ChasingChasing thethe HurricaneHurricane

Full-scale measurements, combined with laboratory testing of hurricanes, improve housing structures so that they can withstand nature’s most powerful storms

By Gregory Kopp, Associate Professor and Canada Research Chair in Wind Engineering, University of Western Ontario Damage to townhouses in Biloxi, Mississippi caused by Hurricane Katrina.

umans build homes and Nature damage makes it difficult to establish active form of research, in which tests them. This has perhaps never standard designs that can withstand a researchers go to the storm, eliminates the Hbeen more true than in 2005, when specified level of risk. Controlled testing is problem of waiting for an unpredictable three of the largest hurricanes on record necessary and should be used with statisti- amount of time to observe a rare event at tore up the coastal regions of the Gulf of cally-based, post-storm damage surveys. a given location. The program, active for Mexico. In 2004, four hurricanes made The link that ties controlled testing with seven years now, is providing valuable landfall in Florida, and 2006 introduced damage surveys is the detailed measure- information about wind speeds houses itself with tornado outbreaks that killed ments collected during storms. face during a storm, the wind loads, and 54 more the 20 people in the U.S. Midwest in A popular joke in the wind engineering damage to adjacent structures under nomi- early April. In Australia, 2006 saw the community is that the best insurance to nally similar conditions. 6 0

0 largest tropical cyclone on the eastern protect a house from a hurricane is to 2 e

n coast in three decades. Tropical Cyclone install instrumentation to measure the WHEREWITHAL TO u J

• Larry was a Category 5 at its peak, with wind loads. Hurricanes are relatively rare DAMPEN DAMAGE a c . r

e wind speeds estimated at 180 mph. Larry events at any given location; recording Gurley and his team have been busy t i r w

r eventually made landfall with Category 4 measurements is therefore a difficult during the past two summers. They have e d n

u wind speeds. endeavour. The Florida Coastal chased storms and performed statistically- n a i d

a Monitoring Program, led by Professor based damage surveys, capturing details of n a c . KEEPING AN EYE Kurt Gurley at the University of Florida, is the damage, once the homeowners have w w

w ON THE STORM making the act of recording wind loads their claims data. This comprehensive Houses are complex structures. There easier with their portable, real-life experi- approach has helped lead to improve- is wide range of shapes, sizes and con- ments. They have placed instrumentation ments in new house construction. struction methods, as well as variability in in more than 20 houses along the Florida Information the team collected, for the materials and installation. There is also coast and have six portable, yet robust, example, led to an update to the Florida variability in the wind speeds and turbu- towers to measure the wind speeds near code in 2001, following the destruction lence houses experience during storms. these houses. The towers are designed to caused by Hurricane Andrew in 1992. Thus, the damage resulting from extreme sustain 200-mph gust wind speeds without Gurley’s subsequent research indicates the winds is also variable. This variability of overturning or sustaining damage. This new houses are holding together better, code compliance is up, but that losses billion of the US$120 billion report- due to water penetration are still ed in total losses. excessive. Wind speeds in coastal David Henderson of the Cyclone Mississippi were less than the design Testing Station at James Cook wind speeds, suggesting not much University in Australia has done sim- wind damage should have been ilar research. His work, based on the observed. There were structural fail- damage caused by Larry, has helped ures, but most damage resulted from to analyze construction standards cladding and roof cover failures. The revised in the mid-‘80s. Henderson cause of these failures seems to be found that older homes performed poor attention to detail. poorly under Larry’s winds, even Townhouse complexes in particu- homes that had been retro-fitted for lar did not fare very well. Stapling improvement (as a result of poor detailing). siding onto a structure, for example, If the 2004 Florida hurricanes is just not a good idea in a hurricane- proved it is possible to improve the prone region. Nails missing trusses resiliency of houses to wind, was a big issue for buildings struck by Hurricane Katrina illustrated there Hurricane Andrew and this problem are still a number of problems con- persists today. cerning where and how houses are The close proximity of many being built. sources of debris (such as the shin- In Katrina, the two largest sources gles and siding), in combination with of losses were the storm surge, which a lack of window protection, may Researchers set up a 10-metre-high tower devastated buildings right on the have exacerbated the problems with just hours prior to Hurricane Dennis mak- coast, and the levee failures in New the townhouse complexes. However, ing landfall. This tower measured a three- Orleans, which led to extensive one positive observation can be second gust speed of about 120 mph as flooding. Wind damage, though made: there appeared to be limited Hurricane Dennis’ eyewall passed over widespread, was not as devastating; wind damage in older neighbour- Navarre, Florida. it accounted for roughly US$10-20 hoods with mature trees.

55 w w w . c a n a d i a n u n d e r w r i t e r . c a • J u n e 2 0 0 6 Gregory Kopp (top left) participated in the Florida Coastal Monitoring Program a day before Hurricane Dennis made landfall. The grey “dishes” on the roof of the house are pressure transducers that collect information on wind loads affecting real houses during hurricane conditions. The houses are pre-wired so a team of four can set-up the “experiment” in about two hours.

DEMYSTIFYING DAMAGE Darwin, Australia in December 1974, BEHIND THE SCENE pioneered the first full-scale tests. These Damage surveys indicate where the tests indicated the strength of complete problems are and laboratory testing structures and their component parts. In allows factors that are widely variable in one important example, they showed the field to be examined in a controlled that drywall provides sufficient resist- way. Modern wind tunnel testing of ance to the horizontal wind forces in structures for wind effects began in the framed construction. ‘60s with a study of the World Trade A new research facility in London, Center towers in New York. The Ontario, supported in part by the approaches developed for that project Institute for Catastrophic Loss led to new methods of experiments and Reduction (ICLR), is called the Three analysis for wind loads along with the Little Pigs project. (See Canadian implementation of improved building Underwriter, April 2006). It is expected codes and design protocols. to make a significant contribution by Forty years later, much has changed; performing tests that will link accurate still, wind tunnel tests represent the wind tunnel loading data with full-scale state-of-the-art in determining wind structural testing. The results are antici- 56 loads notwithstanding Professor pated to close the loop on understanding Gurley’s work. The other side of the and mitigating hurricane wind damage. 6 0

0 equation – how strong the structure The result should be a more resilient 2 e

n actually is – involves full-scale testing. housing stock, built to fend off the u J

• The Cyclone Testing Station, formed worst events that nature can throw a c . r after Tropical Cyclone Tracy destroyed at us. e t i r w r e d n u n a i d a n a c . w w w

Soil characteristics caused this tree to fall on a house in Picayune, Mississippi during Hurricane Katrina. TREES: FRIEND OR FOE? There is an ongoing debate about whether trees offer a net benefit or a net cost to homeowners and insurers. Anecdotal evidence from Katrina indi- cates there is a significant benefit to having many trees around, as they pro- tect houses from the wind. The down- side is that when a tree falls on a house, it is almost certain to cause significant damage. Also, risk of fire increases when trees are in close proximity to houses. Tree blow-over was a significant issue in ‘03 during Hurricane Juan, which made landfall in Nova Scotia just west of Halifax. Millions of trees blew down, including about 400,000 in Halifax alone. This poses a serious chal- lenge to the forestry industry: it needs to harvest the lumber quickly following a hurricane so that it does not go to waste.

Storm surge as high as 30 ft in Biloxi, Mississippi destroyed most of the structures right on the coast.

WILD WATER WEATHER Katrina was really about the water - specifically the storm surge along the gulf coast and the flooding in New Orleans. The storm surge from Katrina was about 30 ft high in Biloxi, Mississippi. Even though the winds had 57

diminished somewhat as the hurricane w w w

made landfall, Katrina’s large size over . c a n a

the Gulf led to Category 5 surge levels. d i a n u

Measurements contribute to prepar- n d e r ing for and understanding natural w r i t e r

catastrophes - which seem to be on the . c a

rise in both frequency and severity - but • J u n

it is important to understand the vari- e 2 0 able reality of nature. One of Kopp’s 0 engineering colleagues, for example, 6 said he would live in New Orleans because, now that it had been hit, it would not be hit again. But recall the area was hit by the storm surge of Hurricane Camille in ‘69, leading to similar destruction. So, as history does seem to repeat itself, an event like Katrina may likely strike again.