Absolute Space or Relational Space, Which Governs Spatiotemporally Extended Effects in Disease Dispersion?

Prevailing disease models typically focus on in-situ effects, that is, the health risks of a location are affected by the transmission-driving factors of the same location. The ex-situ effects, in contrast, extend health risks from neighboring locations and earlier days to focal locations and current dates. These effects could be critical but have not received much attention. This study investigates the extended effects in absolute space and relational space. We examine whether the effects exist, whether they differ between the two spaces, and whether they vary with the order of neighbors and the number of prior dates in both spaces. Results show that extended effects are generally present. Mild effects are identified in absolute space, while greater effects are observed in relational space. The effects vary slightly with the neighbor order in absolute space, but considerably in relational space where the second-order neighbors exert the most prominent effects. In both spaces, the effects diminish at the third-order and last for up to three days. These findings advocate multiple spatializations that offer an in-depth understanding of disease dispersion in specific and dynamic geographic phenomena at large.