A P P L I C AT I O N S A N D I M P L E M E N TAT I O N Collaborative Problem Solving Using an Industrial Ecology Approach The New York/New Jersey Harbor Economy-Wide Substance Flow Case Studies Susan E. Boehme, Marta A. Panero, Gabriela R. Muñoz, Charles W. Powers, and Sandra N. Valle Keywords: cooperative approaches economy-wide material flow analysis (EW-MFA) emission reduction industrial ecology mass balance pollution prevention (P2) Address correspondence to: Susan E. Boehme c/o U.S. EPA GLNPO 77 W. Jackson Blvd. (G-17J) Chicago, IL 60604 sboehme@illinois.edu www.iisgcp.org Summary Over the course of ten years, a consortium of stakeholders (the Harbor Consortium) used a collaborative approach to identify viable pollution prevention (P2) strategies for specific contaminants, namely mercury, cadmium, polychlorinated biphenyls (PCBs), dioxins, and polycyclic aromatic hydrocarbons (PAHs) as well as suspended solids entering the New York/New Jersey Harbor. The project, titled “The New York/New Jersey Harbor Watershed Pollution Prevention and Industrial Ecology Project,” in a very conscious and overt way engaged stakeholders in the process of developing P2 recommendations for the Harbor. The industrial ecology (IE) tools applied by the Harbor Consortium include substance flow analysis (SFA), material flow analysis (MFA), and, to a limited extent, life cycle analysis (LCA) and fate and transport analysis (F&T), to quantify and characterize how the contaminants flow through the regional economy and the Harbor Watershed once released to the environment. The application of these scientific tools to five contaminants at such a large geographical scale, within the context of a broad and inclusive stakeholder process, and with the goal of identifying and implementing pollution prevention strategies, led to a wide range of surprising outcomes and lessons learned. Undertaking this IE research with the key institutions and stakeholders at the table resulted in the identification and the implementation of many P2 opportunities.  c 2009 by Yale University DOI: 10.1111/j.1530-9290.2009.00166_2.x Volume 13, Number 5 www.blackwellpublishing.com/jie Journal of Industrial Ecology 811 A P P L I C AT I O N S A N D I M P L E M E N TAT I O N Introduction and Background The New York/New Jersey (NY/NJ) Harbor watershed is a highly complex environment encompassing urban, suburban, rural, and agricultural regions, a population of nearly 14 million people, several large rivers, and a major port (figure 1). And a decade ago, a group of public and private organizations from this region were concerned with achieving twin goals: to continue improvements in the environmental health of the NY/NJ Harbor, without impairing the Harbor’s major contributions to the regional economy (Lifset 2000). To be sure, contamination levels in the Harbor waters and even the sediments had for several decades been dropping. But they remained too high. At the same time, dredging, a major mechanism for clearing the Harbor (for both environmental and shipping purposes), had become far more difficult since ocean disposal of the sediments had been banned. Hence, it was widely believed that additional improvement would now require reductions in the flow of new contaminants to the Harbor from diverse but largely undocumented sources from throughout the watershed—sources whose size and significance were simply not understood. Missing then were three major ingredients of reform: (1) an understandable articulation of the actual problem for Harbor health posed by the major contaminants, (2) a credible way of tracking down and assessing the sources of the new contaminants entering the Harbor, and (3) some effective way of converting the technical data into understandable steps that could be taken to slow or stop the contaminant flow—steps that would have broad public support. At just the same time, the tools to carry out industrial ecology (IE) were, in the 1990s, becoming much more broadly understood. Could this evolving science somehow be harnessed to address the three-part task just described? A workshop was held by the New York Academy of Sciences (NYAS) to evaluate its application to the NY/NJ Harbor (NYAS 1998). IE was believed to be a promising approach to understand the Figure 1 Map of the New York/New Jersey watershed (indicated by shading). 812 Journal of Industrial Ecology A P P L I C AT I O N S A N D I M P L E M E N TAT I O N ongoing Harbor watershed contamination more holistically, using the research and data previously collected and the ongoing studies of contaminants in the NY/NJ Harbor (e.g., the Contaminant Assessment and Reduction Project [CARP] for the NY/NJ Harbor; the NY/NJ Harbor Estuary Program; U.S. EPA EMAP and REMAP; and others). Early work by Ayers (Ayers and Rod 1986; Ayers 1989) laid the groundwork by applying industrial ecology tools to the NY/NJ Harbor. Industrial ecology would be able to provide the scientific and technical basis to devise mitigating steps but would not be sufficient to bring about the social acceptance of these measures. Previous environmental improvements had come from the top down through a process of regulation and enforcement. In contrast to this historical approach, the NYAS chose a more inclusive approach and convened a consortium of stakeholders composed of individuals and organizations who were likely going to be asked or required to change their practices. Stakeholders were involved throughout the process, from the first steps of choosing the contaminants through the finalization of recommendations for their abatement.1 From this vision, the NYAS Harbor Project was born, and the Harbor Consortium was established (NYAS 1998). The NY/NJ Harbor Consortium worked to achieve two goals: • Apply the analytical tools of industrial ecology such as material flow analysis (MFA)/substance flow analysis (SFA), starting with all available science to understand, quantify, and address this ongoing pollution, emphasizing a systems view. • Identify and implement strategies and policies to prevent the continuing loadings of five contaminants of concern to the Harbor—mercury, cadmium, polychlorinated biphenyls (PCBs), dioxins, and polycyclic aromatic hydrocarbons (PAHs)— and suspended solids to the NY/NJ Harbor watershed. During this ten-year period, there were significant changes occurring in the political, economic, and environmental arenas both locally and in the nation as a whole. Two major Superfund sites in the watershed,2 one contaminated mainly with dioxins (Diamond Alkali Superfund Site) and the other with PCBs (Hudson River PCBs Superfund Site) reached major milestones amid significant public interest and debate. Some of the key players in these milestones were members of the Consortium. The U.S. EPA recognized that fish consumption advisories should be based on methylmercury not total mercury concentrations in fish, and therefore the project’s SFA of methylmercury for the watershed was intensely scrutinized (C. de Cerreño et al. 2002). The Consortium also was acutely aware that any recommendations put forth would likely require funding to implement. When possible, the economic impacts of the recommendations were estimated (Panero 2005). Thus, the SFA was just the first step in getting to the ultimate goal of identifying recommendations for stemming the flow of contaminants to the Harbor. Those recommendations were placed into the context of available P2 and management strategies and within the social and economic context of the Harbor watershed. SFA was the main analytical tool used during these studies. Production, usage, disposal, and recycling patterns, as well as exports and imports, of the five contaminants were quantified to understand their flow through the watershed and into the Harbor. The specifics of the SFA have been previously described in the NYAS publications (C. de Cerreño et al. 2002; Boehme and Panero 2003; Panero et al. 2005; Muñoz and Panero 2006; Valle et al. 2007; Muñoz and Panero 2008). Pollution prevention was defined as all approaches that potentially decrease the amount of a contaminant entering the Harbor watershed, including activities such as recycling and reclamation. Traditional application of SFA/MFA has focused on identifying quantities and pathways of substances through either an economic, commercial, or industrial process, or geographical framework (e.g., Ayers and Rod 1986; Ayers 1989; Stigliani et al. 1993; Hansen and Lassen 2003; Cain et al. 2007). The NY/NJ Harbor project combined several approaches of IE and utilized them in the larger context of pollution prevention and decision making through a stakeholder Boehme et al., Collaborative Problem Solving Using an Industrial Ecology Approach 813 A P P L I C AT I O N S A N D I M P L E M E N TAT I O N process. Some of the key components were as follows: 1. Scale and framework: Although SFA is often applied at the product, facility, or industry level, in this case it was applied to an entire watershed to quantify the flows of contaminants to the NY/NJ Harbor. The same analytical framework (in particular, input/output) also was used to develop a Harbor-wide mass balance (MB) for each contaminant, providing a second estimate of the contaminant flows. 2. Substances: MFA/SFA is typically applied to products or chemicals that have economic value. This was not necessarily the case with the substances evaluated in this study. Two groups of contaminants evaluated in this study (dioxins and most PAHs) are by-products of other processes and were never intentionally produced. The first three contaminants also had unique histories: PCBs have been banned from production for more than 20 years, mercury is being phased out of products, and cadmium is still a commodity, but its uses have changed dramatically over the last 20 years. Conducting this type of analysis for chemicals that are not produced or traded in the regional economy limited the available data (NYAS 2008); however, the analyses did provide useful information in developing P2 strategies. 3. Context: The results of the MFA/SFA were used to evaluate and prioritize recommendations based on loads to the Harbor and likelihood of action based on the current political/social/economic condition. Using the MFA/SFA tools in decision making is exactly what Brunner asked researchers in the field of industrial ecology to do in 2002 (Brunner 2002). 4. Stakeholder process: The Harbor project was an inclusive program where a wide range of potential solutions emerged from a collaborative consultation process involving public, private, and academic institutions as well as citizen representation within the New York metropolitan area. The Consortium’s involvement—as deci814 Journal of Industrial Ecology sion makers, data providers, evaluators, reviewers, supporters, and implementers—is the main reason why this project has led to implementation of P2 and why this approach was identified as a new way of doing science (NYAS 2008). Key Factors in the Harbor Project Process Focus on Pollution Prevention at the Source Assessing the inflow of contaminants to the Harbor via mass balance rendered a broad view of the type and magnitude of contaminant pathways (e.g., air deposition, stormwater runoff) entering the NY/NJ Harbor, but was inadequate to determine their primary sources. SFA was applied to identify specific contaminant pathways and also to paint a more holistic picture of the pollution problem—tracing each contaminant back to where it was generated and/or released, and using that knowledge to prevent or diminish further environmental releases. An Engaged and Curious Harbor Consortium The Consortium was born out of the original invitees to the 1997 Workshop (NYAS 1998) and grew into a broad consultative forum for Harbor pollution-related issues, in which more than 70 institutions participated (see NYAS 2008 for a list of individuals and organizations). The consortium included representatives from academia; consulting firms; industry and labor; professional, business, and trade organizations; environmental, advocacy, and community organizations; and regulatory and other public agencies at the municipal, state, and federal levels. Consortium members learned from the experts (e.g., the industries who used or generated the chemicals, the biologists and chemists who studied their impacts, and the regulators authorized to control them) about a wide range of topics—from contaminants in automobile switches to those found in rechargeable batteries, transformers, driveway sealants, and to the processes used at the region’s wastewater treatment plants (WWTP), A P P L I C AT I O N S A N D I M P L E M E N TAT I O N to name just a few. This opportunity to learn from recognized authorities on a wide variety of issues drew many individuals to the process. Data Sources The first step taken by the consortium was to evaluate data richness for approximately 12 contaminants that were identified in fish consumption advisories or found in high concentrations in Harbor sediments. This allowed the consortium to evaluate if there were sufficient data for a full study of its priority contaminants (mercury, PCBs, dioxins), and it was hoped that by starting with data-rich contaminants such as mercury, the methodologies would be worked out to address contaminants with fewer data available (Asaf-Anid 2003). The consortium selected the contaminants in a stepwise fashion, as the previous contaminant work was nearing completion. Thus, the results reflect the data available in the years prior to the publication of each of the six reports. In a few cases, the Harbor Project took on the collection of new data when deemed imperative to the MFA. For example, there were no available data on how many mercury blood pressure devices (sphygmomanometers) were in a typical hospital. A local New York City hospital was surveyed to estimate the number of devices per hospital bed. Using the data from this hospital to develop appropriate questions, a written survey was submitted to 16 local hospitals, and those data were used to scale to the entire watershed. This survey, as well as other related research (DelConte 1997), clarified the practices by hospitals of buying products such as mercury sphygmomanometers (typically less expensive than their nonmercury counterparts) as part of multihospital purchasing arrangements. At that time, hospital purchasing did not factor in the significant costs associated with the hazardous waste cleaning requirements when a device is broken or the loss of use of the room for a significant amount of time after the break. These results were presented to a meeting of New York hospital equipment purchasers. In 1998, the U.S. EPA signed a P2 memorandum of understanding with the American Hospital Association, including minimizing the use of mercury-containing products (U.S. EPA 1998). The New Jersey Department of Environmental Protection (NJDEP) undertook field studies (Aucott et al. 2003) to estimate the release of mercury from the breakage of fluorescent lamps and from smelting of automobiles containing mercury switches. The New York Department of Environmental Conservation, as part of the Contaminant Assessment and Reduction Project (Litten 2003), was able to respond rapidly to some data gaps identified during the Harbor Project and added samplings of contaminant data of runoff from roads for PCBs and of methylmercury releases from wastewater treatment plants. Analytical Methodology and Products The main analytical tools used for the fivecontaminant industrial ecology analyses were material flow analysis (MFA) and substance flow analysis (SFA), and mass balance (MB) (figure 2). In a few cases, fate and transport (F&T) and life cycle assessment (LCA) also were used (Schechtman 2007). The watershed (and in some cases the larger airshed) provided the boundary conditions to evaluate the flows of the contaminants of concern, but the goal was to focus on the contaminants that were reaching the NY/NJ Harbor itself. A MB assessment and SFA were developed for each of the five contaminants—mercury, cadmium, PCBs, dioxins, and PAHs. The five contaminant-specific reports describe how the assessments were developed and provide an estimate of the size and sources of those flows to the Harbor (C. de Cerreño et al. 2002; Boehme and Panero 2003; Panero et al. 2005; Muñoz and Panero 2006; Valle et al. 2007). For PAHs, an F&T model also was developed to evaluate how the contaminants released throughout the watershed region mobilize toward the Harbor and whether they are stored in permanent or temporary reservoirs. The first step was to develop an inventory of all generic sources of the contaminant. The second step was to estimate the amount mobilized/released through various economic processes and/or sectors and during product use and disposal. The scope of this inventory was the regional watershed economy, and the temporal Boehme et al., Collaborative Problem Solving Using an Industrial Ecology Approach 815 A P P L I C AT I O N S A N D I M P L E M E N TAT I O N Figure 2 Flows estimated to establish the New York/New Jersey Harbor mass balance. CSO = combined sewer overflow. frame was one year. This provided a picture of the total mass of the contaminant that enters the regional environment. Following the recommendation of Wernick and Ausubel (1997), weight (mass) was used as the common denominator to account for physical quantities of materials. Concentration coefficients were applied to processes and products to estimate the quantities used either as raw material or embedded in products, describing how each contaminant is mobilized through various stages of interaction between economic and environmental systems (figure 3), from extraction to production, consumption, reuse or recycling, and disposal. A budget for contaminants released to the environment also was developed, using coefficients to account for different transport rates to receptors (the NY/NJ Harbor) according to the media of primary release. Flows of pollutants that also are commodities (mercury, cadmium, PCBs, and, in a few cases, PAHs) were estimated based on economic census and survey data. For pollutants that are inadvertently produced (dioxins, most PAHs, and some PCBs), estimates were developed by, for example, multiplying the activity level (AL; e.g., 816 Journal of Industrial Ecology metric tons [t] of solid waste burned in barrels) by an appropriate emission factor (EF; e.g., grams of dioxins released per t of waste burned). These ALs and EFs ranged widely in their confidence levels, depending on the emission source. Although the best available EF and AL information was used, some data were incomplete and/or outdated (e.g., vehicle emissions may not reflect the current fleet characteristics), and/or there were uncertainties with the data ultimately impacting the uncertainty of the emission estimate. These estimates were used only to assess the range of loadings to the Harbor. In the case of dioxins, emission estimates were assigned a high, medium, or low confidence rating. For many industrial sources, EFs are fairly well characterized (usually available for types of technologies—e.g., for specific incinerator models and air pollution controls—and even for individual facilities), and the activity level is available from facility reports to environmental agencies. These estimates were assigned a high confidence rating. Sectors for which EFs were based on numerous tests, but where no facilityor technology-specific data were available, were rated medium. Poorly characterized sources with A P P L I C AT I O N S A N D I M P L E M E N TAT I O N Figure 3 Relationship between the material/substance flow accounting and the mass balance analysis. Ind./Comm. Services = Industrial or commercial services. uncertain activity levels and/or EFs were assigned a low confidence rating. This was the case, for example, for structure fires, where the types and amounts of materials burned in any given incident can only be roughly estimated, and the EFs are greatly uncertain because the conditions in any given fire can vary widely. A similar system was applied to PAH emissions. Many of the emission factors used were obtained from the U.S. EPA AP-42 database (U.S. EPA 1995), which gives a rating to emission factors ranging from A to E, with A being the best. The rating is a general indication of the reliability or robustness of the factor. Relying on a simple metric of physical quantities released to support environmental decision making may not result in the best riskmanagement strategies, as was the case for mercury, which has the potential to become more toxic if it is exposed to conditions that promote methylation. Thus, pathways of mercury release were weighted based on probability of methylation (e.g., wastewater treatment facilities provide ideal conditions for methylation). Therefore, a kilogram of mercury discarded to solid waste may not have an effect equal to that of a kilogram released to wastewater. Panero (2005) describes how factors such as the pollutant’s fate and transport or its toxicity characteristics can be used to develop a multicriteria framework to prioritize benefits associated with emissions reduction. As an example of the process used to develop emission inventories, the case of mercury is described: 1. Compile a generic list of potential sources following the blueprint of sources through a literature review. 2. Determine coefficients for use and release by reviewing various data sources such as engineering reports (e.g., Barr Engineering 2001), material flows assessments (Sznopek and Goonan 2000), and national data on anthropogenic mercury emissions (U.S. EPA 1997), and contacting experts on specific products (dental sector release estimates from several organizations, fluorescent light industry, automobile recyclers, and metal scrappers). This stage continued until all leads on data sources were tracked down. 3. Develop a regional inventory of usage based on economic data. Depending on Boehme et al., Collaborative Problem Solving Using an Industrial Ecology Approach 817 A P P L I C AT I O N S A N D I M P L E M E N TAT I O N availability of data, different approaches were used to estimate regional flows: a. Where available, regional census and local survey data were used to estimate specific flows of mercury in the NY/NJ Harbor watershed region. Data sources included regional statistics on production processes (U.S. Bureau of Economic Analysis), data on consumption of various products containing mercury (market research), and survey information (e.g., the number of mercurybearing products used by hospitals). For example, once dental facilities were identified as a generic source, U.S. Census data were used to quantify the number of dental facilities and business volume, and state agencies (NY State Education Department, NJ Division of Consumer Affairs) provided the number of licensed dental practitioners. National survey information provided estimates for the number of dental operatory chairs and rates of mercury amalgam placement and removal per establishment (Water Environment Federation 1999; Berglund and Diercks 2001). Market research data were also used to validate rates of mercury usage. b. Where local data were not available, regional flows were estimated from national estimates adjusted by the level of regional economic activity (e.g., the Harbor watershed region represented approximately 5.8% of the national economy but only 5.2% of the population [U.S. Census 2000]). 4. Determine the discharge and release rates for different products and processes. Contaminant release rates (e.g., for mercury) were determined from several sewerage districts (Lochan 1999; Water Environment Federation 1999; Association of Metropolitan Sewerage Agencies 2000), and the U.S. EPA has published estimates of solid waste composition (U.S. EPA 1992). Engineering reports (Barr Engineering Co. 2001) were used to estimate discharge rates of mercury to wastewater, to air, and as solid waste. 818 Journal of Industrial Ecology The IE analysis described above (and shown in figure 3) provided a picture of the contaminant flows by quantifying the amounts of pollutants in products and processes that are mobilized through different pathways and sinks. The fate of the contaminants at the end of product life was estimated from data on waste handling in the region, via air deposition (when waste is incinerated), from land (runoff, groundwater, or leachate from landfills/contaminated sites), or via wastewater or direct water discharges. In the case of mercury, the inventory estimated that approximately 10,800 kg of mercury from products and processes were likely to be released per year within the Harbor watershed, and a significant proportion of it makes its way into the Harbor (C. de Cerreño et al. 2002). 5. Assess independently how much of each contaminant flowed into and out of the Harbor from different media of release— water, air, and land—with the MB assessment (figure 2). Wastewater treatment facilities are typically required to measure and publish data on releases to waterways and the U.S. EPA, and both New York and New Jersey have rigorous sampling programs for inputs of several contaminants to the Harbor. For example, the MB for mercury estimated contributions to the Harbor via air deposition, tributary inputs, stormwater runoff, combined sewer overflows (combined sewer overflows [CSOs], overflow pipes), as well as effluents from landfills and wastewater treatment facilities (Fitzgerald and O’Connor 2001; and summarized in C. de Cerreño et al. 2002). The data were also used to estimate removal rates of mercury from the Harbor via volatilization, advection, resuspension, and sediment dredging. For all five contaminants, consultants hired by NYAS (research scientists from local universities whose research was related to the specific contaminant of interest) undertook the mass balance assessments. This resulted in an independent estimate of the mass balance from the IE analysis for comparison. A P P L I C AT I O N S A N D I M P L E M E N TAT I O N 6. Compare the IE and MB assessments. This integration of research findings from two different perspectives was quite useful to evaluate the uncertainty in the IE analysis and led to deeper understanding of the fate of several contaminants. For example, in the case of cadmium, the two approaches resulted in different estimates of how much cadmium was actually getting to the Harbor. Looking closer at the chemistry of cadmium showed that it was much less likely to be particle-bound as it moved from freshwater to estuarine/marine conditions. Thus it is likely that the cadmium entering the Harbor may be flushed out to the New York Bight and eventually the Atlantic Ocean. This is in contrast to mercury, where the Harbor is acting more like a bathtub because much of the mercury coming in to the Harbor is particle-bound and is deposited with the sediments (C. de Cerreño et al. 2002; Boehme and Panero 2003). 7. Identify P2 strategies. The results of the IE and MB were used to prioritize P2 recommendation on the major sources of contaminants reaching the Harbor. Typically P2 strategies were identified for as many releases as could be identified; the key recommendations focused on those sources reaching the Harbor. P2 strategies were identified in several ways: Many local and regional organizations, state, and federal agencies exist that have active P2 programs, including the New York State Department of Environmental Conservation and the New Jersey Department of Environmental Protection. These strategies were researched, evaluated, and modified as needed to fit the specific sources and conditions in the NY/NJ Harbor Watershed. The Consortium also developed sector-specific P2 recommendations through consultation with industry representatives. For example, mercury switches prevalent in older-model automobiles were identified as a source of atmospheric mercury when the cars were shredded and smelted at the end of life for new steel production. A subgroup of mercury experts, automobile scrap metal dealers, state regulators. and metal recyclers from the region worked to establish implementable recommendations that would most efficiently achieve removal of switches before scrapping the vehicles at the end of life while more fairly spreading the burden of associated costs (C. de Cerreño et al. 2002). Finally, when no precedent was available on how best to prevent releases from a specific sector, process, product, or activity, the consortium involved various parties and experts in a consultative process to develop recommendations. In general the three-tiered approach of education, implementable alternatives, and regulation were part of many of the P2 recommendations. 8. Recommend actions, and identify who should implement them. The recommendations were debated within the consortium and carefully worded to ensure that they honestly reflected the level of certainty around the research and that each recommended action was prioritized with respect to the likelihood of the pollutant to enter the Harbor (NYAS 2008). The ranking of recommendations was based on as many “valuation lenses” as could be defined in the process. Factors such as environmental benefit, costs (including potential avoided costs), likelihood of success, and long-term impacts were considered. For example, in the case of mercury, the recommendations were coupled with a cost analysis that demonstrated very clearly that it was much less expensive to stem the flow of mercury as close to its source as possible rather than end of pipe (e.g., at the waste water treatment plant). In other cases, the Consortium considered the specific regional conditions to rank its recommendations. For example, recommendations on dioxins minimization from a highly contaminated Superfund site within the watershed are likely unique to this region, and other watersheds might prioritize their recommendations very differently. This suggests that it is the process used rather than our specific conclusions Boehme et al., Collaborative Problem Solving Using an Industrial Ecology Approach 819 A P P L I C AT I O N S A N D I M P L E M E N TAT I O N that will be most useful to other regions and contexts. Results and Discussion Utilizing industrial ecological methods to quantify and identify pathways of contaminants to the NY/NJ Harbor resulted in many new findings. The major findings, recommendations, and actions for the five contaminants are highlighted below. Mercury (Hg) The mass balance analysis considered mercury released to the air, land, and water within the watershed. Pathways to the Harbor’s waters include tributaries, wastewater treatment plants (WWTPs), combined sewer overflows (CSOs), and stormwater runoff. Because bacteria in aqueous environments can convert mercury (Hg) into a potent neurotoxin—methylmercury (MeHg)— a MeHg MB was also developed. To better understand the key sources of methylmercury, NYAS commissioned research and held meetings to evaluate which specific environmental conditions favored the methylation process. Based on research findings, coefficients were developed to estimate MeHg flows from MB vectors (WWTPs, atmospheric inputs, and so on). The coefficient used for inputs and outputs from wastewater treatment plants was shown to be in relatively good agreement with actual measurements from these facilities (C. de Cerreño et al. 2002) and revealed WWTPs as a significant source of MeHg to the Harbor. These results pointed to inputs to wastewater as a priority for P2 and emphasized the importance of understanding releases of mercury at the source (e.g., dental facilities, hospitals, laboratories), each of which was individually discharging small amounts of mercury to wastewater treatment facilities. Furthermore, by focusing on the flows of a substance throughout the region (rather than on individual sources), the analysis showed that cumulatively these “small amounts” were not trivial. Sectors typically characterized to be small quantity generators (SQGs) at the individual facility level turned out to be major contributors of mercury releases when evaluated at the level of the entire industry or at the sector-wide level. 820 Journal of Industrial Ecology This was also the case for other contaminants. The importance of SQGs was a key finding because these generators typically are unregulated, and their releases are not quantified. The three major sectors—dental facilities, hospitals, and laboratories—were given the highest priority to reduce loadings of mercury to the Harbor. The Consortium devoted significant efforts to careful examination of how best to intercept mercury flows from these three sectors. This focus was supported by the economic analyses that demonstrated significant cost savings for controls at the source versus controls at the wastewater treatment plant (C. de Cerreño et al. 2002; Panero 2005). The second priority focused on atmospheric inputs to the Harbor. Mercury deposition onto the Harbor watershed is linked to emissions from incineration of mercury-bearing products and combustion of fuels that contain trace amounts of mercury as well as through volatilization. About one third of the mercury released locally to the airshed is deposited on the watershed and is available to be washed into the rivers and streams, thereby making its way to the Harbor. This region has significantly reduced local power generation from coal and oil combustion and releases from local utilities were estimated to be lower than in other regions. There are extra-regional inputs mainly from coal combustion in the Midwest as well as mercury deposition from global emissions. Thus, the solutions to atmospheric inputs extend beyond the watershed. Finally, the disposal of products and solid waste in landfills constitutes the largest mercury pathway. However, because tentative estimates suggest that this mercury is sequestered (Aucott 2006) and therefore poses a lower direct risk to the Harbor (at least for some length of time), this pool was accorded the lowest priority despite its large size. Implementing P2 for wastewater mercury inputs, however, has the added benefit of decreasing landfill inputs by more than 60% because the three critical sectors (dental, hospital, and laboratory) are major contributors to this pool as well. Cadmium (Cd) The mass balance indicated that there appears to be a rapid and ongoing “cleansing of the A P P L I C AT I O N S A N D I M P L E M E N TAT I O N Harbor” with respect to cadmium because of— its propensity to become dissolved rather than particle-bound when in contact with increasing salinities; improvements in wastewater treatment, which reduced cadmium concentrations in WWTP effluents; regulations controlling the releases of cadmium from industry; and the migration of cadmium-utilizing industries away from the watershed region (specifically, metal plating and metal finishing factories; Boehme and Panero 2003). Nevertheless, cadmium continues to enter the Harbor via wastewater and nonpoint sources, but, based on our current understanding, sediment concentrations in much of the Harbor are below the level where negative impacts are seen in organisms. By contrast, the SFA/MFA paints a picture of continuing flows of cadmium to the region, but there was a paucity of data to understand its fate and/or disposal. The lack of information about the fate of products containing cadmium at end of life (e.g., nickel-cadmium batteries) and as a trace contaminant in biosolids and fossil fuels led to a larger discussion on data collection at the national and state level. This research was underway at the same time as the U.S. EPA was considering collecting even fewer data as part of its Toxics Release Inventory (US EPA 2007) requirements. Undertaking MFA would only become more difficult if data collected by TRI, the U.S. Census, and the U.S. Geological Survey are limited. Not only was there scant data to understand the major cadmium pathways, major shifts were taking place in the usage of cadmium in the previous 15–20 years, from pigments and metal plating to nickel-cadmium batteries (90% in batteries during the early 1990s and then gradually decreasing to about 75%–80% currently). Data collection efforts did not keep pace with this trend. Data collection at the federal level has changed in the last 20 years, resulting in less information to understand battery imports, especially when the batteries are already in products. Despite an industry-supported voluntary battery-recycling program in the United States (Rechargeable Battery Recycling Corporation), there are no published recycling rates and no regional disposal rates. Therefore, two regional surveys were conducted by the project. These surveys showed that more than half of the non- commercial batteries were not being recycled in the watershed. The implications of changing usage and disposal patterns mean that it is difficult to estimate impacts on the Harbor. This was further complicated because New York City exported a good proportion of its solid waste to other regions; therefore, the potential impacts were likely being exported as well. In New Jersey, where the solid waste is being managed locally, the New Jersey Department of Environmental Protection (NJDEP) has seen moderate increases in cadmium levels in the last 10 years in bottom ash from waste-to-energy and electric arc furnace facilities (M. Aucott, personal communication). This may reflect the fact that increasing amounts of products containing nickel-cadmium batteries are reaching the end of life and are being discarded, most likely in the trash. Similarly, research was lacking to fully evaluate the potential environmental impact of biosolids, which are most often land-applied as fertilizer but can also be incinerated in wasteto-energy facilities or landfilled. The land application of biosolids as fertilizer continues to be debated due to concerns about trace contaminants such as cadmium that are always present in biosolids and may be transferred to the food chain. Because there were no longer significant impacts from the inputs of cadmium to the Harbor, the Consortium did not provide recommendations in the same manner as with the mercury report. Pollution prevention strategies were identified for the largest inputs, with the caveat that based on the current state of knowledge, there were not immediate concerns about cadmium flows to the Harbor. Improving batteryrecycling rates in the watershed, however, serves both precautionary goals and the goals of resource conservation. Polychlorinated Biphenyls (PCBs) Despite the 1970s ban on manufacturing and commercial distribution of PCBs, significant regulation, and research, PCBs continue to be redistributed and dispersed through processes such as improper disposal, inadvertent production, mobilization, and volatilization. The molecular weight differences among the 209 different PCB Boehme et al., Collaborative Problem Solving Using an Industrial Ecology Approach 821 A P P L I C AT I O N S A N D I M P L E M E N TAT I O N compounds (congeners) add an additional layer of complexity because those differences affect not only their toxicity but also how they behave and move in the environment. PCBs are a group of manmade chemicals that are identified as persistent bioaccumulative and toxic (PBT), which are found in products, soils, water, and the atmosphere across the globe. PCB mixtures of varying degrees of chlorination were produced for different commercial purposes. Thus data on the PCB congeners found in the environment can be useful in differentiating sources. For example, inputs of PCBs from the Upper Hudson River (and the upstream superfund site) had a different fingerprint than those entering the Harbor from local inputs. It was estimated that the upper Hudson loadings represent greater than 50% of the total inputs to the NY/NJ Harbor and include the Hudson River Superfund site (30%–40% of the ≥50%), other Superfund sites (e.g., Hastings, Fort Edward, and Hudson Falls), other contaminated sites and brownfields, and inputs from floodplains, dredge spoils, and remnant deposits. The MFA was used to describe the other half of the picture, namely that there was significant PCB remobilization and secondary dispersal that enters the Harbor through air deposition, runoff, or from wastewater. Such remobilization may be the result of some or all of these factors that were identified through the MB–MFA process (Panero et al. 2005): 1. Leaking from transformers and capacitors still in use. 2. Improper disposal at end of life of PCBcontaining products. 3. Leaking from mineral oil transformers and/or capacitors contaminated with PCBs during oil replacement because clean oils were transported in the same trucks as PCB oils. 4. No regulation for products containing small amounts of PCBs (e.g., small capacitors used in appliances or fluorescent lamp ballasts) despite significant cumulative volume when all small PCB-containing products are considered. Thirty years ago it was estimated that the U.S. stock of small capacitors was about 870 million units containing approximately 21,000 metric 822 Journal of Industrial Ecology tons of PCBs. Other unregulated products that contained PCBs include caulking, paint, and carbonless paper, and practices such as spraying PCB oils for dust control. 5. Regulatory gaps that prevent adequate monitoring of PCB-contaminated transformers’ proper disposal. 6. Inadvertent generation and release of PCBs during production processes (e.g., manufacturing of certain pigments and dyes and flocculants). The PCBs MFA/SFA/MB demonstrated that the lack of reporting requirements for use and disposal of PCB-containing products has led to huge uncertainties in the amount and location of PCBs still in use, and the fate of PCBs that were already disposed, especially those PCBs in smaller products such as small capacitors. Although it was understood that it was likely too late to require this type of data collection for PCBs (recommendations for ongoing reporting were made for some sources) because dispersal over the last 20 years has been significant, there was an important lesson to learn for other substances that are being considered for regulation. The Consortium recommended that when a ban includes exclusions for certain uses and inadvertent production, as is the case for PCBs, then mandatory and comprehensive inventories, especially of disposal practices, should also be required. Dioxins and Furans (Dioxins) Chlorinated dioxins and furans are a group of 210 compounds, always found as mixtures, which have been classified as persistent organic pollutants (POPs) and PBTs. Of these, 17 congeners are the most studied because of their higher toxicities and were thus the focus of our research. In addition, many other substances possess “dioxin-like” toxicity and are often grouped with dioxins, in particular co-planar PCBs. Dioxins are inadvertently generated in a variety of chemical manufacturing processes involving chlorine and during incomplete combustion. These substances were never intentionally produced and therefore are not a commodity that can be tracked with trade and industrial statistics. Emission inventories and/or releases A P P L I C AT I O N S A N D I M P L E M E N TAT I O N were based on emission factors and level of activity. Dioxins were selected by the Consortium because of their impacts on fish and shellfish as well as fish consumption advisories (NJDEP and NJDHSS 2004; U.S. EPA 2004; NYDOH 2005) in the NY/NJ Harbor watershed, their relatively high toxicity even at low concentrations, their ubiquity in sediments in the Harbor (e.g., the lower Passaic River and Newark Bay), and thus their potential impact on the economy of the region, especially in causing dredging restrictions associated with the Port of New York and New Jersey. The major ongoing sources of dioxins to the Harbor are associated with combustion processes (mainly uncontrolled combustion of waste and other materials), including backyard burning (also referred to as burn barrels), fires at waste management facilities, landfill gas management, structural fires, wood burning by industrial, commercial and electric facilities, steel recycling, and medical waste incineration. The MFA for the NY/NJ Harbor showed (with high uncertainty) that fires at trash management facilities could account for substantial releases of dioxins. This source had been largely neglected in other emissions inventories (Muñoz and Panero 2006). Greater than 90% of the dioxin toxicity in the Harbor is accounted for by three dioxin/ furan compounds (2,3,7,8 tetrachloro dibenzo dioxin [TCDD], 2,3,7,8 tetrachloro dibenzo furan [TCDF], and 2,3,4,7,8-pentachloro dibenzo furan [PeCDF], NJDEP 2006) and these compounds were given highest priority for action. In many cases, because of the relationship between PCBs and dioxins, the same recommendations were highlighted. Three key recommendations were as follows: 1. Proper management of PCB fluids (which always contain furans). 2. Waste minimization and reduction of waste combustion because dioxins are produced during combustion processes. 3. Cleanup of contaminated sites. Much of the dioxin inputs to the Harbor are likely the result of remobilization of former dioxins releases (e.g., a Superfund site affecting sediments within the Harbor and myriad contaminated land sites). As was the case for PCBs, there is a major dioxin-contaminated Superfund site within the NY/NJ Harbor. The Diamond Alkali Superfund site is a significant source of dioxins to the rest of the Harbor. During the Consortium’s recommendation deliberations, there were several legal steps occurring in the Superfund process. The Consortium urged all parties to focus their efforts on achieving early and effective action and implement the Harbor Project’s P2 recommendations to reduce inputs to the Harbor. Polycyclic Aromatic Hydrocarbons (PAHs) PAHs are a group of more than 100 compounds with differing physical-chemical properties. PAHs have been identified as carcinogenic and have been shown to have harmful effects on human and ecological health. The rate at which individual PAHs are generated/released and break down in the environment varies, making tracking and evaluating them even more difficult. PAHs are ubiquitous and have both natural and anthropogenic sources, and are primarily formed during incomplete combustion of fossil fuels and organic substances. The major releases of PAHs to the Harbor watershed identified in the IE analysis were due to combustion processes (wood combustion in wood stoves) volatilization from materials containing PAHs (creosote treated wood), and releases from the use of products containing PAHs (e.g., parking lot sealants) (Valle et al. 2007). In the case of PAHs, sources were categorized according to their intersection with the regional economy. This organization of sources highlighted where P2 strategies could impact more than an individual source. For example, by reducing vehicle use, releases of PAHs from engine combustion, tire wear, and leaking motor oil would be decreased. This perspective also pointed out where technological innovation versus land-use planning/infrastructure could impact PAH releases (Valle et al. 2007). When the mass balance was compared with the primary emissions inventory, the discrepancies highlighted the complexity of what happens to PAHs during their migration to the Harbor. This was not surprising because some Boehme et al., Collaborative Problem Solving Using an Industrial Ecology Approach 823 A P P L I C AT I O N S A N D I M P L E M E N TAT I O N PAH compounds have very short half-lives in the environment and can degrade quickly into other compounds. The previous four chemicals were either elements (Hg, Cd) or highly persistent (PAHs, dioxins). Therefore an F&T was developed (as summarized in Valle et al. 2007) to account for degradation of PAHs in the environment and transport (e.g., deposition, runoff, fluvial transport) and was used to estimate the quantity of PAHs reaching the Harbor. The IE and F&T assessments rendered a different order of P2 priorities based on those having the greatest estimated impact on the Harbor. For example, the IE assessment identified combustion processes as the major primary emission sources of PAHs, while the F&T analysis identified releases to impervious land, such as motor oil leaks, coal tar sealed surfaces, and tire wear, as the major sources to Harbor loadings. In this case, the MB assessment compared well with the F&T model. Barring a coincidence, this comparison helps constrain the uncertainty surrounding the SFA and F&T analysis. Based on the grouping of PAH sources by economic impact noted above, the Consortium called for reducing or removing PAHs from products such as driveway and parking lot sealants, and creosote used to preserve marine pilings, railroad ties, and utility poles. On the demand side, reducing fossil fuel consumption and reducing miles driven were recommended. Product efficiency improvements for combustion devices, implementing best management practices (BMPs) to prevent leakage of used motor oil, and stormwater management measures were also recommended. Advantages to Applying an MFA/SFA Approach to These Five Contaminants Inventories based on MFA/SFA of specific contaminants can help to characterize both largeand small-quantity generators. This is significant, given that cumulative releases from smallquantity generators may be a significant source to a system of concern such as the NY/NJ Harbor (e.g., mercury releases from dental facilities or PAHs releases from coal tar parking lot sealant applications). 824 Journal of Industrial Ecology In general, the majority of large-quantity generators of these five contaminants have been identified, and their releases have been addressed through regulation. The focus of contaminant reduction efforts is now on the smaller and nonpoint diffuse sources. MFA/SFA approaches were shown here to be excellent tools for identifying pollution from nonpoint or dispersed sources. In addition, with the improved analytical methods for detecting contaminants and a growing body of research showing impacts at smaller and smaller exposure levels, there is a greater need to identify all sources, small and large. Coupling IE models with a better understanding of the ecological conditions under which these contaminants are released has also been key to understanding localized impacts to the environment. The approach used here can also aid watersheds that are required to establish total maximum daily loads (a total maximum daily load, or TMDL, is a calculation of the maximum amount of a pollutant that a waterbody can receive and still meet water quality standards). The TMDL process involves first determining the maximum amount of a pollutant that can be discharged from all sources into a waterbody without causing impairment to its designated uses. This load is then allocated among the main vectors contributing to a waterbody (WWTPs, other utilities, CSOs, tributaries, and so on), and any necessary load reductions are then established. Most of these vectors simply convey primary emissions originated elsewhere (e.g., WWTPs may receive mercury discharged by dental facilities) or move historic releases to a new vector (e.g., stormwater may erode and carry dioxins attached to soil from contaminated land sites). This IE approach can be used to identify the primary sources of a contaminant, especially when coupled with F&T models that can predict how and how much primary emissions will reach the Harbor. The states of New York and New Jersey are in the process of developing TMDLs for several toxic pollutants for waterbodies within the NY/NJ Harbor watershed. When data on actual pollution sources are lacking, TMDLs may require communities to resort to less effective and costly end-of-pipe solutions. Trackdown approaches like the one used for this project can be applied to identify the primary contaminant A P P L I C AT I O N S A N D I M P L E M E N TAT I O N sources and quantify input rates much more cost effectively. This approach was used successfully in Camden, New Jersey, where soils from industrial sites were identified as the main contributor of PCBs to the sewer system (Belton et al. 2008). The mercury analysis in this study quantified that the costs of trapping mercury at dental offices was orders of magnitude less expensive than collection at the WWTP (C. de Cerreño et al. 2002; Panero 2005). Challenges to Applying an MFA/SFA Approach to These Five Contaminants The greatest challenge to applying this approach was lack of regional and national data, and, in some cases, data quality. Some examples of data issues were as follows: • For chemicals whose production has been banned (even though many of the products in which they were used continue to be allowed), it is not possible to develop release inventories without specific reporting requirements on use and disposal. For PCBs, very little information was available to construct an inventory of PCBcontaining products that are still in use or are being phased out. • A contaminant may change its chemical characteristics once released to the environment and so all forms must be included in the inventory. For example, mercury is most toxic in the methylmercury form, and all the conditions for that transformation occur in WWTPs. In the cases of PCBs, dioxins, and PAHs—each a group of distinct chemicals—it is necessary to understand how each individual compound reacts in the environment as well as any chemical transformations that might occur. This is a very data-intensive approach. • Regional assessments need to account for transboundary sources (e.g., atmospheric transport of mercury from the Great Lakes region) as well as contaminants that are “exported” to other regions (e.g., transport and disposal of solid waste and dredged sediments outside the watershed). • The religious use of mercury was getting a lot of attention from several agencies interested in mercury release at the time the Harbor Consortium was estimating releases to the Harbor. However, the decision was made not to include any of these estimates of releases to the Harbor because it was discovered that usage estimates were based on a biased data set (C. de Cerreño et al. 2002). This discovery led to a very vigilant review of the literature and an effort to provide error bars on all release estimates, even if the only way to do this was to assign a low, medium, or high confidence estimates. The Role and Importance of the Harbor Consortium In a more traditional regulatory framework, environmental science and risk-assessment models are developed independently of policy decision making. The models are then used to develop threshold emission or exposure levels that can be used for regulatory purposes. In this traditional approach there is scientific discovery in terms of chemical fate and transport, chemical exposure, and health risk from that exposure, but communication between researchers and users of the research is minimal until after the research is completed. The Consortium approach was significantly strengthened by ongoing communication between scientists and the users of the research. As a group, the Consortium was able to develop a flexible, evolving, analytical representation for the watershed system and its boundaries. In this framework, the science is able to inform decision making in a distributed sense. As the process evolves, the group is able to identify key strategic opportunities for intervention and also to identify additional scientific questions to pursue. This type of process has led to a rich array of recommendations from the Consortium, many of which have been adopted because key players were involved from the start. This educational process also highlighted the need for P2 to achieve the goals of a healthy, viable NY/NJ Harbor. Another key role of the Consortium (at any given meeting 50 to 60 attendees were present) was as a source of data and objective evaluators of data quality. The Consortium had knowledge of Boehme et al., Collaborative Problem Solving Using an Industrial Ecology Approach 825 A P P L I C AT I O N S A N D I M P L E M E N TAT I O N and access to local and national data; many were authors of regional reports and provided their own firsthand knowledge that was often not documented (e.g., disposed refrigerators that could contain PCB capacitors are crushed at the curb in parts of NY; Panero et al. 2005). Consortium members helped find experts and new stakeholders when the research was stalled by lack of data or a new line of research needed to be pursued. For example, Consortium members from the local power utility companies provided information on the status of their PCB oil transformers (Panero et al. 2005). The implicit support of this large body of experts helped to move regulatory action forward (New York State used the mercury report to require dentists to install chairside mercury traps). The Consortium developed into a unique forum for evaluating the merits of the data and findings, and at times the refutation of that information (NYAS 2008). The products of the Harbor project were reports containing a wealth of information about specific contaminants and a set of P2 strategies, but the next step to implementation requires human will and action. Consortium participants became the implementers by changing the practices in their own companies, institutions, workplaces, and personal lives (see NYAS 2008 for specific examples). Conclusions Many Consortium members came to believe that the evolved combination of methodologies utilized in developing the analyses described here constitute not just a distinctive approach but perhaps an especially effective way of doing science to gather and interpret the data needed for ecosystem-level evaluations of broad geographical areas (such as watersheds). This can be described as a process that put two kinds of scientific approaches in dialogue with each other. This project utilized both the science of industrial ecology and traditional environmental science to link and quantify how pollution is created, by what sources, and what pathways it follows within the Harbor watershed once released to the environment. The key components were as follows: 1. The comparison of the independent industrial ecology (developed using material 826 Journal of Industrial Ecology flow accounting) and mass balance assessments was key to reducing the uncertainty on the estimates of sources of contaminants to the Harbor. 2. Following material flows throughout the entire cycle from cradle to grave can lead to a very different set of recommendations— which are likely to emphasize actions at the source of contamination, rather than waste flows (end of pipe). 3. An objective approach that allowed for deviation/exploration when outcomes of the process led the research in new directions. For example, when it became clear that a significant percent of the contamination entering the Harbor was being transported from upstream sources attached to sediment particles and that these same particles were part of what was driving the need to dredge the Harbor and spend significant amounts of money to dispose of those contaminated sediments properly, the Consortium chose to look at the source of those sediments. This led to the development of a sixth report, “Sources of Suspended Solids to the NY/NJ Harbor Watershed,” (Muñoz and Panero 2008) to better understand how contaminants were being transported from on-land sources toward waterbodies within the watershed and what land uses and processes were responsible for the movement of particles from the land to the water. This work has led to several open questions. Relating present material flows to historical material flows and the state of the ecosystem could lead to the ability to estimate future inputs and forecast future ecosystem viability. This is the type of effort that the CARP was undertaken to accomplish, and coupling it with the IE assessment allows one to track the contaminants farther up the pipe to initial release. IE methods can complement the hard policy decisions that represent our values toward the environment, society, and economy because it informs and lends support and credibility to the policy process. The Consortium IE approach extends the traditional regulatory capability, which serves as a critical baseline for environmental A P P L I C AT I O N S A N D I M P L E M E N TAT I O N protection. But addressing environmental problems by simply controlling pollution is not enough either today or for future generations. The material flows of mercury, cadmium, PCBs, dioxins, PAHs, and many other contaminants are associated with both historical and current “everyday” choices made by a variety of actors both within and outside of our watershed. The application of systems thinking—IE tools within a Consortium—has been a means to recommend solutions to address these complex environmental problems. Acknowledgments The authors wish to acknowledge the longterm financial support of several of the Harbor Project funding organizations, including the U.S. Environmental Protection Agency, Abby R. Mauzé Trust, and the Port Authority of NY/NJ, as well as grants from J.P. Morgan, NYC Environmental Fund, NYS Energy Research and Development Authority, Harbor Estuary Program, Rockefeller Philanthropy Associates, AT&T Foundation, and the Commonwealth Fund. The authors also would like to acknowledge the support of the Harbor Consortium. Notes 1. During the course of the project, several documents were published by the New York Academy of Sciences. These documents as well as background papers, maps, and outreach materials that were developed for the project can be accessed on the New York Academy of Sciences Web site: www.nyas.org/harbor. 2. Superfund is the federal government’s program to clean up the nation’s abandoned hazardous waste sites. It is also the name of the fund established by the Comprehensive Environmental Response, Compensation and Liability Act (CERCLA) of 1980. References Association of Metropolitan Sewerage Agencies (AMSA). 2000. Mercury source control and pollution prevention program evaluation. Davis, CA, USA: Larry Walker & Associates. March 2002. Asaf-Anid, N. 2003. PAHs, suspended solids, PBDEs, endocrine active substances, and lead: Contaminants of concern to the New York/New Jersey harbor. Presented to the New York Academy of Sciences. November 2003. New York: New York Academy of Sciences. www.nyas.org/harbor (select “List of publications and related materials”). Aucott, M. 2006. The fate of heavy metals in landfills: A review. Presented to the New York Academy of Sciences. February 2006. New York: New York Academy of Sciences. www.nyas.org/harbor (select “List of publications and related materials”). Aucott, M., M. McLinden and M. Winka. 2004. Release of Mercury From Broken Fluorescent Bulbs. Trenton, NJ: New Jersey Department of Environmental Protection. www.state.nj.us/dep/dsr/ research/mercury-bulbs.pdf. Accessed October 8, 2009. Ayers, R. U. 1989. Industrial metabolism. In Technology and Environment, edited by J. H. Ausubel and H. E. Sladovich. Washington, DC: National Academy Press. Ayers, R. U. and S. R. Rod. 1986. Patterns of pollution in the Hudson-Raritan basin. Environment 28(4): 14–43. Barr Engineering Co. (Minneapolis, MN). 2001. Substance flow analysis of mercury in products. Minnesota Pollution Control Agency, <www.pca.state.mn.us/publications/hgsubstance.pdf>. Accessed June 11, 2009. Belton, T., L. Lippincott, E. Stevenson, and J. Botts. 2008. Pollution minimization plans and PCB source trackdown in Camden City. August 2008. Trenton, NJ: New Jersey Department of Environmental Protection. Berglund, P. B. and R. W. Diercks. 2001. Evaluation of amalgam removal equipment and dental clinic loadings to the sanitary sewer. MCES and Minnesota Dental Association Report (MCES Report No. 01-509). Boehme, S. and M. Panero. 2003. Pollution prevention and management strategies for cadmium in the New York/New Jersey harbor. New York: New York Academy of Sciences. www.nyas.org/harbor. Brunner, P. H. 2002. Beyond materials flow analysis. Journal of Industrial Ecology 6(3/4): 8–10. Cain, A., S. Disch, C. Twaroski, J. Reindl, and C. R. Case. 2007. Substance flow analysis of mercury intentionally used in products in the United States. Journal of Industrial Ecology 6(3/4): 201– 219. de Cerreño, A. L. C., M. Panero and S. Boehme. 2002. Pollution prevention and management strategies for mercury in the New York/New Jersey Harbor. Boehme et al., Collaborative Problem Solving Using an Industrial Ecology Approach 827 A P P L I C AT I O N S A N D I M P L E M E N TAT I O N New York: New York Academy of Sciences. www.nyas.org/harbor. DelConte, M. J. 1997. A Mercury Pollution Prevention Study for Medical and Dental Centers: Findings Report and Bibliography. Monroe County Mercury Pollution Prevention Team (May 9, 1997). Fitzgerald, W. F. and J. O’Connor. 2001. Mercury cycling in the Hudson-Raritan River basin. Presented to the Harbor Consortium of the New York Academy of Sciences. New York: New York Academy of Sciences. www.nyas.org/harbor (select “List of publications and related materials”). Hansen, E. and C. Lassen. 2003. Experience with the use of substance flow analysis in Denmark. Journal of Industrial Ecology 6(3/4): 201–219. Lifset R. J. 2000. Full accounting. The Sciences. [New York], May/June 2000. Litten, S. 2003. Contaminant Assessment and Reduction Project. Bureau of Water Assessment and Management Division of Water, New York State Department of Environmental Conservation. August 2003. www.dec.ny.gov/ chemical/23839.html. Accessed October 2009. Lochan, R. 1999. Headworks Analysis Report. New York City Department of Environmental Protection. Muñoz, G. and M. Panero. 2006. Pollution prevention and management strategies for dioxins in the New York/New Jersey harbor. New York: New York Academy of Sciences. www.nyas.org/harbor. Muñoz, G. and M. Panero. 2008. Sources of suspended solids to the New York/New Jersey harbor watershed. New York: New York Academy of Sciences. www.nyas.org/harbor. NJDEP and NJDHSS (New Jersey Department of Environmental Protection and New Jersey Department of Health and Senior Services). 2004. A Guide to Health Advisories for Eating Fish and Crabs Caught in New Jersey Waters. Updated versions of this document are available at <http://www.state.nj.us/dep/dsr/njmainfish.htm>. Accessed October 2009. NJDEP (New Jersey Department of Environmental Protection). 2006. Routine Monitoring Program for Toxics in Fish: Year 2 – Estuarine and Marine Waters. Task II Data. (DRAFT). <www.state. nj.us/dep/dsr/2004data.htm>. Accessed March 23, 2006. NYAS (New York Academy of Sciences). 1998. Industrial ecology and the environment: Applications to the New York harbor. Report on a workshop of the New York Academy of Sciences and subsequent observations on action. February 1998. www.nyas.org/harbor (select “List of publications and related materials”). 828 Journal of Industrial Ecology NYAS. 2008. Safe harbor: Bringing people and science together to improve the New York/New Jersey harbor. Collaborative problem solving using an industrial ecology approach. New York Academy of Sciences. www.nyas.org/harbor. NYDOH New York State Department of Health. 2005. Chemicals in Sportfish and Game: 2005–06 Health Advisories. <www.health.state.ny.us/ nysdoh/fish/fish.pdf>. Accessed May 20, 2005. Panero, M. A. 2005. The economics of mercury pollution abatement in the New York – New Jersey harbor watershed. Ph.D. thesis, New School for Social Research, New York, NY. Panero, M., S. Boehme and G. Muñoz. 2005. Pollution prevention and management strategies for polychlorinated biphenyls in the New York/New Jersey harbor. New York Academy of Sciences. www.nyas.org/ harbor. Schechtman, D. 2007. An economic input output and process based LCA Comparison between unsealed concrete and sealed asphalt parking lots’ upstream environmental impacts. New York Academy of Sciences. (February 2007). www.nyas.org/harbor (select “List of publications and related materials”). Stigliani, W. M., P. R. Jaffe, and S. Anderberg. 1993. Heavy metal pollution in the Rhine River basin. Environtal Science and Technology 27(5): 786–793. Sznopek, J. L. and T. G. Goonan. 2000. The materials flow of mercury in the economies of the United States and the world. U.S. Geological Survey Circular 1197. U.S. Census Bureau. 2000. United States Census 2000. www.census.gov/main/www/cen2000.html. Accessed October 8, 2009. U.S. EPA (United States Environmental Protection Agency). 1992. Characterization of Products Containing Mercury in Municipal Solid Waste in the United States, 1970 to 2000. U.S. EPA, Solid Waste and Emergency Response (OS305) EPA530-R-92-013, or NTIS# PB92-162569. U.S. EPA. 1995. AP 42, Fifth Edition Compilation of Air Pollutant Emission Factors, Volume 1: Stationary Point and Area Sources. Office of Transportation and Air Quality. January 1995. <www. h2e-online.org/docs/h2emou101501.pdf>. Accessed June 10, 2009. U.S. EPA. 1997. Mercury Study Report to Congress, Vol. II. An Inventory of Anthropogenic Mercury Emissions in the United States. Office of Air Quality Planning and Standards and Office of Research and Development. EPA-452/R-97-004. Washington, DC: U.S. EPA. A P P L I C AT I O N S A N D I M P L E M E N TAT I O N U.S. EPA. 1998. Memorandum of Understanding Between the United States Environmental Protection Agency and the American Hospital Association. www.epa.gov/glnpo/toxteam/ahamou.htm. Accessed October 8, 2009. U.S. EPA. 2004. National Listing of Fish Advisories. Fact Sheet. EPA-823-F-04-016. Office of Water. August 2004. <www.epa.gov/ost/fish/ advisories/factsheet.pdf>. Accessed May 20, 2005. US EPA. 2007. Toxics Release Inventory (TRI). <www.epa.gov/tri/>. Accessed June 11, 2009. Valle S., M.A. Panero, and L. Shor. 2007. Pollution Prevention and Management Strategies for Polycyclic Aromatic Hydrocarbons in the New York/New Jersey Harbor. New York Academy of Sciences. www.nyas.org/harbor (select “List of publications and related materials”). WEF (Water Environment Federation). 1999. Controlling dental facility discharges in wastewater: How to develop and administer a source control program. Alexandria, VA: WEF. Wernick, I. K. and J. H. Ausubel. 1997. Industrial ecology: Some directions for research. With the Vishnu Group, The Rockefeller University, and Lawrence Livermore National Laboratory, May 1997. About the Authors Susan E. Boehme is coastal sediment specialist with the Illinois-Indiana Sea Grant College Program in Chicago IL, USA. Marta A. Panero is a Research Scientist and Program Coordinator at the New York University Wagner Rudin Center for Transportation Policy & Management in New York City, NY, USA. Gabriela R. Muñoz is a Program Associate with the New England Interstate Water Pollution Control Commission in New York, NY, USA. Charles W. Powers is a Professor of Environmental Engineering at Vanderbilt University and Co-Principal Investigator of the Consortium for Risk Evaluation and Stakeholder Participation, Nashville TN. Sandra N. Valle currently is an environmental scientist/analyst with Jonas and Associates at Science Applications International Corporation. Susan E. Boehme, Marta A. Panero, Gabriela R. Muñoz, and Sandra N. Valle all worked for the New York Academy during the ten-year Harbor Project. Charles W. Powers was the Harbor Consortium Chair for the length of the project. Boehme et al., Collaborative Problem Solving Using an Industrial Ecology Approach 829