Hurricane Katrina Wind Investigation Report

Abstract

This investigation of roof damage caused by Hurricane Katrina is a joint effort of the Roofing Industry Committee on Weather Issues, Inc. (RICOWI) and the Oak Ridge National Laboratory/U.S. Department of Energy (ORNL/DOE). The Wind Investigation Program (WIP) was initiated in 1996. Hurricane damage that met the criteria of a major windstorm event did not materialize until Hurricanes Charley and Ivan occurred in August 2004. Hurricane Katrina presented a third opportunity for a wind damage investigation in August 29, 2005. The major objectives of the WIP are as follows: (1) to investigate the field performance of roofing assemblies after major wind events; (2) to factually describe roofing assembly performance and modes of failure; and (3) to formally report results of the investigations and damage modes for substantial wind speeds The goal of the WIP is to perform unbiased, detailed investigations by credible personnel from the roofing industry, the insurance industry, and academia. Data from these investigations will, it is hoped, lead to overall improvement in roofing products, systems, roofing application, and durability and a reduction in losses, which may lead to lower overall costs to the public. This report documents the results of an extensive and well-planned investigative effort. The following program changes were implemented as a result of the lessons learned during the Hurricane Charley and Ivan investigations: (1) A logistics team was deployed to damage areas immediately following landfall; (2) Aerial surveillance--imperative to target wind damage areas--was conducted; (3) Investigation teams were in place within 8 days; (4) Teams collected more detailed data; and (5) Teams took improved photographs and completed more detailed photo logs. Participating associations reviewed the results and lessons learned from the previous investigations and many have taken the following actions: (1) Moved forward with recommendations for new installation procedures; (2) Updated and improved application guidelines and manuals from associations and manufacturers; (3) Launched certified product installer programs; and (4) Submitted building code changes to improve product installation. Estimated wind speeds at the damage locations came from simulated hurricane models prepared by Applied Research Associates of Raleigh, North Carolina. A dynamic hurricane wind field model was calibrated to actual wind speeds measured at 12 inland and offshore stations. The maximum estimated peak gust wind speeds in Katrina were in the 120-130 mph range. Hurricane Katrina made landfall near Grand Isle, Louisiana, and traveled almost due north across the city of New Orleans. Hurricane winds hammered the coastline from Houma, Louisiana, to Pensacola, Florida. The severe flooding problems in New Orleans made it almost impossible for the investigating teams to function inside the city. Thus the WIP investigations were all conducted in areas east of the city. The six teams covered the coastal areas from Bay Saint Louis, Mississippi, on the west to Pascagoula, Mississippi, on the east. Six teams involving a total of 25 persons documented damage to both low slope and steep slope roofing systems. The teams collected specific information on each building examined, including type of structure (use or occupancy), wall construction, roof type, roof slope, building dimensions, roof deck, insulation, construction, and method of roof attachment. In addition, the teams noted terrain exposure and the estimated wind speeds at the building site from the Katrina wind speed map. With each team member assigned a specific duty, they described the damage in detail and illustrated important features with numerous color photos. Where possible, the points of damage initiation were identified and damage propagation described. Because the wind speeds in Katrina at landfall, where the investigations took place, were less than code-specified design speeds, one would expect roof damage to be minimal. One team speculated that damage to all roofs in the area they examined was less than 10% when improper installation and deterioration were eliminated as causes. Roofs designed to code and installed according to manufacturers recommendations performed very well.