Conference Agenda

The development of transnational biodiversity monitoring is recognized as a key working area by the EU’s 2030 Biodiversity Strategy, and camera trapping is often the tool of choice for terrestrial mammals. However, despite the widespread use of camera trapping across the continent, efforts to build a European network of standardized camera trapping are still scant. We will first review the multi-faceted complexities associated with building such a network, spanning from those that relate to financial and institutional arrangements to the methodological ones, which include monitoring aims, targets, metrics, data management and sharing routines. We will then introduce the objectives and contributions of the symposium and provide examples of current transnational efforts and collaborations at both global and European levels. Among the promising European projects is Snapshot Europe, the first standardized and coordinated initiative to monitor mammal communities at the European scale. Every year since 2021, volunteer researchers have collected camera-trap data on more than 35 species at 1000+ locations spread across 22+ countries. This talk will set the context towards identifying potential ways forward in terms of how existing and new collaborations can be leveraged to achieve harmonized monitoring across the continent.

Monitoring of lynx Lynx lynx populations in Scandinavia is based around unreplicated minimum counts of family groups, i.e. adult females with dependent kittens. The number of family groups is estimated every year based on confirmed observations of family groups. Traditionally, observations have been tracks from family groups in snow and dead kittens. After experiencing milder winters and decreasing snow cover, a large-scale network of camera traps has been used to increase number of observations (Scandcam). As well as contributing to lynx monitoring we are also exploring ways to utilize the bycatch data on other species to achieve a broader ecosystem monitoring. The last few years the project has also been a central part of monitoring wild boars in Norway. In addition, observations of other species are being used in research on other species.

We will present the design and use of the Scandcam network of camera traps and discuss challenges and lesson-learned from over 12 years work.

Camera trapping has unmatched capability to standardize mammalian monitoring across multiple areas. An important sampling option relates to the placement of camera traps on trails and forestry roads versus random. While the latter potentially provides for estimating density and studying activity patterns under minimum anthropogenic disturbance, the former is more suitable to monitor both human presence and a relatively larger pool of wildlife species, given the preference of many species to move along trails and roads. We deployed systematic sampling on trails and forestry roads to study the effect of humans’ outdoor recreation as a potential source of disturbance on wildlife. We targeted four protected areas in Italy and found that the mammalian meta-community consistently increased nocturnality in response to human passage, with effects mediated by species body mass. In one of the study areas, we monitored over 7 years and estimated occupancy trends both at community- and single-species level: mammals’ occurrence increased over the years in spite of increasing human frequentation, although species tended to temporally avoid humans. The protocol we adopted appears suitable to monitor wildlife populations and communities, and assess their vulnerability to anthropogenic threats, with promising results for broader replication at national and trans-boundary scales.

Combining camera-trap datasets holds promises for large-scale wildlife monitoring, but also comes with challenges. Here, I provide an overview of common challenges in programs that rely on combining data from multiple surveys, and how some of them can be addressed.

First, full standardization of sampling is rarely possible, causing variation across datasets in the spatial extent and resolution of data. Different camera models and setup strategies further affect the data collected. Many of these issues can be addressed by hierarchical models, which can account for the nested data structure, variation in sampling-related parameters, and differences in camera spacing.

Second, focusing on large-scale spatially representative sampling, these programs risk ignoring representative sampling at the local scale, where sampling may be biased towards more easily accessible areas. This may have implications for inference on wildlife communities and requires careful interpretation of results.

Finally, for data analysis large-scale programs frequently use an occupancy framework, as it allows accounting for imperfect species detection. Estimates of occupancy from point-based sampling in continuous habitat, however, are affected by population density and movement behavior and may not be readily comparable across surveys. This issue has received little attention from the camera-trap and statistical modeling communities.

Camera trapping has revolutionized wildlife ecology and conservation by automating data acquisition and generating massive amounts of camera trap data worldwide. However, the management and exchange of this data remain limited, hindering its full potential. To address this, a new data exchange format called Camera Trap Data Package (Camtrap DP) has been developed. Camtrap DP is based on a simple yet flexible data model, allowing users to easily exchange, harmonize, and archive camera trap data at various scales. It supports different camera deployment designs, classification techniques, and analytical use cases, ranging from compiling species occurrence data to distribution, occupancy, activity modeling, and density estimation. The format builds upon existing standards and is developed openly, collaboratively, and with version control from the start. Camtrap DP aims to enable large-scale data interoperability among camera trapping data producers and users, facilitating integration with other biodiversity data sources like GBIF. It also promotes the development of standardized data processing pipelines and the application of AI methods for automatic image recognition and data analysis. By harmonizing camera trap data from large-scale distributed networks, Camtrap DP harnesses the collective power of researchers and conservationists for more effective wildlife monitoring and conservation efforts.

Camera trapping biodiversity monitoring is advanced and broadly deployed worldwide, from a sheer diversity of entities, including research institutions, protected areas, wildlife offices, hunters, and citizens, and for a moltitude of reasons, from base research, to nature enjoyment. Often, camera trapping happens in the context of collective data collection or collaborative initiatives. This huge interest and wealth of data poses a lot of opportunities and some challenges. The recent advances in data standardisation protocols offer the technical possibility to archive camera trapping data in standard way and to communicate outputs between streams of data. Yet, this technical possibility is not paralleled by the emergence of a common 'space' where the different entities and projects are able to easily 'find' themselves, identifying their common purposes or specific objectives, and optimise both data collection and sharing of outputs. As a consequence, ability of camera trapping excercises to direct policy has been so far limited. I discuss these points and possible avenues ahead.