Reducing the uncertainty induced by spatial aggregation in accessibility and spatial interaction applications

- We focus on methods to reduce scale dependencies in estimation of travel time.
- Harmonic-mean methods are used to obtain benchmark travel times and distances.
- Methods are introduced to substantially improve intra-zonal travel time accuracy.
- Mass-weighted centroids should be used to compute extra-zonal travel time.
- Errors in extra-zonal travel times are largely zone-specific, but no useful penalty variable is found.

Analyses of spatial interaction are to some degree plagued by uncertainty regarding the impact of spatially dispersed interaction masses within zones on travel times. In this paper, interaction-weighted travel times are computed from a matrix between regularly distributed points at fine resolution, and used together with secondary data to improve estimates of interaction weighted travel time based on commonly applied methods. The paper proposes a method for computing intra-zonal, interaction weighted travel times that is considerably less sensitive to spatial aggregation than existing approaches, and demonstrates that population-weighted centroids are to be preferred over geographically-weighted centroids.