Characterization of Vulnerability of Road Networks against Fluvial Flooding using Their Spatial Co-location Interdependency with Flood Control Infrastructure

Bahrulla Abdulla, Texas A & M University
Bjorn Birgisson, Texas A&M University
Topic Area
Sustainability and Resilience in Transportation Systems

The objective of this paper is to characterize the vulnerability of road network to fluvial floods using the co-location interdependency indices with flood control infrastructure. Despite the fact that urban critical infrastructure systems share the constrained urban space and deliver their vital functionality in close spatial proximity of each other, until recently, most of the critical infrastructure sectors have been studied in varying degrees of isolation. Achieving sustainability, which requires aiming for a balance among social, economic and environmental domains of development, requires an improved understanding of the interdependencies among different sectors. One such important interdependency (and the subject of this study) is the geospatial colocation between flood control infrastructure (bayous, channels and creeks) and road networks. This study departs from the existing literature by examining this dependency in a more granular fashion and considering the heterogeneity among different neighborhoods. This study first modeled the road network and flood control infrastructure networks in each super neighborhood as primal graphs. Then, colocation interdependencies between flood control and road networks were measured by three indices: (1) the degree of similarity in the orientation of the two networks, measured by discrepancies in overall bearings of the edges in the two networks; (2) density of the flood control infrastructure within road network, measured by the ratio of the total length of edges in flood control and road networks in each super neighborhood; (3) total length of edges in the road network which are located in a short distance (i.e.500m) away from the flood control network. In the end, we examined the relationship between above-mentioned three colocation measures and the observed loss of connectivity in the road network in each neighborhood during a fluvial flooding event. The proposed method was applied to road networks in 92 super neighborhoods in Houston and fluvial flooding event caused by hurricane Harvey was considered. Results indicated that : (1) the more aligned the orientation of flood control and road networks are, the greater the loss in connectivity in road networks during the fluvial flooding; (2) both a higher density of flood control network and their greater proximity to the road network mean heavier flooding to the road network. However, the location of the flooding on the road network matters and severity of flooding doesn't necessarily translate directly into the severity of connectivity loss in road networks. Nodes in the road network with high betweenness centrality have the greatest impact on the loss of connectivity, therefore, maintaining their functionality during the flooding is crucial. Results of this study have important planning and operation implications for both flood control infrastructure and road networks in an urban setting.

Abstract Keywords
road network connectivity
flood control infrastructure
fluvial flood
co-location interdependency