Cette thèse a pour objet la modélisation de systèmes multi-proies–prédateurs. Elle s’intéresse particulièrement à l’influence du comportement d’un prédateur sur les interactions indirectes entre ses proies, i.e. l’effet de l’ajout d’une proie sur la densité des autres. La théorie classique prédit l’occurrence d’effets indirects négatifs entre les proies, ou compétition apparente, résultant de l’interaction avec un prédateur commun ; des résultats plus récents identifient certains mécanismes à même d’atténuer ces effets négatifs. Nos travaux revisitent les hypothèses autour de ces mécanismes dans des systèmes composés de deux proies et de leur prédateur commun. Après avoir fixé le cadre écologique en rappelant les principaux types d’interactions directes et indirectes, nous introduisons les modèles proies-prédateurs classiques. Les travaux se concentrent ensuite sur une famille de modèles présentant de la densité-dépendance négative chez les prédateurs couplés à différentes modélisations du comportement des prédateurs lorsqu’ils sont confrontés à plusieurs types de proies. Nous montrons notamment que les interactions entre ces mécanismes peuvent inverser la compétition apparente et, contre intuitivement, accroître la densité des proies par l’intermédiaire d’un prédateur commun. Nos résultats trouvent pour partie application en lutte biologique, où il est courant de chercher à favoriser les auxiliaires en aménageant leur environnement (apport de nourriture alternative, refuge, ...). Ils suggèrent que de telles pratiques peuvent se révéler contre-productives, le contrôle des ravageurs pouvant être affaibli du fait d’une distraction de leurs prédateurs. / This thesis deals with multi-prey–predators modeling. It is particularly focused on the influence of the behavior of a predator on indirect interactions between its prey, i.e, the effect of the addition of a prey on the other prey. Classical theory predicts the occurrence of negative indirect effects between prey, or apparent competition, resulting from the interaction with a common predator. More recent results identify mechanisms that mitigate these negative effects. Our work revisits the assumptions about these mechanisms in systems composed of two prey and their common predator. After setting the ecological framework by recalling the main types of direct and indirect interactions, we introduce classical predator-prey models. The work then focuses on a family of models with predator negative density-dependence coupled with different models of predator behavior when faced with different types of prey. We show that the interactions between these mechanisms can reverse apparent competition and counter-intuitively, increase prey density through their common predator. Our results are relevant to biological control programs, where a common practice aims at fostering biological control agents by providing them with alternative food or shelters. Our theory suggests that such practices may be counter-productive, pest control being disrupted by a predator distraction effect.

We are currently faced with the global challenge of conserving biological diversity while also increasing food production to meet the demands of a growing human population. Land-use change, primarily resulting from conversion to production land, is currently the leading cause of biodiversity loss. This occurs through habitat loss, fragmentation of remaining natural habitats, and resulting edge effects. Land-sparing and land-sharing approaches have been discussed as alternative ways to engineer landscapes to mitigate biodiversity loss while meeting production objectives. However, these represent extremes on a continuum of real-world landscapes, and it will be important to understand the mechanisms by which adjacent land use affects natural remnant ecosystems in order to make local land-management decisions that achieve conservation, as well as production, objectives. This thesis investigates the impact of juxtaposing production and natural forest on the community-wide interactions between lepidopteran herbivores and their parasitoids, as mediated by parasitoid spillover between habitats. The first and overarching objective was to determine whether herbivore productivity drives asymmetrical spillover of predators and parasitoids, primarily from managed to natural habitats, and whether this spillover alters trophic interactions in the recipient habitat. The study of trophic interactions at a community level requires understanding of both direct and indirect interactions. However, community-level indirect interactions are generally difficult to predict and measure, and these have therefore remained understudied. Apparent competition is an indirect interaction mechanism thought to be very important in structuring host-parasitoid assemblages. However, this is known primarily from studies of single species pairs, and its community-wide impacts are less clear. Therefore, my second objective was to determine whether apparent competition could be predicted for all species pairs within an herbivore assemblage, based on a measure of parasitoid overlap. My third objective was to determine whether certain host or parasitoid species traits can predict the involvement of those species in apparent competition. My key findings were that there is a net spillover of generalist predators and parasitoids from plantation to native forest, and that for generalists, this depends on herbivore abundance in the plantation forest. Herbivore populations across the edge were linked by shared parasitoids in apparent competition. Consequently, an experimental reduction of herbivore density in the plantation forest changed parasitism rates in the natural forest, as predicted based on parasitoid overlap. Finally, several host and parasitoid traits were identified that can predict the degree to which host or parasitoid species will be involved in apparent competition, a finding which may have extensive application in biological control, as well as in predicting spillover edge effects. Overall, this work suggests that asymmetrical spillover between production and natural habitats occurs in relation to productivity differences, with greater movement of predators and parasitoids in the managed-to-natural forest direction. The degree to which this affected species interactions has implications for landscape design to achieve conservation objectives in production landscapes.