Biotic interactions and climate in species distribution modelling

Species have preferred environmental niches 1 and their geographical distributions respond to global climate change 2. Predicting range shifts under climate change has profound implications for conservation of biodiversity 3, provision of ecosystem services, and in the management of invasive species 4. Species distribution modelling (SDM) has largely focussed on climate variations, but biotic interactions, such as predation and competition, can alter potential distributions 5,6 and affect migration rates 7. However, a lack of data on biotic interactions has restricted consideration of these factors for many species 1. Here, we compare the power of biotic and climatic factors as predictors of global distributions of hundreds of crop pests and pathogens (CPPs), for which host preferences are known. We show that host availability is a more important predictor of endobiotic pathogen distributions (fungi, oomycetes, bacteria, viruses and nematodes) than of epibiotic pest distributions (insect herbivores). Conversely, climatic variables are better predictors of epibiotic pest distributions. These results are robust to statistical controls for varying observational capacity among countries. Our findings demonstrate that life history affects global scale species distributions and that SDM should incorporate biotic interactions as well as climate.