Best Practices for Risk-Informed Decision Making Regarding Contaminated Sites: Summary of a Workshop Series (2014)

4


Monitoring

Many of the Department of Energy’s (DOE’s) legacy waste sites will have contamination and wastes remaining on site that will require long-term monitoring and control. The focus of the third session of the workshop was monitoring and institutional controls for long-term stewardship. The topic was introduced by Craig Benson, director of sustainability research and education and chair of civil and environmental engineering and geological engineering, University of Wisconsin and member of the Consortium for Risk Evaluation with Stakeholder Participation (CRESP).

Examples and case studies of long-term monitoring programs were provided by

• Dave Geiser, director and acting deputy director of DOE’s Office of Legacy Management (LM), and
• Mary Flora, environmental compliance and area completion projects for Savannah River Nuclear Solutions, LLC.

Rula Deeb, principal, Geosyntec, moderated the session.

This chapter provides summaries of the key points made by each of these individuals and by participants in the subsequent discussion sessions. These statements reflect the viewpoints of the individual speakers, not the consensus views of the workshop participants or of the National Academy of Sciences.

4.1 STRATEGIES FOR LONG-TERM MONITORING AND STEWARDSHIP

Craig Benson

For containment systems, one must “trust but verify” to confirm that the system is protecting human health and the environment. This can be done by confirming that compliance levels have been met (compliance monitoring) or that systems are performing as designed (functional monitoring). Groundwater wells, which are commonly used for compliance monitoring, are not always the best way to determine whether a system is functioning as designed. Monitoring for compliance is standard practice for many sites (see Table 4-1 below). The regulators dictate (e.g., through tri-party agreements, consent orders, or Federal Facility Agreements [FFAs]) where and when to monitor the containment systems. The problem with this approach is that a failure of the system is detected too late and without enough information to fix the problem.

A preferable approach is to focus on confirming the performance of the containment system through functional monitoring. Monitoring by

TABLE 4-1 Compliance Versus Functional Monitoring

SOURCE: Modified from Benson 2014.

function takes place at critical locations and at frequencies dictated by the design of the system and may include unconventional measurements. This approach identifies underperforming systems before they fail and provides data to improve models, designs, and decision support. Data collected at key locations can add confidence in a system’s design or provide a better understanding of previously unknown mechanisms contributing to underperforming containment systems. The challenge is that functional monitoring may not meet regulatory requirements.

Lysimeters can be used to measure moisture and other parameters in containment systems. For example, a containment system in Sacramento, California, was designed to control water percolation to less than 0.1 mm/year. A functional monitoring program—including lysimeters—provided insight into a previously unknown mechanism that contributed to excessive water percolation within the design.¹

The system was installed and, at the surface, appeared to be functioning well with grass and vegetation growing as expected and with no reported problems. However, lysimeter measurements occasionally showed jumps in percolation rates on the order of 100 mm/year (a factor of 1,000 higher than the system specifications). Additional measurements were made, and historical weather data analyzed. Increased percolation during the spring and summer was correlated with unusually wet, preceding winter months during which the soil did not dry out.

These additional data were used to update models and the source terms for percolation. Modifications were made to the system to function as originally designed. The effect of higher-than-normal precipitation during the rainy season on the soil’s water capacity during dry months has also impacted other designs and has improved the overall understanding of containment systems.

The basic requirements of a sustainable, long-term containment system are a remote monitoring capability, reliable hardware, and flexible deployment. Dr. Benson suggested that remote, automated monitoring should be used whenever possible because relying on onsite visits is problematic. When data are available on-demand (e.g., with frequent updates available online) problems can be diagnosed early.

For example, a successful automated, remote-monitoring program was implemented at the Resource Conservation and Recovery Act (RCRA) cap at the Fernald site. The objectives of the monitoring project were to verify performance of the containment systems, provide early warning of problems, and increase stakeholder confidence in the remedy. The features of the functional monitoring system were

______________

¹ See pages 7-11 of Dr. Benson’s presentation: http://dels.nas.edu/resources/static-assets/nrsb/miscellaneous/benson.pdf.

• surface access,
• a variety of high-reliability sensors (transducers, reflectometer, thermal sensors), and
• data logging and data access via Web-interface.

The Fernald RCRA cap monitoring system provided remote access to many different types of measurement data, enabling analysis to be carried out to find any problems that arose (see Figure 4-1).

The greatest challenge for long-term stewardship is monitoring of remedies. Monitoring must be effective, reasonably priced, and well-defined. Long-term monitoring systems that combine remote monitoring with a design compatible with natural surroundings are more likely to be effective, require minimal maintenance, and, therefore, be reasonably priced.

Two examples highlight the importance of well-designed long-term monitoring systems (e.g., combining remote monitoring with naturally compatible designs). In the first, a Uranium Mill Tailings Radiation Control Act (UMTRCA) disposal cell has been designed with a layer of stone on the cover to inhibit the development of vegetation. Still, removal of vegetation is required to maintain the effectiveness of the remedy. Annual control measures such as the spraying of vegetation will continue as long as the disposal cell is in place (in this case, for perpetuity).

The second example is the Monticello Disposal Cell. The cover for the disposal cell was designed to be compatible with its natural surroundings. Native vegetation was designed into the cover along with functional monitoring. The result is a sustainable cover with water control. The cover has been functioning for more than 20 years and has proven to be effective and performing as designed (< 1 mm/year percolation).(As a side note, a retrofit to the UMTRCA disposal cell example discussed above is under consideration. The retrofitted design introduces a cover compatible with the natural surroundings.)

4.2 SUMMARY OF DISCUSSION SESSION

Fernald Status and Remote Monitoring. Planning committee member Patricia Culligan (Columbia University) asked about the status of the Fernald monitoring system. Dr. Benson responded that, because it was experimental, it has been removed. It was not planned as part of the long-term stewardship of the site, but it did effectively demonstrate the importance of remote monitoring.

Functional Monitoring/Stakeholder Engagement. Carol Eddy-Dilek (Savannah River National Laboratory [SRNL]) works on a technical assistance program for EM in which teams of scientists travel to nuclear legacy

sites to resolve technical remediation problems. Based on her experience, she strongly supports “monitoring by function” and provided several examples that have impacted site remediation. A metal contamination problem at the Savannah River Site (SRS) highlights the usefulness of functional monitoring. The pump-and-treat system remedy for the metal contamination plume costing $1 million/year had become ineffectual. An option to treat the plume by adding a base to strip out the acids was presented by DOE to South Carolina Department of Health and Environmental Control (DHEC) and the Environmental Protection Agency (EPA). One of the challenges to the proposed remedy was monitoring the effectiveness of this new approach on the very large plume. It was decided to measure controlling variables, such as the pH, which were not normally monitored (e.g., through sampling of wells). If the pilot demonstration of the new remedy and its monitoring program show effectiveness, then regulators have agreed to accept and support the new remedy. Functional monitoring of the controlling variables as Dr. Benson has mentioned is critical for understanding of the behavior of the plume and the success of this pilot study.

Another example of functional monitoring is a monitoring and modeling program at an LM UMTRCA site. A uranium mining and tail processing facility has residual contamination from acid and carbonate leaches used during operations. Contamination levels in groundwater have been erratic for more than 20 years and have not been consistent with existing site models. A newly proposed hypothesis is that the large, contaminant source term in the vadose zone contaminates the groundwater as it rises and falls with seasonal flooding. The flooding injects contaminants from the vadose zone, and increased levels of contaminants later appear in the groundwater. Functional monitoring allowed the hypothesis to be developed and models to be updated. However, the physical and chemical processes behind the hypothesis were not initially easy to explain to the stakeholders.

Simplified conceptual models and commonly understood analogies were found to be critical tools when discussing scientific concepts with stakeholders. Block diagrams of the sources and sinks of the contamination connected by weighted arrows illustrate the main processes and contributors within a given site. In the example above, a simplified site model was introduced, and a technical analogy was used when discussing the proposed vadose zone injection hypothesis with stakeholders (in this case, members of tribal nations). The analogy of salt buildup in soil with periodic natural cleansing is a concept and process the stakeholders understood. If the processes can be understood, then the discussions for cleanup options become easier.

Failure of Containment Systems. Planning committee member Michael Kavanaugh (Geosyntec) commented that containment systems are well

understood and asked Dr. Benson what the remaining unknown properties and mechanisms might be. He also recalled that an NRC (2000) report suggested that DOE should plan for containment systems to fail² and wondered whether this is still an appropriate assumption. Dr. Benson replied that the design and construction of containment systems is fairly well understood. However, not much is known about how long they will last or their failure mechanisms. Regulations from the 1970s and 1980s were developed for systems that had not yet been built. The regulatory structure has not progressed along with the knowledge and experience that has been gained with respect to containment systems. As for the assumed failures of systems, care should be taken in using the term “fail.” Failure is interpreted as an on/off condition, when in reality containment systems are likely to degrade gradually; usually they do not suddenly and catastrophically fail. Containment systems designed incompatibly with their natural surroundings degrade faster than compatible systems. Some systems in the arid west of the United States are so compatible with natural systems that they are expected to last on the order of 200 years.

Eric Pierce (Oak Ridge National Laboratory [ORNL]) supported the concept that containment barriers do not instantaneously “fail” and wondered how one might convince regulators about this fact. Dr. Benson suggested that better compilation and sharing of failure rate and failure mode data would increase confidence and acceptance by the regulators.

Robin Anderson (EPA) provided a counter-example to demonstrate that containment systems can fail earlier than anticipated. A permeable reactive barrier (PRB) installed at the Monticello mill tailings site in Utah was designed to last for 70 years but stopped working after 7 years. Dr. Benson offered a different perspective on this example, which highlighted the importance of functional monitoring. Stan Morrison (Morrison 2006) studied the failure of the PRBs at Monticello through careful measurements and monitoring. He is credited for developing better methods to predict precipitation mechanisms in PRBs and to estimate their lifetimes. This is a great example of functional monitoring leading to improved understanding of an innovative remedy.

Monitoring of Sustainability. Paul Black (Neptune and Company, Inc.) pointed out that the types of monitoring systems discussed in this session are for only one part of the sustainability triad. Changes in stakeholder

______________

² “The committee believes that the working assumption of DOE planners must be that many contamination isolation barriers and stewardship measures at sites where wastes are left in place will eventually fail, and that much of our current knowledge of the long-term behavior of wastes in environmental media may eventually be proven wrong. Planning and implementation at these sites must proceed in ways that are cognizant of this potential fallibility and uncertainty” (NRC, 2000, p. 5).

and economic values should also be monitored to identify when decisions should be revisited in the future. Dr. Benson agreed and added that true sustainability analysis starts with remediating the waste but extends beyond that. A life cycle analysis (LCA) may help to understand how to integrate other factors such as stakeholder and economic values. He provided the following example: remediation of the Fox River in northern Wisconsin (which was contaminated by industrial discharges from paper mills) required that the river be dredged and the dredged sediment be transported to a northern Michigan disposal facility. The trucks containing the contaminated sediment passed through four states: Wisconsin, Illinois, Indiana, and Michigan. The remedy protected the Fox River, but an LCA that considered broader sustainability values may have resulted in a different remedy.

Timeframes. Paul Black supported the idea of reducing the timeframes that are routinely used to estimate the lifetime of remedies. He suggested that estimates should be limited to several hundred years because this is as far into the future that can reasonably be predicted. This timeframe should be a “rolling” timeframe, with monitoring to identify changes in the effectiveness of the remedy (or factors exterior to the remedy such as stakeholder values). An exit strategy or “stopping rule” is also needed. Mr. Levitan commented that models and estimates that project several hundreds of years into the future account for a very small fraction of the decay lives of the long-lived radioactive contaminants.

Expense of Functional Monitoring. Marolyn Parson (South Carolina Citizen Advisory Board) noted that compliance monitoring is required by regulation and agreements. How does one convince sites to spend extra money on additional functional monitoring or to replace compliance-based monitors with functional monitors? Craig Benson acknowledged this difficulty. He highlighted Carol Eddy-Dilek’s comments on building confidence with stakeholders to allow for measurements to be made in addition to legally mandated compliance monitoring. It may be time to rethink current regulatory monitoring requirements because they focus on compliance instead of monitoring for function.

Mary Flora (Savannah River Nuclear Solutions) suggested that part of the regulatory requirement to monitor is to understand the contaminants’ movement to confirm an effective remedy. An effective monitoring program needs an accurate “big picture.” Therefore, identifying and measuring leading indicators is an important and useful task because they can provide forewarning and an opportunity to address problems early. Functional monitoring may be of interest to regulators and decision makers for these reasons.

Models. Dawn Wellman (Pacific Northwest National Laboratory [PNNL]) commented that DOE and PNNL have recognized the need to look at the full system of a site using monitoring understand how and confirm whether the remedies are working. A document, “Scientific Opportunities for Monitoring at Environmental Remediation Sites” (Bunn 2012), has been developed by DOE. It is used as a foundational document to better understand the remediation and monitoring systems at the Hanford site.

Bill Levitan (EM) asked how one determines which leading indicators are needed to validate model assumptions. Dr. Benson replied that leading indicators will differ depending on the site and the remedy. Models provide predictions based on assumptions and can identify which variables may be used as leading indicators. However, the models, their assumptions, and their proposed leading indicators need validation with measurements collected through functional monitoring.

Ming Zhu (EM) agreed on the importance of models and their validation. He noted that an interagency working group is considering how to make better use of monitoring data that have been collected over the years from the remediated sites. The Federal Interagency Steering Committee on Multimedia Environmental Modeling (ISCMEM)³ current signatories include DOE, EPA, U.S. Nuclear Regulatory Commission (USNRC), the National Resources Conservation Service, the National Science Foundation, U.S. Army Corps of Engineers, and the USGS.

Importance of Science in Remediation Decisions. Dan Goode (USGS) quoted a NRC study, Science and Technology for DOE Site Cleanup (NRC 2007), which echoed comments by Mary Flora, Carol Eddy-Dilek, and Dawn Wellman on the need for science to guide complex remediation decisions.

There are many impediments to the use of new technologies at the site [Oak Ridge Reservation]…

…The change of contracting approach at the site (from a reservation-wide management and operation contractor to a management and integrating contractor) has also been an impediment because it has severed the direct ties that existed between the cleanup program and ORNL [Oak Ridge National Laboratory] (NRC, 2007, p. 42).

Currently, science has been removed from groundwater issues at ORR. In the 1980s, ORNL had a world-class groundwater monitoring center. But large contracts between DOE and conglomerates of companies have separated science performed at the national labs (i.e., ORNL) from the cleanup. A long-term funding mechanism that would support the type of science

______________

³ See http://iemhub.org/groups/iscmem/, accessed on February 19, 2014.

required to address these challenges is needed. In the 1980s and 1990s the USGS participated with ORR on groundwater issues but the USGS is no longer involved. The USGS is another federal government resource that could help with communication to provide unbiased analysis of data and technical information to stakeholders.

Dawn Wellman agreed that science and functional monitoring are needed. Most monitoring at sites is groundwater based—having additional monitoring information, for example, from the vadose zone would provide another source of information on little-known processes within that region.

4.3 INFRASTRUCTURE ISSUES AND TECHNOLOGY APPROACHES FOR ACHIEVING SUCCESSFUL “ROLLING STEWARDSHIP”

Dave Geiser

The Office of Legacy Management was established in 2003 to carry out the long-term stewardship for DOE sites at which active remediation had been completed. LM’s mission covers long-term surveillance and maintenance of sites, records and information technology (IT) systems, community outreach, and management of lands and assets. LM began with 33 sites, currently manages 90, and is expected to have 120 sites under its management by 2020. The sites are spread over 28 states and territories including Puerto Rico and Hawaii. Nine of these sites require onsite personnel for management, the rest require periodic visits. All sites need record management services. Almost one-third of LM sites have beneficial uses, which include grazing, wildlife, industrial, and/or community use. Previously, DOE focused efforts on keeping the public (and others) away from the nuclear weapon development facilities, but now the public is encouraged to visit many of those same sites. It is a testament to DOE’s Office of Environmental Management’s (EM’s) efforts that some of these sites are now useful community resources (see Fernald case study below).

The largest cost for LM’s long-term stewardship activities is groundwater monitoring—with 2,000 wells and approximately 7,000 samples collected for analysis annually. Record keeping and IT system maintenance is another significant effort—storing 100,000 cubic feet of records, responding to 1,500 requests/year and getting 1,800 hits/day on LM’s Geospatial Environmental Mapping System (GEMS) website.⁴

The following eight sites serve as case studies to highlight the breadth of issues addressed by LM:

______________

⁴ See http://gems.lm.doe.gov/imf/sites/gems_continental_us/jsp/launch.jsp.

• Weldon Springs, Missouri,
• Rocky Flats, Colorado,
• Amchitka, Alaska,
• Pinellas, Florida,
• Tuba City, Arizona,
• Wayne Site, New Jersey,
• Mound, Ohio, and
• Fernald, Ohio.

Several of the case studies highlighted noteworthy aspects of LM management challenges.⁵ LM encourages public use of the sites. The Weldon Springs site is a good example (see Figure 4-2). The site was originally used by DOE for uranium processing and U.S. Army explosives production. The main contaminants were uranium, nitrates, and trichloroethene (TCE). The remediation of the site created a 45-acre disposal cell (the cap can be seen as a large, white pentagon in Figure 4-2). Long-term closure controls include groundwater monitoring, monitored natural attenuation, and institutional controls. Current land use includes a visitor’s center and hiking/biking trail. Weldon Springs is adjacent to a large public high school. The track team uses the disposal facility cap as a hill for part of its workouts. A raised, viewing platform erected by LM is a local sightseeing attraction because it allows for clear views of the surrounding area.

LM also encourages cooperative agreements with local stakeholders. The site in Amchitka, Alaska (see Figure 4-3) (part of the Aleutian Islands) is LM’s most remote site. Three underground nuclear weapons tests resulted in tritium contamination of the site. Long-term controls of the site include groundwater, surface water, and biota monitoring (on a 5-year cycle). The site is within a wildlife refuge managed by the U.S. Fish and Wildlife Preserve. LM has a cooperative agreement with the Aleutian Pribilof Island Association and the local Aleutian tribes for assistance with logistics and travel to/from the site. Another example of a site with a cooperative agreement with a local stakeholder is one in which a local land-owner maintains fencing around the DOE site bordering his land in exchange for harvesting hay from the site.

LM’s most densely populated site is the Pinellas, Florida site (see Figure 4-4). The site was originally used by DOE to manufacture electrical systems and electronics packages. The main contaminants are organic solvents and metals; there are no radioactive contaminants. Long-term controls include groundwater monitoring, monitored natural attenuation, and institutional controls, including the purchase of subsurface rights to

______________

⁵ For more information on each of the case studies, see the full presentation at: http://dels.nas.edu/resources/static-assets/nrsb/miscellaneous/Geiser.pdf.

FIGURE 4-2 Aerial photo of the current Weldon Springs site. The neighboring high school’s tennis courts can be seen in the lower right corner of the picture (see text).
SOURCE: Geiser 2014.

adjacent parcels. The land was sold to Pinellas County for redevelopment purposes in 1997. It is currently an active industrial area with higher employment than when operated by DOE. A large, groundwater plume under the site extends into adjacent properties. The plume is fed by a subsurface contaminant source located beneath one of the largest commercial buildings on the site (Building 100, see Figure 4-4). The plume cannot be remediated without demolition of the building. In this case, a final decision was made to treat the plume without addressing the source term so that the commercial property may continue to generate revenue and support the local economy.

The Tuba City, Arizona, site was presented as an example of a sustainable solution at a remote location with challenging tribal nation rights issues (see Figure 4-5). The site contains a former uranium mill, which processed approximately 800,000 tons of uranium ore resulting in uranium and radon gas contamination. Wastes remaining onsite are stored in a 50-acre disposal cell. Long-term stewardship includes groundwater monitoring

FIGURE 4-3 The Amchitka, Alaska, site is LM’s most remote site.
SOURCE: Geiser 2014.

and a groundwater pump and treat remedy. Located in sparsely populated desert, the site uses photovoltaic power to support groundwater well remediation and monitoring. Cultural challenges include treaty rights and language barriers with the local tribal nations (e.g., Navajo Nation). The desert terrain and climate add to the challenges including severe storms, road washouts, increased sedimentation, and potential impacts to disposal cell integrity.

Lessons learned over the 10 years of LM’s experience with these sites can be separated into “site lessons learned” and “programmatic lessons learned.”

Site Lessons Learned:

• Understand and work within the regulatory structure for each site. Comply with the remedy decision but also consider local, applicable laws and regulations.

FIGURE 4-4 The Pinellas, Florida, site is LM’s most populated site. The groundwater plume is fed by a contamination source located beneath Building 100 (see text).
SOURCE: Geiser 2014.

• Local requirements may conflict with federal ones (e.g., locals want weeds removed from covers but it is not required by federal).

• Stay current with changes in demographics, local land use, and politics. A remedy designed for one set of conditions may not be protective over time.

• Maintain relationships with the local community. During site visits, plan to visit the mayor, the fire department, and/or the police/sheriff. Understand who is buying or selling property near the sites, and whether there are plans to build in the future.

• Create a transparent culture where you invite the public to learn about the site. Engage neighboring land owners, local governments, and the community. Do this at the frequency, and with the communication tools and formality, that the community desires.

• Communicate with the public. An interesting rule of thumb: “The more you communicate, the less interested the public becomes.” This is a challenge when one of LM’s goals is to encourage site use by the public.

• Adjust communication strategies to fit the site and the local community. An example is Riverton, Wyoming, where the local community socializes at the local bars. With some hesitation, the site managers and LM did go to the bars in order to reach the locals (but did not drink alcohol).

• Pursue beneficial reuse of the site within the restrictions of the environmental remedy. Informed site use can enhance the effectiveness of other long-term stewardship activities.
• Establish institutional controls (ICs) that can be effective when properly managed and applied. Legally enforceable controls should be used when necessary. In many cases, administrative ICs are enforced by local governments; effective partnership is important.

• Although they should be used, ICs should never be the first level of protection. Also know that at the local level, the police respond to intruders at the site—not the federal government.

• Long timeframes mean that natural events must be considered in remedy selection and in modeling site performance. Assume the subsurface models are likely to be improved (e.g., when the 100-year flood occurs it will impact the existing site model).
• Plan for, and be able to transition, site managers. Make knowledge management a priority.

• Knowledge management is critical. Educate, train, and empower your staff.

Programmatic Lessons Learned:

• Maintaining the remedy means maintaining infrastructure (roads, vehicles, power) at each site.
• Using common approaches across sites saves cost and improves performance.
• Long-term protection of human health and environment requires knowledge of changes in regulations, policies, knowledge, and advances in technology.
• Encourage cross-site teams, share information, and promote lessons learned. Rotate staff through all sites so they have knowledge of every site and are familiar with the local population. However, the staff rotation period should be no more than 5 years at a single site.
• Records are important—more than most people realize. Consolidate to one location. Make digital copies only when pulling a file (it is not cost-effective to digitize all records). Use commercial software for records management.

• Make all data and reports available to the public (internet and public meetings) following appropriate quality assurance reviews.
• Use data to improve scientific understanding (e.g., RCRA degradation data from 5-year reviews). Invest in science and technology in targeted areas that impact remedy performance over the long term. Share the results with the public and the scientific community.

4.4 SUMMARY OF DISCUSSION SESSION

Institutional Controls and Timeframes. Rateb (Boby) Abu-Eid (USNRC) noted that the USNRC assumes a timeframe in its performance analyses to determine when to move from active to passive remediation. This is currently under debate—whether it should be 100 or 300 years. He asked what timeframes LM has assigned, if any, to its decisions. Mr. Geiser responded that nearly all of the ICs at LM sites are passive. Most are realty instruments such as deed restrictions or access agreements, which provide legally enforceable controls at the sites. The ICs are not assigned a specific duration. Other types of passive ICs, such as fences and signs, require maintenance. Active ICs are the most expensive, for example the visitor centers at the Fernald and Weldon Springs sites. They are part of the ROD and will remain there as long as they are useful to the communities. It is difficult to predict public use and perception into the future.

Models. Dr. Abu-Eid next asked about how monitoring data are used in models. Mr. Geiser responded that in general the models work well and that there is very little movement of the contaminants at their sites.

Successful Public Outreach. Kevin Crowley (National Academy of Sciences staff) noted that one of LM’s lessons learned was that the public seemed to lose interest when information about the sites was made readily available. This was seen at WIPP [Waste Isolation Pilot Plant]. This could be a signal of public trust and could be considered a “success” that should be considered a best practice. Mr. Geiser agreed with the observation about the public’s level of interest.

Mechanisms to Change to Stewardship Activities. Dr. Crowley further noted that cleanup and stewardship are dynamic and that change should be expected. Consequently, the public should be prepared for change as a normal part of the process. However, the public currently sees change as a way to “get around” existing agreements. Mary Flora’s Core Team Process presentation outlined a process that allowed decision makers to revisit issues as more data became available and knowledge increased. He asked

whether LM had a similar process and whether there been any notable changes to stewardship at the sites during its 10 years of operation.

Mr. Geiser responded that LM has a process to drive stewardship change. This process has two fundamental goals that may seem initially to be in conflict with each other. The first goal is to attain 100 percent regulatory compliance at all of the sites. The second goal is to reduce operating and maintenance costs (surveillance and maintenance) by 3 percent every year. All LM site managers, contractors, and program manager have these two goals as common performance measures in their evaluations every year and, as an agency, measures against these goals are reported every year.

These goals are driving the optimization of existing remedies at all LM sites. Numerous examples of stewardship changes have come about as a result. Monticello has been previously highlighted by others in this workshop. Many pump-and-treat systems have been transferred to monitored natural attenuation. Another site has eliminated monitoring altogether. And the total number of wells that are monitored per year at LM sites is declining (more wells are decommissioned every year than are added).

4.5 THE CORE TEAM PROCESS: MAKING RISK-INFORMED DECISIONS FOR ON-SITE MONITORING

Mary Flora

The Core Team Process was introduced at the first workshop. In this workshop, the SRS site-wide monitoring program is presented as a case study to highlight how risk has been used in the decision-making process. The process is designed to promote consensus-based decisions on complex remediation issues at the SRS site. Over the past few years, SRS has been moving from the front-end (determining the remedies) to the back-end (long-term management) of the remediation process.

The SRS cleanup began in the 1980s. In 1989, SRS was added to the National Priority List. The FFA was signed in 1993 by DOE, EPA, and South Carolina’s DHEC. Cleanup activities have focused on soil and groundwater, deactivation and decommissioning of inactive facilities, and closure of tanks. Most of the work is overseen by DHEC and EPA. The USNRC is also involved in waste tank closure.

Although the FFA was signed in 1993 it took several years to figure out how to work together effectively. In 1999, the remediation progress had reached a standstill—communication during negotiations was at an all-time low and a new process was needed. The Core Team Process was established to improve communication between the FFA signatories.

The Core Team operates using four principles of environmental restoration:

1. Building an effective Core Team is essential.
   All participants have authority to make decisions. Their management has entrusted them to make decisions that will not be subsequently overturned within their management structure.
2. Clear, concise, and accurate problem definition is critical.
   Scoping documents are shared among all Core Team members before meetings. These documents lay out the problem, the facts behind it, and the topics on which decisions will need to be made.
3. Early identification of likely response actions is possible, prudent, and necessary.
   Take incremental steps to solve a problem, recognizing that there may be a need to take further action later.
4. Uncertainties are inherent and will always need to be managed.
   This principle allows the team to admit mistakes and modify original decisions when needed.

These principles are posted in the meeting rooms to serve as reminders for all parties. When these steps are followed, the final decisions are embraced by all parties. The benefits include a traceable and documented history of decision making; an increased and common understanding of the link between decisions and technical activities; a clear understanding of knowns and unknowns; and an increased confidence that issues will be identified and managed through time. Decisions can be defended. Decisions “stick” for a longer time because of member buy-in, a common understanding of the problem(s), and well-documented background material. The output of the Core Team Process is timely communication, agreements based on trust, and documentation outlining the scientific and technical underpinnings of each decision.

The current status of the soil and groundwater remediation program at SRS can be gauged by the number of operable units that have been addressed to date. Of the 515 total operable units, 400 have completed remediation. Those that remain have complex and challenging problems (see Figure 4-6). In addressing soil and groundwater contamination, there is continual pressure to move, when appropriate, from active to passive systems. Currently 30 passive and 9 active remediation systems exist at SRS.

More than 2,500 monitoring wells have been installed throughout the site. Some have been in operation for more than 20 years. Their depths vary from 2 to 350 feet. They have been installed during different remediation campaigns without a site-wide plan or systematic approach. For example, Area M has more than 50 compliance wells (see Figure 1-2). Because of the combination of several factors, including the history of the groundwater monitoring program, the amount of data collected, and the lack of a site-wide groundwater monitoring plan, an optimization program

was proposed. DOE approached the other Core Team members, EPA, and DHEC with a proposal to review the well monitoring program, citing operating Principle 4: Uncertainties are inherent and will always need to be managed. The Core Team recognized the opportunity to improve current groundwater monitoring activities and accepted the proposal to evaluate a set of five topics (Principle 2: Clear, concise and accurate problem definitions are critical):

1. Monitoring vs. Regulatory Requirements
   Does the current monitoring program meet regulatory requirements? Do any of the proposed changes meet regulatory requirements?
2. Spatial Optimization
   Are the current well locations optimal for monitoring? Are there locations where no wells exist but should?
3. Temporal Assessment
   Is the data collection frequency appropriate to gauge performance of the remedy?
4. Analyte Assessment
   What analytes should be analyzed to assess remedy performance? Has the list changed over the lifetime of the well? Should any be added or removed from the current list?
5. Reporting Assessment (frequency)
   Should reports be produced annually, semiannually, or at a different frequency?

Each topic was scientifically analyzed, and the results were shared among Core Team members prior to the meetings in which decisions are made. The consideration of each topic during the meetings was based on scientific results, and decisions were made based on a goal to improve the monitoring program. All analyses and decisions were documented.

Consensus was reached, and final decisions were made and documented for all five of the proposed topics.⁶ In some cases, the monitoring requirements were relaxed (e.g., fewer wells were needed in some locations, lower frequency reporting and measurement could be made, and modifications were made to regulatory documents as needed), but in others it was found that additional monitoring was needed (e.g., increased sampling rate for wells near a remedy undergoing change, or installation of additional wells in locations lacking monitoring).

The site-wide groundwater monitoring optimization program is a multiyear effort. Further optimization will continue for the next few years. So far,

______________

⁶ Details on the analysis and results can be found within the presentation: http://dels.nas.edu/resources/static-assets/nrsb/miscellaneous/flora.pdf.

the decisions have held. The lessons learned are the following: engage the team members early, and base decisions on technical evaluations to which all team members have equal access. Based on the success of the Core Team Process for improving the site-wide groundwater monitoring program, it has recently been expanded into liquid waste management at SRS.

4.6 SUMMARY OF DISCUSSION SESSION

Additional Case Studies Involving the Core Team Process. John Tseng (retired DOE) expanded on Mary Flora’s presentation and provided two additional examples of Core Team Process success stories. Tank closure at SRS usually takes 4 to 5 years. (Tanks 18 and 19 at SRS took this long to close.) The Core Team recognized that an alternative process would be needed if FFA milestones were to be met for the remaining tanks. The Core Team established an accelerated process, which reduced the closure time by almost a full year; currently it is being used for Tanks 5 and 6. The other example is the development of a Quality Assurance (QA) program for residue characterization from SRS’s hot cells to meet DHEC expectations. No protocol governed the QA process and analysis for sampling residue from hot cells. The Core Team Process was used to develop a QA sampling program with a step-by-step procedure supported by the full team. The new QA procedure is in the final stages of testing and validation.

Move from Active to Passive Remediation. Michael Kavanaugh (Geosyntec) asked whether there was a well-defined process for moving from active to passive remedies and whether there was push-back from the regulators when passive remediation was proposed. Ms. Flora responded that a change from active to passive remediation is considered when an active system has either met its cleanup goal or has lost its effectiveness. Active systems are expensive; they use power, and they have many moving parts that can fail. In moving to passive systems the Core Team considers options such as solar power or selecting equipment with fewer moving parts. SRS works directly with the regulators and is clear about the desire to move from active to passive systems. Sometimes the regulators will identify active systems that have become ineffective and are no longer a good return on investment.

Models. Rateb (Boby) Abu-Eid (USNRC) asked how monitoring data are used to develop models. Ms. Flora explained that active modeling is performed for all well networks at SRS. SRNL, an independent entity, performs the modeling. The regulators have their own models. The SRNL and regulator models are validated and compared against each other, which

provides further confidence among team members that the data and models are good.

Challenges of Expanding the Core Team Process to Other Sites. Bill Levitan (EM) noted that DOE has expanded the Core Team Process into at least eight different sites, but its successful adoption is dependent on the individual team members. It requires buy-in from everyone on the team for it to work. The Core Team Process was introduced but not successfully implemented at the Hanford site, for example.

Monitoring Data for Improving Site Models. Patricia Culligan (Columbia) tried to bridge the two sets of Session 3 talks by connecting the concept of functional monitoring, testing, and modeling proposed by Craig Benson to LM stewardship and SRS’s groundwater monitoring programs. Dr. Culligan suggested that LM’s data sets might be used to build confidence in existing models for the regulators and the public. Although the increased monitoring and data collection would temporarily increase LM’s budget, it would likely pay benefits in the future. Paul Gilman (Covanta Energy) supported this idea and noted that expanded monitoring would only be needed for systems for which data are not already available. The functional monitoring data for these systems would fill an important knowledge gap.

Dave Geiser responded that additional data are collected on the surface for many of the existing remedies. For example, onsite inspectors walk along detailed transects (inspections following a defined grid) across disposal cells to confirm the containment systems are working as expected. For the Fernald and Weldon Springs sites, leachate measurements are taken to confirm a decrease of leaching in time. Furthermore, there have been studies of soils near disposal cells to monitor changes in contaminant levels. Whereas these individual monitoring efforts are not agency-wide and are not as extensive as the measurements taken by Dr. Benson’s University of Wisconsin group, they do add additional information when the remedy is not working as expected.

This page intentionally left blank.