.png){kind=icon}
.png){kind=icon}
Return
Operational Diagnosis of Urban Road Flood Resilience Using Vehicle Speed
.png){kind=icon}
OA
AI
J
X
R
J
F
A
H
DOI：10.1016/j.ijdrr.2026.106262.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Urban flooding caused by extreme rainfall poses a risk to infrastructure operation. Monitoring of infrastructure resilience is an important adjunct to the targeted, effective management of flood inundation events. Presently, resilience studies utilise socio-economic or infrastructure indicators that are slowly updated and unable to reflect changes in urban operational efficiency (UOE) at high spatiotemporal resolution. We propose a data-driven approach that measures the UOE of each road for each rainfall event through hourly vehicle speed changes. A two-dimensional resilience matrix is derived concerning UOE loss and recovery time. This is paired with machine learning to identify the main drivers behind operation change. We use the method to diagnose the resilience of 121 roads in central Shenzhen, China for 40 storm events during 2021-24. Our findings are as follows: (1) For an extreme 24-h event of 300 mm rainfall, 5%, 38%, 23%, and 34% of roads are classified as robust, high-, medium-, and low-resilience. Block-grid roads which offer abundant alternative routes are more resilient. (2) The order of resilience in recent years is 2023 > 2022 ≈ 2021 > 2024. (3) Heavy rainfall events that occur during peak traffic hours exhibit low resilience. (4) The factors of greatest influence on resilience are rainfall and road attributes. Higher values of the number of road junctions, main pipeline length and outlets per unit area are associated with improved resilience. Our framework facilitates better monitoring of road operation and provides a speed-based diagnosis perspective for resilience.
Keywords:
Urban flood
Road system resilience
Resilience matrix
Transportation big data
Interpretable machine learning
Journal
IF:
4.5
Papers:
6.0K
Citations:
2.1W
Organization
Cited Papers