To the untrained eye, there’s no difference between one type of roof shingle and the next, whether the central air-conditioning unit is located on the first floor or raised on stilts, or if electrical outlets run along the bottom or top of the basement floor.

But to loss-mitigation professionals, every fastener and nail, every joist and beam, every construction detail in building permits reveals clues to how one home survived lashing hurricane winds and another home three doors down was ripped to pieces.

“We are scavengers of data, we seek out all the data we can find,” says Tom Larsen, senior vice president and product architect for CoreLogic EQECAT, a consumer, property, financial catastrophe analytics and modeling vendor in Irvine, Calif.

Larsen and his industry colleagues consume voracious amounts of data.

From ZIP codes to building codes, from topography to geography, from individual details of a house to the characteristics of the neighborhood in which the home is located, no detail goes uncataloged in an attempt to “steer us to a less catastrophe-prone future.”

Studying what failed

Tim Doggett, Ph.D., assistant vice president and senior principal scientist for the Boston-based modeling vendor AIR Worldwide, says that in the aftermath of Superstorm Sandy, the company sent teams of damage experts and scientists to collect information on what failed in buildings.

“Light metal warehouses in the surge zone didn’t do well,” he says.

If the catastrophe modeling industry isn’t quite living through an era of big data, it has come a long way since Hurricane Andrew slammed into south Florida in August 24, 1992, causing $16 billion in insured damage.

The storm, the most expensive natural disaster to hit the U.S. at the time, shattered probable maximum-loss estimates, rewrote the books on hurricane damage, caused nearly a dozen insurers to file for bankruptcy, and turned the Florida homeowners’ insurance market upside-down.

Andrew’s losses would soon pale in comparison with Hurricane Katrina in 2005, which caused an estimated $41 billion in insured losses.

Before Andrew, one insurance hurricane-model would run simulations for 90 different events.

More than 20 years later, larger databases, faster computers, data-intensive mapping techniques and more advanced software allow scientists to layer new information on top of other data have helped catastrophe modelers process wider data sets.

“Now we have models with outputs based on 100 million rows of data or individual records,” Larsen says.

Difference in trees

Data correlations have allowed modelers to note that trees located within municipal boundaries snap more frequently than trees located on private property as municipal trees are maintained less often, Larsen says.

With the Atlantic hurricane season now under way and Hurricane Arthur, the first named storm of the season, already sideswiping the Carolinas, data and analytics experts wonder what this season will bring.

Despite a 50 percent chance of a “below-normal” season and between eight and 13 named storms by the Climate Prediction Center, “go” teams of adjusters, engineers and loss-control specialists stand at the ready to respond at a moment’s notice.

Claims adjusters, statisticians, probability experts, scientists, software programmers and government officials pore over the data, engineering and financial components of the models to anticipate the likelihood and severity of potential future calamities.

Damage near the foundation or on the side of a house indicates water damage from storm surge, Doggett says.

Second- or third-floor damage points to the effects of wind but separating the effects of wind from the effects of water isn’t always easy. “It’s always a gray area,” Doggett says.

Insurance pricing

With every new hurricane that makes landfall in the U.S., advanced catastrophe modeling and analytics allow property-casualty carriers to more accurately price a homeowners insurance policy.

Models also help insurance carriers calculate the amount of capital they need to set aside in reserve to pay claims and how many catastrophe insurance policies insurers can afford to reinsure.

Catastrophe models help insurance companies plan ahead and serve as a tool that contributes to the industry allocating capital more efficiently, Larsen says.

Doggett also says catastrophe models have helped stabilize homeowners insurance prices, allowing experts to smooth out price volatility from one catastrophe-exposed region to another.

Catastrophe modeling is one of several variables used in insurance pricing, but homeowners tend to see insurance prices rise because property values in the long run also go up.

“Values tend to go up every year,” Doggett says. “Catastrophe is just one small layer of what is being covered.”

Source:  Tech Page One