In November 2020, a tropical storm flooded neighborhoods across Tampa Bay, catching residents off guard.
A high tide, elevated by the storm’s effects, meant Tropical Storm Eta packed extra punch, raising the question: As global warming brings rising seas, how much worse will floods get?
To find out, the Tampa Bay Times partnered with the National Hurricane Center at the National Oceanic and Atmospheric Administration. The agency forecasts storms as they approach to warn residents and prepare local governments.
The Times asked the center to use its computer model to estimate Eta’s flooding in 2020, as well as the flooding if sea levels were higher.
Times reporters compared the modeled flooding to data on buildings and property to learn what was at risk. The results: By 2050, if Eta hit again in exactly the same conditions, twice as many properties could suffer flood damage — in a best-case scenario.
Here’s where that number comes from.
Plausible sea-level rise scenarios
For NOAA’s 2017 report “Global and regional sea level rise scenarios for the United States,” researchers produced a range of estimates for how quickly waters could climb in the coming decades. The range considers uncertainty in how much and how fast seas rise. NOAA outlined global scenarios ranging from “Low” to “Extreme.”
In 2019, the Tampa Bay Climate Science Advisory Panel — a group of local environmental experts— settled on its latest recommendations to the Tampa Bay Regional Planning Council. The panel, formed in 2014, recommended planners use the regionally-corrected data from three of those scenarios: “Intermediate-Low,” “Intermediate” and “High.” That ruled out the “Low” and “Extreme” cases.
For this project, the Times used those three projections, plus NOAA’s “intermediate-high” scenario, which was also within the suggested range. Reporters chose to look at the year 2050. The NOAA data is based on change from the year 2000, so to calculate the expected rise of the next three decades, reporters subtracted the projected 2020 rise in each scenario from the corresponding 2050 rise.
For example, in the Intermediate-Low case, NOAA projected St. Petersburg’s 2000 water level to have risen 0.36 feet by 2020 and 0.95 feet by 2050. The expected amount of sea-level rise from 2020 to 2050, then, is 0.59 feet (0.95-0.36).
Authoritative surge flooding models
National Hurricane Center scientists regularly model and forecast surge flooding as hurricanes and tropical storms approach the United States.
The Times asked the agency in 2021 to use its technology to model Tropical Storm Eta in both 2020 and potential future conditions. National Hurricane Center storm surge specialist Cody Fritz did so, using the National Weather Service’s Sea, Lake, and Overland Surges from Hurricanes model.
Fritz used the observed data from Tropical Storm Eta and ensured the modeled surge was in line with water gauge measurements taken as the storm passed.
The agency produced maps showing an estimate of the storm surge flooding all over the Tampa Bay region, calculating flooding using land elevation data from NOAA’s Office for Coastal Management and the United States Geological Survey.
Fritz ran the models using both predicted “normal” tide levels and the observed tides from when Eta hit. (The normal case is based on the present National Tidal Datum Epoch, which covers the period from 1983 to 2001. In 2020, the sea level in St. Petersburg was about 6 inches higher.) For both cases, he raised sea levels by the amounts chosen by the Times and ran the models again.
The agency provided the Times with 10 maps of the maximum flooding over land in feet.
In some ways, these are conservative estimates for overall flooding. The modeling only looks at storm surge, and it does not consider the extra height delivered by waves. It also does not consider flooding from rainfall, which can lead to higher water, especially inland.
In addition, the Times downloaded data produced by researchers at Virginia Tech and the University of North Carolina on the probabilities of storm surge flooding based on hundreds of simulated storms. The data was converted to raster images using linear interpolation.
Relevant local data
To estimate how much development was at risk of flood damage, the Times needed to know more than simply how much land would be under water.
Times reporters downloaded geographic information system data on building footprints from Microsoft, which included the shapes of nearly 7 million Florida buildings. (The dataset now is updated with slightly more records.) The Times also downloaded geographic data of the state’s 2019 property parcels and tax information, compiled by the Florida Department of Revenue, the respective County Property Appraiser’s Office and the University of Florida GeoPlan Center.
The Times used PostGIS to organize the modeled flood maps, building footprints and property parcels in a PostgreSQL database.
Where the National Hurricane Center data showed potential flooding from Eta (in 2020 or in the future), the Times calculated the modeled water height at the geometric center of each building. This shows, generally, how much flooding over the ground there would be in the middle of a building footprint — however, that does not mean the model suggests water inside the building if, for example, the first floor is elevated. In many communities, including flood-prone blocks of South Tampa, St. Petersburg and barrier island cities like Madeira Beach and Treasure Island, new homes are built several feet above ground to meet modern building standards and limit flood damage. Next to those properties often sit slab-on-grade beach cottages that predate strict construction codes. All are included in this analysis, even if the level of flooding — and damage — inside the homes would differ. There is no comprehensive way to measure the elevation of every home across the region, but flooding on a property can cause damage, regardless of whether it reaches inside the house.
Buildings were spatially joined to properties so reporters could look at home values or land use codes and count totals in given cities or neighborhoods.
It was determined a property could potentially flood from Eta in a given scenario if a building on its lot was modeled to flood at least 1 foot. The Times excluded buildings with less than a foot of modeled flooding to eliminate edge cases; the National Hurricane Center also focuses on areas with at least 1 foot of modeled flooding in its advisories.
Often, property parcels overlap or multiple buildings fall on one lot. When reporters counted properties that could flood, only unique building footprints on distinct parcels were used, to avoid double counting.
Reporting beyond the data
Reporters could compare the model of Eta flooding under the real 2020 conditions to what people saw on the ground. Using reports from Pinellas County and Madeira Beach as well as interviews with residents and local emergency officials, the Times could examine how flooding estimates matched up with eyewitness accounts and photos. Sometimes the modeled flood was too high; sometimes it was too low.
Scientists stress that computer models produce estimates and aren’t meant to precisely predict flooding at every home. Data, even collected by government officials, can be imperfect. And where predictions might be accurate, many buildings are still elevated above the ground.
Though the scientific consensus is that seas are rising — and that the rise is accelerating — it is unknown exactly how high waters will reach in the future.