American Scientist

The Golden Gate Bridge might be said to be the West Coast version of the Brooklyn Bridge. They share the basic features of a suspension bridge, most noticeably in the parabolic sag of their main cables that support a shallow arched roadway stiffened by trusses composed of triangles. Yet they do not resemble each other in their more prominent architectural details. The staid Brooklyn is characterized by tall Gothic archways piercing its massive masonry towers, from the top of which radiate its signature diagonal stay cables. The Golden Gate has soaring art deco steel towers that are dominated by crisp horizontal and vertical planes and are painted a bold international orange color.

The Brooklyn Bridge took 14 years to build and opened in 1883; the Golden Gate was under construction for three and a third years and was completed in 1937. Each bridge was record-setting for its time. The total length of the Brooklyn is a tad over 6,000 feet, or about 1.1 miles (1.8 kilometers) from end to end. That of the Golden Gate is just short of 9,000 feet, or about 1.7 miles (2.7 kilometers), in total length, making it about one and a half times as long as the Brooklyn. But the accepted measure of engineering achievement for a suspension bridge is the distance between towers, which for the Brooklyn is within five feet of 1,600 (about 0.5 kilometers) and for the Golden Gate is exactly 4,200 feet (1.3 kilometers), a factor of about two and a third times as long.

Since their respective openings, these bridges have each undergone changes in structure that might be considered by the nonengineer to be so inconsequential in scale as to go unnoticed from a distance. However, each change altered the way the bridge behaved and how it functioned for its users, both vehicular and pedestrian. The original configuration of the Brooklyn Bridge’s roadway contained in each direction two narrow lanes for horse-drawn carriage traffic beside a single railroad track, all separated by an elevated walkway for pedestrians. The tracks carried cable-car–like trains that shuttled passengers between terminals built at the foot of the bridge on both the Brooklyn and Manhattan sides. Over time, the trains were replaced by lighter trolley cars and eventually were eliminated to accommodate three lanes of roadway for automobiles and trucks.

The original layout of the bridge deck had vertical stiffening trusses on each side of the railroad tracks. The inside trusses remain in place, but the outside ones were relocated to the edges of the bridge deck to eliminate a barrier between the motor vehicle lanes, but this change did not significantly alter the dead load on the bridge deck or its stiffness. With the continued growth of traffic in volume and in the weight of motor vehicles, trucks were banned from the bridge altogether. Also, with the growing encouragement and use of bicycles, which shared the elevated walkway with pedestrians, safety concerns led to a dedicated bike lane.

Sitting Steady

The Golden Gate Bridge has undergone more significant structural and behavioral changes. The bridge was designed at a time when the state-of-the-art in bridge engineering was to make the roadway of a suspension bridge look as slender as possible. This aesthetic imperative worked fine for bridges designed to carry multiple lanes of vehicular traffic in addition to railroad trains and pedestrians. Around large cities, this trend led to bridges with three or four traffic lanes in each direction plus tracks for commuter trains. The resulting wide bridge decks were necessarily heavy, and this feature even in the absence of the stiffening trusses made the roadways very stable in the wind. However, as suspension bridges began also to be built in less populated areas, where traffic consisted almost exclusively of cars and trucks, trussless decks carried only two lanes of traffic plus a very narrow sidewalk or two. This layout naturally meant the decks were lighter, which in turn led them to be more flexible, which in time manifested itself as excessive motion in the wind.

John Roebling, the engineer of the Brooklyn Bridge, understood this problem in the middle of the 19th century, when wind was destroying suspension bridges, and his design imperative for a bridge to sit steady in the wind was that its deck be sufficiently stiff by having substantial weight, longitudinal trusses, and diagonal stay cables. His Brooklyn Bridge epitomized the application of this formula. However, beginning early in the 20th century, suspension bridge design trended toward a sleek aesthetic combined with a narrow structure spanning a greater distance. In other words, the bridge deck became lighter and less stiff and less constrained. As the design of any bridge is basically a theoretical hypothesis until it is built and tested by traffic and the elements, bridges designed and constructed in the 1930s tended to be susceptible to movement in strong winds, an old phenomenon that did not manifest itself anew until the second half of the decade. The Golden Gate Bridge fell into this category.

Each change in structure altered the way the bridge behaved and how it functioned for its users, both vehicular and pedestrian.

• Facebook
• Twitter

The bridge was designed and built with a pair of vertical stiffening trusses connected by a horizontal one incorporated into the deck, which carried six traffic lanes and two wide sidewalks. The truss arrangement proved to be inadequate in preventing the roadway from excessive movement. The deck was designed to move almost 30 feet (9 meters) laterally in a crosswind, a distortion that was easily seen by sighting along a line connecting the bridge’s two towers. A certain amount of vertical displacement was also expected, but not to the extent that traveling waves were observed to develop along the length of the roadway. When the Tacoma Narrows Bridge was brought down in 1940 by a 40 mile-per-hour (64 kilometer-per-hour) wind that induced vertical and twisting oscillations along its entire length, there was cause for concern, because winds through the Golden Gate strait can gust to 75 miles (121 kilometers) per hour. Strong winds can also cause vehicles on the bridge to become difficult to control, and can blow over trucks with large, exposed side panels. In 1951, wind caused the bridge deck to sway and roll, and there was concern that these motions could become unstable and develop into the kind of twisting that destroyed the Tacoma Narrows Bridge. To address such issues, an additional horizontal truss was added that connected the bottoms of the two vertical ones of the Golden Gate. By closing the trusswork in this way, the bridge deck was made significantly stiffer, especially against being twisted.

The positive retrofit improved the stability of the roadway, but the additional weight of steel added by augmenting the original trusswork put the dead weight of the deck too close to the limit that the main cables were designed to support. As a result, when there was talk of extending the tracks of the Bay Area Rapid Transit system across the bridge and into Marin County, the plan had to be abandoned. Such can be the kind of unintended negative consequence of making what appears otherwise to be a positive change. It was not the last time a design change authorized by the Golden Gate Bridge, Highway, and Transportation District would come back to bite.

Safety First

In the meantime, another issue with the bridge arose. The dramatic setting of the iconic structure between the roiling of the vast Pacific Ocean to the west and the relative calm of the safe harbor known as San Francisco Bay to the east makes the bridge mark an entranceway of sorts. But the bridge’s location is also prone to fog rolling in, engulfing the bridge so thoroughly that little more than the tops of its towers remain visible, as if they were held aloft by nothing but clouds of mist. For whatever reason, the bridge is also the site of choice for suicides. As of last year, more than 1,800 people had climbed over the sidewalk railing and jumped from 220 feet (67 meters) above the water—an average of 20 suicides per year. Hundreds more were prevented each year by the intervention of bridge and highway patrol crews and concerned citizens. A passive physical barrier was believed to be the definitive solution that was needed.

A debate about adding any kind of suicide prevention system to the bridge had raged for decades, but it came to a potential resolution in 2014, when the region’s Metropolitan Transportation Commission pushed for the funding of a $66 million project to either build a high fence atop the railing or extend a safety net out beyond the edge of the bridge, to be paid for in part by a planned toll increase. (The toll for an automobile would go from $6 to $8, collected only from northbound traffic; the Golden Gate was the first major toll bridge to charge tolls only one way, thereby allowing traffic to flow freely in at least the other direction.) The toll hike was expected to increase revenue by $138 million, which the district had initially planned to use for seismic reinforcement of the bridge, the installation of a moveable barrier to allow the six-lane roadway to be divided unequally during rush hours, and improvements in ferry and bus service. The transportation commission asked the bridge district to divert $12 million of their windfall toward the suicide barrier. Once funding was in place, the project proceeded in earnest.

The so-called Suicide Deterrent Net System, euphemistically referred to as the Safety Net, was designed “to minimize impacts to Bridge views and appearance,” according to the bridge district, and to allow the daily operation and maintenance of the bridge to proceed unimpeded. The net was to be positioned 20 feet (6 meters) below the sidewalk and extend out 20 feet (6 meters) from each railing. Woven from stainless-steel rope, the chain-link-fence–like netting would rest on horizontal struts cantilevered out every 50 feet (15 meters) from vertical struts that would be continuations of the vertical members of the bridge’s stiffening truss, thereby minimizing visual discontinuity. By painting the safety system parts the same color as the bridge, it was expected to be all but invisible from a distance. But anyone striking the net after jumping or falling that distance is likely to sustain injuries, making it imperative that area first responders be trained in rescuing them. The training will be done at a Marin County fire station tower, where a full-scale mock-up of the net has been constructed.

Because the Safety Net will add additional steel and hence additional weight to the already heavily loaded bridge, part of the project involved altering the weight of the railing to compensate. This change was done with a simple railing redesign, which promised not only to compensate somewhat for the added weight of the barrier but also to offer less resistance to the wind. Such a result was clearly desirable, especially because the addition of the suicide barrier was going to add wind-resisting surfaces and so increase the overall wind force on the bridge. It appeared that this addition could be offset by reducing the surface area that the balustrade presented to the wind.

The original railing design was in keeping with the hard edges of the towers: The balusters were steel H- sections attached by rivets to steel channel shapes that constituted the top and bottom members of the balustrade. The 4-inch-wide (10-centimeter-wide) flange of each H-section was oriented parallel to the railing length, giving the impression of a simple fence with pickets spaced 4 inches (10 centimeters) apart. The change consisted of replacing the H-sections with 4-by-¼-inch (10-by-0.6-centimeter) steel plates, with the wider face oriented perpendicular to the length of the railing. By presenting to the wind a face that is only a ¼-inch (0.6 centimeters) wide every 4 inches (10 centimeters), the wind resistance of the balustrade was significantly lowered and so was the force of the wind on the bridge.

The new railing design caused an “eerie hum” to emanate from the bridge on windy days, becoming unbearable for residents living even some distance away.

• Facebook
• Twitter

In 1993, the railing replacement operation attracted the attention of Richard Bulan, a San Francisco native who had fond memories of the bridge. He imagined salvaging some railing to make a headboard for his bed and asked the contractor replacing 12-foot-long (3.5-meter-long) sections of the railing, each of which weighed of the order of 1,000 pounds (450 kilograms), if he could acquire one. While converting the railing section into a headboard, Bulan got the idea of making other pieces of furniture out of salvaged and refinished railing, which in many places had been deteriorating from exposure to the wind and salt air. Because the contractor’s responsibility included carting away the sections of old railing at his own expense, he agreed to let Bulan have all he wanted for his nascent enterprise. His plan was to cut up the rail sections, clean them of paint and rust, and reassemble usable parts into pieces of furniture, such as bookends, lamps, and coffee and end tables, which he now sells as pieces of nostalgia through his Golden Gate Design & Furniture Company.

Look Out for Bicycles

A 2013 study conducted by the Golden Gate Bridge, Highway, and Transportation District determined that the maximum sustained wind speed the original bridge was designed to take was about 70 miles (112 kilometers) per hour. However, at least three times in its history (1951, 1982, and 1983) the bridge has been subjected to gusts near to or exceeding that value, forcing it to be closed to traffic. The new railing design was expected to make the bridge safe in winds up to 100 miles (160 kilometers) per hour. However, as the railing replacement project proceeded, two unanticipated consequences became clear.

One was the development of an “eerie hum”—described as sounding like “chanting monks” and a “wheezing kazoo”—emanating from the bridge on windy days. In 2020 it became unbearable for residents living as far as 3 miles (5 kilometers) away. A study of the phenomenon revealed that the frequency of the sound was about 440 hertz, and it occurred when the wind direction was at a slight angle to the long dimension of the thin slats of the railing. Wind tunnel tests of a full-scale section of railing were conducted to explore ways to quiet the hum.

Meanwhile, during the time when sections of old railing were being replaced by new, cyclists who pedaled regularly across the bridge experienced increased difficulty in maintaining their balance on windier days, especially as they rode past new sections of railing. Whereas the old design had provided somewhat of a barrier against crosswinds, the newer design allowed wind to blow across the sidewalk with much greater velocity. Even experienced bicycle riders reported having difficulty in steering a straight path when traveling at about 8 miles (12 kilometers) per hour, a problem that disappeared as soon as they reached a location where the old railing was still in place.

The Marin County Bicycle Coalition, an advocacy group, challenged the claim that such reports were anecdotal by presenting anemometer readings showing that the velocity of the wind blowing through the new railing could be as much as two or three times what it was through the old. The coalition reminded citizens on its website that “anyone who has ever ridden a bike knows that a 10-mph crosswind pales in comparison to one blowing at 20 or 30 mph.” Further confirmation of the problem came from the results of a coalition survey that found that 84 percent of respondents reported that riding across the bridge had become worse since the new railing was installed, and 62 percent claimed it to be “much worse.” The coalition also asserted that the survey revealed that the bridge district was “vastly underreporting the number of bicycle crashes on the bridge,” because nearly 20 percent of respondents had reported having crashed because of the wind.

In response to such revelations, the bridge district stated that the new railing design was the result of “decades of work,” some done as early as the 1990s, “to make the bridge able to withstand high winds,” evoking the collapse of the Tacoma Narrows Bridge as a warning against not taking action. But whereas the safety of the bridge had been given high priority, it was not clear that the effect on cyclist safety was even taken into account. It remains to be seen what, besides warning cyclists of the wind effect, might be done to mitigate the force of the wind blowing across the sidewalk.

Furthermore, the humming sound was not anticipated to be as loud as it proved to be. In 2010, the final environmental impact report for the bridge modification project stated that the changes would not cause any substantial change in ambient noise. However, once in place, the sound emanating from the new railing arrangement was found to reach 100 decibels. When wind tunnel tests indicated that the noise could be reduced by 75 percent by installing thin, U-shaped clips made of aluminum with rubber inserts over the windward face of the slats, a $450,000 retrofit of the retrofit was approved. The job was supposed to be completed late last year but remains ongoing.

There has also been a delay—and a cost increase—for the installation of the suicide barrier. Contractor bids for the job came in at about double the estimated $76 million cost, which by September 2022 had risen to about $222 million. The metal netting began to be installed last year, and the barrier is expected to be completed this November. Those who have worked so hard to achieve the goal of a safer and quieter bridge have little choice but to remain patiently hopeful that it will be.