We led two of our post-disaster survey teams of engineers dispatched to areas affected by the storm within days of Katrina's landfall on the Gulf Coast of Louisiana.

Typically, our damage survey teams fly in and stay in the hardest hit areas, and subsequently fan out. This was impossible in the case of Katrina, where airports were closed and hotels still standing were either closed or filled to capacity.

Instead, our teams flew into Pensacola, Fla., and worked their way west across Katrina's track. All of the impacted coastal areas of Florida, Alabama and Mississippi were surveyed. New Orleans was still largely inaccessible to any but emergency workers, but we were able to survey from Slidell--across the north side of Lake Pontchartrain-some of the areas with the worst wind damage in Louisiana.

The effort was far from a laboratory exercise, as many of the affected areas lacked power, running water, food and street signs. Moreover, there were curfews in effect, as well as a chaotic flood of evacuees, emergency responders, law enforcement and military personnel.

The primary objective of the damage survey is to validate our modeled scenarios with actual observations of the damage. Our clients can then better serve their customers, by using this information to triage their claims and redeploy their forces to cover areas with the highest average claims severity level.

Our other objective is to record details of the actual damage to use along with detailed claims data to validate and enhance various aspects of the model.

For instance, our damage surveys after each of the four storms in Florida last year found that repair costs were multiplying due to the cumulative effect of the storms on the regional economy. As a result, we modified our 2005 model to enable clients to account for demand surge on an aggregate as well as an occurrence basis.

Hurricane Katrina was both very intense and very large. Katrina's hurricane force winds extended out 120 miles on either side of the storm and more than 150 miles inland.

Our teams observed wind damage as far east as Pensacola, Fla., and as far west as Baton Rouge, La. Traveling inland, damage extended north to Jackson, Miss., and east along Interstate 20 to Meridian, Miss., almost 150 miles from the coast.

Generally, the wind damage was consistent with what we expected from a storm of this size and intensity, and consistent with estimates produced in real time by our hurricane model.

The survey teams selected, at random, multiple three-by-three or four-by-four block sections of each town. They counted the actual number of houses and businesses and estimated the percentage of damage to each structure. Each team typically completes 10-to-15 surveys a day.

The median wind-speed footprint generated by our hurricane model served as a guide for planning our survey routes. Here are some of the highlights of what we found:

o On the fringes of the damage footprint in Pensacola, Fla., and Baton Rouge, La., most of the damage was non-structural. Missing roofing shingles and damage to siding were common.

o There were also cases of fallen trees impacting residential properties, which is typical for the 60-to-70 mph sustained winds experienced in these areas. In the Mobile, Ala., area, with its numerous neighborhoods wooded with mature southern yellow pines, downed trees were a common cause of loss.

o Property damage became more severe as the teams traveled closer to the center of the storm's path.

Hurricane Katrina made landfall southeast of New Orleans, sparing the city its most intense winds. Instead, Mississippi's coast bore the brunt both of Katrina's highest wind speeds and highest storm surge, impacting the well-populated areas of Gulfport and Biloxi.

Pascagoula, Miss., and Slidell, La.--which experienced maximum sustained winds of 90-to-100 mph--suffered significant damage. While parts of Slidell were ravaged by storm surge, the town also sustained significant wind damage.

o Damage to commercial structures in these areas was widespread, particularly to cladding and roofs. As indicated by our hurricane model and confirmed through our observations, we estimate that wind damage will play a large role in the overall insured loss in the city.

o While the most severe wind damage occurred within a 25-to-30 mile swath to the right of Katrina's track, severe wind damage was observed well inland in Wiggins, Miss., and even as far north as Hattiesburg.

o As expected, buildings of inferior construction fared poorly and light metal structures fared the worst. Unreinforced masonry--a widely used construction type in the region--also performed poorly.

o Use of vulnerable cladding materials such as Exterior Insulation and Finish Systems exacerbated damage to hotels, casinos and other commercial structures.

Only three weeks after Hurricane Katrina, the long-term effects on the Gulf States and their economies remain unclear. Our ground-level view confirmed early expectations that Katrina will be not only an unprecedented event for the insurance industry but a humanitarian disaster on an immense as well as individual scale.

While our survey results are intended to assist companies in analyzing the impact and managing their response, some of the more immediate activities we engaged in on the ground--offering a homeowner a gallon of water, a newspaper or a flashlight--were the most valuable services we provided.

Mohit Pande is senior wind engineer at Boston-based AIR Worldwide Corp., and Glen Daraskevich is an assistant vice president in AIR's research and modeling department.

Caption:

Many high-rise commercial structures in the Gulfport and Biloxi, Miss., areas experienced storm surge damage on lower floors and wind damage on upper floors.

Photo By AIR Worldwide Corp.