By Josh Jensen, AScT, CHI, RRO, RRC
For the last few years, the media has highlighted roofing failures caused by wind. Although these problems sensationalize the effects of global warming, they are not new and have occurred since buildings were first constructed. The major contributing factor to many of these failures is the roof or perimeter flashing was not properly designed to meet the project requirements––such as location and occupancy type.

Depending on its direction, wind acts on the building in different ways. As it collides with the structure, it travels up and over, increasing in speed––similar to the effects of an airfoil on a wing. This creates a negative suction wind load on the roof membrane that is not uniform; the result is various areas that have higher wind load (i.e. corners) from lower wind load (i.e. field).

The effects on the flashing are much more dynamic. As the wind travels up the building, it gets behind the flashing and pries it up; at the same time, however, the wind colliding with the flashing exterior is pulling on it from the top. Of course, this only occurs with the ‘perfect’ building––a box with a square roof located in a field. It gets more dynamic when there are interesting building shapes, locations, and configurations.

One often hears about catastrophic failures where the entire roof system becomes detached from the building and ends up in the parking lot or on the neighbouring facility, but failures can also be small where a fastener pulls through a roof membrane or blows off. Unfortunately, wind-related damage due to improper design happens more frequently than most people think. Often, an owner calls a contractor to come repair the damage, but does not dwell on it too much.

The cause
This author has observed many roof failures that, despite being attributed to wind forces, were a result of contractor error. However, if one looks at a specific group of projects with a roof design segment, most consequent failures are as a result of poor design. While improper design pressures are occasionally used, the majority occurs where design pressures and roof assembly are correct, but the various flashings and accessories are not.

The vast majority of catastrophic failures due to wind effects are a result of improper flashing design at the roof’s perimeter. The perimeter flashing, which is most commonly considered moisture protection, is also a significant part of the wind attachment system for the roof assembly. When properly installed, flashing acts like a large termination bar holding the membrane in place. Wind comes up the side of the building and gets under the flashings and membrane at the perimeter. If the flashing is not properly secured and becomes dislodged from the edge of the building, the rest of the roof is exposed and the roof starts to ‘peel’ off the underlying substrate.

Where do designers go to get the proper information to design the roof system to ensure the roofs do not fail? Most specifications provide a reference to a Factory Mutual (FM) Global requirement, such as a FM 1-60 rating. However, FM Global is an insurance company, and its criteria are based on testing and past losses. Is it correct to use FM requirements on buildings they are not insuring? What other documents or standards do designers have at their disposal? Are any of them relevant to Canada?

Canada versus U.S. standards
There is an important difference between Canada and the United States with respect to wind design. Referenced wind velocity pressures within National Building Code of Canada (NBC) Appendix C, “Climatic and Seismic Information for Building Design in Canada,” are converted from referenced wind speeds. The wind speeds are nominally one-hour averages of speeds, representative of a specific height and exposure type. This varies greatly from practice in the United States as its codes provide the representative wind speed rather than a velocity pressure.

Additionally, U.S. wind speed is based on a three-second gust, which will be significantly higher than an averaged hour. For example, if one looks at Windsor, Ont., NBC has a 0.47 kPa (0.07 psi) for a 1/50 occurrence, which converts into 97.2 km/h (60.4 mph). Just across the river is Detroit, which, according to the U.S. wind speed maps, has a wind speed of 144.8 km/h (90 mph). These two locations should have relatively similar wind speeds.

When completing wind design for roofing, most designers start with NBC. Part 4, “Structural Design,” outlines the formula used to determine the pressures applied to each building surface. This is further explained in the NBC Structural Commentary Guide, which provides the pressure applied to the structure without safety factors (i.e. 2.4 kPa [50 psf]). Where does one go from here and what does it mean?

Many designers would then insert a line in the specification stating the roof assembly should meet the requirements of FM 1-60. The problem is FM 1-XX rating wind uplift requirements do not have built-in safety factors. Instead, the calculation completed through the FM Global standards provides the safety factor, so the numbers completed from NBC cannot be used to compare to the wind uplift requirements. Even the FM Global calculation has different variables and factors than the calculation contained within NBC. Currently, there are no standards for designers to refer to in NBC, wherein they may apply the information from the calculations completed through Part 4 of the building code and select a roof system.