In 1991 a group from the U.S. Environmental Protection Agency traveled to the city of Ostrava in what was then Czechoslovakia. They looked at the grimy streets and sky and thought: Pittsburgh, 1940s. The association is easily made. Ostrava represents a mighty union of coal mines, coke ovens, iron smelters, and steel mills. Smokestacks strafe the air pretty much unchecked. On good days, when the wind is blowing, the plumes are swept low across the city and out of town. On bad days, temperature inversions bring high pressure and still air that squeeze the pollution directly onto the 330,000 residents.
The EPA workers were welcomed into Ostrava after the Velvet Revolution that liberated Czechoslovakia from the Soviet bloc in 1989. Like others across Eastern Europe and Russia, the Czechs blamed the pollution created during the Communist era for a raft of ills--among them rising rates of cancer, respiratory disease, and infant mortality. Yet there were no data connecting diseases to specific pollutants, nor was there enough money to clean up all the sources of pollution at once. People recognized that their health was in danger, but they had no systematic way of looking at it, says Peter Preuss, of the EPA’s Office of Research and Development.
The Czechs adopted the set of tools Americans use to identify and rank risks, known as quantitative risk assessment. Both sides were excited by the chance to collaborate--the Czechs were getting expertise, computer programs, and technological know-how that had been created and refined as the Americans went about cleaning their own Ostravas; the Americans were getting the chance to test different ways of doing risk assessment and to do it in a less politicized arena than existed in the United States. In Ostrava, for example, they tried to use data already collected by the Czechs to come up with a comprehensive profile of the risks facing that city relatively quickly. Meanwhile, in the cities of Teplice in northern Bohemia, and Zilina in Slovakia, they started from the ground up, collecting data with their latest equipment. By comparing the results, the researchers have gotten a better sense of the extent to which different techniques can help them gauge risk.
The day for this kind of basic experimentation is over in the United States. During the last five years, as the Czech project has gotten off the ground, risk assessment has become a political lightning rod in this country. Some conservative critics have accused risk assessors of blowing minor risks out of all proportion, leading to useless regulations and wasted government spending. Accordingly, last year Republican members of Congress tried to drastically tighten the criteria for risk assessment. Before banning a chemical or forcing industry or government to pay to clean up a poisoned patch of land, regulators would have to offer extensive scientific and economic evidence that the risk was in fact worth confronting. In this way, the politicians claimed, we’d be able to cope with real threats to our health while leaving illusionary fears alone.
But though it’s under attack by conservatives, risk assessment has hardly been embraced by the liberals. Environmentalists, wary of any attempt to define acceptable limits of pollution, have branded risk assessment a stalking-horse for industrialists trying to fight off environmental regulations.
All parties concerned would do well to look across the waters at the experiences of the EPA in central Europe--particularly in Ostrava. On the one hand, they would understand that risk assessment remains a rough and imperfect science. On the other hand, they would find that even those admittedly imperfect judgments can help us solve environmental problems.
Ostrava used to be the proletarian heart of Communist Czechoslovakia--the city sat at the northern center of the country, near the Polish border and the line that in 1993 would divide the country into separate Czech and Slovak states. The Soviet tanks that crushed President Alexander Dub?cek and the Prague spring of 1968 were made of Ostrava steel. Ostrava’s metallurgical industry long antedated the Communists--it started in the early nineteenth century--but the Communists quickly and brutally expanded it when they came to power after World War II. The czars of the command economy built coke ovens and blast furnaces, and close by they erected row upon row of concrete apartment buildings for the workers. Some of Ostrava’s central district retained the stately stucco look of a nineteenth-century European city, but administrators filled the parks and buildings with sculptures of workers, mothers, and soldiers, all in the heroic style of Soviet Realism. Today these icons are forlorn with grime, ignored by the new capitalists bustling in the streets.
In the early 1990s Czechs made the epa’s brand of risk assessment their national policy. One of their first efforts to implement that policy was the program in Ostrava, dubbed Project Silesia for the region in which the city is located. The project began with Helena Cizkova, the Czech coordinator, and her colleagues rounding up as much data as possible for assessment. They looked for statistics on industrial output, pollution monitoring, incidence of disease--any germane records that might have been kept under wraps by the Communist agencies. Not very much came to light, and some of what did was unreliable. But with these scraps in hand, a screening assessment was initiated, one that was intended to be a snapshot of the environment’s effects on Ostravans’ health. Industrial Economics, a Massachusetts firm, was contracted to begin the examination of how toxic compounds could make their way to Ostravans by various avenues of contamination: through the soil, air, food supply, drinking water, and surface water.
Of the many risk assessments conducted for the screening, the researchers knew that the most critical would be of the particularly noxious pollution from the coke ovens. Coke, a key ingredient in the iron- making process, is made by heating coal to intense temperatures. The gases and smoke that are released during the heating are supposed to stay inside the ovens, but without good equipment and careful work practices, much escapes. Six of the seven coke plants in the Czech Republic are situated near Ostrava, and four are within the city limits.
You can be asleep in a car, says Cizkova, and you can recognize coming to Ostrava by the smell.
The standard procedure for assessing risks involves four steps. The first is to identify the substance in question and determine that it is indeed a health hazard. A good place to see the substance in question-- coke-oven emissions--is Vitkovice, Ostrava’s oldest and most centrally located metallurgical plant.
Vitkovice was founded in 1828 near a coal mine and was operated for more than a century by the Rothschild family until the Communists appropriated it. It is still government controlled today, but since it is unable to meet the demands of the new Czech regulations, Vitkovice’s row of coke ovens, called a battery, will shut down in two years. For now the coke making runs round the clock, 365 days a year--but not because of any great demand. Instead, like a shark that dies if it ever stops swimming, the coke batteries cannot be shut down--with the ovens operating at 2300 degrees, the battery walls would be damaged if they were cooled off and then fired up again. The ovens are parallel slots, alternating with heating flues, and at Vitkovice there are 64 ovens in the battery, each 14 feet high, 43 feet deep, and a foot wide.
Coal funnels into each oven through what look like manholes on top of the battery. When their lids are sealed, the organic compounds in the coal (about a quarter of the mass) vaporize, while some new chemicals, including poisonous pyrolytic compounds, are created. The coal gas, heavy hydrocarbons, ammonia, and tar are drawn off through pipes to a recovery plant, where the mix is cooled and cleansed of the heavier compounds. The coal gas is put back into service as a fuel, not only for the coke making but also for the iron and steel operations.
After 22 hours the oven’s two opposing doors are flung open, and from one side of the battery a ram pushes out the processed coke--more or less pure carbon now--into a V-mouthed railcar. Meeting the air, the coke instantly ignites. It makes for quite a sight when each column of coke, like an incandescent human form, stands for a moment at the oven door and then tumbles out in slow chunks, to be doused by a spray of water and disappear beneath a cloud of white steam.
As this spectacle unfolds, yellow-brown clouds are escaping. These are known as fugitive emissions, which escape from door and lid leaks. The emissions are worse as the ovens are loaded, because the organics in the fresh coal can burst free in the minute or so before the oven can be sealed. Yet Vitkovice’s battery is actually one of the cleaner ones in Ostrava because the coal is loaded through small holes on its top. Others are loaded from the side, through the very doors from which the coke is pushed out, and the large openings produce billows of pollution.
Many of the compounds in this smoke are familiar to toxicologists. One of the most prevalent and toxic is benzoapyrene, proved to cause cancer in animals and probably carcinogenic to humans as well. This substance was first isolated from coal tar in the 1940s. Coal tar, in turn, was the first environmental carcinogen to be identified by science, having caused scrotal cancer among chimney sweeps in eighteenth-century London.
In the 1960s, American epidemiologists began to collect information on the mortality of metallurgical workers in Ostrava’s American twin--Pittsburgh. They found that men who had worked on top of coke-oven batteries were seven times as likely to have died of lung cancer as metallurgical workers in general. Furthermore, the cancer rates indicated that jobs on top of the batteries were more risky than jobs on the sides. Since the emissions were greater on top, one could reasonably conclude that the likelihood of cancer depended on the amount of the exposure. On that basis one might draw a line on a graph of risk, connecting the heavily exposed workers with high mortality rates to the less-exposed ones with lower mortality rates.
To draw that line is to enter the second stage of risk assessment, which is to quantify the hazard of the identified substance. Researchers have taken the mortality data from coke-oven workers and extrapolated the risk out to the much lower exposures that people living around coke ovens experience. Theoretically, the link between exposure and cancer should extend down to vanishingly small doses, with an incidence of pollutant-caused cancer so low that the putative cases are hidden within the background incidence of the disease. Nevertheless, it is assumed that the cases do exist, and, furthermore, that there is no threshold of exposure below which cancers aren’t produced. The toll of a carcinogen on public health need not be observed to be true. A decade ago the EPA established the toll for coke-oven emissions: for each microgram of emissions per cubic meter of ambient air, the data indicated, there could be as many as 62 cancers caused over the lifetimes of 100,000 people. The formula was derived with Pittsburgh data, but it no longer holds much relevance for residents of that city. Since Pittsburgh cleaned up its coke ovens, the risk barely computes there anymore.
Step three in the determination of risk is exposure assessment-- tracking a substance from its source to its victims. The paths of a pollutant through the environment are usually too complex and costly to track--particularly in screening assessments in which many chemicals are being traced at once--so researchers resort to models. These mathematical representations, usually run on a computer, don’t begin to capture the myriad of possible combinations of chemical reactions, weather conditions, and turbulent physics that bring smoke from an Ostrava coke battery into contact with an individual’s lungs miles away. Moreover, the accuracy of dispersion models usually can’t be verified. But if real data exist-- snippets of facts usually collected at the beginning and the end of the route being modeled--the models may acquire a certain validity.
For its study, Industrial Economics selected one of the EPA’s industrial air-dispersion models and plugged in the atmospheric conditions of Ostrava. On a grid, the researchers approximated the movement of the pollutants; then, by overlaying the grid with a map of Ostrava’s population density, they could get an idea of the exposures people would experience. But before sending up the modeled smoke from the coke batteries, the analysts needed to know how much was being released. Nobody had that information--fugitive emissions readings are hard to make and are rare even in the United States. Therefore the analysts assumed for their study that the emissions were like those from poorly controlled American coke batteries. In fact, there are no such coke batteries left in the United States, but there are statistics from a 1972 University of Cincinnati study. Into the model these figures went.
The fourth and final step of risk assessment is to combine the estimated exposures with the calculated potency of the compound to come up with the probability of disease. The result is expressed as the odds of any one person’s contracting or dying of cancer, or the number of cases expected in a population. Industrial Economics presented its final numbers in a preliminary screening in May 1992. The report described the dangers posed by a long list of pollutants, from arsenic in the water Ostravans drank to PCBs in the beef they ate. However, the greatest risk by far, the reported concluded, came in the form of air pollution--specifically in the emissions from the coke ovens, as well as in particulates produced during the production of iron and steel. But the analysts did not stop with a rough ranking of comparative threats: Based on U.S. unit emission factors for coke production, their report read, coke-oven emissions account for over 90 percent of the annual population [lung] cancer risk of 384 estimated cases. This conclusion was for the whole Ostrava region; in Ostrava proper, the risk from the coke fumes was put at 190 cases of lung cancer per year. The other pollutants, even taken together, did not come close.
The announcement that coke ovens could be giving lung cancer to a couple hundred Ostravans every year caused an uproar. On the city council, enemies of heavy industry demanded that all the ovens be shut down. Changes in the political landscape strengthened their cause--a few months after the report came out, Czechoslovakia split into two. To the Czech Republic went the antiquated metallurgical complex at Ostrava. Still heady with freedom and inspired by the humanist ideals of Václav Havel, their playwright president, many Czechs scorned the black behemoths of the Communist age. Ostrava’s plants were producing more coke, iron, and steel than the new country needed, and they were doing so inefficiently and dangerously. Some of the facilities would have to be shut down. Those presenting the highest risks were obvious candidates for closure.
After the Industrial Economics report came out, it took Cizkova and her group many months to get the parties in Ostrava talking constructively again. The managers of the companies, surprised and wounded, attacked Industrial Economics’ estimate as at best unjustified, being based on unsubstantiated and outmoded assumptions--or at worst as some kind of American plot. EPA consultants who came to inspect the coke ovens were rumored to be CIA agents, stealing industrial secrets.
The industrialists made the 20-year-old American emissions study the main target of their ire, but they could just as well have attacked the dispersion model. It showed the plumes from the plants hugging the ground, thus delivering a high dose to the city residents. But the heat of the plumes--which the original model hadn’t been designed to simulate--made them rise into the air. In 1995, when researchers actually monitored the pollution at a number of points around the city, they realized their estimated figures were significantly off. Inspections of the coke batteries also revealed that the fugitive emissions were lower than assumed. With this more accurate data, the latest projection is that coke ovens cause not 190 cases of lung cancer per year in the city but rather only 4.
Today many involved with Project Silesia dismiss the initial calculation of the cancer risk. Needlessly conservative, allows Industrial Economics’ Thomas Walker. The crazy number, Jaroslav Volf, the chief health official in Ostrava, calls it.
The trouble with the number had more to do with the communication of risk than its assessment, and it was a mistake that might have been foreseen. Figures taken from American industry would inevitably lead to wide uncertainties about the actual risks in Ostrava. But Industrial Economics was initially simply trying to rank the dangers in rough order of harm; theoretically, the uncertainties would bump the dangers up (or down) together. Moreover, Project Silesia was about more than quantifying risks. It also involved risk management--reducing risks to public health based not only on the risk assessment but on political and social factors as well-- and for this work the exact numbers were not essential. How good do your data have to be to screen the worst risks? asks Jim Neumann, an analyst for Industrial Economics.
Declaring that 190 people in Ostrava get lung cancer a year from coke ovens has a precise, inescapable grimness. Yet the number had been generated from incomplete data, imprecise models, and a number of assumptions--just as in any risk assessment. Risk assessors are forever juggling unpalatable choices: the more they explain the uncertainties of their methods and their margins of error, the more they may undermine the acceptance of their results. In this case the analysts tried to have it both ways, issuing a precise-sounding number while alluding to the uncertainties, and ultimately taking refuge in the relative ranking of the threats that were examined.
Is risk assessment a science? It draws on scientific methods, but it doesn’t claim to pass tests of scientific proof. A scientific review committee would rip this report to shreds, concedes Neumann. Still, he argues that as a way to manage the environment, it is better than standing on a street corner and looking up at the pollution and estimating, because it gets you closer to a number that both government and industry can form a common ground on.
It’s not a scalpel, says Peter Preuss of the EPA. It’s a crude tool that allows you to make estimates. People don’t read the fine print when they apply risk assessment. But what’s the alternative? If you have information on a chemical and it looks like it might be dangerous, you have to do something to evaluate the risks. This technique is superior to the old ways, which led to the banning or proscribing of any chemical that was a mutagen.
What is often left out of the debate on risk assessment is that its answers are not written in stone. With better data, risk assessments move closer to reality. In the EPA’s other Czech site, in northern Bohemia, for example, researchers brought in new technology, some of which can fingerprint pollution and track it to its source. Although they haven’t been working at this site nearly as long they have in Ostrava, they’ve already shown that power plants there, which were once thought to be the worst threat to health, are not as dangerous as the fumes released by coal stoves in people’s homes. That finding has led to the local government’s giving out subsidies to encourage people to switch from coal to gas heat.
The preliminary success in northern bo-hemia has encouraged the EPA to help the Ostravans upgrade their own pollution monitoring. Researchers have set up four monitoring stations around town. In one trailer, which cowers below the towering ziggurat of the Svoboda plant, Jiri Novak of the Czech Hydrometeorological Institute points out the pattern of pollutants over the past 24 hours. Different-colored lines on a computer screen follow the levels of pollutants such as carbon monoxide, sulfur dioxide, nitrous oxides, total hydrocarbons, and suspended particulates. The hydrocarbons spiked for an hour, then dropped when the wind renewed. The spike is the mark of the coke ovens.
Meanwhile, 30 volunteers walk around Ostrava wearing miniature pump-and-filter systems. These human testers are employees of the Regional Institute of Hygiene--doctors, secretaries, technicians, and an equipment repairman--who have agreed to keep the device on or near themselves for 24 hours, four times in the course of a month, and to provide blood and urine samples each time before and afterward. Researchers will be able to track just how much pollution a person encounters and determine how much gets metabolized by the body, thereby becoming a real cancer threat. Such projects can reveal the biology and physics of pollution; they may someday offer such accuracy that risk assessment can be dispensed with altogether.
One of the Czechs being monitored is the director of the Hygiene institute himself, Jaroslav Volf. Though Volf was a key convert to the EPA program, he is not starry-eyed about risk assessment. I sometimes feel that it’s a good tool for policymakers but a weak tool for scientists, he says. For him risk assessment was effective in highlighting the threat of the coke ovens because now we can discuss how to go, to the epidemiology, or to animal models, or--he holds up the nozzle of his air pump, protruding from his jacket pocket--to science like this.
The monitoring pump in Volf’s pocket seems to chortle in agreement. Science and policy-making--it is not necessary to divide the world between the two, he continues. The artist can be in the middle. Traditionally among the Czechs the physician and the artist are very close. Sometimes I feel like I am an artist doing environmental science. Maybe risk assessment is the language connecting the two.
In the coming years the increasingly sophisticated risk assessment in Ostrava will give officials more precise information, but Project Silesia has already had an effect on the city. After much negotiation, last year the city government and the coke industry signed an agreement. Thanks in part to the project’s results, three coke plants agreed to install emission controls and make other improvements to their batteries by 1998. By this measure Project Silesia will have been a success--even if the rewards of this success cannot be measured for certain.