Flood prediction and stormwater management requires hydrological models to simulate the movement of water from precipitation into streams through hydrological processes under various control and land management scenarios. While physically-based theories have been developed for characterization of individual processes, hydrologic modeling remains inherently data-driven, relying on measurements of precipitation, discharge, land usage, water usage, and hydrography that vary significantly throughout a watershed. The figure below shows hydrological processes simulated with watershed models, including precipitation, streamflow, infiltration, evapotranspiration, soil water movement, etc. Flood susceptibility studies in the Midwest, and nationally, have traditionally been undertaken and communicated on a watershed basis and have been primarily focused on riverine or fluvial flooding. While these assessments are vital to adaptation to changes in precipitation and surface and sub-surface hydrology in the region, they are not sufficient for understanding local-scale flooding, particularly urban or pluvial flooding.

One approach for more robust stormwater data collection and characterization is through citizen science. Researchers typically monitor water quality at select locations or in major river systems but lack sufficient resources to track water quality in every stream throughout a watershed. Employing volunteer citizen scientists is a potentially important approach to monitor flooding and water quality in urban areas. Citizen scientists can meaningfully contribute by collecting samples from nearby streams and lakes, identifying potential sources of pollution in local watersheds, and increasing the amount of water quality data that are available. An organized network of citizen scientists could potentially monitor a whole watershed or the storm infrastructure of a city. Citizen science not only increases spatial and temporal data quantity but also helps achieve public outreach by generating community awareness and increasing scientific literacy.

Novel sensor network and data fusion technologies may also help address some of the issues associated with urban stormwater data collection. Shortages in formal scientific infrastructure for stream and flood monitoring can be addressed by “crowdsourcing” data from sensor and communication systems such as cellular phones, traffic cameras, webcams, and social media. Crowdsourcing allows for near-real-time observations at the right place and time, i.e., locations where localized flooding occurs that affects human activities. For example, an urban flood model was applied to model a storm event in Shanghai that compared favorably with crowdsourced flood data.