Evolutionary landscape epidemiology

International audienceInfectious diseases have long been studied by scientists and public managers from clinical points of view, including causes of diseases, prevention of illness, or means to facilitate recovery. These last decades, our understanding of disease occurrence and dynamics has been improved by the recognition of an unequal distribution of parasites in space and of relationships between environmental factors and disease cases. This has led to the development of ‘landscape or spatial epidemiology’, a fi eld of research which can be traced back to John Snow’s work on cholera in London (Snow 1855 ) . By projecting case reports on maps, Snow highlighted that all cases were clustered around a street pump and deduced that cholera was a water-borne pathogen. Later, considering the infl uence of biotic and abiotic factors on ecological processes, Pavlovsky ( 1966 ) stated that landscape strongly affects the spatiotemporal distribution, abundance, and dispersal of hosts and parasites. Landscape epidemiology was then proposed as an integrative approach that aimed to understand the spatial spread of disease agents by analysing both spatial patterns and environmental risk factors (Pavlovsky 1966 ) . The ‘BAM diagram’ introduced for host–parasite interactions by Soberon and Peterson ( 2005 ) provides a useful conceptual framework for understanding the geography of diseases at the scale of the landscape (Fig. 13.1a ) . It predicts that the distribution of a species (i.e. a parasite) is the overlap between favourable biotic conditions (i.e. presence of competent hosts and other interacting parasites), abiotic factors (i.e. temperature) controlling the survival of free-living stages, and mobility capacities allowing the presence of the species in the appropriate areas. These last years, landscape epidemiology and the modelling of disease dynamics have largely benefi ted from two scientifi c breakthroughs. New computing technologies such as geographic information systems or remote sensing have allowed important advances in infectious disease epidemiology. They are described in Part V of this book. Besides this, the combination of metapopulation theory (proposed by Levins 1969 ) and epidemiological models has greatly improved our understanding of disease dynamics. In these models, a metapopulation of parasites is described as a set of populations distributed over distinct patches represented by either host individuals or host populations, and connected to varying degrees by dispersal (Hess 2002 ; Fig. 13.1b ) . This framework has provided fundamental predictions about the probability of disease diffusion or persistence in different situations (e.g. Ostfeld et al . 2005 f or a r eview)