Investigating the Impacts of Urban Road Network Topology on Street Flooding in New York City

The frequency and magnitude of street flooding are increasing in many urban communities worldwide due to, rapid urbanization, unplanned urban development, and climate change. Street flooding potentially causes serious safety hazards, decreases public transportation accessibility, spreads pollutants, and undermines the surrounding property values. Therefore, it is urgent that urban planners, city managers and stakeholders could integrate street flooding preparation, mitigation, and adaptation in the planning stage to reduce its damage. Particularly, the characteristics of the urban road network, which represent impervious area connectivity have a strong association with street flooding and therefore have become an important topic on street flooding issues in the city.In this study, we combined novel urban road network spatial metrics with street flooding social media big data to investigate their associations with data-driven methods for New York City. Street flooding data was extracted and selected from the New York City 311 complaints platform together with the Rainfall database. A series of regression analyses were then performed to investigate the relationship between road topology factors and street flooding risks. Finally, the hydrological implications of the road topology metrics and urban planning suggestions were drawn from the regression results. Specifically, we implemented the following analytical steps: (1) We quantified urban road network topology based on the traffic system study method, which includes network density, network area density, intersection density, Beta, Alpha, Gama, and so on in 551 sewer sheds in New York City. (2) We screened 4543 street flooding reports, which are related with rainfall, from street flooding complaints during 2011 to 2022 by using precipitation data and KNN cluster analysis. (3) We found the relationship between each road network topology metric and street flooding risks with Poison regression models. (4) We applied the regression results to suggest some best practices and suggested planning scenarios for avoiding street flooding.The preliminary study conclusions include: (1) Street flooding risk has periodic changes characteristics with explicit spatial-temporal heterogeneity. (2) The number of street flooding complaints has a significant negative correlation with rainfall intensity except for the extremely heavy rainfall events. (3) The higher road network density areas are more likely to generate street flooding than lower road network density areas.Reference:Kelleher, C., & McPhillips, L. (2020). Exploring the application of topographic indices in urban areas as indicators of pluvial flooding locations. Hydrological Processes, 34(3), 780-794.Singh, P., Sinha, V. S. P., Vijhani, A., & Pahuja, N. (2018). Vulnerability assessment of urban road network from urban flood. International Journal of Disaster Risk Reduction, 28, 237-250.Wang, R.-Q., Mao, H., Wang, Y., Rae, C., & Shaw, W. (2018). Hyper-resolution monitoring of urban flooding with social media and crowdsourcing data. Computers & Geosciences, 111, 139-147.Agonafir, C., Pabon, A. R., Lakhankar, T., Khanbilvardi, R., & Devineni, N. (2022). Understanding New York City street flooding through 311 complaints. Journal of Hydrology, 605, 127300.