The environment inside the Deer Island Outfall Tunnel—damp, pitch-dark, with no breathable air—was about as inhospitable as you can get and still be on earth. It was into that environment that contractors sent five commercial divers to complete the final stage of the tunnel’s construction—the removal of fifty-five safety plugs, which had been installed to protect workers from being flash flooded by seawater while the structure was being built. The divers’ job was described as “somewhat akin to a spacewalk,” and no less dangerous, as it would require a series of day-long trips into the undersea tunnel, which had but a single entrance and extended ten miles out under the Atlantic Ocean.
Perhaps predictably, the effort ended in tragedy, as two of the five divers perished when their jury-rigged breathing apparatus failed. Yet the accident remains largely unknown, even in Boston, as local and national media were preoccupied with another story at the time—John F. Kennedy Jr.’s fatal single-engine plane crash, which occurred the weekend before.
Ultimately, the tunnel’s contactor, designer, and owner devised an ingenious solution for removing the aforementioned safety plugs, which was accomplished without further loss of life. And today the Deer Island Outfall Tunnel is regarded as one of the world’s most successful engineering projects, as it carries up to a billion gallons of treated sewer water a day, which has allowed for the rebirth of Boston Harbor.
In “Trapped Under the Sea: One Engineering Marvel, Five Men, and a Disaster Ten Miles into the Darkness” (Crown), author Neil Swidey relates the myriad challenges involved with completing the $300 million tunnel—and how an effort to insulate a large group of workers from an exceedingly small risk led the builders to subject five divers to extreme danger. In the process, Swidey imparts a variety of lessons, in effect turning the Deer Island Outfall Tunnel story into a case-study for how parties calculate, discount, and transfer risk. But it’s also “the story of how the modern world gets built,” Swidey says, “and how very smart people can make very bad decisions and leave guys in hard hats to bail them out.”
In the following Failure interview, Swidey explains what the divers were charged with doing, where things went wrong, and the lessons that have been learned from the deaths of Billy Juse and Tim Nordeen.
Why was the Deer Island Outfall Tunnel commissioned?
The larger outlines are that Boston had the filthiest harbor in America. There was a two-part effort to clean it up. The first part was creating a state-of-the-art sewage treatment plant on Deer Island north of Boston. The second part was to build a ten-mile outfall tunnel that would take the treated wastewater and send it far away from the harbor. Not only would the wastewater be treated very vigorously, but it would be diffused into the Atlantic Ocean. So the tunnel was the necessary second component of a two-pronged strategy to transform Boston Harbor into the cleanest urban harbor in America, which it is now.
Why were the fifty-five safety plugs deemed necessary?
The tunnel is horizontal and in its last mile connects with fifty-five vertical diffuser pipes. The diffuser risers rise up from tunnel depth [hundreds of feet down] all the way up to the ocean floor. At the top of each of those was a domed cap, built so it would resist a ship’s anchor being dragged along the ocean floor. Then built into each of those domes were nozzles, where the treated wastewater would eventually exit the tunnel through the nozzles. On each of the nozzles there was a safety cover. But the people in charge of the project were worried that if there was a freak accident and, say, an anchor dragging along the ocean floor ripped off one of the domes, then those two measures wouldn’t be enough and ocean water would flood in and kill the workers.
There was a debate about this because there were already safety measures in place and it was an extremely small risk. But the [powers-that-be] insisted on a secondary system of safety measures, which were these plugs at the elbow where the vertical riser met the horizontal tunnel. At every one of those fifty-five meeting points there was a sixty-five pound plug.
But the contract specified that the tunnel had to be completely done before the safety plugs were removed. And the definition of done meant all the ventilation, lighting, and life support in the tunnel had to be ripped out. That’s why for an exceedingly small risk to a large amount of workers they decided to put a small number of workers at an incredible level of risk by asking them to go in after all life-sustaining utilities had been removed.
Why did the powers-that-be feel it necessary to hire commercial divers to complete this last stage of the work?
The tunnel was not full [of water] yet, but they went with divers because divers know how to deal with extremely dicey conditions and how to bring in their own air and get work done.
Tell me what the five divers had to do if they hoped to be successful in removing the plugs?
The tunnel starts out twenty-four feet wide but gets progressively narrower. The five divers had two Humvees, which they drove into the tunnel as far as possible—about nine miles in. They drove one and towed the other [pointed] in the opposite direction because at a certain point there wasn’t enough room to even open the doors of the Humvees, never mind turn around. Two divers would stay in one of the Humvees and the other three would trudge to the end of the tunnel on foot, dragging their gear with them. At the end of the tunnel they had to crawl into the pipes that connected the horizontal tunnel with the vertical risers. The connector pipes were only thirty inches wide, scarcely wider than the shoulders of the divers. To get a plug out, they had to crawl thirty-five feet into a connector pipe, remove the plug, and then maneuver themselves and the plug out of that space. Then they needed to repeat that process fifty-four times.
In retrospect, the plan seems to have been viable—except for the all-important breathing system for delivering air to the divers.
Here’s the interesting thing—and it has implications for all kinds of complicated infrastructure work. One is how the plan was presented and what actually turned up. The people who signed off on this were given a very impressive plan that said: we can’t take in conventional breathing supplies like tanks with high pressure air. So we have to do something different, and we have this state-of-the-art process of mixing liquid nitrogen and liquid oxygen and vaporizing that and breathing that. Most experienced divers have used liquid gas before so the plan had the appearance of being sound, even though it was exotic and experimental. The problem is that the system was jury-rigged and key components weren’t there until after the job had begun. And the divers weren’t used to using this kind of equipment because the job was so different from anything they had done before.
Also, the parties who signed off on the plan didn’t want to look that closely. They were so tired and over budget and desperate for this job to end that they wanted to believe the plan was better than it really was. They also didn’t want to assume too much risk by commenting too much. When you are dealing with high-level joint venture projects with risk assessments the more you get involved with revising a plan, the more you start to own the risk if something goes wrong. You had a lot of parties who wanted to believe in this solution and didn’t want to engage in rigorous scrutiny for fear of absorbing more risk. But when people get too worried about offloading risk onto other parties, they can unwittingly end up inviting big risks for everyone.
What was the most flawed aspect of the breathing system?
The key flaw was asking the divers to mix their gas and breathe right off it, rather than having the gas pre-mixed and tested and in a volume tank. The other key point is that the mixer turns out not to have been designed for producing air for human consumption.
The divers had considerable difficulties the first two days on the job. Are you surprised that they went into the tunnel for a third and final day?
I’m not—now that I have gotten to know the survivors and spent time in this fraternity that is the diving community. The biggest question I had was: “Why didn’t you walk off the job?” It’s a legitimate question. But then you have to think of several things. If you’re on a small team and you walk off the job you essentially have to have unanimity, otherwise you feel guilty if something happens to your co-workers.
Another thing is that this was such an exotic, one-of-a-kind project that the divers—who are accustomed to very complicated work and a have very low threshold for nonsense—didn’t see right away how screwed up this was. Their basic instincts didn’t kick in because this was new equipment in a new environment. If they had been on an offshore rig using the equipment they were used to, I think they would have called an all-stop to the job.
What was the solution that was devised to get the plugs out?
After the tragedy the big parties put money aside and said: nothing happens if we don’t finish this job. We have this multi-billion dollar cleanup of Boston Harbor ordered by the federal courts and it won’t work unless we figure out how to get these plugs out. So let’s stop battling over ten million dollars here and twenty million dollars there and understand that four billion dollars and the cleanup of the harbor is at stake. They took money off the table and said: What do we do? Let’s have a blue sky meeting and start from scratch.
What’s interesting is that the divers had been called in because the key players couldn’t get along and wanted to go out to a consultant with fresh eyes. But in the end it was parties who had been on the job for a decade who figured out a truly innovative solution, one made possible only because the divers had managed to remove three of the plugs before the tragedy happened. At the blue sky meeting, the executive director of the water and sewer agency that owns the tunnel—who was not an engineer, by the way—said: What if we try ventilating the tunnel from the end? Is there a way we could either pump good air into the tunnel or suck all the bad air out?
The engineers took that out-of-the-box idea and made it real and tested it and re-tested it and war-gamed the entire thing in a very rigorous, smart way. And they were able to do exactly that: They built a giant steel straw—and using an offshore rig—put it over one of the domes on the sea floor. Then they used jet engine fans to suck all the bad air out of the tunnel.
Was anyone held criminally responsible for the deaths of Billy Juse and Tim Nordeen?
State police detective Mary McCauley came to the conclusion that this wasn’t an accident. But how do you prove that? How do you prove it was one party versus another? So many people had some piece of the responsibility, so how do you assign that responsibility? In the same way that assigning risk got them into problems, assigning responsibility got in the way of any kind of prosecution.
I understand that the tunneling industry is now using a new kind of contract structure—a “design-build” approach, as opposed to a low-bid/fixed cost approach.
Rather than have the designer, the contractor, and the third-party construction manager [the eyes and ears of the owner] be three independent agents battling each other [and making sure others assume more of the risk and pay more when things go wrong], you have the designer and builder being the same party. They agree with the owner that there is going to be a maximum price [so that the taxpayers have some protection], but you have flexibility within the early parts of the project to make things work.
There are downsides to every approach and the downside to the design-build approach is that it can end up pushing a lot of the risk onto the owner. But I think there are enough protections in place that it’s much more sensible.
What impact has the completion of the Deer Island project had on Boston Harbor and Massachusetts Bay?
In 1988, when George H.W. Bush was running for president against [Massachusetts governor] Michael Dukakis, one of his most lacerating attacks involved hopping on a barge and having the national media follow him as they looked at this disgusting, polluted, fetid harbor. It’s hard to imagine that it’s only twenty-five years ago that the harbor looked that bad because today the entire center of gravity of Boston has moved from the Back Bay to the Seaport district, an entire neighborhood that didn’t exist before. It was just warehouses and fish markets and muddy parking lots and now it’s the scene of nightlife and development. By cleaning up the harbor the city has reclaimed its historic connection to the sea, and it’s a source of pride rather than a source of embarrassment. It has worked better than anyone could have imagined and I don’t think people have a full appreciation of what it took to get there.
Did writing this book cause you to change the way you think about risk?
It did because there were things I never realized that apply to all of us in our everyday lives. For instance, the concept of normalization: the more often you do a job and don’t suffer negative consequences, the more likely you are to keep doing it and to push the envelope.
The best analogy is something most homeowners have done: cleaning your gutters. If you start cleaning your gutters and you are a sensible person you take the ladder out and clean a three-foot section, then you climb down the ladder and move the ladder six feet and then do it again. And you keep doing that. But by the end, when you are doing the last run of the gutters, you are going to reach farther to get to that last pocket of leaves and save yourself the hassle of moving the ladder one more time. The problem is that as you are moving along you are dulling your senses and you end up taking more risk than you should. If it was sensible to move the ladder every three feet in the beginning, it should be sensible at the end.
That’s why the counterintuitive lesson is that the most dangerous time on a big construction project is not the beginning, it’s the end, when people are feeling cocky because they know how to do things and take chances they wouldn’t have taken at the beginning. They are also feeling pressure to finish because of external forces. That’s why if someone is texting while driving and doesn’t suffer consequences from it, they are more likely not only to text at stop lights, but to text when they are driving through traffic.
The other thing is that you have to make sure that safety measures make sense. You have to understand what the risks are in order to take measures that are sensible, because in the end it was secondary safety plugs—an afterthought on this multi-billion dollar project—that did them in.
Do people learn from tragedies like the one in “Trapped Under the Sea”?
I don’t know. I hope that two things come from people reading about this story. One is that they appreciate what it takes to make modern infrastructure possible. Nothing of substance gets built on paper. It gets built by unnamed workers taking on real risks.
I also hope that it makes people who are involved in complicated undertakings understand how very smart people can make very bad decisions. That takes some humility and some willingness to step back and look at what is going on: That is, if I win the battle with this other party, am I risking the war?
What has been interesting is talking with the people I wrote about. There were no black hats. People were trying to do their best, but some of the best and brightest engineers who were involved told me after reading the book that it was revealing for them to see [the big picture]. At the time they saw through one lens, but to see what they weren’t seeing when they were in the heat of the battle was really informative. This is a painful story but I hope people learn from it and that it has some residual positive effect.