From Barriers to Connections Community Severance and Urban Connectivity in Dubai's Deira Districts: A Network Approach to Measuring, Mapping, and Reducing the Barrier Effect of Major Roads

Major roads connect a city at one scale and divide it at another. The barrier effect of transport infrastructure on local movement, known as community severance, has been recognized since Appleyard and Lintell's San Francisco streets study. This effect has since been linked to reduced walking, weaker social ties, and poorer access to services, yet it is rarely measured in cities where road construction is most rapid. This paper develops and demonstrates a network method for measuring, mapping, and reducing severance in Dubai's Deira districts, where 27 communities defined by Dubai Municipality are separated by expressways, arterials, and the Creek. Two instruments are introduced. The Barrier-to-Bridge Connectivity Framework (BBCF) sequences barrier inventory, permeability scoring, network modeling, and intervention prioritization in a loop that re-scores the network after each investment. The Community Severance and Permeability Index (CSPI) combines a community's permeability-weighted boundary, its impedance-weighted detour ratio, and its expressway boundary share into a single score. Communities are treated as nodes, and shared boundaries as edges of a graph whose weights depend on barrier class and crossing provision. Betweenness and closeness centrality, effective-distance ratios, and a Gini coefficient of closeness describe the structure. A Monte Carlo simulation of 4,000 short cross-boundary trips and a negative binomial model of daily crossings on 66 barrier segments quantify the behavioral consequences. Because crossing inventories and pedestrian counts were not available, permeability scores, crossing counts, and trips are illustrative, and the results are presented as a reproducible template. On this template, expressway boundaries reduce the share of trips walked to 21%, compared to 93% across local streets. The Al Twar, Al Khabaisi, and Al Nahda communities are the most severed. Crossing spacing and the presence of a formal crossing are the dominant predictors of crossings.