The City of Virginia Beach on the US east coast is using big data to predict flooding across the Virginia coastline and allow emergency services to prepare a response. The StormSense project, developed by academics at the Virginia Institute of Marine Science, will feed water levels, wind speeds, air pressure, and other data into a model which can predict which areas will be flooded up to 36 hours in advance. Emergency services will then be able to see the predictions visualised as a map, which the project team eventually plan to make publicly available, allowing residents to take their own measures against flooding.
Results & Impact
Prototypes of the StormSense technology used in the past in Virginia have predicted flooding up to 36 hours in advance. The installation of new sensors across waterways in the region around Virginia Beach will improve predictions of the reach and depth of any flooding, and allow StormSense to make predictions for a larger area. After New Orleans, the Hampton Roads coastal region in Virginia is the second largest population in the US at risk from sea level rise. In 2015, it was estimated that 400,000 properties in the region were at risk from storm surge inundation.
City of Virginia Beach; Virginia Institute of Marine Science, College of William and Mary; US Geological Survey; Christopher Newport University; State of VA Department of Health; Wetlands Watch
Using Internet of Things enabled sensors, StormSense will track water levels across the Hampton Roads region, and combine these with other factors such as wind speed, wind direction, air pressure and precipitation data. The data is aggregated and fed into the StormSense model, which will predict the extent and depth of flooding up to 36 hours in advance. From this, StormSense is able to map precisely which areas are likely to be affected and alert emergency services to problem areas. The aim of the project is to give emergency services time to prepare their response or evacuate areas likely to be heavily affected in advance. Eventually, the team hopes to make the predictions public to allow local residents to take their own precautions against flooding.
Virginia Beach, Virginia, US
Cost & Value
So far the project has been awarded $459,850, with funding coming both from the City of Virginia Beach and in the form of prize money and grants from various other sources. The project team have submitted three proposals for over $1 million to the National Science Fund and other federal agencies for funding.
Although the StormSense team has submitted three proposals for funding to federal agencies, it has had little success so far. Both Hurricane Sandy and Hurricane Matthew largely bypassed the Hampton Roads region on the Virginia coast, which has limited the flood relief funding available to Virginia Beach and StormSense. As a result, the team has had to install new sensors over a longer period of time.
The StormSense team is concentrating on densely populated areas on the Virginia coast first, but plans to expand the program to cover areas further inland over time by installing new sensors. Other hydrodynamic models are currently being piloted in cities across the US, including New York City, Washington DC, Charleston SC and Galveston Bay.
The City of Virginia Beach is using big data to predict flooding during storms, up to the point of warning local residents which streets are expected to be underwater up to 36 hours in advance.
StormSense, a joint project between the City of Virginia Beach and the Virginia Institute of Marine Science (VIMS), is setting up a series of sensors in waterways across the Hampton Roads region of coastline in Virginia to monitor water levels. This data, in combination with wind speed, direction, air pressure and precipitation information, is fed into a model, from which the StormSense team are able to predict the range and depth of potential flooding. StormSense is able to map precisely which areas are likely to be affected and direct emergency services to problem areas, allowing them evacuated citizens, close roads, or permit residents to take precautions to avoid damage to their property.
“The StormSense project is an iteratively predictive model capable of providing emergency managers with the ability to visualise flooding from storm-surge, rain and tides up to 36 hours in advance of the inundation events,” said Dr Derek Loftis of VIMS who leads the StormSense project.
While the data is mainly used to inform the long-term decisions the city makes to plan against flooding, the StormSense project will eventually allow citizens to make their own decisions. “If they need to take a different route home, if you’re dealing with tidal flooding, or if it’s something relating to a hurricane, maybe you’ll get a day or two more advance notice – those types of things usually end up helping quite a bit,” said Loftis.
The Hampton Roads Region, the body of water in Virginia and North Carolina made up of nine separate cities, is the second largest population centre in the US at risk from sea level rise after New Orleans. Based on data collected over the past decade, the Historical Hague region near Newport flooded no less than one cumulative week per year. The region has more the 400,000 properties at risk from storm surge inundation as of 2015, a figure that is only likely to increase as the population in Hampton Roads grows.
StormSense is using Internet of Things-enabled sensors to track the water levels across waterways in Hampton Roads. While many already exist, StormSense will install 24 new sensors through the year in order to widen its range. “Once you have a sensor-sensor network you can start to interpolate between the water levels sensors observe at each of the stations. That can provide you with a better picture of what is going on,” said Lofis. These are then fed into a model in combination with wind speed and other inputs, which crunches the data to predict flood activity.
The team will also collect water-level data crowdsourced from local residents, who can record water-level data in their areas during flooding via the SeaLevelRise app developed by Wetlands Watch, an environmental group in Virginia. This can then be compared with the flood predictions put out by the model, helping to refine it over time.
“The platform for integration is being set up in Amazon Web Services so that we can aggregate data from several cities for rainfall as well as water levels,” said Loftis. From the aggregated data StormSense puts out a map, showing which areas will be affected by flooding and how hard they are likely to be hit. “Ultimately, that’s part of our strategy in trying to disseminate this data to the general public in the later phases of this project. But first, we’ve mostly just been trying to produce maps that mean something to emergency responders and emergency response personnel.”
So far the project has been awarded almost $500,000 in funding and prize money, the majority of which comes from the city of Virginia Beach itself. The StormSense team have applied for funding from the National Science Fund, which would allow them to install more sensors and improve the accuracy of their model.
In 2015, VIMS applied a prototype of the current technology, which was able to predict floods during the September ‘King Tide’ flooding 36 hours in advance. Based on its success, staff from VIMS and Newport News came up with the idea for StormSense in October 2015. Planning continued through 2016, which resulted in StormSense participating in the Global City Teams Challenge, where it won a Replicable Smart Cities Technology Grant of $75,000.
StormSense will focus at first on modelling floods in the densely populated coastal cities in the Hampton Roads region, before spreading out to smaller neighbouring localities. Over the next year, the StormSense team will evaluate the success of its model by comparing its forecasts to actual flooding measured from crowdsourced flood records submitted via the SeaLevelsRise app and statistical analyses.
(Picture Credit: Pexels)