Hazard Ranking System Issue Analysis: Relationship Between Waste Quantity and Hazardous Constituent Quantity MITRE ------- Hazard Ranking System Issue Analysis: Relationship Between Waste Quantity and Hazardous Constituent Quantity Arlene R. Wusterbarth September 1987 MTR-86W141 SPONSOR: U.S. Environmental Protection Agency CONTRACT NO.: EPA-68-01-7054 The MITRE Corporation Civil Systems Division 7525 Colshire Drive McLean, Virginia 22102-3481 ------- 3 Department Approval: -. (L,l / MITRE Project Approval:. ------- ABSTRACT The Hazard Ranking System (HRS) is used by the U.S. Environmental Protection Agency (EPA) to estimate the relative potential hazard to human health and the environment posed by releases or threatened releases of hazardous substances. For HRS purposes, the quantity of hazardous substances present at a site is currently evaluated on the basis of the total quantity of hazardous wastes deposited at the site, and not the total quantity of hazardous constituents present in those wastes. This report examines existing waste composition data to determine whether alternatives to the HRS waste quantity factor can be developed to allow the use of data on total hazardous constituent quantity in evaluating a site when such data are available. The primary concerns with the existing data relate to whether the data are representative of the wastes present at hazardous waste sites and to whether the data adequately characterize those wastes. A statistical analysis of available, but very limited, waste composition data indicates that there are potentially significant differences in the total concentration of hazardous constituents in wastes present in different waste management units at wastes sites. If the existing waste composition data are to be used in developing alternative HRS waste quantity factors, then these alternatives should reflect the different waste management units. Suggested Keywords: Superfund, Hazard ranking, Hazardous waste, Concentration, Waste composition. iii ------- ACKNOWLEDGMENT The author wishes to acknowledge several individuals for their valuable contributions to this research effort. Steve Caldwell of the U.S. Environmental Protection Agency (EPA) provided guidance and oversight. Kris Barrett of the MITRE Corporation supervised and directed the project. Access to the automated Contract Laboratory Program data base was made available by Donald Trees and William Eckel of Viar and Company, Inc. The author also thanks Robert Laidlaw and Joe Lockerd at the National Enforcement Investigations Center (NEIC) for providing the NEIC data. In addition, the author has benefited greatly from the suggestions and assistance of several colleagues. Ming Wang and Thomas Wolfinger gave insight to the statistical analysis. Excellent computer programming was provided by Brian Doty and Kerry Zimmerman. Above all, the author is particularly indebted to Stuart Haus of the MITRE Corporation whose technical contributions and extensive reviews greatly enhanced this report. IV ------- TABLE OF CONTENTS LIST OF ILLUSTRATIONS vii LIST OF TABLES viii 1.0 INTRODUCTION 1 1.1 Background 1 1.2 Issue Description 3 1.3 Scope 5 1.4 Organization of Report 6 2.0 SOURCES AND LIMITATIONS OF WASTE COMPOSITION DATA 7 2.1 Contract Laboratory Program 10 2.2 National Enforcement Investigations Center 12 2.3 Previous Industry Studies of the EPA Office of Solid 13 Waste 2.4 Current Industry Studies Program of the EPA Office of 14 Solid Waste 2.5 Superfund Site Inspection (SI) Reports 16 2.6 Superfund Remedial Investigation (RI) Reports 17 2.7 Other Data Sources 18 2.8 Limitations of Existing Data With Regard to the Current 19 Analysis 3.0 SUMMARIES OF DATA FROM INDIVIDUAL SOURCES 31 3.1 Approach 31 3.2 Summary of Data From Samples of Wastes at Disposal Sites 33 3.2.1 Drum Samples 33 3.2.2 Tank Samples 54 3.2.3 Surface Impoundment Samples 56 3.2.4 Mining Waste Samples 56 3.2.5 Other Samples of Wastes at Disposal Sites 59 3.3 Summary of Data from Samples of Industrial Waste 59 Streams at the Point of Generation 3.3.1 Previous Industry Studies of the EPA Office 60 of Solid Waste (OSW) 3.3.2 Current OSW Industry Studies Program 64 3.3.3 Waste Oil 71 ------- TABLE OF CONTENTS (Concluded) Page 4.0 ANALYSIS OF DATA 73 4.1 Composite Overview of the Total Concentration of 73 Hazardous Constituents Present in Hazardous Wastes by Type of Waste Management Method 4.1.1 Drums 73 4.1.2 Tanks 76 4.1.3 Surface Impoundments 79 4.1.4 Mine Tailings and Waste Piles 82 4.1.5 Landfarms and Landfills 85 4.1.6 Summary of Findings 87 4.2 Comparisons of the Total Concentration of Hazardous 89 Constituents in Liquids, Solids, and Sludges 4.2.1 Liquids 89 4.2.2 Solids 93 4.2.3 Sludges 94 4.2.4 Summary of Findings 95 4.3 Statistical Pairwise Comparisons of Median Total 96 Concentrations 5.0 SUMMARY AND CONCLUSIONS 103 APPENDIX A - Previous Industry Studies of the EPA Office 109 of Solid Waste APPENDIX B - Current EPA Office of Solid Waste Industry 139 Studies Data Base APPENDIX C - Franklin Associates Data Base for Waste Oil 165 Composition APPENDIX D - Definitions of Waste Management Units 171 APPENDIX E - Availability of Waste Composition Data By Site 173 in the Superfund Site Inspection and Remedial Investigation Reports APPENDIX F - Bibliography 177 vi ------- LIST OF ILLUSTRATIONS Figure Number Pag< 1 Frequency Distribution of Hazardous Constituent 39 Concentrations in 580 Drum and High Hazard Samples from the NEIC Data Base (Open Interval) 2 Frequency Distribution of Hazardous Constituent 40 Concentrations in 580 Drum and High Hazard Samples from the NEIC Data Base (Closed Interval) 3 Frequency Distribution of Hazardous Constituent 46 Concentrations in Drum and High Hazard Samples from the NEIC Data Base for Identified NPL and Non-NPL Site Samples (Open Interval) 4 Frequency Distribution of Hazardous Constituent 47 Concentrations in Drum and High Hazard Samples from the NEIC Data Base for Identified NPL and Non-NPL Site Samples (Closed Interval) 5 Frequency Distribution of Hazardous Constituent 49 Concentrations in Drum and High Hazard Samples from the NEIC Data Base for Identified NPL Site Samples 6 Frequency Distribution of Hazardous Constituent 50 Concentrations in Drum and High Hazard Samples from the NEIC Data Base for Identified Non-NPL Site Samples vii ------- LIST OF TABLES Table Number Page 1 Key Features of Identified Sources of Waste 8 Composition Data Used in the Current Study 2 Limitations of Existing Waste Composition Data 20 With Regard to the Current Analysis 3 Descriptive Statistics for the Total 35 Concentration of Hazardous Constituents Present in CLP Drum Samples 4 Distribution of Drum and High Hazard Samples in 43 the NEIC Data Base 5 Summary Statistics for the Total Concentration 44 of Hazardous Constituents Present in NEIC Drum and High Hazard Samples for Identified NPL and Non-NPL Sites 6 Summary of Waste Composition Data Compiled from 52 Superfund Site Inspection Reports 7 Summary of Waste Composition Data Compiled 53 from Superfund Remedial Investigation Reports 8 Preliminary Data on the Concentration of 58 Selected Hazardous Constituents in Indicated Mining Wastes 9 Average Total Concentrations of Hazardous 61 Constituents in Industrial Waste Streams from Previous EPA Industry Studies 10 Total Concentration of Hazardous Constituents 66 Present in ISDB Waste Streams (Unweighted) 11 Total Concentration of Hazardous Constituents 68 Present in ISDB Waste Streams (Quantity-Weighted) 12 Total Concentration of Hazardous Constituents 72 Present in Selected Waste Streams of the Waste Oil Re-Refining Industry viii ------- LIST OF TABLES (Concluded) Table Number 13 Composite Overview of the Concentration of Hazardous Constituents in Drummed Wastes 14 Composite Overview of the Concentration of 77 Hazardous Constituents in Tank Wastes 15 Composite Overview of the Concentration of 80 Hazardous Constituents in Surface Impoundment Wastes 16 Composite Overview of the Concentration of 83 Hazardous Constituents in Mine Tailings and Waste Piles 17 Composite Overview of the Concentration of 86 Hazardous Constituents in Landfarm and Landfill Wastes 18 Summary of Waste Composition Data for Liquid, 91 Solid, and Sludge Wastes 19 Pairwise Comparisons of the Median Total 99 Concentrations of Hazardous Constituents Present in Hazardous Wastes at Disposal Sites (Site Inspection Reports) 20 Pairwise Comparisons of the Median Total 100 Concentrations of Hazardous Constituents Present in Hazardous Wastes at Disposal Sites (Remedial Investigation Reports) IX ------- 1.0 INTRODUCTION 1.1 Background The Comprehensive Environmental Response, Compensation, and Liability Act of 1980 (CKRCIA) (PL 96-510) requires the President to identify national priorities for remedial action among releases or threatened releases of hazardous substances. These releases are to be identified based on criteria promulgated in the National Contingency Plan (NCP). On July 16, 1982, EPA promulgated the Hazard Ranking System (HRS) as Appendix A to the NCP (40 CFR 300; 47 FR 31180). The HRS comprises the criteria required under CERCLA and is used by EPA to estimate the relative potential hazard posed by releases or threatened releases of hazardous substances. The HRS is a means for applying uniform technical judgment regarding the potential hazards presented by a release relative to other releases. The HRS is used in identifying releases as national priorities for further investigation and possible remedial action by assigning numerical values (according to prescribed guidelines) to factors that characterize the potential of any given release to cause harm. The values are manipulated mathematically to yield a single score that is designed to indicate the potential hazard posed by each release relative to other releases. This score is one of the criteria used by EPA in determining whether the release should be placed on the National Priorities List (NPL). ------- During the original NCP rulemaking process and the subsequent application of the HRS to specific releases, a number of technical issues have been raised regarding the HRS. These issues concern the desire for modifications to the HRS to further improve its capability to estimate the relative potential hazard of releases. The issues include: • Review of other existing ranking systems suitable for ranking hazardous waste sites for the NPL. • Feasibility of considering ground water flow direction and distance, as well as defining "aquifer of concern," in determining potentially affected targets. • Development of a human food chain exposure evaluation methodology. • Development of a potential for air release factor category in the HRS air pathway. • Review of the adequacy of the target distance specified in the air pathway. • Feasibility of considering the accumulation of hazardous substances in indoor environments. • Feasibility of developing factors to account for environmental attenuation of hazardous substances in ground and surface water. • Feasibility of developing a more discriminating toxicity factor. • Refinement of the definition of "significance" as it relates to observed releases. • Suitability of the current HRS default value for an unknown waste quantity. • Feasibility of determining and using hazardous substance concentration data. ------- • Feasibility of evaluating waste quantity on a hazardous constituent basis. • Review of the adequacy of the target distance specified in the surface water pathway. • Development of a sensitive environment evaluation methodology. • Feasibility of revising the containment factors to increase discrimination among facilities. • Review of the potential for future changes in laboratory detection limits to affect the types of sites considered for the NPL. Each technical issue is the subject of one or more separate but related reports. These reports, although providing background, analysis, conclusions and recommendations regarding the technical issue, will not directly affect the HRS. Rather, these reports will be used by an EPA working group that will assess and integrate the results and prepare recommendations to EPA management regarding future changes to the HRS. Any changes will then be proposed in Federal notice and comment rulemaking as formal changes to the NCP. The following section describes the specific issue that is the subject of this report. 1.2 Issue Description For HRS purposes, the quantity of hazardous substances present at a site is currently evaluated on the basis of the total quantity of hazardous wastes deposited at the site and not on the total quantity of hazardous constituents present in those wastes. The rationale for this approach is discussed in depth in the Federal ------- Register preamble to the NCP (47 FR 31190, 16 July 1982). In part, this approach has been taken because EPA "has been unable to develop an internally consistent approach for comparing pure hazardous substance quantity at facilities where definitive information is available with hazardous substance quantity at facilities where such information is not available". During both the 1982 NCP rulemaking and subsequent NPL rulemakings, various commenters have indicated that site-specific data on total hazardous constituent quantity should be used in evaluating a site with the HRS when such data are available. The purpose of this paper is to examine existing waste composition data to determine whether, in addition to the current HRS waste quantity factor, alternative HRS waste quantity factors can be developed to allow the use of data on total hazardous constituent quantity in evaluating a site when such data are available. That is, the waste composition data identified under this effort are analyzed to determine whether it is reasonable to express the relationship between waste quantity and the total hazardous constituent quantity in that waste as a constant (e.g., for various types of waste management units, is there a statistically significant value or range for the mean total concentration of hazardous constituents in those hazardous wastes present in each type of waste management unit). If such a relationship exists, then an approach can be developed for consistently and uniformly comparing waste quantities ------- at sites that have data on hazardous constituent quantity available with those sites that do not have such data available. Note that this paper is not concerned with the relative hazard of the individual constituents in hazardous wastes, nor with the relative hazard of entire waste streams. Rather, the issue of concern is strictly the relationship between waste quantity and total hazardous constituent quantity. 1.3 Scope This report describes the major sources of waste composition data that have been identified by this study and provides a compilation and analysis of the waste composition data presently available from these sources. As various data sources were reviewed, it became apparent that concentration data for wastes deposited at disposal sites are very limited. Most available concentration data from disposal sites pertain to environmental samples (e.g., ground water, soil, leachate, surface water), not to waste samples. These environmental samples cannot be used to reliably estimate the composition of the wastes that were deposited at the disposal site. Furthermore, the waste composition data that are available typically pertain to wastes managed in drums, tanks, and surface impoundments. For the purpose of gaining additional insight into the composition of wastes in these and other waste management units (e.g., piles, landfills, landfarms), MITRE also conducted a review and analysis of existing data on the hazardous constituent concentrations present in ------- individual waste streams at their point of generation. While these waste streams are likely to be commingled with other waste streams prior to (or at the time of) disposal, information about their composition could possibly serve as proxies for the composition of some wastes present at disposal sites, subject to the limitations discussed in Section 2. 1.4 Organization of Report Section 2 of this document describes the identified sources of waste composition data. Specifically, the major sources providing waste composition data include the EPA Contract Laboratory Program (CLP), EPA's National Enforcement Investigations Center (NEIC), the EPA Office of Solid Waste (OSW), and the EPA Office of Emergency and Remedial Response (OERR). Section 2 also discusses the limitations associated with the use of these data in the current analysis. Section 3 provides statistical summaries of the waste composition data from each individual data source. Section 4 contains a comparative analysis of all the data. A summary of the analytical findings and conclusions are presented in Section 5. Detailed concentration data for constituents present in various waste streams are contained in Appendices A, B, and C. Definitions of the waste management units considered in this study are included in Appendix D. Appendix E identifies those sites for which waste composition data are available from Superfund site inspection or remedial investigation reports. Appendix F contains the bibliography. ------- 2.0 SOURCES AND LIMITATIONS OF WASTE COMPOSITION DATA This section provides a brief description of the data sources identified as containing data on the concentration of hazardous constituents present in wastes. The various data sources provide information either about the composition of wastes present at disposal sites or the composition of individual waste streams at the point of generation, prior to disposal. Wastes present at disposal sites are generally mixtures of several individual waste streams. Major data sources identified include the EPA Contract Laboratory Program, the National Enforcement Investigations Center, waste characterization studies of the EPA Office of Solid Waste, and the site inspection and remedial investigation reports prepared for the EPA Office of Emergency and Remedial Response. Table 1 briefly outlines the general characteristics of each data source. The type of waste composition information available from each source is discussed directly below. An analysis of the data is presented in Sections 3 and 4. It should be noted that the analytical waste composition data available from these sources have been collected by these sources with objectives different from those of the current study. In many instances, these differences severely restrict the use of the data in the current study, as noted below. ------- TABLE 1 KEY FEATURES OF IDENTIFIED SOURCES OF WASTE COMPOSITION DATA USED IN THE CURRENT STUDY O3 1. Data Source Automated Contract No. of Waste Samples 261 No. of Sites 81 Classification of Sites* 16 NPL; 65 Non-NPL Types of Waste Management Units Analyzed Drums Number of Constituents Analyzed 190 organic constituents and 2. 3. Laboratory Program Data Base 4. EPA National Enforcement Investigations Center 1970s OSW Industry Studies Reports OSW Industry Base Studies Data 327 253 36 1-156*** 14-62*** 173 NA 14 NPL; 22 Non-NPL NA** NA Industrial waste streams from 10 industries Waste streams from 11 segments of the organic chemicals industry Primarily drums; some other high hazard units (e.g., tanks, wasteponds, spills) Industrial waste streams (e.g., sludges, liquids, solids) Industrial waste streams managed in containers, surface impoundments, tanks, piles, landfarms, and landfills**** 28 inorganic constituents (plus 167 other organic constituents also tentatively identified as present) 133 organic constituents, 34 inorganic elements, and cyanide Generally limited to a few metals Several hundred organic and inorganic constituents ------- TABLE 1 (Concluded) Data Source No. of Waste Samples No. of Sites Classification of Sites* Types of Waste Management Units Analyzed Number of Constituents Analyzed 5. Superfund Site 155 Inspection Reports 6. Superfund 109 Remedial Investigation Reports 7. PEDCo Mining 183 Study 44 16 36 NPL; 8 Non-NPL 16 NPL 65 65 Non-NPL Drums, tanks, surface impoundments, piles, mine tailings, landfarms, landfills Drums, tanks, surface impoundments, mine tailings Surface impoundments, mine tailings Variety of organic and inorganic constituents Variety of organic and inorganic constituents 22 inorganic constituents 8. Franklin 1,071 NA NA Waste Oil Study Primarily tanks; some drums 6 metals, 12 organic constituents, and totalchlorine *NPL sites are sites listed on, or proposed for the National Priorities List. Non-NPL sites are sites not listed on, nor proposed for, the National Priorities List. **NA: Not available. ***Range of the number of samples or sites analyzed across the various studies; number of samples and sites was not reported for all industries. See Appendix A for data on the number of samples and sites reported to be analyzed for each of the 10 industries studied. ****Analytical data are from waste streams at point of generation, not from wastes present in the waste management units. ------- 2.1 Contract Laboratory Program The EPA Contract Laboratory Program (CLP) is intended to provide EPA with analyses of field samples collected from hazardous waste disposal sites. The CLP data base consists of hardcopy analytical results from samples analyzed under the CLP since 1980. This data base consists of two major groups of data: Routine Analytical Services (RAS) and Special Analytical Services (SAS). RAS contains data from routine sample analyses (e.g., priority pollutants, metals), while SAS contains data on special compounds and special types of samples (e.g., dioxin, fish samples) for which analyses are tailored to meet particular needs. Of the two groups of data, RAS is the more comprehensive; it contains results from approximately 35,000 samples collected at more than 1,200 uncontrolled hazardous waste sites. Of these 1,200 sites, 19 percent are NPL sites, and 81 percent are non-NPL sites. In 1984 a portion of the CLP/RAS data base (about a 10 percent random sample, stratified by type of site, i.e., NPL and non-NPL) was automated from hardcopy (Friedman et al., 1984).* This automated CLP data base contains information for 358 sites** and over 3,000 samples from these sites. Data are available for more than 218 *The automated CLP data base has been constructed, and is currently maintained, for EPA by Viar and Company, Inc., Alexandria, VA. **As of January 1984, 10 percent of the 358 sites were NPL sites, and the remainder were non-NPL sites. Available data on the identities of the sites in the data base are not adequate to determine for this report whether this distribution has since changed. 10 ------- constituents present in these samples. These constituents primarily consist of 133 organic priority pollutants and other hazardous organics, 24 metals, 4 inorganic ions, and 57 other organics tentatively identified as present in both the samples and at over 1 percent of the sites.* Although the CLP automated data base contains information by site, sample, and compound, almost all analytical results are derived from environmental samples (e.g., ground water, surface water, soils, sediment, and leachate). Very limited data are available on the composition of the wastes as deposited at the site. In fact, drum samples are the only waste form that can be explicitly identified in the automated CLP data base at this time. Data on drummed wastes are available from 81 of the 358 sites in the automated CLP data base. However, most of these drum data cannot be used in the current analysis. For the most part, the analytical results available in the automated CLP data base provide only the organic or inorganic portion of the sample analysis, not both portions. (See Friedman et al., 1984, for a discussion of the development of the automated CLP data base.) Of the 261 drum samples currently in the data base, only 5 include data for both the *0ver 1,000 additional "other organics" have also been tentatively identified as being present in the samples. These tentatively identified organics occur at less than 1 percent of the sites. Concentration data are available for 167 of these tentatively identified organics in drum samples in the automated data base. These concentrations are also included in the data summaries presented in Section 3. 11 ------- organic and inorganic portions of the sample analysis. Analysis of the hardcopy analytical results for the drum samples in the CLP data base is beyond the scope of this effort.* 2.2 National Enforcement Investigations Center The Regulated Substances Laboratory at EPA's National Enforcement Investigations Center (NEIC) operates under contract to the CLP. The NEIC has collected and analyzed samples from drums, tanks, and other waste management units from 221 disposal sites in 41 States and one U.S. territory. This data base consists of approximately 1,600 samples from these sites.** The samples have been analyzed for 113 organic priority pollutants, 20 additional organic substances, and 34 inorganic elements (including 13 priority pollutant elements), and cyanide. Not all samples were analyzed for each constituent. Almost all of the constituents for which analyses were conducted are also included in the CLP analysis discussed in Section 2.1. For the most part, the constituents in the NEIC samples were analyzed using the RAS analytical procedures of the CLP. A few of the samples were subjected to SAS analytical procedures. While the NEIC data consist primarily of drum samples, information is not available to determine which samples are from drums and which samples are from other management units (e.g., *Furthermore, as discussed in Sections 3 and 4, analysis of these data would not be likely to change the conclusions of this study. **A small subset (about 4 percent) of the drum data collected and analyzed by the NEIC is also contained in the automated CLP data base. 12 ------- tanks, waste ponds) or from spills. In addition, the organic and inorganic portions of the NEIC sample analyses are segregated in the data base. The two portions could be matched for 580 of the samples. Consequently, only these 580 samples are included in the subsequent analyses. 2.3 Previous Industry Studies of the EPA Office of Solid Waste In the mid-1970s EPA sponsored several studies to assess the hazardous waste practices of selected industries. Each study provided an economic overview of the industry; a characterization of the wastes generated by the industry; a description of the waste treatment and disposal practices of the industry; and an analysis of costs associated with alternative disposal technologies. A review of these studies was undertaken for the purpose of compiling data on the concentrations of hazardous constituents in individual waste streams which may be present at disposal sites.* Of the 14 studies reviewed, 10 studies contained data on the concentration of constituents present in waste streams at the point of generation. The four remaining studies either did not include sampling programs or did not include the analytical results in the industry study report. The analyses conducted for the industry studies focused almost exclusively on metals present in the waste streams. For three industries, some information was also provided *Appendix A identifies the industries included in the studies and contains the detailed concentration data reported for various hazardous constituents found in the waste streams. 13 ------- on the concentration of cyanides and phenols in waste streams. One industry study also included benzo[a]pyrene in its analysis. In most of the industry studies, samples were not analyzed for PCBs* or other organics. In general, a review of the industry studies revealed that the various studies are not consistent with respect to the number of metals analyzed. Moreover, the type of sampling (e.g., grab, composite) varied within and across the studies. Also, there was a vast difference in the number of samples analyzed for the industrial' t waste streams under study. In one study, for example, only a single sample from one waste stream was analyzed, while in another study 56 samples for the waste stream under consideration were collected and analyzed from a large number of plants. Some studies provide no information about the number of samples analyzed. 2.4 Current Industry Studies Program of the EPA Office of Solid Waste The EPA Office of Solid Waste (OSW) has recently been collecting data on the composition of hazardous waste streams at the point of generation as part of an EPA Industry Studies program which began in 1980. In contrast to the earlier industry studies, the waste samples in the present Industry Studies program are being systematically analyzed for a wide variety of organic and inorganic constituents. To date, the current Industry Studies program has focused exclusively on *Secondary references cited in some of the industry study reports provided some, but very limited, data for PCBs. 14 ------- waste streams from the organic chemicals industry. The data collected have been used to develop the Industry Studies Data Base (ISDB). This data base currently contains information about the facilities, processes, products, waste residuals, and management methods for 11 organic chemicals industry segments. Wastes are characterized by residual types (e.g., sludges), quantities, hazardous constituents and their concentrations, as well as management methods. About 80 percent of the information in the ISDB was derived from mandatory questionnaires. Of approximately 650 questionnaires mailed to organic chemical manufacturers, 633 were completed and returned. Additional sources of information for the data base include: 1) plant visit reports; 2) sampling and analysis site visit reports; and 3) engineering analyses. These additional reports and analyses had formats similar to the questionnaire. A problem in using this data base in the current study relates to the manner in which the waste composition data were reported. Respondents were asked to identify the constituents present in residuals and to specify the concentration of the constituent according to predetermined ranges (e.g., the constituent concentration is 1 to 10 percent, 10 to 50 percent, etc.). Only in cases where the concentration was less than 100 ppm (0.01 weight percent) were the respondents specifically requested to report actual concentrations.* *See RCRA Section 3007 Questionnaire (OMB No. 2000-0396), Question 5, Residuals Characterization Information, Part D. 15 ------- Consequently, a majority of the waste composition data available in the ISDB indicate only the predetermined concentration range in which the constituent concentration falls, not the actual constituent concentration. All ISDB statistical summaries presented in this report are based only upon those waste streams for which actual concentration estimates are available (see Appendix B). For the various waste management methods reviewed in this report, 10 to 60 percent of the relevant samples in the ISDB have actual concentration data, not concentration ranges. 2.5 Superfund Site Inspection (SI) Reports Superfund site inspection (SI) reports were examined for data on the composition of wastes at uncontrolled waste sites listed in the NPL technical data base. Data on the composition of wastes deposited at these sites—including drummed wastes, surface impoundment wastes, and mine tailings, among others—are available for 44 sites.* As of August 1986, 36 of the 44 sites are listed or proposed NPL sites; 2 are non-NPL sites; and the remaining 6 sites are still under review for possible inclusion on the NPL. In many cases the waste samples were analyzed for a wide variety of organics and inorganics. In a few cases, however, the analysis was limited to a single constituent of concern (e.g., PCBs). *Most of the data in the SI reports pertain to environmental samples. Environmental sampling is defined in this report as sampling from ground water (e.g., wells), surface water (e.g., lakes), soil, sediment, and leachate. 16 ------- The waste composition data in the site inspection reports pertain to 8 waste management methods. However, the vast majority of the information relates to only the following 4 waste management methods: surface impoundments* (24 sites); drums (11 sites); tanks (4 sites); and mining operations (3 sites). For 5 sites, waste composition data are available from more than one type of management method (e.g., tanks and lagoons). 2.6 Superfund Remedial Investigation (RI) Reports Fifty remedial investigation (RI) reports prepared for the EPA OERR were also reviewed for data on the composition of wastes deposited at NPL sites. Most of the sampling data in the RIs pertain to environmental samples, not waste samples. Sixteen of the RIs, however, provide waste composition data in addition to environmental sampling data. Primarily, the waste composition data pertain to wastes from surface impoundments (7 sites); drums (5 sites); and tanks (4 sites). At 4 of the 16 sites, waste composition data are available from more than one type of management method. In addition, one of the RIs also contains an analysis of waste water treatment plant sludge that was disposed at the site. Most of the waste samples discussed in the RI reports were analyzed for a wide variety of both organics and inorganics. While "Surface impoundments include lagoons, waste ponds, waste pits, and waste trenches. 17 ------- the waste sampling and analyses were usually performed explicitly as part of the RIs, some of the hazardous constituent concentration data published in the RIs originate from secondary sources. These secondary sources are typically previous studies by various State departments of environmental protection. The number of constituents analyzed in the various State studies varied widely. 2.7 Other Data Sources Three other sources of data on the composition of wastes were identified and reviewed. These sources are: 1) a study of mining wastes prepared by PEDCo Environmental, Inc. for OSW; 2) a data base on waste oil composition developed by Franklin Associates, Ltd. for OSW; and 3) EPA Background Documents for the listing of hazardous wastes under the Resource Conservation and Recovery Act (RCRA). The PEDCo study (1983) characterizes wastes generated by the mining industry. More than 400 waste samples were collected from 65 mining sites across the country- None of these sites are on the NPL. The samples—taken from solid, slurry, and liquid wastes—were analyzed primarily for metals and other inorganics. The data base developed by Franklin Associates (1984) characterizes the composition and concentration of contaminants in waste oil samples collected and analyzed between 1981 and 1984. Analytical results were obtained for more than 1,000 waste oil samples, both automotive and industrial. The samples were taken from establishments (e.g., gasoline stations) storing such waste oil 18 ------- in aboveground tanks, belowground tanks, and drums. The samples were analyzed for 6 metals, 5 chlorinated solvents, 7 other organics, and total chlorine (see Appendix C). The EPA Background Documents for the listing of hazardous wastes under RCRA are based primarily on the industry study reports from the 1970s and were found to provide little additional information to the data already contained in the industry study reports. In one instance, however, a Background Document did supplement the concentration data for waste streams of the paint manufacturing industry. 2.8 Limitations of Existing Data With Regard to the Current Analysis The data sources described in Sections 2.1 through 2.7 provide the most complete and comprehensive data bases available on the composition of wastes present at hazardous waste sites. However, these data bases were developed with different objectives from those of the current study. As a result, the data available from these data bases are not entirely consistent with the data requirements of this study. Consequently, there are limitations in applying the data to the current study and in drawing conclusions based on the data. Table 2 summarizes the major limitations, for purposes of this study, associated with the waste composition data available from each of the data sources. The primary concerns with the existing data relate to whether the data are representative of the wastes 19 ------- TABLE 2 LIMITATIONS OF EXISTING WASTE COMPOSITION DATA WITH REGARD TO THE CURRENT ANALYSIS Data Source Date Type of Wastes Major Limitations 1. Automated Contract 1980-83 Laboratory Program Data Base Drummed Wastes 2. EPA National Enforcement Investigations Center 1980-84 Drummed and High Hazard Wastes Available data pertain only to drummed wastes Incomplete automation of sample results; complete data available for only 5 samples CLP drummed waste samples may not be representative of the wastes at the particular sites from which the samples were taken In general, samples may not be representative of the wastes at uncontrolled hazardous wastes sites Waste composition data for drums combined with other high hazard samples Data limited to samples suspected of having high concentrations of hazardous constituents; therefore, data may not be representative of all drummed wastes Randomness of data indeterminate ------- TABLE 2 (Continued) Data Source Date Type of Wastes Major Limitations 3. 1970s OSW Industry Studies Reports 4. OSW Industry Studies Data Base 1975-78 Industrial • Waste stream data are from the Waste point of generation, not point of Streams disposal from 10 • Analysis of samples limited to a Industries very small number of inorganics; the particular constituents being analyzed varied by industry • Sampling programs varied greatly with respect to the number and types of samples collected and analyzed • Samples not generally random • Samples not likely to be representative of the industry; many samples are single grab samples from a very limited number of facilities 1980- Waste • Data limited to a subset of the Present Streams organic chemicals industry from 11 • Waste stream data are from the Segments point of generation, not point of of the disposal Organic • Large portion of data base only Chemicals indicates a pre-selected range in Industry which the concentration data fall, rather than the actual concentrations ------- TABLE 2 (Continued) Data Source Date Type of Wastes Ma.ior Limitations K) (-0 4. OSW Industry Studies Data Base (Concluded) 5. Superfund Site Inspection Reports 6. Superfund Remedial Investigation Reports • Samples not randomly collected; however, they may be representative of these industries due to the large number of facilities from which data were collected 1980- Variety • Majority of sites are listed or Present of Wastes proposed NPL sites and Waste • Constituents for which samples were Management analyzed varied from one to a few Methods to a large number • Small number of samples available for several waste management methods • Generally no information is available about sampling techniques used; randomness of samples indeterminate 1983-85 Variety • All sites are listed NPL sites of Wastes • Constituents for which samples were and Waste analyzed varied from few to many Management • Small number of samples available Methods for several waste management methods • Generally no information is available about sampling techniques used; randomness of samples indeterminate ------- TABLE 2 (Concluded) Data Source Date Type of Wastes Ma.ior Limitations 7. PEDCo Mining Study 1983 8. Franklin Waste Oil Study 1981-84 Mine Wastes Waste Oil • Data limited to non-NPL sites; may not be representative of all mining sites • Analysis of samples limited primarily to inorganics • Randomness of samples indeterminate • Data limited to waste oil • Limited number (19) of hazardous constituents analyzed for/reported • Randomness of samples indeterminate ------- present at hazardous wastes sites and to whether the available data adequately characterize the hazardous constituents in those wastes present. If the data are not totally representative or do not adequately characterize the wastes, then findings based on the data may be invalid. There are five major factors that affect, to varying degrees, the representativeness of the available data. These are: • Whether the data are based on random samples. • Whether the data pertain to all, or just a limited subset of, wastes present at hazardous wastes sites. • Whether the data pertain to all, or just a limited subset of, the waste management methods used at hazardous wastes sites. • Whether the data pertain to all types of hazardous wastes sites, or just to NPL sites or non-NPL sites. • Whether the data are based on a sufficient number of samples. There is a question as to whether most of the available data are based on random samples. Random samples are important in that all statistical tests require random data for drawing valid inferences. The automated CLP data base was developed by randomly selecting existing samples for inclusion in the data base. However, it is not known to what extent the waste samples themselves were randomly obtained at the hazardous wastes sites. The two OSW Industry Studies data bases contain data that are not based on random samples. The 1970's industry studies primarily contain data from grab samples, while the ISDB consists of data obtained from questionnaires that 24 ------- were not randomly administered. It is not possible to determine whether the samples from the other data sources are random. The data sources vary, to different degrees, in the types of wastes and waste management methods for which data are available. The automated CLP data base provides tdata for only one type of waste management method (i.e., wastes present in drums). These data are, however, based on a wide variety of wastes and a large number of samples. The NEIC data pertain primarily to drummed wastes; however, data from several other management methods are inextricably mixed with the drummed waste data. The NEIC data are also based on a large number of samples. Two other data sources (SI and RI reports) contain data on a wide variety of wastes and waste management methods; however, data on only a small number of waste samples are available for several of the waste management methods in these reports. The PEDCo data base addresses only mining wastes, but provides data for a wide variety of mining wastes and waste management methods used for these wastes. The three other data sources (i.e., the two OSW Industry Studies and the Franklin Waste Oil Study) provide data only about the composition of waste streams at the point of generation, not the point of disposal. Such waste streams are generally commingled with other hazardous and/or nonhazardous waste streams prior to, or as part of, the disposal process. In addition to possibly not being representative of the waste composition at the point of disposal, the 25 ------- data from these latter three sources pertain only to a limited subset of the waste streams present at hazardous waste sites. Furthermore, due to changes in processes and products over time, the data from the two OSW industry studies may not be representative of the composition of waste streams deposited at hazardous waste sites in the past. The data from the OSW ISDB and the Franklin Waste Oil Study are based on a large number of samples while the data from the 1970s industry studies are generally based on an extremely limited number of samples. The data sources also differ as to the types of wastes sites to which their data apply. The data in the SI reports are based on both NPL and non-NPL sites; however, most of these sites are NPL sites.* The data in the RI reports are limited to sites that have been listed on or proposed for the NPL. The PEDCo mining waste data are limited to mining sites that are non-NPL sites.** The data from the other sources generally apply to both NPL and non-NPL sites. *Analysis of the SI data on drummed wastes, the only waste for which sufficient NPL and non-NPL samples are available for statistical analysis, indicates that there is no statistically significant difference between the median total concentration of hazardous constituents in drummed wastes present at NPL and non-NPL sites. This finding is supported by similar testing of the NEIC data (see Section 3.2.1.2). Consequently, there are indications that for at least some types of wastes, the waste composition data from NPL sites may be representative of all sites. **Review of the SI and RI waste composition data for NPL mining sites and the PEDCo waste composition data for non-NPL mining sites indicates considerable differences in mean total concentrations. In fact, the mean total concentration of hazardous constituents estimated from the SI and RI data exceeds the maximum total concentration reported at any site in the PEDCo study (see Section 3.2.4). If these three data sets are truly representative data sets, then this finding is highly unlikely. 26 ------- There is one further concern about the representativeness of the data that applies primarily to the CLP and NEIC data and to a lesser extent to the SI data. Site inspections and the analysis of data by the CLP program and the NEIC are generally integral steps in evaluating whether a site belongs on the NPL. Thus, sites in the CLP, NEIC, or SI data bases that are identified as being non-NPL sites may only be non-NPL sites because they have not yet completed the NPL evaluation process. Upon completion of the process, they may be listed on the NPL. This may cause some anomalies in results when comparisons are made between NPL and non-NPL site data from these data bases. For example, if in this analysis a site with very high waste constituent concentrations is currently a non-NPL site, but at a later date is classified as an NPL site, results presented in the analysis for non-NPL sites will be biased upward, and results for NPL sites will be biased downward. Due to the limited number of samples available, such biases could significantly affect results. It is not possible in this study to determine which non-NPL sites would ultimately be listed on the NPL. In addition to the above factors, there are two other factors that affect how adequately the available waste composition data characterize the hazardous constituents present in the waste samples analyzed. These factors are: • The number of CERCLA Hazardous constituents for which analyses were performed. • The completeness of the available analytical data. 27 ------- The number of hazardous constituents for which waste samples were analyzed varies both from data source to data source, as well as within certain data sources. The number of constituents analyzed in a waste sample ranges from 2 or 3 inorganic constituents in several of the 1970s industry studies to several hundred organic and inorganic constituents in the CLP samples. For the other data sources, the number of constituents analyzed fall within this range. Furthermore, complete analytical results are not available in many cases. The automated CLP data base contains complete sample analyses for only 5 out of 261 drummed waste samples. For the remainder of the samples, only the organic or inorganic portion of the analysis is available in the data base. For the NEIC data, organic and inorganic portions of sample analyses could be matched for 580 out of 1,600 samples. For the ISDB, a majority of the waste composition data indicates only a pre-selected concentration range in which a constituent concentration falls, rather than the actual constituent concentration. The net effect of all these limitations varies from data source to data source and is generally indeterminate. Limitations in the number of constituents analyzed or in the completeness of the analytical data result in hazardous constituent concentrations being underestimated. Limitations in any of the other factors could result either in upward or downward biasing of the results to an unknown degree. 28 ------- Since the identified data are the most complete and comprehensive data available, they are used in the analysis that follows. However, the limitations noted above must be recalled in reviewing the findings of the analysis. Findings that are consistent across several data sources should be considered more reliable than findings that are not supported by more than one data source. Data from each source are analyzed separately in Section 3. To the extent possible, data from the various sources are integrated in Section 4. 29 ------- 3.0 SUMMARIES OF DATA FROM INDIVIDUAL SOURCES Data on the concentrations of hazardous constituents present in wastes at disposal sites and in individual industrial waste streams at their point of generation are provided in this chapter. The approach used to compile the concentration data is explained in Section 3.1. A summary of the hazardous constituent concentrations reported present in wastes at disposal sites is contained in Section 3.2. Section 3.3 provides a summary of the hazardous constituent concentrations reported present in individual industrial waste streams at their point of generation, prior to any commingling with other waste streams as part of the disposal process. Note that Sections 3.2 and 3.3 present the data from each individual data source separately. Integration of the data from the individual sources is discussed in Section 4. 3.1 Approach The general approach used to compile the data involved four steps. First, constituents selected to be included in the analysis were those on the CERCLA reportable quantity list (see 40 CFR 302). This list contains about 700 constituents that are considered hazardous under CERCLA. Second, the total concentration of hazardous constituents in a waste sample was determined by summing the concentrations of the individual hazardous constituents reported to be detected in the sample. For example, if a drum sample was reported to contain 5 parts per million (ppm) of arsenic; 12 ppm of 31 ------- cadmium; 13 ppm of chromium; 30 ppm of lead; and 130 ppm of trichloroethylene, then the total concentration of hazardous constituents in that drum sample, for purposes of this analysis, was 190 ppm. The third step of the approach involved stratifying the data according to waste management practice (e.g., drums, tanks). There are two main reasons for compiling the data this way. First, such classifications are generally inherent to the manner in which waste samples are collected at a site. Second, the total concentration of hazardous constituents in hazardous wastes is expected, a priori, to vary across different waste management units. For example, the concentrations of hazardous constituents in containerized wastes are generally expected to be greater than the concentrations in surface impoundment wastes which typically consist of large amounts of water. (See Appendix D for definitions of the waste management methods included in this report.) The last step of the approach involved identifying NPL and non-NPL site samples, when possible. Generally the data sources provided waste composition data and site names for each of the samples. MITRE then classified these samples as either NPL or non-NPL site samples. A review of the data from the sources discussed in Section 2 generally showed the sampling and analysis programs undertaken by the different studies to be extremely diverse due to the different 32 ------- objectives of each sampling and analysis program. The various studies examined different subsets of hazardous constituents in their sample analyses. Total concentrations in this report are necessarily based only on those constituents reported present by each of the various studies. It is likely that many of the samples also contained other CERCLA hazardous constituents which were either not analyzed for, or which were not reported, by the various studies. With a few exceptions, as discussed below, constituents were excluded from the current analysis only if they did not appear on the list of CERCLA hazardous substances. 3.2 Summary of Data From Samples of Wastes at Disposal Sites Data on the concentration of hazardous constituents present in waste samples collected from disposal sites are described in this section. In particular, the available data relate to the concentration of hazardous constituents present primarily in wastes placed in drums, tanks and surface impoundments, as well as wastes disposed at mining sites. Within each category, data are presented separately for each data source. 3.2.1 Drum Samples Among the identified data sources, the Contract Laboratory Program (CLP) and the National Enforcement Investigations Center (NEIC) data bases provide the most information on wastes in drums. The SI and RI reports also provide data on drummed wastes. This 33 ------- section summarizes the total concentration of hazardous constituents reported in drum samples from these data bases. 3.2.1.1 CLP Drum Data. The CLP drum data, which have been obtained from the automated CLP data base, consist of 261 samples taken from 81 sites. As of January 1984, 65 of the sites (80 percent) were reported to be non-NPL sites, and 16 sites (20 percent) were reported to be NPL sites.* Descriptive statistics for the total concentration of hazardous constituents in CLP drum samples are presented in Table 3. As shown in this table, for a majority of the 261 drum samples (98 percent), the data base contains only partial information on the hazardous constituents in the sample, i.e. analytical results are available either for the organic or inorganic portion of the sample analysis, but not for both. The 5 samples that contain results for both the organic and inorganic portions of the sample are from non-NPL sites and have a mean total concentration of 15 ppm + 147 percent.** This mean concentration is seen to be very low when compared to the concentrations of the drum samples for which only partial analyses are available. The 158 samples for which only the organic analysis *The portion of NPL and non-NPL sites was estimated at the time the automated CLP data base was developed. Precise site names are not available for many sites in the automated CLP data base. Consequently, the reported distribution of NPL and non-NPL sites cannot be verified, nor can the reported distribution be updated to reflect changes in the NPL since January 1984. **15 ppm + 147 percent indicates that the standard deviation about the mean of 15 ppm is 147 percent of 15 (or 22 ppm). 34 ------- TABLE 3 DESCRIPTIVE STATISTICS FOR THE TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS PRESENT IN CLP DRUM SAMPLES Total Concentration of Hazardous Constituents* (ppm) Hazardous Constituents/Site Organics and Inorganics NPL Sites Non-NPL Sites All Sites Organ! cs Only** NPL Sites Non-NPL Sites All Sites Inorganics Only*** NPL Sites Non-NPL Sites All Sites No. of Samples 0 5 5 42 116 158 30 68 98 Minimum __ 0.84 0.84 0.004 0.0002 0.0002 0.07 0.0001 0.0001 Maximum __ 53 53 21,000 464,300 464,300 14,000 60,218 60,218 Mean _ 15 15 760 12,890 9,665 2,343 1,362 1,662 Median — 6 6 21 1 4 95 15 18 Standard Deviation — 22 22 3,249 50,839 43,873 3,704 7,355 6,456 *These estimates exclude sodium whenever reported in a sample. Sodium was excluded to avoid counting its nonhazardous compounds which are likely to be the predominate form present. Concentrations of all other CERCLA hazardous constituents are included when reported. **Samples for which only the organic portion of the analysis is available. ***Samples for which only the inorganic portion of the analysis is available. Note: Dashes indicate not applicable. ------- is available have a mean total concentration of 9,665 ppm + 454 percent, while the 98 samples for which only the inorganic analysis is available have a mean total concentration of 1,662 ppm + 388 percent. In contrast to the 15 ppm estimate, the mean total concentration of hazardous constituents obtained by adding the means developed from the two sets of partial analyses provides a more meaningful estimate. In this case, the mean total concentration is 11,327 ppm. Upon further examination of the information displayed in Table 3, the mean total concentration for organics only at non-NPL sites is 12,890 ppm + 394 percent. This estimate is considerably greater than the corresponding mean (760 ppm + 428 percent) derived for the NPL site samples. There are several possible reasons for this unexpected result. First, an unknown number of the non-NPL sites, especially those with high concentrations of hazardous constituents, may be non-NPL sites only because they are still being evaluated for the NPL. Analysis of site data by the CLP program is one of the preliminary steps in evaluating whether a site belongs on the NPL. Second, for reasons previously discussed, it is possible that some sites were misclassified when the data base was developed as to whether they are NPL or non-NPL sites. Third, the data may not be representative of either NPL or non-NPL sites, or of both. It is possible, for example, that the non-NPL site from which the 464,300 ppm sample comes may, in fact, presently be an NPL site or 36 ------- may become an NPL site in the future. Dropping this one sample from the non-NPL site samples would lower the mean total concentration from 12,890 ppm to 8,965 ppm. Alternatively, adding this one sample of 464,300 ppm to the NPL samples would increase the NPL mean for organics only from 760 ppm to 11,540 ppm. In the case of inorganics only, the mean total concentration for the non-NPL site samples in Table 3 is strongly influenced by the one 60,218 ppm sample. Again, if this one sample were misclassified or subsequently became an NPL site in the future, the mean total concentration of hazardous constituents for non-NPL sites would be reduced considerably—from 1,362 ppm to about 484 ppm, about one-third of the original estimate. Alternatively, the mean total concentration for NPL sites would be increased from 2,343 ppm to 4,210 ppm. Table 3 generally shows that the mean total concentrations are much greater than the estimated medians. This result is a consequence of many drum samples having low total concentrations, and a few samples having very high total concentrations (e.g., 464,300 ppm). In addition, the standard deviations are observed to be generally 2 to 5 times greater than the means. The high standard deviations illustrated in this table indicate a high variability of the total concentration of hazardous constituents among the samples. Information about the physical conditions of the drums is not available in the automated CLP data base. Consequently, it is not 37 ------- known to what extent the drums were open and the contents diluted or concentrated by environmental conditions (e.g., rain, volatilization, evaporation) before sampling and analysis occurred. 3.2.1.2 NEIC Drum Data. The NEIC data base contains information for drum and high hazard samples. The high hazard samples (i.e., samples suspected of having high concentrations of hazardous constituents) were obtained from waste pits or ponds, waste piles, tanks, and heavily contaminated soils. The data base contains the organic and/or inorganic analysis for approximately 1,600 samples. Based on an initial review of the data, both the organic and inorganic portions of a sample analysis could be identified for 284 samples. Information supplied by EPA's Sample Management Office (personal communication with Linda Boynton, March 1986) allowed organic and inorganic analytical results to be matched for an additional 296 samples. As a result, complete information for 580 samples was extracted from the NEIC data base. The mean total concentration of hazardous constituents for these samples is 68,614 ppm + 201 percent. Other descriptive statistics on the concentrations of the 580 NEIC samples are 0.64 ppm (minimum total concentration); 950,535 ppm (maximum total concentration); 12,809 ppm (median total concentration); and 138,184 ppm (standard deviation). Figures 1 and 2 illustrate the frequency distributions for hazardous constituent concentrations in the 580 drum and high hazard 38 ------- PERCENT OF SAMPLES 50 45 40 35 30 25 20 15 10 0- 10,000- 20,000- 30,000- 40,000- 50,000- 60,000- 70,000- 80,000- 90,000- 9,999 19,999 29,999 39,999 49,999 59,999 69,999 79,999 89,999 99,999 TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS (PPM) > 100,000 FIGURE 1 FREQUENCY DISTRIBUTION OF HAZARDOUS CONSTITUENT CONCENTRATIONS IN 580 DRUM AND HIGH HAZARD SAMPLES FROM THE NEIC DATA BASE (OPEN INTERVAL) ------- PERCENT OF SAMPLES 90 80 70 60 50 40 30 20 10 0- 99,999 100.000- 199,999 200,000- 299,999 300,000- 399,999 400,000- 499,999- 500,000- 599,999 600,000- 699,999 700,000- 799,999 800,000- 899,999 900,000- 1,000,000 TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS (PPM) FIGURE 2 FREQUENCY DISTRIBUTION OF HAZARDOUS CONSTITUENT CONCENTRATIONS IN 580 DRUM AND HIGH HAZARD SAMPLES FROM THE NEIC DATA BASE (CLOSED INTERVAL) ------- samples in the NEIC data base. The frequency distribution shown in Figure 1 is based on an open interval scale measured in units of 10,000 ppm up to 100,000 ppm. Alternatively, Figure 2 presents the same information, but the intervals are in units of 100,000 ppm over the entire scale through 1,000,000 ppm. As shown in Figure 1, 47 percent of the samples have total concentrations less than 10,000 ppm and 19 percent have total concentrations greater than 100,000 ppm. While this figure appears to illustrate a bimodal frequency distribution, Figure 2 shows that this observation is merely an artifact of the scale used in Figure 1. In fact, the frequency distribution for the 580 NEIC samples is unimodal. When the open-ended interval in Figure 1 is extended from 100,000 ppm to 1,000,000 ppm in increments of 10,000 ppm, there is no evidence of a multimodal frequency distribution. Figure 2, developed with a closed interval scale, presents this basic finding. Eighty-one percent of the samples have total concentrations of hazardous constituents less than 100,000 ppm. Eight percent of the samples have total concentrations between 100,000 ppm and 200,000 ppm, and 6 percent of the samples have total concentrations between 200,000 ppm and 300,000 ppm. The total concentration of hazardous constituents for the remaining 5 percent of the samples declines almost continuously between 300,000 ppm and 1,000,000 ppm. Although information as to whether the drum and high hazard samples are from NPL or non-NPL sites is not available in the NEIC 41 ------- data base, this information is available from the EPA Sample Management Office (SMO) (personal communication with Linda Boynton, March 1986). A site listing provided by this office allowed 327* of the 580 samples (56 percent) to be identified by site. The 327 samples were taken from 36 sites, of which 14 sites (39 percent) are NPL sites, and the remainder (61 percent) are non-NPL sites. Table 4 shows the distribution of these NEIC samples across the 10 EPA regions. Sixty-nine percent of the samples were obtained from EPA Regions 2 and 5. While EPA Regions 7 and 8 are not represented in Table 4, samples from these regions are included in the other 253 NEIC samples. A statistical summary of the NEIC drum and high hazard sample data for the 327 identified NPL and non-NPL site samples is provided in Table 5. The results presented in this table show that the mean total concentration for all these sites is 82,994 ppm (or 8.3 weight percent). Non-NPL site samples have a mean total concentration of about 76,000 ppm + 180 percent while the mean for NPL sites is approximately 91,000 ppm + 162 percent. A comparison of mean total concentrations in Table 5 shows that the mean derived for the 181 non-NPL site samples is 76,424 ppm + 180 percent, and the mean for all 580 NEIC samples is 68,614 ppm + 201 percent. Intuitively, one *In the course of collecting the raw data from the EPA regions, NEIC discovered that some sample identifications were altered from those identifications originally assigned by the Sample Management Office. This explains, in part, why only 56 percent of the 580 samples could be identified by site. 42 ------- TABLE 4 DISTRIBUTION OF DRUM AND HIGH HAZARD* SAMPLES IN THE NEIC DATA BASE Number of Sites EPA Region 1 2 3 4 5 6 7 8 9 10 NPL 2 1 1 3 3 4 0 0 0 0 Non-NPL 1 1 3 1 4 7 0 0 3 2 Total 3 2 4 4 7 11 0 0 3 2 Number of Samples NPL 5 20 8 9 92 12 0 0 0 0 Non-NPL 1 78 4 1 37 18 0 0 33 9 Total 6 98 12 10 129 30 0 0 33 9 Total U.S. 14 22 36 146 181 327 *High hazard samples are samples suspected of having high concentrations of hazardous constituents at the time of sampling. The samples were obtained from waste piles, waste pits or ponds, tanks, and heavily contaminated soils. 43 ------- TABLE 5 SUMMARY STATISTICS FOR THE TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS PRESENT IN NEIC DRUM AND HIGH HAZARD SAMPLES FOR IDENTIFIED NPL AND NON-NPL SITES Site Samples No. of Samples Total of Hazardous Minimum Maximum Concentration Constituents* (ppm) Mean Median SD** NPL Sites 146 Non-NPL Sites 181 Total 327 14 681,920 91,139 22,945 147,237 1 897,422 76,424 16,529 137,352 1 897,422 82,994 18,936 141,819 All Complete NEIC Samples 580*** 0.6 950,535 68,614 12,809 138,184 *These estimates exclude sodium whenever reported in a sample. Sodium was excluded in order to avoid counting its nonhazardous compounds which are likely to be the predominate form present. **SD: Standard deviation. ***Includes the 253 samples that could not be identified as NPL or non-NPL site samples. 44 ------- would expect the mean for all 580 samples to be greater than the mean for the non-NPL site samples. The unexpected result may be due to properties of the 253 NEIC samples that could not be identified as NPL or non-NPL site samples. One possible explanation for the result is that most of the unidentified samples may be non-NPL site samples with low total concentrations, thereby pushing the mean estimate downwards. Another possible reason is that the proportion of drummed waste samples and other waste samples may be significantly different in the two subsets of data. This would likely affect the results. Similar to the CLP drum data, the median total concentrations estimated from the NEIC data are considerably less than their means. Again, this observation is a consequence of many samples having relatively low total concentrations, and only a few samples having very high concentrations. In addition, the standard deviations for the NEIC data are generally 1.6 to 2 times greater than the means. The frequency distributions for the total concentration of the hazardous constituents reported in these 327 NEIC samples are presented in Figures 3 and 4. These distributions were constructed on the basis of drum and high hazard samples from both NPL and non-NPL sites. As in Figures 1 and 2 for the 580 NEIC samples, the frequency distributions in Figures 3 and 4 present the same information, but use two different interval scales for measuring the total concentration of hazardous constituents. In Figure 3, the 45 ------- PERCENT OF SAMPLES 45 40 35 30 25 20 15 10 0- 10,000- 20,000- 30,000- 40,000- 50,000- 60,000- 70,000- 80,000- 90,000- 9,999 19,999 29,999 39,999 49,999 59,999 69,999 79,999 89,999 99,999 TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS (PPM) FIGURE 3 FREQUENCY DISTRIBUTION OF HAZARDOUS CONSTITUENT CONCENTRATIONS IN DRUM AND HIGH HAZARD SAMPLES FROM THE NEIC DATA BASE FOR IDENTIFIED NPL AND NON-NPL SITE SAMPLE (OPEN INTERVAL) ------- PERCENT OF SAMPLES 80 70 60 50 40 30 20 10 0- 99,999 100,000- 199,999 200.000- 299,999 300,000- 399,999 400,000- 499,999- 500,000- 599,999 600,000- 699,999 700,000- 799,999 800,000- 899,999 900,000- 1,000,000 TOTAL CXDNCENTRAT10N OF HAZARDOUS CONSTITUENTS (PPM) FIGURE 4 FREQUENCY DISTRIBUTION OF HAZARDOUS CONSTITUENT CONCENTRATIONS IN DRUM AND HIGH HAZARD SAMPLES FROM THE NEIC DATA BASE FOR IDENTIFIED NPL AND NON-NPL SITE SAMPLES (CLOSED INTERVAL) ------- units are measured in increments of 10,000 ppm up to 100,000 ppm while in Figure 4 the units are in 100,000 ppm over the entire scale, through 1,000,000 ppm. Figure 3 shows that approximately 41 percent of the samples have total concentrations of hazardous constituents less than 10,000 ppm, or 1.0 weight percent. In contrast, 25 percent of the samples have total concentrations greater than 100,000 ppm, or 10.0 weight percent. (This compares to percentages of 47 and 19 for the entire 580 samples.) Although the frequency in Figure 3 exhibits a bimodal distribution, this result, again, is an artifact of the open interval scale used to develop the distribution. Figure 4 shows the unimodal frequency distribution actually underlying these data. Only 3 percent of the 327 samples have total concentrations exceeding 500,000 ppm (or 50 weight percent). Frequency distributions were also separately developed for concentrations of hazardous constituents in drummed and high hazard wastes at NPL sites and non-NPL sites to determine if any significant differences could be observed. As shown by Figures 5 and 6, the frequencies for NPL and non-NPL site samples do not appear to exhibit any major disparities. Using an interval scale of 100,000 ppm, the frequency distributions in each figure are unimodal. In both cases, the greatest portion of samples (74 percent of NPL site samples and 76 percent of non-NPL site samples) has concentrations less than 100,000 ppm. In addition, 44 percent of the non-NPL site samples and 37 percent of the NPL site samples have concentrations less than 48 ------- PERCENT OF SAMPLES 80 70 60 50 40 30 20 10 0- 99,999 100,000- 199,999 200,000- 299,999 300,000- 399,999 400,000- 499,999- 500,000- 599,999 600,000- 699,999 700,000- 799,999 800,000- 899,999 900,000- 1,000,000 TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS (PPM) FIGURE 5 FREQUENCY DISTRIBUTION OF HAZARDOUS CONSTITUENT CONCENTRATIONS IN DRUM AND HIGH HAZARD SAMPLES FROM THE NEIC DATA BASE FOR IDENTIFIED NPL SITE SAMPLES ------- t_n O PERCENT OF SAMPLES 80 70 60 50 40 30 20 10 0- 100,000- 200,000- 300,000- 400,000- 500,000- 600,000- 700,000- 800,000- 900,000- 99,999 199,999 299,999 399,999 499,999- 599,999 699,999 799,999 899,999 1,000,000 TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS (PPM) FIGURE 6 FREQUENCY DISTRIBUTION OF HAZARDOUS CONSTITUENT CONCENTRATIONS IN DRUM AND HIGH HAZARD SAMPLES FROM THE NEIC DATA BASE FOR IDENTIFIED NON-NPL SITE SAMPLES ------- 10,000 ppm. Furthermore, 3 out of 181 non-NPL site samples (2 percent) were found to have total concentrations exceeding 500,000 ppm. In comparison, for the NPL sites 7 of the 146 samples (5 percent) have total concentrations greater than 500,000 ppm.* In summary, the figures do not indicate any major differences between the frequency distributions for NPL and non-NPL site samples.** 3.2.1.3 SI and RI Drum Samples. Tables 6 and 7, respectively, provide statistical summaries of the drummed waste samples compiled from the SI and RI reports. (See Appendix E for a list of sites providing these data.) Similar to the frequency distributions prepared from the NEIC drum data, the frequency distributions developed from the SI and RI drum samples are also unimodal. For the 58 SI drum samples, 53 percent of the samples have total concentrations less than 10,000 ppm, 83 percent have total concentrations less than 100,000 ppm, and 7 percent have total concentrations greater than 500,000 ppm. As observed from Table 6, these samples have a mean total concentration *A review of the SI and RI reports (see Section 3.2.1.3) identified several sites where drums had been sampled. For comparison purposes, drum data from these these reports were compared to these NEIC data. The mean total concentration of hazardous constituents detected in the SI drummed waste samples was approximately 84,000 ppm—falling towards the higher end of the frequency distribution shown in Figure 5. The RI drummed waste samples had a mean total concentration of about 106,000 ppm. In comparison, the 580 NEIC samples had a mean total concentration of 68,600 ppm. **Moreover, if the NEIC data are assumed to be random, use of the Wilcoxon rank sum median test (see Section 4.3) reveals that there are no statistically significant differences between the median total concentrations for NPL and non-NPL sites. 51 ------- TABLE 6 SUMMARY OF WASTE COMPOSITION DATA COMPILED FROM SUPERFUND SITE INSPECTION REPORTS Waste Sample Source Drums Tanks - Aqueous liquid - Sludge/solid Surface Impoundments** - Aqueous liquid - Sludge/sediment Mine Tailings Sludges Waste Piles Land farm (soil) Landfill (sediment) No. of Sites 11 4 2 2 24 15 17 3 3 2 1 1 No. of Samples 58 6 3 3 55 23 32 19 3 11 1 2 Total Concentration of Hazardous Constituents (ppm) Minimum 4 0.8 3 0.05 0.5 5,556 618 224 — 0.6 Maximum 907,500 484 3,128 98,703 649,251 94,882 19,000 339,759 — 73 Mean 84,261 162 1,971 6,656 62,751 51,904 7,559 182,924 7,221 37 Median 6,357 0.8 2,781 47 12,603 62,649 3,060 198,763 — 37 SD* 198,810 279 1,713 21,540 141,802 29,865 9,983 108,318 — — *SD: Standard deviation. **Includes lagoons, waste ponds, waste pits, and waste trenches. Note: Dashes indicate not applicable. Source: Superfund Site Inspection Reports. ------- TABLE 7 SUMMARY OF WASTE COMPOSITION DATA COMPILED FROM SUPERFUND REMEDIAL INVESTIGATION REPORTS Ui u> Waste Sample Source Drums Tanks - Aqueous liquid - Sludge/solid Surface Impoundments** - Aqueous liquid - Sludge/sediment Mine Tailings Sludges*** Rubber Chips Spent Lime Cells No. of Sites 5 4 3 3 7 5 7 1 2 1 1 No. of Samples 26 53 45 8 20 8 12 1 4 1 4 Total Concentration of Hazardous Constituents (ppm) Minimum 17 2 72 3 175 — 3 — 0.01 Maximum 732,712 12,882 98,339 1,907 161,005 — 2,247 — 10,407 Mean 106,283 3,137 22,444 521 37,089 188,844 1,294 63 2,743 Median 16,082 2,682 5,798 174 8,326 — 1,464 — 283 SD* 180,744 3,290 34,954 710 57,437 — 1,041 — 5,116 *SD: Standard deviation. **Includes lagoons, waste ponds, and waste pits. ***The sludges sampled at one site are actually industrial waste stream samples. Note: Dashes indicate not applicable. Sources: CH2M Hill (1985a, 1985b, 1985c, 1985d, 1985e, 1985f, 1984); Louisiana Department of Natural Resources (1983); NUS Corporation (1985a, 1985b, 1985c, 1985d, 1985e, 1984a, 1984b, 1984c); and TRC Environmental Consultants, Inc. (1985). ------- of 84,261 ppm + 236 percent. For the 26 RI drum samples, 35 percent of the samples have total concentrations less than 10,000 ppm, 77 percent have total concentrations less than 100,000 ppm, and 4 percent have total concentrations exceeding 500,000 ppm. These samples have a mean total concentration of 106,283 ppm + 170 percent. The median total concentrations for both the SI and RI data are considerably less than their corresponding means. This finding is the same for the NEIC data. Such observations suggest that the populations from which the drum samples were drawn are not normally distributed (see Section 4.3). The standard deviations are 1.7 to 2.4 times greater than the means in the two cases. 3.2.2 Tank Samples The SI and RI reports and the Franklin Waste Oil Study provide data on wastes in tanks. These data are briefly discussed below. 3.2.2.1 SI and RI Tank Samples. Four of the SI reports contain data for waste samples from tanks. Generally, the tank samples were classified as either aqueous liquid or sludge/solids. Table 6 shows that sludge has the higher mean total concentration of hazardous constituents for the various tank contents reported. The sludge has a mean total concentration of 1,971 ppm + 87 percent, while the liquids have a mean total concentration of 162 ppm + 172 percent. Four of the 50 RI reports also provided information about the concentration of hazardous constituents in wastes contained in tanks 54 ------- at NPL sites.* Table 7 presents the statistical findings for 45 tank liquid samples and 8 tank sludge samples from these 4 sites. The sludge samples have a mean total concentration of 22,444 ppm + 156 percent, while the liquid samples have a mean total concentration of 3,137 ppm + 105 percent. The medians for the tank samples are less than the estimated means. The median concentration for sludge, in particular, is considerably less than its mean. This is due to the presence of a few samples with extremely high total concentrations. Upon comparing the SI and RI data for tanks in Tables 6 and 7, the concentrations reported in the SI reports are considerably lower than the concentrations in the RI reports. 3.2.2.2 Waste Oil Tank Samples. The waste oil data base constructed by Franklin Associates, Ltd. provides information for the concentration of hazardous constituents in waste oil stored in tanks and drums. According to this data base, waste oil stored in tanks and drums contains, on the average, about 83,000 ppm (8.3 weight percent) of hazardous constituents.** The concentrations of hazardous constituents found in the tank waste samples for aqueous liquids and sludges shown in Tables 6 and 7 are considerably less than the total *See Appendix E for a list of these sites. There were no sites for which both SI and RI data were available. **See Appendix C, Table C-l, for a listing of the specific contaminants and their concentrations found in these waste oil samples. The estimate of 83,000 ppm is not for tanks alone; the concentration is estimated from both tanks and drum samples. The drum data in the waste oil data base could not be separated from the tank data. 55 ------- concentration estimated from the waste oil data in the Franklin data base. 3.2.3 Surface Impoundment Samples The SI and RI reports and the PEDCo study provide data for waste samples from surface impoundments. (See Appendix E for a list of sites for which the SI and RI data are available.) In this section only data from the SI and RI reports are considered. Section 3.2.4 discusses the PEDCo data on mining wastes in surface impoundments (i.e., pond liquid and pond settled solids). Waste samples drawn from surface impoundments were usually categorized as either aqueous liquid or sludge/sediment in the SI and RI reports. Like the tank wastes, Table 6 shows that for the SI data the mean total concentration of hazardous constituents is greater for sludge (62,751 ppm + 226 percent) than for aqueous liquid (6,656 ppm + 324 percent). Seven of the 50 RI reports presented data on the concentration of hazardous constituents in surface impoundment wastes. According to the waste sample data compiled from the RI reports in Table 7, the mean total concentration for surface impoundment sludge is 37,089 ppm + 155 percent. In contrast, the mean total concentration for aqueous liquids is 521 ppm + 136 percent. 3.2.4 Mining Waste Samples Data on the concentration of selected hazardous constituents present in wastes at mining sites have been prepared by PEDCo 56 ------- Environmental, Inc. (1983). These mining wastes were analyzed primarily for metals and other inorganics. Table 8 contains preliminary data on the concentration of hazardous constituents for three types of wastes in 7 mining segments.* The statistics presented in Table 8 show that pond settled solids at mining sites have the greatest mean total concentration (4,340 ppm + 189 percent), and pond liquid the lowest (27 ppm + 367 percent). These mean concentrations are considerably less than the mean concentrations for nonmining wastes in surface impoundments at SI and RI sites. None of the sites sampled in the PEDCo study are NPL sites. Additional data on tailings disposed at mining sites are provided in 3 SI reports and 1 of the RI reports (see Appendix E, Tables E-l, and E-2, respectively). Tables 6 and 7 report the findings from these data sources. The statistics shown in Tables 6 and 7 contrast sharply with those developed using the PEDCo data (see Table 8). The mean total concentration for tailings derived from the PEDCo data is 1,901 ppm + 159 percent for mining sites not on the NPL. Data extracted from the RI and SI reports, respectively, yield mean total concentrations in tailings of 188,844 ppm (single sample for an NPL listed site) and 87,144 ppm + 68 percent (samples from 3 mining sites, one non-NPL and two proposed for the NPL). The pronounced differences in the PEDCo and SI/RI estimates are even *The mining segments include uranium, phosphate, copper, lead/zinc, molybdenum, gold/silver, and iron. 57 ------- TABLE 8 PRELIMINARY DATA ON THE CONCENTRATION OF SELECTED HAZARDOUS CONSTITUENTS IN INDICATED MINING WASTES Total Concentration of Hazardous Constituents* (pom) Waste Sample Source Pond Liquid** Pond Settled Solids*** Tailings**** No. of Samples 60 79 44 Minimum 0.07 33 4 Maximum 706 47,983 16,018 Mean 27 4,340 1,901 Median 2.02 1,613 909 Standard Deviation 99 8,205 3,023 00 *These estimates exclude sodium when reported. Sodium was excluded in order to avoid counting its non-hazardous compounds which are likely to be the predominate form present. **Tailings pond liquid and mine water pond liquid. ***Tailings pond settled solids and mine water pond settled solids. ****Fresh tailings and sand tailings. Source: Preliminary data compiled from PEDCo Environmental, Inc., November 1983. Evaluation of Management Practices for Mine Solid Waste Storage. Disposal and Treatment, (Draft Report), Volume 1, Characterization of Mining Industry Wastes, prepared for the U.S. Environmental Protection Agency, Washington, DC. ------- more evident upon observing that the mean total concentrations for the SI and RI data exceed the maximum total concentration reported at any site in the PEDCo study. 3.2.5 Other Samples of Wastes at Disposal Sites The SI and RI reports provide some additional sampling data for sludges, waste piles, and landfarms, among others, at disposal sites. Tables 6 and 7 include the findings from these reports. The mean total concentration of hazardous constituents in waste piles is extremely large compared to the other wastes in Table 6. This high concentration may be influenced by the specific characteristics of the 1 site from which 10 of the 11 samples were obtained. Historically, that waste site was a manufacturing facility for fertilizer additives. The waste piles—which consisted of virgin flue dust, off-spec fertilizer, micronutrient fertilizer, and mixtures of flue dust and dirt—had high zinc concentrations, ranging between 1,660 ppm and 284,000 ppm for each of the 10 piles sampled. The SI and RI reports also present data on several waste types analyzed at only a single site. The mean total concentrations of hazardous constituents in these wastes range from 37 to 7,221 ppm. 3.3 Summary of Data from Samples of Industrial Waste Streams at the Point of Generation This section discusses the concentration data compiled from two sets of industry studies conducted for EPA. One set of studies was conducted during the mid-1970s. The other is a currently on-going effort that, to date, has focused on the organic chemicals industry. 59 ------- The data reported in both these studies were derived from industrial waste streams at their point of generation, prior to disposal. Such waste streams are likely to be combined with other wastes as part of the disposal process. Consequently, the composition and concentrations of constituents in wastes at disposal sites are likely to be different from those of the individual waste streams. Furthermore, there may be differences due to changes in processes and products over time. Also included in this section is a discussion of the concentrations of hazardous constituents in waste oil and in selected waste streams of the waste oil re-refining industry. 3.3.1 Previous Industry Studies of the EPA Office of Solid Waste (OSW7 Table 9 contains a highly aggregated summary of the average total concentrations of hazardous constituents reported present in selected industrial waste streams during the 1970's. Data presented in this table for the paint manufacturing industry were obtained from an EPA Background Document (see Section 2.7). The remaining data were obtained from the 1970's OSW industry studies. The concentration data are based almost exclusively on analysis of only the heavy metals content of the waste streams. Due to the different ways in which data were reported for different industry segments, the average total concentrations in Table 9 have been derived in different ways. In some cases, the average total concentration 60 ------- TABLE 9 AVERAGE TOTAL CONCENTRATIONS OF HAZARDOUS CONSTITUENTS IN INDUSTRIAL WASTE STREAMS FROM PREVIOUS EPA INDUSTRY STUDIES* (ppm) Solids Paint Spent Waste Untreated Wastes & Clay Industry Sludges** Oils Wastevater Solvents Sand Dusts Filters Residues Miscellaneous 1. Textiles 5,300 2. Plastic 200 Materials & Synthetics 3. Paint 200*** Manufacturing 4. Solvent 3,200 Reclamation 5. Petroleum 2,000 Refining 6. Petroleum 18,800 Re-refining 7. Leather 4,200 Tanning & Finishing 8. Metal 98,300 Smelting & Refining 300 80*** 1,600 10 20 1,100 50,600 7,000 100 32,300 12,100 200a l,100b 7,600C 86,600 139,000d ------- TABLE 9 (Concluded) 10. 11. Solids Industry Waste Sludges** Oils Paint Untreated Wastes & Wastewater Solvents Sand Dusts Spent Clay Filters Residues Miscellaneous Electroplating 93,000 & Metal Finishing Special Machinery Manufacturing Electronic Components Manufacturing 700 18,800 1,600 90 600 *The waste streams were analyzed primarily for metals. The concentrations in the table do not include sodium. This exclusion avoids counting nonhazardous sodium compounds which are likely to be the predominate form present. **Includes silt and still bottoms. ***Estimates are from EPA Background Document, June 7, 1980, Table 6, p. 123 and Table 7, p. 124. aAverage of mean concentrations for sweepings and floor wastes, spent alumina, and waste nylon salt. bAverage of mean concentrations for dissolved air flotation float, coke fines, spent lime, and fluid catalytic cracker catalyst fines. cAverage of mean concentrations for chrome trimmings and shavings, chrome fleshings, leather trim, and sewer screenings. ^Average of mean concentrations for slag and potroom skimmers. Note: Dashes indicate not available or not applicable. Source: Appendix A, Tables A-2 through A-15. ------- reflects the sum of mean concentrations for Individual constituents present in the industrial waste stream (e.g., sludges in the textiles industry). In other cases, the average total concentration is derived as an average of the mean total concentrations generated by the various industry segments within that industry. In such cases, the concentrations were derived as simple arithmetic averages; they have not been weighted, for example, by the quantity of the waste stream generated by each industry segment producing that waste stream.* Table 9 shows that all of the 11 industries generate some type of sludge.** Two industries in particular, Metal Smelting & Refining and Electroplating & Metal Finishing, produce sludges which contain very high concentrations of hazardous constituents. A variety of other waste streams are also generated by the different industries. Among these other waste streams, solids (particularly spent clay filters *For example, Table A-9 (Appendix A) presents average concentration data for hazardous constituents in selected sludges generated by 12 different categories of smelters and refiners (e.g., primary and secondary ferrous and nonferrous smelters). For each of the 12 smelters/refiners, the average concentrations of individual constituents present in these sludges were summed. The estimated total concentrations were summed across the smelters/refiners and then divided by 12 (i.e., the number of categories of smelters/ refiners for which concentration data are reported). The resulting estimate of 98,300 ppm was then tabulated to represent the average total concentration of hazardous constituents present in sludges generated by the Metal Smelting and Refining Industry (see Table 9). **0ther sludge samples from an industrial waste stream were reported in one of the RIs. The hazardous constituents present in these sludge samples, taken from a wastewater treatment plant, were estimated to have a mean total concentration of about 2,000 ppm. 63 ------- and residues) have relatively high concentrations of hazardous constituents. As previously discussed in Section 2.3, there are several caveats associated with the data presented in Table 9. First, the hazardous constituents for which analyses were conducted were almost exclusively metals. Furthermore, the metals being analyzed varied by industry; in some cases analyses were performed for only two or three metals, while in other cases, analyses for as many as 20 metals were performed. Second, grab samples were used to derive estimates for some industries while composite samples were used in other cases. And third, the sampling programs of the different industry studies varied greatly with respect to the number of samples taken to determine representative concentrations. 3.3.2 Current OSW Industry Studies Program More recently, the EPA Office of Solid Waste (OSW) has been collecting data on the composition of hazardous waste streams as part of the EPA Industry Studies program which began in 1980. In contrast to the earlier industry studies, the waste samples in the present Industry Studies program are being analyzed for a wide variety of organic and inorganic constituents. However, it must be pointed out that, to date, the current Industry Studies program has focused exclusively on waste streams from the organic chemicals industry. The data collected have been used to develop the Industry Studies Data Base (ISDB). 64 ------- Tables 10 and 11 provide general summaries of the waste composition data contained in the ISDB.* The waste composition data are presented by the type of waste management method and by whether the wastes are RCRA or non-RCRA waste streams. RCRA waste streams are those waste streams considered hazardous under 40 CFR 261. The descriptive statistics shown in Table 10 have not been weighted by the quantity of the waste stream generated by the industry segment producing that waste stream. Alternatively, the information contained in Table 11 is weighted by such quantities. The quantity-weighted data in Table 11 may provide a more meaningful indicator of the total concentration of hazardous constituents in hazardous wastes at disposal sites than the data presented in Table 10; the quantity-weighted data attempt to account for the eventual commingling of waste streams. Based upon the unweighted data, wastes in piles, landfarms, and surface impoundments typically have lower total concentrations of hazardous constituents than wastes in containers, landfills, and *The constituents used to derive the concentrations in Tables 10 and 11 are those on the CERCLA reportable quantities list, with a few exceptions. Sixteen dilute acids and caustics were not included in order to prevent the concentration estimates from being greatly inflated. The only information available for these 16 acids and caustics was the total concentration of the dilute acid or caustic present in the waste stream, not the concentration of the pure acid or caustic in the waste stream. (For example, if the waste stream contained 200,000 ppm of 1 percent hydrochloric acid, this was reported as 200,000 ppm of hydrochloric acid, not as 2,000 ppm of hydrochloric acid.) Also, sodium was excluded to avoid counting nonhazardous sodium compounds. 65 ------- TABLE 10 TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS PRESENT IN ISDB WASTE STREAMS* (UNWEIGHTED) Total Concentration of Hazardous Management Method RCRA Wastes Containers Landfills Surface Impoundments Tanks Piles Land farms All Management Methods Non-RCRA Wastes Containers Landfills Surface Impoundments Tanks Piles Land farms All Management Methods No. of Samples 14 21 4 36 1 0 69*** 5 33 15 58 2 3 104*** Minimum 3 0.2 0.1 0.03 — — 0.03 4 2 2 0.01 100 38 0.01 Maximum 960,000 984,621 1,006 1,000,000 — — 1,000,000 350,000 500,000 24,000 1,000,000 10,000 40 1,000,000 Mean 364,445 241,540 384 216,675 271 — 235,583 104,140 64,853 4,552 160,850 5,050 39 105,319 Constituents** (Dom) Median 250,000 150,700 265 3,065 — — 10,000 70,000 156 200 3,500 5,050 40 592 Standard Deviation 331,687 299,709 473 358,365 — — 338,144 144,270 138,929 8,530 288,182 — 1 233,586 ------- TABLE 10 (Concluded) Total Concentration of Hazardous Constituents** (ppm) No. of Standard Samples Minimum Maximum Mean Median Deviation Management Method All ISDB Wastes**** Containers Landfills Surface Impoundments Tanks Piles Landfarms All Management Methods 19 54 19 94 3 3 173*** 3 0.2 0.1 0.01 100 38 0.01 960,000 984,621 24,000 1,000,000 10,000 40 1,000,000 295,944 133,565 3,674 182,230 3,457 39 157,274 200,000 804 200 3,500 271 40 1,100 312,970 230,454 7,725 316,192 5,667 1 286,306 *These wastes were defined to include sludges/slurries, spent solvents, solids, liquids, and untreated wastewater. **The concentrations are not weighted by the quantity of waste stream generated by the industry segment producing that waste stream. ***The number of samples for individual management methods does not sum to this figure. Double counting of data may occur between management practices due to a facility reporting more than one management practice for a given waste stream. ****RCRA and non-RCRA wastes combined. Note: Dashes indicate not applicable. Source: Science Applications International Corporation (1986). See Appendix B, Tables B-9, B-16, and B-23. ------- TABLE 11 TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS PRESENT IN ISDB WASTE STREAMS* (QUANTITY-WEIGHTED) Management Method No. of Samples Total Concentration of Hazardous Constituents** (ppm) S tandard Mean Median Deviation RCRA Wastes Containers 12 446,951 500,000 193,450 Landfills 19 103,998 82,733 107,017 Surface Impoundments 4 540 500 157 Tanks 31 518 6 13,129 Piles 1 271 — — Landfarms 0 — — — AU Management Methods 61*** 3,710 20 32,170 Non-RCRA Wastes Containers 4 8,305 293 27,570 Landfills 24 4,562 4 31,446 Surface Impoundments 13 6,305 2,800 8,115 Tanks 38 6,076 300 53,397 Piles 2 9,959 10,000 636 Landfarms 3 39 38 1 An Management Methods 72*** 7,200 2,000 33,224 An ISDB Wastes**** Containers 16 436,833 400,000 202,270 Landfills 43 27,421 100 71,727 Surface Impoundments 17 6,219 2,800 8,085 Tanks 69 2,111 50 30,763 Piles 3 9,695 10,000 1,698 Landfarms 3 39 38 i An Management Methods 133*** 5,542 51 32,774 68 ------- TABLE 11 (Concluded) FOOTNOTES *These wastes were defined to Include sludges/slurries, spent solvents, solids, liquids, and untreated wastewater. **The concentrations are weighted by the quantity of waste stream generated by the industry segment producing that waste stream. ***The number of samples for individual management methods does not sum to this figure. Double counting of data may occur between management practices due to a facility reporting more than one management practice for a given waste stream. ****RCRA and non-RCRA wastes combined. Note: Dashes indicate not applicable. Source: Science Applications International Corporation (1986). See Appendix B, Tables B-9, B-16, and B-23. 69 ------- tanks. As illustrated by the unweighted data in Table 10, this is true for both RCRA and non-RCRA wastes, as well as all ISDB wastes. For all ISDB wastes (RCRA plus non-RCRA) in this table, the mean total concentration is lowest for landfarms (39 ppm + 3 percent), and greatest for containers (295,944 ppm + 106 percent). The mean total concentration for all management methods is estimated as 157,274 ppm + 182 percent. For RCRA wastes, the mean total concentration is 235,583 ppm + 144 percent; for non-RCRA wastes it is 105,319 ppm + 222 percent. When the concentration data are quantity-weighted (Table 11), the relative magnitudes of the total concentrations differ considerably between RCRA and non-RCRA wastes. In the case of RCRA wastes, piles, tanks, and surface impoundments have considerably lower total concentrations of hazardous constituents than containers or landfills. In contrast, for non-RCRA wastes, the mean total concentrations are relatively similar for all management methods except landfarms. However, piles and surface impoundments have the highest median concentrations. Moreover, the quantity-weighted mean and median total concentrations for all management methods are higher for the non-RCRA wastes than for the RCRA wastes. For RCRA wastes the mean and median are 3,710 ppm + 867 percent and 20 ppm, respectively; for non-RCRA wastes these estimates are 7,200 461 percent (mean) and 2,000 ppm (median). 70 ------- 3.3.3 Waste Oil The study prepared by Franklin Associates (1984) provides detailed information about the hazardous constituents present in automotive and industrial waste oil. Between 1981 and 1984, data for more than 1,000 waste oil samples were collected and analyzed by Franklin Associates, Ltd. In general, the waste oil samples were analyzed for a total of 19 constituents (see Appendix C, Table C-l). As previously mentioned in Section 3.2.2, the waste oils contain approximately 83,000 ppm of hazardous constituents, on the average. A sizable portion of waste oils are re-refined for reuse. Table 12 summarizes the total concentration of hazardous constituents present in 4 waste streams of the waste oil re-refining industry. The mean concentrations range from about 1,100 ppm to almost 9,000 ppm. The mean total concentration of hazardous constituents is greatest for settled sludges and distillation bottoms. The sludges from this industry can have a waste oil content greater than 50 percent. 71 ------- TABLE 12 TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS PRESENT IN SELECTED WASTE STREAMS OF THE WASTE OIL RE-REFINING INDUSTRY Waste Stream Settled sludges Wastewater Spent clay** Distillation bottoms** Total Hazardous Minimum 180 126 58 1,188 Concentration of Constituents* (pp Maximum 318,552 21,751 2,834 20,044 m) Mean 8,964 5,673 1,136 8,445 *These estimates are based on a total number of analyzed samples ranging between 2 and 50. Constituents analyzed include a total of 19 metals, chlorinated solvents, and other organics. **Includes only heavy metals (i.e., arsenic, barium, cadmium, chromium, lead, and zinc). Source: Franklin Associates, Ltd. 1984. Composition and Management of Used Oil Generated in the United States, Final Report, prepared for the U.S. Environmental Protection Agency, Office of Solid Waste and Emergency Response, Washington, DC, Tables 41, 42, 43, and 44. 72 ------- 4.0 ANALYSIS OF DATA A comparative analysis of the total concentration of hazardous constituents in hazardous wastes is presented in this chapter. Section 4.1 provides a composite overview of the data by waste management method for all data sources. Section 4.2 provides comparisons of the total concentrations of hazardous constituents in the wastes aggregated as liquids, solids, and sludges. In addition, where feasible, the waste composition data are statistically analyzed to determine if there are significant differences in the total concentrations of hazardous constituents. Section 4.3 presents the nonparametric statistical analysis of the median total concentration of hazardous constituents in hazardous wastes. 4.1 Composite Overview of the Total Concentration of Hazardous Constituents Present in Hazardous Wastes by Type of Waste Management Method In Section 3, descriptive statistics are presented for data from individual data sources. In this section, the waste composition data for drums, tanks, surface impoundments, mine tailings/waste piles, landfarms, and landfills are integrated across the various sources. 4.1.1 Drums Table 13 summarizes the available drum data. The data are grouped in the table according to whether the waste samples were collected at the point of disposal (e.g., a hazardous waste site) or at the point of generation (e.g., an industrial plant). The table shows that the mean total concentration of hazardous constituents in 73 ------- TABLE 13 COMPOSITE OVERVIEW OF THE CONCENTRATION OF HAZARDOUS CONSTITUENTS IN DRUMMED WASTES Total Concentration of Hazardous Constituents (pom) Sampling Point Point of Disposal Point of Generation Data Source CLP NEIC SI RI ISDB Containers (Unweighted) - RCRA - Non-RCRA - All ISDB ISDB Containers (Weighted) - RCRA - Non-RCRA - All ISDB No. of Samples ND* 580 58 26 14 5 19 12 4 16 Mean 11,327** 68,614 84,261 106,283 364,445 104,140 295,944 446,951 8,305 436,833 Median ND 12,809 6,357 16,082 250,000 70,000 200,000 500,000 293 400,000 Standard Deviation ND 138,184 198,810 180,744 331,687 144,270 312,970 193,450 27,570 202,270 *Not able to be determined. **Mean concentration is based on the sum of the mean concentrations from the "inorganics only" and "organics only" portions of the automated CLP data. The "inorganics only" mean concentration is based on 98 samples, the "organics only" mean concentration is based on 158 samples. ------- drummed wastes at disposal sites ranges between 11,327 ppm and 106,283 ppm. The median total concentrations of hazardous constituents developed from the NEIC, SI, and RI data are considerably less than their means. For example, the mean total concentration for the SI data is 13 times greater than its median. For the NEIC, SI, and RI drum data, the associated standard deviations are several times greater than their means. Descriptive statistics have not been developed for the drum data aggregated across the data sources (i.e., CLP, NEIC, SI, and RI). Such an integration would require independent samples taken from the same population of sites. The CLP and NEIC data are not independent; in fact, 4 percent of the NEIC data are contained in the automated CLP data base. Furthermore, it is unlikely that the drum samples from each of the 4 sources come from the same population. For example, the RI data are derived solely from NPL sites while the other 3 data sets are derived from a mixture of NPL and non-NPL sites. In comparison to the means developed from samples taken at the point of disposal, the unweighted means for container samples taken at the point of generation are generally higher, ranging from 104,140 ppm + 139 percent to 364,445 ppm + 91 percent. In this case, as shown in Table 13, the standard deviations are approximately equal to their means, and the means are about 1.5 times greater than their medians. However, when the container data are quantity-weighted, the means range between 8,305 ppm + 332 percent and 446,951 ppm + 43 percent. 75 ------- In comparison to the unweighted data, the standard deviations are reduced, but are still quite large. 4.1.2 Tanks Table 14 provides a composite overview of the available data on the concentration of hazardous constituents between wastes contained in tanks. As in Table 13, the data are organized according to sampling point (i.e., disposal and generation). Note that there is one important difference between the data from the point of disposal and the data from the point of generation. When allowed to remain in a tank for a sufficient period of time, many wastes will separate into solid and liquid components. The data from these different components are those data typically available at the point of disposal. The data at the point of generation do not reflect this separation into various components. Rather, they represent the composite waste stream. At the point of disposal, tank sludges/solids in Table 14 appear to have higher mean and median total concentrations of hazardous constituents than tank liquids. Mean concentrations at the point of disposal appear to be about the same order of magnitude as the mean quantity-weighted concentrations at the point of generation. However, both generally appear to be more than an order of magnitude lower than the mean unweighted concentrations at the point of generation. This may indicate that low-volume high concentration wastes are typically mixed with high-volume low concentration wastes in tanks. 76 ------- TABLE 14 COMPOSITE OVERVIEW OF THE CONCENTRATION OF HAZARDOUS CONSTITUENTS IN TANK WASTES --4 —I Total Concentration of Hazardous Constituents Coom) Sampling Point Point of Disposal Point of Generation Data Source SI - Liquid - Sludge/solid RI - Liquid - Sludge/solid ISDB (Unweighted) - RCRA - Non-RCRA - All ISDB ISDB (Weighted) - RCRA - Non-RCRA - All ISDB Franklin Waste Oil Study No. of Samples 6 3 3 53 45 8 36 58 94 31 38 69 1,071* Mean 162 1,971 3,137 22,444 216,675 160,850 182,230 518 6,076 2,111 82,565 Median 0.8 2,781 _.__ 2,682 5,798 3,065 3,500 3,500 6 300 50 NA** Standard Deviation 279 1,713 3,290 34,954 358,365 288,182 316,192 13,129 53,397 30,763 NA *Total number of samples in the study. **NA: Not available. Note: Dashes indicate not applicable. See Appendix C, Table C-l. ------- The mean concentrations for wastes in tanks at the point of disposal generally appear to be considerably lower than the mean concentrations for drummed wastes at the point of disposal. The median concentrations also appear to be generally lower for wastes in tanks than for drummed wastes. However, the statistical analysis of median total concentrations in Section 4.3 indicates that for the RI data, the median total concentrations for drummed wastes are not significantly different at a 95 percent confidence level from the median total concentration for tank sludges. Both of these median concentrations are, however, shown to be significantly greater than the median total concentrations for tank liquids. (Due to an insufficient number of samples, a similar analysis cannot be performed with the SI data.) The data in Table 14 further indicate that for wastes in tanks at disposal sites the mean total concentrations of hazardous constituents in the wastes are generally less than the respective standard deviations. Furthermore, the mean total concentrations are generally higher than the median total concentrations. However, with regard to SI tank sludge/solid samples, the opposite is true. At the point of generation, the data in Table 14 indicate that the unweighted mean and median total concentrations of hazardous constituents in RCRA wastes going to tanks are less than the unweighted mean and median total concentrations for RCRA wastes going to drums. The same relationship holds for the quantity-weighted mea 78 ------- and median total concentrations. There is no discernable trend for the non-RCRA wastes. Furthermore, at the point of generation, the data in Table 14 indicate that for all classifications of wastes the median total concentrations for wastes going to tanks are considerably lower than their corresponding means. Also, the standard deviations exceed their respective means. Finally, the data in Table 14 indicate that the mean total concentration of hazardous constituents in waste oil (stored in tanks) is considerably greater than the mean for tank wastes in general when the ISDB data are quantity-weighted. In contrast the mean for waste oil is considerably less than the mean total concentration of hazardous constituents in tank wastes in general when the ISDB data are not quantity-weighted. 4.1.3 Surface Impoundments Four data sources provide information on the concentration of hazardous constituents in wastes contained in surface impoundments. Table 15 summarizes the available data. The distinction in the data from the point of generation and the point of disposal noted for tank wastes also applies to surface impoundment wastes (i.e., data at the point of disposal reflect separation into different components; data at the point of generation do not). Similar to the findings for wastes in tanks, at the point of disposal, surface impoundment sludges/sediments have higher mean and median total concentrations of hazardous constituents than surface 79 ------- TABLE 15 COMPOSITE OVERVIEW OF THE CONCENTRATION OF HAZARDOUS CONSTITUENTS IN SURFACE IMPOUNDMENT WASTES 00 o Total Concentration of Hazardous Constituents (com) Sampling Point Point of Disposal Point of Generation Data Source SI - Liquid - Sludge/sediment RI - Liquid - Sludge/sediment PEDCo - Liquids* - Sludges** ISDB (Unweighted) - RCRA - Non-RCRA - All ISDB ISDB (Weighted) - RCRA - Non-RCRA - All ISDB No. of Samples 55 23 32 20 8 12 60 79 4 15 19 4 13 17 Mean 6,656 62,751 ___ 521 37,089 27 4,340 384 4,552 3,674 540 6,305 6,219 Median 47 12,603 __ 174 8,326 2 1,613 265 200 200 500 2,800 2,800 Standard Deviation __ 21,540 141,802 — 710 57,437 99 8,205 473 8,530 7,725 157 8,115 8,085 *Mining wastes defined as tailings pond liquid and mine water pond liquid. **Mining wastes defined as pond settled solids. Note: Dashes indicate not applicable. ------- impoundment liquids. This is true for all three data sources. Furthermore, for the SI and RI samples, the mean and median total concentrations of hazardous constituents in both surface impoundment liquids and sludges/sediments are higher than those for the equivalent PEDCo mining waste samples. For the SI and RI data, the mean total concentrations for surface impoundment sludges/sediments at the point of disposal appear to be higher than the mean total concentrations for tank sludges/solids but generally lower than the mean total concentration for drummed wastes. No other trends are discernible from the data. (In fact, the statistical analysis in Section 4.3 indicates that for the RI data, the median total concentrations for drummed wastes, tank sludges, and surface impoundment sludges are not significantly different from each other at a 95 percent confidence level. For the SI data, the median total concentrations for drummed wastes and surface impoundment sludges are also not significantly different from each other at a 95 percent confidence level, and both are significantly different than the median total concentration for surface impoundment liquids.) At the point of generation, the data in Table 15 indicates that the unweighted mean and median total concentrations of hazardous constituents in wastes going to surface impoundments are less than the unweighted mean and median total concentrations for wastes going to drums or tanks. The quantity-weighted mean and median are also 81 ------- less than those for drummed wastes. The quantity-weighted mean is comparable to that for wastes going to tanks; the quantity-weighted median appears to exceed that for wastes going to tanks. The data in Table 15 further indicates that, for all sampling points and for all classifications of surface impoundment wastes, median total concentrations are less than the respective means.* Furthermore, standard deviations exceed the respective means in all cases but one (i.e., for quantity-weighted RCRA wastes at the point of generation). The standard deviations range from 1.2 to 3.7 times the means in all but the latter case. There, the standard deviation is about 30 percent of the mean. 4.1.4 Mine Tailings and Waste Piles Table 16 presents all the available data for mine tailings and waste piles. The data are grouped by sampling point. From Table 16, the mean total concentrations of hazardous constituents in the SI and RI mine tailings samples are greater than the mean total concentrations for mine tailings from the PEDCo data. Similarly, the median total concentration of hazardous constituents for the SI mine tailings samples is greater than the median total concentration for mine tailings from the PEDCo data. *Curiously, the estimates developed from the SI data for surface impoundments generally are greater than the corresponding estimates developed from the RI data. In all other cases (i.e., drums, tanks and mine tailings), the RI statistical estimates are greater than the SI estimates. 82 ------- oo Point of Generation TABLE 16 COMPOSITE OVERVIEW OF THE CONCENTRATION OF HAZARDOUS CONSTITUENTS IN MINE TAILINGS AND WASTE PILES Sampling Point Point of Disposal Data Source SI - Mine Tailings - Waste Piles No. of Samples 19 11 Total Concentration of Hazardous Constituents (oom) Standard Mean Median Deviation 87,144 112,498 59,006 182,924 198,763 108,318 RI - Mine Tailings PEDCo - Mine Tailings ISDB Waste Piles (Unweighted) 44 188,844 1,901 909 3,023 - RCRA - Non-RCRA - All ISDB ISDB Waste Piles (Weighted) - RCRA - Non-RCRA - All ISDB 1 2 3 1 2 3 271 5,050 3,457 271 9,959 9,695 — 5,050 271 — 10,000 10,000 — — 5,667 — 636 1,698 Note: Dashes indicate not applicable. ------- The mean and mediam total concentrations of hazardous constituents in the SI waste pile samples are greater than the mean and mediam total concentrations both for waste piles from the ISDB data (both weighted and unweighted) and for the SI mine tailings samples. The total mean concentration for the SI waste pile samples, however, is comparable to the total mean concentration for the RI mine tailings sample. For the SI mine tailings and waste piles samples, mean total concentrations are less than median total concentrations, but are greater than the standard deviations. Conversely, for the PEDCo mine tailings samples, the mean total concentration exceeds the median total concentration, but is less than the standard deviations. Comparisons of the mine tailings data in Table 16 with the other data in Tables 13 through 16 do not lead to any consistent findings. Both the mean and median total concentrations of hazardous constituents for the SI mine tailings appear to be higher than the mean and median total concentrations for all other wastes at disposal sites in Tables 13 through 16, except for the mean total concentrations in drummed wastes. The mean total concentration for the SI mine tailings is comparable to the mean total concentration for the drummed wastes at disposal sites. Conversely, both the mean and median total concentrations for the PEDCo mine tailings appear to be lower than the mean and median total concentrations for many of the wastes at disposal sites in Tables 13 tnrough 16 (especially for drummed wastes and surface impoundment sludges). 84 ------- With regard to waste piles, 10 of the 11 waste pile samples are from 10 piles at the same site. Thus, there are too few samples to make meaningful comparisons with wastes in other types of waste management units. Similarly, there are too few samples for waste piles at the point of generation to make meaningful comparisons with wastes in other types of waste management units. 4.1.5 Landfarms and Landfills Table 17 summarizes the available data for landfarms and landfills. Very little information is available on the concentration of hazardous constituents present in wastes in landfarms and landfills. For landfarms, only one sample taken at a point of disposal and only 3 samples taken at a point of generation are available from all the data sources reviewed. These are too few samples to draw any meaningful conclusions. For landfills only 2 samples are available from disposal sites; however, 54 samples are available for the point of generation. At the point of generation, both the unweighted and quantity- weighted mean and median total concentrations of hazardous constituents for wastes going to landfills appears to be less than the unweighted and quantity-weighted median total concentrations for drummed wastes (see Table 13). The unweighted mean total concentration of hazardous constituents for wastes going to landfills appears to be greater than the unweighted mean total concentration for wastes going to surface impoundments. However, except for RCRA 85 ------- 00 TABLE 17 COMPOSITE OVERVIEW OF THE CONCENTRATION OF HAZARDOUS CONSTITUENTS IN LANDFARM AND LANDFILL WASTES Total Concentration Hazardous Constituents Sampling Point Point of Disposal Point of Generation Data Source SI - Landfarm - Landfill ISDB Landfarms (Unweighted) - RCRA - Non-RCRA - All ISDB ISDB Landfarms (Weighted) - RCRA - Non-RCRA - All ISDB ISDB Landfills (Unweighted) - RCRA - Non-RCRA - All ISDB ISDB Landfills (Weighted) - RCRA - Non-RCRA - All ISDB No. of Samples 1 2 0 3 3 0 3 3 21 33 54 19 24 43 Mean 7,221 37 — 39 39 — 39 39 241,540 64,853 133,565 103,998 4,562 27,421 Median — 37 — 40 40 — 39 38 150,700 156 804 82,733 4 100 of (ppm) Standard Deviation — — — — 1 — 1 1 299,709 138,929 230,454 107,017 31,446 71,727 Note: Dashes indicate not applicable. ------- wastes, there are no other discernable treads with regard to wastes going to tanks or surface impoundments (see Tables 14 and 15). For RCRA wastes, both the unweighted and the quantity-weighted mean and median total concentrations for wastes going to landfills are greater than the unweighted and the quantity-weighted mean and median total concentrations for wastes going to tanks or surface impoundments. For all the weighted and unweighted landfill samples at the point of generation, the mean total concentrations exceed the median total concentrations, and the standard deviations exceed the mean total concentrations. 4.1.6 Summary of Findings While the quantity of waste composition data available for drums, tanks, surface impoundments, and mine tailings is relatively extensive, very little data are available on landfarms and waste piles from the sources examined. For landfills, a relatively large amount of data is available for wastes at the point of generation, but not for wastes at disposal sites. Generally, the mean total concentrations of hazardous constituents are considerably larger than the median total concentrations for the data examined. This indicates that a disproportionate number (i.e., greater than 50 percent) of the sample values are less than the mean. Also, the standard deviations tend to exceed the mean total concentrations, often by a considerable 87 ------- amount. This indicates a high variability in the total concentration of hazardous constituents for the samples examined. The total quantity of hazardous constituents present in wastes appears to vary by the type of waste management unit. At disposal sites, the available data show that the mean total concentration of hazardous constituents appear to generally exhibit an approximate ordering, from highest to lowest, as follows: • Drummed wastes • Surface impoundment sludges/sediments • Tank sludges/solids • Surface impoundment liquids • Tank liquids At disposal sites, the median total concentrations appear to generally exhibit an approximate ordering as follows: • Drummed wastes, surface impoundment sludges/sediments, tank sludges/solids • Tank liquids • Surface impoundment liquids Note that these rankings are subjective and are generally not based on statistical analysis for reasons discussed elsewhere in this report. It is likely that if a statistical analysis could have been done, it would have shown that some of the differences present in the rankings were not actually statistically significant. The rankings are also subject to the limitations noted in Section 2.8. Furthermore, the large standard deviations indicate that the rankings 88 ------- may not be meaningful. Additionally, several other types of waste management units (i.e., landfills, landfarms, mine tailings, waste piles) could not even be included in these approximate rankings. At the point of generation, the available data show that the unweighted mean total concentrations appear to generally exhibit an ordering as follows: • Drummed wastes • Tanks, landfills • Surface impoundments The quantity-weighted mean total concentrations appear to generally exhibit an approximate ordering as follows: • Drummed wastes • Landfills • Tanks, surface impoundments These rankings are subject to the limitations noted above. Furthermore, the data at the point of generation are not adequate for the development of rankings based on median total concentrations. 4.2 Comparisons of the Total Concentration of Hazardous Constituents in Liquids, Solids, and Sludges In this section the waste composition data are integrated and analyzed according to whether the data pertain to liquid wastes, solid wastes, or sludges. 4.2.1 Liquids Among the major data sources reviewed, three sources provided explicit information on the concentration of hazardous constituents 89 ------- in liquid wastes at the point of disposal. The SI reports contained 3 tank liquid samples and 23 surface impoundment liquid samples. The RI reports provided additional information—45 tank liquid samples and 8 surface impoundment liquid samples. In addition, the PEDCo study provided 60 pond liquid samples. The ISDB contained 27 samples of liquid wastes taken at the point of generation. Table 18 contains descriptive statistics for the total concentration of hazardous constituents in liquid wastes at the point of disposal and at the point of generation. For comparison purposes, a summary of drum data is also presented in this table. (Since drums may contain solids, liquids, or sludges, the drum data cannot be integrated in any one of the three groups.) The total concentrations of hazardous constituents derived from the PEDCo data are considerably lower than those concentrations estimated from the SI and RI data for liquid wastes. For all 3 data sources, the median total concentrations are much smaller than their means. The standard deviations exceed the means in all instances. The total concentration of hazardous constituents in liquid wastes contrasts sharply with the total concentrations in drummed wastes. The mean total concentration of hazardous constituents in liquid wastes from the SI data is 5,906 ppm + 344 percent; for the SI drummed wastes, the mean is 84,261 ppm + 236 percent. This difference is also observed with the RI data: the mean total 90 ------- TABLE 18 SUMMARY OF WASTE COMPOSITION DATA FOR LIQUID, SOLID, AND SLUDGE WASTES Hazardous Sampling Wastes/ No. of Point Sources Samples Point of Liquid Wastes* Disposal SI Reports RI Reports PEDCo Solids Wastes** SI Reports RI Reports PEDCo Sludges*** SI Reports RI Reports PEDCo Drums CLP NEIC SI RI 26 53 60 39 6 44 32 24 79 __ 580 58 26 Total Concentration Minimum 0.05 2 0.07 3 0.01 4 0.47 3 33 __ 0.6 4 17 Maximum 98,703 12,882 706 649,251 188,844 16,018 477,805 161,005 47,983 — 950,535 907,500 732,712 of Hazardous Constituents (vvm) Mean 5,906 2,742 27 113,368 33,313 1,901 40,327 26,241 4,340 11,327 68,614 84,261 106,283 Median 28 1,907 2 107,849 314 909 7,078 4,903 1,613 — 12,809 6,357 16,082 Standard Deviation 20,317 3,180 99 127,402 76,305 3,023 93,684 46,092 8,205 — 138,184 198,810 180,744 ------- TABLE 18 (Concluded) Sampling Point Point of Generation Hazardous Wastes/ Sources ISDB Liquids Unweighted Weighted ISDB Solids Unweighted Weighted ISDB Sludges/ Slurries Unweighted Weighted Total Concentration of Hazardous Constituents (ppm) No. of Samples 27 22 1 1 89 76 Minimum Maximum Mean 0.01 900,000 157,538 5,262 4 4 0.03 1,000,000 167,043 17,496 Standard Median Deviation 3,000 249,858 2,800 39,113 — — — — 1,538 274,492 10,000 66,650 *The liquid samples were obtained from tanks, surface impoundments, tailings ponds, and mine water ponds. **The solid samples were obtained from tanks, mine tailings, waste piles, spent lime cells, and rubber chips. ***The sludge samples were obtained from surface impoundments, sludge piles, tanks, pond settled solids, landfarms, and landfills. Note: Dashes indicate not applicable or not available. ------- concentration for liquid wastes is 2,742 ppm + 116 percent, and the mean for drummed wastes is 106,283 ppm + 170 percent. The mean total concentration (5,262 ppm + 743 percent) of hazardous constituents in liquids estimated from the ISDB quantity- weighted data is fairly consistent with the mean total concentration (5,906 ppm + 344 percent) for liquids from the SI data. However when the ISDB data are unweighted, the mean for liquids is very much higher (i.e., 157,538 ppm + 159 percent). 4.2.2 Solids Summary statistics for all available solid waste samples are also presented in Table 18. The SI reports provided a total of 39 solid waste samples; 6 solid waste samples were obtained from the RI reports; and the PEDCo study included 44 samples of mining waste solids (tailings). The total concentrations of hazardous constituents for samples taken at the point of disposal differ greatly for each of the three data sources. Estimates derived from the PEDCo data are considerably lower than estimates obtained from the RI data which are in turn much lower than estimates obtained from the SI data. Similar to liquid wastes, the median total concentrations of hazardous constituents in solid wastes tend to be considerably lower than their means. Also, the standard deviations exceed the mean total concentrations. 93 ------- Only one Industrial waste stream sample was available for solid wastes; this sample had a total hazardous constituent concentration of 4 ppm. The total concentration of hazardous constituents in solid wastes appears to be considerably higher than that for liquid wastes at the point of disposal. On the other hand, a comparison of the solid waste samples with the drummed waste samples in Table 18 shows that the mean total concentrations for these two wastes approach the same order of magnitude. For example, the SI solid waste data have a mean total concentration of 113,368 ppm + 112 percent, and the SI drum data have a mean total concentration of 84,261 ppm + 236 percent. 4.2.3 Sludges A total of 135 sludge samples taken at the point of disposal are available from the SI and RI reports and the PEDCo study. The ISDB contained 89 sludge/slurry samples taken at the point of generation. Table 18 summarizes these data1. The estimates derived with the PEDCo data for sludge wastes are much lower than the estimates derived from the SI or the RI data. Relative to the SI and RI data, the PEDCo data also yielded lower total concentration estimates for liquids and solids. For the SI, RI, and PEDCo data, the mean total concentrations of hazardous constituents in sludge wastes are much higher than the medians. Standard deviations are also high relative to their means. 94 ------- The above findings are similar for the ISDB sludge data. Mean total concentrations are greater than their corresponding medians, and standard deviations are high relative to their means— particularly for the quantity-weighted data. The total concentration of hazardous constituents in sludges appears to be greater than that in liquids, but less than that in solids and in drums. 4.2.4 Summary of Findings Among the three types of wastes—liquids, solids, and sludges— the mean total concentration of hazardous constituents in wastes at disposal sites tends to be greatest in solids, somewhat lower in sludges, and lowest in liquids. The mean total concentration in drummed wastes appears to be similar to that of solid wastes. According to the information displayed in Table 18, there is little difference between the mean total concentrations in liquid and sludge waste samples at the point of generation. For all categories of hazardous wastes shown in Table 18, the mean total concentrations are greater than their medians, and standard deviations tend to be high and exceed their means. Estimates of the mean total concentration of hazardous constituents in liquids, solids, and sludges vary according to the data source considered. The SI data show that the mean total concentration of hazardous constituents is greatest for solids, then for drums, then for sludges; it is the lowest for liquids. The 95 ------- RI data reveal that the mean total concentrations are greatest for drums, then for solids, then for sludges; liquids have the lowest concentrations. Finally, the PEDCo data show that the mean total concentrations are greatest in sludges, then solids, and lowest in liquids. The PEDCo data do not include drummed wastes. In general, the waste composition data show that the total concentrations of hazardous constituents in liquid wastes are less than the total concentrations in drummed wastes. Similarly, the total concentrations of hazardous constituents in sludges tend to be lower than the total concentrations in drummed wastes. Finally, the results are indeterminate with respect to whether the total concentrations are greater in solid wastes or in drummed wastes. 4.3 Statistical Pairwise Comparisons of Median Total Concentrations In this section nonparametric statistical methods are used to test whether there are differences in the median total concentration of hazardous constituents among various waste types available from the SI and RI data. Among all the data sources reviewed, the SI and RI reports tend to provide an internally consistent set of waste composition data for wastes at the point of disposal. The data extracted from these two sources are for Superfund sites, and the laboratory analyses of the samples generally included many CERCLA hazardous constituents. Moreover, both sources provide good cross-sections of the data by waste management method. 96 ------- The nonparametrie tests applied in this section avoid making assumptions (e.g., normal populations) that may be unrealistic for the data. As evidenced from Tables 6 and 7, the large departures of the mean and median values for drummed wastes are strong indicators that the distribution is skewed (nonnormal).* Since nonparametric tests assume no shape for the population distribution, the tests are valid for both normal and skewed populations. In addition, the techniques are well suited for small sample sizes (less than 30). A useful nonparametric alternative to the parametric t-test is the Wilcoxon rank sum median test.** This nonparametric test has been applied to data extracted from the SI and RI reports. For purposes of this analysis, the samples are assumed to be random and independent. (For reasons discussed previously, information is not available to assess the validity of this assumption.)*** All samples suspected of being dependent (e.g., samples from different management units containing the same waste stream at a single site) *In fact, neither the raw data nor the log-transformed drum data passed goodness-of-fit tests (Kolmogorov). (A discussion of goodness-of-fit tests is found in Haan, 1979, pp. 174-178.) Consequently, the population from which the drum samples are drawn is neither normal nor log-normal. However, upon checking the sample data for the other wastes identified in Tables 6 and 7, there is evidence based upon the chi-square goodness-of-fit test that some samples (e.g., waste piles, mine tailings, surface impoundment liquid) are from normal or log-normal populations. **This test is also referred to as the Mann-Whitney test, or U-test, in the literature. See Freund, 1971, p. 347. ***While nonparametric tests for randomness are available, the technique requires information about the order or sequence in which the observations were obtained (see Siegel, 1956, p. 52). Such information is not available for these data. 97 ------- have been eliminated from the analysis. Tables 19 and 20, respectively, present the test results for sample data from the SI and RI reports. (For the sample sizes tested, the Wilcoxon rank sum median test is based on an approximation of the normal distribution. Because of this approximation to a normal distribution, a two-tailed statistical test is applied to the z-values calculated from the Wilcoxon rank sum median test to determine if there are differences in the median concentrations. All analyses are conducted at the 0.05 level of significance.) The results in both Tables 19 and 20 generally show that there are significant differences between the median total concentrations of hazardous constituents in wastes contained in different types of management units. From Table 19, for example, drums and waste piles have statistically different median total concentrations of hazardous constituents. However, when waste types are similar—such as tank sludge and surface impoundment sludge in Table 20—the median total concentrations are not statistically different. Both Tables 19 and 20 further indicate that there are no significant differences between the median total concentration of hazardous constituents in surface impoundment sludge and drums. This analytical result is also the same for tank sludge and drums (Table 20). In summary, the analytical results indicate that there are differences in the median total concentration of hazardous constituents in hazardous wastes present in different types of 98 ------- TABLE 19 PAIRWISE COMPARISONS OF THE MEDIAN TOTAL CONCENTRATIONS OF HAZARDOUS CONSTITUENTS PRESENT IN HAZARDOUS WASTES AT DISPOSAL SITES (SITE INSPECTION REPORTS) Wilcoxon Drums (n=58) Surface Impoundment Sludge (n=29) Surface Impoundment Liquid (n=22) Mine Tailings (n=19) Waste Piles (n-11) Rank Sum Median Test Surface Impoundment Drums Sludge (n=58) (n=29) -0.34* -4.79 -4.06 2.55 2.91 3.29 3.39 (Normal Approximation z) Surface Impoundment Mine Waste Liquid Tailings Piles (n=22) (n=19) (n=ll) 4.75 — 4.30 3.06 ^Indicates that medians are not significantly different at the 0.05 level of significance. Note: z is a random variable with an approximate normal distribution, and n is the sample size. 99 ------- TABLE 20 PAIRWISE COMPARISONS OF THE MEDIAN TOTAL CONCENTRATIONS OF HAZARDOUS CONSTITUENTS PRESENT IN HAZARDOUS WASTES AT DISPOSAL SITES (REMEDIAL INVESTIGATION REPORTS) Wilcoxon Rank Drums (n=26) Tank Liquid (n-45) Tank Sludge (n-8) Surface Impoundment Sludge (n-8) Sum Median Test (Normal Approximation z) Surface Impoundment Drums Tank Liquid Tank Sludge Sludge (n=26) (n=45) (n=8) (n=8) __ 3.90 -0.95* 2.12 -0.71* 2.02 -0.16* *Indicates that medians are not significantly different at the 0.05 level of significance. Note: z is a random variable with an approximate normal distribution, and n is the sample size. 100 ------- waste management units. Specifically, the available data reveal that at a 95 percent confidence level there are differences in median total concentrations between: 1) surface impoundment liquid and drums; 2) mine tailings and drums; 3) waste piles and drums; 4) surface impoundment sludge and surface impoundment liquid; 5) surface impoundment sludge and mine tailings; 6) surface impoundment sludge and waste piles; 7) surface impoundment liquid and mine tailings; 8) surface impoundment liquid and waste piles; 9) mine tailings and waste piles; 10) drums and tank liquid; 11) tank liquid and tank sludge; and 12) tank liquid and surface impoundment sludge. Consequently, the same quantities of wastes present in different types of waste management units (e.g., waste piles and drums) are not likely to contain the same quantities of hazardous constituents.* Information is not available to determine how much greater the total concentrations of hazardous constituents are in one waste relative to other wastes. The analytical results shown in Tables 19 and 20 also indicate that at a 95 percent confidence level there are no statistical differences in the median total concentrations between: 1) drums and surface impoundment sludge; 2) drums and tank sludge; and 3) surface impoundment sludge and tank sludge. *Again, such a generalization is not true if the samples used in the statistical analysis are not actually random. 101 ------- 5.0 SUMMARY AND CONCLUSIONS Data on the concentration of hazardous constituents present in hazardous wastes at disposal sites are extremely limited. The existing data bases are program-related and were developed to meet particular objectives of their programs. As a result, the available data are not entirely consistent with the data requirements of this study. None of the data bases reviewed offers a comprehensive set of waste sample information. There are several limitations associated with the existing data bases which severely affect the extent to which the waste composition data in the data bases can be used for analyses in the current study. Among the more serious limitations are: • Most of the data available in the various data bases pertain to environmental samples (e.g., ground water samples), rather than to waste samples. • The analytical results in the data bases are based upon sampling and analysis programs that varied considerably both within and among the various programs. The number of CERCLA hazardous constituents for which samples were analyzed ranges from a couple in some data bases to several hundreds in other data bases. • The various data bases were developed to address specific issues, and the data in them may, consequently, be nonrepresentative of the wastes typically found at disposal sites. • Some of the data bases contain too few samples for specific waste management methods and/or contain samples for only a single type of waste management method. • Several data bases contain incomplete sample results; only the organic or inorganic portion of the waste analysis is available, not both portions. • Many of the data bases may contain data that are not independent, random samples. 103 ------- The primary concerns with the existing waste composition data consequently pertain to whether the data are representative of the wastes at hazardous waste sites and to whether the available data adequately characterize those wastes. The representativeness of the data is affected by: 1) whether the data are based on random samples; 2) whether the data pertain to all, or just a limited subset of, wastes present at hazardous waste sites; 3) whether the data pertain to all, or just a limited subset of, the waste management methods used at hazardous waste sites; 4) whether the data pertain to all types of hazardous waste sites, or just to NFL sites or non-NPL sites; and 5) whether the data are based on a sufficient number of samples. Additional factors which affect how adequately the data characterize the wastes include: 1) the number of CERCLA hazardous constituents for which analyses were performed, and 2) the completeness of the available data. For reasons discussed above, a large, but indeterminate, portion of the existing waste composition data may not be representative of, or adequately characterize, wastes present at disposal sites. Consequently, the findings noted below must be viewed in light of this limitation. A review of the available waste composition data revealed several major findings. First, the frequency distributions developed from the individual sources of drum data (NEIC, SI, and RI) are unimodal. For each source, approximately 75 percent of the samples have total concentrations of hazardous constituents that are 104 ------- less than 100,000 ppm. For these data sources, 35 to 50 percent of the samples have total concentrations less than 10,000 ppm. Twenty-one percent of the 580 NEIC samples have total concentrations less than 1,000 ppm. Similarly, 24 percent of the SI drum samples and 15 percent of the RI drum samples have total concentrations less than 1,000 ppm. Generally only 5 percent (or less) of the samples for each data source have total concentrations exceeding 500,000 ppm. Second, for most waste management practices analyzed, median total concentrations of hazardous constituents tend to be much lower than their corresponding means. This indicates that a disproportionate number (i.e., greater than 50 percent) of the sample values have total concentrations less than the mean value. Moreover, the standard deviations tend to be high and to generally exceed the mean (at times they are as great as 300 percent of the mean). This indicates a very high variability in the total concentration of hazardous constituents across the waste samples examined. Third, the available data indicate that the total quantity of hazardous constituents present in wastes at disposal sites tend to vary by type of waste management unit. Based on the available data, four approximate orderings have been developed and are presented in Section 4.1.6. The rankings are based on the mean and median total concentrations of hazardous constituents present in wastes within different types of waste mangement units. The rankings are also 105 ------- based on whether the waste composition data were obtained from a point of disposal or a point of generation. As noted in Section 4.1.6, the rankings are primarily subjective and are not generally based on statistical analysis. The ordering of waste management units (by total hazardous constituent concentrations) varies across the four rankings. The only generalization possible is that drummed wastes are usually at the top of the rankings (though sometimes in conjunction with other waste management units) and surface impoundment liquids are usually at the bottom. Due to data limitations, several waste management units could not even be included in the rankings (e.g., landfills, landfarms, mine tailings piles). Limitations associated with the rankings are discussed in Section 4.1.6. Fourth, a comparison of liquid, solid, and sludge wastes at disposal sites revealed that the mean total concentration of hazardous constituents in liquids tends to be much lower than the means for solids and sludges, and that the means for sludges are somewhat lower than those for solids. The mean total concentration in drummed wastes also appears to be similar to that of solids. The rankings of these mean total concentrations were, however, found to vary across the different data sources. Finally, nonparametric statistical analyses of the SI and RI data indicated that, at least for those data, there are differences in the total concentration of hazardous constituents present in 106 ------- wastes in different waste management units. For example, differences in the median total concentration of hazardous constituents were found to be significant at a 95 percent confidence level for wastes in: 1) surface impoundment liquids and drums; 2) mine tailings and drums; and 3) waste piles and drums. Other differences are cited in Section 4.3 of the report. Alternatively, the statistical analysis revealed some similarities among the waste types. For example, there were no statistically significant differences at a 95 percent confidence level between the median total concentrations of hazardous constituents for: 1) drummed wastes and surface impoundment sludge; 2) drummed wastes and tank sludge; and 3) surface impoundment sludge and tank sludge. In conclusion, it cannot be determined whether the existing waste composition data are representative of wastes at hazardous wastes sites. Further, only a very limited set of the available data is amenable to statistical analysis. This analysis showed that there are potentially significant differences in the total concentration of hazardous constituents in wastes present in different waste management units at wastes sites. The analysis was performed by stratifying the waste composition data into waste management units. If the existing waste composition data are to be used in developing alternatives to the current HRS waste quantity factor, then these alternatives should reflect the differences in concentrations likely to exist for different waste management units. 107 ------- Finally, any generalizations based on the analytical findings of this report must necessarily be qualified. 108 ------- APPENDIX A PREVIOUS INDUSTRY STUDIES OF THE EPA OFFICE OF SOLID WASTE This appendix provides a compilation of the concentration data for hazardous constituents in industrial waste streams at the point of generation. These data have been extracted from industry studies prepared in the mid-1970s for the U.S. Environmental Protection Agency, Office of Solid Waste. Table A-l lists the specific industries included in the industry studies and indicates which industry study reports provided information on the concentration of hazardous constituents in waste streams. In the following paragraphs, a brief description of the sampling methods and empirical results is presented for each industry reporting such information. Of the 14 industry studies reviewed, 10 studies published sampling and analysis information for land-destined waste streams. The concentration data for the ten industries are reported directly below. 1. Textiles Industry. Wastewater treatment sludges were sampled and analyzed at 14 textile plants.* Two separate sets of samples were taken to analyze heavy metals and chlorinated organics. A total of 112 samples was collected. Of these samples, 56 were analyzed for 13 heavy metals, and 56 for chlorinated organics. Table A-2 contains the ranges of average concentration for 13 metals.* The mean *Specifically, the plants sampled included one wool scouring plant; one wool fabric dyeing and finishing plant; 5 woven fabric dyeing and finishing plants; 3 knit fabric dyeing and finishing plants; and 2 yarn and sock dyeing and finishing plants. 109 ------- TABLE A-l AVAILABILITY OF HAZARDOUS WASTE CONCENTRATION DATA FROM INDUSTRY STUDIES SIC Code 22 281 282 283 285 286, 2879, 2892 2911 2992 3111 33 3471 355 367 3691, 3692 Industry Textiles Inorganic Chemicals Plastic Materials and Synthetics Pharmaceutical Paint and Allied Products Organic Chemicals, Pesticides, and Explosives Petroleum Refining Petroleum Re-refining Leather Tanning and Finishing Metal Smelting and Refining Electroplating and Metal Finishing Special Machinery Manufacturing Electronic Components Manufacturing Storage and Primary Batteries Concentration Data in Report Reference (Yes/No) Versar (1976) Versar (1975a) Snell (1978) A.D. Little (1976) Wapora (1975) TRW (1976) Jacobs (1976) Swain (1977) SCS (1976) Calspan (1977) Battelle (1976a) Battelle (1976b) Wapora (1977b) Wapora (1977a) Versar (1975b) Yes No Yes No Yes No Yes Yes Yes Yes No Yes Yes Yes No 110 ------- TABLE A-2 CONCENTRATION OF HAZARDOUS CONSTITUENTS IN SELECTED WASTE STREAMS OF THE TEXTILES INDUSTRY Hazardous Waste Stream Hazardous Constituents Average Concentration* Minimum Mean (ppm) Maximum Waste water treatment sludges Concentrations **Total does not Source: Versar, Aluminum Arsenic Barium Cadmium Chromium Cobalt Copper Lead Mercury Molybdenum Nickel Strontium Zinc Total** were estimated from 56 include values reported Inc. 1976. Assessment 7.2 <0.1 <12 <0.7 <2.5 <2.8 18 <7 <0.01 <2 <3.7 <2.45 106 131.2 samples. as "less 3,700 <4.9 <65 <6.1 <475 41.6 <416 <63 <0.7 <87.4 <31.9 <33.2 1,522 5,263.6 than" . 12,800 <17 <170 <17 3,969 212 1,130 170 <1.9 <333 88.2 <170 7,791 26,160.2 of Industrial Hazardous Waste Practices, Textiles Industry, prepared for the U.S. Environmental Protection Agency, Publication SW-125c, Washington, DC, pp. C-6 through C-32. Ill ------- concentration of these 13 metals in the sludge is approximately 5,300 ppm. Similar, detailed information for chlorinated organics in the sludge was not provided in the study. However, a numerical average for total chlorinated organics was reported for each of the 6 plants generating sludges. These average concentrations ranged between 0.11 and 64.7 ppm. The average concentration of chlorinated organics for these 6 plants was 24.6 ppm. 2. Rubber and Plastics Industry. For this industry study, a spot sampling program was conducted. Sixty-two plants were visited; the documentation is unclear with regard to whether a single plant was visited more than once. Of the 48 waste samples obtained, 39 samples were analyzed for mercury, lead, cadmium, and chlorine. Results from the sampling analysis are contained in Table A-3. Other metals were also analyzed, but the results were only semi- quantitative. While the analytical protocols used in the sampling of waste streams were discussed in some detail, other information regarding the sampling program was insufficient to explicitly determine the number of samples analyzed for each waste stream identified in Table A-3. 3. Paint and Allied Products Industry. Solvent recovery operations, which were analyzed in this industry study, generate one basic waste stream—still bottoms and sludges. Eight samples of still bottoms were collected and analyzed for lead, chromium, and *Tables A-2 through A-15 follow the text of Appendix A. 112 ------- TABLE A-3 CONCENTRATION OF HAZARDOUS CONSTITUENTS IN SELECTED WASTE STREAMS OF THE RUBBER AND PLASTICS INDUSTRY Hazardous Waste Stream Hazardous Constituents Concentration* (ppm) 1. Rubber - Compounding room sweepings - Used reclaim oil - Warehouse sweepings - Dust collectors - Waste oils 2. Plastics** - Incinerator ash - Warehouse and plant sweepings Mercury Lead Cadmium Chlorine Mercury Lead Cadmium Chlorine Mercury Lead Cadmium Chlorine Mercury Lead Cadmium Mercury Lead Cadmium Mercury Lead Cadmium Mercury Lead Cadmium Chlorine 0.5 72 2.5 612 0.5 3.8 1.0 580 1.0 1.0 0.6 450 0.7 15 2.5 0.1 3.8 1.0 0.1-2.1 16-185 0.9-3.4 0.4-1.2 4-15 1.0-5 1,750 113 ------- TABLE A-3 (Concluded) Hazardous Waste Stream Hazardous Constituents Concentration* (ppm) 3. Plastics** - Sludge Mercury Lead Cadmium Zinc - Spent alumina Mercury Lead Cadmium - Floor wastes Mercury Lead Cadmium - Waste nylon salt Mercury Lead Cadmium 0.2-1.8 1.5-5.0 0.1-0.4 160*** 0.4 1.5 0.2 0.4 1.0 0.1 0.1 1.0 0.1 *The number of samples analyzed for each waste stream was not reported. **In instances where several production processes generate the same waste stream, a range of the contaminant's concentration is given. ***Chemical Assay. Source: Foster D. Snell, Inc. 1978. Assessment of Industrial Hazardous Waste Practices, Rubber and Plastics Industry, Appendices, prepared for the U.S. Environmental Protection Agency, Publication SW-163c.4, Washington, DC, Tables B-l and B-2, pp. B-4 and B-5, respectively. 114 ------- zinc concentrations. The results of these analyses are presented in Table A-4. The industry study in which these findings are reported cautions that the concentrations have been estimated from grab samples. Consequently, the estimates are neither representative of the solvent reclaiming industry as a whole, nor the specific solvent recovery operations from which they were obtained. 4. Petroleum Refining Industry. The study of the petroleum refining industry analyzed samples of 17 waste streams from 16 refineries. Both grab and composite samples were obtained. The grab samples were taken from intermittent waste sources, while the composite samples were taken over a period of 4 to 5 hours from continuous waste sources. An analysis of the waste samples was conducted to determine the concentration of approximately 20 hazardous constituents (including phenols and cyanide, but primarily metals) in each waste stream. Results from this analysis are displayed in Table A-5. 5. Petroleum Re-refining Industry. Although this industry study did not have a waste sampling and analysis program, it does report some concentration data for metals. These data were obtained either from personal communication with industry representatives or from other EPA studies. The three major waste streams of the re-refining industry for which concentration data are reported include sludge (acid and caustic/silicate), spent clay, and process water. The concentration of various constituents for the three 115 ------- TABLE A-4 CONCENTRATION OF HAZARDOUS CONSTITUENTS IN SELECTED WASTE STREAMS OF SOLVENT RECLAMATION OPERATIONS Hazardous Waste Stream Still bottoms and sludges Hazardous Constituents Lead Chromium Zinc Total Concentration* (mg/1) Minimum 100 10 10 120 Mean 1,110 1,820 250 3,180 Median 850 170 130 — Maximum 3,700 730 990 5,420 *Coneentrations were estimated from 8 grab samples. Source: Wapora, Inc., 1975. Assessment of Industrial Hazardous Waste Practices; Paint and Allied Products Industry, Contract Solvent Reclaiming Operations, and Factory Application of Coatings, prepared for the U.S. Environmental Protection Agency, Office of Solid Waste Management Programs, Washington, DC, Table 77, p. 209. 116 ------- TABLE A-5 TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS IN SELECTED WASTE STREAMS OF THE PETROLEUM REFINING INDUSTRY Total Concentration* Waste Stream (mg/kg) Once-through cooling water sludge 518.0 Exchanger bundle cleaning sludge 861.6 Slop oil emulsion solids 1,584.5 Cooling tower sludge 1,617.2 API separator sludge 2,029.0 Dissolved air flotation float 1,023.3 Kerosene filter clays 2,105.0 Lube oil filter clays 140.8 Waste biosludge 380.4 Coke fines 2,078.4 Silt from storm water runoff 856.4 Leaded tank bottoms 10,896.7 Nonleaded product tank bottoms 496.0 Neutralized HF alkylation sludge 133.3 Crude bottom tanks 553.4 Spent lime from boiler feedwater treatment 83.6 Fluid catalytic cracker catalyst fines 1,038.0 *Calculated by summing the mean concentrations of all reported hazardous constituents in each waste stream. Approximately 20 hazardous constituents were analyzed; these were primarily metals, but phenols, cyanide, and benzo(a)pyrene were also included. Source: Jacobs Engineering Co., 1976. Assessment of Hazardous Waste Practices in the Petroleum Refining Industry, prepared for the U.S. Environmental Protection Agency. Publication SW-129c, Washington, DC, Appendix D, pp. 338-353. 117 ------- waste streams are shown in Table A-6. No information is provided with regard to the sampling techniques and methodologies used to prepare these estimates. (Note: Table 12 presents a more recent estimate of the concentration of hazardous constituents in these waste streams.) In a 1978 report for the National Science Foundation, Liroff et al., published their findings on the concentration of constituents in acid sludge samples taken from the petroleum re-refining industry. Their estimates, compiled from five different sources, are displayed in Table A-7. 6. Leather Tanning and Finishing Industry. In order to obtain representative solid waste samples from leather tanning and finishing plants, a detailed field sampling program was undertaken by this industry study. Protocols used in the sampling and analysis were reported. Waste samples were collected from 28 of the 41 tanneries visited. A total of 156 samples was analyzed for various organic and inorganic constituents. Table A-8 presents the concentrations recorded for chromium, copper, lead, and zinc. Although analyses were conducted for other heavy metals, pesticides, and phenols, none were found at potentially hazardous concentrations according to the study. 7. Metal Smelting and Refining Industry. The three general types of waste streams sampled for the metal smelting and refining industry are slag, sludge, and dust. Table A-9 displays the typical 118 ------- TABLE A-6 CONCENTRATION OF HAZARDOUS CONSTITUENTS IN SELECTED WASTE STREAMS OF THE PETROLEUM RE-REFINING INDUSTRY Hazardous Waste Stream 1. Sludge 2. Spent clay 3 . Untreated process water Hazardous Constituents Copper Lead Silver Zinc Barium** Chromium Boron Nickel Tin** Cadmium Molybdenum Arsenic Beryllium*** Cobalt Strontium Vanadium Barium Chromium Copper Lead Nickel Tin Zinc Zinc Copper Barium Nickel Chromium Tin Lead Boron Vanadium Molybdenum Cadmium Titanium Mercury Concentration* (ppm) Minimum 40 1,000 0 200 400 18 10 1 30 NA**** NA NA 0.1 NA NA NA Mean 80 14,625 4 1,475 813 72 30 12 45 9 18 45 1,500 0.8 2.7 18 4,200 1,300 1,200 39,500 60 10 4,300 3 ND*** ND ND ND ND 4 2 ND ND ND ND NA Median 44 15,000 1 1,800 740 39 29 5 35 NA NA NA 1,500 NA NA NA ** Maximum 190 27,500 14 2,100 1,300 190 50 30 70 NA NA NA 3,000 NA NA NA 119 ------- TABLE A-6 (Concluded) FOOTNOTES Concentrations for sludge were estimated from 4 samples. Only one sample was analyzed for spent clay and process water. Excludes concentrations of sodium and magnesium which were reported. **Estimates based on 3 samples. ***Estimates based on 2 samples. ****NA: Not available. *****ND: Not detected. Source: Swain, J.W., Jr., et al., 1977. Assessment of Industrial Hazardous Waste Management Petroleum Re-refining Industry, prepared for the U.S. Environmental Protection Agency, Publication SW-144c, Washington, DC, Table 9, p. 55; Table 11, p. 57; Table 15, p. 62; and Table 17, p. 69. 120 ------- TABLE A-7 CONCENTRATION OF HAZARDOUS CONSTITUENTS FOR PETROLEUM RE-REFINING: SUMMARY OF ANALYSES OF ACID SLUDGE SAMPLES Hazardous Constituents Copper Lead Nickel Zinc Barium Chromium Boron Aluminum Carbon PCB Concentrations (ppm) Study 1 40 20,000 30 2,100 1,300 50 50 40 NR NR Study 2 190 10,000 8 2,100 740 28 18 190 NR NR Study 3 NR* 5,700-28,900 NR NR NR NR NR NR 30-49 NR Study 4 110 NR 200 2,400 NR 75 NR NR NR NR Study 5 150 16,000 <0.02 2,250 120 42 NR NR NR 0.0046 *NR: Not reported. Note: Study 1: Weinstein, N. J., August 1974. Waste Oil Recycling and Disposal, EPA-670/2-74-052, Recon Systems, Inc. Study 2: U.S. Environmental Protection Agency, April 1974. Waste Oil Study, Report to Congress, Washington, DC. Study 3: Esso Research and Engineering Co., October 1972. Research of Oily Wastes. San Diego area for U.S. Navy, Study 4: Putscher, R. F., April 1962. "Separation and Characteristics of Acid Sludge," Armour Research Foundation, ARF-3859-3, Illinois Institute of Technology. Study 5: Unpublished 1977 Study by Teknekron, Inc., Berkeley, CA. Source: Liroff, S. D., et al., March 1978. Management of Environmental Risk; A Limited Integrated Assessment of the Waste Oil Rerefining Industry, prepared for the National Science Foundation by Teknekron, Inc., Berkeley, CA, Table IV-5, pp. IV-13. 121 ------- TABLE A-8 CONCENTRATION OF HAZARDOUS CONSTITUENTS IN SELECTED WASTE STREAMS OF THE LEATHER TANNING AND FINISHING INDUSTRY Hazardous Waste Stream No. of Hazardous Samples Constituents Minimum Concentration (mg/kg) Mean Maximum 1. Chrome (blue) trimmings & shavings 2. Chrome fleshings 10 Chromium 2,200 7,600 21,000 Chromium 4,000 3. 4. 5. 6. 7. 8. Unfinished chrome 9 leather trim Buffing dust 12 Finishing residues 16 Finished leather 4 trim Sewer screenings 17 Wastewater treatment sludges 27 Source: SCS Engineers, Inc., 1976 Waste Practices — Leather Chromium Copper Lead Zinc Chromium Copper Lead Zinc Chromium Copper Lead Zinc Chromium Lead Chromium Lead Zinc Chromium Copper Lead Zinc 4,600 2.3 2.5 9.1 19 29 2 — 0.45 0.35 2.5 14 1,600 100 0.27 2 35 0.33 0.12 0.75 1.2 16,900 90 120 60 5,700 960 150 160 3,500 40 8,400 150 14,800 1,000 2,200 30 60 3,700 370 60 50 37,000 468 476 156 22,000 1,900 924 — 12,000 208 69,200 876 41,000 3,300 14,000 110 128 19,400 8,400 240 147 Assessment of Industrial Hazardous Tanning and Finishing Industry, prepared for the U.S. Environmental Protection Agency, Publication SW-131c, Washington, DC, Table 4, p. 14, 122 ------- TABLE A-9 CONCENTRATION OF HAZARDOUS CONSTITUENTS IN SELECTED WASTE STREAMS OF THE METAL SMELTING AND REFINING INDUSTRY Average Concentration Hazardous Waste Stream Type of Smelter or Refiner 1. Primary Copper 2. Primary Lead 3. Primary Zinc 4. Primary Aluminum Hazardous Constituents Cadmium Chromium Copper Mercury Nickel Lead Antimony S elenium Zinc Cadmium Chromium Copper Mercury Lead Antimony Zinc Cadmium Chromium Copper Mercury Lead Selenium Zinc Fluorine Cyanide Copper Lead Slag 10 110 7,430 0.8 18 143 250 27 4,040 166 166 1,720 — 31,100 59 72,300 — — — — — — — — — — — Sludge 520 50 279,400 0.8 110 8,000 500 30 27,900 6,900 27 5,820 180 143,500 924 79,900 2.1 0.11 6.6 0.06 40 0.17 575 140,000 — — — Dust 520 88 37,260 13 90 1,320 143 33 2,250 — — — — — — — — — — — — — — 28,000 — 10,600 4,600 by (ppm) Potliners & Potroom Skimmings __ — — — — — — — — — — — — — — — — — — — — — — 255,000 1,050 — — 123 ------- TABLE A-9 (Continued) Average Concentration by Hazardous Waste Stream (ppm) Type of Smelter or Refiner 5. Primary Antimony 6. Primary Mercury 7. Primary Titanium 8. Primary Tungsten 9. Primary Tin Hazardous Constituents Arsenic Lead Copper Zinc Nickel Antimony Chromium Cadmium Lead Copper Zinc Nickel Mercury Antimony Cadmium Chromium Vanadium Chromium Zirconium Titanium Chlorine Arsenic Lead Copper Zinc Tin Lead Slag 16 66 50 500 — 18,000 — — — — — — — — — — — — — — — — — — — — 10,000 100 Sludge __ 5 50 2 5 27,000 32 1.0 — — — — — — — — 25,780 11,630 34,770 104,400 187,000 3.5 137 137 26 — — Calcine Digestion Residue Residue __ — — — — — — — 150 430 80 2,600 200 175 — 430 — — — — — — — — — — — — __ — — — — — — — — — — — — — — — — — — — — — — 1.0 377 8.9 — — 124 ------- TABLE A-9 (Continued) Average Concentration Hazardous Waste Stream Type of Smelter or Refiner 10. Primary Cadmium 11. Secondary Copper 12. Secondary Lead 13. Secondary Aluminum 14. Iron and Steel Hazardous Constituents Cadmium Chromium Copper Lead Zinc Cadmium Chromium Copper Nickel Lead Antimony Tin Zinc Cadmium Chromium Copper Nickel Lead Antimony Tin Zinc Chromium Copper Nickel Lead Zinc Chromium Copper Nickel Lead Zinc Slag Sludge Dust __ — — — — 5 20 16,000 260 3,000 100 7,000 10,000 1.7 340 230 3,700 100 19,300 53, 700 1, 6,800 670 60 310 1, 10 300 240 6, 1,780 1, 90 40 180 2, 40 6, — __ — — — — — — — — — — — — — — — — — — — — — — — — 30 20 5 — 000 100 — — 25 20 250 — — 140 500 070 590 720 840 180 180 510 10,900 380 53,100 by (ppm) Plant Residue 280 24 1,150 215,000 39,000 — — — — — — — — — — — — — — — __ — — — — — — — — — 125 ------- TABLE A-9 (Concluded) Type of Hazardous Average Concentration by Hazardous Waste Stream (ppm) Smelter or Refiner 15. Iron and Steel Foundries 16. Ferroalloys Constituents Cadmium Copper Chromium Nickel Lead Zinc Phenol Cobalt Chromium Copper Nickel Lead Zinc Slag 1.0 33.5 77.6 6.7 10 24.6 — 21 832 80 370 8 41 Sludge 2.1 147 49 3.2 132 355 — 40 551 39 1,025 7,500 11,280 Dust 0.9 128 75 47 112 143 — — 55 1,237 1,470 2,167 100 5,533 Sand __ 8.3 4.8 28.1 53.6 6.0 1.1 — — — — — ™«« Note: Dashes indicate not available or not applicable. Source: Calspan Corporation, 1977. Assessment of Industrial Hazardous Waste Practices in the Metal Smelting and Refining Industry, prepared for the U.S. Environmental Protection Agency, Publication SW-145c, Washington, DC, Volume II, Tables, 3, 18, 30, 45, 55, 67, 74, 81, 89, 98, 101, 112, and 120; and Volume III, Tables 5, 12, and 17. 126 ------- concentration of various hazardous constituents found in these waste streams. The documentation for this industry study states that— with the exception of the iron and steel industry—53 industrial plants were visited and sampled only once. In the case of the iron and steel industry, a program which allowed the acquisition and analysis of weekly composited samples was implemented. This program enabled a composite of four weekly samples to be analyzed for the iron and steel industry. The documentation, however, is unclear with respect to the number of samples used to obtain the results presented in Table A-9. 8. Electroplating and Metal Finishing Industry. While there was no waste sampling and analysis program conducted for the job shops study, a semi-quantitative optical emission spectographic analysis of a sample of dewatered sludge was performed in the captive shops study. Results from this analysis are displayed in Table A-10. In addition, the report included responses by industry to a questionnaire requesting information on the concentrations of hazardous constituents in the industry's waste streams. The survey results are contained in Table A-ll. No information is given about respondents reporting no detections. 9. Special Machinery Manufacturing Industries. Table A-12 is derived from an analysis of samples collected from four waste streams at foundry operations in the iron and steel industry. The data presented in this table are derived from combined samples of two 127 ------- TABLE A-10 CONCENTRATION OF HAZARDOUS CONSTITUENTS IN SELECTED WASTE STREAMS OF THE ELECTROPLATING AND METAL FINISHING INDUSTRIES Hazardous Waste Stream Dewatered Sludge Hazardous Constituents Aluminum Zinc Chromium Lead Nickel Titanium Boron Barium Molybdenum Copper Tin Vanadium Cadmium Zirconium Cobalt Strontium Beryllium Average Concentration* (wt %) 2-4 2-4 1 1 0.6 0.3 0.2 0.03 0.005 0.03 0.01 0.005 <0.02 0.1 0.01 0.01 < 0.0001 *Concentrations were estimated from 2 samples. Source: Battelle Columbus Laboratories, January 1976. Final Report on Assessment of Industrial Hazardous Waste Practices— Electroplating and Metal Finishing Industries—Captive Shops, prepared for the U.S. Environmental Protection Agency, Hazardous Waste Management Division, Washington, DC, pp. B-20 and B-21. 128 ------- TABLE A-ll SURVEY RESULTS FOR CONCENTRATION OF HAZARDOUS CONSTITUENTS IN SELECTED WASTE STREAMS OF THE ELECTROPLATING AND METAL FINISHING INDUSTRIES Hazardous Waste Stream 1. Dust 2 . Sludge Hazardous Constituents Chromium Aluminum Cadmium Chromium Copper Cyanide Lead Mercury Nickel Selenium Zinc Phosphates Sulfates Tin Titanium No. of Respondents 1 2 2 16 17 2 5 1 12 1 15 1 2 1 1 Concentration (wt/%) Minimum 0.032 0.0005 0.3 ppm 0.3 ppm 0.001 0.01 2.5 0.01 0.0001 Mean 1 5.7 0.0441 11.7 8.9 2.5 2.5 2.5 1 4.4 69 10-25 15 1-10 Median 5.7 0.0441 2 3.9 2.5 0.24 1.3 1 Maximum 11.1 0.0876 50 50 5 10 10.5 31 3. Finishing Chloride Solvent 30-60 Source: Battelle Columbus Laboratories, January 1976. Final Report on Assessment of Industrial Hazardous Waste Practices—Electroplating and Metal Finishing Industries—Captive Shops, prepared for the u7s. Environmental Protection Agency, Hazardous Waste Management Division, Washington, DC, Table D-l. pp. D-l through D-20. 129 ------- TABLE A-12 CONCENTRATION OF HAZARDOUS CONSTITUENTS IN SELECTED WASTE STREAMS OF SPECIAL MACHINERY MANUFACTURING INDUSTRIES Hazardous Waste Stream Hazardous Constituents Concentration (ppm) 1. Spent foundry core sand 2. Settled dry sand sludge 3. Dryer shake-out sand 4. Settled bond sludge Cadmium Chromium Copper Lead Zinc Cadmium Chromium Copper Lead Zinc Cadmium Chromium Copper Lead Zinc Cadmium Chromium Copper Lead Zinc Nickel 0.5 10.0 23.0 28.0 43.0 <0.2 3.0 6.0 19.0 7.0 <0.2 4.0 6.0 3.0 71.0 0.5 32.0 880.0 51.0 85.0 270.0 Source: Wapora, Inc., March 1977. Assessment of Industrial Hazardous Waste Practice Special Machinery Manufacturing Industries, prepared for the U.S. Environmental Protection Agency, Publication SW-141c, Washington, DC, Table III-ll, p. 113. 130 ------- foundries on the same premise. In contrast to a 1975 study prepared for EPA, the concentrations of heavy metals shown in Table A-12 are low. Analytical results from the earlier study are displayed in Table A-13. The authors of the 1977 industry study for special machinery manufacturing state that the existing data base is too sparse to explain the differences between the two sets of estimates. Table A-14 displays estimates of concentrations for hazardous constituents found in five major waste streams of various special machinery manufacturing industries. Machine shops and food products machinery are among the manufacturers studied in this report. 10. Electronic Components Manufacturing Industry. For this industry study, 16 process waste samples were collected from the manufacturing plants surveyed. All the samples are grab samples, and subsequently represent characteristics of the waste only at the time when the sample was taken. Table A-15 displays the sampling results for five waste streams. 131 ------- TABLE A-13 CONCENTRATION OF HAZARDOUS CONSTITUENTS IN SELECTED WASTE STREAMS OF SPECIAL MACHINERY MANUFACTURING INDUSTRIES IRON AND STEEL CASTING Hazardous Waste Stream Hazardous Constituents 2. Furnace emissions 3. Pouring and shakeout dust 4. Cleaning and finishing dust 5. Waste sand Chromium Copper Nickel Lead Zinc Chromium Copper Nickel Lead Zinc Cadmium Chromium Copper Nickel Lead Zinc Cadmium Chromium Copper Nickel Lead Zinc Phenol Concentration (ug/g) 1. Slag Chromium Copper Nickel Lead Zinc Cadmium 17-150 18-52 10 10-16 10-40 1 21-100 90-300 32-60 130-310 500-7,000 50 150 130 250 2.3 100-200 40-950 130-150 <10-840 20-210 2 3-18 6-26 4-200 <10-400 6-7 1.01-1.73 132 ------- TABLE A-13 (Concluded) Hazardous Hazardous Waste Stream Constituents Concentration (ug/g) 6. Sand reclamation dust Chromium 41 Copper 7 Nickel Lead 15 Zinc 30 Cadmium < 1 Source: Calspan Corporation, 1975. Assessment of Industrial Hazardous Waste Practices in the Metal Smelting and Refining Industry, Volume 3, Ferrous Smelting and Refining, prepared for the U.S. Environmental Protection Agency, Washington, DC. 133 ------- TABLE A-14 CONCENTRATION OF HAZARDOUS CONSTITUENTS IN SELECTED WASTE STREAMS OF THE SPECIAL MACHINERY MANUFACTURING INDUSTRIES Hazardous Waste Stream 1. Quench oil from heat treating* 2. Sludges from treating electroplating wastes* 3. Spent and fresh cutting oils** 4. Degreaser solvent*** 5. Paint sludge**** Hazardous Constituents Cadmium Chromium Copper Lead Zinc Cyanide Cadmium Chromium Copper Lead Zinc Nickel Cadmium Chromium Copper Lead Zinc Cadmium Chromium Copper Lead Zinc Cadmium Chromium Copper Lead Zinc Concentration (ppm) <0.1 <0.1 <0.1 <0.5 <0.1 <10 5 1,840 71,500 5.5 860 708 0.4 247 7,870 435 1,296 <0.1 0.7 2.0 1.3 3.3 1.5 140 45,300 2,275 7,608 134 ------- TABLE A-14 (Concluded) FOOTNOTES *Coneentrations for this waste stream were averaged over 4 plant streams. **Coneentrations for this waste stream were averaged over 6 samples. ***Concentrations were estimated from 2 samples. ****Concentrations were estimated from one grab sample. Source: Wapora, Inc., March 1977. Assessment of Industrial Hazardous Waste Practice Special Machinery Manufacturing Industries, prepared for the U.S. Environmental Protection Agency, Publication SW-141c, Washington, DC, Table 111-12, p. 119; Table 111-14, p. 127; Table 111-15, p. 134; Table 111-16, pp. 135 and 143. 135 ------- TABLE A-15 CONCENTRATION OF HAZARDOUS CONSTITUENTS IN SELECTED WASTE STREAMS OF THE ELECTRONIC COMPONENTS MANUFACTURING INDUSTRY Hazardous No. of Waste Stream Samples 1 2 3 4 Halogenated solvent 2 wastes . Nonhalogenated 3 solvent wastes Wastewater 3 treatment sludges . Oil-containing 2 wastes Hazardous Constituents Cadmium Chromium Copper Lead Zinc Nickel Fluorine Cadmium Chromium Copper Lead Zinc Nickel Cadmium Carbon Copper Lead Zinc Nickel Fluorine Cadmium Chromium Copper Lead Zinc Nickel Concentration (mg/kg) <0.04 0.06 0.62 74 3.6 1.4 — <0.02 0.04 0.20 7.2 153 0.08 36.3 347 3,287 29.6 13,197 1,853 — 0.16 2.3 1,285 229.1 36.8 2.4 136 ------- TABLE A-15 (Concluded) Hazardous No. of Hazardous Waste Stream Samples Constituents Concentration (mg/kg) 5. Paint wastes 2 Cadmium Chromium Copper Lead Zinc Nickel <0.06 216 18.8 380 998.3 2 Source: Wapora, Inc., 1977. Assessment of Industrial Hazardous Waste Practices; Electronic Components Manufacturing Industry, prepared for the U.S. Environmental Protection Agency, Publication SW-140c, Washington, DC, Tables III-ll, 111-12, III-13, III-14, and 111-15. 137 ------- APPENDIX B CURRENT EPA OFFICE OF SOLID WASTE INDUSTRY STUDIES DATA BASE This appendix contains statistical summaries of data on the concentration of hazardous constituents in hazardous waste streams of the organic chemicals industry. Specifically, the data were extracted and summarized from the Industry Studies Data Base (ISDB) which has been developed as part of the Industry Studies program in the EPA Office of Solid Waste (OSW). Table B-l lists the 11 industry segments of the organic chemicals industry contained in the data base. Descriptive statistics were derived for the total concentration of CERCLA hazardous constituents present in the residuals (i.e., waste streams) included in the ISDB. The residuals analyzed in this study are identified in Table B-2. As discussed in Section 3, 16 dilute acids and caustics for which there are CERCLA reportable quantities have been eliminated from the calculations to prevent inflated concentrations. These hazardous constituents are listed in Table B-3. In particular, this appendix contains 20 summary tables. All tables contain the following descriptive statistics: range of hazardous constituent concentrations (i.e., minimum and maximum concentrations); unweighted mean, median, and standard deviation of the concentration data; and quantity-weighted mean, median, and standard deviation of the concentration data. The quantity-weighted 139 ------- TABLE B-l INDUSTRY SEGMENTS OF THE INDUSTRY STUDIES DATA BASE Brominated Organics Carbamate Pesticides Chlorinated Aliphatics Chlorinated Aromatics Dyes and Figments Industrial Organics Miscellaneous Chlorinated Organics Organic Pesticides Organometalllc s Plastics and Resins Rubber Processing 140 ------- TABLE B-2 RESIDUAL CATEGORIES OF THE INDUSTRY STUDIES DATA BASE ANALYZED IN THIS STUDY Category ISDB Residuals Sludge/slurry • Precipitates or filtration residues • Sludges • Heavy ends Spent solvents • Spent solvents Solids • Treatment solids Liquids • Decantates or filtrates • Condensable light ends Untreated wastewater Untreated process wastewater 141 ------- TABLE B-3 CERCLA HAZARDOUS CONSTITUTENTS CONTAINED AS DILUTE ACIDS AND CAUSTICS IN THE INDUSTRY STUDIES DATA BASE Acetic Acid Adipic Acid Benzole Acid Butyric Acid Formic Acid Fumaric Acid Hydrochloric Acid Hydrofluoric Acid Isobutyric Acid Maleic Acid Nitric Acid Phosphoric Acid Potassium Hydroxide Propionic Acid Sodium hydroxide Sulfuric Acid 142 ------- mean, X,,, and standard deviation, SI^, were derived as follows: N E 1=1 xw N N z; cx±- xw)2 Ql N ZQi 1=1 where: X = Total concentration of hazardous constituents in the 1th residual. Q1 = Quantity of ic" residual generated. N = Number of residuals included in each analysis. The quantity-weighted median concentration was derived by rank ordering the residuals by their total concentration of hazardous constituents. The quantity associated with each of the rank ordered residuals was then summed according to the rank ordering. The median quantity-weighted concentration was defined as that concentration for which the aggregated residual quantity was one-half the total residual quantity. The statistical summaries presented in this appendix were prepared by type of residual and by waste management method used for the residual (e.g., solids sent to landfills, solids placed in waste piles). There are three separate analyses of the data: one considers 143 ------- only those residuals that are RCRA hazardous wastes (Tables B-4 through B-9); another considers only those residuals that are not RCRA hazardous wastes (Tables B-10 through B-16); and the third considers all residuals, i.e. RCRA and non-RCRA wastes (Tables B-17 through B-23). 144 ------- TABLE B-4 SUMMARY STATISTICS ON THE TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS, EXCLUDING RANGES, PRESENT WITHIN ALL RCRA HAZARDOUS WASTES FOR EACH RESIDUAL GROUP - BY MANAGEMENT PRACTICE MANAGEMENT PRACTICE CONTAINERS LANDFILLS SURFACE IMPS. TANKS ALL MGHT. 6RPS. MANAGEMENT PRACTICE CONTAINERS LANDFILLS SURFACE IMPS. TANKS ALL M6MT. 6RP9. KE91DIML UAItbUKT: »H<*U1U5» NO. OF MINIMUM MAXIMUM MEAN SAMPLES CONC. (PPM) CONC. (PPM) COHC. (PPM) 1 1 1 * a NO. OF HTO. SAMPLES 1 1 1 S 0 300000 300000 200382 200362 0.1 0.1 30 559000 0.1 559000 9TY -WEIGHTED MEAN CONC (PPM) 300000 200362 0.1 61368 70275 300000 200382 0.1 115818 134934 9TY-HEI6HTED MED. CONC (PPM) 300000 200382 0.1 30 30 MEDIAN CONC. (PPM) 300000 200382 0.1 10000 10000 QTY-HEIGHTED STD. OEV. 0 0 0 174399 173537 STANDARD DEVIATION 0 0 0 247796 206043 SOURCE! Seivnea Application* International Corporation,1986 ------- TABLE B-5 SUMMARY STATISTICS ON THE TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS, EXCLUDING RANGES, PRESENT WITHIN ALL RCRA HAZARDOUS WASTES FOR EACH RESIDUAL GROUP - BY MANAGEMENT PRACTICE MANAGEMENT PRACTICE -RESIDUAL =SLU06E/3LURRY- NO. OF SAMPLES MINIMUM CONC. (PPM) MAXIMUM CONC. (PPM) MEAN CONC. (PPM) MEDIAN CONC. IPPH) STANDARD DEVIATION CONTAINERS LANDFILLS PILES SURFACE IMPS. TANKS ALL MGMT. 6RPS. MANAGEMENT PRACTICE 13 20 1 3 13 43 NO. OF UTO. SAMPLES 3 0.2 271 • 31 0.03 0.03 960000 984621 271 1006 350000 984621 QTY-HEIGHTED MEAN CONC (PPM) 369403 243598 271 512 102293 215916 QTY-HEIGHTEO MED. CONC CPPM) 200000 145350 271 500 106 82733 QTY-HEIGHTED STD. DEV. 344691 307343 0 487 137247 294673 CONTAINERS LANDFILLS PILES SURFACE IMPS. TANKS ALL MGMT. GRPS. 11 18 1 3 13 40 449357 103617 271 541 87501 77823 500000 82733 271 500 106 82733 194104 107056 0 156 127673 123927 SOURCE: Science Application* International Corporation,1986 ------- TABLE B-6 SUMMARY STATISTICS ON THE TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS, EXCLUDING RANGES, PRESENT WITHIN ALL RCRA HAZARDOUS WASTES FOR EACH RESIDUAL GROUP - BY MANAGEMENT PRACTICE -RESIDUAL CATEGORY: »SPENT SOLVENTS- MANAGEMENT NO. OF MINIMUM MAXIMUM MEAN MEDIAN STANDARD PRACTICE SAMPLES CONC. (PPMI CONC. (PPMI CONC. (PPHI CONC. (PPM) DEVIATION TANKS 7 5 1000000 015715 990000 370647 ALL MGMT. GRPS. 7 5 1000000 015715 990000 3706*7 MANAGEMENT NO. Of flTT-HEIGHTED 9TY-HEIGHTEO QTY-HEIGHTEO PRACTICE HTD. SAMPLES MEAN CONC (PPM) MED. CONC (PPM) STD. DEV. TANKS 3 651373 750000 242930 ALL FGMT. GRPS. 3 6*1373 750000 242930 SOURCE: Sci«nc« Applications International Corporation.1906 ------- TABLE B-7 SUMMARY STATISTICS ON THE TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS, EXCLUDING RANGES, PRESENT WITHIN ALL RCRA HAZARDOUS WASTES FOR EACH RESIDUAL GROUP - BY MANAGEMENT PRACTICE —RESIDUAL CATEGORYt =UNTRT. HASTEHATER- MANAGEMENT NO. OF MINIMUM MAXIMUM MEAN MEDIAN STANDARD PRACTICE SAMPLES CONC. (PPMI CONC. (PPM) COHC. (PPMI CONC. (PPM) DEVIATION T; TANKS 11 1 165000 16489 61 49359 oo ALL MGMT. GUPS. 11 1 165000 16489 61 49359 MANAGEMENT NO. OF 4TY-UEIGHTED QTY-MEIGHTED QTY-MEIGHTEO PRACTICE HID. SAMPLES MEAN CONC (PPM) MED. CONC (PPM) STD. DEV. TANKS 10 118 6 1237 ALL MGMT. 6RPS. 10 118 6 1237 SOURCE: Scianc* Application* International Corporation,1986 ------- TABLE B-8 SUMMARY STATISTICS ON THE TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS, EXCLUDING RANGES, PRESENT WITHIN ALL RCRA HAZARDOUS WASTES FOR EACH RESIDUAL GROUP - BY MANAGEMENT PRACTICE MANAGEMENT PRACTICE NO. OF SAMPLES -KC31UUAL LAI til MINIMUM CONC. (PPM) UHT: -ALL LAI . i MAXIMUM CONC. (PPM) MEAN CONC. (PPM) MEDIAN CONC. (PPM) STANDARD DEVIATION CONTAINERS LANDFILLS PILES SURFACE IMPS. TANKS ALL MGMT. GRPS. MANAGEMENT PRACTICE 14 21 I 4 25 58 NO. OF WTO. SAMPLES 3 0.2 271 0.1 0.03 0.03 960000 984621 271 1006 1000000 1000000 364445 241540 271 384 304756 277136 QTY-HEIGHTED MEAN CONC (PPM) QTY-HEIGHTEO MED. CONC (PPM) 250000 150700 271 265 29938 145350 QTY-WEIGHTED STO. DEV. 331687 299709 0 473 399778 353496 CONTAINERS LANDFILLS PILES SURFACE IMPS. TANKS ALL nSMT. 6RPS. 12 19 1 4 21 SI 446951 103998 271 540 83442 78281 500000 82733 271 500 30 62733 193450 107017 0 157 169380 129076 SOURCE: Science Aeoliotform Int«rn«tion«l CorDeration,1986 ------- TABLE B-9 SUMMARY STATISTICS ON THE TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS, EXCLUDING RANGES, PRESENT WITHIN ALL RCRA HAZARDOUS WASTES FOR EACH RESIDUAL GROUP - BY MANAGEMENT PRACTICE Ln O MANAGEMENT PRACTICE CONTAINERS LANDFILLS PILES SURFACE IMPS. TANKS ALL MGMT. 6RPS. MANAGEMENT PRACTICE CONTAINERS LANDFILLS PILES SURFACE IMPS. TANKS ALL M6HT. GftPS. NO. OF MINIMUM MAXIMUM SAMPLES CONC. (PPMI CONC. IPPM) 14 21 1 4 36 69 NO. OF HTD. SAMPLES 12 19 1 4 31 61 3 0.2 271 0.1 0.03 0.03 960000 964621 271 1006 1000000 1000000 MEAN CONC. IPPM) 364445 241540 271 384 216675 235583 QTY-HEIGHTED QTY-HEIGHTED MEAN CONC (PPMI MED 446951 103998 271 540 518 3710 . CONC IPPM) 500000 82733 271 500 6 20 MEDIAN CONC. (PPM) 250000 150700 271 265 3065 10000 QTY-MEIGHTEO STD. DEV. 193450 107017 0 157 13129 32170 STANDARD DEVIATION 331687 299709 0 473 358365 338144 . SOURCEi Set area Application* International Corporation,1986 ------- TABLE B-10 SUMMARY STATISTICS ON THE TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS = EXCLUDING RANGES = PRESENT WITHIN ALL NON-RCRA HAZARDOUS WASTES FOR EACH RESIDUAL GROUP - BY MANAGEMENT PRACTICE MANAGEMENT PRACTICE SURFACE IMPS. TANKS ALL MGMT. 6RPS. MANAGEMENT PRACTICE SURFACE IMPS. TANKS ALL MGMT. 6RPS. NO. OF MINIMUM MAXIMUM SAMPLES CONC. (PPM) CONC. (PI 2 200 2000 17 0.01 900000 19 0.01 900000 NO. OF QTY-HEIGHTED HTO. SAMPLES MEAN CONC (PPM) 2 2754 12 34757 14 4452 MEAN PM) CONC. (PPM) 1500 186532 167055 QTY-MEIGHTEO MEO. CONC (PPM) 2800 72 2800 MEDIAN CONC. (PPM) 1500 3000 2800 QTY-HEIGHTED STO. DEV. 342 141946 33476 STANDARD DEVIATION 1838 280501 270818 SOURCE: Science Application* International Corporation,1986 ------- TABLE B-ll SUMMARY STATISTICS ON THE TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS = EXCLUDING RANGES = PRESENT WITHIN ALL NON-RCRA HAZARDOUS WASTES FOR EACH RESIDUAL GROUP - BY MANAGEMENT PRACTICE MANAGEMENT PRACTICE NO. OF SAMPLES RESIDUAL t MINIMUM CONC. (PPM) AlCbUHi: =3LUUb MAXIMUM CONC. (PPM) MEAN CONC. (PPM) MEDIAN CONC. (PPM) STANDARD DEVIATION Ln t-0 CONTAINERS LANDFARMING LANDFILLS PILES SURFACE IMPS. TANKS ALL MGMT. GRPS. MANAGEMENT PRACTICE 3 3 31 2 4 13 46 NO. OF MTD. SAMPLES 293 38 2 100 2 7 2 QTY-MEIGHTEO MEAN CONC (PPM) 350000 40 500000 10000 600 1000000 1000000 150232 39 58912 5050 170 337231 121357 QTY-MEIGHTED MED. CONC (PPM) 100402 40 150 5050 39 200000 235 100100 1 135520 7000 207 359191 246765 QTY-NEIGHTED STD. DEV. CONTAINERS LANDFARMIN6 LANDFILLS PILES SURFACE IMPS. TANKS ALL MGMT. GRPS. 3 3 22 2 3 13 36 8720 39 5413 9959 91 221380 10170 £93 38 4 10000 38 64000 10000 28187 1 34707 636 164 322665 51195 SOURCE: 3ci*nc« Aoolicatior« Intarmtforal Corporation.1986 ------- TABLE B-12 SUMMARY STATISTICS ON THE TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS = EXCLUDING RANGES = PRESENT WITHIN ALL NON-RCRA HAZARDOUS WASTES FOR EACH RESIDUAL GROUP - BY MANAGEMENT PRACTICE MANAGEMENT PRACTICE SURFACE IMPS. ALL MGMT. 6RP3. MANAGEMENT PRACTICE SURFACE IMPS. ALL MGMT. 6RP9. NO. OF MINI SAMPLES CONC 1 1 NO. OF HTD. SAMPLES 1 1 MUM MAXIMUM . (PPM) CONC. IP * 4 4 4 QTY-HEIGHTED MEAN CONC (PPM) 4 4 MEAN PM) CONC. IPPM) 4 4 4TY-MEIGHTED MEO. CONC (PPM) 4 4 MEDIAN CONC. (PPM) 4 4 QTY-UEIGHTED STO. OEV. 0 0 STANDARD DEVIATION 0 0 SOOOci: Sct*He£ Aoolicatiorw International CorDeration,1986 ------- TABLE B-13 SUMMARY STATISTICS ON THE TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS = EXCLUDING RANGES = PRESENT WITHIN ALL NON-RCRA HAZARDOUS WASTES FOR EACH RESIDUAL GROUP - BY MANAGEMENT PRACTICE MANAGEMENT PRACTICE LANDFILLS TANKS ALL MGMT. 6RPS. MANAGEMENT PRACTICE LANDFILLS TANKS ALL MGMT. 6RPS. NO. OF MINIMUM MAXIMUM MEAN SAMPLES CONC. (PPMI CONC. IPPHI CONC. (PPM) 1 Z 3 NO. OF MTD. SAMPLES 1 1 I 312650 312850 610000 990000 312850 990000 QTY-HEIGHTED MEAN CONC IPPMI 312850 610000 365912 312850 800000 637617 QTY-HEIGHTED MED. CONC (PPM) 312850 610000 312850 MEDIAN CONC. (PPM) 312850 800000 610000 QTY-HEIGHTEO STO. DEV. 0 0 113806 STANDARD DEVIATION 0 268701 339419 SOURCE« Scianca Application* International Corporation.1986 ------- TABLE B-14 SUMMARY STATISTICS ON THE TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS = EXCLUDING RANGES = PRESENT WITHIN ALL NON-RCRA HAZARDOUS WASTES FOR EACH RESIDUAL GROUP - BY MANAGEMENT PRACTICE MANAGEMENT PRACTICE —RESIDUAL CATEGORYl *UNTRT. HASTEMATER- NO. OF SAMPLES MINIMUM CONC. I PPM) MAXIMUM . CONC. (PPMI MEAN CONC. (PPMJ MEDIAN CONC. (PPM) STANDARD DEVIATION Ul Oi CONTAINERS LANDFILLS SURFACE IMPS. TANKS ALL MGMT. 6RPS. MANAGEMENT PRACTICE I I 6 26 35 NO. OF HTD. SAMPLES 4 1009 4 2 2 70000 1009 24000 55000 70000 QTY-MEIGHTED MEAN CONC (PPM) 35002 1009 6074 6701 6110 QTY-MEIGHTED MED. CONC I PPM) 35002 1009 1106 450 300 QTY-MEIGHTED STD. OEV. 49495 0 10668 12636 15611 CONTAINERS LANDFILLS SURFACE IMPS. TANKS ALL MGMT. 6RPS. 1 1 7 12 19 4 1009 9632 1924 6115 4 1009 2000 300 2000 0 0 9466 5616 9215 SOURCEi Science Aoolications International Corporatton,1986 ------- TABLE B-15 SUMMARY STATISTICS ON THE TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS = EXCLUDING RANGES = PRESENT WITHIN ALL NON-RCRA HAZARDOUS WASTES FOR EACH RESIDUAL GROUP - BY MANAGEMENT PRACTICE O\ MANAGEMENT PRACTICE CONTAINERS LANDFARMIN6 LANDFILLS PILES SURFACE IMPS. TANKS ALL MGMT. GRPS. MANAGEMENT PRACTICE CONTAINERS LANOFARMIN6 LANDFILLS PILES SURFACE IMPS. TANKS ALL M6MT. 6RPS. --KC9XUUAL LAICMJN NO. OF MINIMUM SAMPLES CONC. (PPM) 3 3 32 2 7 32 69 NO. OF MTO. SAMPLES 3 3 23 2 6 26 S3 293 38 2 100 2 0.01 0.01 MAXIMUM MEAN MEDIAN CONC. IPPM1 CONC. (PPM) CONC. (PPM) 350000 150232 40 39 500000 66648 10000 5050 2600 526 1000000 266096 1000000 154628 QTY-MEIGHTED 9TY- WEIGHTED MEAN CONC 6720 39 5979 9959 1963 104707 8322 (PPM) MED. CONC (PPM) 293 36 4 10000 2800 72 2800 100402 40 153 5050 40 165000 600 QTY-HEIGHTED STD. OEV. 26167 1 37097 636 1253 244401 46421 STANDARD DEVIATION 160100 1 140671 7000 IOZ5 341251 274209 SOURCEi Solane* Aooltcattoni International Corporation.1966 ------- TABLE B-16 SUMMARY STATISTICS ON THE TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS = EXCLUDING RANGES = PRESENT WITHIN ALL NON-RCRA HAZARDOUS WASTES FOR EACH RESIDUAL GROUP - BY MANAGEMENT PRACTICE 01 MANAGEMENT PRACTICE CONTAINERS LANDFARMING LANDFILLS PILES SURFACE IMPS. TANKS ALL MGMT. 6RPS. MANAGEMENT PRACTICE CONTAINERS LANDFARMING LANDFILLS PILES SURFACE IMPS. TANKS ALL MGMT. 6RPS. NO. OF MINIMUM SAMPLES CONC. (PPM) 5 3 33 2 15 SB 104 NO. OF WTO. SAMPLES 4 3 24 2 13 38 72 4 38 2 100 2 0.01 0.01 ' ICMJKTI 'ALL \.i MAXIMUM CONC. IPPMI 350000 40 500000 10000 24000 1000000 1000000 MEAN CONC. (PPM) 104140 39 64853 5050 4552 160850 105319 QTY-MEIGHTED flTY-HEIGHTEO MEAN CONC 8305 39 4562 9959 6305 6076 7200 (PPM) MED. CONC (PPM) 293 38 4 10000 2800 300 2000 MEDIAN CONC. (PPM) 70000 40 156 5050 200 3500 592 QTY -WEIGHTED STD. OEV. 27570 1 31446 636 8115 S3397 33224 STANDARD DEVIATION 144270 1 138929 7000 8530 268182 233586 SOURCE: Science Aoolication* International Corporation,1986 ------- TABLE B-17 SUMMARY STATISTICS ON THE TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS, EXCLUDING RANGES, PRESENT WITHIN ALL ISDB WASTES FOR EACH RESIDUAL GROUP - BY MANAGEMENT PRACTICE OO MANAGEMENT PRACTICE CONTAINERS LANDFILLS SURFACE IMPS. TANKS ALL MGMT. 6RPS. MANAGEMENT PRACTICE CONTAINERS LANDFILLS SURFACE IMPS. TANKS ALL MGMT. GBPS. NO. OF MINIMUM MAXIMUM MEAN SAMPLES CONC. (PPM) CONC. (PPM) CONC. (PPM) 1 1 3 22 27 NO. OF HTO. SAMPLES 1 1 3 17 22 300000 300000 200302 200382 0.1 2800 0.01 900000 0.01 900000 QTY -WEIGHTED MEAN CONC (PPM) 300000 200302 2754 39576 5262 300000 200382 1000 170461 157538 QTY-WEIGHTED MEO. CONC (PPMI 300000 200302 2800 72 2800 MEDIAN CONC. (PPM) 300000 200382 200 6500 3000 QTT-MEIGHTED STO. DEV. 0 0 342. 148704 39113 STANDARD DEVIATION 0 0 1562 269376 249858 SOURCE I Sc<«nc« Application* International Corporation.1986 ------- TABLE B-18 SUMMARY STATISTICS ON THE TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS, EXCLUDING RANGES, PRESENT WITHIN ALL ISDB WASTES FOR EACH RESIDUAL GROUP - BY MANAGEMENT PRACTICE MANAGEMENT PRACTICE CONTAINERS LANDFARMING LANDFILLS PILES SURFACE IMPS. TANKS ALL MGMT. 6RPS. MANAGEMENT PRACTICE CONTAINERS LANOFARMING LANDFILLS PILES SURFACE IMPS. TANKS ALL MGMT. GRPS. NO. OF MINIMUM MAXIMUM SAMPLES CONC. (PPM) CONC. (PPM) 16 3 51 3 7 26 69 NO. OF HTD. SAMPLES 14 3 40 3 6 26 76 3 38 0.2 100 2 0.03 0.03 960000 40 984621 10000 1006 1000000 1000000 MEAN CONC. (PPM) 328308 39 131338 3457 317 219762 167043 QTY-HEIGHTED QTY-MEIGHTED MEAN CONC (PPM) MED 439512 39 34297 9695 138 162236 17496 . CONC (PPM) 500000 38 4 10000 38 64000 10000 MEDIAN CONC. (PPM) 200000 40 585 271 40 135000 1538 QTY -WEIGHTED STD. OEV. 202716 1 78689 1696 214 266612 66650 STANDARD DEVIATION 327364 1 234962 5667 392 292096 274492 SOURCES Sctmc* Apolle«tlon» International Corporation,1966 ------- ON O TABLE B-19 SUMMARY STATISTICS ON THE TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS, EXCLUDING RANGES, PRESENT WITHIN ALL ISDB WASTES FOR EACH RESIDUAL GROUP - BY MANAGEMENT PRACTICE MANAGEMENT NO. OF MINIMUM PRACTICE SAMPLES CONC. (PPM) SURFACE IMPS. 1 4 ALL MGMT. 6RPS. 1 4 MAXIMUM MEAN CONC. (PPM) CONC. (PPM) 4 4 4 4 MEDIAN STANDARD CONC. (PPM) DEVIATION 4 0 4 0 MANAGEMENT NO. OF QTY-HEIGHTED QTY-UEIGHTED QTY-WEIGHTED PRACTICE HTD. SAMPLES MEAN CONC (PPM) MED. CONC (PPM) STO. DEV. SURFACE IMPS. 14 40 ALL MGMT. GRPS. 14 40 SOURCEi Sei«nc« Aoolications Intarnatfon*! Coroor«tloo,19M ------- TABLE B-20 SUMMARY STATISTICS ON THE TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS, EXCLUDING RANGES, PRESENT WITHIN ALL ISDB WASTES FOR EACH RESIDUAL GROUP - BY MANAGEMENT PRACTICE -RESIDUAL CATEGORY: =3PENT SOLVENTS- MANAGEMENT NO. OF MINIMUM MAXIMUM MEAN MEDIAN STANDARD PRACTICE SAMPLES CONC. I PPM) CONC. (PPM) CONC. (PPM) CONC. (PPM) DEVIATION LANDFILLS 1 312650 312050 312650 312650 0 TANKS 9 5 1000000 612223 990000 334990 ALL MGMT. GRPS. 10 5 1000000 762265 990000 353111 MANAGEMENT NO. OF QTY-UEIGHTEO QTY-HEIGHTEO QTY-UEIGHTEO PRACTICE NTD. SAMPLES MEAN CONC (PPM) MED. CONC (PPM) STD. DEV. LANDFILLS 1 312650 312850 0 TANKS 4 661266 750000 206379 ALL MGMT. GRPS. S 464615 312850 220614 SOURCE: Sci*nc« Application* International Corporation,1966 ------- TABLE B-21 SUMMARY STATISTICS ON THE TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS, EXCLUDING RANGES, PRESENT WITHIN ALL ISDB WASTES FOR EACH RESIDUAL GROUP - BY MANAGEMENT PRACTICE MANAGEMENT PRACTICE CONTAINERS LANDFILLS SURFACE IMPS. TANKS ALL HGMT. GRPS. MANAGEMENT PRACTICE CONTAINERS LANDFILLS SURFACE IMPS. TANKS ALL M6HT. GRPS. NO. OF MINIMUM SAMPLES CONC. IPPM) E 1 8 37 46 NO. OF WTO. SAMPLES 1 1 7 22 29 4 1009 4 1 1 MAXIMUM MEAN CONC. (PPM) CONC. IPPM) 70000 1009 24000 165000 165000 9TY-MEIGHTED MEAN CONC 4 1009 9632 622 2341 (PPM) 35002 1009 60 74 9611 10114 QTY-HEIGHTEO MED. CONC IPPM) 4 1009 2000 50 51 MEDIAN CONC. IPPM) 35002 1009 1106 300 300 9TY-HEIGHTEO STD. DEV. 0 0 9466 3250 6309 STANDARD DEVIATION 49495 0 106M 20429 27264 SOURCE! Scionca Application* International Corporation.1906 ------- TABLE B-22 SUMMARY STATISTICS ON THE TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS, EXCLUDING RANGES, PRESENT WITHIN ALL ISDB WASTES FOR EACH RESIDUAL GROUP - BY MANAGEMENT PRACTICE MANAGEMENT PRACTICE CONTAINERS LANOFARHING LANDFILLS PILES SURFACE IMPS. TANKS ALL MGMT. 6RP3. MANAGEMENT PRACTICE CONTAINERS LANDFARMING LANDFILLS PILES SURFACE IMPS. TANKS ALL MGMT. GRPS. MtSIUUAL CAItbUKIS =ALL tAI . NO. OF MINIMUM MAXIMUM SAMPLES CONC. (PPM) CONC. (PPM) 17 3 51 3 U 57 127 NO. OF HTD. SAMPLES 15 3 42 3 10 47 104 3 38 0.2 100 0.1 0.01 0.01 960000 40 984621 10000 2600 1000000 1000000 M/U M3IMAICK- MEAN CONC. (PPM) 326643 39 136066 3457 475 294260 210576 QTY-HEIGHTED QTY-HEIGHTEO MEAN CONC (PPM) MED 437313 39 34651 9695 1915 99912 13777 . CONC (PPMI 400000 36 4 10000 2600 72 2600 MEDIAN CONC. (PPM) 200000 40 600 271 40 170000 3767 QTY-HEIGHTED STD. DEV. 201661 1 79626 1696 1259 229614 60316 STANDARD DEVIATION 317062 1 231919 5667 636 364756 317587 SOURCES Science Aoolication* International Corporation!1966 ------- TABLE B-23 SUMMARY STATISTICS ON THE TOTAL CONCENTRATION OF HAZARDOUS CONSTITUENTS, EXCLUDING RANGES, PRESENT WITHIN ALL ISDB WASTES FOR EACH RESIDUAL GROUP - BY MANAGEMENT PRACTICE MANAGEMENT PRACTICE CONTAINERS LANOFARMINS LANDFILLS PILES SURFACE IHPS. TANKS ALL MGMT. WPS. MANAGEMENT PRACTICE CONTAINERS LANOFARMIN6 LANDFILLS PILES SURFACE IMPS. TANKS ALL MGMT. GAPS. NO. OF MINIMUM MAXIMUM SAMPLES CONC. (PPMI CONC. (PPMI 19 3 54 3 19 94 173 NO. OF HTO. SAMPLES 16 3 43 3 17 69 133 3 30 0.2 100 0.1 0.01 0.01 QTY-UEI6HTEO 960000 40 984621 10000 24000 1000000 1000000 MEAN CONC. (PPMI 295944 39 133565 3457 3674 162230 157274 QTT-HEIGHTED MEAN CONC (PPMI MEO 436033 39 27421 9695 6219 2111 5542 . CONC (PPM) 400000 30 100 10000 2000 50 51 MEDIAN CONC. (PPM) 200000 40 004 271 200 3500 1100 QTY -WEIGHTED STO. OEV. 202270 1 71727 1690 0005 30763 32774 STANDARD DEVIATION 312970 1 230454 5667 7725 316192 206306 SOURCES Setone* Apolfe«tlon» International Corporation,1906 ------- APPENDIX C FRANKLIN ASSOCIATES DATA BASE FOR WASTE OIL COMPOSITION This appendix contains a brief description of the waste oil concentration data collected, compiled, and analyzed by Franklin Associates, Ltd.* for the EPA Office of Solid Waste. Between 1981 and 1984, waste oil data and samples were collected and analyzed by Franklin Associates, Ltd. Analytical results were obtained for more than 1,000 waste oil samples.** As a result of these analyses, a data base characterizing the composition and concentration of waste oil was developed. The sampling techniques used to determine concentrations, however, were not reported. The major contaminants found in waste oil are: 1) heavy metals, especially lead and zinc; 2) organic solvents, such as benzene, xylene, and toluene; and 3) chlorinated solvents, particularly trichlorotrifluorethane. Table C-l summarizes the number of samples and the estimated concentration of hazardous constituents found in waste oil. Additional information is provided in Table C-2, which shows the concentration of potentially hazardous contaminants at the 75th and 90th percentile. Information about the *Franklin Associates, Ltd., November 1984. Composition and Management of Used Oil Generated in the United States, Final Report, EPA/530-SW-013, prepared for the U.S. Environmental Protection Agency, Office of Solid Waste and Emergency Response, Washington, DC. **Samples were taken from establishments storing waste oil in aboveground tanks, belowground tanks, and drums. 165 ------- TABLE C-l SUMMARY OF ANALYTICAL RESULTS FOR POTENTIALLY HAZARDOUS CONSTITUENTS FOUND IN WASTE OIL* Total Analyzed Samples Samples in which Contaminant Detected Number Percent Mean Concentration** Concentration Range (ppm) Low High Median Concentration (pom) Metals Arsenic Barium Cadmium Chromium Lead Zinc 537 752 744 756 835 810 135 675 271 592 760 799 25 89 36 78 91 98 17.26 131.92 3.11 27.97 664.5 580.28 0.01 0 0 0 0 0.5 100 3,906 57 690 21,700 8,610 5 48 3 6.5 240 480 Chlorinated Solvents Dichlorodifluoromethane 87 Trichlorotrifluoroethane 28 1,1,1-Trichloroethane 616 Trichloroethylene 608 Tetrachloroethylene 599 51 17 388 259 352 58 60 62 42 58 373.27 62,935.88 2,800.41 1,387.63 1,420.89 1 20 1 1 1 2,200 550,000 110,000 40,000 32,000 20 160 200 100 106 Total Chlorine 590 568 96 4,995 40 86,700 1,600 ------- TABLE C-l (Concluded) Total Analyzed Samples Samples in which Contaminant Detected Number Percent Mean Concentration** (pom) Concentration Range (ppm) Low High Median Concentration (ppm) Other Organics Benzene Toluene Xylene Benzo(a) anthracene Benzo(a)pyrene PCBs Naphthalene 236 242 235 27 65 753 25 118 198 194 20 38 142 25 50 81 82 74 58 18 100 961.2 2,200.48 3,385.54 71.3 24.55 108.51 475.2 1 1 1 5 1 0 110 55,000 55,000 139,000 660 405 3,800 1,400 20 380 550 12 10 5 330 *1,071 different waste oil samples were included in the Franklin Associates waste oil study. **Calculated for detected concentrations only. For the purpose of determining mean concentrations, undetected levels were assumed by Franklin Associates to be equal to the detection limit. Source: Franklin Associates, Ltd., November 1984. Composition and Management of Used Oil Generated in the United Statesf Final Report, prepared for the U.S. Environmental Protection Agency, Office of Solid Waste and Emergency Response, Washington, DC, Table 2, pp. 1-12. ------- TABLE C-2 SUMMARY OF WASTE OIL CONCENTRATION AT THE 75TH AND 90TH PfiRCENTILE* Concentration Concentration Number at 75th at 90th of Percentile** Percentile*** Samples (ppm) (ppm) Metals Arsenic 537 5 18 Barium 752 120 251 Cadmium 744 8 10 Chromium 756 12 35 Lead 835 740 1,200 Zinc 810 872 1,130 Chlorinated Solvents Dichlorodifluoromethane 87 160 640 Trichlorotrifluoroethane 28 1,300 100,000 1,1,1-Trichloroethane 616 1,300 3,500 Trichloroethylene 608 200 800 Tetrachloroethylene 599 600 1,600 Total Chlorine 590 4,000 9,500 Other Organics Benzene 236 110 300 Toluene 242 1,400 4,500 Xylene 235 1,400 3,200 Benzo(a)anthracene 27 30 40 Benzo(a)pyrene 65 12 16 PCBs 753 15 50 Naphthalene 25 560 800 168 ------- TABLE C-2 (Concluded) FOOTNOTES *For the purpose of determining percentile concentrations, undetected levels were assumed by Franklin Associates to be equal to the detection limit. **75 percent of the analyzed waste oil samples had contaminant concentrations below the given value. ***90 percent of the analyzed waste oil samples had contaminant concentrations below the given value. Source: Franklin Associated, Ltd., November 1984. Composition and Management of Used Oil Generated in the United States, Final Report, prepared for the U.S. Environmental Protection Agency,* Office of Solid Waste and Emergency Response, Washington, DC, Table 2, pp. 1-12. 169 ------- concentration of these contaminants by the type of unit in which they are stored (i.e., aboveground tanks, belowground tanks, and 55-gallon drums) is not available. 170 ------- APPENDIX D DEFINITIONS OF WASTE MANAGEMENT UNITS This appendix provides definitions for each of the major waste management practices discussed in this report. These definitions are based on the RCRA definitions (40 CFR 260.10). Drum or Container: Any portable device in which a material is stored, transported, treated, disposed of, or otherwise handled. Injection Well: A well into which fluids are injected. Underground injection is the subsurface emplacement of fluids through a bored, drilled, or driven well; or through a dug well, where the depth of the dug well is greater than the largest surface dimension. Landfarm: Landfill: Pile: Surface Impoundment: A facility or part of a facility at which hazardous waste is applied onto or incorporated into the soil surface. A disposal facility or part of a facility where hazardous waste is placed in or on land and which is not a landfarm, a surface impoundment, or an injection well. Any noncontainerized accumulation of solid, nonflowing hazardous waste. A facility or part of a facility which is a natural topographic depression, manmade excavation, or diked area formed primarily of earthen materials (although it may be lined with manmade materials), which is destined to hold an accumulation of liquid wastes or wastes containing free liquids, and which is not an injection well; examples of surface impoundments are holding, storage, settling, and aeration pits, ponds, and lagoons. 171 ------- Tank: A stationary device, designed to contain an accumulation of hazardous waste which is constructed primarily of nonearthen materials (e.g., wood, concrete, steel, plastic) which provide structural support. 172 ------- APPENDIX E AVAILABILITY OF WASTE COMPOSITION DATA BY SITE IN THE SUPERFUND SITE INSPECTION AND REMEDIAL INVESTIGATION REPORTS This appendix explicitly references the sites for which waste composition data were available from the site inspection and remedial investigation reports. Table E-l lists the 44 sites for which site inspections reports provided waste composition data. The table also identifies the particular type of waste management units/wastes for which data are available for each site. Table E-2 provides similar information for the 16 NPL sites for which data were obtained from remedial investigation reports. 173 ------- TABLE E-l AVAILABILITY OF WASTE COMPOSITION DATA IN SUPERFUND SITE INSPECTION REPORTS Site Name Waste Management Units/Wastes Revere Textile Arky Property Cosden Chemical Coating Corp. Kane and Lombard Hunterstown Road Middletown Airfield Shriver's Corner Road L.A. Clark and Sons First Piedmont Rock Quarry Saunders Supply Co. Mobay Chemical Corp. Morgantown Ordnance Works Sydney Mine G.C. Singleton By Pass 601 Sangamo/Twelve Mile/Hartwell NL Industrial Tara Corp Lead Site Parson's Casket Hardware Torch Lake Motor Wheel Disposal Site Rockwell Int'l Corp. Ormet Corporation City of Stoughton Landfill Midland Products Drums Drums Drums Drums Drums, surface impoundment Surface impoundment, tanks Drums Surface impoundment Drums Tanks Surface impoundment Drums Surface impoundment Drums Surface impoundment Sludge, landfill Waste piles Surface impoundment Mine tailings Tanks Surface impoundment Surface impoundment Drums Surface impoundment 174 ------- TABLE E-l (Concluded) Site Name Waste Management Units/Wastes Sand Springs Petrochemical Co. Passes Co. Brio Refining Stewco, Inc. Texarkana Wood Preserving Sheridan Disposal Services Midwest Manuf./No. Farm Site Lawrence Todtz Farm Site Lindsay Manufacturing Monroe Auto Equipment Co. Smuggler Mountain Burlington No. Tie Treating Facility Montana Pole and Treating Plant Mayflower Mountain Tailings Tooele Army Depot Micronutrients International Apache Powder Co. Queen City Farms Midway Landfill Naval Undersea Warfare Eng. Station Surface impoundment Drums Surface impoundment Surface impoundment Surface impoundment Surface impoundment, landfarm Surface impoundment Surface impoundment Surface impoundment Surface impoundment Mine tailings Surface impoundment Sludge, tanks Mine tailings Surface impoundment Waste piles Surface impoundment Surface impoundment Surface impoundment Sludge 175 ------- TABLE E-2 AVAILABILITY OF WASTE COMPOSITION DATA IN SUPERFUND REMEDIAL INVESTIGATION REPORTS Site Name Reference Waste Management Units/Wastes Beacon Heights Landfill Hocomonco Pond Bridgeport Rental & Oil Services Krysowaty Farm Swope Oil & Chemical Co. Drake Chemical McAdoo Associates Mill Creek Dump Arcanum Iron & Metal Laskin/Poplar Oil Co. Old Mill Cecil Lindsey Bayou Sorrell Cleve Reber Old Inger Oil Refinery Celtor Chemical Works NUS Corporation, 1985d TRC Environmental Consultants, 1985 NUS Corporation, 1984b NUS Corporation, I984a NUS Corporation, 1985c NUS Corporation, 1985e NUS Corporation, 1984c NUS Corporation, 1985b CH2M Hill, 1985c CH2M Hill, 1985d CH2M Hill, 1984 CH2M Hill, I985a CH2M Hill, 1985b CH2M Hill, 1985e Louisiana Department of Natural Resources, 1983 CH2M Hill, 1985f Sludge Surface impoundment Surface impoundment, tanks Drums Surface impoundment Surface impoundment Tanks Drums Rubber chips Surface impoundment, tanks Drums Drums Surface impoundment, spent lime cells Drums Sludge, surface impoundment, tanks Mine tailings 176 ------- APPENDIX F BIBLIOGRAPHY Arthur D. Little, Inc., 1976. Hazardous Waste Generation, Treatment and Disposal In the Pharmaceutical Industry, prepared for the U.S. Environmental Protection Agency, Publication SW-508, Washington, DC. Battelle Columbus Laboratories, September 1976a. Assessment of Industrial Hazardous Waste Practices—Electroplating and Metal Finishing Industries—Job Shops, prepared for the U.S. Environmental Protection Agency, Publication SW-136c, Washington, DC. Battelle Columbus Laboratories, January 1976b. Final Assessment of Industrial Hazardous Waste Practices—Electroplating and Metal Finishing Industries—Captive Shops, prepared for the U.S. Environmental Protection Agency, Washington, DC. Blackman, W. C., R. L. Garnas, J. E. Preston, and C. M. Swibas, March 1985. "Chemical Composition of Drum Samples from Hazardous Waste Sites," U.S. Environmental Protection Agency, National Enforcement Investigations Center, Denver, CO. Boynton, L., March 1986. U.S. Environmental Protection Agency, Sample Management Office, personal communication, Washington, DC,. Calspan Corporation, 1977. Assessment of Industrial Hazardous Waste Practices in the Metal Smelting and Refining Industry, Volumes 1, 3, and 4, prepared for the U.S. Environmental Protection Agency, Publication SW-145c, Washington, DC. CH2M Hill Ecology and Environment, July 1985a. Draft No. 2 Remedial Investigation Report for Cecil Lindsey Site, Newport, Arkansas, prepared for the U.S. Environmental Protection Agency, Washington, DC. CH2M Hill Ecology and Environment, July I985b. Remedial Investigation Report Bayou Sorrel, Louisiana, Draft, Volume 1 of 2, prepared for the U.S. Environmental Protection Agency, Washington, DC. CH2M Hill Ecology and Environment, June 1985c. Draft Remedial Investigation Report, Volume 1, Arcanum Iron and Metal Site, Arcanum, Ohio, prepared for the U.S. Environmental Protection Agency, Washington, DC. 177 ------- CH2M Hill Ecology and Environment, May 1985d. Draft Remedial Investigation Report, Volume 1, Laskin-Poplar Oil Co. Site"7 Jefferson, Ohio, prepared for the U.S. Environmental Protection Agency, Washington, DC. CH2M Hill Ecology and Environment of May 1985e. Remedial Investigation Report, Volume 1 2, Cleve Reber Site, Ascension Parish, Louisiana, prepared for the U.S. Environmental Protection Agency, Washington, DC. CH2M Hill Ecology and Environment, April 1985f. Remedial Investigation Report, Celtor Chemical Works Site, Hoopa, California, prepared for the U.S. Environmental Protection Agency, Washington, DC. CH2M Hill Ecology and Environment, December 1984. Remedial Investigation/Feasibility Study, Old Mill Site, Rock Creek, Ohio, prepared for the U.S. Environmental Protection Agency, Washington, DC. Eckel, W. P., D. P. Trees, and S. P. Kovell, May 1985. "Distribution and Concentration of Chemicals and Toxic Materials Found at Hazardous Waste Dump Sites," proceedings of the National Conference on Hazardous Wastes and Environmental Emergencies, Cincinnati, OH. Foster D. Snell, Inc., 1978. Assessment of Industrial Hazardous Waste Practices, Rubber and Plastics Industry, prepared for the U.S. Environmental Protection Agency, Publication SW-163c, Volumes 1-4, Washington, DC. Franklin Associates, Ltd., November 1984. Composition and Management of Used Oil Generated in the United States, Final Report, EPA/530-SW-013, prepared for the U.S. Environmental Protection Agency, Office of Solid Waste and Emergency Response, Washington, DC. Freund, J. H., 1971. Mathematical Statistics, Prentice-Hall, Inc., Englewood Cliffs, NJ. Friedman, P. H., W. P. Eckel, D. P. Trees, and B. Clemens, November 1984. "Construction of a Data Base from Hazardous Waste Site Chemical Analyses," The 5th National Conference on Management of Uncontrolled Hazardous Waste Sites, Washington, DC. Haan, Charles T, 1979. Statistical Methods in Hydrology, The Iowa State University Press, Ames, IA. Hollander, M. and D. A. Wolfe, 1973. Nonparametric Statistical Methods, John Wiley and Sons, NY. 178 ------- Jacobs Engineering Co., 1976. Assessment of Hazardous Waste Practices in the Petroleum Refining Industry, prepared for the U.S. Environmental Protection Agency, Publication SW-129c, Washington, DC, Liroff, S. D., M. Hoffman, G. Sessler, and C. Shulock, 1978. Management of Environmental Risk: A Limited Integrated Assessment of the Waste Oil Rerefining Industry, prepared for the National Science Foundation by Teknekron, Inc., Berkeley, CA. Louisiana Department of Natural Resources, October 1983. Remedial Investigation-Phase One Element I, Old Inger Abandoned Hazardous Waste Site, Darrow, Louisiana, (Final Report), Baton Rouge, LA. NUS Corporation, August 1985a. Feasibility Study for Bog Creek Farm Site, Howell Township, New Jersey, prepared for the U.S. Environmental Protection Agency, Washington, DC. NUS Corporation, August 1985b. Remedial Investigation/Feasibility Study Report, Mill Creek Site, Erie County Pennsylvania, prepared for the U.S. Environmental Protection Agency, Washington, DC. NUS Corporation, June 1985c. Remedial Investigation/Feasibility Study of Alternatives, Swope Oil Company Site, Pennsauken Township, Camden County, New Jersey, prepared for the U.S. Environmental Protection Agency; Washington, DC. NUS Corporation, April 1985d. Remedial Investigation Report, Beacon Heights Landfill Site, Beacon Falls, Connecticut, Appendix, prepared for the U.S. Environmental Protection Agency, Washington, DC. NUS Corporation, April 1985e. Remedial Investigation Report, (Phase II), Drake Chemical Site, Lock Haven, Clinton County, Pennsylvania, (Revised), prepared for the U.S. Environmental Protection Agency, Washington, DC. NUS Corporation, October 1984a. Remedial Investigation Report and Feasibility Study for Krysowaty Farm Site, Township of Hillsborough, Somerset County, New Jersey, prepared for the U.S. Environmental Protection Agency, Washington, DC. NUS Corporation, July 1984b. Remedial Investigation Report, Bridgeport Rental and Oil Services Site, Logan Township, New Jersey, (Draft), prepared for the U.S. Environmental Protection Agency, Washington, DC. NUS Corporation, May 1984c. Remedial Investigation Report, McAdoo Associates Site, Kline Township, Schuylkill County, Pennsylvania, prepared for the U.S. Environmental Protection Agency, Washington, DC. 179 ------- PEDCo Environmental, Inc., November 1983. Evaluation of Management Practices for Mine Solid Waste Storage, Disposal, and Treatment. Volume 1, Characterization of Mining Industry Wastes, (Draft), prepared for the U.S. Environmental Protection Agency, Office of Research and Development, Cincinnati, OH. R. E. Wright Associates, Inc., July 1985. Draft Remedial Investigation Report and Feasibility Study of Alternatives, Helen Kramer Landfill Site, Mantua Township, Gloucester County, New Jersey, prepared for the NUS Corporation, Pittsburgh, PA. Science Applications International Corporation, 1986. Unpublished Data from Industry Studies Data Base (ISDB), McLean, VA. SCS Engineers Inc., 1976. Assessment of Industrial Hazardous Waste Practices — Leather Tanning and Finishing Industry, prepared for the U.S. Environmental Protection Agency, Publication SW-131c, Washington, DC. Siegel, S., 1956. Nonparametric Statistics for the Behavioral Sciences, McGraw-Hill Book Company, Inc., NY. Swain, J. W., Jr., et al., 1977. Assessment of Industrial Hazardous Waste Management Petroleum Re-refining Industry, prepared for the U.S. Environmental Protection Agency, Publication SW-144c, Washington, DC. TRC Environmental Consultants, Inc., 1985. Hocomonco Pond Site Westborough, Massachusetts, Remedial Investigation Report, (Draft), prepared for NUS Corporation, Pittsburgh, PA. TRW, Inc., 1976. Assessment of Industrial Hazardous Waste Practices, Organic Chemicals, Pesticides and Explosives Industries, prepared for the U.S. Environmental Protection Agency, Publication SW-118c, Washington, DC. U.S. Environmental Protection Agency, 1984. Uncontrolled Hazardous Waste Site Ranking System; A Users Manual, HW-10, Washington, DC. U.S. Environmental Protection Agency, 1980. Subtitle C, Resource Conservation and Recovery Act of 1976 (RCRA), Final Environmental Impact Statement Part I—Appendices, Office of Solid Waste, Washington, DC. Versar, Inc., 1976. Assessment of Industrial Hazardous Waste Practices, Textiles Industry, prepared for the U.S. Environmental Protection Agency, Publication SW-125c, Washington, DC. 180 ------- Versar, Inc., 1975a. Assessment of Industrial Hazardous Waste Practices; Inorganic Chemicals Industry, prepared for the U.S. Environmental Protection Agency, Publication SW-l04c, Washington, DC. Versar, Inc., 1975b. Assessment of Industrial Hazardous Waste Practices; Storage and Primary Battery Industries, prepared for the U.S. Environmental Protection Agency, Publication SW-102c, Washington, DC. Wapora, Inc., 1977a. Assessment of Industrial Hazardous Waste Practices; Electronic Components Manufacturing Industry, prepared for the U.S. Environmental Protection Agency, Publication SW-l40c, Washington, DC. Wapora, Inc., 1977b. Assessment of Industrial Hazardous Waste Practices; Special Machinery Manufacturing Industries, prepared for the U.S. Environmental Protection Agency, Publication SW-141c, Washington, DC. Wapora, Inc., 1975. Assessment of Industrial Hazardous Waste Practices: Paint and Allied Products Industry, Contract Solvent Reclaiming Operations, and Factory Application of Coatings, prepared for the U.S. Environmental Protection Agency, Office of Solid Waste Management Programs, Washington, DC. 181 ------- |