P 2.17 WIND EFFECTS ON ASPHALT SHINGLES

1 P WIND EFFECTS ON ASPHALT SHINGLES . Timothy P. Marshall*, Scott J. Morrison, Richard F. Herzog, Jeffrey R. Green Haag Engineering Co. Irving, Texas 1. INTRODUCTION. While wind damage to ASPHALT SHINGLES is usually ASPHALT SHINGLES have been utilized as a roofing obvious ( torn and missing SHINGLES ) there remain a material for more than one hundred years. The first number of issues with ASPHALT SHINGLES ( cupping, ASPHALT SHINGLES were manufactured in 1901 but were clawing, splitting, lack of bonding) which some people not mass produced until 1911 (Snoke 1941). ASPHALT falsely attribute to wind. Thus, it is the subject of this SHINGLES were originally made from cotton rags that paper to review and discuss the modes of wind damage were coated with ASPHALT and surfaced with slate to ASPHALT SHINGLES . particles (McNulty 2000). In the early 1960s, glass fiber mats were introduced as the base material for 2. WIND DAMAGE MECHANISM. ASPHALT SHINGLES which made the SHINGLES lighter and less apt to retain detrimental moisture (Cullen 1993).

2 Wind interacting with a roof is deflected over and Relatively inexpensive ASPHALT along with around it. As a result, uplift pressures develop on the advancements in mat and sealant technology quickly led roof. However, uplift pressures are not uniform and are to ASPHALT SHINGLES becoming the primary choice for highest along the windward corners, rakes, eaves, and steep slope roofing. In 2009, ASPHALT SHINGLES ridges (Fig. 1). It is at these locations that wind uplift comprised 57 percent of the roofing market covering damage initiates especially with ASPHALT SHINGLES that million roofing squares (The Freedonia Group are not well bonded (Fig. 2). Wind flow in these areas 2010). (Note: A square in roofing is 100 sq. ft. or is quite turbulent. Thus, it is important that the roof m2.) Today, there are literally billions of roofing covering receive additional anchorage in these high squares covered by ASPHALT SHINGLES in the United States wind uplift regions.

3 The Federal Emergency (Noone and Blanchard 1993). Management Agency (FEMA 2005) has published Not surprising, a large percentage of steep slope guidelines on how to attach ASPHALT SHINGLES in these roofing in hurricane prone regions is ASPHALT SHINGLES areas which involve the application of adhesive dollops because ASPHALT SHINGLES are more economical than their along the roof edge and between the ASPHALT SHINGLES in tile or metal counterparts. Further, stringent building high wind zones. codes and standards have led to tougher demands on the ASPHALT shingle industry. Today, ASPHALT SHINGLES can be rated for high wind zones up to 67 m s-1 (150 mph) as a three-second gust. However, there remain problems at the basic level with the design, manufacture, installation, and durability of ASPHALT SHINGLES . Assessments of ASPHALT shingle roofs by the authors after hurricanes have revealed deficiencies in each of these areas. Similar findings have been made by McDonald and Smith (1990) after Hurricane Hugo, Smith (1995) after Hurricane Andrew, Rash (2006) after Hurricane Ivan, and the Roofing Industry Committee on Weather Issues (RICOWI) (2006, 2007, and 2009) after Hurricanes Charley and Ivan, Katrina, and Ike, respectively.

4 _____ Figure 1. Peak negative uplift pressure coefficients *Corresponding author address: Timothy P. Marshall, (Cp) on a gable roof from a wind tunnel model. Wind 4041 Bordeaux Circle, Flower Mound, TX 75022- directions were at 10 degree increments. From Liu et 7050. email: al. (2006). Figure 2. Uplift of ASPHALT SHINGLES on the corner of a Figure 4. Idealized diagram showing the increase in gable roof during Hurricane Gustav. Image by Tim the projected area (blue line) as the shingle is uplifted in Marshall. the wind. One mechanism of ASPHALT shingle uplift is specific 3. TYPES OF ASPHALT SHINGLES . to the shingle itself. Peterka et al. (1997) and Jones et al. (1999) have shown that wind-induced uplift of The two most common types of strip ASPHALT SHINGLES SHINGLES can be different from the mechanism that are three-tab and laminated (Fig. 5). Both shingle types causes uplift on impervious roof sheathing. Due to are comprised of glass-fiber or paper mats that are small dimensions of the SHINGLES and their pervious saturated in hot ASPHALT and top coated with granules.

5 Design, pressure variations across the shingle surfaces With three-tab SHINGLES , two joints are cut out of the can be relatively small. Thus, it is the profile of the bottom half of the shingle primarily as a design feature. shingle that determines the extent of wind uplift. A Laminated SHINGLES have a top laminate and a bottom critical part of the shingle is that region cantilevered laminate. The lower portion of the top laminate has downslope from the sealant strip (Fig. 3). Wind trapezoidal-shaped cut-outs whereas the bottom stagnates at the base of the shingle while aerodynamic laminate is a solid shingle. The bottom laminate is only uplift occurs on top of the shingle, a situation similar to half the width of the top laminate and is adhered to the what occurs on an aircraft wing. The resulting lift can lower portion on the back side of the top laminate. The deform the shingle, producing greater projected area design of the laminated shingle gives the appearance of and therefore, more lift (Fig.)

6 4). Noone and Blanchard wood SHINGLES . Both shingle types have sealant strips. (1993) indicated: 1) the quality of fastener installation, 2) the strength of the sealant, and 3) the physical properties of the ASPHALT shingle ( fastener pull-out strength and stiffness) are critical factors in resisting wind uplift damage. Figure 3. Idealized wind flow (blue lines) and resultant Figure 5. The two most common types of ASPHALT positive (plus signs) and negative pressures (minus SHINGLES are the three-tab and the laminated. At least signs) at the bottom edge of a shingle. From Peterka et four fasteners must be installed per shingle (standard al. (1997). application) as noted by locations of the red dots. Other types of ASPHALT SHINGLES include solid strip TABLE 1. (without the cut-outs or laminates) and large format. PERFORMANCE OF ASPHALT SHINGLES . Less common types are the T-lock, hexagonal, and IN HURRICANE FRANCES. three or more laminates.

7 The interlocking SHINGLES typically do not have sealant strips. SHINGLE Little 10-50% >50% Total TYPE to no Damage Damage number SHINGLES must be fastened properly to the roof in Damage of roofs order to achieve the maximum wind resistance. In non- Three-tab hurricane regions, four fasteners are required to secure 57 (44%) 11 (9%) 60 (47%) 129. three-tab and laminated SHINGLES . However, in Laminated 132 (87%) 6 (4%) 13 (9%) 151. hurricane prone regions, some building codes require six fasteners per shingle. The vast majority of shingle manufacturers require that fasteners be installed below Failures of the three-tab SHINGLES typically occurred the sealant strip and above the cut-outs, and within one when the tabs were not bonded to the sealant strips and inch of the ends of the SHINGLES . Fasteners must extend were lifted, breaking at the tops of the tabs. Failure was through the shingle as well as the underlying shingle. more likely in those tabs which bridged the butted end Fasteners also must be long enough to penetrate joints of underlying SHINGLES .

8 The result was a "zipper". plywood and oriented strand board (OSB) roof decking. like pattern of missing tabs that extended upslope from It is critical that fasteners be driven flush with the the eaves to the ridges (Fig. 6). shingle surfaces and not over- or underdriven. 4. HURRICANE FRANCES STUDY. According to the Beven (2004) at the National Hurricane Center, the eye of Hurricane Frances made landfall on Hutchinson Island, FL as a Category 2 storm on the Saffir-Simpson scale around 0430 UTC on 5. September 2004 and traveled slowly west-northwest across the Florida peninsula. The Army Corps of Engineers (USACE) station at Port Mayaca reported sustained winds of 38 m s-1 (85 mph) at 0500 UTC, while a portable instrumented tower operated by the Figure 6. Wind damage to ASPHALT SHINGLES installed in Florida Coastal Monitoring Program (FCMP) at Ft. the racking or straight-up pattern after Hurricane Pierce reported 36 m s-1 (81 mph) sustained winds at Frances.

9 0402 UTC along with a peak gust of 48 m s-1 (107. mph). It was found that three-tab ASPHALT SHINGLES did not Shortly after the hurricane, the lead author perform as well as laminated SHINGLES during the conducted a survey of roof damage along the eastern hurricane. Only 44 percent of roofs with three-tab Florida coast. The purpose of the survey was to SHINGLES had little to no damage compared to 87 percent evaluate the performance of common roofing systems in of roofs with laminated type ASPHALT SHINGLES . This was strong winds. The survey was performed along the due in part to lighter weights of the tabs and their coastline via Highway A1A from Cocoa Beach to design which acted like a flap on an aircraft wing. Jupiter Beach. Three coastal subdivisions were selected Once the sealant bond was broken, the tab would fold for examination of roof performance. These back and crease along the top of the tab. Multiple subdivisions were located in Vero Beach, Stuart, and flipping of the tab in the wind eventually caused some Ft.

10 Pierce. tabs to tear and blow away. The degrees of shingle damage were categorized A total of 11asphalt shingle roofs were selected for as: (1) little to no damage, (2) between 10 and 50 detailed examination. On 10 of the 11 roofs, the percent damage, and (3) greater than 50 percent SHINGLES had been fastened improperly with nails or damage. A total of 280 ASPHALT roof coverings were staples installed in or above the sealant strips, thereby evaluated. Of these, 129 roofs had three-tab SHINGLES missing the underlying SHINGLES . In some instances, and the remaining 151 roofs had laminated type fasteners interrupted the transfer of the sealant to the SHINGLES . Many three-tab shingle roofs were installed in overlying shingle. The numbers of fasteners ranged the straight-up or racking patterns whereas laminated from three to six per shingle. SHINGLES were installed in diagonal patterns. Table 1. summarizes the results of the survey: Four primary modes of ASPHALT shingle failures Dupuis and Graham (2002) conducted laboratory occurred during Hurricane Frances: 1) creasing, 2) tests on various three-tab SHINGLES and found that heat flipping, 3) tearing/removal, and 4) abrading from and moisture affect tear strength, leading to premature flying or falling debris (Fig.