De l'échelle locale à l'échelle globale : l'importance d'une approche intégrée pour la conservation d'une espèce de reptile en déclin, la vipère péliade (Vipera berus, L. 1758) | From the local to the global scale: the need for an integrative approach for the conservation of a declining reptile species, the common adder (Vipera berus, L. 1758)

Today, evidence points to a global decline in reptiles. However, while the main causes of this decline are now well known, the mechanisms by which ecosystem disturbances affect reptiles remain poorly investigated. Indeed, ecosystem disturbances can act on multiple spatial scales and have cascading consequences on numerous parameters affecting reptiles. In such a context, a comprehensive analysis of the complex mechanisms affecting reptile populations cannot be confined to a single spatial scale. The aim of this thesis is therefore to highlight the multi-scale aspect of the mechanisms (micro-meso-macro influence spatial scale) governing the dynamics of reptile populations in Western Europe. To meet this objective, we decided to build our research strategy around a sentinel species, the common adder (Vipera berus). First, our research strategy focused on the role played by the local context of vegetation structure (micro-influence scale) in the modulation of the predation pressure exerted on the adder. Using 2400 artificial adders, we were able to demonstrate that the local context and especially the increase in the local vegetation complexity had a positive impact by creating microhabitats in which thermoregulatory behaviour presents a lower risk of predation. In addition, our results highlighted the increased risk of mammalian predation in the vicinity of linear elements structuring habitats (hedgerows and forest edges). Our results showed the modulating role played by the local vegetation structure and microhabitats availability on predation risks. This study has significant management implications for snake conservation leading to the formulation of practical recommendations. We then focused our research on the influence of the local vegetation context (micro-influence scale) on the abiotic conditions available within microhabitats (microclimates) and on the abundance of reptiles found within forest edges. Our results showed that high local canopy cover induces a general cooling of microhabitats, resulting in a decrease of microhabitat quality within interior forest edges. Our analyses based on repeated count data of common lizards and adders present in our network of 55 edges also highlighted the importance of the local structure of the vegetation within forest edges. Indeed, reptile abundance was positively related to the amount of light reaching microhabitats at ground level. Consequently, internal forest edges with an intermediate vegetation stratum along the exploitation driveway harbour higher estimated abundances than narrow or steep forest edges. Taken together, these results demonstrate that abiotic conditions defined by local vegetation structure are of paramount importance for reptiles within habitats with a high degree of canopy closure such as forests. Based on these results, several practical recommendations have been formulated to combine forestry activities and reptile conservation. Finally, we have left the micro-scale to focus on the influence of the global context (meso- and macro-influence scale) in which populations evolve, and the impact of global context on the local dynamics of the populations of common adder found in Western Europe. To achieve this goal, we have compiled the largest database of standardised monitoring data on adder population in western Europe. This large database, comprising 251 populations monitored for at least 7 years, was then used to propose a multi-site analysis in which we examined, among others, the role played by habitat parameters (mesp-influence scale) and macroclimatic parameters (macro-influence scale). Our analyses have confirmed the significant decline of the adder on a sub-continental scale, with an estimated annual loss in population size of 2.05%. Our analyses also highlighted the positive influence of a high availability of suitable habitats and of a high diversity of habitats encountered on meso scale (3 km radius around populations). Finally, our results demonstrated the important role played by the anthropic management of semi-natural habitats to maintain open environments suitable for the species. However, this positive effect of anthropic pressure on natural areas seems to reach a breaking point when anthropic pressure becomes too high and no longer allows the development of microhabitats favourable to the species. However, our data did not allow us to demonstrate a significant effect of measured climatic variables with a macro-scale influence on population dynamics. Taken as a whole, the results of this thesis underline the importance of adopting a more holistic and integrative multi-scale approach to better understand the mechanisms impacting reptile populations. Our results clearly indicate that habitat characteristics, both on a micro scale (microhabitats-microclimates) and on a meso scale (landscapes), play a crucial role in explaining the dynamics observed. However, our analyses were unable to significantly determine the role of climatic variables (macro-influence scale) on population dynamics. However, the local context in which individuals evolve is proving to be of considerable importance in modulating the exposure of reptiles to constraints on a larger scale. Considering the significant decline in reptile populations, it is essential to develop a conservation strategy adapted to the 21st century modern landscapes. To this end, we suggest a reassessment of the traditional dichotomous choice between the concepts of land-sharing and land-sparing. Our proposal is to develop conservation strategies that combine the concepts of land-sharing and land-sparing within unique agro-ecological landscapes. Consequently, within agricultural landscapes, maintaining and protecting lenses of favourable habitats is essential to ensure the role of core areas for reptiles. In parallel, a major effort needs to be made to improve the landscape matrix permeability to make it more reptile-friendly by extensification of farming and forestry practices. In these landscapes combining elements of land-sharing and land-sparing, movement between core areas (including migration) is possible, enabling the establishment of metapopulations. These metapopulations of reptiles are then more resistant to disturbances acting at various spatial scales, as well as to changes in the local and global context.