers of the phenology of births in large herbivores. We extracted the outcomes of 124 articles and assessed the reliability of their findings using a general index based on the quality of the data and methods used. 4. Results and DiscussionWe identified a major research gap in Asian and South and Central American species. The effect of population and individual characteristics remained marginally accounted for, even though there was evidence of their importance in the determination of the phenology of births. The two dominant hypotheses (seasonality and predation) received good to moderate empirical support regarding their effect on birth timing and synchrony, respectively. Given the strong geographic and methodological biases, these hypotheses should be considered as most consistently supported in the existing literature rather than definitively established across systems. 5. ConclusionsWe encourage research in these understudied domains, as well as large scale multi-species comparative analyses to fill in these gaps and improve our understanding of reproductive phenology. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=97 SRC="FIGDIR/small/694258v2_ufig1.gif" ALT="Figure 1"> View larger version (41K): org.highwire.dtl.DTLVardef@4f1a9org.highwire.dtl.DTLVardef@95c5a6org.highwire.dtl.DTLVardef@e3324forg.highwire.dtl.DTLVardef@1fc1a6f_HPS_FORMAT_FIGEXP M_FIG C_FIG This semi-systematic review supports the two dominant drivers of birth phenology, the seasonality and predation hypotheses. Even though there is evidence of their importance, the effects of female, offspring and population characteristics remain marginally accounted for. Asian and South and Central American species are currently understudied.

ecology↗

bioRxiv2025

How big is enough? Movement-informed zoning for African swine fever mitigation

O_LIAfrican swine fever (ASF) poses a serious threat to domestic pigs and wild boar populations. Wild boar can disperse the virus, making effective containment crucial. One of the main control strategies involves establishing restricted zones around detected cases, i.e., areas with temporary restrictions on access and hunting; however, determining the appropriate size of these zones remains a major challenge. C_LIO_LITo inform the size of restricted zones, we analyzed GPS data from 527 wild boar across 46 European study sites using a two-step approach combining first-passage time analysis and survival modelling to quantify the risk of wild boar leaving areas of different radii (i.e., spatial scales). We investigated how the risk of leaving varied over time and across environmental gradients. To go further, we used our model findings to develop an online application that generates predictive maps of optimal buffer sizes for ASF management at the European scale, based on a given risk threshold (the maximum acceptable probability that a wild boar leaves the area). C_LIO_LIWe found that the relationship between radius and the risk of leaving is negative exponential, and the risk of leaving increased over time, with a more rapid increase for smaller radii. Landscape homogeneity, terrain ruggedness and human impact increased the risk of leaving, with stronger effects at small scales. Contrary to other predictors, agricultural cover exerted a strong effect on risk of leaving over large spatial scales, especially when it was abundant. C_LIO_LIAcross Europe, a buffer radius of [~]8 km is likely sufficient around high-risk infection zones in most areas (considering an infectious period of 14 days and a risk threshold of 5%); however, in certain areas, a radius of up to 20 km may be needed to effectively limit wild boar movement. C_LIO_LISynthesis and applications: Our results highlight the need for adaptive, context-specific restricted zones. Buffers of 8 km around ASF-affected areas can limit the risk of infected wild boar dispersal, but they may be reduced to 5 km in highly heterogeneous landscapes or high-human impacted areas. Larger buffers may be required in agricultural landscapes. We provide spatially explicit outputs (optimal buffer sizes) that can directly inform policy and wildlife disease response strategies. The approach can be adapted to any other infectious disease. C_LI

ecology↗

bioRxiv2025

Using social network analysis and non-invasive antibody detection to explore pathogen exposure in wildlife communities

Understanding the dynamics of infectious diseases in ecosystems shared with wildlife is a priority for preventing future threats to human, animal health and conservation. However, the collection of ecological and epidemiological data on wild populations has relied on invasive and costly methods which limited the capacity of investigation. Recent technological developments have changed this trend. In Hwange national park and its surrounding area in Zimbabwe, we combined camera trap-based ecological monitoring over a period of 12 months, covering 14 water holes with a fecal-based antibody survey in 16 large herbivore species. A survey involving 52,829 pictures as well as 629 faecal samples collected every 15 days, has shown Through the modelling of multispecies contact networks, the risk of infection exposure at species-level is predicted. Coupled with foot-and-mouth disease virus (FMDV) antibody information, the role of each species in the dynamics of infectious diseases is explored. This study highlights how community networks can provide valuable insights into the functional epidemiological role of wildlife populations. Rather than establishing transmission routes, our aim is to propose a scalable and non-invasive surveillance framework that identifies priority species and areas for epidemiological monitoring in complex ecological systems. Significance StatementMonitoring disease circulation in wildlife is often hindered by the difficulty of collecting epidemiological data. We propose a novel non-invasive approach that combines species interaction networks derived from camera-trap data with non-invasive antibody detection to explore the exposure patterns of large herbivore communities to foot-and-mouth disease virus (FMDV). By linking the position of species in contact networks with their immunological status, we demonstrate the potential of using ecological centrality as a proxy for identifying key hosts in transmission or indicator species for pathogen circulation. Beyond FMDV, this framework can be adapted to other pathogens for which non-invasive immunological assays are currently being developed, making it broadly relevant for wildlife disease surveillance in remote or protected areas.

ecology↗

bioRxiv2025

Paying Attention to Other Animal Detections Improves Camera Trap Classification Models

O_LIIn ecological studies, automated species classification models are increasingly used to process large volumes of camera trap images. Most current classification models rely on a two-step pipeline: a detector first locates and crops animals, followed by a classifier that predicts species independently for each crop. While effective, these models still struggle under challenging conditions and ignore temporal context or information from nearby animals available in sequences of camera trap images, whereas human annotators often use it in difficult cases. C_LIO_LIWe propose to leverage self-attention, a core mechanism of Large Language Models and Vision Transformers, to enable the model to learn relationships between crops occurring in similar contexts. Our self-attention module operates directly on the set of crop embeddings, producing new representations enriched with information from other crops, improving species classification. The module fits into the two-step pipeline without requiring structural change and adds only minimal computational overhead. To address the lack of annotated multi-species sequences, we develop a training strategy that synthetically generates multi-species sequences from monospecific ones. C_LIO_LICompared to an independent classifier baseline, our multi-crop model achieves higher accuracy on mono-specific sequences, both real and synthetic. In multispecies settings, evaluated using synthetic test sets, we also observe a substantial improvement in accuracy. Using real but weakly annotated multi-species sequences, we reformulate the task as multi-label set classification, and conduct a visual analysis, to highlight the benefits of our approach. C_LIO_LIBy leveraging the information brought by all detections of animals in the image and others of the same sequence, our approach reduces species misclassifications and enables more accurate estimates for downstream ecological analyses focusing on, for instance, species richness, occupancy, and species interaction. C_LI

ecology↗

bioRxiv2025

Spatial covariation between wild boars and other mammals in peri-urban landscapes: insights for management from southern France

Global urbanization is rapidly increasing, transforming ecosystems and favoring generalist species that adapt to human-modified environments. The wild boar, a highly adaptable species, is expanding its range and increasingly observed in urban and peri-urban areas, where it can cause accidents, damage, and health concerns. To mitigate their presence, non-lethal control strategies such as vegetation clearing in peri-urban parks and wastelands are being implemented to reduce the number of suitable resting sites. However, the ecological consequences of such habitat modifications on other wildlife species remain poorly understood. This knowledge gap limits the development of effective urban wildlife management strategies that seeks to balance ecological, social and economic considerations without harming wild boar populations and urban ecosystems. In this context, we used camera trapping to study the covariation between wild boars and other mammal species in a peri-urban area, with the aim of guiding targeted and effective management strategies. Here, we show a widespread presence of wild boars but no spatial or temporal covariation between the wild boars and the other species. In contrast, we found strong spatial structuring in mammal communities, primarily driven by small carnivores (mustelids, genets, cats, foxes) and a minimal influence of temporal variation, confirming that spatial factors, rather than seasonal changes, largely shaped species distributions. The study of the temporal variation across months indicated minimal seasonal structuring but revealed substantial within-site variability in month-to-month species abundance, with a general abundance gradient driven by common species like wild boar, cat, and fox. In conclusion, our results show that the lack of spatial covariation between wild boar and other species prevents reliable predictions, supporting a targeted rather than general approach to vegetation management.

ecology↗

bioRxiv2025

Are all waterholes equal from a lion's view? Exploring the role of prey abundance and catchability in waterhole visitation patterns in a savannah ecosystem

Prey abundance and catchability shape the spatial ecology of predators. Predators can select habitats where prey are more abundant to maximize encounter rate with prey or habitats where prey are more catchable to maximize prey capture. These hypotheses are commonly referred to as prey-abundance and prey-catchability hypotheses. Although these hypotheses are often tested at the landscape scale, little is known about how between-patch variations in prey abundance and catchability determine the space use of predators. In many savannah ecosystems, large herbivores aggregate around waterholes, which become hotspots of prey and their selection by predators is classically interpreted as supporting the prey-abundance hypothesis. Here, we investigated whether between-waterhole variations in prey abundance and catchability influence the frequency and duration of lion visits to waterholes, testing the prey-abundance and prey-catchability hypotheses at the resource-patch scale. We combined datasets on (1) lion movements recorded from GPS collars deployed on 20 adult males and 16 adult females between 2002 and 2015, (2) prey abundance evaluated from long-term, regular monitoring of waterholes and (3) prey catchability evaluated from remote-sensing satellite imagery of vegetation cover around waterholes in Hwange National Park (Zimbabwe). Lions did not use all waterholes in their territory equally: there was a high variability in the frequency and duration of visits. Surprisingly, between-waterhole variations in prey abundance and catchability only slightly explained these variations in frequency - and even less in duration - of lion visits to waterholes. Yet, the frequency of lion visits to waterholes decreased with the number of waterholes within their territory, and male lions more frequently visited the waterholes surrounded by more open habitats. We discuss the limits of our work, but also the ecological mechanisms that may explain these findings. First, lions and their prey are involved in a shell-game that leads them to adopt unpredictable movement strategies. Second, lions have only access to a limited number of waterholes amongst which to distribute their hunting effort. Lastly, lions visit waterholes not only to hunt but also to interact with social mates and competitors. This work challenges the implicit assumption that all waterholes are the same from a lions view and calls for further studies investigating the drivers of the variability in lion visits at the resource-patch scale.

ecology↗

bioRxiv2024

Spatio-temporal dynamics of attacks around deaths of wolves: A statistical assessment of lethal control efficiency in France

The lethal control of large carnivores is criticized regarding its efficiency to prevent hotspots of attacks on livestock. Previous studies, mainly focused on North America, provided mixed results. We evaluated the effects of wolf lethal removals on the distribution of attack intensities in the French Alps between 2011 and 2020, using a Before After Control-Impact approach with retrospective data. We built an original framework combining both continuous spatial and temporal scales and a 3D kernel estimation. We compared the attack intensities observed before and after the legal killings of wolves on a period of 90 days and on a range of 10 km, and with control situations where no removal occurred. The analysis was corrected for the presence of livestock. A moderate decrease of attack intensity was the most common outcome after the lethal removal of a single wolf. This reduction was greatly amplified when removing two or three wolves. The scale of analysis also modulated this general pattern, with decreases being generally amplified at a small spatio-temporal range. Contextual factors (e.g., geographical or seasonal variations) could also lead to deviations from this general pattern. Overall, between 2011 and 2020, lethal control of wolves in France generally contributed to reduce livestock attacks, but mainly locally and to a minor extent. Our results highlight the importance of accounting for scale in such assessments and suggest that the evaluation of the effectiveness of lethal removals in reducing livestock predation might be more relevant in a local context. As a next step, we recommend to move forward from patterns to mechanisms by linking the effects of lethal control on wolves to their effects on attacks through analysis of fine-scaled data on wolves and livestock.

ecology↗

bioRxiv2024

Do mixed-species groups travel as one? An investigation on large African herbivores using animal-borne video collars

Although prey foraging in mixed-species groups benefit from a reduced risk of predation, whether heterospecific groupmates move together in the landscape, and more generally to what extent mixed-species groups remain cohesive over time and space remains unknown. Here, we used GPS collars with video cameras to investigate the movements of plains zebras (Equus quagga) in mixed-species groups. Blue wildebeest (Connochaetes taurinus), impalas (Aepyceros melampus) and giraffes (Giraffa camelopardalis) commonly form mixed-species groups with zebras in savanna ecosystems. We found that zebras adjust their movement decisions solely to the presence of giraffes, being more likely to move in zebra-giraffe herds, and this was correlated to a higher cohesion of such groups. Additionally, zebras moving with giraffes spent longer time grazing, suggesting that zebras follow giraffes to forage in their proximity. Our results provide new insights on animal movements in mixed-species groups, contributing to a better consideration of mutualism in movement ecology.

ecology↗

bioRxiv2024

Zero-shot animal behavior classification with image-text foundation models

1. Understanding the behavior of animals in their natural habitats is critical to ecology and conservation. Camera traps are a powerful tool to collect such data with minimal disturbance. They however produce very a large quantity of images, which can make human-based annotation cumbersome or even impossible. While automated species identification with artificial intelligence has made impressive progress, automatic classification of animal behaviors in camera trap images remains a developing field. 2. Here, we explore the potential of foundation models, specifically Vision Language Models (VLMs), to perform this task without the need to first train a model, which would require some level of human-based annotation. Using an original dataset of alpine fauna with behaviors annotated by participatory science, we investigate the zero-shot capabilities of different kind of recent VLMs to predict behaviors and estimate behavior-specific diel activity patterns in three ungulate species. 3. Our results show that using these methods, it is possible to achieve accuracies over 91% in behavior classification and produce activity patterns that closely align with those derived from participatory science data (overlap indexes between 84% and 90%). 4. These findings demonstrate the potential of foundation models and vision-language models in ecological research. Ecologists are encouraged to adopt these new methods and leverage their full capabilities to facilitate ecological studies.

ecology↗

bioRxiv2024

Can overlooking "invisible landscapes" bias habitat selection estimation and population distribution projections?

Species future distributions are commonly predicted using models that link the likelihood of occurrence of individuals to the environment. Although animals movements are influenced by physical landscapes and individual experiences (for example space familiarity), species distribution models developed from observations of unknown individuals cannot integrate these latter variables, turning them into invisible landscapes. In this theoretical study, we address how overlooking invisible landscapes impacts the estimation of habitat selection and thereby the projection of future distributions. Overlooking the attraction towards some invisible variable consistently led to over-estimating the strength of habitat selection. Consequently, projections of future population distributions were also biased, with animals tracking habitat changes less than predicted. Our results reveal an overlooked challenge faced by correlative species distribution models based on the observation of unknown individuals, whose past experience of the environment is by definition not known. Mechanistic distribution modelling integrating cognitive processes underlying movement should be developed.

ecology↗

bioRxiv2023

Beyond accuracy : score calibration in deep learning models for camera trap image sequences

In ecological studies, machine learning models are increasingly being used for the automatic processing of camera trap images. Although this automation facilitates and accelerates the identification step, the results of these models may lack interpretability and their immediate applicability to ecological downstream tasks (e.g occupancy estimation) remain questionable. In particular, little is known about their calibration, a property that guarantees that confidence scores can be reliably interpreted as probabilities that a models predictions are true. Using a large and diverse European camera trap dataset, we investigate whether deep learning models for species classification in camera trap images are well calibrated, or in contrast over/under-confident. Additionally, as camera traps are often configured to take multiple photos of the same event, we also explore the calibration of predictions at the sequence level. Finally, we study the effect and the practicality of a post-hoc calibration method, i.e. temperature scaling, for predictions made at image and sequence levels. Based on five established models and three independent test sets, our findings show that, using the right methodology, it is possible to enhance the interpretability of the confidence scores, with clear implication for, for instance, the calculation of error rates or the selection of confidence score thresholds in ecological studies making use of artificial intelligence models.

ecology↗

bioRxiv2023

MoveFormer: a Transformer-based model for step-selection animal movement modelling

The movement of animals is a central component of their behavioural strategies. Statistical tools for movement data analysis, however, have long been limited, and in particular, unable to account for past movement information except in a very simplified way. In this work, we propose MoveFormer, a new step-based model of movement capable of learning directly from full animal trajectories. While inspired by the classical step-selection framework and previous work on the quantification of uncertainty in movement predictions, MoveFormer also builds upon recent developments in deep learning, such as the Transformer architecture, allowing it to incorporate long temporal contexts. The model predicts an animals next movement step given its past movement history, including not only purely positional and temporal information, but also any available environmental covariates such as land cover or temperature. We apply our model to a diverse dataset made up of over 1550 trajectories from over 100 studies, and show how it can be used to gain insights about the importance of the provided context features, including the extent of past movement history. Our software, along with the trained model weights, is released as open source.

animal behavior and cognition↗

bioRxiv2023

Hogs sleep like logs: wild boars reduce the risk of anthropic disturbance by adjusting where they rest

Many animals living in anthropized landscapes try to avoid encountering people by being active at night. By doing so, however, they risk being disturbed while at rest during the day. To mitigate this risk, diurnally resting species may be highly selective about where they rest. Here, we used GPS and activity sensors to study how wild boars (Sus scrofa) adjust their resting site selection and visitation patterns to the risk of disturbance by people. The data was complemented with audio recordings from animal-borne loggers to evaluate the efficacy of wild boars resting strategy in reducing the probability of encountering people while at rest. Generally, we found that wild boars did not specifically avoid resting near villages or roads, i.e. where the risk of encountering people is higher, if they could find sites with suitable vegetation cover. Wild boars could actually rest close to villages with very little risk of being disturbed. Resting sites located close to villages were visited more repeatedly that those located further away, suggesting that focusing on a few familiar and quiet resting sites was a successful strategy for resting undisturbed in an anthropized landscape.

animal behavior and cognition↗

bioRxiv2023

Time-varying habitat selection analysis: a model and applications for studying diel, seasonal and post-release changes

Resource selection functions are commonly employed to evaluate animals habitat selection, e.g. the disproportionate use of habitats relative to their availability. While environmental conditions or animal motivations may vary over time, sometimes in an unknown manner, studying changes in habitat selection usually requires a priori time discretization. This limits our ability to precisely answer the question when is an animals habitat selection changing?. Here, we present a straightforward and flexible alternative approach based on fitting dynamic logistic models to used/available data. First, using simulated dataset, we demonstrate that dynamic logistic models performed well to recover temporal variations in habitat selection. We then show real-world applications for studying diel, seasonal, and post-release changes in habitat selection of blue wildebeest (Connochaetes taurinus). Finally, we provide the relevant R scripts to facilitate the adoption of the method by ecologists. Dynamic logistic models allow to study temporal changes in habitat selection in a framework consistent with resource selection functions, but without the need to discretize time, which can be a difficult task when little is known about the process studied, or may obscure inter-individual variability in timing of change. These models should undoubtedly find their place in the movement ecology toolbox.

ecology↗

bioRxiv2022

The DeepFaune initiative: a collaborative effort towards the automatic identification of the French fauna in camera-trap images

Camera traps have revolutionized how ecologists monitor wildlife, but their full potential is realized only when the hundreds of thousands of collected images can be readily classified with minimal human intervention. Deep-learning classification models have allowed extraordinary progress towards this end, but trained models remain rare and are only now emerging for European fauna. We report on the first milestone of the DeepFaune initiative (https://www.deepfaune.cnrs.fr), a large-scale collaboration between more than 50 partners involved in wildlife research, conservation and management in France. We developed a classification model trained to recognize 26 species or higher-level taxa. The classification model achieved 0.97 validation accuracy and often >0.95 precision and recall for many classes. These performances were generally higher than 0.90 when tested on independent out-of-sample datasets for which we used image redundancy contained in sequence of images. We implemented our model in a software to classify images stored locally on a personal computer, so as to provide a free, user-friendly and high-performance tool for wildlife practitioners to automatically classify camera-trap images.

ecology↗

bioRxiv2021

Does timing of birth affect juvenile and mare survival in wild plains zebra?

In large herbivores, the timing of births often coincides with the seasonal peak of food resources availability, likely to improve juvenile survival and reduce reproduction costs. Some species however, breed year-round, even in seasonal environments. Demographic processes, such as to what extent being born during the lean season reduces survival of juveniles and reproductive females, remain understudied in large mammals inhabiting tropical ecosystems. We investigated survival rates in plains zebras (Equus quagga) in Hwange National Park (Zimbabwe), a highly seasonal savanna ecosystem. We used capture-recapture models to analyse long-term demographic data (2008-2019). We investigated the effect of seasonality as a categorical (wet versus dry season) and continuous (duration of the dry season) variable on survival. We found little variability in early juvenile survival ({varphi} = 0.458 {+/-} 0.044 SE, < 6 m.o.), whereas late juvenile and yearling survival was higher and decreased with increasing length of the dry season (from 0.850 {+/-} 0.095 SE to 0.480 {+/-} 0.120 SE). Female survival was high (> 0.703 {+/-} 0.057 SE and up to 0.995 {+/-} 0.006 SE) but decreased with increasing exposure to the dry season in non-reproductive females. The probability of females becoming reproductive in the following year was not affected by the length of the dry season (0.423 and 0.420 for reproductive and non-reproductive females respectively). Our results highlight the importance of individual quality in reproductive performance, as reproductive females seem to buffer the effect of environmental variability on their own survival and that of their foal.

ecology↗

bioRxiv2021

How to describe and measure phenology? An investigation on the diversity of metrics using phenology of births in large herbivores

Proposed in 1849 by Charles Morren to depict periodical phenomena governed by seasons, the term "phenology" has spread in many fields of biology. With the wide adoption of the concept of phenology flourished a large number of metrics with different meaning and interpretation. Here, we first a priori classified 52 previously published metrics used to characterise the phenology of births in large herbivores according to four biological characteristics of interest: timing, synchrony, rhythmicity and regularity of births. We then applied each metric retrieved on simulation data, considering normal and non-normal distributions of births, and varying distributions of births in time. We then evaluated the ability of each metric to capture the variation of the four phenology characteristics via a sensitivity analysis. Finally, we scored each metric according to eight criteria we considered important to describe phenology correctly. The high correlation we found among the many metrics we retrieved suggests that such diversity of metrics is unnecessary. We further show that the best metrics are not the most commonly used, and that simpler is often better. Circular statistics with the mean vector orientation and mean vector length seems, respectively, particularly suitable to describe the timing and synchrony of births in a wide range of phenology patterns. Tests designed to compare statistical distributions, like Mood and Kolmogorov-Smirnov tests, allow a first and easy quantification of rhythmicity and regularity of birth phenology respectively. By identifying the most relevant metrics our study should facilitate comparative studies of phenology of births or of any other life-history event. For instance, comparative studies of the phenology of mating or migration dates are particularly important in the context of climate change.

ecology↗

bioRxiv2020

Can citizen science analysis of camera trap data be used to study reproduction? Lessons from Snapshot Serengeti program

Ecologists increasingly rely on camera-trap data to estimate biological parameters such as population abundance. Because of the huge amount of data, the assistance of non-scientists is often sought after, but an assessment of the data quality is necessary. We tested whether volunteers data from one of the largest citizen science projects - Snapshot Serengeti - could be used to study breeding phenology. We tested whether the presence of juveniles (less than one or 12 months old) of species in the Serengeti: topi, kongoni, Grants gazelle, could be reliably detected by the "naive" volunteers vs. trained observers. We expected a positive correlation between the proportion of volunteers identifying juveniles and their effective presence within photographs, assessed by the trained observers. The agreement between the trained observers was good (Fleiss {kappa} > 0.61 for juveniles of less than one and 12 month(s) old), suggesting that morphological criteria can be used to determine age. The relationship between the proportion of volunteers detecting juveniles less than a month old and their actual presence plateaued at 0.45 for Grants gazelle, reached 0.70 for topi and 0.56 for kongoni. The same relationships were much stronger for juveniles younger than 12 months, reaching 1 for topi and kongoni. The absence of individuals < one month and the presence of juveniles < 12 months could be reliably assumed, respectively, when no volunteer and when all volunteers reported a presence of a young. In contrast, the presence of very young individuals and the absence of juveniles appeared more difficult to ascertain from volunteers classification, given how the classification task was presented to them. Volunteers classification allows a moderately accurate but quick sorting of photograph with/without juveniles. We discuss the limitations of using citizen science camera-traps data to study breeding phenology, and the options to improve the detection of juveniles.

ecology↗