Resolving the accessibility dilemma: Comparing cumulative and gravity-based measures of accessibility in eight Canadian cities

There is a lack of agreement regarding the theoretical framework that practitioners should use for accessibility assessment - a measure of transport and land-use systems performance. Cumulative measures are simple and easy to interpret, while gravity-based measures are more sophisticated, resourceful, and less intuitive approaches. As such, this study aims to investigate whether the estimates of a simple cumulative opportunity measure are significantly different from those made using advanced gravity-based measures to understand if the former can be a substitute for the latter in practice and if a certain threshold of travel time can be recommended for different regions. We estimated cumulative and gravity-based accessibility using decay-probability density functions, decay-cumulative density functions, Gaussian, and a Log-Logistic decay-cumulative density functions using census commuting flows, car congested travel time and public transit schedules from eight metropolitan regions across Canada - Toronto, Montreal, Vancouver, Edmonton, Quebec City, Winnipeg, London, and Halifax. These measures of accessibility were tested for correlation, and we found that a coefficient of approximately 0.90 is reached when the threshold to calculate cumulative opportunities accessibility is set to the average commute time for both low- and non-low-wage jobs accessibility analyses by transit and motor vehicles. The paper provides evidence to support the reliability of cumulative accessibility, facilitates its broader adoption for evaluation of transport and land use interactions in North American cities, as well as opens opportunities to advance the equitable distribution of transport system benefits.

Automated summary

There is a lack of consensus on the theoretical framework for assessing accessibility in transport and land-use systems. This study compares a simple cumulative measure with advanced gravity-based measures to determine if the former can be a substitute for the latter. The results indicate a high correlation between the two measures, suggesting that cumulative accessibility can be reliable and widely adopted for evaluating transport and land use interactions in North American cities.