Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
APPENDIX A Six Cases in Risk Analysis and Characterization r ~ his appendix presents brief ac- counts of six risk decision processes that illustrate some of the points we make about risk characterization. The examples are diverse in terms of the kinds of risk decisions and decision makers they involve and the points they illustrate about analysis and deliberation in informing risk decisions. We have chosen these six chiefly because they are not readily available in published sources. In describing these cases, we do not imply that any one of them successfully characterized the relevant risks. In- deed, some of these efforts were themselves controversial, a fact that underlines the difficulty of designing an effective analytic-deliberative process for informing risk decisions that are likely to become contentious. We believe, nevertheless, that several of the cases illustrate approaches to risk characterization that responsible agencies might find useful to adapt to suit their situations. The cases are presented in the order in which they are mentioned above: ecosystem management in South Florida, incinera- tion siting in Ohio, regulatory negotiation for a disinfectant by-products rule, siting a power plant in Florida, the California Comparative Risk Project, and future land use for a former nuclear waste site. APPLICATION OF ECOSYSTEM MANAGEMENT PRINCIPLES FOR THE SUSTAINABILITY OF SOUTH FLORIDA The US Man and the Biosphere Program (US MAB) Human-Domi- nated Systems Directorate is conducting a 4-year project on ecosystem 167
168 UNDERSTANDING RISK: INFORMING DECISIONS IN A DEMO CRITIC SOCIETY management for the sustainability of South Florida ecological and associ- ated societal systems (Harwell et al., in press). Although this effort was not envisioned as a project in risk analysis or risk characterization, it addresses important public decisions in which risks are a significant com- ponent. The project is interesting in terms of risk characterization for several reasons. One is its strong emphasis on problem formulation: the project appears to have changed the dialogue on the future of the South Florida environment by redefining the issues into an ecosystem manage- ment framework. Another is its effort to use an analytic-deliberative process to define policy goals that would in turn generate questions for analysis. A third is the project's use of a diverse group of natural and social scientists to represent the concerns of the spectrum of interested and affected parties. The project focuses on the essential issues related to achieving eco- logical sustainability for the Greater Everglades and the South Florida region and has involved more than 100 scientists representing academic and government sectors in both the natural and social sciences (Harwell and Long, 1995~. The project uses the concept of ecosystem management as the framework for harmonizing and integrating the diverse but mutu- ally dependent sustainability needs of society and the environment. This paradigm was not used in South Florida during years of large-scale ma- nipulations of the environment. Quite the contrary: during this century, South Florida has been managed to satisfy human-centered needs with little regard for the sustainability of the ecosystem. The ecosystem man- agement perspective presumes that this approach must evolve to one that explicitly recognizes the mutual interdependence of society and the envi ronment. Ecosystem management is a goal-driven framework that integrates scientific understanding of ecological relationships within societal con- texts and emphasizes the need to protect ecosystems and species of con- cern, manage for ecological fluctuations, and employ core reserve/buffer zones to protect the ecosystem over the long term. Because ecosystem management focuses on human and natural systems at regional scales and across intergenerational time periods, it is inherently integrative and adaptive in nature. The US MAB project adopted and applied a set of generic ecosystem management principles (see box, page 171~. These principles emphasize the long-term maintenance and sustainability of biological populations, ecosystem structures, functions, and processes, but also explicitly recognize that humans are an integral part of ecosys- tems. This last point cannot be overemphasized. The US MAD project seeks to replace the divisiveness created by po- larized viewpoints with a spirit of cooperation that fosters the develop
~_. ~- IIi ant!! !di~-I!ll~i~n,I~ll~olingl~l~l^ i 1!11~11 1~1 1ku4!l!~ll_dllllu~l11l~1 ~= ~l!_l~dlll~0 ~!~!~S!S!~i~!~!~ss!:~ ~!S~ss ;s~s!~!~i~sS~iS:~:~i~!~S~!~!~!~!~ IS ~!~!~!~s~!!!~!~!!! ~I1~ ~ S~-l~1~l1~1l1l4^1ll1llsl~I~Je~i~lll~l~l~s~1! 1 !~!!~!~!~ ~!~!~!~!~!~!!~!!~SS!!:S s~!!!~!!~Si~sii~!!!!~!~i!!sssis~ !!!!!~!i~!~!!~: s~!~!~!~!!~!~:~!!~:~S!~!!~!!~!~i~!s s s Hi!! s a!!!! Italic 1 := ~ I IS 1 =, I 1 I IIII=~II-? !!~!!!~!S!~!~!!~!~!!!~!~ ~!!!~i~!!:~!!~!!!~:~!~!!!!~!!!!~:~!~!~SS S S!!!ss s~!!!!!~:~S:~!!!!~!~!!~!!!~i!:~!~!~!!~!~!!!!sSs~!!~!!!!! ~!~!!~!!~!i!~!!!!~!!~!!!!~!!~!!~!!!~!~!!~i~!!!!!! I II _ 1~ I_~I~I-II lll!~ndinS l!~llll~!lll~illlllll~-llll~nt!!~umesi~ !1 ~1 ~s#~1~Ik1!1i! nh~n~111u pet Uteri b[1~ici~11 @_ l A I III ~Ih!!IIphIhb<!!!sh~ldIII ~ !~-II °n iI!g~lII ~I!I-UlI(ngII!!nawI!~!~! #~ =#~"-''' 1~-~·~='''~'"~!'~'''" .~_I:__.__^..#.-,p,~I._,.I-._I,_. mat 1~1 n~#me~pi;~ll:.dll~:: 11 11 !1 11! 11111~ ~I~-Is<-~I~!I~J!IlIIllIlIII!!II~IIl~iiIl~IIIlI ~Il~iII~I:II~l~ :::: s s ~ sass ...... S:S:S:S:S:S: ~S~!~ss~s SSSsSsSsSsS~SsS~S SS~S~S~S:S.S MISS .. ~!~!~S~s~ Isis s~sssss ....... ....... SASS ~!~ ~1~ S:':S:S:S:S:S 1 !! sssI ssssssssss~sss~s~s~sss~s ...~!~!!!!~!~!~!~!~!~!~!~!~!~ ~!~i~!~!~!~!~ SSSsSsSsSsSsS~SsSsSsSsSsSsSsSsSsS~SsSsSsSsSsSsSsSsSsSsS llT ?!~? ~!~!~!~!~!~!~!~!~!~!~!~!!~!~!~!~!~!~!!~!~!~!~ssss~sss~s~s~sssssss~s~s~s~s~sss~s~sss~sss~s~s~sssssssssss~s~sss~sssssss~sss~sssssssssssss~s~sssss~sssss~s~sssss~s~sssssssssss~ss ..~ ..~.~.~.~ .. ~? ~ . ~.~ ~.~ ~ ~.~ ~ ~( ~.~ ~. ~ .. ~ ~ ~ ~.~ .~ ~ ~ ~ ~ ~ a ~ . ~ ~ i. ~ ~.~ ~.~ ~ i ~.~ . ~ ~ ~ ~ . ~.~.~ i. ~ ~ ~.~ ~ ~ . ~ ~ ~ i. ~ ~ 1 ~!~!~is~s~ S:S:S:S.'S'S::: ~!~!~ ! ~!~!~!~!~!~!~!~ Ha!! sea ~!~ I I ~S~S~S~S~SS~S~ sss s s SSS:S:S:::S:S:: 1 sass !!!!! !~!~!~ ~!~ ~1~ ~ sssssssssssss~1 so s s sits s saws sit ~s:1 ! ~!~!~! .
1 70 UNDERSTANDING RISK: INFORMING DECISIONS IN A DEMO CRITIC SOCIETY ment of solutions that are beneficial to both society and ecosystems. The project's operating principles raise several management issues: . the need for a shared vision for ecosystem use and development- although the interests of every group cannot be accommodated, a shared vision can include access to decision-making and a common perception of broad, long-term goals for the region; · the positive linkage of the environment to sustained economic de- velopment explicit coupling of environmental and economic security rather than the more commonly held view of competition; · the imperative for adaptive management recognizing that each decision is simply the best one that can be made under present under- standing and that it can be modified and adjusted as new knowledge is gained and uncertainties are reduced; and · the need for a system of ecosystem governance suitable for imple- menting ecosystem- and landscape-level sustainability goals where the hierarchies and complexities of the natural and the human systems are recognized and are directly coupled in governance just as they are in reality. The organizers of the project presumed that an ecosystem manage- ment approach would require integration of theory and knowledge from the natural sciences with analyses of societal and ecological costs and benefits of ecosystem restoration. It would require several kinds of analy- sis: to identify the defining physical, chemical, and ecological characteris- tics of the natural, unperturbed Greater Everglades; to use these defining characteristics to develop ecological sustainability goals for the ecosys- tems of importance in the Greater Everglades; to select methods and eco- logical characteristics (called ecological endpoints) for assessing and monitoring change; to evaluate the patterns of human uses of environ- mental resources (such as land and water) and identify other human- caused stresses; to examine the societal and institutional factors influenc- ing ecological sustainability; and to assess the compatibility of ecosystem management with societal policies and institutions and the ability of these policies and institutions to achieve ecological sustainability goals. And it would require deliberation to identify the concerns of the region's inter- ested and affected parties and the social and economic outcomes or end- points that the project's analyses would have to address. All potentially interested and affected parties did not participate di- rectly, but efforts were made to have their concerns represented by in- cluding in the project a number of experts (primarily social scientists) who were sensitive to those concerns and outcomes because of having worked in and with local groups.
APPENDIX A 17 Application to South Florida The Everglades of South Florida are unique in the world, originally spanning vast open spaces between the coastal ridges of Florida, covering a total area of about 20,000 square kilometers (Bottcher and Izuno, 1994~. Wading birds, alligators, sawgrass plains, mangroves, and tropical for- ests are among the region's most recognizable features, but the essence of the Everglades is the abundance and diversity of species that once lived among the diverse range of habitats (Douglas, 1947; Davis and Ogden, 1994~. This "river of grass" flowed from Lake Okeechobee, through sawgrass, hardwood hammock, and pineland communities, to the estuar- ies of the southern tip of the peninsula (see Figure Am. The defining features of the natural Everglades consist of the large spatial scale of the system, the highly variable seasonal and interannual patterns of water storage and sheet flow across the landscape, and the very low levels of nutrients in the surface waters. These characteristics led to a unique assemblage of wading birds, large vertebrates, and fish and plant com- munities patterned in a mosaic of habitats over the landscape and sea- scape of the region (Davis and Ogden, 1994~. Since the early 1900s the regional environment has undergone exten- sive habitat degradation associated with hydrological alterations by hu- mans. These were made initially to drain land for agriculture and human settlements, and somewhat later to protect against flooding (Light and Dineen, 1994~. The resultant Central and South Florida Project of the U.S. Army Corps of Engineers has created one of the most massive engineered hydrological systems in the world. Additionally, the human population of South Florida is now 4.5 million and growing at a rate of almost 1 million per decade, mostly perched on the narrow coastal ridges. As a consequence of these changes, only half of the original Everglades re- mains, and only a mere 20 percent of the ecosystem falls within the pro- tective boundaries of the Everglades National Park. The ecosystem con- tinues to degrade, and ecological sustainability cannot be achieved without fundamental changes (Davis and Ogden, 1994; US Man and the Biosphere Program Human-Dominated Systems Directorate, 1994~. The US MAB project included nine steps (Harwell and Long, 1995~: (1) define the geographical boundaries of the regional ecological sys tem; (2) identify the types of ecosystems that exist within those boundaries and that are of management concern to humans; (3) identify the natural and anthropogenic stressors on the regional system, including their spatial and temporal components; (4) identify ecological endpoints for each ecosystem type, where eco
172 UNDERSTANDING RISK: INFORMING DECISIONS IN A DEMOCRATIC SOCIETY Scenario C US MAB Human-Dominated Systems Directorate June 1994 Isle at Hact. Mane L/\/-U Canals :~ Countes L\: Everglades National Purl; ~ Buffer Area gi4.: i.\ .... FIGURE A-1 A scenario for management of the greater Everglades ecosystem. SOURCE: Harwell, Long, Bartuska et al., in press. logical endpoints are defined as the ecological characteristics across a range of hierarchy (population, community, ecosystem, and landscape levels) that can be used to evaluate the health or change-of-health of the ecosystem;
APPENDIX A 173 (5) specify the ecological and hydrological characteristics of a sustain- able ecosystem, defined in terms of the selected ecological endpoints; (6) characterize the human factors affecting the ecosystem, including stressors, feedbacks to society, and societal values of the ecosystems; (7) define ecological sustainability goals for each component of the landscape, with focus on core areas of maximal ecological goals and buffer areas to support the attainment of those coals; tem; and ~ 1 `_ (8) establish plausible scenarios of management of the regional sys (9) examine those scenarios for their implications for the desired goals for sustainability of the regional ecological and societal systems. The study presumed that sustainability for the South Florida regional ecosystem would require the reestablishment of enough of the natural hydrological system to provide water quantity, timing, and distribution over a sufficiently large area to support the ecological components, such as wading birds and the mosaic of habitats, that constitute the essence and uniqueness of the Everglades (Harwell et al., in press). The study concluded that the environment of South Florida has more than sufficient water except in severe drought years to support all anticipated urban, agricultural, and ecological needs, but that the major portion of that fresh- water is lost directly to the sea through the engineered system of drainage canals. The critical issue then is not competition for resources, but the storage and wise management of this renewable resource. The US MAB project used the scenario-consequence analytical ap- proach to examine environmental effects of human actions. This approach involves developing a hypothetical set of conditions (scenarios) that are internally consistent and scientifically defensible and that specify all im- portant factors needed to evaluate effects (Harwell et al., 1989~. A sce- nario is meant to be neither a prediction of the future nor a proposed plan of action; rather, it is meant to cover the range of situations that are suffi- ciently plausible to warrant further evaluation. The relative risks, costs, and benefits of the plausible management strategies can then be evalu- ated comparatively, providing a much stronger basis for selecting among options. The US MAB examined three scenarios in terms of how associated changes in land use and hydrology would affect the defining characteris- tics of the natural Everglades: spatial scale, dynamic storage and sheet flow, and habitat heterogeneity. Scenarios were aggregated according to land use designations: Scenario A: management involving only existing publicly held wet- lands in the Everglades basin;
174 UNDERSTANDING RISK: INFORMING DECISIONS IN A DEMOCRATIC SOCIETY Scenario B: the addition of contiguous, privately owned areas within the original Everglades that remain as functional wetlands; Scenario C: the further addition of contiguous, privately owned ar- eas that no longer are functional wetlands, but that could provide water storage and management functions or could be recovered as functional wetlands. Land use was classified into core and buffer areas: core areas are the wetlands that would be managed to recover predrainage hydrological patterns in terms of water quantity, distribution, and timing (where maxi- mum recovery of the defining characteristics of the natural ecosystem is assumed to occur); buffer areas are parts of the system to be used for water supply for both ecological and human needs, flood protection for urban and agricultural areas, enhancement of water quality, and as hy- drological transition zones between natural conditions and managed ar- eas. Variations in scenarios were developed that differed in the location and extent of core and buffer areas. All scenarios presumed flood protec- tion and other support required for the human-occupied zone. It was concluded that Scenarios A and B contain insufficient spatial extent of the core area to provide for the defining characteristics of the Everglades at population to landscape levels or contain insufficient buffer area in order to provide for hydrological storage and release similar to natural hydrological cycles. That is, neither scenario was considered eco- logically sustainable. Scenario C (see Figure A-1) involved using portions of the Everglades Agricultural Area (EAA) for dynamic water storage while it remains en- tirely or in part under private ownership. The EAA presently consists of 280,000 hectares, primarily under sugar production, with total annual economic activity of approximately $1.2 billion (Bottcher and Izuno, 1994~. Scenario C was considered sufficient to achieve the ecological goals for the core area, but variations in the scenario relating to the amount of EAA lands that would be acquired publicly and therefore taken out of agricul- tural production had major societal implications. Complete acquisition of the EAA was concluded to have too high an economic and social cost for the communities of this historical agricul- tural area (Bottcher and Izuno, 1994~. Yet the sustainability of the sugar industry in the EAA itself is at risk for several reasons: extensive soil degradation, which has been caused by the lowering of the water table and extensive microbial oxidation and loss of the peat soils; potential changes in sugar price supports because of liberalization of international trade; likely political changes relating to Cuba that would affect current import bans on sugar; political forces aligned against sugar production in the EAA, including efforts to tax the sugar industry exclusively for funds
APPENDIX A 175 to restore the Everglades; and economic pressures to acquire EAA lands for residential development. Consequently, bringing at least part of the EAA into a buffer function (water storage and management) in support of the ecological systems of the region might counteract the risks to the sustainability of the agricul- tural system. The US MAB scenario suggested possible uses for the EAA that would allow for sugar production to continue and for the water management needs to be met, thereby linking the sustainability of the ecological system with the societal sustainability of the local community. One possibility is the development of sugar cultivars that would be highly productive under flooded conditions; another is the creation of economic incentives for water storage by sugar farmers, such as subsidizing flooded-system sugar prices or paying farmers to store water rather than grow crops. An interesting result of the analyses is that if any agriculture is to remain in the EAA, sugar is probably the most desirable ecologically, as its nutrient demands and nutrient exports to the Everglades are per- haps an order of magnitude lower than those of vegetable crops. Further- more, sugar agriculture is much preferable to the alternative of housing developments in terms of the impact on the Everglades system down- stream. Further detailed analyses remain to be done, but the risk characteriza- tion using an ecosystem management framework in the US MAB project has now suggested that a solution may be feasible that achieves ecological sustainability of the regional ecosystem and is consistent with the eco- nomic, cultural, and other societal sustainability goals for the agricultural community of the region. Implications of the Case Study The ecosystem management case study in South Florida illustrates some points that have more general importance for environmental risk characterization. First, it illustrates a way to address seemingly more complex issues than traditionally addressed by risk assessments (i.e., single chemical, single health effect) by examining the many outcomes of a few plausible scenarios as a way of understanding the risk situation. Second, it illustrates the critical role of problem formulation in the ecological risk assessment paradigm (U.S. Environmental Protection Agency, 1992a) and its constituent elements. These elements include iden- tifying the at-risk systems or populations; selecting the full range of out- comes to those at-risk systems or populations that must be characterized; identifying the types of information, analyses, analytical methodologies, and other tools needed to characterize the risk situation; making explicit the contextual issues and their implications; and mapping the problem
176 UNDERSTANDING RISK: INFORMING DECISIONS IN A DEMOCRATIC SOCIE~ onto a risk characterization landscape to provide guidance to the decision makers. Third, it illustrates the use of an adaptive management framework for risk decision making and the place of risk characterization within that approach. Adaptive management is fundamentally a problem-driven approach. It begins with explicit objectives (in this case, ecological sus- tainability that is consistent with societal sustainability); takes a long- term perspective, recognizes that long-term achievement of environmen- tal goals affects and is affected by the societal context; and adopts a policy strategy of making interim decisions, monitoring consequences, and al- tering the decisions as conditions warrant. It relies on analyses that are interdisciplinary (especially across boundaries between natural and so- cial sciences) and that focus on reducing uncertainties. It also relies on deliberation, beginning with efforts to develop a shared vision or prob- lem formulation that might be accepted by many affected parties. In the South Florida case, that effort began only after the initial analytical stages of developing options and scenarios, and it involved participation by analytical experts with varying disciplinary perspectives and familiarity with the parties' perspectives, although not by the parties themselves. It is too early to tell whether a shared vision will be widely accepted by the people of South Florida. The key result of the analysis and deliberation was a set of scenarios and consequence analyses that essentially charac- terized the risks and other outcomes of various plausible management strategies and served as input from the program to the decision partici- pants. The adaptive management framework, as applied in South Florida, faces limitations involving such issues as data needs, the potential for adversarial processes to interfere or derail the process, the perception that it is controlled by elites, and the time required for adequate development of the interdisciplinary research team. The approach is geared for ad- dressing large and complex problems, those that are multidimensional and cover a long time frame or large spatial scale. APPROVAL OF THE WASTE TECHNOLOGIES, INC. INCINERATOR AT EAST LIVERPOOL, OHIO Controversy surrounding the Waste Technologies Industries (WTI) hazardous waste incinerator, located in East Liverpool, Ohio, reflects a number of issues common to hazardous waste facility siting, and par- ticularly incinerators. Planning for the incinerator on a 20-acre plot of land next to the Ohio River, previously owned by the county port author- ity and zoned for heavy industrial development, began in 1980. A long history of permit applications, challenges, appeals, disputed approvals,
APPENDIX A 177 and legal maneuvering preceded construction work from 1990 to 1992, and assessment and dispute has continued since then. A series of air collusion exposure and risk studies were conducted during the early plan , , , . , ,- , ,~,. , ,. . ~ . · ... . ~ ~ . 1 mng period, with a new formal effort tor risk assessment 1mtlatea oy tne U.S. Environmental Protection Agency (EPA) in 1992 and continuing with review and updates to the present. These studies have been supple- mented by a series of controversial test burns that began at the end of 1992. The site has been the subject of much scientific, technical, and political debate on the safety and appropriateness of incineration, captur- ing the attention of the national media and producing impacts on national EPA policy and presidential politics. The EPA efforts at risk assessment of the WTI facility initially focused on the cancer risks to the population and to the hypothetical maximally exposed individual that would be associated with permitted, routine air releases from the incinerator stack and subsequent exposure through in- halation and ingestion of water and locally grown food, including meat and dairy products. Contaminants of concern included dioxin and other stack emissions. More recently, following review, the risk assessment was broadened to include health, safety, and ecological concerns, includ- ing releases during startup, shutdown and malfunction/upset conditions, and on-site and off-site transportation accident risks (U.S. Environmental Protection Agency, 1993a; A.T. Kearney, Inc., 1993; Johnson, 1996~. Draft results of the risk assessment, still undergoing peer review, suggest that cancer risks are less than 1 in 106 for a nearby resident, and that noncancer health risks and risks from accidents and ecological impact are low and acceptable Johnson, 1996~. These results, and the belief by many that the WTI facility is "world class," with the best state-of-the-art pollution con- trol equipment available, form the basis for arguments by proponents that continued operation will be both safe and profitable. Yet significant controvery remains. For many, the debate over the specifics of the WTI risk assessment has served only as a surrogate for continuing societal concerns over broader issues, including the role and suitability of incineration as a haz- ardous waste management option. The uncertainy and controversy sur- rounding dioxin has also affected the WTI risk debate, as it has at other incinerator sites. Beyond this, some people believe that approving the operation of any hazardous waste management technology provides an incentive for continued production and distribution of chlorinated or other potentially hazardous compounds. These systematic concerns ebb and flow into any local debate over incineration, but they are often pushed off the screen in formal risk assessment and risk characterizations for a proposed project. The current risk assessment efforts are for a facility already built and
178 UNDERSTANDING RISK: INFORMING DECISIONS IN A DEMOCRATIC SOCIETY undergoing test burns. Indeed, proponents of incineration maintain that this is the only way to conduct an effective risk assessment: site-specific burn efficiencies and operating parameters are needed to effectively evalu- ate emission rates. However, opponents argue that such a procedure, allowing extensive costs to be sunk into the construction and testing of a facility prior to the risk assessment, virtually ensures that approval will be granted. Effective and cooperative stakeholder participation in the risk analysis process is indeed difficult when some participants feel that the decision has already been made that it is "a done deal" and that the results of the risk analysis must perforce conform to this decision. This timing virtually assures that many will view the risk assessment as inherently biased or irrelevant. A number of participants in the WTI debate have noted the need for inclusion of a broader range of considerations in the risk evaluations. In particular, they have argued that the assessment of risks to a population due to a single source cannot be conducted absent the context of other risks imposed on that population and that particular subgroups within the population may be subjected to an especially high combination of exposures with implications for both health risk calculations and environ- mental equity concerns. These issues are often raised as part of the broader agenda of environmental groups in debates over hazardous waste facilities. In public comments during the December 1993 EPA review meeting of the planned risk assessment for WTI, Rick Hind, representing Greenpeace, posed the following questions (U.S. Environmental Protec- tion Agency, 1993a:Appendix H. Observer Technical Questions): When assessing risks to sensitive subpopulations near WTI, how will you consider the effects and sensitivity (body burden) on the African American community closest to WTI? How will you address WTI risk posed to these and other children living and attending school (nearby to the site)? Will you look at body burden already found in these subpop- ulations from other sources including possible immune suppression from toxic chemical pollution and radiation? East Liverpool resident Marilyn Parkes further noted the need to consider the WTI emissions in the context of other industrial exposure sources, including those from an area power company that had purchased excess emission permits from another region. These requests for more detailed analysis raise legitimate issues. For some of those who raise them, how- ever, doing so may serve to advance in-principle objections to incinera- tion that are allowed no legitimate place in the discussion when the prob- lem is formulated in terms of risk assessment for a particular incinerator (see lasanoff, 1986). Another strong current in the WTI debate is the need to address local
APPENDIX A 179 public health concerns by monitoring measurable environmental out- comes in the community, rather than using only theoretical, model-based risk predictions. In a letter dated December 6, 1993, to the EPA Region 5 Administrator, the City of East Liverpool Board of Health states (U.S. Environmental Protection Agency, 1993a:Appendix A: The only real data which EPA proposes to use in the Phase II Risk As- sessment involve trial burn emission data. This is not sufficient to gain public confidence in East Liverpool.... Community monitoring pro- grams need to be expanded in scope and frequency. This must include monitoring of air, soil and crops (i.e. foodchain) at numerous sites with- in the community.... Any assessment that involves only theoretical assumptions on com- munity exposures stands little chance of gaining public support. The East Liverpool Board of Health advocates the development of an ex- panded and accelerated emission and environmental monitoring pro- gram rather than a quantitative risk assessment based on little real data. The East Liverpool Board of Health encourages cooperation with Feder- al EPA in further development of specific data collection programs which can provide a realistic basis for protecting public health. These comments illustrate the difference between a public health paradigm (Burke, 1995) and that implicit in the current prevailing ap- proach to scientific risk assessment. Such concerns have remained de- spite the fact that the scientific studies in support of the risk analysis were detailed and extensive. Indeed, the extent of mistrust in risk assessment and those performing it has been indicated by local residents' suggesting "the possibility of scientists 'fudging' risk assessment data" (Johnson, 1996~. The long-term impacts of continued operation of the WTI incinera- tor may not be known for many years. Assuming the results of the cur- rent risk assessments are representative and correct, the facility could provide a safe, effective, and economically beneficial mechanism for haz- ardous waste management. It is clear, however, that the concerns of all participants have not as yet been addressed by these assessments, and for many, the full dimensions of risk from the WTI facility have yet to be characterized. REGULATORY NEGOTIATION FOR A DISINFECTANT BY-PRODUCTS RULE Regulatory negotiation is a process in which representatives of for- mal stakeholder groups work consensually with government regulatory bodies to draft a proposed rule (see Appendix B for more detail). Regulat- ing negotiation can provide a way in which analysis and deliberation can be coordinated to inform risk decisions. Without endorsing the specific
180 UNDERSTANDING RISK: INFORMING DECISIONS INA DEMO CRITIC SOCKET process used in this case study, we present it in detail as an illustration of how respect for a few very simple principles can help characterize risks in a way that meets the needs of public officials and the interested and affected parties. Chlorination of drinking water is the most common technique used by water suppliers to reduce the risk of microbial infectious pathogens. However, although chlorine and other disinfectants do reduce microbial risks, they also react with organic compounds already present in the wa- ter- disinfectant by-products (DBPs) some of which are carcinogens. It is these by-products that EPA wanted to regulate. Under the Safe Drink- ing Water Act, EPA was obliged to set standards for drinking water con- taminants. The proposed DBP rule represents a portion of this effort. EPA turned to regulatory negotiation because any DBP regulation would have been contentious using the customary rule-making process, for two main reasons. First, EPA had previously concluded that the avail- able data were inadequate to address concerns about DBP and microbial risks associated with disinfection of drinking water. Any proposed rule would have to invoke judgments that EPA did not believe it could ad- equately substantiate with evidence. Second, the Safe Drinking Water Act was soon to come before Congress for reauthorization. The water supply industry was already on a collision course with environmentalists and consumer advocates. The water industry hoped to gain relief from the additional and, in its view, unreasonable regulatory burdens de- manded in the act and environmentalists wanted to preserve the act's central mandate. Meanwhile, EPA had already entered into a consent order with a litigant, obliging itself to propose a DBP rule by June 1994. Regulatory negotiation offered a way to bring would-be adversaries to- gether, to give them a chance to speak with each other, to work with EPA to devise a proposed rule, and to give all parties involved a chance to recognize that the conflict over drinking water regulation was not likely to be won by any single party, but, instead, might be resolved by coopera tion and compromise. In September 1992 EPA announced its intention to organize a regula- tory negotiation rule-making process to develop a DBP rule. Between November 1992 and June 1993 a negotiating committee met eight times, reviewed evidence, ordered analyses, negotiated, and developed pro ~ ... posed rules that all participants might accept. It consolidated 16 policy options to 3 and then asked a technical assistance group to perform cost- benefit analyses and compliance assessments for the 3 options. The nego- tiating committee resolved disagreements and issues of uncertainty in creative and unexpected ways. For instance, instead of producing a single DBP proposed rule, the committee proposed three rules. Only one ad- dressed DBP regulation directly. Furthermore, the proposed rules made
APPENDIX A 181 allowances for not burdening small water supply systems with complex and costly measurements and treatment technologies by arranging for regulations to be phased in over time, and dealt with the issue of informa- tion gaps by proposing that DBP rules be refined as more data becomes available. Forming the Negotiating Committee and Technical Support Under the Negotiated Rulemaking Act of 1990, certain criteria are used to establish if regulatory negotiation would be productive and worthwhile. A main objective is to determine how feasible it is to com- pose a negotiating committee that adequately represents the full spec- trum of interests on the issue. In this case, EPA hired an outside firm, which conducted more than 40 interviews with agencies, water suppliers, environmentalists, equipment manufacturers, and consumer groups. From these interviews it was established that the number of affected in- terest positions was relatively small, the factual base for holding delibera- tions was well developed, there was a strong degree of "good faith inter- est" in resolving the issue through negotiation, and the agency was willing to commit the necessary resources. These results satisfied the agency's criteria for initiating regulatory negotiation. On the basis of the preliminary interviews, EPA proposed a negotiat- ing committee of 17 individuals. After public review, one more member was added. A dispute arose about whether or not chlorine chemical and treatment equipment suppliers deserved a seat at the table. Clearly these stakeholders had valuable knowledge and advice to offer, but they were also strongly committed to the use of chlorine as the primary drinking water disinfectant. A creative compromise was struck. The chlorine industry would not be granted a seat on the negotiating committee, but would instead participate in an advisory role by serving on the Technolo- gies Working Group (TWG). A representative was invited to state the group's position at the very first negotiating session. The TWG provided a formal opportunity for chemical and equip- ment suppliers who had not been named to the negotiating committee to contribute input to the negotiations, while also supplying much-needed information about the cost and performance of drinking water treatment technologies. In addition, EPA arranged for three experts to provide ongoing scientific advice and technical support for those members of the negotiating committee whose organizations were not able to provide this kind of expertise.
182 UNDERSTANDING RISK: INFORMING DECISIONS INA DEMO CRITIC SOCIETY Defining the Problem and Informational Needs In all, six negotiating meetings were planned. The general strategy was to follow the following sequential steps: gather information, develop evaluative criteria, generate options, evaluate options, draft agreement, and obtain closure. In practice, the process was much more iterative and integrated. For example, the committee defined criteria at the first and second meetings, gathered information throughout the entire process, and reached closure on some items earlier than others. It took eight meetings to reach partial closure. Over the next 8 months, participants did not meet, but drafted rules, reviewed them, and finally gave ap- proval. EPA initially framed the problem in a 1991 EPA Status Report (59FR 38675) that mentioned the need to regulate DBPs by assigning maximum contaminant levels for individual compounds or for groups of com- pounds. It also framed the inherent problem as a risk-risk tradeoff. Major difficulties were expected to be the reduction of DBP risk without a simul- taneous increase in microbial risk and the introduction of new or un- known risks from changes in treatment technology. In the status report, EPA clearly stated a position against adoption of a rule that would pro- mote a shift to non-chlorine disinfectant technologies. The report sug- gested recommending continued use of chlorine disinfectant by specify- ing that as the preferred choice of technology best available technology or BAT. When the negotiations began, not all members of the negotiating com- mittee shared the opinions expressed in the 1991 status report. There were substantially different perceptions of the problem among members of the committee, including differing perceptions of the risks associated with DBPs. Many members expressed the opinion that available scientific and technical information was inadequate to justify proposing specific rules. Some members believed that only slight improvements in disinfec- tant technologies were justifiable until more became known about the relative risks. Others believed that it would be best to specify treatment techniques and not specific contaminant levels. Still others advocated a radically new approach of regulating DBP precursors, namely the elimi- nation of organic material before treatment through a combination of filtering and enhanced watershed protection. The negotiating committee was not free to define the problem as it wished because it was constrained by law. The Safe Drinking Water Act requires that contaminants be regulated with maximum concentration levels. Treatment techniques can be the basis of a regulation, but only when it is not economically or technologically feasible to measure the contaminant. Watershed protection was an attractive target for the com
APPENDIX A 183 mittee, but it is covered under a different rule (the enhanced surface water treatment rule). The committee frequently found these constraints to be barriers to a creative agreement and did not always accept them unquestioningly. In fact, the committee pursued several alternatives that seemed to run contrary to the letter of the law. For example, the commit- tee asserted that DBP rulemaking should be consistent with other rules, and it proposed a new enhanced surface water treatment rule as well as the DBP rule. The negotiating committee did not begin by attempting to define the problem. Instead, it sought to define the characteristics of a good solu- tion. One of the first tasks was to define the value objectives of a "good" rule. In facilitated open discussions, the committee produced an unranked listing of value objectives to be considered in the decision-making pro- cess. The list included whether a potential rule was protective of human health, flexible to source water quality and existing treatment facilities, protective of environmental equity, sensitive to needs of susceptible popu- lations, consistent with EPA rules, explainable to the public, and afford- able. These criteria would be used in later sessions to evaluate competing proposals. At the first formal negotiating session, most participants agreed that some type of DBP rule was needed. There was less acceptance of EPA's definition of the problem as a need for maximum contaminant levels. Some believed that the committee should consider a pollution prevention approach that aimed at eliminating the DBP precursors. (Precursors are naturally occuring humic and fulvic acids that react with disinfectants to produce dangerous DBPs.) Theoretically, most of the DBP problem would disappear if precursors could be removed before treatment. Precursors can be reduced though watershed protection measures (which bring other additional benefits), granular activated carbon, membrane filtration, or enhanced coagulation. The participants who defined the problem in this manner advocated the development of a rule to enhance the existing surface water treatment rule (which is basically a watershed protection rule). The committee early recognized the need for more data and analysis to inform its discussion of the competing problem formulations. To ad- dress these needs, the committee consulted with the TWG in order to specify its informational needs. A technical workshop was organized for the purpose of informing negotiating committee members on the range of scientific opinions about health risks, treatment technologies, costs, and modeling efforts. Twenty-three nationally recognized experts on drink- ing water treatment gave presentations and participated in panel discus- sions for the benefit of the negotiating committee. As questions arose
184 UNDERSTANDING RISK: INFORMING DECISIONS INA DEMO CK TIC SOCIETY during the following months, additional presentations or testimony were given to the committee. Integrating Deliberation and Analysis Throughout the negotiations, the negotiating committee requested studies, evaluations, data, and advice from the TWG. At the very first negotiating session, for example, the TWG was asked to organize the available information about treatment technologies. One participant sug- gested the results be presented as a matrix comparing treatment alterna- tives along several performance criteria. These criteria were defined by the negotiating committee only in the vaguest sense. It remained for the TWG to specify the criteria, to develop indicators, and to make assess- ments on the measures. Three weeks later, at the second negotiating session, the TWG presented a schema of 31 treatment scenarios, a stan- dardized coding form, and results for 10 completed evaluations. Negotiating sessions usually began with members of the TWG re- sponding to requests from the previous session. Often, these exchanges raised more questions than they answered and resulted in more tasks being assigned to the TWG. For example, after receiving the analysis of cost to each household, the negotiating committee then asked the TWG to estimate the national costs of each treatment technology. The TWG re- sponded that that question could not be answered without first gathering more baseline information about the quality of pretreated water across the nation. This need for more baseline information would arise repeat- edly throughout the negotiations. In some instances, the TWG was able to provide clarification that helped resolve misunderstandings or disagreements among the negotiat- ing committee. One definitional issue that arose concerned a technology known as granular activated carbon. Different committee members had different understandings of what the technology was and what it was capable of accomplishing. A TWG member cleared up the misunder- standing by explaining that there were basically two major types of granu- lar activated carbon systems. Later, the negotiating committee would ask the TWG to write an official definition of the two types of systems to resolve the confusion. The TWG also performed risk analyses and cost- benefit and other economic analyses at the request of the negotiating committee. At times, members of the committee were assigned to work with the TWG in these analyses. At the third negotiating session, the committee asked the TWG to develop information on the distribution of source water characteristics across the nation. This information was needed to satisfy the objective of environmental equity, giving each community the same level of risk pro
APPENDIX A 185 section. Rules that require 99.99 percent reduction of viruses do not pro- vide equivalent risk protection when source contamination varies over several magnitudes. The negotiating committee wanted to know how much variation in source water contamination existed and how that was distributed among different communities. Most of the TWGs effort went into computing cost and compliance figures for the different approaches. Sensitivity analyses were also re- quested. These data and analyses were needed to meet the stated objec- tive of an affordable rule. Cost analyses that were originally ordered in January were presented at the next meeting in February and revised, and more sophisticated results were presented at later meetings. This cycle of request/report/revise/report was characteristic of this regulatory nego- tiation process. Information and results of analyses from the TWG were also used by the negotiating committee to help resolve fundamental differences in regulatory approaches. For example, by the third session, the committee was considering three basic approaches. A pollution prevention approach involved reducing the entry of DBP precursors into the water treatment system or removing them before the disinfectant was added. In order to investigate the hypothesis that precursor treatment could provide ad- equate risk protection, the committee asked the TWG to analyze the effect of precursor content on DBP levels in treated water. The TWG had to address the question of which variables to measure in the untreated wa- ter. A Creative Solution to Handling Uncertainty During the deliberations, the committee determined that there was still insufficient information available to make many appropriate deci- sions. Specifically, it needed more field information about the existing nature of risks from microbial pathogens and the capabilities of different types of treatment technologies to control those pathogens. The negotiat- ing committee responded to this need by preparing, with the assistance of the TWG, an Information Collection Rule (59 FR 6332 [February 10, 1994~. The proposed rule requires large public water supply systems to monitor source water for microorganisms and viruses, test for the pres- ence of DBP precursors before and after treatment, test for DBPs, and provide engineering information about treatment techniques. In some instances, systems reporting high levels of precursors would be required to conduct special pilot tests to reduce them. This amounts to a pretest of the field performance of treatment technologies being considered for the DBP rule. The purpose of the pilot testing is to obtain more information about cost effectiveness of precursor control technologies. Language in
186 UNDERSTANDING RISK: INFORMING DECISIONS IN A D~MOC^TIC SOCIETY the proposed rule specifies minute technical details referring to the de- scription of acceptable pilot technologies, their implementation, and test- ing and reporting protocols. The Information Collection Rule represents the breakthrough com- promise that made the proposed DBP rule politically feasible. Although there was wide agreement at the very start of the process that a DBP rule was needed, informational gaps stood in the way of resolution and final agreement. One thing the committee did agree on was the need to reduce uncertainties by gathering more data. Data collected under the rule are to be used to inform the final adjustments of the proposed rules and the need for and content of long-term rules. Rather than put off the entire rule-making process until after all rel- evant information had been collected, the negotiating committee decided to agree on provisional rules with the understanding that a second stage of negotiations would be necessary some few years in the future. Stage 2 rules would be revised on the basis of the data gathered in the meantime. Committee members advocating stricter rules demanded that Stage 2 "straw man" provisions be set, and set intentionally low in order to en- sure that all committee members would return for those negotiations. In addition the committee announced it would meet again should addi- tional information become available on acute health effects that would necessitate more expeditious changes in rules. Conclusions The DBP regulatory negotiation represents a creative approach for dealing with conflict and uncertainty by reducing uncertainty over time. It also illustrates how regulatory decisions can benefit from processes that promote meaningful deliberation among experts, public officials, and in- terested and affected parties. A variety of deliberation techniques were used to help build mutual understanding and consensus among the nego- tiating committee members. Deliberation was integrated with analysis through requests to and reports from a technical advisory committee. The DBP regulatory negotiation was a success in the sense that it resulted in the publication of three proposed rules, including the DBP rule, that are in the process of final review. Before the process began, there was considerable antagonism among the stakeholders in the pro- cess. EPA was clearly obligated to propose a rule, but it expected that any such rule would be challenged in the courts. Meanwhile, the Safe Drink- ing Water Act was scheduled to be reauthorized by Congress, and it appeared possible that Congress would entertain proposals to alter the act. Several parties were hopeful that Congress would amend it so as to not require such demanding regulation. For all these reasons, the general
APPENDIX A 187 expectation was that a DBP rule, if passed, would probably not be very comprehensive. The negotiation also led to a good-faith effort to cooperate in funding research into disinfectant by-products and their risks. The American Wa- ter Works Association Research Foundation committed several million dollars, which was matched by EPA and further supplemented by Con- gress. The consensus is that this money and cooperation will bring about a lot of good, relevant research that otherwise would never have been done. One member of the negotiating committee, the representative of small, mostly rural water suppliers, did drop out of the process when he de- cided that he would be able to protect his group's interests better by acting outside the negotiation. He concluded that the group's concerns for high costs of new regulation could best be served by appealing di- rectly to Congress. This group never acknowledged there was a need for DBP regulation. Once its proposals were rejected by the other committee members, its representative withdrew. The groups efforts to influence Congress apparently bore fruit in 1994 in special amendments in the Sen- ate to the Safe Drinking Water Reauthorization that eased the regulatory burden on small water suppliers. (The bill did not clear the conference process because of unrelated issues.) As of early 1996, the small water suppliers had not commented on the proposed rules or challenged them in court. The disinfectant by-products regulatory negotiation provides an ex- ample of how analysis and deliberation can be coordinated in making risk decisions. Throughout the negotiations, committee members repeatedly drew upon technical advisors to interpret and analyze data, to model risk scenarios, and to estimate costs and feasibility of regulatory options. Three positive aspects of this case are worth reiterating: the use of a tech- nical advisory committee and expert assistance; the manner in which the negotiating committee dealt with the issue of data inadequacies; and the iterative nature of the negotiations and rule-making process. In addition, the process proved widely acceptable to almost all of the interested and affected parties (U.S. Environmental Protection Agency, 1992c, 1992d, 1992e, 1993b, 1993c, 1993d, 1993e, 1993f, 1993g).1 (For a more detailed account of the role of analysis in the regulatory negotiation, see Roberson et al., 1995.) It is also worth emphasizing one negative aspect: the possi- bility that a single party may be able to trump the entire process of analy- sis and deliberation with an intervention in the political arena. This pos 1See also Federal Register 59(28):6332-6444 (Feb. 10, 1994); 59(145):38668-38829 July 29, 1994); 59(145):38832-38858 (July 29, 1994).
188 UNDERSTANDING RISK: INFORMING DECISIONS IN A DEMOCRATIC SOCIETY sibility casts a shadow over the entire process of risk characterization and decision making in government agencies. It is especially significant for approaches that make special efforts to broaden participation, because the success of an interest group in going outside a deliberative process can increase alienation and mistrust on the part of participants who were coaxed, perhaps reluctantly, to participate in a process they did not ini- tially fully trust. SITING A POWER PLANT WITH THE AID OF DECISION ANALYSIS TOOLS In the late 1980s, the Florida Power Corporation (FPC) began a search for one preferred and two alternate sites for a 3,000 megawatt coal-fired power plant. Knowing that public scrutiny would be keen, FPC decided to avoid the traditional failings of a decide-announce-defend approach to siting by drawing multiple perspectives into the site selection process early on. Two groups a committee made up of FPC managers and staff and an Environmental Advisory Group (EAG) composed of eight high- profile figures in Florida engaged in a decision-making process that employed multiattribute utility analysis to help make the issues transpar- ent and logical and to focus the debate in a constructive manner. By all apparent measures, the process was successful. The EAG had clear and specific influence in the site selection process, and licensing is under way without opposition at the selected site. Multiattribute Utility Analysis Multiattribute utility analysis is a technique for formally drawing multiple perspectives and evaluations into a decision making process. It begins with the working assumption that decisions have multiple goals, not all of which are equally important. Differences are attributed to the particular interests of the interested and affected parties and the decision makers. The technique makes these differences (and similarities) lucid by eliciting from participants their subjective judgments about the impor- tance of outcomes (subjective utilities) and using these as a basis for com- parison. The goal is not to reach a forced "consensus" through averaging, but to clarify positions and to test the feasibility of various policy objec tives. A typical approach in multiattribute utility analysis begins by asking each participant or group of participants to list and set priorities among the criteria that they would use to evaluate the decision options. For example, in a siting case, one person may value visibility of the facility from the town center as paramount, while another may emphasize noise
APPENDIX A 189 or emission levels. With the help of a decision analyst, these criteria ar organized into a hierarchy called a "value tree." Value trees can be com- posed for individuals and for groups. The next step is to assign relative weights to each criterion. As the criteria are arranged hierarchically, each "branch" is given a relative weighting. For example, all criteria having to do with aesthetics might be grouped as one branch of the tree, while another branch may include all criteria concerning emissions. There is no single right way to group or weight the criteria; rather, the goal is to make certain that the result accurately reflects the concerns and judgments of the participants. "Twigs" on each branch are indicators or measures of how well a decision option performs on that criterion. Overall scores for decision options from the perspective of a participant can be computed by first applying the indicators and then summing each weighted branch using simple algebra. Multiattribute utility analysis requires coordination between delib- eration and analysis at two levels. First, the computation of scores, an analytic procedure, follows deliberation in which values are elicited, dis- cussed, peer-tested, and revised. Deliberation plays a major role when this analysis deals with groups of people rather than individuals. For example, in the Florida Power Corporation case, two different groups did separate analyses, and in each group, the members deliberated to reach agreement on the evaluative criteria and their relative weightings. The different results allowed for further interaction and learning. Analysis and deliberation are also intertwined in quantification. De- pending on the nature of the indicators and the available data, measures may be made by technical procedures (such as measuring the permeabil- ity of subsurface soils), or quantification that may be highly subjective (such as nuisance from additional traffic). Analysis provides the former measures; deliberation arrives at ways to quantify factors that are not readily measured. Deliberation is also important in reading and inter- preting numbers that combine such disparate kinds of indicators. The Decision Process Florida Power Corporation was assisted by consultants specializing in decision analysis techniques. Their task was to elicit evaluative criteria from the FPC group and the EAG, operating separately, and then to help the groups come to consensus on how to weight the criteria relative to each other. This was accomplished in a step-wise manner, beginning with a large search area and a few general criteria and working toward more specific criteria and a narrower search area. In between EAG meet- ings, the consultants would apply the criteria to the potential sites. The results were presented at the following meeting. Exclusionary and screen
190 UNDERSTANDING RISK: INFORMING DECISIONS INA DEMOCRATIC SOCIETY ing criteria were used to eliminate unworthy sites from future consider- ation. Potentially suitable sites were ranked using each group's weighted criteria. Results from the FPC group were shared with the EAG and vice versa. As a result, the FPC group reconsidered certain weightings, and the two groups' judgments came closer together as the project progressed. The final site chosen was highly recommended by both groups. The siting study was conducted in five phases and took more than 18 months to complete. Each phase was based on successively more refined data and criteria. Input from the EAG was used only in Phases II, IV, and V, due to limitations on the amount of work the EAG was able to under- take. It is in these phases where the interplay of analysis and deliberation can be seen most clearly. The EAG, as noted above, was made of eight high-profile individuals, including local and state government officials, academics, an environ- mental attorney, and representatives of environmental and business groups. Its purpose was to bring multiple values and perspectives into the siting process early on. This approach demanded significant contri- butions from the sponsoring company and the individuals who agreed to serve on the EAG. Over a year's time, the EAG convened regularly for all-day meetings. Ideal MUA procedure would be to elicit the evaluative criteria from the EAG and the FPC group from scratch. Because of time constraints the consultants, working with staff from FPC, developed a preliminary set of working criteria and measurement indicators. These were used as a start- ing point for the EAG. Using a structured elicitation protocol, the con- sultants' first step in Phase II was to have the EAG review and modify the preliminary list to better suit their needs. Phase II proceeded in six steps: 1. The consultants elicited a set of criteria and weights from the FPC panel. 2. Using the FPC criteria (though not its weights) as a starting point, a set of criteria and weights was elicited from the EAG; see Table A-1. The EAG supplemented its list with justifications and explanations. 3. The FPC group reviewed the results from the EAG and modified its criteria set to match the EAG set. As Table A-1 shows, it also modified many weights in the direction of the EAG weights. Wetlands disruption was revised from a weight of 54 to 14, very close to the EAG's value of 15. Based on argument given by the EAG, the FPC group decided to prefer sites that were already near a disturbed area, and they added this new criterion to their list. The FPC group then increased weighting on several other criteria (effects on county tax base, disturbed areas, sensitive land uses, surface water, and terrestrial communities), bringing them closer to
APPENDIX A TABLE A-1 Criteria and Weighting for Power Plant Site Selection 191 FPC Initial EAG FPC Final CriteriaWeight Weight Weight Impact on Surface Water Quality11 15 19 Wetlands Disruption54 15 14 Proximity to Disturbed Areasnot used 19 11 Proximity to Sensitive Terrestrial Communities6 15 11 Proximity to Sensitive Land Uses6 15 11 Proximity to Class I Air Quality Areas10 12 11 Impact on County Tax Base3 10 11 Proximity to Urbanized Areas6 0 11 Existing Air Sources6 0 0 Total102 101 99 the EAG weights. On only one criterion, proximity to urbanized area, did the FPC group's weighting grow further apart from the EAG's. 4. The consultants used the EAG criteria and weights to rank the 172 areas and select 59 sites. 5. The consultants used the FPC criteria and weights to rank the 172 areas and select 57 sites. 6. The consultants combined the two sets of selected sites and set a cutoff threshold in collaboration with FPC staff and after considering graphical plots of the site evaluations. Because of the FPC group's shift in weights, there was agreement on 55 areas sites: 4 were selected only by the EAG and 2 were selected only by the FPC group. All 61 sites were sent to Phase III for further consideration. The single criterion added by the EAG was quite significant. FPC's set included criteria that selected for sites away from people and areas of development. Fearing that a remote site would promote sprawl, the EAG advocated sites on already "disturbed" land near, but not directly on, existing settlements, which could house and support the plant's employ- ees. After much deliberation, the EAG discovered it could capture these concerns with a criterion called "Proximity to Disturbed Areas." That criterion was the highest weighted by the EAG and the third highest
192 UNDERSTANDING RISK: INFORMING DECISIONS IN A DEMOCRATIC SOCIETY weighted (ultimately) by the FPC. The effect of combining that criterion with the others favoring sites far from people was to select for remote, but not environmentally attractive areas. In fact, the site ultimately selected was just such a site. In sum, the EAG affected the Phase II ranking in three ways. First, it added an important criterion that FPC had not originally included. Sec- ond, it assigned different weights among the criteria than did the FPC group, and the resulting different rankings were used to select a different set of sites than would have been selected by the FPC group alone. Third, it encouraged the FPC group to modify its criteria set to match the EAG set and to modify its weights in the direction of the EAG weights. Those modified criteria and weights were used in the FPC ranking and selection of potential sites for Phase III. The EAG role in the Phase IV ranking and Phase V selection closely paralleled its role in Phase II. The EAG added two criteria in Phase IV: "Potential for Broader Purpose" and "Regional Urbanization." It recog- nized that at some sites in particular, the power plant could fulfill "broader purposes," such as: the use of damaged land that would other- wise remain a blight; the basis for conservation easements or other forms of land preservation; the use of waste for fuel or sewage for cooling; and heat cogeneration. A three-point scale was developed for rating potential for broader purpose. The EAG also recognized that a power plant could be a large burden on the infrastructure of a remote area. Therefore, it suggested a criterion favoring sites with development near enough to provide infrastructure support. The measure selected was size of popula- tion within 9 to 25 miles. Phases III and IV reduced the number of sites from 61 to 21 to 6. In Phase V, each group confirmed the ranking of the six candidate sites and then selected the preferred and two alternate sites on the basis ofthat ranking. Rankings in Phases IV and NT were done with 11 criteria that were elicited, along with their weights, jointly from the FPC group and the EAG. Additional Features of the Case The iterative process used in this instance had several benefits. First, it allowed an efficient targeting and use of data collection and analysis resources. The entire study area was first examined at a coarse scale of data, and more promising potential sites were examined at a finer level of detail. Money and time were saved by not collecting detailed data about unpromising areas. Second, iteration between the FPC group and the EAG allowed the former to revise its own criteria and weights. This process moved the two groups' judgments closer together. Third, the
APPENDIX A 193 process built trust among the members of the EAG, who initially sus- pected that their inputs would not make a difference. The use of EAG input to revise the map of areas to be considered for further study con- vinced EAG members of the intention of FPC to actually use their inputs in the selection of the sites. The value elicitation procedure resulted in a useful value tree graphic that looks like an organization chart. The graphic helped the participants by clearly relating the overall mission of the project to the criteria and the specific indicators used to assess the adequacy of possible sites. The graphic and the process that generated it made clear multiple perspec- tives. Building a weighted list of concrete criteria and indicators focused discussion, thereby discouraging posturing and vague rhetoric. People directed their concerns toward the goal of improving the collective value tree. In the FPC case, multiattribute utility analysis techniques helped in- corporate multiple concerns into the site-selection decision and resulted in a transparent decision-making process that has acquired a high degree of public acceptance. The licensing for the preferred site is proceeding in a relatively smooth way. Using this analytic approach does not, however, guarantee that all participants' ideas are meaningfully included in an analytic-deliberative process or that the resulting rankings are meaning- ful (Brody and Rosen, 1994~. Neither does this approach, even when used sensitively, guarantee noncontroversial siting of contentious facilities when it is becoming increasingly difficult to license any large, locally unpopular facility. Nevertheless, it is encouraging that people from di- verse perspectives could reach agreement on the siting of a 3,000 mega- watt coal-fired power plant. Another encouraging sign is that two of the members of the EAG-in fact, two who would in other circumstances be expected to oppose actions by FPC have gone on public record on their own initiative in support of the siting process. THE CALIFORNIA COMPARATIVE RISK PROJECT When California began its comparative risk project (CCRP), four states had already completed similar studies and 10 others were under way. Comparative risk projects work from the assumption that policy priori- ties of environmental problems should be determined, at least in part, by the magnitude of the risk each problem presents (U.S. Environmental Protection Agency, 1987~. While the CCRP workplan (California Envi- ronmental Protection Agency, 1992:i) emphasized ranking risks using "good scientific minds together to help establish the 'best science,"' it also realized that deliberation among scientists would not be a sufficient basis to set policy priorities. The California project paid explicit attention to the
194 UNDERSTANDING RISK: INFORMING DECISIONS IN A DEMOCRATIC SOCIETY need for broader participation in its initial process design (California En- vironmental Protection Agency, 1992:2~: The responsibility for protecting California's environment applies not only to government, but also depends on the involvement of individuals with academic, industrial, business, activist, residential, and political ~n- terests within the State.... We will conduct the comparative risk project in such a way as to allow for all opinions to be accounted for, since the project is dedicated to expanding the public's ability to make important decisions about the fate of their environment. Direct citizen participation did not play as large a role as this state- ment might suggest. Still, the CCRP was significant because it recognized the need for broadly participatory deliberation and for a broad agenda for risk analysis it presumed that social, economic, and equity concerns would have to be included in any risk ranking scheme. This was evident in both the way the problem was formulated and the fact that the CCRP was willing to modify its process design as a result of deliberation about whether it would produce the needed information. The initial process design developed by the California EPA as a result of its diagnosis of the situation called for three technical committees on human health, ecological health, and social welfare- to work indepen- dently to rank risks in their categories, while three other committees would consider management options and the legal and economic con- straints associated with making those choices (California Environmental Protection Agency, 1992~. In a second phase, the risk rankings of the three committees were to be integrated during a two-day statewide sympo- sium. This design seemed to have analysis preceding deliberation and to confine the participation of nontechnical people to the decision-making phase of the project. Soon after the project began, this process design was challenged on the grounds that the technical committees could not be expected to pro- duce purely objective risk rankings. The technical analysis would be permeated, some said, with policy considerations at various levels. One argument was that the use of population-risk estimates left risk managers blind to the inequitable distribution of risk among subpopulations. This concern led to the creation of an environmental justice committee to ad- dress such issues. The original process design was also altered to include more feedback and interaction among the technical committees and to include critics of a conventional risk analysis approach. The new process design comprised two "components" operating in parallel with significant cross-fertilization. One involved the original three technical committees, using their knowledge of quantitative risk assessments in order to rate the impacts of various environmental condi
APPENDIX A 195 lions on the broad areas of human health, environment, and social wel- fare. The other comprised three committees that supplemented the tech- nical risk assessments with social and economic concerns. One of these committees, the environmental justice committee, raised additional con- cerns (e.g., social equity) for inclusion in the risk rankings and also addi- tional options (pollution prevention) for consideration (California Envi- ronmental Protection Agency, 1994~. In doing these things, the committee called into question the initial problem formulation as too narrow. Its perspective found its way into the criteria finally used to rank the social welfare and human health impacts of environmental stressors (equity was one of the criteria adopted, although pollution prevention was not), and into the CCRP's final report in the form of committee recommenda- tions and a "critique of the risk-ranking model." Each technical committee reached a surprising level of agreement on rankings considering the great diversity of backgrounds among each committee's members. Nonetheless, the CCRP generated considerable controversy, particularly over some of the more subjective social welfare outcomes, such as "peace of mind," that it endorsed as important (Stone, 1994~. The CCRP's statewide Community Advisory Committee decided not to integrate all three rankings into a common list because of time constraints, concerns over technical issues involved in reducing the three very different rankings into a single ranking, and discomfort about in- cluding in any "technical" ranking what some characterized as "subjec- tive" and "nonscientific" social welfare rankings. Some trade organiza- tions went directly to the press and the governor's office with the concern that "California has come up with a new and controversial method of evaluating environmental risk that downplays the traditional role of sci- ence and takes into account people's values, opinion, fears, and anxieties" (Clifford, 1994:A1~. In election year politics, the state government dis- tanced itself from the CCRP report: although it released the report, it did not publicize its results or apparently use its rankings in setting priorities. Despite this immediate outcome in state politics and the fact that the project did not fulfill all of its own objectives, the project is instructive because it brought together conventional forms of risk analysis with analy- sis and deliberation about various social, economic, equity, and other concerns. It demonstrates the importance of iteration in process design, particularly the use of deliberation to revisit the provisional problem for- mulation and the process design that emerged from the diagnostic phase, and the links between process design and problem formulation. The new process did not reduce controversy, but it did make explicit some of the different views of the nature of environmental problems underlying envi- ronmental policy conflicts in the state. The longer-term effects of the project remain to be seen. One possibil
196 UNDERSTANDING RISK: INFORMING DECISIONS IN A DEMOCRATIC SOCIE~ ity is that the project's methods and findings may be reconsidered at a later date. Another, less sanguine, possibility is that events may have made it more difficult to organize future broadly based risk deliberations in California because many participants who worked hard on the project may have been alienated. Some saw concerns they judged to be legiti- mate and essential successfully painted as unscientific and therefore un- worthy of consideration by others (some of whom also participated), working outside the process. Some volunteered considerable time and effort to technical analyses and deliberations and, acting on good faith, agreed with others (with whom they publicly disagreed in other forums) on rankings, only to see their hard work set aside by the political process. PLANNING FUTURE LAND USES AT HANFORD, WASHINGTON Hanford is a 560-square-mile site on the Columbia River that was used for decades by the U.S. Department of Energy (DOE) and its prede- cessors for the production of nuclear materials for national defense. Air, water, and soil at the site have been contaminated with radioactive mate- rials. Since the mission of the site ended in the 1980s, attention has turned toward restoring its environment and preparing the location for future uses. As part of this project, the DOE is required to complete an environ- mental impact statement (EIS) that would determine potential impacts associated with the cleanup and restoration projects. DOE decided to seek participation by representatives of a wide range of governments (federal, state, tribal, local) and various other organizations and interest groups in planning the EIS and identifying alternative scenarios for fu- ture use of the site. Customarily, an EIS begins with a scoping process to identify all envi- ronmental issues that are likely to be significant in the assessment. Scoping is meant to make the impact analysis more efficient by focusing attention on realistic issues and concerns. It involves defining assessment objectives and is similar to problem formulation as described in this re- port. Risk assessments are likely to be part of any EIS concerning a hazardous waste site, and the scoping process is a way of ensuring that the policy options and possible harms addressed in the assessments are consistent with the needs of the decision makers and affected parties. DOE, in collaboration with EPA, the Oregon and Washington state governments, and county and tribal governments of the region composed a list of potential participants for the Hanford Future Site Use Working Group, including both public officials and interested and affected parties, to advise on the EIS (Hanford Future Sites Working Group, 1992~. One of the most noteworthy features of this effort was that the process design for
APPENDIX A 197 the working group was not imposed in advance, but was created only after conducting interviews with the prospective members. Each candi- date was asked what would make the process successful, and each was also asked to nominate other potential participants. By obtaining this feedback at the start, the organizers bettered the chances that the eventual EIS would serve the needs of the interested and affected populations. Three things emerged from the interviews as important: the process should consider a wide variety of viewpoints, it should provide a com- mon base of information about the site, and the decision-making agencies should commit to using the products of the process in their decision making. With the assistance of a professional facilitation team, the working group drafted its own charter, specifying its purpose and the scope of its work. The main task was to identify alternative scenarios for cleanup and future site use, based on the participants' visions and influencing factors. The group agreed to focus on how cleanup and future site uses would be connected to each other essentially specifying some outcomes to be sub- jected to analysis in the EIS. The working group also decided not to seek consensus on a preference for future site uses, but to emphasize common- alities among the options generated. It also decided to make all decisions by consensus. A significant feature of the Hanford Future Site Uses Working Group was that the participants identified the information they needed. Since each participant represented a specific constituency, the information needs were linked with the participants' objectives. After reviewing ex- isting information and visiting the site, the group composed a list of needed educational seminars and information. This resulted in 34 pre- sentations by expert teams at various meetings of the working group. A huge amount of information is available on the history, use, and present condition of the Hanford site. While the working group noted important gaps in data and potential uncertainties, it did not have the resources or time to initiate any specific studies. However, it did identify needs for future data collection and analysis. The working group met five times over a 6-month period, and partici- pants were expected to consult with their constituencies in the time be- tween meetings to get wider input. The group issued draft recommenda- tions with consensual support and brought them to the public for review and comment at eight "open houses" in various locations in Washington and Oregon. The draft report was revised to reflect the comments re- ceived. The working group made nine major recommendations in addition to specifying general future use options and cleanup scenarios for different areas of the Hanford site. Several of the general recommendations had to
198 UNDERSTANDING RISK: INFORMING DECISIONS INA DEMOCRATIC SOCIETY do with possible outcomes that the working group asserted should be strongly considered in any impact assessment: treaty rights of Native American tribes (especially access to religious sites); stopping and pre- venting contamination of the Columbia River and cleaning up groundwa- ter; access to contaminated areas; public awareness of waste shipments; and jobs and regional economic impacts. The working group did not offer a complete list of possibly significant outcomes, nor did it rank their importance. The bulk of the working group's recommendations concerned: (a) defining eight general future use options for each subarea of the Hanford site, (b) listing some specific and important possible outcomes relevant to that area (e.g., effects on cultural sites, wildlife, and industry), and (c) defining the cleanup scenarios necessary to assure the safety of each fu- ture use. Cleanup levels were delineated according to access restrictions, which would be based on health risk assessments. This case illustrates how a risk decision-making process can use tech- niques for deliberation and public review in combination with analysis to arrive at a widely acceptable formulation of the problem and to design a process that is likely to characterize risks in a way that meets the needs of the interested and affected parties and public officials. DOE is seriously considering the working group's recommendations. Before preparing the draft EIS, DOE staff prepared an implementation plan that assimilated the recommendations of the working group. In a truly iterative process, the DOE staff reconvened the working group 1 year after its dissolution to confirm that its recommendations had been properly interpreted by DOE in the EIS implementation plan. The entire working group recommenda- tion document was included in the draft EIS scoping chapter, as the work- ing group had requested. DOE has found the process valuable enough to require similar future use working groups at all of its cleanup sites. DOE's reliance on local public involvement, however, remains controversial (see, e.g., Blush and Heitman, 1995), and the ultimate effect of the process on cleanup at Hanford remains to be seen. Because the underlying issues remain intensely controversial, a policy consensus may remain elusive even if the EIS addresses all the parties' questions.