Abstract
In light of the increasing hydrological challenges posed by climate change, aging infrastructure, and urbanization, effective stormwater management is crucial. The presentation, 'From Data to Action: Utilizing Innovative Technologies to Facilitate Adaptive Management in Stormwater,' reviews an advanced approach using innovative technology and data analytics to refine and continually adapt stormwater management strategies for the greatest benefit to the community. Following the broader discussion, we discuss the application of these approaches to managing quantity and quality impacts of stormwater in a Florida community. A comprehensive program starts with the utilization and analysis of static data layers in the development of foundational flood/impact models. These layers can include, among others, historical precipitation data, which provides insights into past rainfall patterns and their intensity; topography, which outlines the natural and man-made contours of the land; land use, which indicates areas of urban development, agriculture, or natural reserves; and soil type, which can influence water absorption rates and runoff patterns. To further refine these models, we employ GIS technology to spatially analyze and overlay these data layers, thereby enabling a detailed examination of areas prone to flooding. With this precision, vulnerability assessments become more accurate, guiding the development of mitigation strategies that are both effective and targeted to specific at-risk zones. While static data are invaluable in understanding the context within which we are working, the significance of IoT sensors in stormwater management cannot be overstated. These devices are pivotal in capturing real-time water level data from various sources, including rights-of-way (ROWs), surface waters like lakes and rivers, and intricate stormwater systems. The strategic placement of these sensors is crucial. For instance, utilizing publicly available data from USGS and NOAA to monitor river and tidal levels provides early warnings for potential riverine and coastal flooding, while sensors in storm drains can detect urban runoff surges and highlight bottlenecks within the system that must be addressed via maintenance or retrofit. Similarly, sensors within ROWs can monitor water levels within roadways. Combining these data with local weather stations allows us to capture microclimate conditions and hyper-local precipitation data. Integrating real-time water quality sensors with existing rainfall and water level data systems allows us to detect and evaluate various parameters, from pH levels to the presence of specific contaminants. When combined with rainfall data, we can assess how precipitation events might dilute or exacerbate pollutant concentrations. Similarly, water level data can indicate the flow and spread of these pollutants. By continuously monitoring these parameters, we can gauge the effectiveness and operational status of Best Management Practices (BMPs) implemented to treat and manage stormwater. If anomalies or inefficiencies are detected, it signals the need for maintenance or, in some cases, a complete replacement of the BMP. These analyses provide an opportunity to move beyond the traditional run-to-failure management approach toward a more forward-thinking strategy. The former, as the name suggests, involves operating systems until they break down, often leading to unplanned downtimes and potential damage. In contrast, the latter emphasizes maintaining a consistent Level of Service (LOS) and bolstering system resilience. This proactive approach ensures that stormwater systems remain operational and effective, even under challenging conditions. Central to this strategy is the use of predictive analytics that allow us to analyze and detect patterns and trends, enabling the prediction of potential system issues before they manifest. These algorithms have the potential to process vast amounts of data, identifying anomalies and providing insights, which in turn allow for timely interventions, reducing the likelihood of system failures. While machine learning plays a critical role in analyzing complex datasets and predicting nuanced patterns, the value of basic algorithms remains significant. These traditional algorithms provide consistent and reliable evaluations of system performance and BMP effectiveness. They can be used to calculate flow rates, assess sediment accumulation, or determine the efficiency of filtration systems. Their straightforward nature ensures that results are replicable and easily interpretable, making them indispensable in many stormwater management scenarios. When combined with the predictive capabilities of machine learning, which can analyze both historical and real-time data to assess risk factors like potential blockages or structural weaknesses, we achieve a comprehensive analytical toolkit. This dual approach ensures that stormwater systems, be it in small neighborhoods, larger Municipal Separate Storm Sewer Systems (MS4s), expansive watersheds, or broader regional setups, are evaluated holistically, ensuring comprehensive coverage and effective management. As an example of the above, Neptune Beach regularly struggles with the challenge of its sewer system flooding, leading to traffic disruptions even during storms that are within the system's design standard. To address this, the city took the initiative to deploy a network of sensors that monitored water levels and flow in real-time. The primary metric or Key Performance Indicator (KPI) they focused on was Pipe Capacity, with the objective of ensuring it remained below 50% during storms that met the design standard. Upon analyzing the data, the project team discerned two main factors. First, while tidal influences did affect capacity, they weren't the predominant cause of flooding. The more pressing issue was the combination of heavy rainfall events with an undersized stormwater system. With these insights in hand, Neptune Beach adopted an adaptive management approach. Instead of dispatching crews during every storm, the city's teams operated based on data-driven insights. They received alerts when the system reached over 50% capacity, preparing them for potential issues. If capacity exceeded 80%, the data was then used to make informed traffic management decisions. To further enhance their adaptive management strategy, Neptune Beach utilized mapping to visualize flood risk based on a series of static physical parameters. This mapping provided a clear picture of areas most susceptible to flooding. The real-time data from the sensors not only validated the accuracy of these flood risk models but also corroborated observations from residents and city officials. This holistic approach, combining mapped predictions with real-time validations and community feedback, ensured that the city's interventions were both data-driven and grounded in real-world observations. The long-term benefits of this comprehensive approach were substantial. Neptune Beach successfully secured funding from the state to design and construct a stormwater system better equipped to handle peak flow volumes. Furthermore, the city ensured that data-driven capacity planning was a cornerstone of the new system, equipping it to manage both standard and extreme weather scenarios effectively. In summary, 'From Data to Action' underscores the importance of innovative technology and data analytics in modern stormwater management. The methodologies and technologies discussed offer a pathway to both address current challenges and proactively prepare for future hydrological events. The projects highlighted demonstrate the practical application and benefits of these technologies in the field of stormwater management.
This paper was presented at the WEF Collection Systems and Stormwater Conference, April 9-12, 2024.
Author(s)E. Rothman1
Author affiliation(s)Stormwater Investment Group 1
SourceProceedings of the Water Environment Federation
Document typeConference Paper
Print publication date Apr 2024
DOI10.2175/193864718825159390
Volume / Issue
Content sourceCollection Systems and Stormwater Conference
Copyright2024
Word count14