Synthetic storms provide supplemental data for hurricane damage predictions
bodnarchuk/Getty Images
Emergency managers no longer have to rely on limited historical datasets to prepare for hurricane-related infrastructure damage.
A new tool gives state and local officials a closer look at potential property-level damage caused by tropical storms—and what it could cost—based on computer-modeled hurricane tracks.
A major obstacle for any research into hurricanes is a lack of historical data, one expert said. Throughout U.S. history, 38 tropical storms have reached Category 5 level, and only four—Andrew (1992), Labor Day (1935), Camille (1969) and Michael (2018)—made landfall.
“That’s not a huge dataset to work from … If you’re trying to compare the last decade or two of [hurricane behavior] versus hurricane behavior from 30 to 40 years ago, it is very difficult to do,” said Ed Kearns, chief data officer at First Street Foundation, a nonprofit dedicated to climate technology and research. “When we’re looking at climate change impacts on hurricanes, that makes the dataset even smaller.”
To fill in the gaps, FSF partnered with WindRiskTech, a hurricane modeling company, to program synthetic storms that mimic hurricanes’ typical characteristics such as precipitation levels, wind patterns and storm speed and direction. The model relies on high-resolution topography, property data provided by federal, state and local governments and hurricane records to project hyper-local wind hazards throughout the U.S.
The synthetic hurricanes can be manipulated to show how climate change will affect major weather events, which often goes overlooked, Kearns said during a webinar late last month. “This gives us much greater statistical certainty in the patterns that we're going to be seeing from climate change today and … looking out 30 years.”
For instance, the Florida Commission on Hurricane Loss Projection Methodology, an initiative created in 1995, leverages computer modeling to determine appropriate residential insurance rates for flooding and hurricane damage but largely relies on historical data without considering the evolving climate, FSF founder and CEO Matthew Eby said.
Using historical landfall rates in Florida, the commission found that the probability of a Category 5 hurricane touching land in Florida was 4.3% in 2023, but with the model’s synthetic data, FSF projected the likelihood to be 7.6%—nearly a 77% increase, according to a recent report. The former method creates a “blind spot” for disaster planning, Kearns said, which could lead to insufficient insurance coverage and recovery resources.
The foundation’s open-access website that provides climate risk assessments for flooding, heat and fire events added its findings on wind to its lineup late last month. The website can help government leaders make data-driven decisions to better prepare for and recover from severe weather events.
The site provides profiles on states and specific addresses that details a location’s susceptibility to wind damage, weaknesses in existing infrastructure, projected wind patterns such as speed and duration of wind gusts and other critical factors.
Profiles also suggest solutions to build community resilience such as planting more trees to reduce wind force or enhancing building codes to withstand disturbances. Property-level assessments further evaluate which building features are at risk of destruction and their estimated property loss costs.
Property parcel data was provided by Lightbox, a provider of location intelligence and workflow solutions. Wind damage, costs and repair time calculations were provided by Arup, a global engineering consulting firm, and Mapbox, a data and mapping services provider.
Over the next three decades, the annual average loss due to wind-related damage from tropical storms and hurricanes will increase from $18.5 billion to $19.9 billion, and 13.4 properties currently not at risk will become vulnerable, the report stated.
“If you don’t understand the level of risk, you may not prepare adequately,” Kearns said.