1
2	Acknowledgements Several slides and all of the model development described in this presentation reflect a multi-year collaboration involving EPA OSW, EPA ORD (Athens), RTI, HydroGeoLogic (HGL), and consultants. Special thanks to this outstanding team ! Gerry Laniak and Justin Babendreier (ORD, Athens) Jo Ellen Brandmeyer and Robert Truesdale (RTI) Dua Guvanasen and Ted Lillys (HGL) Defne Apul (University of Toledo)
3	Industrial “Waste” Materials and Highway Uses
4	Potential COCs in Industrial “Waste” Materials
5	Risk Assessment and Beneficial Use of Byproducts The Resource Conservation Challenge (RCC) and RCRA 2020 vision requires a broader, systems-level perspective to evaluate potential risks associated with wastes/byproducts along multiple points in the life cycle. The scope of beneficial use questions suggests a national scale risk assessment; however, practical, timely decisions are more likely to be regional in nature, and depend greatly on economics and local/regional risk considerations. Whereas other risk-based programs at OSW such as listing and delisting have well-established methods, models, and data sources; and several states have or are developing guidelines and standards for beneficially used materials, such a national framework does not yet exist for risk assessment of beneficial uses.
6	Overview of Presentation Present a case study example of chat (mine tailings) potentially used in roadway construction. Describe (briefly) new tools/models that will soon be available specific to road construction and land application of byproducts. Present a framework that is beginning to emerge as risk-based approaches continue to evolve to address beneficial use questions. Consider the practical implications of beneficial use problems, what the next steps should be to move us forward, and who should be involved (e.g., HQ, states, regions, other stakeholders).
7	Roadway Construction – Chat Example What is chat? Granular mine tailings (Tri-State Mining Area) Metals above background levels (Pb, Cd, Zn) Washed and unwashed (separates metal-enriched fines) What are the proposed uses? Encapsulated (e.g., aggregate for asphalt, concrete) Unencapsulated (e.g., fill material, gravel) What is the question? Can chat safely be used in roadway construction?
8	Chat – A Tiered Approach Tier I – Review existing information, studies, and open literature to develop conceptual model of exposures and prepare a qualitative characterization of potential risks. Tier II – Based on limitations discussed under Tier I and external peer review comments, conduct focused screening-level modeling of additional aspects of the chat life-cycle. Tier III – Unnecessary, but could have included multimedia modeling, single media modeling, probabilistic modeling ….
9	Tiered Approach and Decision-Making
10	Tier I – Hazard Characterization Reviewed and distilled thousands of pages of reports and studies of unencapsulated and encapsulated uses of chat Chat life cycle from processing to application Physical characteristics (e.g., particle size distribution) Engineering characteristics (e.g., %chat in hot mix asphalt) Leach, total test samples from the lab Some leach, total test results from “field” Superfund RA of residential use of unencapsulated chat Some environmental samples for metals Open literature (e.g., ecological impacts) Considered EPA guidance (e.g., SSLs), and regulatory levels (e.g., MCLs) for comparative purposes.
11	Conceptual Model of Chat Use in Roadway
12	Tier I – Analysis Identified lead, cadmium, and zinc as constituents of concern. Identified potentially complete pathways, specifically, leaching to groundwater. Characterized attributes of chat (particle size) related to potential exposure concentrations. Evaluated sufficiency of data from multiple sources relevant to unencapsulated and encapsulated uses. sampling data? physical characteristics? human health risk? environmental impacts? Compared sample data (e.g., TCLP, SPLP) to appropriate EPA regulatory (TC limits) and total concentrations to non-regulatory (generic SSLs) levels.
13	Tier I – Risk Characterization Determined that unencapsulated uses of chat in roadways could result in potentially significant risks to health and the environment. Determined that, for hot mix asphalt, chat formulations for either “new” or “weathered” asphalt would not result in metal releases at levels of concern. Explained that, although studies on other encapsulated uses were generally unavailable (e.g., asphalt concrete), the metals would likely be tightly bound in these matrices. Identified a potentially significant lack of study data on milling operations of asphalt containing chat, specifically for the air pathway.
14	Tier II Screening Assessment Peer reviewers in general agreement with conclusions, e.g., Unencapsulated uses of chat should be prohibited unless specific studies are conducted demonstrating that risks are de minimis Encapsulated uses, specifically HMA, pose de minimis risks via the groundwater pathway; however, an analysis of milling operations of the chat life cycle is needed. EPA determined that existing regulations were sufficiently protective of workers for milling operations, so we focused on residential exposures. The exposure scenario for milling operations included direct inhalation of chat from street sweeping operations following milling direct inhalation of emissions from temporary storage waste piles Incidental ingestion of soil from air deposition from the above
15
16	Tier II – Screening Approach Lead exposures were assessed for children using IEUBK; this includes all default background exposures (e.g., water) and uses the predicted Pb soil concentration and air concentration from milling operations Zinc and cadmium were assessed for ingestion of soil by comparing Superfund soil screening levels to predicted soil concentrations Cadmium assessed for direct inhalation (cancer and noncancer) Zinc was not assessed for direct inhalation because no EPA-approved benchmarks were identified
17	Tier II – Inhalation Screening The inhalation screening was conducted in six basic steps Select appropriate air deposition model (SCREEN3) Parameterize scenario, model (e.g., geometry, dimensions, metal concentrations, health benchmarks) Determine emission rates from street sweeping Determine emission rates from storage pile Loading/unloading and windblown Run SCREEN3 deposition model to estimate maximum air concentrations Calculate exposure/risk using current EPA inhalation risk methodology (e.g,. EPA Region/ORD workshop, 2003)
18	Tier II – Ingestion Screening Following the inhalation modeling, the ingestion screening was conducted in six basic steps (used simple spreadsheet) Parameterize scenario, model (e.g., geometry, dimensions, metal concentrations, soil screening concentrations) Calculate total mass of metal in the volume of roadway that is milled Calculate the total emissions from the storage pile and roadway sweeping that occur over these respective source areas Assuming that emissions occur for one week, determine the percent loss of metals released Calculate the metal concentrations in the buffer soil assuming that all metal emissions are deposited in the buffer area Compare predicted metal concentrations in the buffer soil to Superfund SSLs and Western U.S. background concentrations
19	Tier II – Lead Screening The lead screening was a multi-source, multi-pathway screening conducted using IEUBK. Using the maximum air concentration and soil concentration for lead estimated with SCREEN3 and the mass balance approach for soil, the IEUBK was run at default values for all other parameters, including the target blood lead level of 10ug/dL. Note that this is for children only.
20	Tier II – Risk Characterization Inhalation Highest noncancer HQ of 4.1E-03 for the sweeping scenario at a receptor distance of 39 m based on a Cal REL of 0.02 ug/m³. Highest cancer risk of 2.1E-09 predicted for the sweeping scenario at a receptor distance of 39 m based on the IRIS Air Unit Risk of 1.8E-03 (ug/m³)^-1. Maximum concentrations for loading/unloading operations for the storage pile scenario were at a distance of 76 m and somewhat (~with a factor of 10) below the sweeping scenario estimates. Maximum concentrations for windblown emissions from the storage pile were also at a distance of 76 m and were several orders of magnitude below the estimates for loading/unloading operations.
21	Tier II – Risk Characterization Ingestion Predicted concentrations of metals in the buffer soils were 37.6 (Zn), 3.2 (Pb), and 0.2 (Cd) mg / kg soil. All concentrations were well below the respective Superfund SSLs of 23,000 (Zn), 400 (Pb), and 70 (Cd) mg / kg soil. All concentrations were below the median background concentrations in the Western U.S. of 55 (Zn), 20 (Pb), and 0.4 (Cd) mg / kg soil presented by EPA (Eco SSLs, 2003).
22	Tier II – Risk Characterization Lead Using the maximum air concentrations at 39 m (3.5 ug/m³) and the predicted Pb concentration in buffer soil of 3.2 mg / kg soil, the highest blood lead levels were predicted for children in the 1-2 and 2-3 age cohorts (4.9 ug/dL). For the default scenario in IEUBK at the background Pb soil concentration of 20 mg / kg soil and the screening air concentration of 3.5 ug/m³, the IEUBK predicts a blood lead level of 5.1 ug/dL for these same age cohorts. The resulting blood lead levels for the IEUBK default scenario that incorporates the screening air and soil concentrations from the milling scenario were below the current target blood lead level of 10 ug/dL. (Note that the target level is likely to change.)
23	BUT, what if chat failed the Tier II screening assessment??? As suggested above, Tier III was not yet defined even though the intent was clear (i.e. to increase the level of resolution and, therefore, decrease the level of conservatism). Next steps would have been developed “on-the-fly,” and would likely have involved existing tools (e.g., ISC3, AERMOD, CRAMM) and/or tools under development such as 3MRA.
24	Screening Tools and Models The Line Source Model (LSM) is under development as part of the FRAMES-3MRA modeling system state-of-the-science, site-based model core is based on the general soil column model (GSCM) has the capacity to handle multiple layers in road construction simulates changes in roadway due to wear (e.g., cracking) includes major loss terms (e.g., particulate emissions) supports connectivity among compartments can be run as a stand-alone model or plug into 3MRA
25	Line Source Model – Conceptual Approach Line Source Model (LSM) developed as a source term for multi-pathway transport of chemical constituents Three major components pavement column compartment shoulder/median column compartment drainage compartment Heterogeneity allowed in each compartment (multiple layers) Mass transfer between compartments is permitted Flexible: one- or two-dimensional representation
26	LSM – Simulates Key Processes Variably saturated flow and transport in fractured porous media (potentially multi-dimensional) Infiltration from runoff and precipitation Advection and dispersion (aqueous phase) Interphase mass transfer (sorption and volatilization) Degradation and losses (photolysis, hydrolysis, etc.) Speciation reactions (for metals, ongoing research) Change in environment due to chemical reactions Deposition of metals and organics on highway surface Source of wind-borne contaminated particulates
27	LSM – Allows InterCpt Communications and IntraCpt Heterogeneity
28	LSM – Simulates Losses from Road Wear
29	LSM – Segmental Application to Roadway
30	LSM Application Currently in verification/testing phase to ensure robustness of numerical solution under wide range of conditions. Fully-tested version should be available in the March-April time frame as a stand-alone model that can also feed FRAMES-3MRA. Pilot study is planned to validate LSM predictions using available research and study data, a key step to ensure relevance of the results to real-world problems. Development of input files will allow the stand-alone version to predict leachate fluxes from the roadway; this could feed single models like EPACMTP or AERMOD or …. Enter FRAMES-3MRA …..
31	FRAMES-3MRA Multi-media, Multi-pathway, Multi-receptor Risk Assessment modeling system originally designed to support national-scale analyses using site-based information on environmental characteristics. Extensive database includes over 700 parameters, including geospatial characteristics (e.g., soil type), chemical/physical properties (e.g., solubility), exposure factors (e.g., water consumption rates by age), census data, and health and ecological benchmarks. Extensively peer reviewed and reviewed by EPA’s Science Advisory Board, the 3MRA modeling system represents a fully integrated assessment tool capable of analyses at multiple scales. The 3MRA system is continuing to evolve into a user-friendly application that will be available in an open source environment that provides extensive tools to populate the complete database (e.g., DATA4EM).
32	3MRA Core Model
33	FRAMES-3MRA Designed to Support a User Community of Risk Assessors
34	DATA4EM Provides Significant Data Collection Capabilities to Feed 3MRA
35	Application of 3MRA-LSM to Beneficial Use RA (Human) Greater than A% of the people living within B distance of the roadway with a risk/hazard of C or less (Ecological) Greater than D% of the habitats within E distance of the roadway with an ecological hazard less than F (Regional) For G% of similarly constructed roadways in a specific area, region, or state (could also be rolled up nationally) (Uncertainty) With confidence H% accounting for empirical model input uncertainty, and confidence I% accounting for output sampling precision error.
36	3MRA Produces an Explicit Estimate
37	So, what is emerging? There is a continuum of technical options available to evaluate and characterize potential risks associated with byproducts used in road construction or other applications, from qualitative review to state-of-the-science models (aka FRAMES-3MRA). For the chat assessment, a weight-of-evidence approach and focused screening-level modeling were sufficient to support a high level of confidence in the decision. So, pulling out the “big guns” was not necessary or even desirable. Perhaps the chat framework could be expanded and serve as a template for identifying next steps and determining best practices in the risk assessment of beneficial uses.
38	An Emerging Framework with Explicit Criteria ??
39	Practical implications for risk assessment? A tiered framework as described for the chat analysis has the advantage of being flexible, but does not lend itself immediately to a standardized approach. What are the tradeoffs and how do we find the right balance point between flexibility and practicality? How do we coordinate/integrate efforts by states, regions, and other stakeholders to shape the risk assessment framework into a practical approach for a wide variety of beneficial use questions? Would the investment in model development and application to create regional scenarios to apply to specific beneficial use questions (e.g., roadways in North Carolina) provide a solution that his sufficiently flexible, or should it remain a Tier III option?