Eur. Phys. J. B 84, 579-587 (2011)
https://doi.org/10.1140/epjb/e2011-20485-2

Regular Article

Recurrent host mobility in spatial epidemics: beyond reaction-diffusion

¹ Department of Civil and Environmental Engineering, Massachusetts Institute of Technology, 02139 Cambridge, MA, USA
² Max Planck Institute for Dynamics and Self-Organization, Göttingen, Germany
³ Department of Physics, University of Göttingen, Göttingen, Germany
⁴ Northwestern Institute on Complex Systems and Department of Engineering Sciences and Applied Mathematics, Northwestern University, Evanston, IL, USA

^a e-mail: belik@mit.edu

Received: 17 June 2011
Received in final form: 23 October 2011
Published online: 5 December 2011

Abstract

Human mobility is a key factor in spatial disease dynamics and related phenomena. In computational models host mobility is typically modeled by diffusion in space or on metapolulation networks. Alternatively, an effective force of infection across distance has been introduced to capture spatial dispersal implicitly. Both approaches do not account for important aspects of natural human mobility, diffusion does not capture the high degree of predictability in natural human mobility patters, e.g. the high percentage of return movements to individuals’ base location, the effective force of infection approach assumes immediate equilibrium with respect to dispersal. These conditions are typically not met in natural scenarios. We investigate an epidemiological model that explicitly captures natural individual mobility patterns. We systematically investigate generic dynamical features of the model on regular lattices as well as metapopulation networks and show that generally the model exhibits significant dynamical differences in comparison to ordinary diffusion and effective force of infection models. For instance, the natural human mobility model exhibits a saturation of wave front speeds and a novel type of invasion threshold that is a function of the return rate in mobility patterns. In the light of these new findings and with the availability of precise and pervasive data on human mobility our approach provides a framework for a more sophisticated modeling of spatial disease dynamics.