3
Monitoring and Data Analysis to Support Adaptive Site Management

Adaptive site management (ASM) is dependent on the the development of analytical tools to help site managers determine when, and to what degree, a change of remedy will better achieve the goals of cleanup. At the same time, these tools should help demonstrate to diverse stakeholder groups that changes are warranted. It is important to gain support from the affected public and from public or private transferees prior to making changes in remedial strategies, even when an agreement has already been reached between the lead regulatory agency and the responsible party. Consensus can best be achieved if there are objective methods that help evaluate the potential changes.

This chapter considers analytical tools and monitoring techniques that can aid in the assessment of remediation performance and help site managers decide if the current remedy-in-place should be reevaluated. Monitoring programs supply the information required to support the four management decision periods (MDP) described in Chapter 2. For example, analysis of monitoring data is needed to determine whether performance standards and operational expectations have been met, whether remedial goals have been achieved, and ultimately whether site closeout can occur.

ANALYTICAL TOOLS FOR EVALUATING REMEDY EFFECTIVENESS AND NEED FOR CHANGE

Both graphical and tabular techniques exist to help make decisions about the effectiveness of remedies and the need for change. Tabular



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 106
3 Monitoring and Data Analysis to Support Adaptive Site Management Adaptive site management (ASM) is dependent on the the development of analytical tools to help site managers determine when, and to what degree, a change of remedy will better achieve the goals of cleanup. At the same time, these tools should help demonstrate to diverse stakeholder groups that changes are warranted. It is important to gain support from the affected public and from public or private transferees prior to making changes in remedial strategies, even when an agreement has already been reached between the lead regulatory agency and the responsible party. Consensus can best be achieved if there are objective methods that help evaluate the potential changes. This chapter considers analytical tools and monitoring techniques that can aid in the assessment of remediation performance and help site managers decide if the current remedy-in-place should be reevaluated. Monitoring programs supply the information required to support the four management decision periods (MDP) described in Chapter 2. For example, analysis of monitoring data is needed to determine whether performance standards and operational expectations have been met, whether remedial goals have been achieved, and ultimately whether site closeout can occur. ANALYTICAL TOOLS FOR EVALUATING REMEDY EFFECTIVENESS AND NEED FOR CHANGE Both graphical and tabular techniques exist to help make decisions about the effectiveness of remedies and the need for change. Tabular

OCR for page 106
methods attempt to characterize the various objectives and attributes of interest for alternative remediation plans and display them on a single table so that they may be considered together. These objectives could include human health and ecosystem risks (or risk reductions), contaminant mass remaining (or removed), projected time and cost to completion of remediation, projected land use and property values at or near the site, and a qualitative indication of the likely extent of support or opposition among different stakeholder groups. This presentation should help illuminate major advantages and disadvantages of each alternative, and indicate the tradeoffs between the desired objectives that occur in switching from one remediation plan to another. More formal analysis is also possible using various techniques of multiattribute utility theory (Keeney and Raiffa, 1976; Keeney, 1980; Merkofer and Keeney, 1987; Edwards and Barron, 1994; Clemen, 1996; Farber and Griner, 2000). Examples include the assignment of weights to different objectives (both by the site manager and by different stakeholders) to see how sensitive preferred alternatives are to these differing weights. As a hypothetical example, the eight objectives identified in Chapter 2 could be used, with differential weight being given to them to reflect laws and regulations and stakeholder preferences. The outcomes of different remedies can be ranked in an attempt to identify the most promising alternative. Such techniques have been employed to help facilitate stakeholder deliberations and decisions for other environmental management problems (Jennings et al., 1994). Often, such deliberations are best supported with simple and effective graphical presentations for each alternative, as discussed below. One weakness of this approach is that it can be difficult and costly (in terms of time and resources) to obtain quantitative values for all objectives. In addition to the tabular approaches, a number of graphical options can be developed to illustrate when changes in a remedy might be necessary. For remediation operations based upon contaminant extraction (e.g., pump-and-treat or soil vapor extraction), the most straightforward graph would be one that displays mass removal over time, as shown in Figure 3-1. Indeed, such graphs are already commonly prepared in practice, as discussed in Chapter 2 in the Lawrence Livermore case study (and other case studies described later). Recent Navy guidance (NAVFAC, 2001) advocates preparation of performance plots of monthly operation and cost data similar to Figure 3-1. Although mass removal is one objective measure of the remediation performance, cleanup goals are normally based upon reduction of total pollutant concentrations to health-based standards. [Such cleanup goals

OCR for page 106
contain an implicit assumption that total concentration levels determine risk, which may or may not be accurate depending on the bioavailability of the contaminant (NRC, 2003)]. Therefore, another way to assess the progress of remediation is to plot the temporal changes in concentration at chosen “sentinel” monitoring wells (e.g., wells located at the downgradient property boundary or adjacent to critical receptors). Such a plot is represented by Figure 3-2, which shows both hypothetical contaminant concentration over time as well as the reduction in contaminant concentration (or reduction in risk) over time. This second measure is more reliable, because calculation of the baseline risk associated with the initial contaminant level is fraught with uncertainty, whereas there is less uncertainty about the risk reduction (as measured by the surrogate concentration reduction). The hypothetical graphics shown in Figures 3-1 and 3-2 are drawn to represent a single remediation technique (e.g., pump-and-treat, soil vapor extraction). Analogous curves using different measured parameters could also be drawn to describe containment technologies (e.g., sediment capping), which aim to limit the contaminant mass flux through a “compliance” boundary. Of course, a containment technology would have FIGURE 3-1 A hypothetical plot of contaminant mass removed over time or over cost, for a remedy based on extraction of mass. Both the cumulative mass removed and the rate of mass removed are shown.

OCR for page 106
FIGURE 3-2 Hypothetical plot of (A) contaminant concentration over time or over cost and (B) reduction in contaminant concentration over time or cost. little or no impact on mass removal (Figure 3-1), but would achieve dramatic reduction in risk; this is an example of an exposure reduction (“E”) strategy as previously discussed in reference to Figures 2-1 and 2-2. Ideally, there would be a set of performance curves like those in Figures 3-1 and 3-2 for different remediation methods or management options such that the curves could guide decisions as to which option to select and when to change from one approach to another. As an illustration of such curves, consider Figure 3-3, which shows a family of hypothetical curves for the risk reduction over time for various types of remediation systems. Curves A, B, C, D, E, and F within Figure 3-3 suggest a wide range of potential results from different remedies. Attaching specific strategies to any given curve is not possible without more informa-

OCR for page 106
tion on the type of contamination, the predominant exposure pathway, and the affected receptors. However, one can speculate that Curve A represents a mass removal strategy such as in situ chemical oxidation of dense nonaqueous phase liquids, where a high percentage of mass must be destroyed before a significant reduction in groundwater concentrations and thus risk is achieved (see Box 5-12 for more explanation of this behavior). Curves B, C, and E could represent any number of strategies where risk is reduced incrementally over time from the source zone, including monitored natural attenuation. Curve F may represent a strategy like containment or a landfill cap where no contaminant mass is reduced, with the dotted line representing the possibility of future catastrophic failure. The “effectiveness” of any particular remedy could be based on the ratio of risk reduced per unit of time or cost. (Keep in mind that it is difficult to quantify risk, and thus the ordinate axis may actually represent reduction in concentration.) Higher ratios would be desirable, and any remedy that provided the higher ratios may be considered well suited for the particular risk reduction goal. Lower ratios would suggest that either the remedy is not appropriate for meeting the risk reduction goal or the remedy needs to be optimized. FIGURE 3-3 Hypothetical graphical representations of the change in risk with time or cost for different remedies.

OCR for page 106
Of these curves, only Curve E suggests a remedy that is totally ineffective in meeting cleanup goals and perhaps would provide the strongest graphical illustration of the need to change or modify the remedy. Clearly, remedies for which costs are increasing without any noticeable reduction in risk should not be continued. The case for change may be less clear for Curve A, which will eventually result in risk reduction albeit at longer time periods and higher costs than Curves B and C. The shape of Curve B is analogous to several case studies (see Box 2-3 and Appendix B) in that there is a relatively large initial reduction in contaminant concentration followed by a long period of relatively small reduction. Curve D represents a unique case in the sense that risk is seemingly being reduced quite effectively, yet as the remedy continues longer, the risk increases. This type of result may occur when source materials are drawn into an area or aquifer as a result of the remedy, increasing the concentrations of the contaminants to such a degree that higher risk results. This may also be the same type of curve that would result when an effective remedy is turned off and a rebound in concentrations occurs as the plume continues to move through the monitoring wells (but only if time, not cost, is the x-axis; if you turned off a remedy, presumably the cost disappears). In addition to the qualitative assessments of the various curves described above, graphical tools could provide more quantitative guidance, assuming that reliable and accurate values for cost and risk reduction can be measured. For example, if there is a desired target goal for risk reduction, then a horizontal line can be drawn from this target to find the “least cost” remediation scheme. Using the example illustrated in Figure 3-3, Curve B would be conceptually the most desirable over the mid term, although Curve D achieves the target risk reduction at the least cost over the short term. Conversely, if there is a target remediation budget, then a vertical line can be drawn from this target to find the most effective remediation scheme. Using the example illustrated above, Curve A would be selected. These examples are intended to be illustrative, and more detailed quantitative assessments are possible. For example, the slopes of the curves in Figure 3-3 measure the marginal risk reduction per unit investment, and these can be used in principle to optimally switch from one curve to another. Of course there may be other constraints that preclude such flexibility, and the difficulties in generating the risk reduction estimates must also be appreciated. As discussed in Chapter 2, risk reduction may not be the sole objective of a site remediation strategy. For example, if both contaminant

OCR for page 106
mass removal and risk reduction objectives are sought, then the problem becomes more complicated to visualize; however, graphical tools such as the one illustrated in Figure 2-2 could be developed. In this case each remediation system is represented by two curves, one measuring its performance for the risk reduction objective, and the other for the mass reduction objective. The time horizon for remediation is another objective that is often not considered explicitly during the remedy selection phase. However, short remediation times would be highly desirable in scenarios where the property is to be transferred for economic development. In most cases a single remediation strategy will not be capable of simultaneously satisfying all the objectives. The value of such a multidimensional graphical plot is that tradeoffs among objectives and strategies become evident, thus establishing a framework for stakeholder input and negotiation. Although development of performance curves is advocated in recent Navy guidance (NAVFAC, 2001), they are not routinely developed at most sites, particularly for soil and groundwater contamination. Rather, the general sequence of events is to determine a remedial goal and then choose a technology that will meet the goal at lowest cost. For sediment contamination, it is more typical to use the type of predictive models that could generate these performance curves in choosing the remedy (e.g., see Figure 3-4). The committee strongly recommends that the Navy make a concerted effort to collect the appropriate performance data so that these curves can be generated for various types of remedial actions and hydrogeologic settings. Indeed, data likely exist from Department of Defense (DoD), Department of Energy (DOE), and Superfund sites, as well as from government demonstration programs like the Environmental Security Technology Certification Program. The goal is to develop a set of models for broad classes of remedies, contaminants, exposure pathways, and receptors that can then be calibrated (most logically during the feasibility study) with site-specific data to generate performance curves applicable to a specific site. Developing the models in the first place will require data collection at sites where remedies are already in place, including data on contaminant concentrations at compliance or receptor locations if risk reduction is a desired metric. The benefits of this exercise are accrued later when the resulting models are calibrated with site-specific information and then used to inform remedy selection. Because such models reflect our current understanding of subsurface processes, which in some cases is limited, the models should be updated as performance monitoring data become available.

OCR for page 106
FIGURE 3-4 Model projections for polychlorinated biphenyl (PCB) concentrations in Thompson Island fish from 1998 to 2068 for various remedial alternatives as outlined by EPA Region 2. The noise in the "no action" projection is due to year-to-year variability in the projected flow record, which reflects the statistics of historical flows. SOURCE: Reprinted, with permission, from the National Research Council (2001). © (2001) National Academies Press. Graphical tools can also be used to make decisions after implementation of a remedy, particularly in conjunction with the specific management decision periods of ASM. In addition to answering the three questions of MDP2 (is the remedy meeting performance standards, is it meeting operational expectations, and is it meeting the remedial goal), graphical analysis of monitoring data can enable identification of asymptotic conditions where concentrations are not low enough at the site to achieve the health-based remedial goal, and operation and maintenance costs have become high enough to raise concerns. Interpretation of the graphs to provide yes or no answers to these questions will be subjective, because there will likely be disagreement about various critical performance criteria (e.g., at what dollar value does the cost per pound removed become cost-inefficient, or at what slope of the concentration versus time curve should the remedy be changed). Nevertheless, the graphs will indicate trends that provide information needed by the remedial project managers (RPMs), regulators, and stakeholders for decision making.

OCR for page 106
Several case studies already exist demonstrating how graphical tools can aid in making decisions to modify remedies and in evaluating remedial objectives. In almost all these examples, concentration is the measured parameter and is used as a surrogate indicator of risk. The first study is from the set of volumes published by the U. S. Environmental Protection Agency (EPA) under the auspices of the Federal Remediation Technologies Roundtable (EPA, 1998a). At Pope Air Force Base, as much as 75,000 gallons of JP-4 free product are floating on top of the water table; some dissolved volatile organic compounds (VOCs) have also been detected in groundwater samples. The remediation system consists of a free product cut-off trench and a dual pump recovery system. Figure 3-5 shows a decreasing removal rate over time for free product at this site, such that the cumulative recovery curve begins to flatten after April 1995. (Note that the EPA report includes another case study for a different free product removal site at Pope AFB where the cumulative removal continues to increase approximately linearly over time.) As of the last reported date (October 1996), approximately 3,500 gallons had been recovered, less than 5 percent of the estimated spill volume of 75,000 gallons. FIGURE 3-5 Monthly and cumulative free product removal at Site SS-07, Pope AFB. SOURCE: EPA (2000a).

OCR for page 106
An interesting aspect of this case study is that estimates of operation and maintenance (O&M) costs are combined with the above data to produce a graph showing cumulative costs versus cumulative pollutant mass removed. This graph, presented in Figure 3-6, illustrates the economic impact of the “tailing” behavior—as the remediation progresses, it becomes increasingly more costly to remove a given unit of contamination. However, it should be noted that Figure 3-6 was produced by making the simple assumption of constant average monthly O&M costs. Because the monthly costs are constant, and the monthly removal rate decreases over time as shown above, the cost per unit gallon removed will increase. This graph indicates that the performance of the remediation system has declined because little additional mass is being removed as funds continue to be spent, signaling that the system should be reevaluated. A reevaluation may or may not lead to a change in remedy, depending on the expected performance of the technology and other factors. FIGURE 3-6 Free product removal versus cumulative operation and maintenance costs at Pope AFB. SOURCE: EPA (2000a).

OCR for page 106
A second case study is for the Campbell Street Fuel Farm groundwater pump-and-treat system located at Marine Corps Air Station New River, which is co-located with Marine Corps Base Camp Lejeune, North Carolina (NAVFAC, 2001). The fuel farm is an active storage facility for JP-5, and release of fuels at the site has led to contamination of soil and groundwater by benzene, toluene, ethylbenzene, and xylene (BTEX) and VOCs. Contaminated groundwater is limited to the upper portion of a surficial aquifer with its water table 6 to 7 feet below ground surface. Initial remedial actions at the site were excavation of contaminated soil and removal of measurable free product. A groundwater pump-and-treat system began operation in July 1996; the system includes interceptor trenches and several extraction wells that were installed in plume hot spots. The trenches are downgradient of the contaminant plume, and all intercepted water is directed toward sumps for removal. Figure 3-7 shows that the VOC mass removal rate has decreased significantly over time; while 3.5 pounds were removed during July 1996 through March 1999, less than 0.5 pounds have been removed since December 1997. Figure 3-8, a plot of the cumulative cost versus cumulative mass removed, dramatically displays the tailing behavior of the system. It can be seen that approximately $175,000 was spent to remove the first 3 pounds of VOCs, but an additional $325,000 was spent to remove the next 0.5 pounds. The graphical data below were used in conjunction with other analyses and assessments to recommend that the trenches be shut down and that monitoring data be collected to evaluate the degree to which the plume was being affected by natural attenuation processes. Figure 3-8 suggests that caution and knowledge of the chosen treatment are needed when interpreting such graphs for the purpose of making changes to the remedial system (as discussed in Chapter 2 with respect to MDP2). It would have been premature to abandon the pump-and-treat system at the first sign of cost inefficiency in late 1996. Fortunately, site managers recognized that such systems generally take years before performance reaches an asymptote; continued operation resulted in a substantially longer period of effective mass removal. When the graphical tools indicate the remediation system should be reevaluated, changing the remedy can improve the system, as illustrated graphically in Figure 3-9. This figure schematically depicts contaminant concentration versus time when changing from a suboptimal remedy (such as that depicted by Curve E in Figure 3-3) to another remedy (Curve B in Figure 3-3). Changing the remedy should alter the concentration versus time curve such that the target contaminant level is reached sooner.

OCR for page 106
could occur downstream beneath the creek channel, transporting contaminants to an estuary of the Chesapeake Bay without discharge through wetland sediments where biodegradation of the chlorinated solvents occurs. The Hoverprobe allowed investigation at 13 sites along the stream channel that were previously inaccessible because of mud and shallow water (Phelan et al., 2001). Continuous sediment coring and water-quality profiling for chlorinated volatile organic compounds and redox-sensitive constituents were conducted without installation of wells, providing data to define plume boundaries and to refine the hydrogeologic parameters in a groundwater flow model used to assist in evaluating remedial alternatives. The Hoverprobe and a support hovercraft during drilling and water-quality profiling along the West Branch, Canal Creek, Aberdeen Proving Ground, MD. The support hovercraft was used for transport of samples to nearby laboratory facility for immediate analysis and in case emergency exit was needed.

OCR for page 106
ited monitoring well information. Although these types of technologies may never be appropriate for deep vadose zone sites or sites with fractured rock flow systems, they would be appropriate for the majority of coastal Navy facilities with relatively near-surface saturated zones and contamination events. In the case of traditional monitoring wells, techniques for obtaining less expensive and more representative groundwater samples have also been developed. These include low purge technologies and passive diffusion samplers. Passive diffusion samplers can eliminate altogether the need for purging monitoring wells before sampling. Diffusion samplers are a class of samplers, developed by Don Vroblesky at the U. S. Geological Survey, that are based on the laboratory and field confirmation that VOCs can diffuse through low-density polyethylene films and reach equilibrium concentrations that correlate well with actual subsurface contaminant concentrations (USGS, 2001). Types of diffusion samplers include water-to-water samplers and vapor-to-vapor samplers. Both types are applicable to the sampling of groundwater (via wells), the groundwater/surface water interface, pore water in sediments, surface water, and water from treatment systems. Vapor-to-vapor samplers are also effective for measuring in situ soil gas and vapor phase concentrations in confined spaces. The effectiveness of diffusion samplers is dependent upon the samplers being in direct contact with volatile organic compounds. Diffusion samplers should not be deployed in monitoring wells where sand packs are less permeable than the surrounding formation. In addition, diffusion samplers are not recommended for the quantitative measurement of methyl-tertiary butyl ether (MTBE) or acetone. Multiple diffusion samplers deployed in a vertical array can provide an effective method of vertical contaminant profiling in monitoring wells. Optimal conditions would consist of the diffusion sampler or groundwater monitoring well screen being in direct contact with the surrounding formation, but correctly designed monitoring well sand packs are also appropriate. The presence of vertical gradients across the sampling interval will compromise the resolution of vertical contaminant profiling. The most promising application for diffusion samplers appears to be for long-term groundwater monitoring in wells, with the potential to reduce long-term monitoring costs by 20 percent to 50 percent. Detailed information regarding the appropriateness, construction, deployment, handling, and analysis of diffusion samplers can be found in USGS (2001).

OCR for page 106
Dynamic Work Plans The last opportunity for developing a more flexible and adaptive approach to subsurface performance monitoring is to base a characterization or monitoring program on dynamic work plans. Dynamic work plans differ from more traditional sampling and analysis plans in that they identify the decision logic that will be used for determining the appropriate analytical techniques and sample numbers, locations, and frequency as work proceeds, rather than pre-specifying those data collection characteristics. As alluded to above, dynamic work plans rely at least in part on direct push technologies and field analytic techniques. With these technologies, data collection can be adapted in response to the changing information needs of a remedial action, and the remedial action itself can be adjusted or adapted based on feedback from the data collection. The concept of developing hazardous waste site characterization programs based on dynamic work plans has been implemented under a variety of names, including expedited site characterization (DOE, 1998) and adaptive sampling and analysis programs (DOE, 2001). The EPA TIO has been advocating the Triad approach (EPA, 2001) to environmental data collection, which adds systematic planning to the dynamic work plan/field analytic mix. The EPA Superfund program is currently preparing draft guidance on the development of dynamic work plans. Case studies that document characterization cost reductions associated with these types of approaches usually report savings on the order of 50 percent or more. These savings are derived from reductions in per-unit analytical costs and in the overall number of samples collected. Although the emphasis has historically been on site characterization, dynamic work plan concepts and associated technologies (field analytics, sensors, direct push, etc.) are equally applicable to the remediation phase of site restoration. In fact, the potential impacts on overall costs and remediation performance are greater during remediation than they are during characterization because savings can be realized both from reductions in data collection costs and from improved remedial action performance. In this context, dynamic work plans are a natural component of ASM. Dynamic work plans are particularly applicable to contaminated soil excavations or contaminated sediment dredging operations. Box 3-7 describes the adaptive nature of a removal project for soils contaminated with radionuclides. A similar example, but in the context of pesticide-contaminated soils, was reported in USACE (2000). In this example,

OCR for page 106
BOX 3-7 Precise Excavation at the Ashland 2 Site The U.S. Army Corps of Engineers (USACE) is conducting cleanup of radiologically contaminated properties as part of the Formerly Utilized Sites Remedial Action Program (FUSRAP). The largest cost element for most of the FUSRAP sites is the excavation and disposal of contaminated soil. Conventional approaches to the design of soil excavation/disposal programs delineate excavation boundaries based on existing characterization data. Excavation then proceeds using these design drawings as the basis for determining which soil must be excavated and which can remain. There is considerable evidence that in fact most pre-remediation characterization datasets are inadequate for precisely delineating contamination footprints. The result can be overexcavation of clean soil at considerable unnecessary expense. A precise excavation approach was implemented at the Ashland 2 FUSRAP site. Data collection was embedded into the excavation program, with data collection consisting of real-time in situ sensors, global positioning system units, and an onsite laboratory. Excavation work proceeded in lifts that ranged from 0.5 to 2 feet in depth, with dig-face screening occurring before excavation continued. A pre-excavation estimate of contaminated soil volumes based on RI/FS data placed the total at 14,000 cubic yards. By the time the work was completed, approximately 45,000 cubic yards of soil were identified as being contaminated at levels that were above the cleanup criteria and were excavated for offsite disposal. A post-excavation analysis specifically of the initial surficial lift showed that if excavation of surficial soil had been based solely on pre-existing data, it would have removed 4,000 cubic yards of minimally contaminated soil (i.e., where soil contaminant concentrations were below the cleanup criteria), and it would have missed 8,000 cubic yards of soil that had contamination is excess of the cleanup immunoassay kits were used to better define excavation footprints and verify dig-face cleanup guideline compliance at the Wenatchee site. In its cost and performance report, the USACE indicated that overall remediation costs were half of what would have been incurred if excavation had proceeded on the basis of existing historical datasets alone. There is also a place for dynamic work plans within groundwater remedial action monitoring. A simple example is a plan that samples a traditional network of monitoring wells. In this instance a dynamic work plan might rely on passive diffusion samplers for generating samples and on field analytics for screening those samples. Based on the results, a decision might be made to replicate analyses using an offsite laboratory, to expand sampling to adjacent wells that would not have otherwise been

OCR for page 106
criteria. Preliminary cost estimation work indicated that the additional cost of the excavation support data collection was approximately $168,000 over six months of excavation. Over $1.5 million in cost savings were achieved by avoiding unnecessary offsite disposal costs for just the initial surficial lift (Durham et al., 1999). sampled in that round, or to increase sampling frequency in the short term. In the situation where a technology such as direct push was available for quickly acquiring groundwater samples from new locations, or for installing temporary monitoring points, the decision might be to expand the network in the short term to address unexpected trends or results in datasets. Alternatively, a monitoring system might include real-time data acquisition from dedicated in situ sensors. A dynamic work plan would identify the types of result scenarios that would require a response, either by requiring additional data collection or by revisiting the remedial system. An example would be real-time monitoring of a leachate collection system for parameters that might indicate a containment cell failure. A second example would be continuous depth-to-water-table sensors posi-

OCR for page 106
tioned around a barrier wall whose relative potentiometric results might indicate loss of groundwater capture. These latter examples do not represent current practices for monitoring system design, but they do suggest ways that dynamic work plans and adaptive sampling techniques could be used to facilitate an ASM approach to remedial action performance evaluation. MAJOR CONCLUSIONS AND RECOMMENDATIONS This chapter was meant to provide general guidance on how to assess remedial performance monitoring with graphical tools and on some of the new monitoring tools available to do so. A major challenge in implementing adaptive site management will be to design the information-gathering efforts to support the management decision points fleshed out in Chapter 2. Thus, monitoring plans should be developed from clearly articulated objectives (such risk reduction, reduction in some indicator of risk, or mass removal), they should support the evaluation of remedial operations performance (MDP2), and they should validate or refine site conceptual models. More specific recommendations that link monitoring to the ASM process are provided below. Plots of mass removal or concentration versus time or cost (or other metrics depending on the remedy) are objective and transparent tools for illustrating remedial effectiveness that should trigger when to either modify or optimize the existing remedy or to change the remedy. Such graphs should be used after remedy selection to address management decision periods 2 and 3 of ASM. Graphical representations should serve both to enhance stakeholder understanding of the options and to make better decisions about implementing or modifying remedies. At individual sites under investigation, the Navy, in consultation with all stakeholders, should select a unit cost for the continued operation of the remedial action, above which the existing remedy is no longer considered a tenable option. The Navy should collect and analyze data to develop and validate predictive models of remedy performance. The remedy selection process could be made more quantitative and transparent with the provision of design guidance, charts, and models that summarize technology applications and predict their performance in different environmental settings.

OCR for page 106
Uncertainties in hydrogeologic data, contaminant concentrations, and rates of remediation should be explicitly recognized in the development and application of performance plots. There are many sources of uncertainty in estimating the mass or risk reduction achieved by any remediation scheme. When sufficient site data are available, statistical methods can be used to estimate error or confidence bands on the performance plots. Site monitoring plans should be developed to ensure that the collected data serve to reduce uncertainty. A concerted effort should be made to increase monitoring program effectiveness (and to reduce costs) by optimizing the selection of monitoring points, incorporating field analytics and innovative data collection technologies such as direct push, and adopting dynamic work plans and adaptive sampling and analysis techniques. Real-time in situ monitoring technologies should also be considered as they mature. These techniques enhance the collection of information upon which ASM decision making is based. DoD should continue to support and foster research in chemical, physical, and biological techniques that would provide more rapid and adaptive approaches for monitoring remedy effectiveness. REFERENCES Aziz, J. J., C. J. Newell, H. S. Rifai, M. Ling, and J. R. Gonzales. 2000. Monitoring and remediation optimization system (MAROS) software user’s guide. Brooks Air Force Base, TX: Air Force Center for Environmental Excellence. Berthouex, P. M., W. G. Hunter, and L. Pallesen. 1978. Monitoring sewage treatment plants: some quality control aspects. Journal of Quality Technology 10:139–149. Box, G. E. P., and G. M. Jenkins. 1994. Time series analysis: forecasting and control (3rd Edition). Upper Saddle River, NJ: Prentice Hall. Cameron, K., and P. Hunter. 2000. Optimization of LTM networks using GTS: statistical approaches to spatial and temporal redundancy. Brooks Air Force Base, TX: Air Force Center for Environmental Excellence. Clemen, R. T. 1996. Making hard decisions: an introduction to decision analysis (2nd Edition). Belmont, CA: Duxbury Press. Department of Energy (DOE). 1998. Expedited site characterization. Innovative Technology Summary Report DOE/EM-0420. DOE. 2001. Adaptive sampling and analysis programs (ASAP). Innovative Technology Summary Report DOE/EM-0592.

OCR for page 106
DOE. 2002. Induced fluorescence sensors for direct push systems. Innovative Technology Summary Report DOE/EM-0638. Durham, L., D. Conboy, R. Johnson, and T. Sydelko. 1999. Precise excavation—an alternative approach to soil remediation. Pp. 93–98 In: Proceedings of the National Defense Industrial Association, Denver, Colorado, March 19–April 1. Edwards, W., and F. H. Barron. 1994. SMARTS and SMARTER: improved simple methods for multiattribute utility measurement. Organizational Behavior and Human Decision Processes 60:306–326. Environmental Protection Agency (EPA). 1994. Assessment and remediation of contaminated sediments (ARCS) program—final summary report. EPA 905-S-94-001. Washington, DC: EPA. EPA. 1995. Site characterization analysis penetrometer system (SCAPS). Innovative Technology Evaluation Report. EPA/540/R-95/520. Washington, DC: EPA. EPA. 1998a. Remediation case studies: groundwater pump and treat (nonchlorinated contaminants). EPA 542-R-98-014, Vol. 10. Washington, DC: EPA Office of Solid Waste and Emergency Response. EPA. 1998b. EPA’s contaminated sediment management strategy. Washington, DC: EPA Office of Water. EPA. 2000a. FRTR cost and performance remediation case studies and related information. EPA 542-C-00-001. Washington, DC: EPA Office of Solid Waste and Emergency Response. EPA. 2000b. Contaminated sediment news. EPA-823-N-00-002. Washington, DC: EPA. EPA. 2000c. Test methods for evaluating solid waste, physical/chemical methods. SW-846 Draft Update IVB. Washington DC: EPA Office of Solid Waste and Emergency Response. EPA. 2001. Using the Triad approach to improve the cost-effectiveness of hazardous waste site cleanups. EPA-542-R-01-016. Washington, DC: EPA. Farber, S., and B. Griner. 2000. Using conjoint analysis to value ecosystem change. Environ. Sci. Technol. 34(8):1407–1412. Freeze, R. A., and S. M. Gorelick. 1999. Convergence of stochastic optimization and decision analysis in the engineering design of aquifer remediation. Ground Water 37(6):934–54. Gibbons, R. D., and D. E. Coleman. 2001. Statistical methods for detection and quantification of environmental contamination. New York: Wiley. Gilbert, R. O. 1987. Statistical methods for environmental pollution monitoring. New York: Van Nostrand Reinhold. HydroGeoLogic, Inc. 2000. Final base-wide groundwater sampling and analysis plan, NAS Fort Worth JRB, Texas. March 2000. James, B. R., and S. M. Gorelick. 1994. When enough is enough: the worth of monitoring data in aquifer remediation design. Water Resources Research 30(12):3499–3513. Jennings, A. A., N. Mehta, and S. Mohan. 1994. Superfund decision analysis in

OCR for page 106
presence of uncertainty. J. Environ. Engr.120(5):11321150. Johnson, V. M., R. C. Tuckfield, M. N. Ridley, and R. A. Anderson. 1996. Reducing the sampling frequency of groundwater monitoring wells. Environ. Sci. Technol. 30(1):355–358. Johnson, R., J. Quinn, L. Durham, G. Williams, and A. Robbat, Jr.1997. Adaptive sampling and analysis programs for contaminated soils. Remediation, Summer:81–96. Kampbell, D. H., P. E. Haas, R. N. Miler, J. E. Hansen, and F. H. Chapelle. 1998. Technical protocol for evaluating natural attenuation of chlorinated solvents in ground water. Washington, DC: EPA. Keeney, R. 1980. Siting energy facilities. New York: Academic Press. Keeney, R., and H. Raiffa. 1976. Decisions with multiple objectives. New York: Wiley. Kirtay, V. J., J. H. Kellum, and S. E. Apitz. 1998. Field-portable x-ray fluorescence spectrometry for metals in marine sediments: results from multiple sites. Water Sci. Technol. 37(6-7):141–148. Lettenmaier, D. P. 1976. Detection of trends in water quality data from records with dependent observations. Water Resources Research 12:1037–1046. Lettenmaier, D. P. 1977. Detection of trends in stream quality: monitoring network design and data analysis. Technical Report No. 51, NTIS PB-285 960. Seattle, WA: C. W. Harris Hydraulics Laboratory, Department of Civil Engineering, University of Washington. Loáiciga, H. A., R. J. Charbeneau, L. G. Everett, G. E. Fogg, B. F. Hobbs, and S. Rouhani. 1992. Review of ground-water quality monitoring network design. Journal of Hydraulic Engineering 118(1):11–37. Lorah, M. M., and L. D. Olsen. 1999a. Degradation of 1,1,2,2-tetrachloroethane in a freshwater tidal wetland: field and laboratory evidence. Environ. Sci. Technol. 33(2):227–234. Lorah, M. M., and L. D. Olsen. 1999b. Natural attenuation of chlorinated volatile organic compounds in a freshwater tidal wetland: field evidence of anaerobic biodegradation. Water Resources Research 35(12):3811–3827. Lorah, M. M., L. D. Olsen, B. L. Smith, M. A. Johnson, and W. B. Fleck. 1997. Natural attenuation of chlorinated volatile organic compounds in a freshwater tidal wetland, Aberdeen Proving Ground, Maryland. U.S. Geological Survey Water-Resources Investigations Report 97-4171. Maxwell, R. M., S. D. Pelmulder, A. F. B. Tompson, and W. E. Kastenberg. 1998. On the development of a new methodology for groundwater-driven health risk assessment. Water Resources Research 34(4):833–847. Maxwell, R., S. F. Carle, and A. F. B. Tompson. 2000. Risk-based management of contaminated groundwater: the role of geologic heterogeneity, exposure and cancer risk in determining the performance of aquifer remediation. Proceedings of the 2000 Joint Conference on Water Resources Engineering and Water Resources Planning & Management, ASCE, July 30– Aug. 2, Minneapolis. Merkofer, M. W., and R. L. Keeney. 1987. A multiattribute utility analysis of

OCR for page 106
alternative sites for the disposal of nuclear waste. Risk Analysis 7:173–194. Meyer, P. D., A. J. Valocchi, and J. W. Eheart. 1994. Monitoring network design to provide initial detection of groundwater contamination. Water Resources Research 30(9):2647–2659. Minsker, B. S., and J. Bryan Smalley. 1999. Cost-effective risk-based in situ bioremediation design. Pp. 349–354 In: Proceedings of the 5th International In Situ and On-site Bioremediation Symposium, April 1922, 1999, San Diego, CA. Bruce C. Alleman and Andrea Leeson (eds.). National Research Council (NRC). 1999. Environmental cleanup at Navy facilities: risk-based methods. Washington, DC: National Academy Press. NRC. 2000. Natural attenuation for groundwater remediation. Washington, DC: National Academy Press. NRC. 2001. A risk management strategy for PCB-contaminated sediments. Washington, DC: National Academy Press. NRC. 2003. Bioavailability of contaminants in soils and sediments: processes, tools, and applications. Washington, DC: National Academy Press. NAVFAC. 2000. Guide to optimal groundwater monitoring. Prepared for the Naval Facilities Engineering Research Center by Radian International. NAVFAC. 2001. Guidance for optimizing remedial action operation (RAO). Special Report SR-2101-ENV. Prepared for the Naval Facilities Engineering Service Center. Research Triangle Park, NC: Radian International. Phelan, D. J., M. P. Senus, and L. D. Olsen. 2001. Lithologic and groundwater-quality data collected using Hoverprobe drilling techniques at the West Branch Canal Creek wetland, Aberdeen Proving Ground, Maryland, AprilMay 2000. U.S. Geological Survey Open-File Report 00-446. Reed, P., B. Minsker, and A. J. Valocchi. 2000. Cost-effective long-term ground-water monitoring design using a genetic algorithm and global mass interpolation. Water Resources Research 36(12):3731–3741. Reible, D. D., K. T. Valsaraj and L. J. Thibodeaux. 1991. Chemodynamic models for transport of contaminants from sediment beds. Pp. 187–228 In: Handbook of environmental chemistry. O. Hutzinger (ed.). Heidelberg, Germany: Springer-Verlag. Ridley, M., and D. MacQueen. 1995. Cost-effective sampling of groundwater monitoring wells: a data review and well frequency evaluation. Pp. 14–21 In: Proceedings of the Hazardous Materials Management Conference and Exhibition, April 4–6, 1995, San Jose, California. Rizzo, D. M., D. E. Dougherty, and M. Yu. 2000. An adaptive long-term monitoring and operations system (aLTMOs™) for optimization in environmental management. Proceedings of 2000 ASCE Joint Conference on Water Resources Engineering and Water Resources Planning and Management, Minneapolis, MN. ISBN 0-7844-0517-4. R. H. Hotchkiss and M. Glade (eds.). Reston, VA: American Society of Civil Engineers. Russell, K. T., and A. J. Rabideau. 2000. Decision analysis for pump-and-treat design. Ground Water Monitoring and Remediation, Summer:159–168. Sohn, M. D., M. J. Small, and M. Pantazidou. 2000. Reducing uncertainty in

OCR for page 106
site characterization using Bayes Monte Carlo methods. J. Environ. Engr. ASCE. 126(10):893–902. Stallard, M. O., S. E. Apitz, and C. A. Dooley. 1995. X-ray fluorescence spectrometry for field analysis of metals in marine sediments. Mar. Pollut. Bull. 31:297–305. Stansbury, J., I. Bogardi, and E. Z. Stakhiv. 1999. Risk-cost optimization under uncertainty for dredged material disposal. J. Water Resour. Plan. Manag. ASCE 125(5):342–351. Starks, T. H., and G. T. Flatman. 1991. RCRA ground-water monitoring decision procedures viewed as quality control schemes. Environmental Monitoring and Assessment 16:19–37. Stiber, N. A., M. Pantazidou, and M. J. Small. 1999. Expert system methodology for evaluating reductive dechlorination at TCE sites. Environ. Sci. Technol. 33(17):3012–3020. Thoms, S. R., G. Matisoff, P. L. McCall, and X. Wang. 1995. Models for alteration of sediments by benthic organisms. Project 92-NPS-2. Alexandria, VA: Water Environment Research Foundation. U.S. Army Environmental Center (USAEC). 2000. The Tri-service site characterization and analysis penetrometer system-SCAPS: innovative environmental technology from concept to commercialization. Report Number SFIM-AEC-ET-TR-99073. U.S. Geological Survey. 2001. User’s guide for polyethylene-based passive diffusion bag samplers to obtain volatile organic compound concentrations in wells. Part 1-4. Water-Resources Investigations Report 01-4060. U. S. Army Corps of Engineers (USACE). 2000. Expedited characterization and soil remediation at the test plot area, Wenatchee Tree Fruit Research Center, Wenatchee, Washington. Cost and Performance Report. U.S. Army Corps of Engineers Hazardous, Toxic, Radioactive Waste Center of Expertise . Wagner, B. J. 1999. Evaluating data worth for ground-water management under uncertainty. J. Water Resour. Plan. Manag. ASCE 125(5):281–288. Wiedemeier, T. H., H. S. Rifai, C. J. Newell, and J. T. Wilson. 1999. Natural attenuation of fuel and chlorinated solvents in the subsurface. New York: John Wiley and Sons. Wisconsin Department of Natural Resources. 2001. Remedial investigation and feasibility study for the Lower Fox River, October.