Improved central place foraging modeling: importance of landscape context and modeling strategies in the case of a French sweat bee

Insect pollinators, including wild bees, contribute to the pollination of >90 % of wild flowering plants. A realistic simulation of their movements during foraging behaviour is essential to predict their pollination capability. The aim of this study was to develop a new individual based model called SimOïko_CPF to tackle some of the limitations of current models of wild bees’ foraging movement. Its performance was compared with a distance decay model: InVEST crop pollination (InVEST_Poll). For central-based foraging species like wild bees, the SimOïko_CPF model makes it possible to use the SMS algorithm as a support for the movements of individuals during foraging. The SimOïko_CPF model increases the capability to model the interactions between individuals and their landscape components such as physical friction, the attractiveness of floral resources, and collection probability. Sensitivity analysis of the parameterization of the species’ life-history traits performed in this study highlights the importance of two parameters in the landscape visitation: the spatial short-term memory size, and the resource collection probability. The two models were used to simulates the foraging pattern of a small species of wild bee (Lasioglossum marginatum). The outputs of the two models showed similar correlations with the capture data. The comparison between SimOïko_CPF and InVEST_Poll revealed discrepancies in the simulated foraging patterns on virtual landscapes varying in patch size, aggregation and composition. It’s showing that SimOïko_CPF can be use in various heterogenous landscapes and allows the investigation of a large set of parameters including the effect of physical friction and resource attractiveness.