Journal of
Geography and Regional Planning

  • Abbreviation: J. Geogr. Reg. Plann.
  • Language: English
  • ISSN: 2070-1845
  • DOI: 10.5897/JGRP
  • Start Year: 2008
  • Published Articles: 391

Full Length Research Paper

Why populations are not planets_ gravity and the limits of disease modeling by analogy

Tom Koch
  • Tom Koch
  • Department of Geography (Medical) Vancouver, University of British Columbia, Canada.
  • Google Scholar
Ken Denike
  • Ken Denike
  • Department of Geography (Medical) Vancouver, University of British Columbia, Canada.
  • Google Scholar

  •  Received: 18 May 2021
  •  Accepted: 10 August 2021
  •  Published: 31 August 2021


Wellness depends on health and that, in turn, depends on the absence of disease. Analogous models based on physical laws have long been utilized by researchers to understand epidemic expansion in urban communities.  Perhaps the most significant of this class is the gravity model in which population size is equated with planetary mass and distance between cities to that separating planets. While the model assumes homogeneity among different bodies, cities or planets, in epidemiology the likelihood of disease spread may depend on other heterogeneous, non-constant factors. The study used a public dataset of H1N1 Influenza in 2009 as the focus. A natural log regression was applied in an attempt to sort the relative importance of gravity model variables as predictors of influenza occurrence and diffusion. It was found that while the model population size serves as a general predictor of disease expansion that distance failed as an indicator of disease dynamics. Furthermore, findings from the study show that disease progression was irregular and not, as one might expect from the gravity model, consistent in space or over time. The study concludes that the gravity model may serve only as a coarse predictor of disease expansion over time. By extension, this raises similar questions about other models in which homogeneity between populations or network of populations is assumed.


Key words: Gravity model, H1N1influenza, regression, spatial epidemiology.