Helene

Post-Event All Perils | Summary
Posting Date: October 8, 2024, 2:00:00 PM

Verisk estimates that insured losses from Hurricane Helene for wind, storm surge, and hurricane precipitation-induced flood to the private insurance market will fall between USD 6 billion and USD 11 billion.

Of the modeled private market insured losses, wind contributes about a majority, 60 to 75% depending on the scenario, with a bit more than half of the remaining losses driven by inland flood, and the rest attributable to storm surge.

Verisk estimates that total ground up damage due to wind, storm surge, and hurricane precipitation induced inland flooding from Helene could exceed USD 30 billion.

ALERT™ subscribers can now download Touchstone® and Touchstone Re™ event sets, loss-based similar stochastic events (SSEs), and a modeled flood extent shapefile for the combined wind, storm surge, and inland flood impacts from Hurricane Helene from the Downloads tab for this posting on the ALERT Website. Please see the commentary in the download for the loss-based similar stochastic events for guidance on how to best utilize these. The information provided herein is strictly confidential and is solely for the use of Verisk clients; disclosure to others is prohibited unless noted in your Verisk software license.

The three scenarios for Helene provided with this ALERT posting include updated modeled wind footprints for Events 1 and 2. The wind footprint for Event 3 is the same as was provided in the last Helene posting on October 2.  All three scenarios have precipitation-induced flooding added; this sub-peril was not included in the October 2 posting. No changes have been made to storm surge footprints; these are the same as provided in the October 2 posting.

Users may note that the low end of the industry loss range quoted in this summary is higher than the loss the lowest modeled scenario (Event 3) will directly produce in Touchstone Re when analyzed against indexed insured take up rates.  This difference is due to the following reasons:

• It is extremely challenging to consider the role of antecedent soil conditions across the footprint and its impact on wind damage estimation. Significant rainfall preceded Helene’s arrival, which saturated soils and led to greater than expected treefall and resulting damage. The higher loss scenarios (Events 1 and 2) account for this possibility across wide swaths of Georgia, North and South Carolina.
• Uncertainty in the estimated take-up rate of private flood insurance
• Given regions impacted by floods also experienced strong winds, there could be possible leakage of flood losses in wind policies not otherwise accounted for.

Modeling Information and Assumptions

Today’s modeling resources for Helene do not include:

• Losses paid out by the National Flood Insurance Program
• Losses exacerbated by litigation, fraudulent assignment of benefits, or social inflation
• Storm surge or inland flood losses paid on wind only policies due to government intervention
• Explicit modeling of losses due to landslides or mudslides
• Losses to inland marine, ocean-going marine cargo and hull, and pleasure boats
• Losses to uninsured properties
• Losses to infrastructure
• Losses from extra-contractual obligations
• Losses from hazardous waste cleanup, vandalism, or civil commotion, whether directly or indirectly caused by the event
• Losses resulting from the compromise of existing defenses (e.g., natural and man-made levees)
• Loss adjustment expenses
• Other non-modeled losses, including those resulting from tornadoes spawned by the storm
• Losses for U.S. offshore assets and non-U.S. property

Modeling of Inland Flooding

On September 25, while the storm center of Helene was still off the coast of Yucatán Peninsula, it was already raining over southern Appalachia from a predecessor rain event (PRE) from a stalled cold front fed by abundant moisture from the Gulf of Mexico. Extreme rainfall continued through September 26 and 27 as Helene made landfall at Florida's Big Bend region and moved inland. Record rainfall had fallen in 3 days (September 25-27) over southern Appalachia and in particular western North Carolina. The largest recorded 3-day rainfall of 31.33 inches was observed at the North Carolina Forest Service’s RAWS station in Busick. Asheville, which was worst hit, received 13.98 inches in 3 days, well beyond its 1,000-year return period quantile.

To produce the loss estimates, the Verisk team analyzed hourly precipitation fields over the U.S. derived from the National Weather Service (NWS) and National Water Model Analysis. As the vast majority of losses for this event come from the USGS Hydrologic Unit Code (HUC) regions, 3, 5 and 6 (https://water.usgs.gov/GIS/huc.html), the spatial extent of this event set, posted through Verisk’s ALERT service, is limited to these three regions that included the states of North Carolina, South Carolina, Georgia, Tennessee, eastern parts of Virginia, West Virginia, Florida, Kentucky, and Alabama. Perturbations based on the spatial correlation in precipitation accumulations were introduced to the adjusted precipitation fields to generate a large set of precipitation scenarios and, more importantly, to quantify the uncertainty associated with the precipitation input. This ensemble of precipitation fields (a total of 50 scenarios) was then used to model the runoff and river flow.  Each precipitation scenario generated a loss scenario resulting in an ensemble of loss estimates. Based on the estimated loss pattern across the event footprint, three scenarios from the ensemble catalog were selected to represent the range of estimated losses and made available as a part of the custom event set.

The observed instance of precipitation (base) and its 50 perturbations were used to drive Verisk’s hydrological model to create different possibilities of flooding patterns for both fluvial and pluvial flooding. The hydrological model was run beginning January 1, 2024 and soil moisture was continuously tracked all the way through the event to appropriately account for the antecedent soil saturation conditions. The streamflow gauge data from the USGS stations across Helene’s path were used to validate the model’s streamflows and where necessary, to bias correct the modeled flows. Along with the base case, two other perturbations were then selected to represent the range of hazard extremes.

Extensive validation of the modeled flood footprints was performed using observations that were collected by the Verisk team from public news media, crowdsourcing along with aerial imagery. Figure 2 shows one such validation exercise in Asheville, NC.

Additional Considerations

The National Flood Insurance Program (NFIP) and Flood Insurance Take Up Rates

Homeowners who choose to purchase flood insurance most commonly do so through the National Flood Insurance Program (NFIP), and Florida has the highest proportion of NFIP policies of any state – about 35%, according to Fitch Ratings. These policies typically cover up to $250,000 in damages and do not include additional living expense coverage, suggesting that even for some homeowners with flood insurance, there may be a gap between coverage and rebuilding costs.   

The flood insurance take-up on residential risks in Florida is variable depending on the county. Recent data from the NFIP (as of 8/31/2024) shows that the risk count based take-up rates of policies-in-force varies from 48-50% in Pinellas, Charlotte and Lee County to as low as 5-15% in a majority of counties along the Gulf coast including Taylor, Citrus, Dixie, Hernando, Levy, Pasco, Hillsborough counties. Within those counties, homes inside the government-designated floodplain, the area most exposed to flooding, the take-up rate tends to higher in comparison to areas outside the floodplain — many of which are still likely to have been damaged by flooding from Helene. This is the likely story in the Carolinas, Georgia, and elsewhere impacted by inland flooding. Data as of October 2024 from NFIP show the statewide risk count based take-up rates for policies-in-force in North Carolina around 4%. The take-up rates across counties in Western North Carolina where precipitation induced flood impacts were rampant is less than 1%. An analysis of Census Bureau and NFIP data by the Raleigh News and Observer indicates the same trend:  less than 1% of the homes in the counties in western North Carolina most impacted by flooding are not covered; the same is true in the most impacted counties in Tennessee. 

Most homeowners in these areas will not have flood insurance, given in most cases it is not required as these areas are generally outside of the designated Special Flood Hazard Areas (SFHAs). 

Verisk’s modeling of the damage from precipitation induced inland flooding indicates that total ground up damage from this aspect of Helene alone could well exceed $11 billion; only a fraction of this will be paid by private insurance. Users of this ALERT should therefore note that there is upside in the estimate of insured industry losses due to uncertainty in the actual private flood insurance take-up rate; while information on NFIP take-up is generally available, actual estimates of private insurance take-up is not well documented.

Additional ALERT Plans for Helene

The Verisk ALERT team will continue to study the impacts of Hurricane Helene; however, no additional ALERT postings are currently scheduled.  Please contact your Verisk representative with any questions about Hurricane Helene.

Post-Event All Perils | Downloads
Posting Date: October 8, 2024, 2:00:00 PM

The information provided herein is strictly confidential and is solely for the use of Verisk clients; disclosure to others is prohibited.

Simulated Event Sets

These event sets contain three custom modeled scenarios for Hurricane Helene's wind, storm surge, and precipitation induced flood, allowing for uncertainty in key model parameters such as radius of maximum winds and central pressure.  All scenarios follow the official NHC track.

Product	Description	Download
Touchstone Re	Selected set of simulated scenarios
Touchstone	Selected set of simulated scenarios

Loss Based Similar Stochastic Event IDs

These stochastic events were selected as best matches to the insured industry loss footprints generated by the modeled scenarios in the Simulated Event Set

Product	File Type	Description	Download
Touchstone / Re	Excel (xls)	Selected set of stochastic events, with Readme in Excel format
Touchstone / Re	CSV (csv)	Selected set of stochastic events, formatted for import into Touchstone/Re

Additional Downloads

Note: Additional downloads related to the posting are listed below. Please use the appropriate application to view these files.

Title	File Type	Description	Download
Hurricane Helene Modeled Flood Extent Shapefile	.zip	A shapefile for the base modeled flood footprint from the Simulated Event Set
Updated Hurricane Helene Modeled Wind Shapefile	.zip	A shapefile for the median loss (Event #2) wind footprint
Hurricane Helene Custom Modeled Scenarios Information	.xlsx	Event parameters and modeled insured industry loss estimates with sub-peril breakouts for the three custom scenarios for Hurricane Helene found in the Simulated Event Sets.