USGS Home
Contact USGS
Search USGS

U.S. GEOLOGICAL SURVEY
Scientific Investigations Report 2007–5106

Back to Table of Contents

Guidance on the Use of Benchmarks in Screening-Level Assessments

In screening-level water-quality assessments, estimates of potential contaminant exposure (concentrations or concentration statistics determined from measurements of contaminants in ground water or surface water) are compared to water-quality benchmarks derived for the protection of human health. Screening-level assessments provide an initial perspective on the potential significance of detected contaminants to human health and help prioritize further investigations (Toccalino and others, 2006b). They also provide a perspective on where adverse effects are more likely to occur and what contaminants may be responsible for those effects (Gilliom and others, 2006). Screening-level assessments are not designed to evaluate specific effects of contaminants on human health, and are not a substitute for comprehensive risk assessments, which generally include many additional factors, including multiple avenues of exposure (Toccalino and others, 2006b).

Initial guidance on the use of benchmarks for evaluating water-quality data in the context of human health was provided in a state-scale pilot study (Toccalino and others, 2004). This report updates that guidance, which also is periodically updated on the HBSL website (Toccalino and others, 2006a).

Selection of Benchmarks

In screening-level assessments, contaminant concentrations or concentration statistics for regulated contaminants—those with USEPA Maximum Contaminant Levels (MCLs) (sidebar 3)—are compared to their MCLs as benchmarks, and concentrations of unregulated contaminants—those without USEPA MCLs—are compared to their HBSLs as benchmarks, when available (Toccalino and others, 2003). For local and state-scale water-quality assessments, concentrations of contaminants that are regulated by the USEPA and (or) the state in which the assessment takes place, are compared to USEPA and state drinking-water standards (Toccalino and others, 2003).  State MCLs, when they exist, are equal to, or more stringent than, USEPA MCLs.

The most current USEPA MCL values are published in the Code of Federal Regulations (U.S. Environmental Protection Agency, 2002) and in periodic USEPA compilations of drinking-water contaminants and their MCLs (U.S. Environmental Protection Agency, 2006a and 2006c). State MCLs typically are published on state drinking-water program web sites.  The most current HBSL values are posted on the HBSL web site (Toccalino and others, 2006a).

Comparison of Contaminant Concentrations to Benchmarks

HBSLs, as well as most MCLs, are maximum contaminant concentrations that are not expected to cause adverse health effects over a lifetime of exposure to drinking water (sidebars 1 and 3). Therefore, contaminant concentrations or concentration statistics indicative of long-term exposure are most appropriate to compare to MCLs or HBSLs in most applications. In ground water, contaminant concentrations tend to change slowly over time, and therefore it is appropriate for the purpose of screening-level assessments to compare ground-water contaminant concentrations measured in individual well samples to MCLs or HBSLs. Examples of ground-water assessments for which the frequency distribution of contaminant concentrations was compared to MCLs or HBSLs are provided in various USGS reports (see figures 4 and 5 in Toccalino and others (2004) and appendix 7 in Zogorski and others (2006)). Such analyses indicate the proportion of detections that are greater than, or within a certain fraction of, MCLs or HBSLs, and also may present the number of samples collected and detection frequencies for each contaminant.

In contrast to concentrations in ground water, surface-water contaminant concentrations generally change substantially over time and have strong seasonal patterns. For screening-level assessments of surface water, annual or long-term mean (average) concentrations (determined from multiple samples over a period of time and time-weighted) generally are most appropriate for comparison to MCLs or HBSLs because mean concentrations provide a more reliable indication of long-term exposure than concentrations from individual samples. If surface-water data are insufficient for the calculation of a reliable mean concentration, then comparison of contaminant concentrations from individual samples to MCLs or HBSLs can be a useful first step if caution is exercised (see the section “Evaluation of the potential significance of contaminant occurrence to human health”). Chapter 6 of NAWQA’s national assessment of pesticides in streams and ground water (Gilliom and others, 2006) provides examples of surface-water assessments where annual mean concentrations of pesticides are compared to water-quality benchmarks for human health.

When evaluating the potential significance of contaminant occurrence data to human health, the analytical minimum reporting level for each contaminant should be less than its MCL or HBSL. This ensures that the laboratory methodologies are adequate to detect concentrations relevant to human health. If the minimum reporting level for a contaminant is greater than its MCL or HBSL, then (a) the contaminant may be present at a concentration greater than a benchmark but not be detected (U.S. Environmental Protection Agency, 1989b), and (b) there is greater uncertainty in evaluating the contaminant concentration or concentration statistic in the context of human health (Toccalino and others, 2004).

Benchmark Quotient (BQ) values can be calculated to aid in evaluating water-quality data in the context of human health.  BQ values are ratios of the contaminant concentrations or concentration statistics to their respective MCLs (for regulated contaminants) or HBSLs (for unregulated contaminants).  For unregulated carcinogens, BQ values are first calculated using the low end of the HBSL range, which corresponds to a 10^-6 cancer risk. If a BQ value for a carcinogen is greater than 1 using the low end of the HBSL range, then a BQ value using the high end of the HBSL range, which corresponds to a 10^-4 cancer risk, also is calculated. BQ values are rounded to one significant figure (Toccalino and others, 2004). Figures 6, 7, and 8 in Toccalino and others (2004) show examples of how distributions of BQ values can be graphed to interpret water-quality findings for a state-scale ground-water assessment.

Evaluation of the Potential Significance of Contaminant Occurrence to Human Health

Benchmark Quotients are useful in screening-level assessments for determining the potential significance of contaminant occurrence in water to human health (table 4). A BQ value greater than 1 signifies a contaminant concentration of potential human-health concern if the computed BQ value is indicative of a long-term average concentration and if the water were to be ingested, without treatment, over a lifetime as the primary drinking-water source (Toccalino and others, 2006b). The likelihood for adverse effects generally increases as a contaminant concentration increases above its benchmark (and its BQ value increases above 1) (Gilliom and others, 2006). Contaminants with concentrations or concentration statistics of the greatest potential human-health concern typically are those that both have BQ values greater than 1 and are frequently detected.  Drinking-water standards (MCLs) are not violated, however, if concentrations of regulated contaminants are greater than MCLs (BQ values are greater than 1) in ground-water or surface-water samples, because samples collected by the USGS are not collected for compliance with the Safe Drinking Water Act (Toccalino and others, 2006b).

Ingestion of water containing a contaminant with a BQ value less than 1 is unlikely to result in adverse human-health effects resulting from that contaminant alone (table 4) (Toccalino and others, 2006b). For surface-water assessments, in particular, if only one or a few samples are available and data are insufficient to calculate a reliable mean concentration, BQ values less than 1 do not lead to a definitive screening-level conclusion because the probability is low that a small number of samples would include high contaminant concentrations that occur infrequently.

A BQ value greater than or equal to 0.1 can be used to identify compounds that may warrant additional monitoring (table 4) (Toccalino and others, 2006b).  Although a variety of BQ values may be selected as threshold values to identify contaminants that may warrant further monitoring, a threshold BQ of 0.1 is consistent with various state and Federal practices (for example, see U.S. Environmental Protection Agency, 1998). Regular monitoring of such contaminants would enable analysis of trends in their occurrence and provide an early indication of contaminant concentrations that approach MCLs or HBSLs (Toccalino and others, 2006b).

Interpretation of the potential significance of contaminant occurrence to human health also should consider the type of water sampled.  For example, ground water from domestic wells, which typically is used by a single family, often is consumed with little or no treatment. Ground water from public-supply wells and surface water from water-supply intakes usually is treated or blended before distribution and (or) consumption, potentially reducing contaminant concentrations, but often is used by many people.  Ground-water data from monitoring wells are not directly relevant to human health because this water is not consumed, but contamination in monitoring wells can indicate the potential for future contamination in deeper aquifers used for drinking-water supplies (Toccalino and others, 2004). Likewise, surface-water samples that are not collected at water-supply intakes are not directly applicable to drinking-water supplies. Results from such surface-water samples, however, can be used to provide a perspective on the potential significance to human health for source water with similar watershed land uses (Gilliom and others, 2006), or if the sampled water body is considered a potential future source of drinking water.

If information about the potential health effects associated with specific contaminants in drinking-water supplies is needed for a particular water-quality assessment, such information may be obtained from sources such as the USEPA and the Agency for Toxic Substances and Disease Registry (ATSDR). the ATSDR is a federal public-health agency of the U.S. Department of Health and Human Services. The USEPA publishes fact sheets for many compounds with MCLs (U.S. Environmental Protection Agency, 2006c) and the ATSDR publishes Toxicological Profiles and ToxFAQs (summaries about chemical exposure and the effects of exposure on human health) (Agency for Toxic Substances and Disease Registry, 2007a; Agency for Toxic Substances and Disease Registry, 2007b).

Limitations of Screening-Level Assessments

Screening-level assessments are a useful first step toward evaluating contaminant occurrence in the context of human health, but such assessments have limitations. For example, MCLs and HBSLs generally are developed using toxicity information for single contaminants, whereas it is common to detect contaminant mixtures in ground water and surface water. The long-term cumulative effects of low concentrations of multiple contaminants on human health currently are unknown for most contaminant combinations. An additional limitation is that HBSLs have not been developed for some unregulated contaminants because of a lack of toxicity information; the potential human-health significance of the occurrence of contaminants without benchmarks cannot, therefore, be evaluated at this time (Toccalino and others, 2006b). It is, however, useful to identify those unregulated contaminants that most frequently occur in a given assessment, but do not have HBSLs. As improved data on toxicity and environmental concentrations are developed, HBSLs and exposure estimates can be updated, and screening-level assessments can be improved and expanded. the USGS works closely with the USEPA to assist them with incorporating water-quality findings into their risk assessments (Gilliom and others, 2006).

Examples of HBSL Applications

Several published studies have used HBSLs to help provide an initial perspective on the potential human-health significance of contaminant concentrations in water. These studies also provide examples of analyses and language applied to the use of HBSLs and MCLs as water-quality benchmarks:

• National assessment of VOCs in ground water (Zogorski and others, 2006) and a corresponding Fact Sheet about what those findings may mean to human health (Toccalino and others, 2006b).
• National assessment of pesticides in streams and ground water (Gilliom and others, 2006). This assessment did not use HBSLs, but it contains examples of the application of MCL, Lifetime HA, and Cancer Risk Concentration values as water-quality benchmarks in the context of human health.
• State-scale analyses of the occurrence of anthropogenic organic compounds in ground water and finished water of community water systems in Nevada (Rosen and others, 2006) and the northern Tampa Bay area in Florida (Metz and others, 2006).
• Fact Sheet providing background on the HBSL effort (Toccalino and others, 2005).
• State-scale analyses that applied HBSLs to ground-water quality data in New Jersey (Toccalino and Norman, 2006; Toccalino and others, 2004). Both of these documents use HBSLs calculated using the original methodology, and not the revised HBSL methodology described in this report.
• Original HBSL methodology and history of the HBSL effort (Toccalino and others, 2003).
• National assessment of VOCs in ground water (Zogorski and others, 2006) and a corresponding Fact Sheet about what those findings may mean to human health (Toccalino and others, 2006b).

Back to Table of Contents