CASE STUDY : Comparing Apples and Supercomputers: Evaluating Environmental Risk in Silicon Valley

Community Risk Profiles: A Tool to Improve Environment and Community Health

• Summary
• Background Paper
• Evaluating Environmental Equity in Allegheny County – Theodore S. Glickman
• Comparing Apples and Supercomputers: Evaluating Environmental Risk in Silicon Valley – Lenny Siegel
• List of Participants

LENNY SIEGEL
Pacific Studies Center, Mountain View, California

Silicon Valley, California, is recognized globally as the number one address in high technology. It is not only the world’s most significant concentration of high-tech industry and talent, but the term “Silicon Valley” itself has become synonymous with a new way of doing business, based upon constant innovation and industrial flexibility. President Clinton and Vice-President Gore, among others, have embraced the Silicon Valley experience as the model for U.S. industrial growth in the new millennium.

Unfortunately, the Silicon Valley success story has been blemished by serious environmental degradation. It contains the greatest concentration of “Superfund” National Priorities List sites in the country. Industrial e ffluent from high-tech firms has contaminated the San Francisco Bay. Air pollution, principally from motor vehicles, masks the view of the mountain ranges that frame the Santa Clara Valley–the historic name of the region.

Silicon Valley is an excellent place to pioneer and test new mechanisms for community-based risk evaluation because its diverse population is aware and concerned about environmental contamination. The highly educated engineers, scientists, and health professionals that make Silicon Valley unique are in a position to choose where they live. Attracted to a climate that encourages year-round outdoor living, they feel threatened by anything, including industrial pollution, that threatens their qualify of life. The largely immigrant production and service workforce, including many who have labored on California’s farms, are sensitive to environmental threats for another reason. They have no choice but to labor in workplaces where they are daily exposed to harmful chemicals.

For statistical purposes, Silicon Valley is often defined as Santa Clara County, but in reality it corresponds to the northern urban section of the County, located at the southern end of the San Francisco Bay, as well as adjacent sections of San Mateo, Alameda, and Santa Cruz Counties. The County’s 1994 population totaled nearly 1.6 million people. County residents live in 15 cities, as well as unincorporated areas such as Stanford University. San Jose, with 836,000 residents, is today the largest city in the entire Bay Area.

Local jurisdiction over environmental issues is diffuse. Cities, private water companies, and a Valley-wide water district re all responsible for supplying water. Three of the County’s four sewage treatment plants serve multicity districts. Land use planning lies principally with the cities but the County operates bus and light rail systems.

Furthermore, Silicon Valley is inseparable from the wider Bay Area. Air pollution and water quality are regulated by regional agencies. A regional board has principal responsibility for transportation planning. Daily, some 151,000 people commute into the County. Nearly 77,000 residents commute out.

The population of Silicon Valley is, on the average, unusually well educated and wealthy. More than 12 percent of the white adults have graduate degrees and over 21 percent more have bachelor’s degrees; among Asians the rates are even higher: 16 and 24 percent, respectively. The median household income in Santa Clara County is higher than any other California metropolitan area. In fact, according to 1990 Census data, it ranked seventh in the nation, with an average annual family income approaching $54,000.

Table 1. High-Tech Industry in Silicon Valley–Mean Annual Earned Income by Gender and National Origin, 1989

(These data come from my own analysis of Public Use Microdata Sample files from the 1990 U.S. Census. I averaged the annual earned income of individual workers, not family or household income. Differences therefore could be due to three factors: 1) different pay rates (wage or salary), 2) varying lengths of workweek (hours), and 3) duration of employment during the year. The analysis does not determine whether pay differentials are due to varying educational back-grounds, differing skill levels, discriminatory employment practices, or other factors. The Census does not ask whether workers chose part-time work or voluntarily worked only a portion of the year, so one cannot determine whether any group’s low income was by choice.)

The averages conceal a sharply polarized society, however. In 1989, 31 percent of the Valley workforce, nearly 362,000 people, had annual earned incomes under $15,000. With average rents nearly $800 a month in 1990, a household earning $30,000 annually before taxes–that is, two of the $15,000 income earners–would have to devote a third of its gross income just for shelter. In practice, the working poor pool their incomes so they can stay in the Valley. The Latino, Southeast Asian, African-American, and other working poor may work on the same products or even in the same buildings as the male white and North Asian professionals who embody the Valley’s success, but they lack the income, power, and opportunity that have drawn engineers, scientists, and managers to the Valley from around the country and the world. For example, only 19.4 percent of the white men in the Valley earned less than $15,000 in 1989, while 55.6 percent of the Mexican women and 45.3 percent of Filipina women, for example, reported earnings at that level.

A SENSITIVE ENVIRONMENT

Silicon Valley’s sensitivity to industrial contamination is perhaps an accident of history. It moved from agriculture directly to high-tech, as sprawling apricot orchards were uprooted to make way for tilt-up electronics industry offices, laboratories, and factories. It is one of the nation’s leading manufacturing areas, but it never had many traditional, heavy “smokestack” industries.

Prior to the 1980s, only a handful of activists warned of the toxic threat of high tech, and we were ignored. Based upon reports from high-tech factory workers, we formed what is now the Santa Clara Center on Occupational Safety and Health (SCCOSH) in 1977. However, most political leaders, high- tech executives, the press, and the public at large all believed that high-tech was a clean industry, and they considered worker health and safety efforts to be an opportunistic attempt to unionize the nonunion electronics labor force.

Then, in January 1982, officials disclosed that six weeks earlier they had closed down a drinking water well near Fairchild Semiconductor’s south San Jose plant. Nearby residents had been drinking water contaminated with toxic solvents, including trichloroethane and dichloroethene, from a Fairchild underground storage tank.

Upon hearing the news, neighbors concluded that the company was to blame for what they believed to be an alarmingly high incidence of birth defects and miscarriages. Eventually, in 1984, the State Department of Health Services released an epidemiological study confirming that the neighborhood did indeed suffer a cluster of birth defects and miscarriages and that a somewhat larger area surrounding the plant experienced a cluster of congenital heart defects (1). In 1986, the company agreed, in an out-of-court settlement, to pay an undisclosed “multimillion-dollar” sum to more than 500 claimants from the neighborhood. Fairchild and its owner at the time, Schlumberger, have spent more than $30 million on cleanup at that facility alone.

More importantly, the Bay Area Regional Water Quality Control Board ordered that manufacturers throughout the Valley check for leaks, and most found them. There are literally hundreds of contamination sites in Silicon Valley, 29 of which are considered significant enough to be on the Superfund list. Public concern mounted. Most public water supplies in the area do not contain industrial pollution–many, however, have toxic byproducts of water purification–but large numbers of people started buying bottled water or filtering their own. Newspapers and television news programs have discovered that toxic releases in the Valley are nearly as sensational as kidnaping and murders.

When the Fairchild story hit the front page, community, labor, and environmental activists created the Silicon Valley Toxics Coalition (SVTC), which today has nearly fifteen thousand members. Originally formed as a spin- off of SCCOSH, the Toxics Coalition has from the start had strong participation from organized labor, as well as traditional conservationists in the Valley. As a result, the Coalition has consistently sought to simultaneously promote the Valley’s economic and environmental health. The Toxics Coalition not only pressed for cleanup of contaminated areas, it spearheaded a series of Valley initiatives, which continue today, to prevent future contamination.

Whereas leaders of the electronics industry have at times resisted particular regulatory proposals, they have accepted restrictions, not only because they were “caught in the act” of pollution but because their ability to recruit professionals from a global labor market depends upon the perceived quality of life in Silicon Valley. While the petroleum industry fought against local laws to control underground storage, the electronics industry lobbied to make them easier to comply with. Stung by environmentalist reports based on the Toxics Release Inventory of the U.S. Environmental Protection Agency (EPA), the Santa Clara County Manufacturing Group, the organization of the Valley’s largest manufacturers, now issues its own annual report, touting its members’ success in reducing emissions and discharges.

In 1983, a task force led by fire chiefs from the County’s cities developed a model ordinance for the storage and handling of liquid and solid hazardous materials. Adopted by all of the County’s industrial communities, the ordinance pioneered the requirement for double-containment for underground tanks and piping.

On Earth Day 1989, demonstrators demanded that IBM’s San Jose plant, the nation’s number one emitter of ozone-depleting Freon 113, cut its releases. IBM reports that it was already developing alternatives before the demonstration, but I’m convinced that public pressure helped make commercial electronics/computers one of the first industries to phase out ozone- depleting solvents.

In 1990, Santa Clara County governments developed a model ordinance for controlling lethal and other toxic gases, substances essential to semiconductor and other electronics production.

In 1992, the Air Quality district ordered United Technologies Corporation (UTC), a rocket producer with a test facility in the hills above San Jose, to phase out the open burning of waste solid rocket fuel. The principal products of rocket fuel combustion are aluminum oxide and hydrogen chloride, but long after the district had outlawed the open burning of waste agricultural products, UTC had been permitted to openly burn its toxic mixture.

In 1992, SVTC and its allies from other high-tech regions successfully urged Congress to order Sematech, the Semiconductor Manufacturing Technology Consortium, to spend one tenth of its Federal funding on pollution prevention research.

Also in 1993, local sewage treatment plants agreed to environmentalist demands for technical assistance and financing for small electronics shops–in printed circuit board fabrication and other electroplating businesses–to cut discharges of heavy metals into the San Francisco Bay.

In 1990, SVTC joined the newly formed Technical Review Committee at the Moffett Naval Air Station Superfund site, and in 1993 it received a Technical Assistance Grant covering both Moffett and the adjacent electronics industry Superfund sites. The Navy, EPA, Cal-EPA, and the electronics companies have all considered the Toxic Coalition’s participation to be constructive. In fact, the Federal Facilities Environmental Restoration Dialogue Committee–an official EPA advisory group, of which I am a member–used the Moffett experience as a model for public participation in the oversight of cleanup. That model is not only being implemented at Energy, Defense, and other federal agencies, but it has been adapted and written into the Superfund Reauthorization legislation section on public participation.

Local concern about environmental contamination has not been confi ned to releases by the electronics industry, however. In 1981, large numbers of people opposed, unsuccessfully, government plans to spray Malathion bait from helicopters to combat an infestation by the Mediterranean Fruit Fly. Most of the local governments in the area have passed stringent ordinances restricting public smoking. The Bay Area Air Quality Management District has instituted voluntary no-burn days to discourage the use of wood-burning fireplaces and stoves. Storm drains in the area are identified by “No Dumping, Empties into Bay” signs. Prompted by Valley cities, the State Department of Pesticide Regulation recently proposed rules to outlaw the sale of root-killers and cooling system biocides, the use of which causes discharges of heavy metals into the San Francisco Bay.

ASK THE COMMUNITY

Silicon Valley is probably in the forefront of local initiatives to reduce the public health threat caused by environmental contamination. Would decision- makers–private, as well as all levels of government–make better decisions if the relative risk of each form of contamination were better understood? Silicon Valley activists are skeptical. They believe that the national trend toward comparative risk assessment threatens to weaken regulation and divert resources away from environmental protection. However, I believe that trend is irreversible, so communities such as Silicon Valley must figure out how to influence the way that bureaucracies analyze risk or face a future in which their concerns are ignored.

Therefore, I propose the creation of a community task force, representing an ethnically and socially diverse cross-section of educational, economic, political, and religious leaders, to evaluate the full range of environmental threats to public health in Silicon Valley. The task force would build on past studies, as well as a wide range of medical and scientific expertise, but it would be directed by representatives of the impacted population, not by experts.

In addition to the Fairchild studies, a handful of other major studies have looked directly at the public health impact of toxic releases on Silicon Valley. In 1985, the EPA conducted an integrated environmental study of public health in Silicon Valley, and it found that smog, produced primarily by automobiles throughout the Bay Area, and treated, imported drinking water posed greater threats to public health than the risk posed by industry-contaminated groundwater (2). Considering that tens of millions of dollars had been spent confining toxic groundwater plumes, I never found that controversial result surprising.

In 1988, a state study found no statistical link between Valley birth defects and the aerial spraying of malathion (3). The Semiconductor Industry Association investigated miscarriage rates among its production workers, and in late 1992 it found a substantially increased incidence among those exposed to ethylene glycol ethers, even at legally acceptable levels (4).

I believe that a community-based environmental risk evaluation in Silicon Valley would show that an informed community has done a reasonable ad hoc job of characterizing and combatting environmental risks. A more formal process, however, would not only fine tune activities in Silicon Valley, but would help other areas identify, rate, and respond to similar threats in a comprehensive fashion. It would also help determine the different views of risk held by educationally, economically, and culturally diverse populations within the same geographic area.

It is conceivable, of course, that a formal risk evaluation process in Silicon Valley would surprise me–that is, the process might find that activists and policy-makers have been barking up the wrong tree. If so, the process must be robust enough that the result is respected and not viewed as the private tool of polluters attempting to avoid paying for their crimes. Finally, the results of the task force study could be used by decision-makers, at many levels, to develop public policy. Because the Valley’s cleanup strategy and regulatory framework have been in place for many years, it is unlikely that a new risk evaluation would make a major difference in those areas.

In the field of pollution prevention, however, a great deal remains to be done. The concept of pollution prevention is widely understood and supported within the Valley, so if funds were available–in the form of a federal grant, for example–the task force could help officials channel resources to solve problems currently identified by the community as a serious threat to environmental and public health.

STRUCTURING A STUDY

The recognized success of the Moffett Technical Review Committee, in which polluters, their public antagonists, and regulators work constructively at the same table, should help participants work together without the history of conflict and mistrust significantly delaying the process. Because the task force would be fact-finding and advisory, it would not be stealing authority away from elected officials or appointed regulators.

Each constituency should be represented by a person focused on these issues, but members need not be technical, medical, or legal experts. The task force should be informed by two sets of “experts.” First, a quasi-neutral group of scientists would act as staff to the task force. Second, representatives of advocacy groups would be invited to explain their concerns and priorities. Furthermore, other members of the public could put forward their points of view, initially and in response to interim reports from the task force, at a series of public forums.

I suggest that the task force attempt to reach consensus on a descriptive matrix, evaluating each risk against a list of criteria such as the fifteen I list below. Once the task force is established, it would be free to revise the matrix in the furtherance of its assigned task.

I have considered, and rejected, the creation of any type of overall numerical ranking system. Any formula that tries to compare diverse types of risks numerically will be unreliable and laden with hidden values–it’s like comparing Apples and supercomputers. How does one compare one cancer against 10 miscarriages or a thousand respiratory diseases? What if a response action designed to protect one species destroys the habitat for another? Is it fair to measure the risks from groundwater contamination, released without the permission or knowledge of the local community, against the risks voluntarily accepted by cleanup workers. Formulas designed to make such broad comparisons quantitatively are analogous to electoral reapportionment: everyone judges the tool against his/her preconceived preferred results.

The goal is to provide a tool to make it easier for decision-makers to consider specific courses of action based upon their jurisdiction and values. From such a matrix, a single agency–such as a regulatory body or a department with a unified budget–might effectively develop and utilize a numerical ranking system for risks or priorities, but the premature ordering of actions is counterproductive.

For example, when the city of Seattle ranked its environmental priorities, it excluded problems beyond its jurisdiction, such as worker safety and health (5). A project designed broadly to protect public health in Silicon Valley, however, would be an environmental justice debacle. Declaring the subject off limits would undermine working conditions for immigrants and other people of color at the bottom of the Valley’s industrial hierarchy. Furthermore, declaring worker issues irrelevant or unimportant would reduce the value of pollution prevention–reduced use of chemicals in high-tech production–as a solution to off-site environmental problems.

The scope of such an evaluation is somewhat arbitrary. One could consider a number of significant public health and safety threats that are not principally environmental, such as auto accidents, poor nutrition, earthquakes, family violence, or infectious disease. But that would enlarge the magnitude of the task enormously, with the results probably having little impact on what today we call environmental policy.

I suggest a focus on environmental hazards that pose a threat to public health or the survival of the remaining natural ecosystems in the area. The following is not a complete list. In fact, the task force may choose to divide the problems in a different way. But it is exemplary of the types of environmental problems that have been addressed, or at least considered, by local policy-makers.

• Agriculture chemicals (pesticides, fertilizer, etc.)
• Automotive air pollution
• Fireplace smoke
• Fish/game consumption
• Friable asbestos in buildings
• Greenhouse gas generation
• Groundwater contamination (fuel)
• Groundwater contamination (industrial)
• Industrial air pollution
• Industrial gases/accidents
• Habitat encroachment
• Lead from plumbing in drinking water
• Lead paint in residences
• Release of ozone-depleting chemicals
• Soil contamination (industrial)
• Surface water contamination
• Tobacco smoke
• Treated water

Perhaps the greatest tension in risk assessment, as traditionally practiced, is geographic. People who are directly exposed to toxic substances in their workplace or drinking water don’t like being told, by experts or bureaucrats, that their problem is minuscule in relation to environmental risks that threaten the general population, such as global warming or automotive air pollution.

Therefore, I propose an initial screening to divide the risks into two lists. One list would deal with environmental risks to the bulk of the population of the region or beyond; the other list would identify problems or sites where the receptor population is localized. On the localized impact list, each source site would be identified and evaluated individually.

Each risk would be rated according to the following criteria. If certain criteria prove confusing or impractical to apply, then the task force could make modifications.

EVALUATION CRITERIA

1. Health–Toxicity Level. How hazardous is the substance (or other hazards)? The U.S. Defense Department (DOD) has proposed a risk evaluation methodology that helps communities to transfer numerical, research-based tables into three categories: significant, moderate, and minimal. DOD defines a Contaminant Hazard Factor (CHF) that compares maximum site contaminant concentrations in each media to risk-based standards. It presents formulas that allow the evaluator to combine carcinogenic and noncarcinogenic risks from more than one substance at the same site (6). I don’t think DOD always uses the right numbers, but I find the process helpful.

2. Health–Type of risk. Although cancer risks are often used to rate the seriousness of exposures, other types of risks do not necessarily track cancer risks. It is often useful to describe the potential impact of the substance: lethal, acute, chronic, carcinogenic, mutagenic, etc. I don’t even know the term for substances, such as lead, that retard mental development.

3. Ecology–Toxicity. The Defense Department methodology calls for a separate calculation of its CHF for ecological receptors.

4. Ecology–Persistence. It isn’t only important to know the immediate impact of a contaminant. How long will it remain in the environment? Is it bioaccumulative?

5. Migration. DOD’s evaluation methodology provides simple, comprehensible definitions for three levels of pathway: evident, potential, and confined.

6. Human Receptors–Type and Quantity. Sometimes it makes sense to lump together the entire affected population, but often it is important to break it down by age (i.e., infants, elderly, etc.) ethnicity, gender, or other factors appropriate to the risk. For example, in evaluating the risk of chemical releases in semiconductor manufacturing, it is important to know how many women of childbearing age are involved. The DOD methodology, one should note, does not attempt a quantitative measure. Instead, it gives three possible Receptor Factors: identified, potential, or limited.

7. Ecological Receptors–Type and Quantity. Here, it may be more important to know what share of the population of a species or other defined group is affected. A threat to a small number of clapper rail, an endangered bird species found in Silicon Valley baylands, would be considered serious, because so few remain. The particular risks to migrating species may need to be identified.

8. Range of source. In this category, each risk is described by its breadth, not of exposure, but by where it comes from. Recall that I have proposed that two matrixes be prepared, one for risks that have a local impact, another for those of general impact. That screening would be completed before any of these criteria are applied.

Global: ozone-depletion is a good example;

Regional: such as automotive air pollution;

Multiple: underground storage tank leaks;

Unique: burning of solid rocket fuel at United Technologies;

Combined: Are there other types sources in the same area?

(This has been a concern in many poor communities, outside of Silicon Valley, where polluting industries are concentrated.).

9. Timing. When do exposures occur? Are they constant, frequent, occasional, or rare? Do they only occur during work hours? Or perhaps only during the summer? Heavy metal discharges from the electronics industry through the sewage system pose the greatest threat during the dry season (most of the time during the drought), when Bay flushing is minimal.

10. Economic Impact. Does the health threat or the activity proposed to reduce it pose a threat to jobs, taxes, or other economic objectives? Contamination of the southern San Francisco Bay has destroyed commercial shrimp harvesting. On the other side, some small electronics plants have been closed because of their toxic releases.

11. Quality of Life. Many health threats also undermine the quality of life. Visible air pollution is an obvious example. Noise pollution may have a minor impact on health, but it can be a major irritant. Some health risks themselves both contribute and detract from the quality of life, depending upon who is making the judgement. Cigarette smokers claim to enjoy smoking, although I would consider the smell of tobacco smoke a major irritant, even if it were not a threat to my health. On the other hand, I love the sight, smell, and sounds of a wood fire, even though it too may pose a health risk. In some areas–particularly recognized native lands–environmental hazards pose a threat to an entire way of life. I don’t know of any such major cultural impacts in Silicon Valley, but it’s worth exploring.

12. Jurisdiction. In deciding whether to take action to reduce a risk, decision-makers must establish both legal and practical jurisdiction. Does the body considering action have legal responsibility? Is it in a position to take action? Is it owner or responsible party or does it merely play a regulatory/oversight role? Levels of jurisdiction include local, regional, state/tribal, and federal. In some instances, international law, such as trade agreements, may also come to bear. And even if a problem is local, it may belong to or be better addressed by another local jurisdiction.

13. Causality. Can the problem be prevented, is it reversible, or is it irreversible? Is it already largely under control, or is there little hope of controlling it? (EPA does not consider groundwater contamination to be as serious as air pollution in Silicon Valley but that is largely because protective action, such as closing down wells, has already been taken.) Are we at a turning point that would magnify any action?

14. Responsibility. We tend to support action based upon our ability to assign responsibility. Can we identify the perpetrators or is the blame generic or assignable only to the taxpayer for remediation? Which of the following factors are responsible for the problem: ignorance, negligence, malevolence, or bad luck? Or is the problem natural or unavoidable? Are those people principally at risk volunteering to take known risks (such as smokers)? Are addicts volunteers?

15. Cost. Theoretically, decisions to reduce risk are supposed to be based upon the cost per unit of reduced risk. But it’s more complicated. Requiring an electronics company to clean up its groundwater may represent a slight cost for the regulatory agency, but it may represent an enormous cost for the polluter or its insurers. On the other hand, it’s easier to ask Intel or IBM to spend millions of dollars on cleanup than a small, independent factory owner. Some cleanup activities, such as source removal or containment, make sense because they reduce long-term costs. And in some instances, pollution prevention can actually save money by saving energy, water, or other materials. That is, risk reduction can actually be profitable. (More broadly, the high-tech electronics industry would benefit from the perception of a clean local environment because their chief factor in competition is the ability to attract brainpower.)

My proposed risk evaluation matrix is not simple, because I believe that simplistic analysis has been used to justify pollution or, at least, continued exposure to pollution. It is my hope, however, that the method is transparent or at least comprehensible. I have attempted to include factors that most of us use in setting environmental priorities, but which we do not always list, particularly in the context of risk evaluation.

Our political system, imperfect as it is, takes into account these external factors but not necessarily in a conscious way. By channeling input through a representative task force of people focused on these issues–not just leaders elected or appointed based on other issues–it is possible either to reach consensus on evaluated risks or at least to narrow the differences. If such a body is given access to and control over its own expert consultants, its conclusions would be more “correct,” in a Platonic sense, than if it were poorly informed.

Figure 1. San Jose Metropolitan Statistical Area Santa Clara County

Risk-Based Site Evaluation Concept Summary

1. Susan Yoachum, “How Researchers Hunted Down Birth-Defect Cluster,” San Jose Mercury News, January 21, 1985. A later state study argued that the contaminated drinking water was not the cause, however. (Mitch Benson and Rebecca Salner, “State Writes Off Tainted Water,” San Jose Mercury News, May 24, 1988.

2. Mitchel Benson and David Kutzmann, “EPA Calls Valley Water Treatment, Air Pollution the Chief Cancer Risks,” San Jose Mercury News, October 12, 1985

3. Associated Press, “Birth-Defect Tie to Medfly Spray Weak, Study Says,” San Jose Mercury News, June 8, 1988.

4. Rebecca Smith, “Chip-Making Health Risks Found,” San Jose Mercury News, December 3, 1992.

5. “Environmental Risks in Seattle: A Comparative Assessment,” City of Seattle Office for Long-Range Planning, October, 1992.

6. Risk-Based Site Evaluation Primer, Office of the Deputy Under Secretary of Defense (Environmental Security), Summer, 1994 (Draft No. 2).