Research interests

Level- and scale-dependent habitat selection

A primary objective of community ecology is to understand the conditions that allow species to coexist by identifying how co-occurring species use and share space and resources. The European pine marten Martes martes (PM) and the stone marten Martes foina (SM) are syntopic mustelids with similar morphology and ecology. In a previous telemetry based study, we documented segregation of space use in Bresse, a fragmented rural area in France. The differential use of resting habitat, a critical resource for mustelids, could theoretically explain syntopy between these species. However, habitat selection differs from habitat use or association, it implies choice, and is commonly measured as ‘use relative to availability’ or as ‘use versus non-use’. We thus undertake a fine-scale study of habitat selection for resting sites by these species to complement previous results.

As organisms commonly respond to their environment across a range of scales, we performed an optimized multiple-level and multiple-scale resting-site habitat selection study using multivariate analysis (Outlying Mean Index and K-select). Given our previous knowledge of habitat use of both species, we predicted that differences in habitat selection and the spatial scale at which it occurs could explain the coexistence of these species in the Bresse area.

Meso-carnivores landscape genetics

Connectivity has been defined as “the degree to which the landscape facilitates or impedes movement among resource patches”. A variety of techniques have been used to measure directly the connectivity between populations such as tracking organisms through field observations. Mark recapture techniques have been extensively used in the past and continue to be the primary way to assess the connectivity. However, these methods can be challenging when monitoring large populations or considering broad spatial extents, particularly when studied organisms are cryptic and do not provide information over evolutionary time scales.

Landscape genetics provides a tool to overcome these limitations and assess connectivity over many generations. While the traditional spatial approach in population genetics consists in testing the influence of the Euclidean geographical distance on genetic variations (sensu isolation by distance), landscape genetics takes into account the quality and the permeability of the landscape matrix by using a more functional geographical distance that takes into account the cost of movement for the animal across the landscape, the so-called least-cost distance.

Using a landscape genetics framework, we investigated the consequences of habitat fragmentation and harvesting on the connectivity of several highly secretive meso-carnivore species in human dominated landscapes. The results may help conservation biologists unravel contemporary and historical ecological and anthropogenic factors that influence rare and threatened species with similar ecology.