Alarmist, exaggerative, storm surge report by CoreLogic
May 6, 2011
As noted by a commenter below, the Florida-Times Union has published a story based on this report as well, causing me to notice an error regarding Hurricane Dora, see last bullet point.
A few days ago, the Virginian-Pilot posted a brief story “Hampton Roads among most vulnerable to hurricanes“. The story did not link to the report that it was based on, 20111 Corelogic Storm Surge Report, but I hunted it down and dropped a couple of clarifiying remarks on comments that readers had posted to the story and was prepared to leave it at that. However, after reading the report thoroughly, I had enough concerns to motivate me to compose this post.
Corelogic describes itself as “a business partner that provides high-value information, analytics and business services that help you gain dynamic insights into your most complex business problems—insights that help you reduce risk and improve performance” and states that they “develop original proprietary research, and track current and historical trends, in a number of categories, including consumer credit, capital markets, real estate, mortgages, housing, fraud, regulatory compliance, default, natural hazards and disaster projections. ”
The report covers 10 cities/metro areas: Charleston, Corpus Cristi, Houston-Galveston, Jacksonville (FL), Long Island, Miami-Dade, Mobile, New Orleans, Tampa, and Virginia Beach. For each city, they gave a categorical score for hurricane probability, storm-surge vulnerability, and residential density, along with a dollar value for residential homes exposed to storm surge from a category 5 hurricane (category 4 for Long Island).
It was their “hurricane probability” ratings that caused me alarm, upon further review. Jacksonville, Florida, an area which despite being in the Sunshine state, is rarely affected by hurricanes, much less direct strikes, was rated as having “extreme” hurricane probability. Mobile, Alabama, on the other hand was rated as merely “high” probability. This is incomprehensible on first look.
From NOAA’s revamped hurricane hurricane tracks site, we see that 10 hurricanes have passed within 65 nautical miles of Jacksonville in the 1908-2008 timeframe, with only one actually making landfall on the coast of the metro area.
From the same site, we see that 17 hurricanes passed through the 65 mile radius around Mobile, Alabama, with only a couple making incidental “pass-bys”.
From this, it is clear that the probability of a hurricane bringing storm surges to the respective areas is much higher in Mobile than in Jacksonville. This is shown in the probability calculations from the “United States Landfalling Hurricane Web Project“. They calculate that in any given year, Mobile has a 2.2% chance of 1 or more hurricanes making landfall and a 1% chance of 1 or more major hurricanes doing so. The respective probabilities over a 50 year period are 67.6% and 39.7%. The probabilities for Jacksonville are substantially lower, with annual probabilities of 0.5% and 0.1% and 50 year probabilities of 21.5% and 6.4%, respectively.
So given this, how on earth does Jacksonville rate having a higher hurricane probability than Mobile? The answer seems to come in CoreLogic’s methodology:
The storm surge model utilizes a numerical index equated to
a categorical score of low to extreme to assess relative vulnerability to hurricanes for U.S. coastal counties. The index includes measures of both incidence and exposure. Incidence is measured by the number of landfalling hurricanes impacting a county over the past century. Exposure is quantified by both population and property value subject to hurricanes.
THIS HAS NOTHING TO DO WITH HURRICANE PROBABILITY! By factoring in population size, they are saying all other things being equal, a city with more houses has a higher hurricane probability! This is absolute non-sense!
Another problem with this, is hurricane incidence doesn’t necessarily reflect storm surge incidence. Given that storm surge is the focus of the report, this is important. As seen in the 100 year history for Jacksonville, the ratio of surge-generating hurricanes to hurricanes that are just passing by is low. Compare that to Long Island/Nassau County:
Even though more hurricanes pierced Jacksonville’s circle (ten compared to six), more surge generating storms affected Long Island. Yet, the report only reflects Long Island as having “low hurricane probability”. Also, the afforementioned US Landfalling Hurricane Project does not concur with the relative assesment and shows higher probabilities for Nassau County (Long Island) than Duval (Jacksonville) both for being affected by hurricane force winds and being affected by a landfalling hurricane. Given that Jacksonville had a higher probability than Mobile, which seems to be explainable due to factoring in population size, it is inexplicable to me how Long Island was given a “Low Probability” while Jacksonville was given “Extreme”.
Compared to the wrongness of the “hurricane probability” the other issues I have are small, yet still worth of note:
- In asessing storm surge damage, worst case scenarios are assumed. I.e., hurricane making landfall at high tide, perpendicular to the coast, with the city in the right-front quadrant and the storm at the top end of the Saffir-Simpson category. Only a small percentage of storms affecting an area would meet this criteria. On top of that, the report focuses on what damage would be from a category five hurricane landfall, an event that has only been recorded three times in U.S. History. (Also, it makes one wonder what they defined as the maximum wind speed for category five as the Saffir-Simpson scale, defines it as wind speeds greater than 155 mph.) The report does not give any details for (more) likely scenarios.
- In a couple of places where cities’ vulnerabilities are discussed, storm surge flooding is confused with flooding from rainfall. Discussing Virginia Beach, the report states “Virginia has not been seriously hit since Hurricane Floyd in 1999. That hurricane had
storm surge between 9 to 10 feet with total damage of $30–40M caused by flood damage from rain, surge and hurricane-spawned tornados.” As the storm made landfall near Cape Fear (well south of the Virginia state line), storm surge was not an issue for the Commonwealth. The damage was from inland flooding in places like my mother’s hometown, Franklin. (Also, the tornadoes only affected North Carolina, per the National Hurricane Center report. The same error is made in describing Tropical Storm Erin’s effect on Corpus Christi; rainfall is not storm surge!
- Bizarrely, the discussion on Tampa mentions Hurricane Andrew: “Tampa was also impacted by Hurricane Andrew in
1992 as it crossed the southern peninsula of Florida. It hammered the Tampa/Miami area of Florida, causing $43.7B worth of damage, one of the most destructive U.S. hurricanes on record.” Andrew had no affect on the Tampa metropolitan area! (see the damage listing in NHC report on Andrew ; none of the Florida counties are close to even bordering Tampa). As over 200 miles separate Tampa (on a southeast-northwest line) one can’t consider them to be part of the same area .
- In the Jacksonville discussion, the report incorrectly states that Hurricane Dora of 1964 made landfall as a “category one storm”. In fact, as shown by the historical track, Dora was at category three strength prior to landfall and the first advisory after landfall rated it at category two. In its list of costliest storms (table 3b, page 9) , NOAA reports it as a category two landfall.