A PROPOSED FRAMEWORK FOR MULTINATIONAL COMPARATIVE RISK ANALYSIS: PESTICIDE USE, IMPACTS AND MANAGEMENT Report of the Comparative

Risk Assessment Methods Workgroup J.A. SHATKIN The Cadmus Group, Inc., Watertown, Massachusetts 02472 USA, I. ANDREAS Romanian Association for Science and Progress, Bucharest, ROMANIA D.S. APUL University of New Hampshire, Environmental Research Group, Durham, USA A. ATTIA Institute of Graduate Studies & Research, Alexandria University, EGYPT M. BRAMBILLA, F. CARINI Universita Cattolica del Sacro Cuore, Piacenza, ITALY Y. ELSHAYEB Cairo University, Faculty of Engineering Mining Dept, Giza, EGYPT S. GIRGIN Middle East Technical University, Ankara, TURKEY G. IGNATAVITUS University of Vilnius, Vilnius, LITHUANIA T. MANDARSZ University of Miskolc, Miskolc-Egyetemvaros, HUNGARY M. SMALL Carnegie Mellon University, Pittsburgh, PA, USA O. SMIRNOVA Research Center of Spacecraft Radiation Security, Moscow, RUSSIA J. SORVARI Senior Research Scientist, Lic. Tech. Helsinki, FINLAND
A. TAL The Arava Institute for Environmental Studies, D, Eilot, ISRAEL

149

I. Linkov and A. Bakr Ramadan (eds.), Comparative Risk Assessment and Environmental Decision Making, 149168. 2004 Kluwer Academic Publishers. Printed in the Netherlands.

150

Abstract Comparative risk assessment is a natural tool for decision making regarding transboundary environmental issues. A workgroup of environmental experts met during the NATO Workshop over three days and addressed the state of the practice of comparative risk assessment and its applicability in an international context. Discussion of the application of comparative risk assessment led to a proposed case study on transboundary risks from pesticide use. Pesticides are of international concern because of transboundary trade of agricultural goods, impacts of applied pesticides on international environmental resources such as water bodies, differences in local and national safety practices, and concerns over the transferability of assessment models developed under different conditions than where pesticides are ultimately applied. Thus, a comparative assessment of potential impacts of pesticides on an international scale is proposed to be illustrative of the utility of the methodology for international environmental policy purposes. Workgroup discussions included the need to defining terminology, data needs, methods and tools for comparing risks, development of evaluation criteria, key issues regarding transboundary comparisons, and unique communication issues for international collaboration. A proposal for future work, including recommendations for international risk comparisons are also offered.

1. Introduction Over the past fifteen years, comparative risk assessment (CRA) has emerged as a central tool in evaluating public health concerns, environmental management strategies and especially prioritizing environmental and ecological issues for communities and countries. In the United States alone, more than half of all states have conducted comparative risk exercises of sundry scopes and formats which frequently drive public policy and resource allocation within the public sector. Owing to its risk-based scientific ranking approach, CRA is widely used for identification of higher risk problems and setting priorities for research and action. CRA provides a powerful methodology to improve resource allocation, particularly if costs are explicitly introduced or risks are normalized for a given benefit. Policies based on comparative risk assessments could lead to more efficient use of resources and greater protection of public health and the environment. CRA can also be used for analysis and comparison of risks from two or more risk management alternatives that might be applied to the same environmental problem [1], Increasingly, CRA is being used outside the U.S. but generally is applied within countries and even cities [2-4]. However, at the recent World Summit for Sustainable Development, comparative risk analysis was not only absent from the multi-lateral agenda that was manifested in the Johannesburg Declaration and Plan of Implementation, but also from the numerous bi-national and regional partnerships spawned by that international summit. This stands in contrast with the increasing awareness that many environmental hazards require a multi-lateral intervention to be successfully addressed. According to one estimate, there are over 100 multi-national watersheds in the world. Because

151

wildlife are unaware of geopolitical borders, diversity protection often constitutes a regional challenge. Tropospheric ozone formation, acid rain and vehicle emissions are just a few of the issues in the area of air pollution which require a transboundary strategy, and could benefit from an international scale comparative assessment of risk. Thirty-seven risk analysts and researchers from nineteen countries recently gathered in Anzio (Rome, Italy) under the auspices of NATO in a professional workshop to consider developments in comparative risk analysis in general and their applicability in international frameworks. This article offers the conclusions of a working group of experts that met at the workshop to this end. The overall questions that we addressed are: How can comparative risk analysis be used in an international context in reaching a common set of environmental priorities and objectives? If so, what form, format, geographical scope and approach might be appropriate for such initiatives? What issues are important considerations? To answer these questions we present a brief review of the range of methodologies presently utilized by risk practitioners and criteria for evaluating them within an international context. As data gaps pose a threshold obstacle to a multinational comparative risk effort, information requirements are discussed with regard to their likely availability and the potential for models and other efforts to supplement empirical data when needed. Finally, the group considered the example of agricultural pollution, in particular contamination and exposures from pesticides, as an example of an area in which regional or transboundary comparative risk assessment could aid in the identification of international priorities for sound pesticide management. While other areas of interest, such as air pollution, might be more natural transboundary environmental media, efforts to consider a conventional, but concrete environmental problem such as pesticide applications reveal problems, both practical and conceptual, and demands which a multi-lateral comparative risk exercise may pose. At the same time, this evaluation suggests that multi-lateral risk exercises can lead to more efficient utilization of public resources by the participating nations, assist donor agencies in assembling a more cost-effective funding strategy, and most importantly, lead to greater environmental improvements as a result of logical allocation of resources and direction of energies in the environmental sphere. In the remainder of the report we discuss prior applications of comparative risk assessment and available tools. We then present the results of our discussion on key international issues associated with the application of pesticides in a comparative risk framework. First, we define some key terminology.

2. Definitions As in any international context, terminology may be used to express a range of ideas. Others have developed international glossaries of risk assessment terminology (e.g. Duffus [5]). Because we ourselves were an international group, we found it necessary

152

to agree on definitions of our key terminology. In this work, we define the term hazard as an event with adverse consequences, and risk as the probability of a hazard. Finally, we defined comparative risk assessment (CRA) as the simultaneous analysis, evaluation or ranking of multiple hazards and their associated risks.

3. Applications of Comparative Risk Assessment Comparative Risk Assessment (CRA) provides a general framework for evaluating environmental problems affecting humans and ecosystems. Use of CRA for environmental problems started in U.S. with Environmental Protection Agencys (EPA) Unfinished Business: A Comparative Assessment of Environmental Problem report in 1987 [1]. The report, that evaluated more than 30 environmental problems in a comparative manner, reached the conclusion that the priorities of the environmental program at that time did not reflect the priorities determined by scientific methods. Since then EPA has been promoting the use of CRA and related environmental planning tools by states, regions, cities, and native tribes to help communities in addressing their environmental concerns. Many comparative risk assessment projects and work for formulating methods to make broader use of CRA are currently in progress [7]. Outside the United States, U.S. Agency for International Development has commissioned about ten CRAs since 1990 for selected developing cities, countries, and regions in the world [2]. CRA has been also used in a number of developed countries. In Europe, both the EU and individual countries are working to adjust risk assessment techniques for application within their contexts [3]. CRA was employed in the preparation of the 1993 Environmental Action Programme for Central and Eastern Europe [4]. Reviews listing and comparing many CRA applications are available in the literature. The Green Mountain Institute for Environmental Democracy has a resource guide that lists hundreds of documents that reflect and/or discuss the various aspects of planning, implementing, and using the results from comparative risk assessment [8]. Morgenstern et al. [7] examines the experience with CRAs conducted in various developing countries in transition and compares both the methodologies and the results. The World Banks Pollution Prevention and Abatement Handbook [4] also includes a summary of risk assessment projects in developing countries and transition economies. A review of the first ten years of comparative risk assessment in U.S., primarily at the state and local level, is given by Jones [9] with the aim to provide an informed historical perspective on experiences with comparative risk approaches in legislative, regulatory, and policy contexts, and to address uses and misuses of these approaches. Comparative Risk Assessment Primer software, developed by Purdue University - Center for Technology Transfer and Pollution Prevention, includes summaries of 36 CRA projects conducted in U.S. and also detailed information on CRA methodology [10]. Several applications of comparative risk assessment, not listed in these reviews, are of interest. One example is the World Health Organization Global Burden of Disease study, which estimates disease and injury burden attributable to different

153

risk factors using the CRA methodology and models health outcomes from the distribution of exposure to the risk factors [11]. Another example is assessment of potential human exposures from the consumption of contaminated drinking water and related health risks. The results of such a CRA can provide useful information on the incidence and relative risks of different drinking water contaminants in a region; this analysis can be used to prioritize public health hazards and aid in the development of appropriate risk mitigation efforts [12]. Similarly, air pollution risks can also be evaluated using CRA [13]. Setting priorities for remediation of environmental contamination is another area where CRA can be extensively used. CRA can help to identify the best way to allocate financial resources for the cleaning up of the environmental contamination associated with industrial activities. It is also used to make informed decisions on the decontamination priorities of the sites, the extent of the remediation, and the techniques to be used for this purpose. The World Bank Environment, Industry, and Mining project in Bolivia, which includes CRA for the adverse effects of mining activities, like heavy metal contamination, acid generation, and physical hazards, is a satisfactory application in this area [14]. CRA can be used to evaluate the impacts of industrial waste disposal and the adequacy of and priorities for waste treatment policies can be highlighted, identifying attributes important to human, ecosystem health and to decision making for priority setting in the early stages of the environmental planning [15]. Agricultural pest management is a global, serious problem as approximately 50% of the worlds food supply is destroyed each year by pests while the human population continues to expand rapidly [16]. Many potential advantages may be gained by including comparative risk assessment in the management of pesticides including that CRA may result in increased commercial incentive to develop less hazardous products. In addition to this, data gaps may be filled in response to the commercial incentives to be able to demonstrate a products lower risk in a comparative risk assessment [17]. CRA can be used also to compare the risks arising from chemical, organic, genetic engineering, and other pest management methods. Comparative assessment of pesticides has already been shown to be effective and successful within several EU countries, such as Sweden. This principle is also included in Directive 98/8/EC concerning the placing of biocidal products on the market. Recently a report prepared by EC addressed the need to modify the directive concerning the marketing of plant protection products (pesticides) in certain respects and comparative assessment is listed as one area for consideration. Comparative risk assessment for pesticides features in the 5th European Community Environmental Action Programme (EAP) and is currently included in the proposal for the 6th EAP [18]. In addition to health risk assessments, comparative risk assessment can be used for ecological risk assessments. For example, CRA can be integrated in estuary management programs. The Lower Columbia River Estuary Partnership, a two-state public-private initiative, is a successful example of application of transboundary of CRA for priority setting. Using a watershed approach, the Estuary Partnership cuts across political boundaries, integrating 28 cities, 9 counties, and the states of Oregon and Washington in U.S. [19]. Another application area of CRA is marine environment risk assessment: the banning: the International Maritime Organization has developed a

154

global initiative that will eventually result in the ban of all antifouling systems exhibiting harmful effects on the marine environment. Given the number of alternative antifouling paints being developed, a process is necessary to determine the antifoulant expected to have the fewest impacts on the environment. CRA is a useful technique for such a purpose [20]. Rankings obtained from a CRA study may be useful for resource-allocation decisions. The risk-based process being introduced by the Department of Energys Environmental Management Program at the nations nuclear-waste sites is testing the effectiveness of translating the identification, analysis, and comparison of risks and remedies into budget decisions. The Commission encourages federal regulatory agencies to use comparative risk assessment for priority-setting on an experimental or demonstration basis [21]. CRA is also used in the energy sector. By applying CRA, the diverse characteristics, problems and requirements of energy technologies can be determined and the most efficient technology in terms of cost, public health and environment can be found [22]. There are also military applications of comparative risk assessment. CRA was used to assess different process arrangement alternatives and minimize the human health and ecosystem risks risk for managing and treating chemical agent stockpiles [23]. The remainder of this paper describes the discussions of the workgroup, and reports on key findings in on topics relative to the development of a proposed case study of the comparative risks of pesticide use and management in an international context.

4. Case Study Pesticides are applied globally, but political and natural variation within nations affects their impact on health and the environment, both internally, and across ecological and trade barriers. The workgroup opted to address pesticides in an international context because of the international implications for food safety, for ecological impacts which do not follow political boundaries, opportunities for international cooperation on pesticide management, because of the range of expertise within the group in this area, and the opportunity to elucidate important factors in successful comparative risk assessment for the evaluation of international environmental issues. Further, the example made concrete the conceptual issues, such as the key data needs, comparative criteria, and key issues to be addressed. Pesticides may pose transboundary problems when they are introduced into the environment. Because of their persistence in the environment, they can be transported through movements of waterbodies, air masses, or ocean currents. Some pesticides can be redistributed at a global level from warm-temperate to cold areas of the planet, and can be accumulated in aquatic or terrestrial organisms and transferred through the food chains. Pesticides may also pose a transboundary problem when contaminated food items such as fruits and vegetables are exported from areas of production to other areas and countries. Land use patterns and regional geography affect the migration of pesticides across political boundaries.

155

The remainder of this paper describes the discussions of the workgroup on topics relative to the development of a case study on the comparative risks of pesticide use and management in an international context. First, data needs and available and useful tools for comparing pesticide risks are described. Next, key criteria for comparing risks are evaluated. We then raise important issues for international risk comparisons, and suggest further work in this area. Finally, recommendations from the process are highlighted.

5. Data, Analysis and Tools Numerous presentations during the four day workshop addressed important data issues and analytical tools useful for comparative risk analysis. For example, Schumann (this volume) discussed the effects of regional differences in soil type and irrigation/rainfall in Nepal, highlighting the need for local measurements of pesticides in addition to laboratory testing and models based on measures in other climates and soil types. Below types of data required for an international comparison of pesticides are discussed. 5.1. DATA NEEDS FOR COMPARATIVE RISK ASSESSMENT For an international comparison of health and environmental risks from pesticides, data must be gathered from each source area. Many factors will affect exposure, and factors will vary by local practice and geography. Toxicity data is needed, but will not vary much across places. Specific types of data are needed to evaluate and compare exposure, such as residue levels in workers, food items, the environment, and more nationally specific data on tolerances, food consumption habits, importation/ exportation rates, worker standards, and historical information. Data must be of good quality and be comparable, measured with comparable apparatus and techniques. That is, the types of information must be on similar scale, frequency, aspect of pesticide use, and gathered by similar methods. Data for evaluating pesticide risk across boundaries includes quantitative and qualitative information. Types of information include: biological, chemical, physical and environmental aspects of the system under consideration, socio-economic, political, and demographic data. The availability of data on pesticide use, residues, and exposure levels to workers and consumers varies across nations. National laws and customs dictate the gathering of information, and the level of safety with which pesticide use is managed. Gathering of new data for this study would improve comparability. Data useful for comparing risks from pesticide use include identification of the sources of a pesticide, its characteristic features, environmental fate and possible adverse effects. Physical-chemical and fate and effect parameters may vary with environmental conditions. Data are needed on geographical/ climatic similarities and differences within and between countries, including: hydrogeology, meteorology, geotechnical properties, as well as land use, and local measurements. Identification, geographical distribution, and classification of vulnerable ecosystems and exposed groups of populations is needed to compare risks across

156

political boundaries. This classification can help policy makers to make decisions on the areas and groups of population for which the measures have to be implemented. Biomarkers could be used to compare exposures across pesticide exposed populations. Some biomarkers have been associated with specific environmental exposures. One workshop presentation, Attia, focused on the use of biomarkers to evaluate pesticide exposure in agricultural workers [24]. Biomarkers are used to indicate the presence of or level of activity of a given process, ideally, through the measurement of a reactant that is consumed or produced by that process, thus directly related to the mechanism-of-action [25]. Unfortunately, most biomarkers are not mechanistically derived but rather arise from an observed correlation with the existence of the process. At the genomic level, changes in the level of ribosomal nucleic acid or protein expression are markers that signify changes in biological activity, although they only comprise a single component of the overall process involved in determining biological activity. Another workshop participant, Smirnova, discussed the use of mathematical models for identifying hypersensitive individuals chronically exposed to low levels of unfavorable factors (toxic substances, ionizing radiation and so on) [26]. To resolve this problem, the development of new approaches to risk assessment are needed, due to the ambiguity of effects of such exposures. In particular, hormetic effects were observed in a number of experiments with low doses of poisons and radiation. Therefore the new approaches must not ignore intrinsic properties of exposed organism. The implementation of such approaches calls for development and investigation of mathematical models describing mortality as an ultimate result of damage of mammalian organisms induced by an unfavorable factor. While the models developed were for evaluating risk from radiation exposure, biologically-based models could also be used to estimate risks for exposed individuals and populations in international comparative risk evaluations. . 5.2. TOOLS FOR COMPARING RISKS Comparative risk assessment studies work best when conducted by multidisciplinary working groups including scientists, technical experts, industry representatives, government officials, citizens, and professional communicators, each having different backgrounds and viewpoints. Effective discussion and information flow both within and between groups are important for obtaining satisfactory results and meaningful comparative risk rankings. Advanced data visualization capabilities may facilitate this communication. Data in the form of a map is easier to interpret and communicate about than is tabulated data, and this affects the connections made and the conclusions drawn from it, particularly in a multilingual context. Drawing and layout tools are helpful in this respect and can be used to present the outputs from the risk assessment study to the decision and policy makers. The first step of a CRA is the determination of the set of environmental problem areas to be analyzed and compared. For an international comparison of risks associated with pesticide use and management, a lot of information must be collected and analyzed. A non exhaustive listing of available tools that could be used in any risk

157

assessment study is presented, followed by few paragraphs specifically addressing tools for CRA. 5.2.1 GIS Models The nature of any environmental or economic activity with a spatial dimension cannot be properly understood without reference to its spatial qualities. Therefore during the risk assessment process spatial dimension should be taken into the consideration. A Geographic Information System (GIS), with its advanced data integration, query, analysis, visualization and modeling capabilities, can be an effective and efficient platform for this purpose. Actually, GIS can be used as a tool for all phases of a CRA, including the determination of environmental problems to be addressed, scale of the evaluation, analysis of the risks they pose, and ranking them based on their relative importance. A database management system, an integrated part of a GIS, provides means of rapid data access and query based on geographic location or attribute data. Using mapping functions of GIS, it is possible to superimpose two or more data layers and to relate otherwise disparate data on the basis of common geographic location. For example, exposed populations can be identified from a population density layer on a map identifying land uses. Also GIS makes it possible to explore and analyze data by location, revealing hidden patterns, relationships, and trends that are not readily apparent in spreadsheets or statistical packages. Because GIS products can be produced quickly, multiple scenarios can be evaluated efficiently and effectively. 5.2.2 Quantification Models Numerous models are available for quantifying qualitative parameters (ecological factors, qualitative soil and pesticides parameters, etc.). Other workshop papers described a range of these tools, including Elbashyeb, Shatkin, and others, found elsewhere in this volume. These models include Fuzzy systems, Grey techniques, and other quantification techniques. 5.2.3 Statistical Models Statistical models are used to improve and better understand subsequent patterns within the available data in order to be able to draw general conclusions. Statistical techniques can be applied for classification, estimation, prediction, clustering, and data description. It goes without saying that statistical models include all basic statistical techniques, such as probability distributions and inferencing methods. 5.2.4 Analysis Models Interdisciplinary System Analysis models are useful when dealing with CRA applications due to the interdisciplinary nature of the system dealt with in such applications. System simulation are valuable techniques when dealing with poorly understood systems, and where small changes in the system could have multiple big impact on the subsequent system and are usually irreversible. It is though useful to build simulation models in order to test different configurations, and choose the most

158

suitable one. Many CRAs rely on analyses with if/then rules, or expert judgment comparison. In addition to these approaches, many of the review volumes discussed in the applications section also summarize available tools for CRA.

6. Comparative Decision Criteria In a CRA, the criteria by which risks are compared may be based on risk to human health, ecology, or other endpoints such as impact on economy, or quality of life. For any comparison, the criteria must allow fair evaluation of each risk, and must be applied consistently to the data. The comparative decision criteria of an international CRA with a focus on pesticide use could evaluate affected populations at a regional, national or international scale. Broad participation by those affected by the comparison can help to ensure that important criteria are used for risk comparisons. Below, key criteria are discussed that should be considered in a multinational CRA. Section 7 highlights how the workgroup proposes to address these criteria. It is not straightforward to determine how to compare different types of risks. For example, how does one weight risks to ecological receptors versus human health risk? Similarly, in characterizing the risk to human health, how does one compare different toxicological consequences such as cancer and non-cancer endpoints? Some of the decisions required to compare different risks are political rather than scientific, and therefore must include decision makers. The complexity of deriving comparative risk criteria increases if the comparison is across boundaries where many nations with differing priorities are involved. This section discusses options for comparing risks and the factors contributing to the complicated nature of risks comparisons. 6.1 EXPOSURE CRITERIA RELATED TO HUMAN AND ECOLOGICAL RISKS It may be helpful to evaluate the relative distribution of any xenobiotic between the media or compartments (water, soil sediment, air and biota) at equilibrium [27]. Comparing the relative distribution of a contaminant can provide a good indication of the environmental compartment in which further CRA on effects should be conducted. Given that the exposure of a foreign substance primarily depends on its fate in the environment, the way mobility (transport/translocation), transformation (metabolism) and degradation (mineralization/formation of irreversible bounds) are taken into consideration needs to be comparable as well because of the space and time scale implications of these three features that may change according to the type of source as well as the physic-chemical properties of a substance. This comparison should consider the relationship between dose or concentration of a substance and/or its metabolites and the incidence and severity of an effect on organisms or ecosystems. 6.2 COMPARING ECOLOGICAL RISKS When assessing the ecological (ecotoxicological) hazard potential, an attempt must be made to weight the large number of different data required to yield an overall result that

159

will permit risks to be classified in an understandable univocal way. A useful way to weight, manage and sum up all these factors is the use of models to study and forecast the environmental fate of any xenobiotic. A pragmatic approach (making use of the existing ecological knowledge) is needed to create foundations enabling the benefits and risks of a substance for human beings and the environment to be assessed [28] and overcoming the statistical uncertainty entailed by the many different ecological methods. Comparing ecological models can be useful for CRA because of their iterative process of investigating real conditions, provided that assumptions and prerequisites on which models are based on are disclosed by the risk assessor. These are usually based on many different procedures and criteria that belong not only to the adopted modeling techniques but also to the risk assessment procedures of each country. Validation of the risk assessment procedures should be taken into consideration as well. 6.2.1. Effect of Bug Resistance Tremblay [29] defines resistance as a drop of the sensitivity of any organism to a certain active substance. It is the outcome of changes occurred in the penetration, activation, degradation and excretion of usually a toxic active substance. Being hereditary, resistance to toxic substances is maintained by the next generations: usually the more frequently the same product is used at high concentrations on the same area, the sooner it sets up. Bug resistance comparison criteria in the case of trans-boundary problems should pay attention not only to the kind of active substance together with the formulation but also to its rate of use and to the reproduction rate and mobility of the most dangerous bug. 6.3. TRANSPARENCY, SUSTAINABILITY, AND RISK PERCEPTION Different types of risk (e.g. ecological versus human health) cannot be normalized to the same unit. Faced with this problem, the option is to clearly state the factors that affect the risk analysis so the comparison is made as transparent as possible. For example, institutional differences such as different countries having different methodological approaches, regulations and law (e.g. difference in analysis of pesticide problems, difference in national tolerance levels) cannot be expressed in any defined units but should be considered and clearly stated. The temporal and spatial scales should also be clearly expressed and if at all possible set to same scales for comparison. On a global scale of multidisciplinary CRA, where cross boundary comparisons of the risk in presence of political, geographical differences are inherent, clear explanation of these differences is most needed to normalize the comparisons to the extent possible. In cases of international comparison of risk assessments where the results cannot be easily normalized for direct comparison, the results can be analyzed within their context. In other words, for non-comparable units, the risks can be analyzed in parallel. For example, risks related to welfare, human health, and ecological health in different countries may be compared in their own categories as an alternative to translating the units in their own categories to a consistent unit, such as dollars.

160

Risks expressed in terms of cost may or may not be an advisable strategy. Using economics or another normalizing metric to compare the risks from different categories may create issues across boundaries, where cultural preferences create different weightings for categories of risk. If costs are used as a comparative criterion, these will have to be normalized with the participation of representatives from each study location. Sustainability is a management criterion of risk assessment, and should be kept within this focus in order to be able to compare methods, tolerances, criteria, etc. In the case of pesticides, the sustainable management and assessment of risk are related to other criteria such as tolerance and perception. In order to achieve a sustainable risk management policy, a focus on the importance of the life cycle of pesticides, disposal management and the possibility of using auto-biodegradable products is suggested.

6.4. INCORPORATING VALUES DEMOCRATICALLY

Values are inseparable from comparative decision criteria. Yet, whose values should be used in risk comparison? Ideally, in comparing risks in a democratic society, the people who will be directly affected by the decision should make the comparison. A decision maker distant from the endpoints of the risk may have a limited basis for decision making. Effective public participation in decision making will pose a challenge in an international context and offers an opportunity for new collaborative work. Understanding peoples preferences, and helping them form their own opinions which may temporally and regionally fluctuate can be difficult, even more challenging if different nations are involved. Many times, town meetings and public hearings may not be effective in educating people and eliciting peoples opinions. For democratic decision criteria, significant progress is needed in risk communication. Methods are needed for incorporating values in comparative decision criteria. Decision makers and risk communicators roles as facilitators for helping people form educated opinions and documenting their opinions should be more clearly defined to express this need. Considering that science develops in areas where there is a need or a demand, communicating this need is essential for potential development of a technique to scientifically incorporate decisions and values of people who will be affected by the risk. Perhaps, in the future, there may be a branch of science or trade that regularly records peoples preferences and regionally and periodically updates a happiness scale similar to stock exchange indices. Such a happiness scale may incorporate Vermonts quality of life criteria such as aesthetics, economic well being, fairness, future generations, peace of mind, recreations, and sense of community [2].

7. Issues and Communication A Case Study Proposal In the development of a multinational CRA of pesticide use, impacts and management, a number of challenges must be addressed to allow for a coherent study that provides a clear understanding of current risks, and also serves as a guide for ongoing national and multinational data collection and risk management efforts. These challenges involve:

161 (a) the different temporal and spatial scales over which pesticide use and impacts occur; (b) the high degree of variability in pesticide use and application practices and in the environments and populations that are impacted; (c) differences in the way that data are collected and stored across different countries (and even within some countries), hindering their transfer and combination for a unified evaluation; (d) identifying and obtaining input from decision makers that influence pesticide use and management in different nations, including regulatory authorities, intergovernmental agencies, agricultural aid and outreach organizations, manufacturers, distributors, and trade organizations; and (e) ensuring the comparability of current and projected risk estimates across different health, safety and ecological endpoints. The workgroup outlined a proposed approach for addressing these issues for evaluating pesticides in an international context. We do not expect that these issues can all be fully addressed in a single study. Rather, we propose to begin with a broad-scale effort that clearly delineates these challenges and measures the extent to which they are met, identifying the ongoing and future research, data-collection and management programs that could best assure improvements in subsequent assessments. Effective communication during the planning, implementation and dissemination of the study will be essential to ensure that progress is made in overcoming these challenges. The following summarizes some of the key features of these issues that our proposed study will address, as well as the communication methods that will be used to ensure its success.

7.1. SCALE ISSUES IN PESTICIDE ASSESSMENT Pesticide use and impacts occur over a wide range of spatial and temporal scales. Risks of pollutant release to the environment occur during the manufacture and transport of compounds, during application, and following application to the field. Very near-field exposures and health risks occur first to workers involved in the manufacture and transport of pesticides, then (usually to a greater extent) to pesticide applicators in the field. These direct dermal and inhalation exposures can result in serious acute health effects to these workers. Subsequent dermal, inhalation and ingestion exposures can also occur following work hours to applicators and their families. These effects can be both acute and chronic for those who are involved in this activity over a period of many years. At a minimum, a survey of pesticide application practices and worker protection, education and exposure avoidance programs will be needed for countries included in a multinational study. This survey should be linked to available datasets on worker exposures, biomarker concentrations (in blood, urine, etc.), and documented health effects. A second major route of pesticide exposure occurs to consumers due to their ingestion of chemical residuals in food. Depending on the type of food distribution network and the particular use pattern of the crops and foodstuffs (or other consumer products) made from them, exposures and impacts can range from local to national to multinational, over time scales of a growing season to a few years. The proposed study

162

will attempt to establish a source-receptor matrix for agricultural products for the nations included in the study, recognizing that some portion of the crops produced will be exported out of the region, and some portion of the food consumed will originate from exogenous imports. Within each country, the sub-categorization of locallyproduced-and-distributed crops, vs. those distributed on the national market, will also be made. This will be done first for major grains, vegetables and fruits, with subsequent consideration of other products involved in the manufacture of moreprocessed foods, dairy and meat products (for an example of a methodology for estimating pesticide residuals in processed foods, see Hengel and Shibamoto [30]). The matrix will be linked to available information on trade and diet for each country, as well as a database of observed1 and allowable2 pesticide residuals in foods (this will allow a comparison of ingestion exposures and risks estimated under current actual vs. current ideally-managed conditions). The final major route of pesticide exposure comes about through their general release and transport through the environment. This occurs primarily following application, but can also occur due to routine or accidental releases during manufacture, transport, or on-site formulation. The spatial scales of these effects can range from local or regional impacts on the ambient atmosphere, streams, groundwater aquifers, vegetation, fish and wildlife3, to the widespread global impacts now associated with persistent organic pollutants (POPs) and persistent bioaccumulating toxics (PBTs)4. Of the 12 pollutants now identified by international treaty as POPs requiring oversight and control, eight are pesticides (DDT, chlordane, endrin, heptachlor, mirex, toxaphene, dieldrin, and aldrin).5 The temporal scale of these impacts can range from months to decades, depending on the properties of the compound and the environmental receptors ultimately exposed. To provide a first assessment of risks resulting from environmental release, the suite of pesticides used in the study countries will be characterized in terms of their fate-and-transport properties for long-range transport, persistence and bioaccumulation. A number of recent multi-media environmental modeling assessments will be used to inform this evaluation.6 This analysis will be supplemented by the development of a database of reported pesticide concentrations in
1 European Union reports on pesticide residuals in foods in selected countries are found at: http://europa.eu.int/comm/food/fs/inspections/fnaoi/reports/annual_eu/index_en.html. Examples of recent research studies on pesticide residuals in food include Lazoro et al. [31], Cabras and coworkers [32-35], Krol et al. [36], Saitta et al. [37] and Holden et al. [38]. 2 Reports describing maximum residue levels prescribed for pesticides in food in the European Union are found at: http://europa.eu.int/comm/food/fs/ph_ps/pest/index_en.htm . 3 For examples of studies documenting these effects, see Qian and Anderson [39], Falandysz et al. [40], Meijer et al. [41], and Papastergiou and Papadopoulou-Mourkidou [42]. Observed datasets can also be complemented by the use of environmental fate-and-transport models, for example, Woodrow et al. [43, 44] and Barra et al. [45]. 4 Documentation of global transport of pesticides is found, for example, in Cortes et al. [46] and de Wit et al. [47]. 5 For more information on pesticides as POPs, see, for example: http://www.sierraclub.org/toxics/factsheets/pops.asp and http://www.epa.gov/oppfead1/international/negotiation.html. 6 See, for example, Scheringer [48, 49], Bennett et al. [50,51], Eisenberg and McKone [52], and Hertwich and McKone [53].

163

water, soil, air, vegetation, fish and wildlife for the study region and adjacent areas. The effect of alternative application and field management practices on pesticide release to the environment will also be considered (e.g., Gan et al. [54]). 7.2 VARIABILITY IN PESTICIDE USE AND IMPACT There is a high degree of nation-to-nation and site-to-site variability in pesticide use practices, regulatory policies and landscape conditions that affect the distribution of pesticides to food and different environmental compartments. To the extent possible, we will document these differences and include them in our assessment models in an explicit manner, using location-specific factors linked to a GIS map and database. We will also explore the use of statistical models to generate distributions of pesticide use; time-activity patterns for pesticide applicators and their families; meteorological, soil, and geo-hydrologic properties affecting environmental transport; resulting pesticide residuals and ambient concentrations; exposure factors for human and ecological receptors; receptor sensitivity; and subsequent health and ecological effects. These distributions will be used to supplement missing data at specific grid cells in the model, and also to represent system component variability within grid cells. 7.3 DATA COMPATIBILITY Data on pesticide use, food residuals, ambient concentrations, exposures, biomarker concentrations, and health effects have been collected by government agencies and researchers in many countries. However, many of these data records are difficult to integrate because of differences in chemical names (and of course language), measurement methods, units in which the measurements are reported, the amount of information provided on the sample type and locations, and the degree of temporal and spatial averaging used in the measurements. We will attempt, when possible, to convert all data to a comparable basis and a common set of units. The data will be maintained in a common GIS database agreed upon by project participants during the first three months of the study period. We will determine whether protocols can be developed from those established in response to the July 1, 2002 European Union Communication Towards a Thematic Strategy on the Sustainable Use of Pesticides (COM 2002-349, see: http://europa.eu.int/comm/environment/ppps/home.htm). We will also examine other cross-national studies of environmental health for possible data-reporting and archiving protocols.7 A particular concern will be to ensure that data collected on pesticide use, environmental conditions, chemical transport, food residuals, and human and ecological exposure factors can be interfaced at common spatial and temporal scales. To accomplish this, a single spatial unit will be selected for use in the GIS system and all

7 Such as APHEIS: A European Information System on Air Pollution and Health (see: http://europa.eu.int/comm/health/ph/programmes/pollution/ph_poll_fp00_en.html), which includes cooperation by the European Environmental Agency, the Joint Research Centre, EUROSTAT and the World Health Organisation / European Centre for Environment and Health.

164

available data layers will be averaged (if originally at smaller scales) or disaggregated (if originally at larger scales) to match the spatial unit selected.
7.4 OBTAINING STAKEHOLDER INPUT

Effective participation and input is needed from the regulatory bodies, intergovernmental agencies, agricultural aid and outreach organizations, businesses, and trade organizations8 from across the countries participating in the proposed study. Project leaders from each participating country will be asked to identify and provide contact information for those stakeholders in their nation and region, and they will also be asked to establish a mechanism by which project study plans, information requests, and progress reports can be shared with them. Project groups will also be asked to characterize the flow of information and authority among their decision-making bodies and stakeholders, so that the institutional settings for pesticide management can be properly characterized for each participating nation.
7.5 COMPARABILITY OF RISK ESTIMATES

The assessment of risks from pesticides to workers, consumers, and the general population will include estimates for a significant number of human health, ecological and economic impacts. Two approaches will be explored for summarizing and presenting these results. First an attempt will be made to combine the different risk estimates for human health, ecosystem impacts, and economic effects into three aggregate measures of effect. Second, techniques for multi-attribute risk comparison and ranking will be used to explore the tradeoffs among these three major areas of impact. The argument for seeking a common unit of impact for risks is that, while decision makers must address problems of different scale, nature and origin, action on these problems must be prioritized, ranked and compared to each other. In order to make such comparison, common units are required. Unless the common unit is achieved the problems cannot be compared and the decisions will be made instead based on perceptions and subjective opinions. Although risk assessment cannot always eliminate subjective aspects, it is currently the best available tool to compare different threats and their consequences. The common unit that is the basis of the comparison is risk. A number of approaches have been explored for combining different risks to human health. One such approach is through the development of estimates of the quality of life-years lost, which can be used to evaluate both premature mortality and morbidity effects [55]. Similarly, environmental and economic impacts may be evaluated using a common set of economic measures, such as the value of the ecosystem services lost as a result of the environmental damage and impairment, or the

8 For example, through CropLife International and the Global Crop Protection Foundation, see: http://www.gcpf.org/

165

willingness-to-pay to avoid the ecological effect.9 Discussions among participating researchers and national and international decision makers and stakeholders will be conducted to identify the most pertinent impacts to include and the most appropriate methods for evaluating them. The discussions will then lead to the formulation of a final multi-attribute list of impacts for which risk estimates will be developed. Some of these estimates will be quantitative, while others will, due to limitations in available data and assessment methods, remain more qualitative in nature (for these, impacts will be estimated as either low, medium or high). Risk assessment is a tool that pushes the point of comparison from the concentrations in the media to levels of risk at the receptors in threat. (See figure 1. in [57]) This give three unique feature to the risk assessment. 1. The decision is made not based on concentration levels but on risk levels. 2. This allows decision makers to compare and rank environmental threats of different environmental media (e.g., the threat of an air pollution problem with a ground water contamination.) 3. The comparability of the risks is even wider, by providing comparable information not just between different environmental media, but between different types of hazards (eg. earthquake risks with risks due to flooding, or risks caused by chemical spills.)

8. Summary and Recommendations The workgroup considered the practice of CRA in an international context, and reviewed applications, tools, methods, data needs, and issues regarding the implementation of a study to evaluate risks associated with the use of pesticides. We generated a proposal for an international CRA, to address methodological challenges and answer an important environment policy question: how best to address risks from regional scale activities within a nation that potentially impact human and ecological resources across political boundaries, and at variable scale. In summary, we recommend: Data be gathered from available sources, and supplemented with original efforts to ensure comparability. Data needs for an international survey of pesticides include geophysical measures, residue measures, political information on practices and tolerances, and exposure measures reflective of the range of potential receptors (e.g. workers, consumers, ecological receptors). 8.1. TOOLS Spatial tools, such as GIS, are an obvious choice for comparing risks among geographical units including countries and ecosystems, as well as numerous statistical tools, modeling approaches, and deliberation as discussed during the presentations of workshop participants and summarized in other papers in this volume. We propose
9 A collection of different approaches for economic evaluation of environmental impacts is found in an April 15, 2000 special issue of the journal Environmental Science & Technology. The paper by Mourato et al. [56] specifically addresses the evaluation of health and environmental impacts from pesticide use.

166

incorporating many of these tools in an international comparison of risks associated with the use of pesticides on food crops. 8.2. ISSUES Key issues for comparing risks include scaling issues, cross boundary variability, the need for comparable data, obtaining common metrics for comparison, and ensuring representative participation in decision making. One issue flagged for further development in our proposed international CRA is the need for approaches to compare data across endpoints, both in quantitative and qualitative terms. Other issues include the temporal relevance of such an evaluation, because any data collected reflect current activities, but need to be characterized by their past and future effects on exposed receptors. Our investigation proposes to address these issues. 8.3. BENEFITS The workgroup members undertook this effort to identify resources because of a shared belief that CRA is a tool of potential import in the international environmental policy context. As international cooperation on environmental issues develops, tools can aid in identifying priorities, and analyzing alternatives. The risk framework offers a transparent way to consistently evaluate environmental practices with their associated human and environmental impacts across political boundaries. Comparative evaluations require consistent reporting mechanisms be developed, and fosters cooperation and communication among national representatives in decision making.

9. References
Murphy, P.A. and Rice, G.E. (2001) Overview of Comparative Risk Integration of Scientific Ideas and Approaches, Proceedings of the Society for Risk Analysis 2001 Annual Meeting. 2. Davies, J.C. (1996) Comparing Environmental Risks: Tools for Setting Government Priorities, in Davies, J.C. (ed), Resource for the Future, Washington D.C. (USA). Brantly, E. (1999) Comparative Risk Assessment: Lessons Learned, Environmental Health Project, 3. U.S. Agency for International Development. 4. World Bank Group (1998) The Pollution Prevention and Abatement Handbook, World Bank Group editions. Duffus, J.H. (2001) Risk Assessment Terminology, Chemistry International 23(2), 5. (http://www.iupac.org/publications/ci/2001/march/risk_assessment.html). 6. United States Environmental Protection Agency (1987) Unfinished Business: A Comparative Assessment of Environmental Problems, Washington, DC (USA). 7. Morgenstern, R.D., Shih, J., and Sessions, S.L. (2000) Comparative Risk Assessment: An International Comparison of Methodologies and Results, Journal of Hazardous Materials 78, 19-39. 8. Green Mountain Institute for Environmental Democracy (1997) Comparative Risk Resource Guide, Third Edition. 9. Jones K. (1997) A Retrospective on Ten Years of Comparative Risk, American Industrial Health Council. 10. Embleton, K..M., Jones, D.D., and Engel, B.A. (1996) Comparative Risk Assessment Primer, Environmental Software 11(4), 203-207. 11. Ezzati M., Lopez A (2002) WHO Global Burden of Disease: Comparative Risk Assessment, in Ezzati, M., Lopez, A., Rodgers, A.D., and Murray, C.J.L. (eds.) (2002) Comparative qualifications of Health
1.

167
Risks: the Global Burden of Disease Attributable to Selected Major Risk Factors, Geneva: World Health Organization. 12. Williams, P.R.D. and Sheehan P.J. (2001) A Comparative Risk Analysis of Drinking Water Contaminants in California, Society for Risk Assessment 2001 Annual Meeting. 13. Bykov, A.A., Akimov, V., Revich, B.A. (1998) Assessment and Comparative Analysis of Health Risks Caused by Contaminated Environment in Moscow, Society for Risk Analysis Europe 1998 Annual Meeting. 14. Ijjasz, E., and Tlaiye, L. (1999) Comparative Risk Assessment, Pollution Management In Focus, Discussion Note No: 2. 15. Shatkin, J.A. and Palma, J.M. (2001) A Multi-attribute Comparative Risk Assessment for Informing Priorities in the Treatment of Portuguese Industrial Hazardous Waste Streams, Society for Risk Analysis 2001 Annual Meeting. 16. Fisher, S. and Marta, M., A Comparative Risk Assessment of Chemical Genetic Engineering and Organic Approaches to Pest Management. 17. WWF-UKs Response to the Pesticide Safety Directorates Consultation on Comparative Risk Assessment (Substitution) in the Regulation of Pesticides, WWF-UK. 18. Consultation on Comparative Risk Assessment (Substitution) in the Regulation of Pesticides, (http://www.pesticides.gov.uk). 19. Lower Columbia River Estuary Partnership, (http://www.lcrep.org). 20. Comparative Aquatic Life Risk Assessment of Tributyltin and Tin-Free Biocides, ORTEP Organotion Environmental Programme Association, (http://www.ortepa.org). 21. Comparative Risk Assessment for Risk Management, Risk Report, (http://www.riskworld.com/Nreports/1996/risk_rpt/html/nr6aa001.html). 22. Ramanathan R., Comparative Risk Assessment of Energy Supply Technologies: A Data Envelopment Analysis Approach, Energy The International Journal, 26, 197-203. 23. Crawford-Brown, D.J., Comparative Risk Assessment of Alternative Management and Treatment Options for the Army Chemical Weapon Incineration Program, (http://www.cwwg.org/ComparativeRA.html). 24. Attia, A.M. (2003). Risk assessment of occupational exposure to pesticides. In: Linkov, I. and A. Ramadan, eds., Comparative Risk Assessment and Environmental Decision Making. Kluewer, Amsterdam (in press). 25. Ferriman, A. (2002). Advocates of PSA testing campaign to silence critics. Br. Med. J. 324, 2555. 26. Smirnova, O.A. (2000). Mathematical modeling of mortality dynamics of mammalian populations exposed to radiation, Mathematical Biosciences 167, 19-30. 27. Rmbke J., Moltmann J.F. (1996) Applied Ecotoxicology, Lewis Publishers, Boca Raton, New York. 28. Kettrup, A., Steingberg, C., and Freitag, D. (1991) kotoxikologie - Wirkungserfassung und Bewertung von Schadstoffen in der Umwelt, kotox, UWSF-Z, Umweltchem, 3, 370-377. 29. Tremblay E. (1985) Entomologia applicata, volume primo (generalit e mezzi di controllo), Liguori Editore. 30. Hengel, M.J., and Shibamoto, T. (2002) Method Development and Fate Determination of PesticideTreated Hops and Their Subsequent Usage in the Production of Beer, J. Agric. Food Chem. 50(12), 3412-3418. 31. Lazaro, R., Herrera, A., Arino, A., Conchello, M.P., and Bayarri, S. (1996) Organochlorine Pesticide Residues in Total Diet Samples from Aragn (Northeastern Spain), J. Agric. Food Chem. 44(9), 27422747. 32. Cabras, P., Angioni, A., Garau, V.L., Pirisi, F.M., Brandolini, V., Cabitza, F., and Cubeddu, M. (1998) Pesticide Residues in Prune Processing, J. Agric. Food Chem. 46(9), 3772-3774. 33. Cabras, P., Angioni, A., Garau, V.L., Melis, M., Pirisi, F.M., Cabitza, F., and Cubeddu, M. (1998) Pesticide Residues on Field-Sprayed Apricots and in Apricot Drying Processes, J. Agric. Food Chem. 46(6), 2306-2308. 34. Cabras, P., Angioni, A., Garau, V.L., Melis, M., Pirisi, F.M., Cabitza, F., and Pala, M. (1998) Pesticide Residues in Raisin Processing, J. Agric. Food Chem. 46(6), 2309-2311. 35. Cabras, P., and Angioni, A. (2000) Pesticide Residues in Grapes, Wine, and Their Processing Products, J. Agric. Food Chem. 48(4), 967-973. 36. Krol, W.J., Arsenault, T.L., Pylypiw, H.M., Jr., and Incorvia Mattina, M.J. (2000) Reduction of Pesticide Residues on Produce by Rinsing, J. Agric. Food Chem. 48(10), 4666-4670.

168
37. 38.

39.

40.

41.

42.

43. 44.

45.

46.

47.

48. 49. 50.

51. 52.

53. 54. 55. 56.

57.

Saitta, M., Di Bella, G., Salvo, F., Lo Curto, S., and Dugo, G. (2000) Organochlorine Pesticide Residues in Italian Citrus Essential Oils, 1991-1996, J. Agric. Food Chem. 48(3), 797-801. Holden, A.J., Chen, L., and Shaw, I.C. (2001) Thermal Stability of Organophosphorus Pesticide Triazophos and Its Relevance in the Assessment of Risk to the Consumer of Triazophos Residues in Food, J. Agric. Food Chem. 49(1), 103-106. Qian, S.S., and Anderson, C.W. (1999) Exploring Factors Controlling the Variability of Pesticide Concentrations in the Willamette River Basin Using Tree-Based Models, Environ. Sci. Technol. 33(19), 3332-3340. Falandysz, J., Strandberg, L., Puzyn, T., Gucia, M., and Rappe, C. (2001) Chlorinated Cyclodiene Pesticide Residues in Blue Mussel, Crab, and Fish in the Gulf of Gdask, Baltic Sea, Environ. Sci. Technol. 35(21), 4163-4169. Meijer, S.N., Halsall, C.J., Harner, T., Peters, A.J., Ockenden, W.A., Johnston, A.E., and Jones, K.C. (2001) Organochlorine Pesticide Residues in Archived UK Soil, Environ. Sci. Technol. 35(10), 19891995. Papastergiou, A. and Papadopoulou-Mourkidou, E. (2001) Occurrence and Spatial and Temporal Distribution of Pesticide Residues in Groundwater of Major Corn-Growing Areas of Greece (19961997), Environ. Sci. Technol. 35(1), 63-69. Woodrow, J.E., Seiber, J.N., and Baker, L.W. (1997) Correlation Techniques for Estimating Pesticide Volatilization Flux and Downwind, Environ. Sci. Technol. 31(2), 523-529. Woodrow, J.E., Seiber, J.N., and Dary, C. (2001) Predicting Pesticide Emissions and Downwind Concentrations Using Correlations with Estimated Vapor Pressures, J. Agric. Food Chem. 49(8); 38413846. Barra, R., Vighi, M., Maffioli, G., Di Guardo, A., and Ferrario, P. (2000) Coupling SoilFug Model and GIS for Predicting Pesticide Pollution of Surface Water at Watershed Level, Environ. Sci. Technol. 34(20), 4425-4433. Cortes, D.R., Hoff, R.M., Brice, K.A., and Hites, R.A. (1999) Evidence of Current Pesticide Use from Temporal and Clausius-Clapeyron Plots: A Case Study from the Integrated Atmospheric Deposition Network, Environ. Sci. Technol. 33(13), 2145-2150. de Wit, C.A, Muir, D.C.G., and Svavarsson, J. (2000) POPs in the Arctic: An Update, in AMAP Report on Issues of Concern: Updated Information on Human Health, Persistent Organic Pollutants, Radioactivity, and Mercury in the Arctic, Arctic Monitoring and Assessment Programme (AMAP), AMAP Report 2000:4, (http://www.amap.no/amap.htm). Scheringer, M. (1996) Persistence and Spatial Range as Endpoints of an Exposure-Based Assessment of Organic Chemicals, Environ. Sci. Technol. 30(5), 1652-1659. Scheringer, M. (1997) Characterization of the Environmental Distribution Behavior of Organic Chemicals by Means of Persistence and Spatial Range, Environ. Sci. Technol. 31(10), 2891-2897. Bennett, D.H., McKone, T.E., Matthies, M., and Kastenberg, W.E. (1998) General Formulation of Characteristic Travel Distance for Semivolatile Organic Chemicals in a Multimedia Environment, Environ. Sci. Technol. 32(24), 4023-4030. Bennett, D.H., Kastenberg, W.E., and McKone, T.E. (1999) General Formulation of Characteristic Time for Persistent Chemicals in a Multimedia Environment, Environ. Sci. Technol. 33(3), 503-509. Eisenberg, J.N.S. and McKone, T.E. (1998) Decision Tree Method for the Classification of Chemical Pollutants: Incorporation of Across-Chemical Variability and Within-Chemical Uncertainty, Environ. Sci. Technol. 32(21), 3396-3404. Hertwich, E.G., and McKone, T.E., (2001) Pollutant-Specific Scale of Multimedia Models and Its Implications for the Potential Dose, Environ. Sci. Technol. 35(1), 142-148. Gan, J., Yates, S.R., Papiernik, S., and Crowley, D. (1998) Application of Organic Amendments To Reduce Volatile Pesticide Emissions from Soil, Environ. Sci. Technol. 32(20), 3094-3098. Hammitt, J.K. (2000) Valuing Mortality Risk: Theory and Practice, Environ. Sci. Technol. 34(8), 1396 1400. Mourato, S., Ozdemiroglu, E., and Foster, V. (2000) Evaluating Health and Environmental Impacts of Pesticide Use: Implications for the Design of Ecolabels and Pesticide Taxes, Environ. Sci. Technol. 34(8), 1456-1461. Mandarsz. T. (2003). Risk-Based Evaluation Of The Surface Cover Technology Of A Red Sludge Waste Disposal Site In Hungary, in: Linkov, I. and A. Ramadan, eds., Comparative Risk Assessment and Environmental Decision Making. Kluewer, Amsterdam (in press).