David Grann - The Devil and Sherlock Holmes

The construction of the first water tunnel required no fewer than eighty men to drill and blast for at least a week in order to advance a hundred feet. The mole, with a fraction of the manpower, can tunnel that far in a day.

Yet, even with the mole, the third water tunnel has already taken six times as long as either City Tunnel No. 1 or No. 2; some people think it won’t be completed, as scheduled, by 2020. “We should’ve been done with this thing twenty years ago,” Jimmy Ryan said. “But the city keeps fucking around.”

Conditions above ground have proved almost as difficult as those below. After the initial phase of a billion-dollar contract to build the tunnel was awarded to a consortium of companies, costs began to exceed estimates by the millions. When the city balked at the rising costs, the companies sued and the work stalled. Then, in 1974, when the city went bankrupt, construction was halted altogether. In all, nearly a decade was lost, and in 1981, with work proceeding only piecemeal and the ever-growing demand for water forcing the old tunnels to carry sixty per cent more capacity than intended, city officials were so desperate that they pleaded with the federal government to fund the project.

Meanwhile, charges began to surface that Tammany Hall-like machinations were contributing to the delays. The once vaunted Board of Water Supply, which oversaw the construction, had become a “Democratic patronage plum tree,” as one critic put it. Stanley M. Friedman, the Bronx Democratic power broker who was later convicted of racketeering, was given a lifetime position on the board, with a salary of twenty thousand dollars, as well as an office, a secretary, a chauffeured car. “When I came in as mayor, it was a lifetime job given to retiring politicians,” Koch told me. “They didn’t do anything.”

The board was dismantled. But in 1986 the man in charge of supervising purchasing for the water tunnel at the D.E.P., Edward Nicastro, warned that contracts were still not being properly monitored. “You’d be amazed at how easy it is to steal in the system,” he told a reporter at the time.

In recent years, the greatest delays seem to be caused not by efforts to defraud the public but by attempts to placate it. Where the old water board once plowed over communities, the D.E.P. is now impeded by them. In 1993, when it tried to sink a shaft on East Sixty-eighth Street, Councilman Charles Millard protested that his office had received calls from parents whose children were “finding it difficult to concentrate.” NUMBY,or “not under my back yard,” movements sprang up. In 1994, after engineers had spent two years planning a new shaft site, residents in Jackson Heights held a protest, carrying signs that said, “DON’T GIVE US THE SHAFT.” Engineers were forced to find a new location. “When we want to choose a shaft site, everyone says, ‘Oh, the water system is a miracle, but please find another place,’” Ward told me. “‘We’re building a co-op’-or hotel or park-‘there.’” A D.E.P. engineer and geologist, Scott Chesman, added, “Instead of taking seven years to finish, we’re on thirty years, and hardly any of it’s been done. It’s like the eighteen-hundreds again.”

Indeed, for the first time the historic Delaware Aqueduct-the eighty-four-mile underground pipeline that carries the water from reservoirs upstate down to Yonkers, where it connects to City Tunnels No. 1 and No. 2-has begun to crack. According to some D.E.P. reports, in 1995 the aqueduct was losing about five hundred million gallons a month from leaks, which were creating massive sinkholes in Ulster and Orange Counties; in 2000, the monthly loss sometimes exceeded a billion gallons. An investigation by Riverkeeper warned of a potential “collapse” of the aqueduct, which would cut off as much as eighty per cent of the water flowing into the city.

In the spring of 2000, the D.E.P. decided to send a team of deep-sea divers down to do repairs on one of the original bronze valves in the Delaware Aqueduct, in the Dutchess County town of Chelsea, which had cracked, spewing a torrent of water through a hole the size of a quarter at eighty miles per hour. “For about two or three months, we built a mockup of the valve and a mockup of the bottom of the shaft,” said John McCarthy, the engineer who oversaw the project. “We took the crew and experimented in a tank of about fifty feet of water, without any light, trying to simulate the conditions.”

After practicing for days, the engineers transported a diving bell and a decompression chamber to the leak site. Four divers, who were hired from the same company that had helped to salvage the Russian nuclear submarine Kursk after it sank in the Barents Sea in August, 2000, had to remain inside the decompression chamber for twenty-four hours, in order to adjust to the intense water pressure underground. The chamber was about the size of a van, only round. On the outside were valves and hoses and an air-lock door to send in food (mostly fluids and peanut butter) and to remove human waste. The pressure in the chamber was gradually brought to the same pressure as that of the water seven hundred feet underground.

After breathing a mixture of ninety-eight per cent helium and two per cent oxygen for twenty-four hours, two of the divers crawled into a thirteen-foot diving bell that was attached to the top of the chamber. Once they had sealed themselves inside, the bell was lifted by a crane and lowered down the shaft that led into the aqueduct. There were only inches between the bell and the walls of the shaft. When the divers reached the bottom, one climbed out and swam toward the leak. (The other diver remained in the bell in case of an emergency.) He wore a wetsuit, a mask, and scuba equipment, and carried a small waterproof tool set. While struggling to stay in position against the pressure of the escaping water, he placed a brass plug in one of the holes, then sealed it with a clamp and an epoxy compound.

Each shift lasted at least four hours, then the bell was lifted up and two other divers went down. “It was not for the faint of heart,” McCarthy said. The men spent ten days finishing the repairs, and fifteen more in the decompression chamber.

Still, far greater leaks are suspected somewhere between the Rondout Reservoir, in the Catskills, and a reservoir in Putnam County. In June, 2003, the D.E.P. sent a custom-made two-million-dollar submarine through forty-five miles of the Delaware Aqueduct. (The job was deemed too dangerous for a human.) The eight-hundred-pound craft, which was nicknamed Persephone, took three hundred and fifty thousand photographs. “The sub looks like a torpedo with catfish antennas,” Commissioner Ward told me. “While a motor pushes it through, the antennas help it bounce back off the walls to stay within the middle of the tunnel.” The Woods Hole Oceanographic Institution, on Cape Cod, and the D.E.P. are examining the pictures to evaluate the structural integrity of the pipeline.

But even if the locations of the leaks are determined, and if engineers can then concoct some way to plug them, most D.E.P. officials I spoke with do not consider this section of the aqueduct the most vulnerable. They are more worried about pipelines closer to the city-in particular, Tunnels No. 1 and No. 2, which, because of their greater depth and buried valves, are far less accessible, even to a self-piloted submarine. Some sand-hogs believe that the only thing preventing these sections from collapsing is the pressure of the water pushing against their walls. A former chief engineer on the water system, Martin Hauptman, has noted, “We see headlines in the streets frequently where a 24-inch water main breaks and the street’s flooded, basements are flooded, the subway is flooded, and people think that is a horrible situation. Failure of a tunnel is an entirely different situation. What bothers me most… is the element of time. You cannot buy time with a situation like that.”