Safety Alert: Study Reveals Sharp Increase in Deck Failures A landmark study reveals that there have been 179 reported deck collapses from January2000 through December 2006, killing 33 and injuring 1,122.By: Michael Morse, Brittney Corwin, Robert Morse and Andrew JohnsonIn September 2006, a particularly violent deck collapse occurred. What made this collapse so frightening was it occurred with surprisingly little impetus. The potential purchasers, a family of five, were doing a final walkthrough on a single-family house in a quiet, older subdivision in Lawrenceville, Georgia. The house was situated on a lot that sloped away from the street. The back sliding door was about 12’ above grade. There was a well built and well maintained 12’ by 15’ deck overlooking a wooded backyard. When the husband, wife, and two of their teenage children stepped out onto the deck, it pulled off the house… but it did not just fall down. As the deck pulled away from the house, the deck swung underneath, struck the outside support columns, and landed upside down. The family fell to the ground and the deck, now upside down, landed on top of them. They were trapped underneath the deck, injured and traumatized. Emergency services were called and the family was transported to the hospital. Fortunately, the deck’s guardrail acted like a roll bar and held the deck off the ground, preventing the victims from being crushed.Why did this happen? How could a solidly built structure just detach and collapse?
While the deck itself was structurally sound, the connection of the deck to the house was not. When the family walked onto the deck and then stopped, their momentum was transferred to the fasteners that held the deck to the house. That little bit of movement was just enough to overcome the friction holding the fasteners to the house rim joist. The deck simply pulled away from the house.Since the deck was built to the standards prescribed by national building code, it was supposed to be capable of accommodating the weight of 48 people, but it only took four people to bring it down. Why?Scope of the problemThere is no reliable source for statistics on how many decks there are in the United States, how many decks are being built annually, or by whom. Several indirect approaches were used to generate reasonable estimates relating to deck construction and the number of decks there are in the United States. Information was collected referencing housing starts, home design trends, and the do-it-yourself market.Home BuildersGrowth of the decking industry is partly driven by sales of new homes. The National Association of Home Builders (NAHB) estimated housing starts in 2006 to be nearly 2 million units houses being built today new home options in 2006. This number does not include new decks installed on older homes, or renovations of older decks.Deck BuildersThe North American Deck and Rail Association (NADRA) estimates the annual retail installed value of deck components and accessories in the United States to have been between $9 and $10 billion in 2005 coming years. In fact, NADRA was started in response to this terrific growth. Based on a 2005 survey performed by national retailers, the average cost of a deck is approximately $11,300 decks were constructed in 2005 alone.HomeownersDeck construction appears to be simple and straight forward, and many homeowners undertake the project themselves. With readily available calculators, guides, and premade construction plans, it is easy to see why deck building has become one of the most common ‘do-it-yourself’ projects. Although this group makes a significant impact on the decking market, there was no apparent way to quantify its contribution.House Design InfluenceIn 2005, more than 60% of all new homes either came with a deck, or incorporated the opportunity for future outdoor living space delay the construction of a deck, the layout of the house includes, and anticipates, this future addition.While some homeowners may choose to Subdivisions, such as this town-house community, can be found across the United States. It is clear that these homes were designed to include a structure outside of the rear door. The rim joist located just below this door was intended to be the point of attachment for the future deck. Building CodesThe International Residential Code (IRC) is one of the primary references for both deck builders and code officials. It contains instructions on how to build reliable and safe structures. General requirements for all structures, including decks, are in Chapter 3, Section R301 Design Criteria. This section states that: “Buildings and structures, and all parts thereof, shall be constructed to safely support all loads, including dead loads, live loads, roof loads, flood loads, snow loads, wind loads and seismic loads as prescribed by this code. The construction of building and structures shall result in a system that provides a complete load path capable of transferring all loads from their point of origin through the load-resisting elements to the foundation.”Specifications concerning decks are found in the chapter on floors (Chapter 5) of the IRC. One section that is written specifically for decks (R502.2.2 Decks), provides a very brief and non-prescriptive recap on how decks must be attached and supported. “Where supported by attachment to an exterior wall, decks shall be positively anchored to the primary structure and designed for both vertical and lateral loads as applicable. Such attachment shall not be accomplished by the use of toenails or nails subject to withdrawal. Where positive connection to the primary building structure cannot be verified during inspection, decks shall be self-supporting.”All other guidance must be derived from sections that do not specifically cite deck construction. They are written with house building, not deck building, in mind.The Study of Reported Deck CollapsesThis study seeks to better define the scope of deck failures in the United States by providing statistical evidence of the problem. This report includes deck, porch, and associated collapses that were reported from January 2000 through the end of 2006. Although a few Internet sites referenced deck failures, no central source of data was found.MethodologyUntil now, conclusions drawn on deck collapse were based on a very limited sample size. This report hopes to establish a database with a statistically significant sample size, and to then identify trends, characteristics, or weaknesses. In order to draw accurate conclusions on deck failure, there must be adequate data to analyze.Deck collapse or deck failure, for the purpose of this report, is defined as a single negative structural event that renders a deck non-functional. If a deck either detaches or shifts away from the primary structure, it would be included in the study.An incident report form was developed, completed, and archived for every collapse. Data sought for each event includes: the height and size of the deck, construction materials, occupancy and activity at time of collapse, and cause of failure. These data points were then analyzed to identify trends or patterns.The data used for this report was gathered through comprehensive searches of Internet and periodical archives using key phrases including deck collapse, deck injury, rail collapse, etc. Great care was taken to include all legitimate events; that is, events based on construction technique rather than an unrelated accident.Source of InformationThe primary source of information detailing deck collapses is the news media. News reports focus on injuries sustained rather than the actual cause of the event. Reporters reflect the statements of emergency responders or eye witnesses, neither of which are focused on the physics or engineering of the deck structure. The cause that was initially reported may be different than what a subsequent investigation would find. This lack of complete information can lead to flawed assumptions as to the cause of deck failure.Correcting the design defects that cause deck failure is impossible if solutions are based on inaccurate information. Deck collapses are reported as isolated events. Very little background information is provided on the scope of this problem. Subsequent reporting could include the reasons why decks go down in the United States. There have been news segments on building a better deck; however, there also should be reports on the actual causes of failures and segments on preventing deck collapse.Data, Trends, and Analysis DataFrom January 2000 through December 2006, there were 179 reports of deck and railing failure. In these events 1,938 people were exposed to injury; they were either on or under the deck when the failure occurred. Of those involved, 1,122 sustained injuries, and 33 people died. This translates into 58% of the people involved in reported deck and railing failure were injured or killed.Virtually no municipalities perform an investigation that documents the cause of the deck collapse. A smaller sample was used to investigate this trend. Out of all the collapses included in the subcategory, only one such report was found.TrendsAn examination of the data on reported deck collapses reveals several interesting trends. Deck collapses are increasing at an average rate of 21% per year. Twice the number of deck collapses occur as compared to the rest of the year.There is a well-defined deck collapse season (June through August) in which over. Virtually all reported deck collapses occur while the deck is occupied. AnalysisAn analysis of the data and trends for deck collapse has led to the following observations.DECK CONSTRUCTIONReported deck collapse. This type of failure includes the separation of the rim joist from the house floor joists, the separation of the ledger board from the rim joist, and the separation of the ledger board from deck joists.Failure of the house to deck ledger connection accounted for over 90% of all deck collapses. Decks are built to the same codes and standards that houses are, yet decks are more prone to collapse. Current deck connections and/or the deck components are subject to failure long before the end of the service life of the deck. Follow up investigative reports on the cause of collapses are rarely generated.Ninety-five percent of reported collapses occurred when the decks were occupied.ConclusionAre there right ways (or, more importantly, wrong ways) to build decks? Is anyone watching?An analysis of deck collapse data indicates that deck connections are subject to failure long before the end of the service life of the other components of the deck. The lack of structural redundancy, especially at the critical connection points, leads to deck collapse.Specifically, deck collapse is related to the connection assembly of the deck ledger board to the house substructure. In the scenario where the house was built to accommodate a deck, it is assumed that the floor joist system of the house was constructed to support a deck. This intention was known by the architect, the builder, and the homeowner. Was it known by the framing contractor? A better question is: did the framer prepare the point of attachment (the rim joist) for the loads associated with a deck in use? Can this rim joist resist the pull out force exerted by a deck, no matter how large?When a deck collapses, people are injured, or worse. A design flaw in deck construction may lead to these catastrophic events. One can disagree with the process used to estimate the number of decks being built, but the underlying fact is that the growing number of decks in existence will directly translate into a proportionally larger number of collapses, which presents a significant public health risk.When a deck is bolted to a house, the strength and durability of this attachment depends primarily on the ability of the rim joist of the house to transfer the load to the house foundation. Unfortunately, the rim joist was not designed to resist the pullout action imposed by a deck. Current building codes provide details for the installation of rim boards. These details are the same whether or not a deck is to be attached. Building codes should require additional anchoring for the rim joist when constructing decks and/or houses onto which decks will most likely be attached.As with any new development, there is a learning curve to discover the long-term performance of a product. The shortcomings of current deck construction must be studied, understood, and addressed. A factor that complicates this process is the incredible number of new decking products and techniques that are constantly being introduced. There is no time to slow down and evaluate the effect that one individual product or new technique has on the overall structural performance. With new products and new techniques being introduced so quickly, there is no easy way to generate a base line of deck performance against which to evaluate change.The popularity of outdoor living space is growing despite the increasing number of deck collapses. There is a perception that each collapse is an isolated event that is dependant on the quality of the deck builder, as opposed to part of a larger trend predicated on a design or structural flaw of the deck’s critical connections.The public is not aware of the increasing danger associated with deck failure, largely because of a lack of conclusive data. A central database is needed to collect and analyze reports on deck collapses, the cause of the collapse, the number and severity of injuries, and the associated costs. Only after this information is assembled can the effect on public safety be evaluated and addressed.A central database is now being created to allow for the archiving of deck collapse events. The purpose of this new database is to provide information for the further study of deck collapses.A Bad WinterThe winter of 1996 saw an unusual amount of snowfall in North Dakota. As it continued to snow in the Fargo area, a repetitive phenomenon occurred: deck after deck collapsed under the weight of the accumulating snow.The number of collapses intrigued a local deck builder. Mr. Todd Funfar, President of Deck Masters, began keeping a photographic log of deck failures. He cataloged over eighty separate collapses that occurred during the winter of 1996.A review of his photographs leads to the following observations:.The deck detached from the house foundation, either from the house band board or from the outside support beams, and then dropped. This suggests that the deck joist system was capable of carrying loads greater than the capacity of the ledger connection and/or outside load beam connection.Rather than the deck floor joists breaking mid-span, decks simply detached from house.The loads increased very slowly and over an extended period of time. As it snowed, the weight grew greater, exerting a sustained load on the decks throughout the winter months. Normally, loads on a deck are applied quickly and for a much shorter period.Due to adverse weather conditions, this series of collapses occurred without the condition. Heat from the house would induce snowmelt adjacent to the ledger board. .The most common point of failure was the connection of the ledger board to the house. While the mechanics of ledger failure varied, the result was the same; the assembly that links the deck and the house floor joists failed, causing the deck to collapse. With the Fargo collapses, the conclusion that can be drawn is that the connection of the deck to the house was not adequate to transfer the loads from the deck, through the boards and hardware, to the foundation of the house.As an aside, while researching reports of deck collapses (covering hundreds and hundreds of hours in national archives and internet search engines) not one of the collapses from the North Dakota winter was discovered.
The deck connections and deck components were in a prolonged wet service.