Steep-Slope Roofs in the Wind David Roodvoets DLR Consultants | Montague, MI 1dlrconsul@charter.net IIBEC International Convention & Trade Show | SeptembeBEr 15-20, 2021 R Roodvoets | 141 142 | Roodvoets II BEC International Convention & Trade Show | September 15-20, 2021 ABSTRACT Do you work on projects that may involve steep roofs? The Roofing Industry Committee on Weather Issues (RICOWI) has inspected hundreds of steep roofs as part of its Wind Investigation Program (WIP). These posthurricane investigations have led to a better understanding of the strengths and weaknesses of the wind performance of metal panel, tile, and asphalt shingle roofs. This presentation will discuss the systems and their strengths, and describe where consultants and inspectors can focus to ensure a wind-resistant steep roof. Since the inception of the RICOWI WIP in 1996, there has been wind-tunnel, full-scale, and laboratory-designed wind research and testing. This research has resulted in better products and systems, but the fury and duration of hurricanes test every aspect of design and installation. Using photos from six WIP investigations, the presenter will show participants how systems have improved, and where more attention is needed. This information can be used to design roofs for all climates and wind zones. David Roodvoets DLR Consultants | Montague, MI David Roodvoets has had a leadership role in all six RICOWI hurricane investigations. He has worked with Asphalt Roofing Manufacturers Association on asphalt shingle wind resistance and test standards, and has been involved with poststorm wind damage investigations for 40 years. Roodvoets has worked with wind research engineers in testing of asphalt shingles, vegetative systems, photovoltaic systems, and single-ply and ballasted roof systems. As a result of this work, manufacturers have incorporated improved design and installation requirements, and changes have been developed, advocated for, and incorporated into the International Code Council’s family of codes. SPEAKER Buildings have roofs to separate the interior of the building from the exterior; their main function is to keep water out of the building. So, why aren’t all roofs low-slope? Can you prove that a roof with a 6/12 pitch will be better at keeping the water out than a roof with a 2/12 pitch? Is a 6/12 pitch roof less expensive than a 2/12 pitch roof? Thought not. But there are more steep-slope roofs in North America than low-slope roofs. The reason is current aesthetics. The slope of the roof really depends on when and where the building was built. Before World War II, most houses had roof pitches of 6/12 or greater. Starting in the 1950s, 3/12 slopes became popular, and since the 1990s, 10/12 and steeper roofs have been popular. Slope is unlikely to play a role in the wind performance of steep-slope roofs. Field data from the six hurricane studies conducted by the Roofing Industry Committee on Weather Issues (RICOWI)1–5 are inconclusive. Studies conducted in the 1990s by Cermak Peterka Petersen Inc. for the Asphalt Roofing Manufacturers Association indicated that the wind speed at 5 mm off the top of the shingles was somewhat faster over shingles on a 9/12 pitch roof than on a 3/12 pitch roof.6 This produced slightly higher pressure differentials but was not considered to be a significant factor in shingle damage. Because there are so many variables, it is not possible to sort out slope in the RICOWI Wind Investigation Program data. Wind tunnel research on steep-slope roofs demonstrates that the uplift pressures vary greatly over the roof and depend significantly on wind direction and, to a lesser extent, slope. Peterka et al.7 found that the highest uplift was when the slope was 2/12, occurring near the peak of the rake edge of gabled roofs. The peak uplift is at the ridge and just on the leeward side of the ridge. The RICOWI inspections clearly support the wind tunnel data, with ridge cap or similar damage being the most commonly witnessed. After the ridge, the corners have higher uplift pressures and are most vulnerable on all types of steep-slope roofs. However, there are clear factors that can be used to differentiate the performances of the systems. This paper reviews the changes in installations and products with the expectation that the remaining weak links for wind resistance can be reduced or eliminated. STATE OF THE ART IN 2004 RICOWI was formed in 1990 after Hurricane Hugo and wind design conferences held at Oak Ridge National Laboratory (ORNL) in 1989. An immediate goal was to get more industry coordination on building code issues. Following Hurricane Andrew in 1992, many industry members believed the information on roofing performance was incorrect. This created a sense of urgency to provide a mechanism to inspect roofing systems after hurricane events and responsibly report the facts. Consequently, many talented individuals in the industry and academia developed a training course that was delivered to nearly 150 participants at ORNL in 1996. For the first time, a large number of roof consultants and manufacturers’ technical personnel had an understanding of the fundamentals of wind engineering and a standardized way to document storm damage. No major hurricanes came inland in a populated area of the United States until 2004, when four major hurricanes made landfall in Florida. All of them met the RICOWI requirements for deployment; however, due to the time constraints of volunteers and financial resources of RICOWI, assessment teams were only deployed after hurricanes Charley and Ivan. What did they see and learn? TILE ROOFS IN 2004 Hurricane Charley hit the Punta Gorda area of Florida with 140 mph winds. Tile roofs are popular in this area, and many single-family homes there were built before 1997 with tile roofs installed using mortar paste as the adhesive to hold the tiles to the deck (Fig. 1–3). For tiles to II ii B E C International Convention & Trade Show | September 15-20, 2021 R Roodvoets | 143 Steep-Slope Roofs in the Wind Figure 3. Hurricane Charley: 130 mph winds. Ridge damage on this roof. Figure 1. Hurricane Charley: Likely 140 mph winds. Typical damage in this older subdivision. Tiles using mortar as the adhesive. Figure 2. Hurricane Charley: 130 mph and greater winds. Typical ridge damage with downslope breakage due to tile debris. Mortar adhesive. adhere properly, 27 things that had to be done correctly or the system would become vulnerable to wind damage. Hurricane Charley’s 140 mph winds were more than adequate to take advantage of these vulnerabilities, resulting in serious damage to all types of roofs and buildings including many of the roofs with mortar adhesive. An alternative to mortar adhesive is the use of nails or screws to secure the tile. This did not appear to be a common practice, however, and tiles that were installed with only one nail were often damaged by Charley’s 140 mph winds. Foam adhesion of tile had been developed in response to the poor performance of mortar adhesive found after Hurricane Andrew. In 2004, the practice of adhering tiles with expandable foam was still a relatively new process (Fig. 4). It requires fewer steps than the mortaring process; the main requirements are the correct amount and placement of foam under each tile. Most of the roofs that had been installed with the foam process were mostly intact. The weak link was inadequate foam on the ridge and hip tiles. Hip and ridge tiles were vulnerable when the foam had not risen sufficiently to adhere. When the ridge tiles were dislodged, they often tumbled down the slope, breaking tiles and creating missiles. FEMA 4888 provides more details on the impacts on tile roofs from these hurricanes. Mirmiram et al.9,10 provide more details on tile installation and testing. Numerous other FEMA reports look at the whole building and are available at www.fema.gov/multimedia-library. In 2004, the weak link for tile roofs was the use of mortar adhesive, which has now been mostly abandoned in the United States. Other problem areas were inadequate nailing and insufficient foam adhesive. TILE ROOFS IN 2017 AND 2018 In 2017, RICOWI investigations after Hurricane Irma in Florida found tile roofs in Naples and the Keys with no observed catastrophic losses. There were many roofs with dislodged ridge and hip tiles, which almost always resulted in damage to other tiles on the roof (Fig. 5–6). In most cases the damage was minor; however, some roofs were significantly damaged, and investigators observed missile impacts in residential areas. Missile damage is confirmed by random broken tiles with intact roof ridges; the source of the missiles is not always evident. It did not appear that any of the inspected tile roofs were leaking. One large tile roof had been installed with two nails per tile; this roof had significant ridge damage and subsequent downslope breakage. It appeared that nails used to secure the ridge were not long enough to grip the deck (Fig. 7). Hurricane Michael inspections focused on Panama City, Fla., where there were fewer tile roofs. Those that were inspected were in the 130 to 140 mph wind zone. They had lost a few ridge tiles, again due to inadequate foam adhesive. The remaining weak link in tile installation is using enough adhesive, primarily on the ridge tile. Because of the underlayment used, roofs with damaged tiles were not found to be leaking. METAL-PANEL ROOFS State of the Art in 2004 The 2004 RICOWI wind investigations viewed buildings that ranged from more than 30 years old to recently built. In general, there was a significant amount of progress in strengthening the structures, with newer buildings having improved performance. This was due to the use of computer-aided design to build most commercial metal buildings and significant wind studies that identified many of the weak links in previous designs. One building in downtown Punta Gorda had no damage, which stood out as everything around it was severely damaged. 144 | Roodvoets IIiiBEC International Convention & Trade Show | September 15-20, 2021 Figure 4. Hurricane Charley: Likely 130 to 140 mph winds. New units, foam adhesive. Very minor damage. Figure 7. Hurricane Irma: Winds 100 to 110 mph. Inadequate adhesion. Figure 5. Hurricane Irma: 110 to 120 mph winds. Typical tile damage. Foam adhesive; damage limited to tiles missing from ridge. Figure 6. Hurricane Irma: 100 to 110 mph winds. Tiles installed with two nails; displaced ridge tiles damaged downslope tiles. State of the Art in 2018 Metal buildings and metal-panel roofs may not show environmental damage due to age for many years. A building that was built many years ago can appear to be relatively new unless inspected carefully. Then corroded fasteners may be the first thing noticed. Corroded fasteners have been an issue in the past and continue to be a weak link in buildings that are not carefully constructed or buildings for which the manufacturer’s requirements were not followed. Electrogalvanized fasteners do not have the longevity required for most modern environments. Hot-dipped galvanized fasteners are a minimum requirement and corrosion-resistant fasteners may be required. Fastener corrosion appears to be a more significant issue on roofs that are near saltwater coasts and remains one of the weak links that limit the service life of metal buildings. Metal-panel roofs are often chosen to recover existing systems and often the old roof remains. Inspectors have frequently found new metal panels that were not solidly fastened to the structure; instead, they were fastened to old metal panels or poorly fastened to substitute structures (Fig. 8, 9) A fastening deficiency is usually not found until there is a highwind event. Attachment details need to be carefully considered when retrofitting any metalpanel roof over an existing system. Fasteners for roof re-cover need to be attached to the purlins for best wind performance. The breach of a wall or window can increase internal pressure significantly. This breach can result in loads that exceed the design loads of the system. Without internal pressurization, metal-panel systems most often survive above-code winds. However, when the weak point is on a windward face, the metal roof can fail in peel, similarly to the failure of membrane roofs. Lack of solid perimeter securement was found when a peel failure was observed (Fig. 9–10). II BEC International Convention & Trade Show | September 15-20, 2021 R oodvoets | 145 Corroded fasteners have been an issue in the past and continue to be a weak link in buildings that are not carefully constructed or buildings for which the manufacturer’s requirements were not followed. Figure 8. Hurricane Irma: 100 to 100 mph winds. Missile impact damage. Figure 9. Hurricane Michael: 140 mph winds. This metal-panel roof was blown off. One of the weak links was the edge fasteners totally missing the edge nailer. This roof also had significant internal pressurization from failed soffits. Metals roofs are sometimes punctured by debris. Debris can open seams, and falling trees and other objects have damaged metal-panel roofs. Damage from debris appears to not be a major factor in metal roofs. The weak link found in recent inspections was lack of attachment of retrofit roofs and corroded fasteners (Fig. 12). The best structures can still fail when the walls are breached, and unfortunately this can be the weak link. A metal roof manufacturer gives the following advice: • Lower sloped rooflines reduce wind shear. • Consider limiting the roof pitch to less than 6/12. • Hip roofs deflect wind better than gable roofs. • The deeper the roof overhang of a building, the greater the lifting force of strong winds on that structure. • The most wind-resistant buildings have no overhang at the eave at all, as in most prefabricated steel buildings. ASPHALT-SHINGLED ROOFS State of the Art in 2004 Florida was a unique place for roofing installations in 2004. Whereas most asphalt-shingled roofs in the United States are installed by going across the roof, in Florida, an installation method called racking was most often used. In this method, the roofer installs the shingles by going vertically up the roof slope, usually three shingles across. This requires the shingle furthest to the left to be lifted to attach the final (fourth or sixth) nail. This frequently results in having no nail or a permanent bend of the shingle, preventing the shingle tab from adhering. In 2004, many asphalt-shingle suppliers tested their shingles at wind speeds up to 110 mph according to ASTM D3161 Class F.11 However, most shingles installed before 2004 had likely been tested according to ASTM D3161 Class A (60 mph). Although high wind speed testing was not required by code, many roofs had limited damage. The code also called for six nails in high-wind areas. Research conducted in the 1990s by Peterka et al.7 helped the industry understand the value of shingle stiffness as well as the importance of sealant strips. Peterka’s research demonstrated the importance of having strong adhesive strips, and that nailing, while absolutely required, was not the primary factor in shingle wind-resistance. In 2002, the industry began developing ASTM D7158,12 which provides for classification of shingles to Class H or 150 mph and measures shingle stiffness and sealant-strip resistance to wind uplift load. The 2006 International Residential Code13 required shingles installed in high-wind areas to be Class F. The 2009 edition14 required 146 | Roodvoets IIiiBEC International Convention & Trade Show | September 15-20, 2021 Figure 11. Hurricane Michael: 140 mph winds. The brick wall was inadequately supported and collapsed, allowing this building to be overpressurized. The roof damage stopped at a floor-to-roof separator. Figure 10. Metal panels where the perimeter was not fastened to the edge nailer. Figure 12. Hurricane Michael: 140 mph winds. The roof on this building, located less than 2 miles from the building shown in Fig. 10, appears to be 100% intact. asphalt shingles to be ASTM D7158 Class H if installed in wind zones of 120 mph or greater. Nontabbed shingles were required to meet ASTM D3161 Class F in these wind zones. State of the Art in 2018 In Florida, roofers frequently installed asphalt shingles using the racking method. As previously noted, this system is vulnerable to high-wind damage (Fig. 13–17). Assessment teams observed 132 asphalt-shingled roofs after Hurricane Michael, which was typical of the area where winds reached 130 to 150 mph. About 90% of the roofs in this area were at least partially covered with blue tarps and were expected to leak if uncovered. Data indicated that about 75% of the observed roofs were at least 12 years old and therefore unlikely to be in accordance with current codes. Most newer roofs, that is, those installed since 2015, were performing well. The ridge remained one of the weak links in the system. Missing ridge vents were often a source of potential leakage. The lack of sealant on the initial row of shingles was one of the damage initiators (Fig. 15). When the edge is not secured, the shingles can peel up the roof slope, and is more likely to happen when the shingles are installed using the racking method. Synthetic roof underlayment was common; however, in the worse cases of shingle blow-off, the underlayment was also dislodged. Adequate nailing is important for all nonadhered underlayments. The Insurance Institute for Business & Home Safety’s (IBHS’s) FORTIFIED program recommends three approaches to underlayment that are likely to prevent a leaky roof if the shingles are blown off.15 The simplest approach is to tape all the seams of the roof deck using plywood or oriented strand board as the primary water barrier. The second option is the use of two layers of ASTM D226 #30 felt installed with multiple secure nail tabs. The third option is the use of a membrane that is fully adhered to the entire wood deck; however, not all roofing manufactures approve of this method. The remaining weak links for asphalt shingles are proper installation of the hip and ridge shingles and ridge ventilation. Adequate resistance to the uplift at these points requires nails that are long enough to penetrate the deck. High-wind areas require additional adhesive in the starter row and at the rake edges. OTHER ROOFING SYSTEMS Two other steep-slope roof types were observed: metal shingles and cedar shakes and shingles. Metal shingles, being relatively new, had little corrosion, but often lacked attachment to the substrate. The cedar roofs had corroded fasteners. II ii B E C International Convention & Trade Show | September 15-20, 2021 R Roodvoets | 147 Figure 15. Hurricane Michael. Corner of roof with damage. Excessive overhang of starter course. First course not sealed. Figure 13. Hurricane Michael: 140 mph winds. This open metal building had cladding and roof loss. Many of the connections were corroded. Figure 14. Hurricane Michael. Racking installation. CONCLUSION RICOWI, the Special Interest Group on Dynamic Evaluation of Roofing Systems, and IBHS have contributed substantial data on the performance of roofing systems in high winds. Hurricanes and documented roof inspections after windstorms have confirmed wind tunnel data, contributing to a substantially better understanding of steep-slope roof performance in high winds, and these data have been used by manufacturers to develop better products. Codes have required better testing, and contractors continue to be better trained. This has resulted in more wind-resistant roofs, not only for hurricane-prone regions but everywhere roofs are installed. REFERENCES 1. Roofing Industry Committee on Weather Issues (RICOWI). 2006. “Hurricanes Charley and Ivan Investigation Report.” https://www.ricowi.com/reports/REPORT-2006-HRCN-CHARLEY-DL. 2. RICOWI. 2007. “Hurricane Katrina Investigation Report.” https://www.ricowi.com/reports/REPORT-2007-HRCN-KATRINA-DL. 3. RICOWI. 2010. “Hurricane Ike Investigation Report.” https://www.ricowi.com/reports/REPORT-2010-HRCN-IKE-DL. 4. RICOWI.2018. “Hurricane Irma.” https://www.ricowi.com/reports/REPORT-2018-HRCN-IRMA-DL. 5. RICOWI. 2019. “Hurricane Michael.” https://www.ricowi.com/reports/REPORT-2019-HRCN-MICHAEL-DL. 6. Jones, J., R. Metz, and C. W. Harper. 1997. “ARMA Wind Uplift Load Model for Assessing Asphalt Shingle Performance.” In Fourth International Symposium on Roofing Technology Proceedings, September 17–19, 1997, Gaithersburg, MD. Rosemont, IL: National Roofing Contractors Association. http://docserver.nrca.net/pdfs/technical/5832.pdf. 7. Peterka, J. A., J. E. Cermak, L. S. Cochran, B. C. Cochran, N. Hosoya, R. G. Derickson, C. Harper, J. Jones, and B. Metz. 1997. “Wind Uplift Model for Asphalt Shingles.” Journal of Architectural Engineering 3 (4): 147–155. 8. Federal Emergency Management Agency (FEMA) Mitigation Assessment Team. 2005. Hurricane Charley in Florida: Observations, Recommendations, and Technical Guidance. FEMA 488. Hyattsville, MD: FEMA. 9. Mirmiran, A., T-L Wang, C. Abishdid, P. Huang, D. L. Jiménez, and C. Younes. 2006. “Performance of Tile Roofs Under Hurricane Impact − Phase 1.” Department of Civil and Environmental Engineering, Florida International University. 10. Mirmiran, A., T-L Wang, C. Abishdid, P. Huang, D. L. Jiménez, and C. Younes. 2007. “Performance of Tile Roofs Under Hurricane Impact − Phase 2.” Department of Civil and Environmental Engineering, Florida International University. 11. ASTM Subcommittee D08.02. 2020. Standard Test Method for Wind Resistance of Steep Slope Roofing Products (Fan-Induced Method). ASTM D3161. West Conshohocken, PA: ASTM International. 12. ASTM Subcommittee D08.02. 2020. Standard Test Method for Wind Resistance of Asphalt Shingles (Uplift Force/Uplift Resistance Method). ASTM D7158. West Conshohocken, PA: ASTM International. 13. International Code Council (ICC). 2006. 2006 International Residential Code for One- and Two-Family Dwellings. Country Club Hills, IL: ICC. 14. International Code Council (ICC). 2009. 2009 International Residential Code for One- and Two-Family Dwellings. Country Club Hills, IL: ICC. 15. Insurance Institute for Business & Home Safety. “Fortified Programs.” https://ibhs.org/fortified/. ADDITIONAL RESOURCES • Shoemaker, W. L. 2007. “The Case for the Engineer of Record for a Metal Building System” Structure Magazine March 2007: 9–12. • FEMA. 2010. Home Builders Guide to Coastal Construction. Technical Fact Sheet G.1. Hyattsville, MD: FEMA. https://www.fema.gov/sites/default/files/2020-08/fema499_2010_edition.pdf. • New Jersey Institute of Technology. 2007. Home Shapes And Roofs That Hold Up Best In Hurricanes.” ScienceDaily 21. https://www.sciencedaily.com/releases/2007/06/070619155735.htm. 148 | Roodvoets IIiiBEC International Convention & Trade Show | September 15-20, 2021 Figure 16. Hurricane Michael. Relatively new roof 2018 with some ridge and corner damage, older roofs in background with damage. Figure 17. Typical damage to a roof installed using the racking method.
