نبذة مختصرة : Earth-bound transportation infrastructures have numerous impacts on ecosystems. Each of the causes of biodiversity loss identified by IPBES is represented by the road network: habitat destruction and degradation, invasive alien species, and the direct exploitation of animal and plant species. In particular, collisions between wildlife and vehicles are a source of mortality which leads to direct losses to wild populations. This mortality, which adds to other anthropogenic pressures, can jeopardize the persistence of many species. Therefore, it is crucial to be able to quantify the losses caused by collisions, and to know how to predict them in time and space. Collision monitoring projects based on citizen science are gaining traction in many countries: road users are asked to contribute their own reports of animal carcasses encountered on roads. However, citizen science data is more difficult to exploit than data generated through standardized scientific surveys, and is currently under-exploited in scientific publications. This thesis is structured around three main axes: a mechanistic model, which replaces the movement of animals and vehicles at the center of the wildlife-vehicle collisions system, allowing us to demonstrate that certain predictors, such as road traffic or the way animals move, have a real causal link with collision risks. Secondly, the project addresses the issue of roadkill data collection, particularly from the perspective of citizen science: roadkill persistence, detection rates and reporting rates by contributors to the projects. With novel methodologies, the persistence on roads of different species are studied, as well as detection and reporting rates in the context of citizen science. Finally, a data correction method is developed and applied to the Faune-Auvergne-Rhône-Alpes citizen science database, providing estimates of the actual number of collisions for several wildlife species. The mechanistic model proposed here, as well as the efforts put in place to rigorously evaluate citizen ...
No Comments.