Health and Environmental Protection Standards for Uranium and Thorium Mill Tailings
Groundwater is one of our nation’s most precious resources. A significant portion of the U.S. population draws on groundwater for its potable water supply. In addition to serving as a source of drinking water, people use groundwater for irrigation, stock watering, food preparation, personal health and hygiene, and various industrial processes. When that water is radioactively contaminated, each of those uses becomes a radiation exposure pathway for people. Groundwater contamination is also of concern to us because of potential adverse impacts upon ecosystems, particularly sensitive or endangered ecosystems. For these reasons, it is a resource that needs protection.
A number of federal and state laws have been passed through the years to protect drinking water. At the federal level, the SDWA (discussed in detail in Section II.F.1) establishes the basic framework for protecting the drinking water used by public water systems in the United States. This law contains requirements for ensuring the safety of the nation’s public drinking water supplies. At the state level, many similar drinking water and water use laws have been passed.
Groundwater is also a valuable and dwindling resource, particularly in western states where most ISR activities are anticipated. EPA views protecting groundwater as a fundamental part of its mission. Particularly in cases where groundwater is directly threatened by an activity, as it is by the ISR technology, EPA believes it has a special duty to ensure that the authority of all applicable federal statutes (e.g.,
UMTRCA and the SDWA) are used to help protect the groundwater and that appropriate standards to protect public health, safety and the environment are developed and implemented.
We anticipate the objection that the presence of uranium deposits typically results in groundwater of poor quality, and not a pristine source of drinking water. We recognize that this is often the case, and that the volume of water affected by the mineralized zone may be significant. We do not, however, see this as a reason to allow this groundwater to be further degraded. The increasing scarcity of groundwater is leading some communities to consider using sources of water that previously would have been considered non-potable, using advanced treatment to make it suitable for livestock or human consumption. Since such advanced treatment may not be economically feasible for some communities, it is all the more important to prevent, as much as reasonably possible, additional degradation of the groundwater.
A guiding philosophy in radioactive waste management, as well as waste disposal in general, has been to avoid imposing burdens on future generations for clean-up efforts as a result of management approaches that are reasonably anticipated to result in pollution in the future. Adhering to the concept of sustainability, we should not knowingly impose undue burdens on future generations. Imposing performance requirements that avoid polluting resources that reasonably could be used in the future, therefore, is a more appropriate choice than imposing clean-up burdens on future generations. ISR facilities use significant volumes of water during both operations and restoration. We believe it is reasonable to make every effort to ensure that ISR activities leave groundwater in no worse condition than pre-ISR operational status.
A. How does today’s proposal relate to existing 40 CFR part 192?
In 1983, EPA promulgated regulations at 40 CFR part 192 in response to the statutory requirements of UMTRCA. At the time, uranium recovery from ore was done almost exclusively by conventional milling processes, where at most a few pounds of uranium were recovered for each ton of ore mined and processed. The wastes from the milling process (the tailings and raffinates, i.e., uranium byproduct materials) were disposed of in large piles on the surface at mill sites, posing contamination risks to surface water, groundwater, and soils, both on and off site. Liquid wastes were often discharged into rivers. Contaminants of concern consisted primarily of radionuclides and non-radioactive metals, radon gas and organics. Concerns that these tailings piles would be a continuing source of radiation exposure and environmental contamination unless properly reclaimed and managed were the driving force behind the passage of UMTRCA. The statute’s intent was to contain tailings in engineered impoundments to prevent the further dispersion and misuse of the material. This measure would also protect uncontaminated aquifers from becoming contaminated by the uranium mill tailings impoundments and prevent radon emissions through performance specifications for radon barriers (covers). Because the major environmental risk at that time was perceived to come from the conventional uranium mill tailings, which already existed in large volumes, other uranium recovery technologies, including ISR, received little attention.
As stated earlier, ISR has surpassed conventional milling as the dominant form of uranium extraction in the United States and is expected to predominate in the future. The ISR process presents different environmental concerns from conventional milling. ISR does not generate large volumes of solid waste materials or require permanent tailings impoundments. The ISR process does, however, directly alter groundwater chemistry, posing the challenge of groundwater restoration and long-term subsurface geochemical stabilization after the ISR operational phase ends. With ISR, the “milling” of uranium ore is performed within the ore zone aquifer by injection of lixiviants. As stated earlier, the lixiviants can also liberate other elements, particularly metals that are often found co-located with uranium deposits. Their migration outside the production zone can potentially contaminate surrounding aquifers. Furthermore, when processing of the ore zone is no longer economically viable, ISR operators can release the site for future use, either by selling the land or returning the property to the original owner. The operators are required to restore the aquifer to its original geochemical conditions, to the extent possible, and to show some level of stability in the geochemistry of the production zone before terminating the license and making the site available for other uses. Whereas conventional mill tailings piles are under perpetual institutional control, current NRC regulations allow for ISR sites to terminate their licenses, essentially ending regulatory oversight of the site.
Today, EPA is reaffirming that ISR facilities are subject to the 40 CFR part 192 requirements. We seek to provide clear direction on how to monitor groundwater in and around the production zone during all phases of the ISR facility’s lifecycle, and how to demonstrate geochemical stability at these sites.
We believe there has been some uncertainty about how to apply the current standards, which are more targeted to conventional mills, to ISR sites. In addition, there has been confusion about applicability of UMTRCA restoration requirements at aquifers that have been exempted from the standards of the SDWA. With the prospect of additional ISR facilities beginning operations, we believe it is necessary to clarify these issues. Therefore, we are proposing additional groundwater protection provisions to 40 CFR part 192 that are specific to uranium ISR facilities and consistent with the SDWA and RCRA. We believe these provisions are necessary to ensure that ISR sites are not released from regulatory control until it can be reasonably demonstrated that groundwater will not degrade over time.
Specifically, we are proposing provisions that will result in long lasting protection of surrounding aquifers. The provisions specify how to determine preoperational background conditions that will be used to set appropriate restoration goals, applicable standards and alternate concentration limits. We are also proposing specifications for long-term groundwater stability monitoring and a corrective action program that is triggered if excursions/exceedances do occur. We view these as the key elements in ensuring that ISR sites do not become a source of continuing or widespread contamination after the ISR operation is terminated.
Sufficient data must be collected to characterize the conditions existing within and outside the proposed production zone to set appropriate groundwater protection standards (i.e., restoration goals) that account for the variability in geochemistry frequently encountered in mineralized regions. Subsequent to the end of uranium production, the regulator must ensure that alternate standards are approved only after restoration has been attempted and it is clearly demonstrated that the initial groundwater protection standard(s) cannot be achieved, or once achieved, cannot be maintained. Such approval should take place only after the operator has made reasonable and satisfactory efforts to achieve and maintain the initial standard(s) and fully considered a number of factors. Whether the initial goals are met or alternate concentration limits are approved, conditions must be shown to be stable and groundwater quality must not degrade over time, as is possible when: lingering amounts of lixiviant solution remain in isolated pockets within the wellfield; reducing conditions are not fully reestablished; and/or the long-term stability monitoring period is too short compared to the time it takes for groundwater to move through the aquifer. Therefore, the operator must monitor groundwater at the site for a sufficiently long period after restoration is complete and use statistically significant results to provide a reasonable demonstration that long-term stability has been achieved. This demonstration can include geochemical modeling to confirm the persistence of stability of the groundwater chemistry. Geochemical modeling can provide a defensible demonstration of an aquifer’s natural capacity to maintain stability, which statistics alone cannot provide. Although the selection and application of geochemical models will be on a site-specific basis, geochemical models that have been used to predict the fate and transport of uranium at ISR facilities include PHT3D, PHREEQC, and PHAST.
We intend for today’s proposal to eliminate any confusion about the relationship of the aquifer exemption process to restoration requirements at ISR sites. We further recognize that the application of the existing standards in 40 CFR part 192 to ISR sites is not as straightforward as it could be. Nevertheless, we believe there is sufficient information available to indicate that practices related to groundwater protection at ISR facilities have not been sufficiently rigorous to provide confidence either that groundwater is being restored appropriately or that such restoration will persist into the reasonably foreseeable future. 51 52 53
We believe today’s proposal addresses these issues in a manner that is both logical and implementable;
we solicit comment on our view of the current situation and the overall approach of our proposal.
B. What groundwater protection standards are we proposing for ISR facilities?
We are proposing today to establish groundwater protection standards consistent with those applied to conventional mills in 40 CFR part 192, subpart D. That is, the licensee will use as the applicable standard during restoration and long-term stability monitoring either (1) the background concentrations of groundwater constituents measured prior to the start of the ISR operational phase; or (2) a specified regulatory level, whichever is higher. In certain circumstances, the licensee may request that the regulatory agency approve an alternate concentration limit.
1. Generally Applicable Groundwater Standards
We emphasize again that the groundwater protection standards currently found in 40 CFR part 192 apply to ISR sites. These standards address both radiological and non-radiological constituents. The standards applicable to non-radiological constituents adopted the requirements for groundwater monitoring at RCRA hazardous waste sites.
These generally applicable standards were originally based upon EPA’s 1976 Maximum Contaminant Levels (MCLs) in drinking water (40 CFR part 141).
See section II.F.1 of the preamble and footnote for background. EPA further specified radiological and non-radiological constituents of concern at mill tailings sites. Following the same approach, we are proposing today to specify, as Table 1 to subpart F, the constituents that must be monitored at ISR sites, as appropriate. The required constituents mirror those included in Table 1 to subpart A of 40 CFR part 192, with the exception of the six pesticides.
We are not proposing to establish new numerical standards in the rule. EPA’s preferred option for carrying over and updating the groundwater protection standards in the new ISR-specific subpart F is to incorporate, by reference, the most protective standards issued under the SDWA (40 CFR 141.61, 141.62, 141.66, 141.80 and 143.3), values from RCRA standards (40 CFR 264.94), and the maximum constituent concentrations found in Table 1 to subpart A of 40 CFR part 192. By incorporating these standards by reference, the new subpart F would automatically update if those concentration values change in the standards under SDWA or RCRA and thereby, be self-implementing. Upon promulgation, licensees currently in restoration, stability monitoring or long-term monitoring at a given wellfield at a licensed facility would continue to be held to the standard(s) in place at the time of licensing for those given wellfield(s), unless the regulatory agency determines otherwise. Operating wellfields, new wellfields and expansions of wellfields would be required to meet the newly promulgated standards. This option would make the groundwater protection standards under the proposed subpart consistent with all relevant current and future standards under SDWA and RCRA. We believe that this approach will more effectively keep the groundwater protection standards current with the Agency’s policies while providing for regulatory certainty. The standards in the existing portion of 40 CFR part 192 are outdated for arsenic and uranium,
Today’s proposal would update the standards for arsenic and uranium as they apply to ISR facilities. Should the Agency propose to update its MCLs or RCRA standards at some point in the future, stakeholders will have the opportunity to comment on the potential impacts to ISR activities.
We are also considering the alternative approach of placing a static table of restoration goals in the new subpart F. The table would list the 13 required constituents for which groundwater protection standards must be met, and also provide the specific numeric concentration value associated with each constituent. If this option is promulgated in the final rule, the standards would not automatically update with any future changes to standards under the SDWA or RCRA but would remain static. Under this approach, the Agency would initiate future changes to standards through a notice-and-comment rulemaking specifically for 40 CFR part 192.
In order for an ISR operation to proceed, a UIC permit is required and typically, an aquifer exemption is needed as well. The exemption effectively removes from the protection of the SDWA, an aquifer or portion of an aquifer that would otherwise meet the definition of an underground source of drinking water. The wellfield used by the ISR operation to extract the uranium deposit may constitute only a portion of the overall exempted area. As noted in Section II.E.1 of this document, there is no similar exemption for the aquifer from the requirements of UMTRCA, nor does UMTRCA contemplate such a concept. We emphasize again that the SDWA-based aquifer exemption does not relieve the operator of an ISR facility of the obligation to remediate environmental contamination resulting from activities regulated under UMTRCA, both within and outside the exempted portion of the aquifer.
2. Alternate Concentration Limits (ACLS)
Consistent with RCRA, EPA currently allows the use of ACLs if the operator is unable to restore groundwater to either preoperational background conditions/concentration levels or the applicable restoration goals. Today we propose to clarify the requirements for requesting and granting ACLs in the production zone, after restoration efforts have taken place. While the 19 criteria to be considered in granting ACLs are spelled out for Title II sites in 40 CFR 192.32(a)(2)(iv) through incorporation of 40 CFR 264.94(b), they have not always been implemented as intended. 59 60
In the past, NRC and Agreement States have issued secondary class-of-use restoration goals at ISR sites, but these goals were typically less restrictive than meeting background concentration levels.
NRC no longer recognizes class-of-use as an appropriate standard for restoration of groundwater at uranium ISRfacilities;
secondary class-of-use restoration goals are inconsistent with the requirements of 40 CFR part 192 and 10 CFR part 40, Appendix A. There is evidence that relaxed restoration standards have been granted in Agreement States,
and some instances where ACLs have been identified and approved by the regulator before restoration efforts have been initiated and/or completed. 64 65
We believe these situations can result in insufficient protection of groundwater; in particular, we believe it only is appropriate to establish restoration goals based on a thorough characterization of the preoperational environment and not to approve ACLs unless it has proven impracticable to achieve or maintain the initial restoration goals or return to background conditions after restoration. With this proposal, we specify the conditions that must be met prior to requesting an ACL and emphasize the factors that must be considered in establishing and approving ACLs. These factors specify that, if ACLs are deemed necessary or appropriate after all best practicable restoration activities have been completed, they must not pose a substantial present or potential hazard to human health or the environment.
ACLs can be established for carcinogens and/or non-carcinogens. When considering the potential for health risks caused by human exposure to known or suspected carcinogens, ACLs should, where practicable, be established at concentration levels that represent a cumulative excess lifetime risk to an average individual at no greater than 10 −4 (one chance in ten thousand).
The regulatory agency may face situations in which the operator will request ACLs. If after extensive effort the operator determines that the initial restoration goals for one or more constituents cannot be achieved as required in the license, the operator may request and the regulatory agency may approve the levels that have been achieved as provisional ACLs and determine that restoration is complete (i.e., that there is no statistically significant trend in the concentrations of regulated species over time). Then, the operator may request and the regulatory agency may approve final ACLs if post-restoration monitoring indicates three consecutive years of stability at the 95 percent confidence level. The approval of final ACLs, however, would not by itself satisfy the requirements for long-term stability monitoring.
In the second case, after restoration is complete, the operator may find that post-restoration monitoring detects increases in the concentration of one or more constituents of concern. Depending on the statistical significance of the increase, the regulatory agency may determine that further attempts at restoration or corrective action are needed. If the situation persists after such action is taken, the regulatory agency may choose to wait and see if the increase levels off (i.e., stabilizes). If stabilization does occur, the operator may request and the regulatory agency may approve final ACLs if post-restoration monitoring indicates three consecutive years of stability at the 95 percent confidence level.
An additional consideration is the potential effect of ACLs on groundwater downgradient of the wellfield. The granting of ACLs could be viewed as inconsistent with the purpose of groundwater restoration, which is to prevent contamination of groundwater resources beyond the production zone. However, NRC has in recent years adopted an approach defining the “point of exposure” as the aquifer exemption boundary, where the initial restoration goal must be met. We propose to adopt a similar approach today.
This will ensure that the non- endangerment condition of the UIC permit will be sustained. We believe the decision to grant an ACL is among the most important that the regulatory agency can make. We believe our proposal appropriately clarifies the situations in which ACLs can be considered and emphasizes the factors that must be considered, thereby making the overall process more rigorous. However, we also recognize that the regulatory agency may need to spend additional effort to evaluate the full record of activities at the site in order to determine whether an ACL is appropriate, and at what level. Because the long-term protectiveness of this decision may not be fully understood until well after site activities conclude and the license is terminated, we encourage the regulatory agency to inform and seek input from the affected public when ACLs are being considered. We believe this request would constitute a license amendment significant enough to warrant an opportunity for public comment, if not public hearings.
C. Adequate Characterization of Groundwater Prior to Uranium Recovery
To design and operate an ISR facility, the chemical composition and hydrology of the groundwater in and around the ore body must first be rigorously characterized. Defining the configuration of the ore zone and designing the production zone for uranium recovery requires detailed subsurface information obtained from geophysical investigations, including but not limited to logs and cores.
In addition, the groundwater in the production zone is also characterized to determine the proposed chemical composition of the lixiviant and to determine background groundwater chemistry by which to set restoration goals for the post-production phase of the ISR operation (i.e., the efforts to return the groundwater chemical conditions in the production zone to those that existed prior to the uranium recovery efforts). The preoperational chemical composition of the groundwater in the production zone is called “baseline” in practice within the ISR industry and by NRC. In EPA documents and regulations the term “background” is used to indicate the original state of groundwater before activities take place that may introduce contamination into the groundwater, such as leakage of contaminants from a surface or near-surface waste disposal cell or an underground source of contamination such as leaking storage tanks or disposal wells.
For the ISR method, there are a number of “backgrounds” involved, the most important being the preoperational background within the portion of the ore zone where uranium production will take place (i.e., the production zone). Knowledge of this background is necessary to design the leaching process and set restoration goals—two very important steps in the ISR operation. “Background” groundwater composition data are also needed in portions of the aquifer surrounding the wellfield and in overlying and underlying aquifers that may have communication with the uranium ore-bearing aquifer to determine whether excursions occur during operations, and to determine whether seasonal variations in groundwater chemistry are occurring in shallow aquifers. Background data are also needed for geochemical modeling of the groundwater in the production zone and downgradient to support assessments of the natural capacity of the restored production area and downgradient portion of the exempted aquifer to maintain long-term stability of the restored wellfield.
There are spatial and temporal designations for the various “backgrounds” measured in relation to an ISR operation. For instance, preoperational background is determined above, below, around and within the wellfield in the exempted aquifer. The preoperational background downgradient of the wellfield and in aquifers above and below the production zone are needed to detect any excursions that may occur during the ISR operational phase or restoration phase. The uses of the various “backgrounds” are described in the technical background information document supporting this rulemaking.
NRC requires establishment of background at uranium recovery sites in its regulations at 10 CFR part 40, Appendix A, Criterion 7;
most of the implementing requirements are found in guidance and license conditions. Today’s proposal includes provisions to ensure that operators adequately characterize preoperational conditions inside and outside the wellfield. This characterization is necessary to establish appropriately protective restoration goals that are representative of the wellfield, accounting for natural variability. There is evidence that regulators and operators have at times used high-end values to represent the overall wellfield or have used a generalized “class-of-use” for the groundwater to set restoration goals.
We do not believe this is appropriate, as we explain below.
Today’s proposal also specifies that the preoperational groundwater monitoring program must account for the effects of well installation and development on the groundwater characteristics. The physical act of penetrating the aquifer to install the well can cause localized changes in constituent concentrations or chemical parameters, which can lead to a misleading picture of background conditions. This can, in turn, result in selection of artificially high restoration goals. It is important that the operator allow a sufficient interval of time between well installation and sampling to allow localized disturbances to dissipate and ensure that background conditions are accurately characterized.
1. Establishing Restoration Goals
The successful protection of groundwater at ISR sites begins with the selection of rigorous and appropriate restoration goals. As described in Section III.B of this preamble, restoration goals will be established as the preoperational background concentration or as a specified regulatory level for that constituent, whichever is higher.
This is more complicated than it might seem. ISR wellfields may cover areas of 10 acres or more, and the presence of mineralized zones often means that there is significant variability within the proposed production area. As a result, background concentrations in one area of the wellfield may diverge significantly from those measured elsewhere. The question, then, is whether it is possible to select a single level that is representative of the entire wellfield and, if not, how measurements should be evaluated.
We stated previously that we do not believe it is appropriate to select among high-end measurements as representative values for restoration. It might be argued, however, that restoring a given well to its preoperational values would be an indication that restoration would be equally successful in the rest of the wellfield. This might be the case at sites where remediation of groundwater is focused on removing a contaminant that has been introduced from outside the system; however, we question the general application of this assumption at ISR sites. As discussed earlier, the initial deposition (precipitation) of uranium mineralization is uneven and alters the porosity and permeability of the host rock. The extraction and restoration processes at ISR sites are imperfect and further alter the distribution of the uranium in the deposit. Flow paths and velocities in local areas are altered as a result of changes in porosity and permeability that occur from the removal of material from pore spaces and later re-precipitation. It is possible that areas of heavy and lighter mineralization or groundwater concentrations can change from the distribution existing before uranium recovery to that after restoration, reflecting the degree to which the oxidizing and reducing agents contact the mineralization. As a result of these changes, “hot spots” may be found at wells that initially registered lower constituent concentration measurements, and vice versa.
Because of the site-specific nature of this variability, we are proposing today that operators utilize background measurements from across the wellfield, combined with appropriate statistical techniques, to determine restoration goals. As appropriate, goals may be developed for individual wells, groups of wells, or the entire wellfield. The point(s) of compliance for restoration will be determined by the operator and regulatory agency after a thorough technical evaluation of the operator’s geophysical investigation.
During the operational and restoration phases at an ISR wellfield, it is possible that lixiviant or byproduct fluids can escape the capture zones of the extraction wells and move toward the monitoring well ring surrounding the production zone. The placement of the injection and extraction wells, combined with their relative pumping rates, are designed to prevent such movement,
but heterogeneities in the aquifer characteristics and difficulties in maintaining perfect performance of the wellfield can lead to lateral excursions as well as excursions into overlying and underlying aquifers (i.e., vertical excursions). Detecting these excursions is a prime focus of regulatory attention. Indicators of excursions (e.g., increases in concentrations of certain indicator parameters, such as, but not limited to, chloride ion concentrations above the preoperational background) are typically defined in the license conditions, as are requirements for reporting excursions to the regulatory authorities and corrective action requirements once an excursion is detected. The excursion monitoring wells are positioned far enough away from the injection and extraction wells so as to not be affected by the production processes, but close enough to detect excursions in a timely manner. The spacing of wells within the monitoring ring must prevent contaminants from passing between the wells. The excursion monitoring wells should also be far enough from the aquifer exemption boundary to ensure that any necessary corrective action can be taken before a USDW is adversely impacted. We have seen instances where the outer monitoring ring is essentially coincident with the boundary of the exempted aquifer. We do not believe this practice is appropriate. While it may allow the operator to limit the amount of land dedicated to the ISR facility, it provides little margin for error in preventing contaminants from reaching protected aquifers (i.e., USDWs), and may hamper corrective actions should they be needed.
Today we are proposing to adopt a definition of “excursion” consistent with that used by NRC in license conditions. Under this definition, an excursion is identified when two or more indicator parameters are measured at levels exceeding their upper control limits (essentially, background levels) at perimeter monitoring wells or in monitoring wells in overlying or underlying aquifers. Thus, an excursion can take place vertically between aquifers as well as horizontally within the aquifer from which uranium is being extracted.
This approach differs somewhat from that taken under RCRA to detect releases of hazardous constituents, so it is important that we distinguish between the two approaches and explain why our proposed approach is more suitable in the ISR context and consistent with law.
Monitoring under RCRA is conducted to detect any evidence that an engineered hazardous waste unit (e.g., a landfill or impoundment) has failed. To that end, the detection monitoring program includes not only indicator parameters that might signal a change in groundwater chemistry or quality, but also hazardous constituents contained in the waste unit.
The statistically significant detection of any monitored parameter or constituent triggers further investigation and potentially corrective action. Because the engineered unit has been introduced into the environment and the monitoring takes place at the edge of the unit, it is unlikely that a detection can be attributed to the natural variability in the groundwater at the site. Detection of a single parameter or constituent appropriately triggers action in this case because, in addition to remediating groundwater, the failure of the unit itself must be addressed to prevent further releases.
By contrast, at an ISR site all constituents that may be “released” are part of the natural setting, and their presence in groundwater may vary over time. Only the lixiviant is introduced from outside the natural system. Therefore, the “indicator parameters” are typically those that most reflect the lixiviant properties. For example, chloride is often incorporated into the lixiviant as a tracer; similarly, because the lixiviant mobilizes uranium by increasing alkalinity, a significant increase in alkalinity at excursion monitoring wells may signal that lixiviant has escaped the production zone extraction wells. Because the lixiviant typically moves more rapidly than the mineral constituents, increases in the properties associated with the lixiviant will most likely be detected well before the other constituents reach the excursion monitoring wells. The presence of these parameters in the natural groundwater accounts for the reliance on detecting two such parameters at levels above their upper control limits to signal an excursion, rather than only one.
We believe this approach to defining excursions (i.e., relying on two indicator parameters) is reasonable and has been shown to be workable in practice. We are also proposing to define “upper control limit” consistent with NRC’s use of the term. The “upper control limit” defines the level of an indicator parameter that, when two of which are detected at excursion monitoring wells, would signal an excursion; as described above, indicator parameters will typically be identified in the facility license.
It is important that the upper control limits be set appropriately to account for both background levels of indicator parameters and the characteristics of the lixiviant. We agree with NRC that “upper control limit concentrations of the chosen excursion indicators should be set high enough that false positives (false alarms from natural fluctuations in water chemistry) are not a frequent problem, but not so high that significant groundwater quality degradation could occur by the time an excursion is identified.”
We have heard some concerns that upper control limits have in some cases been established at levels that would be unlikely to be exceeded under any conditions, thereby eliminating the possibility of detecting an excursion altogether. Such a situation must be avoided.
Upper control limits can be calculated using various statistical techniques, but are often derived by adding a multiple of the standard deviation to the mean of a distribution. EPA’s Unified Guidance
covers methods that can be used to develop control limits or prediction limits, which serve a similar function. NRC staff describes its current view of acceptable practice in NUREG-1569
: “The staff has decided that in areas with good water quality (total dissolved solids less than 500 mg/l), setting the upper control limit at a value of 5 standard deviations above the mean of the measured [background] concentrations is an acceptable approach.”
The potential for excursions may also be a factor in the facility’s decision to stop operations and enter the restoration phase. In some cases, conventional mills may enter a standby period, in which they stop processing ore with the intent to resume operations at some point in the future (the price of uranium is often the decisive factor in these decisions). In some cases, mills have remained on standby for years at a time. For an ISR facility, however, such a “standby” period is inappropriate because the migration of constituents mobilized by the prior injection of lixiviant continues even if the decision is made to stop extracting uranium. Excursions beyond the production zone are more likely to occur if the gradient within the wellfield is not maintained. In our view, stopping the extraction cycle must be interpreted as an end to the operational phase and should trigger initiation of the restoration phase. We are interested in stakeholder views on this interpretation.
E. Long-Term Stability Monitoring
Perhaps the most significant aspect of today’s proposal involves the actions to be taken by the operator after groundwater restoration is complete. If insufficient monitoring is conducted, either in duration, frequency, or in the number of wells used to sample the wellfield, it is very possible to reach premature conclusions of stability. In such cases, residual lixiviant or localized areas within the production zone that have not stabilized may cause continued mobilization of uranium and other constituents after monitoring is terminated, potentially leading to contamination downgradient or beyond the boundary of the exempted aquifer. Today’s proposal contains provisions related both to the duration of the monitoring and to the sufficiency of the data necessary to determine that stability has been achieved.
After the ISR operational phase ends, the altered chemical state has to be returned to the preoperational conditions, to the extent possible, so that uranium and other contaminants do not migrate outside the wellfield. Treatments to re-establish chemically reducing conditions (which greatly reduce the uranium concentration in the ore zone groundwater) can restore groundwater constituents to preoperational background levels to a large extent, although experience has shown that restoration of all constituents to the preoperational background level is seldom 100 percent successful.
In addition, the chemically reducing conditions initially present, and the mechanisms that maintained these conditions originally, may not be restored sufficiently to persist over the long-term. Re-oxidation of treated groundwater-host rock systems in other situations has been observed, and post-restoration monitoring at ISR locations has historically been relatively short, typically six months to periods of no more than a few years. A slow re-oxidation process with the resulting potential for enhanced migration of uranium and other contaminants may not be detected during a relatively short post-restoration monitoring period. Such an event could occur if the oxidizing agents in the lixiviant significantly removed the reducing agents originally present in the ore zone (e.g., organic material and iron sulfide minerals) that were responsible for sequestering the uranium to form the ore deposit in the first place. Over time, naturally oxygenated waters entering the ore zone from up gradient could re-oxidize the uranium removed from solution during the restoration process, mobilizing it once again and transporting it downgradient beyond the wellfield. To determine whether a trend of increased concentrations is occurring, it is necessary to monitor over long periods of time and use statistical techniques to analyze the data. This is particularly important if the trend in increased concentrations is relatively slow and the natural variability in the well samples is relatively high. These difficulties point to the need for longer post-restoration monitoring periods than historically performed. However, as discussed earlier, the choice of appropriate statistical techniques to determine the presence or absence of trends in monitoring data can be complicated by shortcomings in the monitoring database, such as missing measurements, “nondetects,” analytical errors and other causes that are difficult to avoid in practice for long timeframe monitoring efforts.
We have considered several options for the length of the long-term stability monitoring period as described below.
1. Thirty-Year Long-Term Stability Monitoring Period, With Provisions for Shortening That Time Period
The initial part of our proposal for long-term stability monitoring addresses the duration of monitoring. Specifically, we are proposing that a facility must demonstrate three consecutive years of stability monitoring and then maintain long-term stability monitoring for an additional period of 30 years; this timeframe can be shortened by demonstrating long-term geochemical stability through modeling, as described below. In determining the appropriate length of long-term stability monitoring to provide confidence that the restored wellfield conditions will remain stable over time, and considering our statutory direction for consistency with RCRA requirements, we find that some direction can indeed be found in the RCRA regulatory framework. For RCRA hazardous waste disposal facilities, a post-closure monitoring period of thirty years is required before permit termination can occur.
Since an engineered RCRA disposal facility for the containment of chemically hazardous waste is similar in concept to relying upon a chemically treated ISR wellfield to contain the potential spread of contaminants, we believe it is reasonable to conclude that a thirty-year long-term stability monitoring period for ISR activities is a consistent application of RCRA requirements. We have examined various statistical techniques for determining the presence or absence of trends in monitoring data, under assumed levels of natural variability and extent of trending, and concluded that, under reasonable values for these variables, a thirty-year monitoring period is a reasonable length of time to detect upward trends in constituent concentrations.
We recognize that a thirty-year monitoring period would be significantly longer than current practice and that stability may be achieved in a shorter timeframe. Therefore, we are also proposing a provision that would allow the regulatory agency to shorten the monitoring period
if the operator can both demonstrate geochemical stability through monitoring and support a conclusion of long-term stability through geochemical modeling. We believe that modeling, which can provide confidence that a geochemical environment exists to prevent uranium and other constituents from re-mobilizing, is an important element of any decision to shorten the monitoring period. Further, we believe this provision will encourage operators to expend more effort in preoperational site characterization, which will improve their modeling efforts.
We are proposing that three consecutive years of stability be demonstrated through monitoring as a prerequisite before the modeling would be considered as justification for reducing the monitoring period. The three-year stability demonstration begins when sufficient monitoring data have been collected to allow a showing of statistical significance at a specified level of confidence. This three-year demonstration period has its roots in the RCRA framework, where it is a metric for the success of corrective action after groundwater contamination has been detected.
We also see this situation as analogous to the restoration of the ISR wellfield. Stability would be demonstrated statistically at the 95 percent confidence level, which we believe will help to ensure that operators collect data of sufficient quantity to support regulatory judgments. Stability would be demonstrated using statistical tests with sufficient power to detect trends with a false negative rate no higher than 5 percent. We believe this will ensure that operators collect data of sufficient quantity and quality with adequate power to support regulatory judgments. As noted in Section II.E.2 of this document, a 95 percent confidence threshold can also be found in the RCRA monitoring program.
2. What Other Options Did EPA Consider For the Long-Term Stability Monitoring Period?
In addition to the option described above, EPA considered two alternatives related to the duration of long-term stability monitoring. We are interested in receiving comments and data on all three options described.
a. Required Thirty-Year Long-Term Stability Monitoring Period
The second option we considered also relies on the RCRA regulatory framework. In this alternative, no provision for shortening the long-term stability monitoring time frame is permitted; thirty years of groundwater monitoring is required. This alternative provides a significant increase in the monitoring period over current industry practice, and the extended time would provide added confidence that the restored wellfield chemistry is remaining stable through this period of time. Thirty years of consistent statistical performance (i.e., no upward trending) would provide strong support for concluding that groundwater systems will remain in a chemically reduced state over time. If upward trending of contaminant concentrations was observed during the monitoring period under this approach, the operator would be required to perform additional corrective action, after which the monitoring period would begin again. We ultimately decided not to pursue this option because it does not sufficiently recognize the site-specific aspects of aquifer restoration or give operators the incentive to reach license termination sooner by conducting geochemical modeling.
b. Narrative Standard With No Fixed Monitoring Period
We also considered the option of a performance-based standard without explicitly calling for a long-term monitoring period. We considered requiring that two conditions be met (i.e., return of the physical hydrologic system to a condition similar to the preoperational flow regime and stability of the geochemical environment) before license termination. To meet the first condition, return of the physical hydrologic system, no significant residual influences from the injection-extraction restoration cycle could remain after restoration. This would include conditions such as hydraulic head and flow direction. Depending on the site, this would likely take many months and perhaps a year or more. To meet the second condition, stability of the geochemical environment, the operator would have to show that the groundwater chemistry is statistically stable at a 95 percent confidence level for a duration of time sufficient to account for site conditions. These site conditions would include such things as variability of constituents in the wellfield, groundwater velocity, constituent travel times and any seasonal influences. We expect it to take at least several years to collect data sufficient to achieve the 95 percent confidence level. With this approach, the regulatory agency would have maximum flexibility in determining whether to establish general requirements or approach each site on an individual basis.
Ultimately, we decided against this approach for several reasons. Statistical analyses alone, without the added requirement of long-term monitoring or the option of geochemical modeling, would provide no assurance that groundwater systems will remain in a chemically reduced state over a longer time frame than that used for data collection. Furthermore, this option does not incorporate RCRA’s thirty-year post-closure period. As previously stated, UMTRCA requires that generally applicable standards promulgated under its authority by EPA for non-radiological hazards be consistent with the standards issued under Subtitle C of RCRA. Based on these two reasons, we feel that this approach has greater potential for premature termination of the license. Furthermore, ambiguity in the narrative nature of such standards has the potential to provoke litigation and make implementation difficult.
3. How will groundwater stability be determined?
The success of a groundwater restoration effort will be measured ultimately not only by whether the restoration goals are achieved, but also by whether those levels can persist and the geochemistry of the groundwater remain stable in the long term. The primary intent of the restoration effort is to return the chemical condition of the groundwater in the production zone to the state that existed prior to the initiation of the ISR operations; restoring the hydrologic regime is also important. The persistence in time (i.e., stability) of the chemical condition developed during restoration is the ultimate measure of success for the aquifer restoration effort. We define stability as the state in which the concentrations of the constituents in the groundwater remain relatively constant over time, with no significant upward trending. The key factor in determining stability, then, is developing a meaningful measure by which to determine whether trending is occurring. Such a measure must address the sufficiency of the data collected, both over time and in its spatial distribution within the wellfield. We discussed the proposed monitoring timeframes in the previous section. The remainder of this section describes how we propose to determine whether groundwater chemistry is stable and where we propose to apply this method.
a. What do we propose for determining stability?
There are some similarities between a hazardous waste land disposal situation and an ISR operation that allow us to draw on the RCRA experience for developing standards. Both the RCRA disposal technology and the post-operation aquifer restoration efforts for an ISR operation are intended to prevent contaminants from migrating in the environment. However, there are some differences that apply to developing ISR standards. An ISR production zone differs from a hazardous waste disposal situation in that the contaminants of potential concern (largely uranium and radium) were present at significant levels entrained within the host rock of the aquifer before ISR operations began and will still be present, to some extent, in the groundwater after the aquifer restoration effort has ended; the process will not completely remove them. The concentrations of contaminants of potential concern are subject to natural temporal variations both before and after ISR operations, and this variability must be taken into consideration in setting standards for judging the adequacy of aquifer restoration. Because of this natural variability, repeated sampling of the post-restoration groundwater must be done to judge the adequacy of the restoration process. To assess when the chemical condition in the wellfield groundwater has become stable, statistical measures and analyses are necessary for examining temporal variations in the water composition data collected over a period of time. Today we are proposing to establish a statistical level of confidence as the standard for determining stability of post-restoration groundwater. We believe this is a relatively simple and straightforward way to represent the level of rigor we believe is necessary to conclude that concentrations of important constituents in the groundwater are not increasing significantly over time.
Determining when groundwater compositions have achieved temporal stability will be a site-specific decision, dependent on the natural variability at the site, which is in turn dependent on many site-specific factors (e.g., spatial variations in uranium mineral distribution within the aquifer, variations in other chemical constituents that affect uranium dissolution), the frequency of sample collection, and the magnitude of any trends in composition that may be present relative to the magnitude of natural variability. Chapter 7 of the technical background information document supporting this rulemaking discusses these aspects of stability monitoring in much greater detail and illustrates the relationships between sampling frequency and data trends with time.
Because of the site-specific interplay between the variables that affect stability, we are not proposing to specify what statistical methods the operator should use to make this determination. There are a variety of methods available that could prove appropriate given the specific conditions at each site. These would include both parametric and non-parametric methods. We recommend that readers consult EPA’s “Statistical Analysis of Groundwater Monitoring Data at RCRA Facilities—Unified Guidance” (2009), which provides exhaustive discussion of methods that have been considered for use in the RCRA program. Further discussion of statistical methods for determining trends in groundwater data may also be found in EPA’s technical background information document, which was prepared to support this proposal.
We emphasize that the choice of statistical method must be based on the quantity and quality of the available data and must be justified accordingly to the implementing regulatory agency.
The intent of the statistical analysis of groundwater monitoring data is to avoid a situation where a wellfield that is unstable is judged to have reached temporal and spatial stability. An appropriate statistical analysis will help to ensure that the regulatory decision reflects a high degree of confidence in the interpretation of the monitoring data. We are proposing that a statistical confidence level of upper 95 percent confidence limit be used to determine stability over time. This level of confidence is often used in regulatory applications, including in the RCRA groundwater monitoring framework.
We believe that an equivalent level of confidence, and its implications for sampling and analysis of groundwater composition data, is appropriate for consistency with RCRA approaches and requirements and the statutory direction applicable to this rulemaking. We believe a confidence level of this rigor will make it necessary for operators to collect an appropriate amount of data that clearly demonstrates that the restored ISR aquifer is geochemically stable and that UMTRCA requirements have been met. The frequency of sampling that will provide meaningful data must be determined from site-specific conditions, such as groundwater flow rates. Another consideration is that stability sampling may be misleading if the operator has not allowed sufficient time for the natural system to recover to the point where the injection-extraction cycle is no longer influencing groundwater flow parameters in the wellfield, particularly in the immediate area around the monitoring wells.
b. Where will the determination of stability be made?
We have noted that a restored ISR wellfield essentially functions as a RCRA hazardous waste management unit. In this sense, when restoration is completed successfully, and the chemistry of the groundwater has been returned to a reducing environment, the uranium and other constituents that were mobilized are essentially “locked in” to the subsurface, as are hazardous constituents that are contained by RCRA engineered units. Following this reasoning, it might be considered appropriate for the outer boundary of the restored ISR wellfield to be designated as the point of compliance with the groundwater standards. However, we are not proposing to take this approach.
Today we are proposing that each well within the wellfield be considered for use as a point of compliance for purposes of determining stability after restoration is determined to be complete (note that today’s proposal does not address the point of compliance for the regulatory agency’s determination that restoration is complete, which may be a more complicated matter). We believe that this is appropriate given the size of some wellfields (on the order of hundreds of acres) and the significant variability that is typically present in the mineralized zone. We believe such an approach will more readily inform both the operator and regulatory agency of localized trending, which may then be remedied as appropriate. If the licensee is able to demonstrate that a particular well is sufficiently representative of groundwater conditions in a larger area, the regulatory agency may approve the use of one well to demonstrate stability in the area covered by a larger number of wells.
F. Institutional Control
Institutional controls are intended to maintain long-term cognizance of the nature and location of particular activities that were done at a specific site, in this case the location of the uranium ore zone exploited by an ISR process. Institutional controls can prevent inadvertent intrusions or adverse consequences for future use of the site. Institutional controls are commonly described as active or passive. Active controls are measures such as guards and fences posted around a site. Passive controls could be the erection of signs or placards at a site.
We are not proposing to establish institutional controls for ISR facilities. Active maintenance of the site will cease with the termination of the license, which will occur when the regulatory agency determines that all license conditions have been met. In this sense, we do not view the long-term stability monitoring period as an institutional control following the ISR restoration phase; rather, we view it as a period of active surveillance to determine the long-term success of the restoration effort.
Nor are we proposing to establish passive controls, either at the site or in documents such as local land records. Requirements for survey plats or other records to be maintained would be consistent with RCRA requirements for hazardous waste facilities; however, these typically apply when waste management units remain at the site and are intended to restrict disturbance of the site.
Though we are not proposing that such records be established for ISR sites, we strongly encourage NRC and Agreement States to include such provisions in ISR licenses since ISR sites will not be restricted from sale or further development. Such provisions could simply inform the subsequent owner of the previous ISR, groundwater restoration activities and aquifer exemption on the property.
G. Other Proposed Amendments
EPA has identified several non-ISR related provisions within 40 CFR part 192 that should be updated and amended. The issues that we propose to address today include:
Amending § 192.32 to address a ruling of the Tenth Circuit Court of Appeals;
Deleting reference to Grand Junction Remedial Action Criteria (10 CFR 712) at § 192.20(b)(3) since the criteria have been removed from the Code of Federal Regulations (CFR); and
Correcting minor typographical and grammatical errors found in §§ 192.31 and 192.32.
1. Judicial Decisions
Section 192.32 has been affected by a ruling from the Tenth Circuit Court of Appeals. Under § 192.32(a)(2)(v), NRC was required to obtain EPA concurrence for approval of ACLs in groundwater restoration. This provision was effectively struck down by the Tenth Circuit Court of Appeals in Environmental Defense Fund v. U.S. Nuclear Regulatory Commission, 866 F.2d 1263, 1268-1269 (10th Cir. 1989), when the Court ruled that NRC has authority under AEA section 84(c) to independently make these site-specific ACL determinations, and that NRC has no duty to obtain this EPA concurrence. Therefore, today we are proposing to revise 40 CFR 192.32(a)(2)(v) by deleting this EPA concurrence requirement.
2. Miscellaneous Updates and Corrections
EPA is proposing an amendment to address an area of part 192 where reference is made to another environmental regulation that has since been removed from the CFR. EPA is also proposing several technical corrections to address known typographical and grammatical errors.
a. Outdated Cross-Reference
Section 192.20(b)(3) refers to criteria that no longer exist in the CFR. Because of this, EPA is proposing to eliminate reference to the Grand Junction Remedial Action Criteria, which once existed at 10 CFR part 712.
In addition, language in § 192.20(b)(3) cites methods that did not prove effective during the Grand Junction Remedial Action Program.
The final report for the program clearly states that filtration (by high efficiency filters or by electrostatic precipitators) and sealants (mainly epoxy-based resins) were not effective over the long term, and were not recommended as remedial options for radon mitigation.
EPA proposes to eliminate the language referencing sealants and filtration.
b. Technical Corrections
Since promulgation of 40 CFR part 192, several typographical and grammatical errors have been identified. Today, EPA is proposing amendments in §§ 192.31(a), 192.31(f) and 192.32(a)(2)(v) to address these technical errors (e.g., spelling mistakes, misplaced comma).