3.0 Screening Assessment
Screening is the process by which the universe of COPCs and exposure pathways is evaluated and those that are below a threshold of concern are eliminated from further evaluation. Proper screening is critical because it focuses resources on key site issues, optimizing risk assessment and risk management.

To avoid eliminating potentially relevant exposures, screening is deliberately designed to be inclusive and overly conservative. For this reason, the incorporation of bioavailability at the initial screening stage of sediment evaluation is limited. However, a brief section on screening has been included here because of the importance of this step in the overall site evaluation process. Where bioavailability-related factors can be considered, these are highlighted. Note that screening typically occurs in a tiered fashion, with site-specific considerations integrated as the initial component of pathway evaluation. These types of more focused screening tools are covered in the pathway-specific chapters of this document, Chapters 4–7.

The screening process for sediments occurs when the initial COPC selection process and CSM development are complete. Thus, COPCs and pathways have generally been identified. The purpose of the screening process is therefore to assist the user in determining the need for, and nature of, further investigation and to facilitate further refinement of the CSM and COPC list.

As discussed in Section 2.2.3, COPCs are initially selected based on general knowledge about the site’s current and historic activities. This initial list of COPCs is distinguished from final list of chemicals, which may be subject to further detailed evaluation. Screening is the process by which the universe of chemicals that are known or suspected to be released from a site are reduced to a final list of COPCs that require further evaluation.

Screening is often treated as a mechanical process structured to err on the side of inclusion to provide conservatism. In practice, however, this approach tends to pull in contaminants that are secondary or may even be insignificant. As the number, complexity, and cost of sediment assessment tools rapidly expands, the most important site contaminants need to be identified so that finite resources can be optimally allocated. Thus, the screening process is a critical step in sediment assessments and must be undertaken with care and judgment. Refining the COPC list increases the efficiency and relevance of subsequent studies and thus contributes to better risk-management decisions.

Text Box 3-1. Wisconsin Department of Natural Resources Sediment Quality Guidelines (2003)

Nonpolar Organic Compound
and TOC Relationships

In the case of nonpolar organic compounds such as PAHs, PCBs, dioxins/furans, and chlorinated pesticides, the bulk sediment concentrations can be normalized to the TOC content for site-to-site comparison purposes by dividing the dry weight sediment concentration by the percent TOC in the sediment expressed as a decimal fraction. This normalization helps account for differences in contaminant bioavailability among sites due to varying sediment concentrations of TOC.

Example

TOC-normalized PCB concentration of 7 mg/kg with 3.5% TOC is 200 mg PCB/kg TOC (i.e., 7 mg PCB/kg ÷ 0.035 kg TOC/kg = 200 mg PCB/kg TOC).

The Wisconsin sediment quality guidelines for total PCBs range 6–67.6 mg/kg TOC.

Normalization of nonpolar organic compounds to TOC content is valid only if the TOC content in the sediments is >0.2%. At TOC concentrations <0.2%, other factors that influence partitioning to the sediment pore waters (e.g., particle size and sorption to nonorganic mineral fractions) become relatively more important (Di Toro et al. 1991). This number is then compared to the appropriate screening level.

Although screening benchmarks do not typically account for bioavailability, some screening levels do consider binding properties that affect the degree of exposure associated with a particular bulk sediment concentration. For organic compounds, concentrations may be normalized to TOC content; extractable metals concentrations can be normalized to AVS. These tools are summarized in Section 3.1 and described in more detail in Chapter 4. TOC-adjusted screening levels can be applied within the screening process under many state regulatory programs (e.g., Text Box 3-1).

3.1 Screening Approaches
There are two major categories of screening. The first compares site-related concentrations to concentrations that are representative of prevailing conditions in the site area. Background evaluations may not always be successful at eliminating nonsite contributions because of specific provisions of the applicable regulatory framework, which may prohibit these considerations or require consideration of total risk. The second involves comparison of observed concentrations to one or more sets of reference values. Typically, these sets of benchmarks are specified by the overseeing agency. In some cases where benchmarks are not available, it may be appropriate to seek literature.

Figure 3-1 provides an overall flowchart for the screening process. State environmental agencies are encouraged to provide as much flexibility as possible to enhance the value of the screening process.

The following discussion provides an overview of the screening benchmarks commonly in use; these benchmarks typically apply only to benthic endpoints (i.e., protection of benthic invertebrates). Additional screening approaches for other endpoints and pathways appear later in this document where the respective pathways are discussed.

Figure 3-1 Sediment screening process.

3.1.1 Background Evaluations
A critical step in refining a COPC list, which can also be useful at later stages of site assessment, is the determination of background values. By definition, COPCs are chemicals that are present due to some release that results in concentrations that are distinct from prevailing conditions. Therefore, the process of identifying COPCs must include a mechanism to distinguish elevated concentrations from background. This step should occur early in the site assessment, before other types of screening techniques are applied.

The process of determining a background sediment concentration can be complex and depends on the type of contaminant, the similarity between the background area sediments and site sediments, and confirmation that site sources are not negatively affecting the area(s) used to establish background concentrations. Consideration of background is important at several stages of the site evaluation process. In addition to serving a screening role, characterization of background is necessary for the appropriate identification of reference/control areas in almost any kind of toxicity study. In the risk management phase, the sustainability of a proposed mitigation must be evaluated in the context of background concentrations of COPCs. Background samples can be collected from any sediment not affected by site activities but should include areas with similar physical, chemical, geological, and biological characteristics as the site being investigated (USEPA 2005a). Some agencies will allow the elimination of a COPC by comparing it to background. Be sure to confirm the requirements with your regulatory agency.

When used in sediment characterization studies, the term “background” refers to both the concentrations of COPC that are not a result of the activities at the site undergoing assessment and the locations of the background areas (MacDonald and Ingersoll 2002). USEPA and many states recognize two types of background. According to USEPA (MacDonald and Ingersoll 2002), compounds are naturally occurring or anthropogenic based on the following definitions:

• Naturally occurring background compounds are present as a result of geochemical processes that have not been influenced by human activity. Naturally occurring organic and inorganic background chemicals in soil and groundwater are attributable to the natural physicochemical characteristics of the area.

• Anthropogenic background compounds are synthetic constituents or natural substances that have been released to the environment as a result of human activities that are not related to specific activities conducted at the site. USEPA (1991) considers the following to be sources of anthropogenic background compounds: agricultural runoff, septic systems, irrigation (former agricultural and residential application of pesticides), air pollution, historical industrial discharges and landfills not associated with the site, and urban pollution (lead and PAHs from automobiles and combustion process).

Users should investigate whether their state and/or USEPA region has different definitions and requirements for assessing background conditions as part of environmental site assessments.

The selection and analysis of background sediment samples is important in the determination and use of bioavailability within the site characterization process. Background comparisons often necessitate analysis of hydrodynamic conditions as well as the fate and transport potential of COPCs at the investigated site. Guidance for Comparing Background and Chemical Concentrations in Soil for CERCLA Sites (USEPA 2002b) provides guidance on the determination of background concentrations and how to compare these to site concentrations. Although this document was originally intended for soils, it also has applications for sediments.

DQOs should be established and followed when preparing a background comparison (USEPA 2006a). Regardless of the intended use, it is critical to determine data needs prior to conducting the background assessment. This document is not intended to provide a recipe for conducting these types of investigations. The references and links provided in this section will inform and guide the user in making the appropriate decisions regarding background for a particular scenario.
It is up to the user to decide what type of background evaluation is necessary for the scope of the assessment and whether the collection of site-specific data is justified. The user must also decide whether the background evaluation is necessary only as a screening step or whether the data will also serve in the scoping of later evaluations, such as identifying a reference/control area for, say, a toxicity study. In cases where site-specific background data are not available or practical to obtain, regulatory agencies may allow use of representative concentrations obtained from the literature or agency sources, e.g., National Oceanic and Atmospheric Administration (NOAA) Screening Quick Reference Table (SQuiRT) values (see http://response.restoration.noaa.gov/environmental-restoration/environmental-assessment-tools/squirt-cards.html).

3.1.2 Numeric Screening Criteria for Sediments
A number of agencies have established sets of quantitative screening values for contaminated sediments, including USEPA (federal), USEPA regions, foreign environmental agencies, NOAA, U.S. Department of Defense agencies, Oak Ridge National Laboratory (ORNL), and some state environmental agencies. Screening values vary widely and can apply to freshwater or marine but not both. In most cases, states have adopted one set of values; however, some states (e.g., Oregon) apply more than one screening value per analyte to address multiple endpoints (i.e., receptors). Other agencies provide tables containing multiple ecological screening values (see http://www.esd.ornl.gov/programs/ecorisk/tools.html).

The USEPA (2001a) states that, since screening values “are based on conservative endpoints and sensitive ecological effects data, they represent a preliminary screening of site contaminant levels against these screening values to determine if there is a need to conduct further investigations at the site. Ecological screening values should not be used as remediation levels.” Although these measures alone might not always accurately reflect risk to the environment, SQGs have been developed as numerical chemical concentrations intended to be either protective of biological resources, predictive of adverse effects to those resources, or both (Wenning and Ingersoll 2002).

3.1.2.1 Screening values in relation to bioavailability
There are generally two types of screening values currently in use by states and other regulatory agencies: site specific and benchmark based (i.e., established screening levels). Site-specific values generally capture critical variables related to bioavailability and rely on back-calculation to sediments from identified ecological or human health endpoints to determine the need for further evaluation. This type of screening is not common as an initial step because it is labor-intensive and costly, although it may be applied later in the process.

The vast majority of agencies use benchmarks for screening purposes. Most benchmarks are based on bulk sediment concentrations (total individual COPC concentrations in sediment). Benchmark values are generally based on observational studies that correlate biological effects (i.e., biological response) with concentrations (i.e., exposure) in samples potentially affected by multiple chemical and physical stressors. These benchmarks do not provide causal links between individualcontaminants and effects but do indicate whether a potential correlation exists between individual contaminants concentrations in sediments and effects on benthic organisms. Some sets of benchmark values do include some laboratory-derived values representing effects of a single contaminant on a single species (such as spiked sediment bioassays) or are based on predicted interstitial concentrations of specific contaminants that have been studied under controlled laboratory conditions (such as USEPA’s equilibrium sediment benchmarks [ESBs]). For the most part, screening benchmarks are not valid for predicting effects associated with individual contaminants. They can, however, be used as conservative, lower-bound estimates of individual contaminant concentrations in sediment that are likely to cause effects on benthic organisms.

Benchmark values are typically based on effects to benthic organisms; however, some state agencies have benchmark values for multiple potential receptor groups, including aquatic invertebrates, wildlife, plants, birds, and humans. Bioaccumulation- or food web–based benchmarks are typically calculated on an individual contaminant basis. Approaches using these types of benchmark values are discussed in more detail within Chapters 6 and 8.

All benchmarks are by nature simplistic since they are intended to apply to a universe of sites with vastly different characteristics. Factors generally not considered include the following:

• site setting (i.e., heterogeneous populations, dietary changes, weather conditions)
• multiple chemical exposures (i.e., distinguishing effects of individual contaminants or considering additivity/synergism)
• sediment physiochemical characteristics (i.e., changes in OC, redox, pH, etc.)

As a result, the screening process rarely incorporates site-specific bioavailability of COPCs.

Most of the published sediment screening tables are cross-referenced to two basic sets of values: the lowest effects levels (LELs) and severe effects levels (SELs) published in Guidelines for the Protection and Management of Aquatic Sediment Quality in Ontario by the Ontario Ministry of the Environment (https://projects.itrcweb.org/contseds-bioavailability/References/guide_aquatic_sed93.pdf, Persaud, Jaagumagi, and Hayton 1993) and the effects range low (ER-L) and effects range median (ER-M) values published by Long et al. (1995). While it has been argued that these screening levels were partly derived from laboratory aquatic toxicity studies (some using sediments spiked with the COPC of interest), none of these benchmarks account for site-specific conditions that may influence the bioavailability of individual contaminants. In fact, sediment screening levels are derived to conservatively predict the absence of a toxic effect and generally do not provide adequate information to predict toxic effects. The prediction of toxicity is better addressed through the incorporation of bioavailability in subsequent stages of the site investigation. A brief discussion of the background and basis of the most commonly used sets of screening benchmarks for sediments, including LELs/SELs, ER-Ls/ER-Ms, and the USEPA ESBs, is presented below.

3.1.2.2 Screening values commonly used by the states and federal agencies
A variety of sources of screening benchmarks are applied by states and other jurisdictions for evaluation of contaminated sediments; however, many cite the same basic sets of screening levels or other compilations that are cross-referenced to these screening levels. The following provides a brief basis and background summary of the most popular benchmarks. The use of a particular set of benchmarks should consider the basis and background of that set of benchmarks in determining the weight to assign to the result provided by the benchmark value.

Lowest Effects Levels and Severe Effects Levels. Persaud, Jaagumagi, and Hayton (1993) described a methodology for developing LELs and SELs for metals,¹ nutrients, polar organics, and nonpolar organics. The authors reviewed a range of protocols for setting SQGs, including background, EqP, apparent effects threshold (AET), screening level concentration (SLC), and spiked bioassay. To set the LELs and SELs, only the SLC method was considered. This method involves first plotting the frequency distribution of the contaminant concentration across multiple (at least 10) sites where an individual benthic species is present. From this plot, the 90th percentile of the concentration distribution, or species screening level concentration (SSLC), is estimated; that is, 90% of sediment concentrations where the species was observed are lower than this SSLC. Then the 90th percentiles for all species are plotted in order of increasing concentrations. The LEL and SEL are set at the 5th and 95th percentiles, respectively, of this distribution. Nonpolar organics are normalized to OC, thus incorporating a component of the EqP process.

The application of the SLC approach in setting chemical-specific criteria was detailed by Jaagumagi (1993), and three sets of companion documents addressing metals, PCBs and organochlorine pesticides, and PAHs ultimately led to the Ontario screening benchmarks (Guidelines for the Protection and Management of Aquatic Sediment Quality in Ontario, https://projects.itrcweb.org/contseds-bioavailability/References/guide_aquatic_sed93.pdf, Persaud, Jaagumagi, and Hayton 1993). The data were derived from several hundred samples collected in and adjacent to the Great Lakes region. The SLC method provides an overview, within the Great Lakes area, of statistical associations between concentrations of COPCs and absence of species but does not address toxicity (causality) due to any specific constituent. The method does not accommodate integration of other variables that are important for benthic organisms, such as dissolved oxygen concentrations or substrate particle size. Furthermore, no consideration was given to the fact that LELs for most of the metals (cadmium, copper, chromium, iron, and nickel) are below the identified background concentrations for these metals identified from the Great Lakes area.

Effects Range-Low and Effects Range-Median Concentrations. The most frequently cited set of benchmarks for saline/estuarine sediment was published by Long et al. in 1995. The criteria were developed by assembling a biological effects database consisting of studies in which both sediment concentrations and adverse biological effects were reported. For each target constituent, concentrations were arranged in ascending order. The ER-L and ER-M were identified as the 10th and 50th percentiles, respectively.

The 10th percentile (ER-L) represents the COPC concentration that was associated with a low probability of effects. The ER-L therefore serves as a useful indictor of a concentration below which toxicity from that contaminant is not likely to occur. However, the ER-L cannot be used to infer whether impacts due to that constituent would occur at a higher concentration since any number of toxicants or conditions could be the cause of observed effects in those samples. The ER-M is the COPC concentration below which 50% of the samples did not exhibit toxic effects. Given the presence of multiple contaminants in the samples used to develop the ER-L/ER-M benchmarks, the inference that the median concentration of any specific contaminant is associated with “probable effects” is not supported. Without a detailed understanding of the toxicity of other COPCs in any sample in conjunction with the physical characteristics of the sample, no conclusion regarding effects from an individual contaminant is possible. Rather, the ER-M is simply the median concentration of a COPC developed from a spectrum of contaminated sediments.

Threshold Effects Concentrations (TECs) and Probable Effects Concentrations (PECs). MacDonald, Ingersoll, and Berger (2000) evaluated several sets of SQGs. The final sets of values include the LECs/PECs and ER-Ls/ER-Ms discussed above, as well as three other sets of values: threshold effects levels/probable effects levels (TELs/PELs) (Smith et al. 1996), minimum effects thresholds/toxic effects thresholds (METs/TETs), and USEPA sediment quality advisory levels (SQALs) (USEPA 2008b). Of these, the METs and TETs were developed using a screening level concentration approach and field effects–based approach, respectively. TEC/PEC-HA28 values are based on 28-day Hyalella azteca toxicity tests from sites affected by multiple contaminants. The SQAL values (used by MacDonald, Ingersoll, and Berger [2000] for only the TECs) were derived based on EqP theory to predict interstitial water concentrations of COPCs. Of the methods above, only the SQAL values consider bioavailability (i.e., through the use of organic carbon normalization), but these values are available for only a limited number of nonionic organic contaminants.

MacDonald Ingersoll, and Berger (2000) derived the TECs and PECs by calculating the geometric mean of three to five other benchmarks. They reported them to three significant figures, suggesting a 99.9% confidence in the validity of these resulting screening levels. In fact, these “consensus” benchmarks are a variable combination of other values and do not have a consistent background and basis. Furthermore, the data sets are not strictly independent, since the TELs were developed using data from datasets that were also used to develop the ER‑Ls/ER‑Ms by Long et al. (1995). The consensus values of MacDonald, Ingersoll, and Berger (2000), as well as other sets of benchmarks, are overlapping in basis and background and ultimately are derived from many of the same datasets.

Overall, “field effects–based” screening levels imply a statistical association/correlation between COPC concentrations and the likelihood of a benthic species presence/effect in samples containing multiple contaminants. However, these types of benchmarks do not reflect causality between any individual contaminant and observed effects. While in situ effects naturally reflect the bioavailability of COPC in sediments, there is no way for the field effects–based benchmarks to determine which of the many COPCs present in contaminated sediment sites may account for the observed effects or whether physical-chemical factors such as dissolved oxygen concentrations or substrate particle size distribution could account for toxicity or species presence/absence. Furthermore, as stated by Long et al. (1995), factors that are typically likely to influence bioavailability, such as grain size, sulfides, and carbon, were not reported in most of the studies used to develop the field effects–based screening benchmarks.

Overall, therefore, the following caveats apply to the use of these field effects–based benchmarks:

• The thresholds values are conservative and useful to help eliminate individual constituents from further consideration.
• The values do not indicate causality between any individual contaminant and biological effects.
• Since contaminants co-occur, “probable” or “likely” effects levels of individual contaminants are not necessarily predictive of effects and should not be used to infer toxicity by individual contaminants. These benchmarks are not appropriate for predicting effects at sites.

Apparent Effects Thresholds. AETs were developed (Beller and Simoneit 1986, Barrick et al. 1988) for the Puget Sound Estuary using four types of biological indicators: amphipod, oyster larvae, Microtox bioassays, and benthic infaunal abundance. Both impacted and nonimpacted stations were considered in establishing these values. Validation of the AETs indicated between 50% and 96% reliability in predicting effects in Puget Sound. However, the authors caution that the values are not necessarily applicable to other aquatic systems. The AETs are sediment management values and not intended as screening levels; however, they do provide a useful reference for concentrations that may associated with effects in estuarine systems.

Equilibrium Partitioning Sediment Benchmarks. USEPA used EqP to develop sediment benchmarks for selected hydrophobic organic chemicals (HOCs, e.g., PAHs, pesticides, etc.). ESBs for metals are based on binding to AVS. The partitioning equations assume that sequestration occurs by partitioning (binding) to OC or sulfides, rendering the contaminant unavailable for biological uptake. ESBs offer several advantages over field effects–based benchmarks because they are contaminant specific, address causal relationships between COPCs and the potential for toxicity, encompass site-specific conditions that affect bioavailability (e.g., binding ligands), and address additivity within contaminant groups (e.g., PAHs). However, ESBs address only direct toxicity. Other endpoints that may be more sensitive, such as bioaccumulation, require additional evaluation.

3.2 Screening Summary
Screening is a critical step in the sediment assessment process but does not offer much opportunity for bioavailability adjustments. It is important to remember that the initial screening process is, in most cases, simply an approach to conservatively and cost-effectively determine whether additional site investigation is necessary. Screening should never be mechanical and must include professional judgment. The initial screening flow chart (Figure 3-1) suggests how users in jurisdictions that require consideration of specific benchmarks can enhance the process to obtain the most focused results and ensure that important endpoints are not obscured. States are encouraged to provide a flexible framework for site screening assessments and to collaborate with assessors in identifying COPCs so that resources can be directed to the most important contaminants.

An effective screening process focuses further evaluation on endpoints and pathways of concern and assists the user in refining the baseline CSM. In cases where benchmarks are set conservatively low and the screening process yields borderline results (marginal exceedances of benchmarks), decision makers need to carefully consider whether the site warrants the dedication of additional resources for further risk evaluation. Bioavailability generally becomes a more important factor and bioavailability assessment costs can be better justified after the initial screening process has determined that additional site investigation is necessary.

¹By convention and for convenience, the nonmetal arsenic is considered as part of the metals group.