Robustness response of the Zurich road network under different disruption processes

Road systems provide a set of essential services to modern societies, and as such, their functioning becomes crucial to sustaining urban ecosystems. Understanding infrastructure robustness is one with understanding how major disruptions affect the functionality of the overall infrastructure network. Preparing for and responding to unexpected disruptions call for a broad understanding of robustness and a variety of disruption scenarios. The purpose of this study is to analyze how a broad range of disruption magnitudes affects the robustness of the Zurich road network under three factors of influence: network growth in time, disruption strategy, and topological metrics. Our case study was the road system of Zurich in 1955 and 2012, modeled with roads as edges and junctions as nodes. We applied the percolation theory framework to analyze fragmentation processes on the network structure. We introduced two strategies to evaluate the fragmentation processes, the “targeted-randomized” and “attractor-guided” strategies, which use two betweenness centrality (BC) metrics as proxies of the value at risk for each road. The analysis yielded four major findings. First, the Zurich road network demonstrated a phase transition with a loss of approximately 90% of its functionality at a range of 0.1–0.4 of disrupted edges. Second, the distribution function of the robustness index was more robust in 2012 than in 1955. Third, the attractor-guided and targeted-randomized disruption strategies did not yield significant differences in how they affected the network robustness. Fourth, the two BC metrics also yielded a minor influence on the overall robustness. Future studies can extend the temporal analyses to a wider timeframe and combine commuter trips to assign a new value at risk of roads in fragmentation analyses. •The distribution of the robustness index indicated changes in the robustness of the road network structure with time.•Betweenness centrality metrics demonstrated a minor influence on the overall robustness.•The two disruption strategies, namely, the attractor-guided and targeted randomized, did not yield significant differences.