For several reasons butterflies are choice models for the study of phenomena occurring in the real world. As many insects, numerous butterfly species are facing with a strongly diminished viability due to human induced perturbations. Moreover, their short life cycle causes a rapid reaction to such perturbations, allowing a more immediate study, with minimal alteration due to the extinction debt -delay between the cause (perturbation) and its consequence (effect on the viability). Finally, despite this decline, data that can be gathered on these butterfly systems are numerous and of a high quality because the ecology and life history traits of butterflies are in general well known owing to the long naturalist and scientific tradition of lepidoptera studies, because even threatened species are often lcoally abundant enough to allow quantitative studies, and because many species are specialist restricted to well delimited biotopes, which makes their study easy. For all these reasons, butterflies are often considered as an umbrella group for biodiversity conservation.

Currently, our research on butterflies focuses on the study of two main determinants of viability: dispersal and habitat quality. 
- Dispersal, generic term for the movements of individuals between habitat patches at the local or regional scales, is currently recognized as one of the fundamental processes in the functioning of populations, a fortiori given the increasing habitat fragmentation following human activities. It is a key parameter in the design of ecological networks (e.g. Natura 2000), but poorly known in conservation biology. The bulk of previous studies of dispersal in fragmented landscapes focused on the patterns of dispersal, but the current need lies in the understanding of processes that create these patterns. The aim of our research is to use computer modelling to determine how movement and dispersal behaviours are translated in patterns at the scale of a fragmented landscape.
- Habitat quality has often been ignored or highly simplified in the modelling of dynamics and viability of natural (meta)populations, despite its functional importance. Due to the difficulties of its estimation, its integration is essentially binary: the biotope is either a suitable habitat or an inhospitable matrix. Nevertheless, its variation is in reality gradual. Currently, the vision of the habitat as being a given type of biotope or vegetation zone is gradually disappearing, being replaced by a resource-based definition: the habitat is the area containing all the resources necessary to the population; it may correspond to a single biotope or to the union of several biotopes, each containing some part of the resources. To determine habitat quality, it is necessary to measure reources and their influence on population demography through its impact on individual performance and fitness. The objective of our research is to establish habitat quality maps from the distribution of the resources necessary to the life cycle of the species and to include them into models of population viability.
We mainly work with wetland butterfly species:
     the Bog Fritillary (Proclossiana eunomia)
     the Violet Copper (Lycaena helle)
     the Marsh Fritillary (Euphydryas aurinia)
     the Cranberry Fritillary (Boloria aquilonaris)