When I was still an architecture student, a teacher told me, "We learn more from buildings that fall down than from buildings that stand up." What he meant was that construction is as much the result of experience as of theory. Although structural design follows established formulas, the actual performance of a building is complicated by the passage of time, the behavior of users, the natural elements—and unnatural events. All are difficult to simulate. Only selected building components are tested in furnaces for fire resistance, for example, or analyzed on vibrating platforms for earthquake resistance. Similarly, only mock-ups of facades are pressure-sprayed to test their ability to keep out driving rain. Buildings, unlike cars, can't be crash-tested.
The first investigative report of the World Trade Center collapse, by the Federal Emergency Management Agency and the American Society of Civil Engineers, was released last May. Disappointingly, the six-month inquiry was inconclusive. It did not reveal any design deficiencies, or any "specific structural features that would be regarded as substandard," nor did it make any definite recommendations to safeguard tall buildings. Such recommendations will come. The National Institute of Standards and Technology is beginning a two-year, $16 million study of the twin-tower collapse that will address three pressing questions: whether current testing standards and building codes are adequate to resist catastrophic fires; whether building codes sufficiently take into account what engineers call progressive collapse, the chain reaction that leads to extremely rapid collapse (in the case of the World Trade Center towers, about 10 seconds); and how existing buildings can be made less vulnerable to terrorist attack.
Post-disaster investigations are commonplace. It's not a question of apportioning blame—despite the calls of some politicians and some victims' families—but rather of learning what worked and what didn't. This does not necessarily imply that mistakes were made or that corners were cut. It's normal for building failures to lead to changes in practice. Codes have often been rewritten following disastrous fires. The 1911 Triangle Shirtwaist factory fire, in which 146 women workers lost their lives, led to the creation of New York City's Bureau of Fire Investigation. And just in time. The next year, New York's 130-foot-high Equitable Building, considered fireproof at the time, was so severely damaged that the entire structure had to be demolished. It, too, left an important legacy—the realization that tall buildings had to be designed and built differently. As buildings continued to reach for the clouds, critics became more and more nervous about their susceptibility to fire. And although fires did break out in tall buildings, until the World Trade Center attack, no high-rise blaze had led to the actual structural collapse of an entire building. Even when a fire broke out in 1970 at 1 New York Plaza, a 50-story office tower in Lower Manhattan, it affected only two floors. Like others before it, though, there were lessons to be learned and codes to be changed. That blaze is memorable for disclosing a shortcoming of heat-activated elevator-call buttons. The heat from the burning 33rd floor summoned an elevator, and when the doors opened, the unsuspecting occupants of the cab were suddenly exposed to intense heat and flames. Two men died. Heat-sensitive call buttons were discontinued shortly afterward. In 1988 a large fire destroyed four floors of a 62-story downtown Los Angeles office building, but only one person was killed. Subsequently, the city passed a municipal ordinance requiring sprinklers, pumps, and standpipes in all high-rise buildings over 75 feet. The worst high-rise fire occurred in 1991, when flames raged out of control for more than 19 hours and gutted eight floors of a 38-story Philadelphia office tower. That fire was eventually stopped by automatic sprinklers, which apparently had been installed by one of the tenants on an upper floor. This incident underlined the value of sprinklers, which were subsequently required by the city building code.
The first important lesson of the World Trade Center collapse is that tall buildings can withstand the impact of a large jetliner. The twin towers were supported by 59 perimeter columns on each side. Although about 30 of these columns, extending from four to six floors, were destroyed in each building by the impact, initially both towers remained standing. Planes have hit New York skyscrapers before. In 1945 an Army Air Force B-25 Mitchell bomber strayed in the fog and struck the 78th and 79th floors of the Empire State Building, killing 11 occupants and setting a fire that destroyed two floors. However, there was no overall structural damage, and firemen extinguished the blaze in 40 minutes. The low number of fatalities and relatively slight damage were mostly because the collision occurred on a Saturday morning, when there were few people in the building. The floors above the crash site were largely vacant, so there was little combustible material to fuel the fire. Moreover, the B-25 is a small aircraft, weighing 14 tons fully loaded, and flies at a cruising speed of 230 miles per hour. A Boeing 767 has a maximum takeoff weight of 198 tons, and the one that hit the south tower of the World Trade Center was going about 590 mph. When the twin towers were built, in the late 1960s, they were among the first skyscrapers specifically designed to withstand the impact of an airplane. The Boeing 767 is slightly larger than the Boeing 707, which at 168 tons was the standard for large commercial aircraft then flying, but the difference is well within the margin of error.
The north tower remained standing for 1 hour and 43 minutes and the south tower for almost an hour. Thanks to the robustness of the construction, and what engineers refer to as the redundancy of the structure, most of the twin towers' occupants had enough time to escape. According to the Federal Emergency Management Agency, almost everyone who was below the impact areas was able to leave the buildings safely. This is the second lesson: Tall buildings can be quickly evacuated in an emergency, even from 80 or 90 floors up. The combination of emergency fire stairs and fire-drill training works. However, there is no cause for complacency. The entire daytime population of the World Trade Center complex was approximately 58,000, but because the attacks occurred in the early morning, it was likely the number of occupants in the twin towers was well below that number. According to some estimates, there may have been as few as 14,000 people in the two buildings when the first plane struck at 8:46 a.m. Evacuation started immediately from both towers. Although people were told they could return to the as-yet-unharmed south tower, many wisely disregarded this advice. There may have been as few as 2,000 people in that building when it was struck 17 minutes later. Had the attacks occurred later in the day, or simultaneously, the evacuation would have been more difficult and the number of casualties much greater.
Even so, the death toll was appalling. In addition to 421 firefighters, police officers, and other emergency responders, as well as 157 jetliner crew members and passengers, 2,245 people lost their lives. The majority of the casualties worked in the twin towers—more than 1,400 in the north tower and more than 600 in the south. Approximately 70 percent of these people worked on upper floors, but it has been estimated that 800 people in the north tower and 300 in the south tower survived the initial crashes and were trapped in or above the impact zones. Why couldn't they get out? Each tower had three sets of fire stairs (two 44 inches wide and one 56 inches wide), all clustered together in the service core at the center of the building, which also contained elevators, air-handling shafts, and bathrooms. High-rise buildings have always been designed with centrally located cores, which provide a convenient place for structural support and bracing. The design hides mechanical functions in the least desirable part of the building and leaves the perimeter next to the windows free for human use. The vertical shafts—stairs, ducts, and elevators—tend to act as chimneys during a fire and have to be specially protected. Although the cores of the World Trade Center towers were built of closely spaced, massive steel columns and beams, the fire stairs themselves were encased only by gypsum wallboard attached to metal studs: two 5/8-inch-thick layers of wallboard on the exterior and one on the interior. Such an assembly can withstand fire for two hours, but it offers little resistance to even a hammer blow, never mind the avalanche of debris that assaulted it on September 11. The failure of the fire stairs was almost total. All three sets of stairs in the north tower, and two of three in the south tower, were completely destroyed. Only 18 people in the south tower managed to escape from the floors above the crash zone (tragically, some people used the surviving stairs to climb up, believing that safety lay in the upper floors, away from the fire). It's impossible to know the extent of the destruction on the impacted floors of the towers following the crash, but it's easy to conclude that more robust emergency stairs, of reinforced concrete, spaced far apart rather than clustered together, would have been more effective. It's likely that new codes for the design of fire stairs in tall buildings will result from this experience. Reinforced concrete is much tougher than gypsum, but it is not fireproof. The heat of a fire dehydrates the concrete, and it eventually crumbles. The heat inside the impact zones of the World Trade Center towers was intense. Each jetliner carried an estimated 10,000 gallons of fuel, 2,000 less than half capacity. The fuel produced a giant fireball, but it probably burned off entirely in a matter of minutes—not enough time to weaken the structure but more than enough time to ignite a vast fire across entire office floors on several levels at once. Such a conflagration would have completely overwhelmed the sprinkler system had it been working, but the water supply lines were severed by the crash. The Federal Emergency Management Agency estimates that temperatures reached as high as 2,000 degrees Fahrenheit and that at its peak, the overall fire generated three to five gigawatts of energy. Even if temperatures were lower—in the 1,200°F to 1,300°F range, as some experts believe—the effect of sustained heat on the structure was devastating. Steel begins to soften and bend when temperature differentials are only 300°F. Exposed to a sustained temperature of 1,200°F, steel loses about half its strength. Once the trusses that supported the floors failed, the exterior walls, which depended on the floors for lateral bracing, buckled. This brought the entire weight of the upper portion of the tower above the fire to bear on the floor below, starting a process of progressive collapse, one floor falling on the next.
But isn't steel protected against the heat of fire? Building codes require that a layer of noncombustible material insulate the steel from the fire's heat for a given period, preserving its structural integrity long enough so that the building can be evacuated. That is why fire protection is rated in hours: two hours, three hours, and so on. Until the 1960s, structural steel was encased in poured concrete or brick, whose heavy mass absorbed the heat and dissipated it through dehydration. Because the weight of such fire protection added significantly to the cost of tall buildings, lightweight substitutes were developed, most commonly spray-on coatings of mineral fibers. The structural steel of the World Trade Center towers was originally fire-protected with sprayed-on asbestos, later abated and replaced by a 3/4-inch coating of inorganic fibers. This coating was in the process of being thickened to 11/2 inches (not all the floors in the south tower impact zone had this augmented fire protection).
At the World Trade Center, sprayed-on fire protection was effective in some buildings. A fire was started in 7 World Trade Center, a 47-story high-rise, by debris and heat from the collapsing towers. By then, all underground infrastructure in the area had been destroyed, and there was no water for fire fighting, so the blaze, fed by diesel fuel from generators in an electrical substation, raged out of control. Nevertheless, it was seven hours before the building collapsed, and no lives were lost. On the other hand, the Federal Emergency Management Agency concluded that much of the spray-on fire protection in the twin towers was probably dislodged by the jarring impact of the planes and by flying debris, leaving the steel exposed and vulnerable. There is no doubt that it's time to take a long, hard look at the actual performance of spray-on fire protection. If more effective coatings cannot be developed, perhaps we should return to heavier methods of building.
Before examining whether spray-on coatings need to be replaced by heavier materials, the standardized fire tests used to evaluate fire endurance ratings will have to be reconsidered. "The current testing standards are based on work carried out at the National Institute of Standards and Technology in the 1920s," the institute's director, Arden L. Bement Jr., told a congressional committee. "They do not represent real fire hazards in modern buildings." Fire-resistance ratings measure how individual building components should perform in a fire, but they do not evaluate the performance of the entire structural system, including the crucial connections. While it will never be practical to test high-rise buildings to destruction, more elaborate tests are required. It is likely that predictive computer models will be used to simulate the performance of entire buildings in different types of fires. In the long run, the realization that fire, especially in tall buildings, is a dynamic-design condition, like winds and earthquakes, may be the most significant engineering lesson of the World Trade Center collapse.
A little less than an hour after the first jetliner flew into the north tower of the World Trade Center, a Boeing 757 struck the Pentagon. It's estimated that no more than 140 people died in the Pentagon, whereas more than 1,400 occupants lost their lives in the north tower. The Pentagon is not smaller. It actually houses about 25,000 office workers, compared with an estimated 20,000 in the tower; its area is 6.6 million square feet, compared with the north tower's 4.7 million square feet. The Pentagon, dating from the 1940s, is a bearing-wall structure of reinforced concrete. Its mass was much more successful in absorbing the impact from the aircraft than the lightweight steel structure of the World Trade Center towers. Also, the section hit by the plane had recently been renovated and a number of security measures installed: strengthened walls, blast-resistant windows, a new sprinkler system, fire dampers in the ducts, and accordion fire doors that slammed shut immediately after the crash. These improvements helped to protect the occupants, sped up evacuation, and slowed the spread of fire. Less than a fifth of the building suffered severe structural damage and had to be demolished, another fifth was affected by water and smoke; the rest was undamaged. In sum, the crash of one jetliner killed fewer than 140 people in the Pentagon and put approximately 2.5 million square feet of office space out of service, whereas at the World Trade Center towers, the crash of two jetliners killed more than 2,200 civilians and 421 rescuers and wreaked so much collateral damage that it put about 30 million square feet of office space out of service.
Even accounting for the fact that the plane that hit the Pentagon was flying more slowly—345 mph—the damage in the two cases was entirely disproportionate. Of course, the Pentagon is a building surrounded not by a city but by parking lots. And it is a very low building—five stories instead of 110. All the engineering evaluations of the World Trade Center collapse I have read, including the Federal Emergency Management Agency report, take the extreme height of the World Trade Center towers for granted. But just as their extreme height made the towers symbols—and targets—there is no getting around the fact that it was also their chief liability after they were struck. Once a building reaches a hundred stories, the weight of construction materials becomes a crucial factor, and there is a tendency to make the building as light as possible. If emergency systems—stairs, water pipes, elevators, phone lines—are clustered in a compact core, as they usually are in tall buildings, damage to the core means damage to all the systems. In a tall building, if the fire stairs are destroyed, there is simply no way for people to escape or for firefighters to reach the fire. And once the structure has been severely weakened by the heat of fire, the entire building—not merely a portion of it—is destined to collapse. One can't stop bad people from doing bad things, but when bad things happen to tall buildings, the situation quickly becomes desperate—for the occupants, for firefighters, and for adjacent buildings—in ways that are an order of magnitude worse than in low buildings.
I was once asked, during a panel discussion about the future of commercial real estate development, how tall buildings should be designed given what we'd learned from the World Trade Center collapse. My answer was, "Lower." Fortunately, the member of the audience who asked the question did not follow up with, "How much lower?" That would have put me on the spot.
On the one hand, the question of when a tall building becomes unsafe is easy to answer. Common aerial fire-fighting ladders in use today are 100 feet high and can reach to about the 10th floor, so fires in buildings up to 10 stories high can be fought from the exterior; if necessary, the buildings can also be accessed and evacuated from the exterior. Fighting fires and evacuating occupants above that height—whether the building is 20 floors, 50 floors, or 100 floors—depend on fire stairs. (Although some cities use elevators for emergency access, New York firefighters consider them dangerous and use only stairs.) The taller the building, the longer it will take for firefighters, hauling heavy equipment, to climb to the scene of the fire. Before September 11, it was assumed that in the unlikely event that people occupying floors above a fire were cut off from escape, they would simply wait until the fire was extinguished, a matter of hours at the most. After September 11, it is hard to be this sanguine. So the simple answer to the question "What is a safe height for tall buildings?" is "Lower than 10 stories."
Eliminating high-rise buildings is a radical proposal but not necessarily impractical. Washington, D.C., has had a height limit since 1910, when Congress mandated that buildings could be no higher than the width of the street right-of-way plus 20 feet. Thus buildings facing avenues, which are generally 110 feet wide, can be no more than 130 feet high—about 12 floors. Washington, D.C., has a thriving office market with rents comparable to those in other downtowns with high-rise office buildings, so it's hard to say that height restrictions have had a negative effect on commercial real estate. Developers have adjusted their plans to the rules of the game. In any case, there's nothing intrinsically more efficient about a tall building. Indeed, it's possible to argue that in some ways, very tall buildings are inefficient. Not only are they more expensive to build, but more than a third of every floor in buildings as tall as the World Trade Center towers is given over to the service core. In a 12-story building, with fewer elevators, fewer fire stairs, and smaller service ducts, the core occupies less than half that area. So why don't cities impose lower height limits? A 60-story office building does not have six times as much rentable space as a 10-story building—the core is larger and greater setbacks are required than for a lower building, but the taller building probably has four times as much space. All things being equal, such a building will produce four times more revenue and four times more in property taxes. Because most cities have lost their industrial and manufacturing functions, commercial downtowns have become the chief contributor to tax revenues. Cutting building heights would mean cutting city budgets.
Still, cities would bite the bullet of height limits if the public demanded them. But according to a recent opinion poll of New Yorkers' attitudes since September 11, considerably more people are uneasy about traveling in the subways than about going into skyscrapers. I've spoken to architects and developers about the perceived risks of working or living in tall buildings. There is agreement that many people are nervous about occupying the upper floors of prominent landmarks such as the Sears Tower or the Empire State Building, and no one seriously suggests replacing the World Trade Center towers. But I couldn't find any developers who were shelving high-rise projects in the 30- to 60-story range because they thought people wouldn't move into them, or any architects who were designing skyscrapers radically differently. Perhaps the strongest evidence that nothing much has changed is the recently unveiled design of the replacement for 7 World Trade Center: At 52 stories, it is actually taller than its predecessor.
The most important lesson of the World Trade Center collapse is not that we should stop building tall buildings but that we have misjudged their cost. We did the same thing when we underestimated the cost of hurtling along a highway in a steel box at 70 miles per hour. It took many years before seat belts, air bags, radial tires, and antilock brakes became commonplace. At first, cars simply were too slow to warrant concern. Later, manufacturers resisted these expensive devices, arguing that consumers would not pay for safety. Now we do—willingly.
Terrorists probably have something quite different in mind for their next attack, and it is not feasible to harden all existing buildings against their threats. On the other hand, it would be a mistake to conclude that we have not learned important lessons from the World Trade Center collapse. Architects and engineers must pay more attention to fire protection in tall buildings, treating it not as a code requirement but as a design problem. There have been many suggestions for ways of making tall buildings safer: more redundancy in fire exits, more robust fire stairs, extra emergency stairs, hardened elevators solely for use by firefighters, heavier construction techniques, and so on. These measures will undoubtedly raise the cost of building tall buildings, and they will probably make people think twice before building extremely tall buildings. Which would not be a bad thing. A detailed summary of the collapse of the WTC buildings can be viewed at www.house.gov/science/hot/wtc/wtcreport.htm. Read the congressional statement of Arden Bement Jr., "Learning From 9/11: Understanding the Collapse of the World Trade Center," at www.nist.gov/testimony/ 2002/abwtc.html. For an overview of various engineering approaches, see "Why Did the World Trade Center Collapse? Science, Engineering, and Speculation": www.tms.org/pubs/journals/ JOM/0112/Eagar/Eagar-0112.html. NOVA's "Why The Towers Fell" is an interactive online journey into the impact of 9/11: www.pbs.org/wgbh/nova/wtc.
