Risk-Based Waste Classification in California (1999)

The NRC Committee commends the Department of Toxic Substances Control (DTSC) of the California Environmental Protection Agency (Cal/EPA) for its effort to improve the scientific structure of its regulatory system and increase the use of modern risk-analysis procedures and environmental models in this undertaking. Clearly DTSC devoted a substantial effort to the development of the proposed approach for classification of hazardous wastes, and made extensive use of many of the most up-to-date scientific procedures.

The committee based its review on the documentation and spreadsheets that DTSC provided, supplemented by DTSC's response to questions submitted by the committee, public presentations made to the committee by staff from DTSC, comments from other individuals interested in the proposed regulations, and materials available to the public on DTSC's Internet homepage [http://www.dtsc.ca.gov]. Although these materials are extensive, the committee was left with substantial uncertainty and confusion about exactly what DTSC was attempting to accomplish and the procedures and assumptions it adopted in developing the proposed regulations. Among these uncertainties are the protection goals for human health and the environment, definitive exposure scenarios (both those modeled and those unmodeled), and a logical step-by-step explanation of the actual waste classification methodology specifying how a waste is assessed based on DTSC's flowchart (DTSC 1998a, p. 36). DTSC needs to explicitly state the advantages and disadvantages of

their proposed system compared with both their current system and the federal system. The advantages in terms of regulatory flexibility, improved health and environmental protection, and any economic benefits for both the regulated community as well as the general public should be clearly articulated and documented.

In its review, the committee adopted the following criteria to guide its evaluation of the proposed methodology for classifying waste. Some of these criteria are more important to the committee's deliberations, conclusions, and recommendations than others. However, they are all important to the development and implementation of regulations designed to protect human health and the environment. The committee recommends that DTSC consider using similarly defined set of criteria when revising and evaluating their proposed approach.

• Statement of Goals. The effort should begin with a clear, coherent statement of the health and environmental protection goals that the waste classification system is attempting to achieve.
• Realistic Exposure Scenarios. The analyses should be based on exposure scenarios that are realistic and consistent with the health and environmental protection goals established.
• Multimedia and Multipathway Assessment. The risk assessments should incorporate, in a consistent and integrated fashion, the full range of potential exposures to human and nonhuman receptors through different media and along different pathways.
• Valid Science. The analyses should incorporate data, assumptions, and relationships that represent the best current scientific knowledge at the appropriate level of detail. In addition, the data and analyses should be thoroughly validated and peer-reviewed. The variability and uncertainty associated with the data, assumptions, and relationships should be explicitly recognized in the analyses.
• Transparency. The entire process should be clearly explained and thoroughly documented. Policy assumptions and decisions should be clearly distinguished from scientific data. The bases for making the many scientific and policy assumptions that are inevitable in such an undertaking should be clearly explained.
• Flexibility. The analytic tools and regulations should incorporate sufficient flexibility so that the parameters and approaches can be easily modified when it is appropriate to do so. Then situations that

• differ in assumptions from the initial analyses can be realistically represented, and advances and improvements in scientific information and knowledge can be easily incorporated.
• Implementation Practicality and Evaluation. The proposed waste-management system needs to be designed with a continuous concern for the feasibility, practicality, and costs (to all stakeholders) of its implementation. It also requires a process for evaluating its success in achieving its goals and whether the program impacts are consistent with the assumptions that were made in its formulation.

The remainder of this chapter presents the committee's conclusions of how the approach adopted by DTSC satisfies those criteria. The remaining chapters provide more detailed comments about specific elements of the proposed classification system.

In any undertaking directed at protecting human health and the environment, it is important to clearly set forth, at the beginning of the regulatory process, the level of protection that is being sought and for whom. These protection goals become the basic policy that drives the whole process. These goals guide the project staff on what analyses should be carried out and what types and stringency of management requirements should be imposed. They provide the information the project managers need to ensure that regulations are coherent and consistent. They also include the fundamental information the public needs to intelligently understand, review, comment on, and respond to the process and results.

The clearer and more explicit the statement of the protection goals, the more coherent the project will be. There will be less tendency for different staff members to establish their own goals, and less effort spent by the public in divining the goals of the project from its results. Clear, explicit goals should result in a more efficient development process and more focused and relevant public comments.

The statement of the protection goals should address such questions as

• What population is to be protected and what level of protection is to be provided to this population?

• What level of protection is to be provided to the most-exposed individuals?
• Which sensitive populations are to be considered and what level of protection is to be provided to them?
• What level of protection is to be provided to various forms of wildlife and other ecological components?

For illustrative purposes, the committee offers the following statements to show how the human health protection goals might be more clearly and precisely stated. For risks from hazardous waste disposal all the following protection goals would be required:

• Ninety-five percent of California's population will be protected from carcinogenic risks to a level of 10^-6 or better (lifetime risk).
• The reasonably most-exposed children will be protected from carcinogenic risks to a level of 10^-5 or better (lifetime risk).
• All other reasonably most-exposed individuals will be protected from carcinogenic risks to a level of 10^-4 or better (lifetime risk).
• No reasonably most-exposed individual will be subject to other chronic health risks at a level greater than a hazard index¹ of 1.0.
• No reasonably most-exposed individual will be subject to acute health risks exceeding a margin of exposure of 10.

''Reasonably'' might be defined as assuming that existing regulatory programs are implemented effectively, exposure assumptions are consistent with existing identified conditions, and populations and individuals

are considered to have no unusual sensitivities to the substances to which they are exposed.

The last suggested statement of a health protection goal concerns acute health risks. DTSC's inclusion of acute toxicity measures in its classification scheme was confusing, as no goals involving acute effects were given in the DTSC documentation. It appears, however, that based on the proposed classification schematic (DTSC 1998a, p. 36) DTSC does have a distinct goal of protecting against acute effects, in addition to the partially stated goal of protecting against chronic effects. The committee recommends that DTSC establish protection goals at some level for both acute and chronic effects.

Similarly, for wildlife (flora and fauna) protection, the goals might be stated as protecting 95% of the individuals in 90% of the species by not exceeding a critical concentration 90% of the time (or over 90% of the state's area). To make such a goal operational, DTSC would probably have to identify, for example, by species sensitivity and exposure assessments, certain "indicator species" that would represent the entire exposed wildlife population.

California law requires DTSC to provide protection from "substantial present or potential hazard" even when the wastes are "improperly treated, stored, transported, or disposed, or otherwise improperly managed" (California Health and Safety Code §25141 (a)(2)). Such a prescription might be built into DTSC's goals by stating the allowable levels of risk to individuals, under the assumption that other controls are not implemented effectively.

The documentation presented to the committee does not contain a clear exposition of goals for either the human health or ecological protection. The document refers to a "reference carcinogenicity risk level" of 10^-5 (DTSC 1998a, p. 3), and, in an appendix, states that "all health-based values are 10th percentile estimates of the concentrations that would correspond to the stated level of risk or hazard" (DTSC 1998a, p. 43). In oral and written responses to questions from the committee, DTSC indicated that "the intent is to be 90% certain that the true risk or hazard index for an individual taken at random from the population is not greater than 10^-5 or one, respectively." (DTSC, personal commun., October 29, 1998, see Appendix C, No. 28, question 71). This is equivalent to setting a goal of protecting 90% of the population of California from risks equal to or greater than those indicated.

However, most of the critical-exposure estimates (i.e., those used

for determining regulatory levels) appear to be based on the most-exposed individuals. Thus, it is unlikely that 10% of California's population consists of landfill operators with 30% of their skin exposed, or that 10% of the population lives within 100 meters (m) of a landfill (assumptions made in the exposure scenarios). The analyses apply the 10^-5 carcinogenic risk protection and the hazard index of 1.0 to these most-exposed individuals without explicitly enunciating any health protection goals for them. It is also unclear whether the individual maximum exposures used are reasonable for the chronic exposure scenarios, and there are no clear scenarios propounded for acute exposures.

With respect to wildlife, the analyses fail to establish clear wildlife protection goals, and they appear to assume in many cases that risk levels established to protect human health will also protect wildlife. DTSC provides no justification for this assumption.

The committee notes that definitive data are rarely available to answer risk-related questions unambiguously. As a result, many risk-based decisions are made on the bases of science-policy decisions. These science-policy decisions can include, but are not limited the selection of protection goals, exposure scenarios, sensitive subpopulations, and default assumptions. While evaluation of policy is not within the committee's task, such decisions must be recognized as forming one foundation of DTSC's proposal. They cannot be ignored when reviewing the scientific merits of the proposal.

The existing DTSC waste-classification system is broader than the federal waste-classification system under the Resource Conservation and Recovery Act (RCRA), in that the DTSC system considers acute toxicity, effects on aquatic life, and human exposure pathways other than groundwater. Although the current waste-classification system is broader than RCRA, it is based only on the potential toxicity of the waste and its constituents, and does not factor in potential exposure, including environmental fate and transport. The waste-classification system being proposed by DTSC attempts to correct those deficiencies, making use of multimedia, multipathway, risk-assessment models to classify wastes.

DTSC's proposed approach differs from the current federal RCRA waste classification in three ways. First, the proposed system has two

thresholds for the classification of wastes, rather than a single threshold. Second, the proposed approach is based on risk, rather than solely on the toxicity of the chemical of interest. Thus, both the toxicity of the chemical and the potential for exposure are considered in the proposed waste-classification system through the use of multimedia, multipathway risk-assessment models. Third, the proposed system is based on individual chemicals rather than on sources or types of waste. RCRA has a list of approximately 40 chemicals that are regulated on a specific-chemical basis, but the rest of the classifications are based on sources or types of waste.

DTSC's efforts to improve the state's environmental management by using a more comprehensive, multimedia, multipathway, risk-based approach is commendable. However, no waste-classification system can be entirely risk-based. DTSC should reconsider two of the decisions it made in the selection and use of risk-assessment models. The CalTOX model incorporates many of the appropriate factors for undertaking a good multimedia, multipathway risk assessment, and its components can be considered to approximate the state-of-the-art when correctly applied. A disadvantage of CalTOX is that it does not include exposure scenarios that are explicitly appropriate for ecological receptors. Given the overall advantages of such a model, however, the committee questions why it was not used more extensively. In the analyses, exposures to water-soluble contaminants and non-water-soluble contaminants are evaluated separately in soluble or extractable regulatory thresholds (SERTs) for groundwater pathways and total threshold limit concentrations (TTLCs) for air, surface water, and soil pathways. However, human and ecological receptors are typically exposed to contaminants by more than one pathway. It is desirable, therefore, to consider multimedia, multipathway exposures using a single model rather than to evaluate different pathways separately; DTSC should consider integrating the exposure calculations to simultaneously include all pathways and environmental media. For example, benzene is a constituent of many wastes and is readily mobile within and between most environmental compartments. To determine the risks associated with benzene contamination of a waste, its movement through and between air, soil, groundwater and surface waters should be assessed simultaneously.

DTSC has decided to classify wastes on the basis of single chemical components. The approach has a clear practical advantage in that

information on the toxicity of specific chemicals is generally available, whereas toxicity information about wastes from various sources is generally unavailable. Also, the chemical composition of waste from specific sources or processes can vary from time to time both qualitatively and quantitatively. Classifying waste on the basis of individual chemicals allows the use of available toxicity data, response to changes in waste composition, and addition of new information.

However, the committee questions DTSC's single-chemical approach. The disadvantage of regulating on a chemical-by-chemical basis is that people and ecological receptors are rarely exposed to individual chemicals; instead they are typically exposed to mixtures of chemicals. The proposed classification system does not address the additive or potential synergistic effects of exposure to multiple toxicants. The U.S. Environmental Protection Agency (EPA) handles the risk of mixtures by adding the cancer risk estimates from each component and, for noncancer risk, by adding the hazard indices for each component, at least for a first approximation. DTSC has chosen not to add the risk estimates from each chemical in a mixture when defining TTLCs or SERTs. Such an approach may underestimate the risk from a mixture. It is recommended that DTSC, lacking better information about the effects of chemical mixtures, adopt the same default assumption as EPA, and linearly add the risk estimates for the mixture's components to estimate the risks associated with the entire mixture. DTSC is encouraged to monitor any scientific advances in the evaluation of chemical interactions at low exposures and to incorporate such advances into their approach. DTSC's proposed approach does, however, consider whole wastes, or all waste components, when assessing risks from acute toxicity (see Chapter 4, Toxicity Tests).

Federal regulatory decisions are typically a blending of risk-based procedures together with a multitude of policy decisions. DTSC's approach is no different. For example, the structure of the risk-assessment model to be used (in this case a model of the migration of a chemical between environmental compartments), the particular parameters to be used in the model, and the exposure scenarios to be employed all involve policy decisions by DTSC. A completely risk-based approach to regulations is impractical, if not impossible. Chapter 4 of this report contains substantial detailed comments on DTSC's selection and application of the risk-assessment models.

The problem of inadequate explication and justification of policy assumptions incorporated in DTSC's analyses was particularly noticeable in the selection of the exposure scenarios. The selection of realistic exposure scenarios is fundamental to carrying out reasonable risk assessments. However, the committee had great difficulty in examining the appropriateness of the scenarios selected by DTSC, because there was no discussion of the policies that DTSC wished to implement, or documentation of any connection between policy and scenario. The DTSC staff, in written and oral responses (DTSC, personal commun., October 9, 1998, see Appendix C, No. 28; DTSC, personal commun., November 20, 1998), suggested that there had been some process by which the scenarios were selected; however, they provided no information on what the selection process was and how or why some scenarios were rejected and others were not. DTSC also suggested that the scenarios were selected to be representative not only of the explicitly described situation, but also representative of other situations (e.g., the landfill scenario was also supposed to apply to waste piles). However, no documentation was provided to elucidate the other situations that each scenario was supposed to represent. The description of the acute exposure scenarios was extremely limited for oral and dermal exposures and non-existent for inhalation exposures, and the committee is unsure whether DTSC intended to protect separately against acute effects, or whether the acute toxicity part of the classification scheme was somehow also supposed to protect against chronic effects.

The justification of the wildlife exposure scenarios was even weaker than that of the human exposure scenarios. No wildlife-specific scenarios were provided by the DTSC, and it seems likely that human exposure scenarios are inappropriate for wildlife. The DTSC does not provide a basis for their assumption that the lower TTLCs or SERTs derived for the protection of humans would be protective of aquatic or terrestrial wildlife. Although there are some differences in sensitivities between species, the greatest difference in toxic effects of a chemical on wildlife often results from differences in exposure of wildlife. Some wildlife, because of restricted home ranges, are likely to have higher exposure than humans. This is particularly important for soil-dwelling organisms or those at the top of food chains where the biomagnification potential of chemicals could come into play.

The DTSC report does not provide any details about how ecological assessments will be conducted. The report does indicate the general steps that would be taken for conducting a de novo ecological risk assessment, but it does not specify the scenarios or the methods that would be used to derive reference doses for specific chemicals to compare with estimated environmental or tissue concentrations (DTSC 1998a). Chapter 3 of this report contains other observations about DTSC's selection of realistic exposure scenarios.

The committee is pleased that DTSC has incorporated a number of up-to-date and valid scientific methods in developing its proposed regulations. These include not only the multimedia, multipathway risk assessments mentioned earlier, but also sophisticated modeling techniques and a statistical method for incorporating factors whose values vary for different members of a population. In making these selections, DTSC has established an overall scientific approach to its regulatory effort which, in many ways, represents the best current scientific knowledge. However, in many of the details of its effort, DTSC has failed to realize the potential promised by this approach. These failures are the focus of Chapters 3 and 4. In general, DTSC's approach would have benefitted substantially from a careful internal review for content, presentation, and implementation prior to external peer review.

The information presented to the committee regarding the scientific analyses is very uneven, and the poor documentation has made it difficult to judge the validity of the scientific methodologies adopted. The documentation for the principal exposure model used in the analysis, CalTOX, appears to have an adequate level of detail, although the committee has not performed an extensive peer review of all model details. The computations in the lead spreadsheet are adequately documented, but the tables are confused by the inclusion of extraneous materials, such as the inclusion of non-zero values for adults when the receptors selected by the scenario description are children. In addition, the structure of the model might not be clear to the uninitiated reader. The documentation of the preliminary endangerment assessment (PEA) model is inadequate and contains many errors. It is difficult to ascertain what values were used for which scenarios, and there are substantial differences between

the documentation and the spreadsheet implementation. The documentation for calculation of the SERTs is adequate, although it is impossible to determine what distribution was supposed to be used for the liner protection factor.

The documentation for the acute exposure scenarios is so limited that the committee was unable to discern DTSC's reasoning, and hence was unable to tell whether those scenarios were appropriate or appropriately used. If the acute scenarios were indeed designed with a goal of protecting against acute risks (separately and independently of the goal of protecting against chronic risks), then the approach taken might be reasonable. Even then, use of an endpoint such as an acute lowest observed adverse effect level would be preferable to extrapolating from an acute LD₅₀ or LC₅₀ value. If, however, DTSC is attempting to use the acute toxicity measures for protection against chronic risks (and the acute toxicity approaches are currently all that are applied for any wastes that do not contain any of the 38 TTLC chemicals), then the measures of toxicity used by DTSC are scientifically inappropriate.

DTSC's efforts to adopt a statistical method for characterizing variables is to be commended. However, the application of this method was incomplete and uneven. Many factors, particularly those characterizing exposure scenarios and other policy decisions, which could have been characterized as statistical distributions, were, instead, included as point estimates. DTSC also failed to address adequately the distinction between variability and uncertainty. It was stated that the 90th percentile of the calculated distributions was to be used; however, there was no distinguishing variability and uncertainty distributions in the discussion or in the analyses. Such analyses may have meaning only if the calculations are for randomly selected individuals (so that the variability becomes an uncertainty). In response to the committee's questions, however, DTSC indicated that they were aiming at the 90th percentile of both the variability and the uncertainty distribution; this is not achieved using the statistical method they selected.

Another major failure, discussed in more detail in Chapters 3 and 4, is the inadequacy of the sensitivity analyses. DTSC's proposed approach is complex, and involves many assumptions. Sensitivity analyses are essential to determine the most important parameters of the models, and afford a mechanism to focus on those important parameters. During the presentations to the committee, comments on several of the more unreasonable scenarios presented in public comments were dismissed by

DTSC by allusion to sensitivity analyses that indicated a single or small number of pathways dominated the risk. However, the presentation of such sensitivity analyses was weak and the analyses themselves were insufficient to justify the exclusion of some pathways. It is important that DTSC determine how sensitive the results of the analyses are to the numerous factors and assumptions in the model. DTSC has failed to do this in a comprehensive manner, and has failed to present and interpret the sensitivity analyses it did undertake in a meaningful and readily understood manner.

The documentation of the values used in DTSC's analyses is uneven. In some, there is substantial explanation of the values used; in others, very little. Sometimes a value is selected based on a reference to its use elsewhere, although the referenced document does not provide any indication of the source of the specified value (e.g., air-dispersion mixing height above a landfill, where the value selected might be reasonable as used in the referenced document but is incorrect in the DTSC context).

In its attempt to use the most current scientific approaches in its analyses, DTSC made use of models, data, and scientific assumptions that have not been adequately validated or peer-reviewed. In one such case, DTSC used methodologies and data from EPA's proposed hazardous waste identification rule that were found to be inadequate by EPA's Scientific Advisory Board when subjected to peer review. If DTSC believes that it needs to make use of information or methodologies that have not yet undergone full peer review, it should thoroughly document these cases and, to the extent possible, subject them to peer review before releasing the proposal for public comment.

DTSC also should have included the following information in its documentation:

• An analysis of the physical processes involved in the transport of waste from source to receptor in each scenario, and the models that are to be used to represent such physical processes (model requirements).
• A complete (mathematical) description of the models that are used; and a specification of which models are used in which scenarios to match which physical processes.
• A complete specification of each parameter value used in the document, together with references allowing retrieval of the description of its measurement or its method of calculation.

DTSC has undertaken a very sophisticated and complicated analysis in support of its proposed methodology. This creates a burden on DTSC to clearly explain the complex models and risk-assessment processes used to establish its regulatory limits. Most of the community affected by these regulations has little capacity (including time and training) to fully understand the basis for or implications of the proposed regulations. After reviewing almost 2,000 pages of materials (including public comments), the committee concludes that DTSC has failed to provide a transparent explanation or justification of the proposed methods, including a logical step-by-step explanation of how a waste is classified (DTSC 1998a, p. 36).

The documentation provided is incomplete, disjointed, overly long, contradictory in places, and confusing. It appears to be predominately an assemblage of disparate documents, at least some of which were still only in draft form. Although DTSC claimed to be under a tight schedule, this schedule was largely self-imposed. The number of years that it has taken to prepare this proposed approach should have provided an opportunity to prepare adequate documentation. Preparing documentation should not be postponed until the last minute; it should accompany the release of the proposal for public comment. Evidence of the inadequacy of DTSC's explanation of its approach lies in the fact that many of the public comments provided to the committee appear to be based on incorrect interpretations about what DTSC is proposing and why. The principal documentation failures, apart from missing elements, are the lack of logical order to the documents, the lack of a complete overview with adequate cross-referencing, and the lack of adequate copyediting for accuracy and clarity. The absence of table of contents, index, and pagination, incorrect cross-referencing, and inadequate referencing to source documents are also indicative of poor documentation.

Assuming that the selected scenarios match the policy requirements, a link between the models and the scenarios in which they were used, including a clear and complete description of each scenario, is lacking. Although in some cases, it is clear which models were used in which scenarios, there is a failure to match model requirements with the model used (e.g., dust emission and dispersion, landfill organic emissions, liner protection factor, and selection of many parameter values, such as dust-deposition rate, and mixing height above landfill). This indicates that

DTSC's process for matching model requirements to scenarios was flawed. It appears that DTSC used existing models that were close to what was required for the scenarios, perhaps after modifying them slightly. No analysis of model requirements for each scenario is documented, however, so the committee can only note where the models used do not match the scenarios selected.

The selection of values used in the models is often not adequately described or referenced; and sometimes the values used do not correspond to the descriptions or references cited (e.g., the values for the mixing depth of soil, and the deposition rate of dust). In some cases, the mismatch occurs because the scenarios were not adequately described. For example, the potential dust-emission area for the landfill was set equal to the whole landfill area. DTSC indicated at the second public meeting (DTSC, personal commun., November 20, 1998), that this value was selected because of an assumption that special waste could be used as landfill cover; or perhaps because this scenario was also supposed to account for the waste pile. There is no documentation of either possibility; such documentation belongs in the missing discussion of the relationship between policy and scenarios.

The most serious documentation failures relate to the policy decisions that are necessarily a fundamental element of these analyses. These flaws range from what level of health and environmental protection is intended, to why twice the level of detection was established as the regulatory level when the risk level falls below the detection-level criterion. The committee is still unsure about the DTSC's intended approach to background concentrations because different approaches appear to be suggested at different points in the documentation, and the practice appears to be different from any of them.

It is recommended that the documentation for the proposed regulations be substantially improved to clearly set forth:

• An explanation and justification for the policy decisions incorporated into the proposal.
• An explanation of the regulatory context or scientific basis for these decisions.
• A description of the process used to relate the policy decisions to the scenarios to be examined. In particular, how are all exposure scenarios reduced to a finite set of exposure scenarios (which may be a set that represent the worst case in some sense)?

• A complete description of all the selected scenarios to be examined, including the assumptions related to any policy decision.
• Several examples of known wastes that illustrate the differences; between the current and the proposed classification systems.

In addition, it is strongly recommended that the documentation include a table of contents, consistent pagination and cross-referencing, perhaps an index, and an introduction that summarizes the DTSC waste classification process in toto, and the layout of the documentation, with cross-references.

DTSC's proposal intends to establish a waste classification system that is more flexible than the one currently used. For example, the use of three classifications of wastes instead of two provides additional flexibility in how wastes are handled. It is not economically reasonable to treat all hazardous wastes as if each one posed the same risk to humans or the environment, and even among the hazardous wastes there are degrees of hazard and risk. Different requirements for the disposal of the three waste types should allow flexibility in the stringency of the waste-disposal requirements. However, in practice, the one ''bright line'' currently used to separate hazardous from nonhazardous waste is usually a "fuzzy line" due to the uncertainties associated with the calculation of the value for the line. Therefore, it is not clear if one could truly distinguish two separate values for classification of wastes without an overlap of the confidence intervals around the lines.

However, inadequate consideration has been given to the need for flexibility in design and implementation of the regulations, given the high variability in the conditions to which the regulations are to apply, and the changes that are likely to occur in the scientific underpinnings of the regulations. Indeed, in some instances the proposal appears to make it more difficult to respond to such changes. DTSC needs to clearly articulate the advantages and disadvantages of the proposed system and to demonstrate where the proposed system provides greater scientific and regulatory flexibility and who will be affected by it.

For example, the proposal appears to lock in place a list of 38 chemicals of concern that DTSC apparently created some years ago. No

criteria for selection of these chemicals are given in the DTSC documentation (DTSC 1998a); however, during the public meetings, DTSC did indicate that these chemicals were selected solely for historical reasons. The list of chemicals for which TTLCs and SERTs are being proposed is very constricted (see Chapter 1). Although economically important at one time, the production and use of many of the organic chemicals (most of which are pesticides, such as DDT) have ceased for several years. Therefore, although still environmentally relevant because of their persistence in contaminated materials, such as soils and sediments, they are unlikely to be present at significant concentrations in current industrial wastes. The list of TTLCs and SERTs does not include values for other potentially hazardous materials, such as the xylenes or phthalates, that could and do commonly occur in wastes.

Although one of DTSC's goals is to regulate hazardous chemicals other than those on the TTLC and SERT lists, there is currently no proposed methodology to add additional chemicals to those lists. The documents submitted to the committee refer to a list of 800 chemicals, but there is no indication of whether the risks of these 800 chemicals would be further evaluated, nor of how such an evaluation would be conducted. The application of a limited list of, at most, 38 TTLCs or SERTs to the classification of wastes is a significant flaw in the proposed waste-classification system. Application of the proposed waste-classification system to a much larger list of chemicals would probably require significant modification of the classification methodology to properly account for the wide range of properties of those chemicals.

DTSC should establish a well-defined process for adding substances on its list of regulated chemicals, and for incorporating advances in scientific understanding regarding the risks that these substances might pose. Perhaps DTSC can begin the process of including additional substances before issuing the final regulations because making frequent changes in regulations such as these is expensive and is likely to leave the public confused about what is being regulated and what is not.

The committee is also concerned about how flexible the proposed regulations will be in dealing with special wastes and the particular characteristics of specific waste-management proposals. DTSC has indicated that it intends to provide such flexibility through a variance process. However, variance processes are rarely as responsive, quick, and efficient as anticipated. This is particularly true when no special provisions have been made in the regulations to establish a responsive

process for reviewing and issuing variances. In the absence of such a process, variances are likely to be subject to an often burdensome management process. The relatively few variances that DTSC has apparently issued in the past reinforces this concern about the lack of such flexibility.

The reliance on a variance process appears to introduce another weakness into DTSC's proposal. Variances are usually requested by waste generators or managers when they believe that the characteristics of their waste or its proposed management warrant a relaxation from the stringent regulatory requirements. However, these parties have no incentive to request a variance when the characteristics of the waste or its proposed management warrant a tightening of the requirements. Individuals that might have such an incentive, for instance families living adjacent to a landfill, lack the information and resources to request a variance. Thus, an efficient variance process might adequately respond to situations in which the regulation imposes excessive controls (e.g., reclassification of a hazardous to a special waste) but not in which it provides inadequate controls (e.g., reclassification of a non-hazardous waste to a special waste).

Ultimately, the most important concern regarding the quality of a waste-management system is whether it provides adequate protection to human health and the environment? Other important concerns include: how effective and efficient is its implementation? does its implementation impose excessive delays and other costs on the public? can it be easily understood and followed by those who have to comply with it? and can it be easily enforced? The answers to such questions cannot, of course, be fully known until the proposed regulations have been adopted and implemented. However, fundamental questions such as these should guide the regulatory development.

The committee considered these questions and asked DTSC staff to address them at the public meetings. The DTSC answers were not always clear. To a large extent this is because DTSC apparently distinguished between the current regulations that classify wastes and any future waste-management regulations issued pursuant to this proposed waste-classification scheme. Based on this distinction, the implementa-

tion questions relating to the waste-classification regulations of the program are less important than those relating to the subsequent waste-management regulations. The primary implementation questions regarding the waste-classification procedures are the following:

• Can waste generators and managers clearly understand how they are to classify their wastes (e.g., what tests have to be conducted)?
• Are these waste-classification procedures likely to provide accurate classifications?
• Are these the least costly and most efficient procedures for obtaining the necessary information at the desired degree of accuracy?
• Can compliance with the regulation be quickly and easily determined?

No information has been provided to the committee that would allow it to make judgments on any of the above questions. However, the committee notes the following:

• The distinction between the waste classification regulations and management regulations is apparently not complete. The exposure scenarios necessarily incorporate assumptions on how the wastes will be managed, and thus presume the results of the waste management regulations. Indeed, according to an explanation given in response to a question (DTSC, personal commun., October 9, 1998, see Appendix C, No. 28), DTSC presumes that all hazardous wastes will be disposed to a class I landfill, and all special wastes will be disposed to a class I landfill, or a class II landfill if DTSC issues a variance. Thus, it appears that these classification regulations actually incorporate the waste-management decisions, although no pertinent information about the proposed waste-management regulations was provided by DTSC.
• The analyses and scenarios are limited to the disposal of solid wastes. But the definition of solid wastes, at least in federal law, also includes many liquids. This is also the case in California because the proposed classification system explicitly includes sewage sludge, which has a high liquid content. It is not clear whether the analyses on which the proposed regulations are based are appropriate for liquid wastes. It is clear, however, that the landfill disposal assumptions used to determine the special waste and hazardous waste thresholds are not. Under federal law, most liquid wastes containing hazardous components cannot be disposed to landfills.

• It is not reasonable to assume that the volume of waste classified as non-RCRA hazardous waste will not change significantly under the proposed regulations. Enough questions have been raised by public commenters to call into question DTSC's projected estimates of current versus future regulated-waste volumes (DTSC, personal commun., October 9, 1998, see Appendix C, No. 28, question 15 and attachment 3).

DTSC should give further consideration to the following factors in implementing the regulations:

• Establish a public process for evaluating the effectiveness and efficiency of the regulations. Are the regulations providing the desired degree of protection at the least possible cost?
• Establish incentives and processes within the regulatory scheme for improving the information on which the regulations are based. For example, some of the regulatory levels have been determined by current limits of analytical detection rather than human health or environmental risks. It would be advantageous to create, within the regulatory framework, incentives for improving analytical procedures so that the regulations can become more truly risk based. Without such consideration, the regulations might create strong incentives against such analytical improvements.
• Establish, as discussed above, efficient processes for incorporating new information into the regulations and for allowing the regulations to be tailored to site- and waste-specific characteristics. This is particularly important if the regulations do, in fact, incorporate waste management requirements and are not solely a waste-classification scheme.