«This page left intentionally blank. United States Environmental Protection Agency EPA-540-R-05-012 Office of Solid Waste and Emergency Response OSWER ...»
• Are there likely to be other factors, such as non site-related releases, besides the cleanup that will influence the monitoring results, and are these well understood?
• How often should monitoring take place, and how long should it continue?
• Can the monitoring results be readily placed into searchable, electronic databases and made available to the project team and others?
• Is it clear who is responsible for reviewing the monitoring data and what the triggers are for identifying important trends (positive or negative) in the results?
• What are the most appropriate methods for analyzing the monitoring data? Should these be based on statistical tests or other quantitative analysis? Will there be sufficient data to support these statistical measures?
• Is there agreement on what actions will be taken based on the results of the monitoring data?
• How will the results be communicated to the public, and who is responsible for doing this?
Although sediment sites vary widely in size and complexity, monitoring typically requires a
higher degree of planning than at some other types of sites for the following reasons:
• Sediment sites often involve more than one affected medium (e.g., sediment, surface water, biota, floodplain soils, and ground water) and multiple contaminants of concern;
An especially important issue for project managers at large sites with more than one response action is the need to monitor both the effectiveness of individual sediment actions and the ability of achieving overall site RAOs. Frequently, the monitoring parameters at large sites are different. For example, where contaminants from multiple sources are indistinguishable, it may be necessary to use unique parameters for monitoring effectiveness of individual actions. However, it also may be very important to monitor parameters (i.e., some fish species), which may be responding to multiple sources or areas of a site.
8.2 SIX RECOMMENDED STEPS FOR SITE MONITORING
When developing a monitoring plan, it is important to review the ROD and supporting documents for the site. The ROD generally should contain numerical cleanup levels and/or action levels for sediment and sometimes for other media, and narrative RAOs that relate more directly to reducing risk.
Generally, these form the basis of the monitoring plan. RODs or other site documents may also contain specific performance criteria or objectives for the short-term and long-term performance of the remedy that should be incorporated into the monitoring plan.
EPA’s Monitoring Guidance (U.S. EPA 2004c) describes six key steps that are recommended in developing and implementing a monitoring plan. These steps are listed in Highlight 8-3 and explained briefly along with sediment site examples in the following text. This guidance was developed for use at all hazardous waste sites, not just Superfund sites, and therefore, uses the term “site activity” to apply to implementation of removal actions, remedial actions, ICs, or habitat mitigation.
Step 1. Identify Monitoring Plan Objectives
Generally, the most important element in developing an effective monitoring plan is for the project manager to identify clear and specific monitoring objectives. Identifying appropriate monitoring objectives normally includes examining the intended outcomes of the action and the methods used to achieve that outcome at the site. Inadequate or vague monitoring objectives can lead to uncertainty about why the monitoring is being conducted and how the data will be used. Furthermore, funding for monitoring is often limited. Specifying objectives can help to focus the experimental design and ensure that the most useful information is collected. When identifying monitoring objectives other than those already established in decision or enforcement documents, the project manager should involve participants from all concerned stakeholders (e.g., public, natural resource trustees, state agencies, potentially responsible parties).
8-4 Chapter 8: Remedial Action and Long-Term Monitoring
Step 2. Develop Monitoring Plan Hypotheses • Develop monitoring conceptual models • Develop monitoring hypotheses and questions Step 3.
Formulate Monitoring Decision Rules Step 4. Design the Monitoring Plan • Identify data needs • Determine monitoring plan boundaries • Identify data collection methods • Identify data analysis methods • Finalize the decision rules • Prepare monitoring quality assurance project plans (QAPPs) Step 5. Conduct Monitoring Analyses and Characterize Results • Conduct data collection and analysis • Evaluate results per the monitoring of data quality objectives (DQOs), developed in Steps 1-4, and revise data collection and analysis as necessary • Characterize analytical results and evaluate relative to the decision rules Step 6. Establish the Management Decision • Monitoring results support the decision rule for site activity success S Conclude the site activity and monitoring • Monitoring results do not support the decision rule for site activity success but are trending toward support S Continue the site activity and monitoring • Monitoring results do not support the decision rule and are not trending toward support S Conduct causative factor and uncertainty analysis S Revise site activity and/or monitoring plan and implement
Source: U.S. EPA 2004c
Physical, chemical, and/or biological endpoints should be identified to help evaluate each monitoring objective. In general, physical and chemical endpoints are less costly and more easily measured and interpreted than biological endpoints and, therefore, may be more appropriate where quick decisions are needed. However, the ability of physical and chemical endpoints to quantify changes in ecological risk reliably may be less direct than biological measurements, for example where risk is due to direct contact with multiple contaminants. In this case, toxicity tests or bioassessments may provide an integrated measurement of the cumulative effects of all contaminants and, therefore, can be a better
assessment of ecological risks in some situations. Conversely, where the primary risk is due to humans and wildlife eating fish, chemical endpoints in fish may be most appropriate.
When identifying appropriate endpoints, it is important for the project manager to ensure that the measure employed matches the time frame established for the criteria. For example, acute toxicity tests quantify short-term effects on an organism; therefore, this type of test may be appropriate for operational monitoring (e.g., monitoring during remedial dredging), where it can be performed in a short period of time. Other biological endpoints, such as changes in species diversity, typically occur over long periods of time and may be more appropriate for use in a long-term monitoring program designed to look at ecological recovery. Although no single endpoint can quantify all possible risks, a combination of physical, chemical, and biological endpoints usually provides the best overall approach for measuring risk reduction.
Example: In the ROD, EPA established a RAO of reducing polychlorinated biphenyl (PCB) concentrations in fish tissue to levels that would eliminate the need for a fish consumption advisory for PCBs (for this site, 0.05 ppm). To achieve this objective, EPA selected a cleanup level of 0.5 ppm total PCBs in sediment. The short-term objective of the monitoring program is to monitor PCB concentrations in sediment until the cleanup level is met and the long-term objective of the monitoring program is to monitor PCB concentrations in fish tissue until the RAO is met.
Step 2. Develop Monitoring Plan Hypotheses
Typically, monitoring hypotheses represent statements and/or questions about the relationship between a site activity, such as sediment remediation, and one or more expected outcomes (U.S. EPA 2004c). The development of the monitoring hypotheses is analogous to the problem formulation step (Step 1) of the DQO process (U.S. EPA 2000a). The monitoring hypothesis may be generally stated as “The site activity has been successful in reaching its stated goals and objectives,” or in question form, as “Has the site activity reached its stated goals and objectives?” As described in EPA’s Monitoring Guidance (U.S. EPA 2004c), the concept of a monitoring conceptual model may be helpful in identifying and organizing appropriate hypotheses. This model, frequently a flow chart or graphical display, consists of a series of working hypotheses that identify the relationships between site activities and expected outcomes.
Example hypotheses: The PCB concentration in sediment has reached the cleanup level of 0.5 ppm. The PCB concentration in fish tissue has reached the remedial goal of 0.05 ppm.
Step 3. Formulate Monitoring Decision Rules Once monitoring objectives and hypotheses are agreed upon and stated explicitly, the next step should be to identify specific decision rules that will be used to assess whether the objectives are met.
A decision rule is normally an “if... then...” statement that defines the conditions that would cause the decision maker to choose an action. In a monitoring plan, the decision rules should establish criteria for continuing, stopping, or modifying the monitoring or for taking an additional response action. Four main elements of a decision rule usually are: 1) the parameter of interest; 2) the expected outcome of the 8-6 Chapter 8: Remedial Action and Long-Term Monitoring remedial action; 3) an action level, the basis on which a monitoring decision will be made; and 4) alternative actions, the monitoring decision choices for the specified action (U.S. EPA 2004c).
Another factor the project manager should consider when developing decision rules is the time frame under which they will operate. For example, when dredging highly contaminated sediment, a realtime monitoring program could be established to analyze water samples before proceeding with the next day’s dredging. In contrast, the time frame required to assess a long-term monitoring objective (e.g., to lower fish tissue concentrations) would be longer. In either case, the time frame should be explicitly stated and understood by all the participants.
Examples: A decision rule could be established to require certain actions if suspended sediment or contaminant concentration in the surface water due to releases from dredging exceed certain criteria. A decision rule could be established to assess whether the sediment cleanup level of 0.5 ppm PCBs has been reached, defined as an average of 0.5 ppm PCBs in each of ten grids over the site. A decision rule could be established to assess whether progress is being made toward the remedial action objective of reduced PCB concentrations in fish tissue by establishing an interim goal of achieving 0.8 ppm in fish tissue within five years, after which monitoring frequency will be revisited. PCB concentrations in fish species “A” will be measured on a specific frequency (e.g., annually) that is commensurate with the relevant species’ uptake and depuration rates.
Step 4. Design the Monitoring Plan
The fourth recommended step for the project manager is to identify the monitoring design for collecting the necessary data. Design considerations include identifying data needs; determining monitoring boundaries (frequency, location, duration); identifying data collection methods; and identifying data analysis methods, including uncertainty analysis. EPA recommends that a systematic planning approach be used to develop acceptance or performance criteria for all environmental data collection and use. The Agency’s DQO process is a planning approach normally appropriate for sediment sites (U.S. EPA 2000a). Quality assurance project plans (QAPPs) or their equivalent are also recommended for environmental data collection and use.
The spatial and temporal aspects of a monitoring plan typically define where and when to collect samples. In general, sampling locations should be based on the areal extent and magnitude of the contaminated sediment and the propensity for the contaminants to move, either through transport (e.g., remediation, natural events) or through the food chain. Generally, the more dynamic the conditions, the more frequently sampling is necessary to represent conditions accurately. However, a less costly alternative can be to use data endpoints which respond to cumulative, longer-term conditions, where appropriate. Additional factors that should be considered in establishing sampling locations include locations of baseline or pre-remediation sampling stations and spatial gradients in concentration. For example, generally greater sample density is needed where concentration gradients are high.
Selecting a statistical approach to use in evaluating the data is another important aspect of the monitoring program design. Data are sometimes collected in a manner that is incompatible with or insufficient for the statistical tests used to analyze the data. Although the amount of data needed to compare point-in-time data may be less than that needed to reliably establish a trend in data, both types of analyses may be needed to draw conclusions reliably. Especially for critical decisions, project managers
should seek expert advice in order to design a sampling program that will yield statistically defensible results. One potential method, power analysis, is described in Biostatistical Analysis (Zar 1999).
Another crucial element of developing a monitoring plan typically is cost. Generally, it is more cost-effective to collect less data, providing they are the “correct” or most useful data than it is to collect more of the “wrong” data. Following the key steps outlined in this guidance to design a monitoring plan should help project managers determine what are the “correct” data. Project managers may also find it useful to consider the use of indicator or surrogate parameters that correlate with those of primary interest, as a supplement to primary parameters that are especially costly or problematic to collect.
Finally, this step of monitoring plan development should ensure mechanisms are in place for modifying the plan based on new information.
Example: From the remedial investigation data, we know that smallmouth bass spend most of their time in the contaminated area and spawn in late spring. The proposed sampling plan would consist of overlaying an unbiased sampling grid onto a map of the contaminated area of River X as well as in the areas upstream and downstream of the site.
It is decided that 30 four-year old female bass will be collected in the early spring, before spawning, in each of these areas. A power analysis on baseline data indicated 20 fish would allow the project team to discern a 0.5 ppm or greater change in tissue concentration with 0.25 ppm confidence intervals (90 percent). However, given cost considerations, only ten samples will be analyzed immediately and the other 20 archived for further analyses pending the results.
Step 5. Conduct Monitoring Analyses and Characterize Results
The next recommended step in developing a monitoring plan includes data collection and analysis, evaluating analytical results, and addressing data deviations from the monitoring DQOs. At this point, the project manager should evaluate the data with regard to the monitoring hypotheses, the DQOs, and the monitoring decision rules developed in previous steps. At this step, the project manager should implement decision rules that may call for continuing, stopping, or modifying the monitoring or for taking additional action at the site.