The Blue Death

by Dr. Robert D. Morris

  The plant was an unmitigated disaster. The muddy water clogged the filters so quickly that they had to be cleaned every four hours. George G. Earl, who became superintendent of the New Orleans Sewerage and Water Board in 1892 after the filters were already under construction, concluded that the water they produced was “not fit for any use except lawn sprinkling.” Much of the sediment slipped through the filters, clogged the pipes, and resulted in water flowing from the taps of New Orleans that was “often muddier than the river itself.” The New Orleans Water Company, a private company under contract with the city of New Orleans, refused to pay for the filters and National Water Purifying collapsed after losing a court battle to force payment. In this infamous bit of engineering hubris, National Water Purifying failed to grasp a lesson almost as old as Egypt.

  Three hundred years earlier, Prospero Alpini, an Italian botanist and physician, had described watching as an Egyptian man filled his camel-skin bag from a cloudy tributary of the Nile and poured it into an oblong clay vessel. The man took a handful of crushed almonds, thrust his arm into the vessel, and stirred it violently. After three hours most of the fine mud had settled and the water was ready to drink.

  The Egyptians were not adding almonds for flavor. The oils from the almonds helped suspended particles clump together, a phenomenon known today as coagulation. The practice was so common in Egypt that cakes of almonds were sold in the market just for water purification. By the late nineteenth century, a range of chemical coagulants had been introduced that provided far more effective treatment than crushed almonds. In designing their treatment plant, National Water Purifying had used a coagulant, but tried to save money by not building sedimentation basins, which would have allowed the particles to settle before reaching the filters. Instead dense clumps of Mississippi mud clogged their filters. The disastrous results dissuaded New Orleans from filtering its water for another fifteen years.

  Faced with the relentless threats of rampant waterborne and mosquito-borne disease and with no viable alternative to the water from the Mississippi, the city embarked on a bold two-part plan in 1906. The most visible and dramatic element of the plan involved the installation of massive pumps intended to operate almost constantly to drain the swamps around New Orleans. The drainage would surround the city with thousands of acres of new land. The second part of this plan involved the construction of one of the country’s largest water treatment plants on seventy of those acres. The city fathers were determined to avoid any repeat of their disastrous first attempt to drink the Mississippi. The Carrollton Water Treatment Plant would use coagulation and sedimentation in conjunction with sand filters, an approach now standard for the treatment of water from an unprotected source.

  Today a cryptosporidium oocyst that has washed off a dairy farm in northern Louisiana and floated down the Mississippi into the intake of the Carrollton plant would face a daunting obstacle course if it hoped to reach a glass of drinking water and infect a new human host—daunting, but not impassable.

  The pumping station for the Carrollton Treatment Plant can suck water out of the Mississippi River fast enough to fill a swimming pool in about ten seconds. The water boils with thick clouds of particles, the accumulated debris of the river basin. As soon as the water leaves the pump, coagulant flows into the huge pipe and treatment begins.

  Packs of long, snakelike molecules of coagulant wrap themselves around particles, neutralizing their electrical charge. Ordinarily that electrical charge would help to keep particles apart. Now, as the uncharged particles collide, they begin to stick together, like drops of oil in a glass of water. By the time the water reaches the treatment plant, the clumps have grown so large that they begin to sink and form a layer of sludge on the bottom of the plant’s sedimentation basins. If our oocyst avoids getting stuck to one of these aggregates, it flows onto the plant’s filters.

  These beds of sand lie at the heart of the obstacle course. From a human perspective, it is hard to imagine anything significant passing through this thick layer of fine-grained sand, but this perception simply reflects the stubborn myopia of our human-scale world. To help us view this filter with a microbe’s eyes, imagine an oocyst enlarged to the size of an average human adult. When we enlarge the filter to match, the smallest grain of sand would grow to the size of a four-story building. The largest would become taller than the Sears Tower.

  At this scale one can easily imagine our supersize oocyst tumbling through the gaps between gargantuan grains of sand. Some of the oocysts would get stuck, but many others would find a way through. Bacteria, slightly smaller than an oocyst, could also slip through this maze. For a virus bypassing a filter is even easier. Relative to our human-size oocyst, this protein-encrusted virus would be like a child’s marble tumbling through a chaotic stack of boulders the size of apartment buildings.

  The coagulant makes the oocyst stickier and makes the filter more effective. As microbes and other debris accumulate on the top layers of sand, they begin to form dense layer of microorganisms and fine sediment known as the schmutzdecke (from the German, “filth cover”). The schmutzdecke is so essential to the effective operation of a water filter that a clean filter cannot be put online until it forms and must be taken offline before that critical film becomes old and thick and cracks begin to form that would allow pathogens to pass through again. A filter with a functional schmutzdecke will remove ninety-nine out of a hundred oocysts.

  One would think that removing 99 percent of the problem would eliminate it, but oocysts can be present in large numbers. If we took the ten thousand oocysts that would fit into the period at the end of this sentence, added them to a liter of water, and poured it over a filter bed, the liter that trickled out would still contain a hundred oocysts. This explains how Rose and others often found cryptosporidium oocysts in treated water.

  If we remember that viruses are far smaller and harder to remove than oocysts, it is clear that filters alone provide only limited protection against pathogens. The engineers who designed the treatment process for New Orleans understand this too. Just before filtration they bombard the oocyst with chlorine. A powerful oxidant chlorine burns through the outer skin of pathogens like molecular napalm. The bacteria are usually the first to die. With a tougher shell, viruses last longer. As they can pass through filters more easily, it is critical the operators add enough chlorine to destroy them. But with the arrival of our oocyst, plant operators have a problem.

  When microbiologists who study cryptosporidium want to isolate oocysts from a stool sample, they add bleach to it. The extremely high concentration of chlorine in the bleach wipes out most other microbes, but the shell of the oocyst is so tough that the bleach can’t cut through it. What remain are viable, infectious oocysts. In other words, there is no way for operators to add enough chlorine to kill cryptosporidium. This explains how cryptosporidium devastated Milwaukee even though plant operators maintained required chlorine levels at all times.

  To varying degrees, we all drink from the Mississippi. Water supplies around the country face some if not all of the problems confronted by pre-Katrina New Orleans. Our source waters are universally tainted. In other words, most Americans consume treated sewage.

  Almost every water supply in the United States relies on treatment processes similar those used in New Orleans. These filtration systems do not remove 100 percent of pathogens and many of those pathogens are to some degree resistant to chlorine, the chemical used almost exclusively to disinfect drinking water in the United States. To make matters worse, the chlorine used to protect us from waterborne disease may threaten our health in other ways including cancer, stillbirths, and birth defects.

  When the EPA convened the Federal Advisory Committee on Disinfection By-products and Microbial Risk in 1996, these were the problems laid before them. With the events in Milwaukee, it seemed that the writing was not just on the wall, it was written in bold, flashing neon.

  The tiny oocyst certainly got the industry’s attention. No one wa
nted a disaster like Milwaukee’s to occur on his or her watch. In the drinking water industry, however, big changes mean massive capital projects with long lead times. Over the next nine years, I would watch drinking-water industry lobbyists do whatever they could to slow down and dilute the regulatory efforts that followed.

  At one of the committee’s regular meetings, I came face-to-face with the industry’s resistance to major change. Ed Means, the representative from the American Water Works Association (AWWA), the largest industry trade group, was laying out its position on turbidity limits. Means, smooth and quick-witted with surfer-boy good looks, allowed that the AWWA would accept a reduction in the maximum allowable turbidity for treated water from 5.0 NTU to 2.0 NTU.

  I listened in disbelief. The turbidity of the water in Milwaukee during the spring of 1993 never exceeded 1.7 NTU. In other words, the drinking-water industry was proposing a standard that would continue to put a federal stamp of approval on the water responsible for the largest waterborne outbreak in United States history. I had been appointed as a technical adviser to the committee and it seemed that some technical advice was clearly needed. At the next pause in the discussion, I reminded the committee members of the turbidity level in Milwaukee. I assumed they had simply forgotten the exact numbers and would appreciate the correction. I assumed wrong.

  At the next break, the professional mediator who was running the meeting pulled me aside. Committee members from the drinking-water industry, she told me, were extremely upset that I had spoken out. Technical advisers, according to strict protocol, should only give information if they are specifically asked to do so by members of the committee. I had served as a technical adviser to many different committees and had never before been chastised for offering advice. Moreover, the previous meetings of this committee had all been relatively informal. My comments, it seemed, had inspired not gratitude, but a new desire for strict adherence to protocol.

  The first round of negotiations by the committee stretched on for almost a year as the negotiators struggled with scientific uncertainty and the inevitable tension between costs and ideal public health protection. The many sides inched their way toward consensus. In large part that consensus was driven by an analysis of what improvements could be made by refining the operation of existing treatment plants rather than introducing any fundamental change in the treatment of drinking water. Even a consensus limited by those constraints proved fragile. At the committee’s final meeting, it was almost destroyed.

  As the committee members prepared to sign off on a proposed change to the rules for treating drinking water, the moderator who had brokered the deal went around the table asking all the members if they could go back to their organization and sell the deal. From Ed Means of the American Water Works Association to Erik Olson of the NRDC, one committee member after another agreed to the new rules. Just as this began to take on the air of a formality, the question passed to Roy Heald.

  Tall and angular, Heald represented the National Rural Water Association. With far shallower pockets than the large cities, the small communities he represented were wary of major rule changes. Nonetheless, he had remained remarkably quiet throughout a year of meetings.

  Heald, it seems, subscribed to an extreme economy of words. Uttering the first words I could remember falling from his lips in almost a year of meetings, he said simply, “No, I can’t sign this.” Apparently, he had been sitting through the entire meeting for the sole strategic purpose of blocking the process at the point of maximum leverage.

  The thud of jaws hitting the floor was almost audible. After months of tough negotiations, the other committee members seemed ready to string Heald up by his heels. The moderator called for a break and she and the EPA administrators huddled with Heald in an effort to assuage his concerns. They brokered a deal that would allow the agreement to survive and the process to proceed. A process that had begun with preliminary discussions in 1992, even before the Milwaukee outbreak, would produce rules that were not scheduled to go into effect until 2002. As the process inched forward, problems continued to plague the drinking-water industry.

  13

  DEATH IN ONTARIO

  F rank Koebel just ignored the instructions. A brooding bull of a man with short black hair and the ruddy complexion of an alcoholic, Frank didn’t like being told what to do. As foreman for the Public Utility Commission (PUC), he was responsible for maintaining the water supply for Walkerton and the surrounding towns. The chlorinator for well 7, the main source of water for the city, hadn’t worked for almost a week. His boss, who happened to be his older brother, Stan, had just left town for a weeklong conference. Before he left, Stan told Frank to install the new chlorinator. Frank decided he had more important things to do.

  Installing the new chlorinator just never seemed urgent. After all, it had been sitting in storage for over a year. A few days more or less wouldn’t matter. Besides, chlorine didn’t seem to do much besides ruin the taste of the “cool, clear, crisp water” that flowed from the aquifers deep beneath southern Ontario. Even when the automatic chlorinator was running, Frank rarely set it to add the required amount. If he did, customers complained. In fact, the wells had run for decades without chlorine until a new regulation required it. When he or Stan was thirsty, they liked to drink the raw, unchlorinated water straight from the well.

  To be precise, Frank had already taken the first step in installing the new chlorinator. Two days earlier he and two PUC employees had opened the bypass valves and removed the old chlorinator. Up until that point, the old equipment, while fraught with problems, could still chlorinate the town’s water. They left it sitting idle, disconnected, and useless. As they shut the door, the steady whir of the pump meant that well 7 was still running, sending raw water into Walkerton.

  On May 8, 2000, three days after Stan Koebel left, rain began to fall. Over the next two days, a steady downpour drenched the town and the farms around it. Along highway 9, the main highway through Walkerton, construction workers slogged through a growing sea of mud. They had exposed the rotten vasculature of pipes and had laid new pipes to replace them. To do that, they would need to cut into the existing water mains to connect the new pipe.

  Even Frank Koebel, who had gotten most of his training on the job, understood that opening the pipes in the distribution system posed a risk of contamination, just as a cut in a vein or artery raises the threat of infection. As the spring mud poured into the trenches, the danger increased. If pathogens got into the system, the only thing that would stop them was chlorine.

  If the rain worried Frank, it did not convince him to install the chlorinator. It appears that he was not much for vigorous exercise. At forty-one he had already had two heart attacks. The town’s two other wells, 5 and 6, both had chlorinators. (Wells 1–4 had been closed long before.) It is not clear whether Frank Koebel or a lightning bolt did it, but, in the dark of night, wells 6 and 7 shut down leaving just 5 to supply the city.

  A small, pious farming community parked halfway between Lake Huron and Lake Erie, Walkerton is the kind of town where nothing ever happens. But Frank Koebel and the storm had just planted a bomb beneath the peace and quiet.

  The Koebel brothers’ faith in the quality of groundwater was not entirely misplaced. Thick layers of soil, sand, and clay usually filter out most pathogens from water on its way to a deep aquifer. In fact, the beds of sand at the heart of our water purification systems are little more than minuscule imitations of the immense filters that overlie our groundwater. There is evidence that some microbes, particularly viruses, can survive deep below the surface, but they do not appear to pose a major disease risk. The wells in Walkerton, however, were not as safe as the Koebels imagined.

  On May 12, in the midst of a ferocious rainstorm, a newborn foal struggled to his feet under the watchful eyes of David Biesenthal, a farmer and veterinarian. After five days of rain, the skies had opened wide, dumping three inches of water onto the fields where newly planted corn was just breaking
the surface. Biesenthal had taken advantage of the warm weather just two weeks earlier to sow his fields. The rain, together with the rich manure from his cows and their young calves he had spread on his fields, would give the new crop a great start.

  Most of the rain that fell on the Biesenthal farm drained toward Silver Creek. One field, however, sloped gently toward a clump of trees growing on a piece of land owned by the town. David Biesenthal had never noticed the small cinderblock building hidden in those trees. He never heard the pump that ran inside that building. He never realized that the shaft for well 5 reached down into the groundwater just thirty feet below that pump. He never imagined that the rain that watered his corn was carrying soil and manure and bacteria toward the source of Walkerton’s drinking water.

  Stan and Frank Koebel had water in their blood. They had grown up around the pump houses of the wells of the PUC where their father had risen to the rank of foreman. As soon as Frank finished school, he went to work for the Walkerton PUC. Twenty-five years later, in May 2000, he had his father’s old job.

  Five years older, with gray hair, a mustache, and a nervous expression, Stan Koebel had always been Frank’s boss in one way or another. After thirty years at the PUC, he was now the general manager. In addition to the water, Stan was responsible for Walkerton’s electricity. The imminent deregulation of the electrical industry meant that electricity, not drinking water, had occupied most of his time. He tended to leave the operation of the water supply to his younger brother.

  Stan returned to the offices of the PUC at six A.M. on Monday morning, May 15. His first task was to deal with an employee at the PUC shop with whom he was having a bitter dispute. Stan planned to have him fired. Before he left for the shop, he looked at the computer control system and noticed that well 7 was not running. Assuming the chlorinator was installed and it had been turned off in error, he turned it on.