Modest flooding can trigger catastrophic road network collapse due to compound failure

Urban infrastructure networks are vital for community access and economic activity, yet they face increasing pressure from population growth and flooding. Population increases lead to congestion, while climate change exacerbates flood risks. Transportation networks are particularly vulnerable, as floods disrupt infrastructure and cause cascading failures. Existing research often focuses on single types of failures (e.g., topological or functional), neglecting the complex interactions that occur during compound failures, where flooding coincides with traffic congestion. This study addresses this gap by developing a network-theory-based framework to analyze compound structural, functional, and topological failures in transportation networks during urban flooding. The study uses Hurricane Harvey in Harris County, Texas, as a case study to illustrate the catastrophic consequences of these interconnected failures. Understanding these complex interactions is crucial for designing robust and resilient transportation systems capable of withstanding future challenges.

Previous research on transportation network robustness has primarily focused on either single failure types or solely on either network topology or traffic flow. Studies have modeled disruptions by altering road topology or functionality. However, the consequences of simultaneous disruptions, such as flooding and congestion, have received less attention. Existing studies often utilize topological perturbation analysis, but lack a comprehensive framework to capture the compound effects of structural, functional, and topological failures occurring simultaneously during a flood event. This research bridges this gap by developing a novel approach to analyzing these compound effects.

This study employs a network-theory-based framework that integrates structural, functional, and topological failure modes. The researchers used high-resolution traffic data (15-minute interval road speeds) for Harris County, Texas, during and after Hurricane Harvey (August-October 2017). The Harris County transportation network was represented as a graph, with links representing roads and nodes representing intersections. Three types of failures were defined: 1. **Structural Failure (SF):** Road closures due to flood inundation. 2. **Functional Failure (FF):** Reduced travel speed due to congestion. 3. **Topological Failure (TF):** Isolation of road segments from the main network due to SF and FF. Link quality was derived from the ratio of observed travel speed to a reference speed, with links below a certain threshold considered to have low quality (congested). Network percolation theory was applied to analyze how the size of the largest connected component (giant component, GC) changed with varying levels of SF, FF, and TF. The researchers performed percolation modeling by iteratively removing links based on their quality, observing the resulting GC size, and identifying critical percolation thresholds. This allowed for the quantification of the impact of different failure types and their combinations on network connectivity. The study also compared empirical findings against synthetic scenarios (no flooding, random flooding) to isolate the specific impact of flooding on network robustness. Regression analysis was used to explore the relationships between SF, FF, TF, and GC size.

The key findings of this research are: 1. **Compound Failure Amplification:** Even a small percentage (2.2%) of flood-induced road closures (SF) significantly amplified the effects of functional failures (FF) and topological failures (TF), resulting in a substantial reduction (up to 17.7%) in the size of the largest connected component (GC) of the Harris County transportation network during Hurricane Harvey. This highlights the catastrophic effect of the combined failures. 2. **Temporal Dynamics:** The critical percolation thresholds (qc), indicating network vulnerability, exhibited temporal variation during Hurricane Harvey, showing a significant decrease during the height of the flood and a gradual recovery afterward, but not to pre-flood levels. This underlines the long-term impact of floods on network resilience. 3. **Non-linear Relationships:** Regression analysis revealed non-linear relationships between different failure types and network connectivity. A small increase in SF disproportionately increased FF and TF, leading to a significant decline in GC size. This implies that network vulnerability sharply increases even with modest flooding. 4. **Flood Impact Comparison:** Comparing the empirical data with synthetic scenarios (no flooding and random flooding) demonstrated that flooding significantly reduced network robustness compared to situations with congestion alone. Random flooding resulted in even more severe network disruptions than observed during the actual event, suggesting that the spatial distribution of flooding plays a crucial role. 5. **Spatial Distribution of Failures:** Visualization of the spatial distribution of different failure types during Hurricane Harvey showed how flooding concentrated failures, leading to widespread network fragmentation and severely reducing connectivity.

The findings of this study underscore the importance of moving beyond assessments of single failure modes when evaluating the resilience of urban transportation networks to flooding. The significant reduction in network connectivity observed even with modest levels of flooding emphasizes the amplifying effects of compound failures. The non-linear relationship between different failure types highlights the need for considering indirect consequences, such as changes in traffic patterns, when estimating flood impacts. The temporal analysis of network vulnerability demonstrates the long-term consequences of flooding events, extending beyond the immediate period of inundation. These findings are relevant for urban planning and hazard mitigation, advocating for interventions aimed at enhancing network robustness against compound failures. For example, reducing congestion through measures like improving public transit could mitigate the amplification effects of flooding.

This paper demonstrates the devastating impact of compound failures on urban transportation network resilience during flooding. The study highlights the critical need for considering the combined effects of structural, functional, and topological failures in flood risk assessments and planning. Future research should expand this analysis to other cities and disaster scenarios, investigate the vulnerability of different populations, and explore strategies for enhancing network resilience against compound failures. The framework presented here provides a valuable tool for informing the design of more robust and resilient urban infrastructure.

The study's focus on Harris County, Texas, during Hurricane Harvey limits the generalizability of the findings to other geographic locations and flood characteristics. The reliance on traffic speed data for assessing functional failures might not fully capture all aspects of network performance. Additionally, the study did not explicitly model human behavior changes in response to flooding, which could affect traffic patterns and network resilience.