IN LATE JUNE, the First Street Foundation, a nonprofit research and technology group, publicly released the results of its new First Street Foundation Flood Model, providing free flood-risk data for more than 142 million U.S. homes and properties nationwide. According to the foundation, its model reveals that millions of properties not classified as flood prone by the Federal Emergency Management Agency (FEMA) are in fact at substantial risk of flooding. The findings attracted significant media attention, garnering headlines in such publications as the New York Times, the Washington Post, and the Wall Street Journal.
The First Street Foundation Flood Model uses a probabilistic approach to evaluate anticipated flooding depths for events of various return periods, ranging from 1-in-2-year to 1-in-500-year events. The resulting inundation depths are tracked at a 3 m spatial resolution for the entirety of the contiguous 48 states. To account for anticipated climate change, the methodology includes the output of 21 global climate models, enabling projections of changes in flood risk for five-year periods through 2050.
The methodology accounts for flooding from riverine (fluvial), stormwater (pluvial), and coastal sources, both tidal and storm surge. In this way, the methodology creates an “estimated flood depth from any source at any likelihood at a high resolution across the country,” according to The First National Flood Risk Assessment: Defining America’s Growing Risk, a report released by First Street in June that summarizes the results of its findings. “Modeling the data using this approach allows for the estimation of flood probabilities and likelihoods for any depth of water for any location in the country,” the report states.
The methodology accounts for flooding from riverine (fluvial), stormwater (pluvial), and coastal sources, both tidal and storm surge.
For its flood model, First Street incorporates data from multiple sources, including Fathom–USv2, a program developed by the modeling company Fathom to model inland and coastal flooding throughout the United States. Other sources include stream gauge data from the U.S. Geological Survey (USGS) as well as precipitation frequency estimates and tide gauge data from the U.S. National Oceanic and Atmospheric Administration (NOAA).
To represent actual conditions to the greatest extent possible, First Street created a database of traditional gray infrastructure as well as green infrastructure that is designed to mitigate flooding. “We ended up getting over twenty-three thousand individual features,” says Mike Amodeo, P.E., the director of data science at First Street. The foundation will continue to add more features as it becomes aware of them, he notes.
First Street’s flood model uses the various data sources to determine inundation depths on every property in the country. First Street then uses building data provided by the digital mapping company Mapbox to indicate the maximum inundation depth within the building footprint. If no building is present, the inundation depth is recorded at the center of the property.
“We’re sharing this individual data for free with the country so that people can understand their risk and learn more about flood awareness.” — Mike Amodeo, P.E.
For individual properties, the flood model develops a “flood factor,” which is a score from 1 to 10 indicating its cumulative risk of flooding during the course of a typical 30-year mortgage. Scores for individual properties may be found at FloodFactor.com. “We’re sharing this individual data for free with the country so that people can understand their risk and learn more about flood awareness,” Amodeo says.
First Street maintains that nearly 6 million U.S. properties not designated as flood prone by FEMA have substantial risk of flooding. (See the map above.) The foundation defines substantial risk as inundation of 1 cm or more within a building in a 100-year event, comparable to the Special Flood Hazard Area (SFHA) designation used by FEMA to indicate areas subject to flooding during a 100-year event.
“At the national level, the First Street Foundation Flood Model identifies around 1.7 times the number of properties as having substantial risk compared to the FEMA 1-in-100 SFHA designation,” according to First Street’s report. “This equates to a total of 14.6 million properties across the country at substantial risk, of which 5.9 million properties and property owners are currently unaware of or underestimating the risk they face because they are not identified as being within the SFHA zone.”
The top five states having the highest percentage of properties currently at risk of substantial flooding are West Virginia (24.4 percent), Louisiana (21.1 percent), Florida (20.5 percent), Idaho (14.8 percent), and Montana (14.2 percent).
Discrepancies between First Street’s findings and those of FEMA result largely from key differences between the modeling approaches used by the organizations. For example, First Street includes pluvial flood risk, while FEMA does not. “FEMA does not incorporate flooding due to stormwater in their flood hazards, which leaves many urban areas showing flood zones only along major rivers and streams,” Amodeo says.
Similarly, “our model provides full coverage of the lower forty-eight states,” Amodeo says. “FEMA has not mapped the entire country, [instead] prioritizing its resource expenditures in the most critical areas. However, forty percent of streams in the country are not gauged, making them harder to model. Through statistical inference and similarity scoring of catchments, our model imputes flow characteristics for ungauged streams, allowing us to quantify flood risk in previously unmapped portions of the country.”
In addition, the flood model accounts for changing environmental factors, rather than simply relying on existing conditions to represent current risk.
In addition, the flood model accounts for changing environmental factors, rather than simply relying on existing conditions to represent current risk. To examine future flood risks, First Street incorporates such expected environmental changes as rising sea levels, changing precipitation patterns, and warming sea surface and atmospheric temperatures.
Because of these anticipated changes, the number of U.S. properties having substantial risk will grow by 10.9 percent, reaching 16.2 million by 2050, First Street says. Coastal areas in particular will bear the brunt of this increased risk. For example, nearly 70 percent more properties in Louisiana will face substantial risk of flooding by 2050. Other coastal states expected to see significant increases in properties at substantial risk include Delaware (21 percent), New Jersey (19.1 percent), Florida (18.6 percent), and South Carolina (16.7 percent).
Amodeo emphasizes that First Street is not seeking to compete with FEMA when it comes to defining flood risk throughout the United States. “It is not our goal to replace FEMA floodplain maps,” he says. “They serve a different purpose than Flood Factor, and our work is very much complementary to theirs. Because they are used for regulatory purposes, FEMA maps are still the gold standard for flood projections in the United States. Flood Factor’s goal is to supplement FEMA maps where there are none and provide a greater level of information where they do exist by producing a gradation of both likelihood and depth of flooding.”
FEMA agrees with this assessment. “The FEMA Flood Insurance Risk Maps and First Street Foundation maps do not conflict with each other; rather, they complement one another by depicting different types of risk and will be helpful in different ways for an individual to determine flood risk,” said a FEMA spokesperson who replied in writing to questions from Civil Engineering. “Users should explore the differences between the maps to build a more comprehensive understanding of flood risk. We’re encouraged to see First Street Foundation building on federal agency datasets and the information included in our flood maps to bring together new data and technology in order to visualize flood hazard information in a different and useful way.”
“The FEMA Flood Insurance Risk Maps and First Street Foundation maps do not conflict with each other; rather, they complement one another.” — FEMA spokesperson
The Association of State Floodplain Managers also welcomes the introduction of the First Street Foundation Flood Model, says Meg Galloway, P.E., M.ASCE, a senior policy adviser for the association. “It has a lot of useful purposes,” Galloway says, including providing general property-level flood-risk information, especially in areas that have not been mapped by FEMA. That said, First Street’s approach involves a “big-data” model that may generate findings that are not as precise as those from FEMA, she says. “The FEMA maps generally have more focused data than you can get from the big-data scale of Flood Factor,” she notes. Similarly, FEMA maps must undergo a public vetting as part of their adoption process, whereas Flood Factor does not. “That doesn’t make Flood Factor bad,” Galloway says. “It just doesn’t make it suitable for regulatory purposes.”
To develop its flood model, First Street partnered with researchers and hydrologists from Columbia University; Fathom; George Mason University; the Massachusetts Institute of Technology; the research firm the Rhodium Group; Rutgers University; the University of California, Berkeley; and the University of Bristol.
As part of a partnership known as the First Street Foundation Flood Lab, the foundation is sharing its model and data with about 100 researchers at 20 academic institutions. The researchers will use the information to analyze such issues as “flooding’s impact on the U.S. housing market; its implications for lower-income and minority communities; its cost to federal, state, and local taxpayers; climate gentrification; and fairness in federal flood mitigation spending,” according to a June 29 news release from First Street. For-profit entities may pay to purchase access to the data.
This article first appeared in the September 2020 issue of Civil Engineering.