Abstract
Accessibility, the number of time-decayed jobs available to each zone within a region, can help prioritize candidate transportation investments. This paper demonstrates how to compute auto accessibility using commonly available resources and identifies strategies needed to render calculations feasible and transparent. (The scope excludes transit and pedestrian impacts.)
For the first objective, computational solutions included developing a semi-automated method to import legacy transportation networks, automating turn prohibitions, and using an algorithm to check for inconsistently formed service areas that sometimes occur in a random fashion with geographic information system software. Failure to exercise quality control using these approaches gives erroneous results: not solving the problem of inconsistently formed service areas led to a region within 50 mi of a 1-mi corridor (where improvements are proposed) having an accessibility almost 40 times higher than the correct value. For the second objective, the influence area (i.e., catchment radius) mattered most: for one project, the forecast accessibility improvement dropped by 80% when an area within 45 mi of the project, rather than an area within 15 mi, was the basis of the analysis. Other decisions affected forecast accessibility improvement less: the choice of the number of centroid connectors affected forecasts by an average of 23% (with a 10-mi influence area). Choosing to eliminate negative net accessibility contributions, attributed to geometric approximations in the software, affected forecasts by less than 21% (35-mi influence area or smaller). Ranking five proposed investments in relation to their forecast accessibility benefit demonstrated the importance of documenting users’ computational choices.
Get full access to this article
View all access options for this article.