As the risk of flooding continues to wreak havoc on transportation infrastructure systems, enhancing the ability to effectively monitor flood inundations in road networks becomes increasingly critical. We learned this the hard way… Traditional approaches to flood monitoring have relied primarily on physical sensor networks, but these systems often suffer from limited coverage and blind spots. Fortunately, a wealth of innovative technologies and data sources are emerging that can supplement physical sensors and revolutionize how we monitor and manage flood events.
Now, this might seem counterintuitive…
Limitations of Physical Flood Sensor Networks
Physical flood sensors, such as stream gauges and water level monitors, provide invaluable data on real-time water levels and flow rates. However, these sensors are not always installed systematically or extensively, leaving many areas uncovered and unobserved. Additionally, physical sensors can only provide hydrological information for the immediate vicinity, failing to capture the broader spatial propagation of floodwaters across the transportation network.
“Physical sensors are not installed extensively and systematically, and therefore, the network of sensors may not be able to provide sufficient information needed to monitor the flooding status of entire region or infrastructure networks,” explains Dr. Hamed Farahmand, an expert in flood risk management and resilient infrastructure systems.
Leveraging Crowdsourced Data for Flood Monitoring
To overcome the limitations of physical sensor networks, researchers have begun exploring the potential of crowdsourced data to enhance flood monitoring capabilities. Crowdsourced data, such as citizen-reported 311 calls or social media posts, can provide real-time, location-specific insights on flood impacts that physical sensors may miss.
“Clusters of crowdsourced data can be referred to as ‘social sensors’ for collecting localized information to improve flood monitoring,” says Dr. Miguel Esparza, a specialist in disaster informatics and urban resilience. “While crowdsourced data are generated by people with personal experience from a flood, the data needs proper validation to be used for improving flood monitoring.”
By integrating crowdsourced flood reports with physical sensor data, cities can gain a more comprehensive understanding of flood propagation and impacts across the transportation network. For example, a study in Harris County, Texas found that leveraging 311 call data could reduce the need for physical flood sensors by up to 32% in areas lacking existing sensor coverage.
“The framework includes in four steps: 1) determining the cases in which clusters of 3-1-1 reports can signal flood inundation and be used as a social sensor; 2) employing a graph-theoretic approach to identify the minimum number and locations of additional sensors in a road network that can double-check that monitoring the flood inundation in the entire network; 3) identifying the road intersections from step 2 that can be supplemented by 3-1-1 reports based on the number of reports that are within a buffer of these nodes; and 4) developing a multi-criteria criticality-based approach to identify the set of sensors that take advantages of social sensors while double-check that that the additional sensors are located on the critical regions of the road network.”
Integrating Remote Sensing and Predictive Analytics
While crowdsourced data can enhance flood monitoring, remote sensing technologies offer another powerful tool for understanding the spatial and temporal dynamics of flooding events. Satellite imagery, aerial photography, and unmanned aerial vehicles (UAVs) can provide high-resolution data on flood extents, water depths, and infrastructure impacts.
“Remote sensing data can help fill gaps in spatial and temporal coverages, which prioritized vulnerable roads for improving response initiatives,” notes Dr. Xueming Liu, an expert in transportation network modeling and observability analysis.
By combining remote sensing data with predictive analytics, cities can take flood monitoring to the next level. Advanced hydrologic and hydraulic models, fed by real-time sensor data and weather forecasts, can forecast flood inundation paths and warn residents and emergency responders of impending dangers.
“Proper flood monitoring provides information for detecting areas with a higher risk of inundation in near future and identifying areas that require emergency response due to flood damage,” says Dr. Ali Mostafavi, director of the Urban Resilience AI Lab. “Therefore, it is imperative to devise proper methods and tools to monitor the inundation propagation in transportation infrastructure.”
A Holistic Flood Monitoring Framework
To effectively leverage these innovative technologies and data sources, flood management programs might want to adopt a holistic, integrated approach to flood monitoring. This framework should incorporate the following key elements:
1. Observability Analysis: Employ graph-theoretic techniques to identify the minimum number and optimal placement of physical flood sensors required to fully monitor the transportation network.
2. Crowdsourced Data Integration: Assess the reliability of crowdsourced flood reports and determine how they can supplement physical sensors to enhance observability in blind spots.
3. Remote Sensing Integration: Incorporate high-resolution remote sensing data to provide comprehensive spatial coverage and aid in damage assessment.
4. Predictive Analytics: Develop advanced forecasting models that integrate sensor data, weather information, and crowdsourced insights to predict flood inundation and inform emergency response.
5. Multi-Criteria Optimization: Use a combination of factors, such as infrastructure criticality, social vulnerability, and flood exposure, to determine the optimal sensor placement that maximizes the effectiveness of the overall monitoring network.
By adopting this holistic approach, cities and transportation agencies can build robust, resilient flood monitoring systems that leverage the strengths of various technologies and data sources. This will enable better decision-making, faster response times, and more effective mitigation strategies in the face of increasingly frequent and severe flooding events.
“Integrating crowdsourced data, such as 3-1-1 reports, into flood monitoring has the potential to supplement physical flood sensors which would reduce the need to install these sensors and offer new insights on the flooding status of the road network,” concludes Dr. Esparza. “This approach can help city managers improve flood monitoring by leveraging socially sensed data to supplement physical sensors, especially in blind spots where no flood gauge exists.”
To learn more about innovative flood monitoring technologies and strategies, visit Flood Control 2015.
Example: Manchester Advanced Flood Control Project 2024
