EARTHQUAKE HAZARDS IN THE INTERMOUNTAIN U.S.: ISSUES RELEVANT TO URANIUM MILL TAILINGS DISPOSAL

PAPER ON SEISMICITY AND TITLE II SITES

Earthquake hazards in the Intermountain U.S.: Issues relevant to uranium mill tailings disposal

Ivan G. Wong and Susan S. Olig
Seismic Hazards Branch, Woodward-Clyde Federal Services, Oakland, CA USA

Bruce W. Hassinger
Smith Environmental Technologies Corporation., Englewood, CO USA

Richard E. Blubaugh
Environmental and Government Affairs, Atlas Corporation, Denver, CO USA

ABSTRACT:

In the past two decades, a tremendous amount of new information and data has emerged on seismic sources in the Intermountain United States and their associated processes of earthquake generation. Consequently, the seismic safety of U.S. uranium mill tailings sites, which are located almost exclusively in this region, are being reviewed by the U.S. Nuclear Regulatory Commission (NRC). Based on a deterministic and probabilistic re-evaluation of potential seismic hazards at a Title II site in southeastern Utah, three significant issues have been raised which will impact other sites in the Intermountain U.S. required to revisit their seismic design criteria by the NRC. These issues are: (1) whether the NRC's required use of a deterministic approach for assessing seismic hazards is appropriate for Title II uranium mill tailings sites in a region such as the Intermountain U.S.; (2) is the alternative approach of probabilistic seismic hazard analysis acceptable to the NRC for uranium mill tailings sites; and (3) what is the appropriate return period that should be used. Based on our evaluation, we conclude that deterministic ground motion approaches such as the NRC's 10 CFR 40 Appendix A can result in overly conservative seismic design criteria for Title II sites in the Intermountain U.S. and that instead, probabilistic seismic hazard analysis should provide the bases for such criteria. Additionally, as in all decisions of this nature, the selection of a return period for a specific site should be based on what is deemed an acceptable level of risk;. Such levels may vary from site to site depending on the consequences of radionuclide release into the environment. However, the values of 200 and 1000 years cited in the Environmental Protection Agency's (EPA) 40 CFR 192.02 and NRC's Appendix A Criterion 6(1) should form the basis for the selected return period.

1 INTRODUCTION

Many portions of the Intermountain region of the western United States (Figure 1) exhibit geologic evidence for large prehistoric earthquakes although they may lack even low levels of historical and/or contemporary seismicity. Such areas are subject to future seismic hazards. Large events such as the 1959 magnitude (M) 7.3 Hebgen Lake, Montana and 1983 M 6.8 Borah Peak, Idaho earthquakes attest to the earth's potential to damage both natural and man-made environments. The recurrence intervals of such large events on a specific fault in the Intermountain U.S., however, may span from a few thousands to more than 100,000 years. Hence, one of the most significant problems

Figure 1. Seismicity of the western U.S. (1808 to 1996) and physiographic provinces and major seismic source zones located in the Intermountain U.S. Also shown is the study area around the Moab site in southeastern Utah. Earthquake data courtesy of the National Earthquake Information Center.

Please click on the image for a larger better quality version of it!

facing the community involved in earthquake hazard mitigation is how to address the hazard from large but infrequent earthquakes. In contrast, there also exist portions of the Intermountain U.S., such as the interior of the Colorado Plateau, where the earthquake potential is low based on both recent geologic and seismologic data.
In 1978, Congress enacted the Uranium Mill Tailings Radiation Control Act (UMTRCA) to provide for the disposal, long-term stabilization, and control of uranium mill tailings. The NRC, which regulates UMTRCA uranium mill tailing sites, has initiated a program of re-evaluating the seismic design criteria of Title II (licensed) sites based on the results of a recent study performed by Lawrence Livermore National Laboratory (LLNL) (Bernreuter et al. 1995). In the LLNL study, "simplified" site specific probabilistic seismic hazard analyses were performed for 19 Title II sites located in Utah, Wyoming, South Dakota, and New Mexico based on readily available information. Bernreuter et al. (1995) concluded that at most sites, their estimates of probabilistic peak ground acceleration at return periods of 2,000 years and more were higher than the values used in design.
In a recent re-evaluation of a Title II site in Moab, Utah, three key seismic hazard issues have emerged in our interactions with the NRC. These issues will significantly impact most, if not all, other sites in the Intermountain U.S. This paper describes these issues and our approach to resolving them.

2 EARTHQUAKE HAZARDS IN THE INTERMOUNTAIN U.S.

The Intermountain U.S., as defined in this paper, consists of the states of Idaho, Nevada, Arizona, Utah, Montana, New Mexico, Colorado, and Wyoming. Physiographically, the region consists principally of the Basin and Range province, Colorado Plateau, Rocky Mountains, and Great Plains. Four major seismic zones are located within or border the Intermountain U.S. including: (1) the Sierra Nevada-Great Basin boundary zone; (2) the Intermountain seismic belt including the Centennial Tectonic Belt; (3) the Central Nevada seismic zone; and (4) the Rio Grande rift (Wong et al. 1982) (Figure 1). Elsewhere, away from these zones, the level of historical seismicity is more subdued but there still exists the potential for the occurrence of large but infrequent earthquakes as indicated by the presence of late-Quaternary faults. For example, the 1887 Sonoran earthquake of estimated M 7.4 occurred as a result of rupture along the Pitaycachi fault just south of the Arizona-Mexico border (Bull and Pearthree 1988) in an area characterized by a low level of historical and contemporary seismicity.
Of greatest relevance to the Intermountain Title II sites are the Intermountain seismic belt and Rio Grande rift. The Intermountain seismic belt is one of the most extensive zones of seismicity within the continental United States (Figure 1). It trends 1300 km northward from- northwestern Arizona through central Utah, straddles the Idaho Wyoming border, and turns northwestward through Montana in the vicinity of Yellowstone National Park (Smith and Sbar 1974; Smith and Arabasz 1991) Much of the Intermountain seismic belt is characterized by generally north- to northwest-trending normal faults. Prominent fault zones include the Sevier and Hurricane faults in northern Arizona and southern Utah, the Wasatch fault zone in central Utah, and the Madison and Hebgen faults near Yellowstone. Since the beginning of the historical record in the mid 1800's, about 25 earthquakes of M 6 or greater have occurred along the Intermountain seismic belt (Smith and Arabasz 1991). The largest event in historical time was the 1959 Hebgen Lake earthquake.
The Rio Grande rift extends for approximately 600 hen from south-central New Mexico northward to south-central Colorado (Figure 1). Most of New Mexico's population is concentrated along the Rio Grande rift in cities such as Albuquerque and Santa Fe. The earliest report of earthquake activity was a sequence of 22 events felt in 1849 to 1850 near the town of Socorro (Sanford et al. 1991). The largest earthquakes observed to date are three events that occurred on 12 and 16 July and 15 November 1906 near Socorro. The estimated size of the latter event, the largest of the trio, is about M 6.

3 SEISMIC HAZARD EVALUATION OF THE MOAB SITE

In response to a request by the NRC, an up-to-date seismic hazards evaluation of the Title II Moab site was performed (Wong et al. 1996). This site, owned by Atlas Corporation, consists of a 130-acre pile consisting of 10 1/2 million tons of processed tailings derived from the past operation of the Atlas uranium mill. The tailings were emplaced over alluvial soils and the disposal area was developed from 1956 to 1984. The site is in the process of final closure and the Remedial Action Plan (Reclamation Plan) requires NRC approval.
According to the Standard Review Plan (SRP June 1993), "there are no NRC regulatory guidelines directly applicable to the geologic and seismologic aspects of the UMTRA Program". However, the basic acceptance criteria pertinent to the geologic and seismic stability aspects are provided in the EPA's 40 CFR Part 192, Subpart A and according to section 192.02, "control of residual radioactive materials and their listed constituents shall be designed to be effective for up to 1000 years, to the extent reasonably achievable, and in any case, for at least 200 years". NRC staff has interpreted this standard to mean that certain geologic and seismic conditions must be met in order to have reasonable assurance that the long-term performance objectives will be achieved (NRC 1994).
The SRP states that NRC staff review of seismotectonic stability must conclude whether the information and investigations in the Remedial Action Plan provide an adequate basis for selection of the Maximum Credible Earthquake (MCE) and determination of the resulting vibratory ground motion at the site. The NRC defines the NICE as the "earthquake which would cause maximum vibratory ground motion based upon an evaluation of earthquake potential considering the regional and local geology and seismology and specific characteristics of local subsurface material" (10 CFR 40 Appendix A). The NRC's Appendix A approach, which basically requires the determination of the 84th percentile MCE ground motions, is a deterministic approach. It requires the use of the worst case earthquake with no consideration for its frequency of occurrence. Although Appendix A stipulates that a tailings pile be designed for the MCE, the Introduction to Appendix A allows for alternatives to be proposed by the licensee. These alternatives "may take into account local or regional conditions, including geology, topography, hydrology, and meteorology. The commission may find that the proposed alternatives meet stabilization and containment of the site concerned, and a level of protection for public health, safety, and the environment from radiological and non-radiological hazards associated with the sites, which is equivalent to, to the extent practicable. or more stringent than the level which would be achieved by the requirements of this Appendix and the standards promulgated by the EPA in 40 CFR Part 192." Furthermore, Appendix A Criterion 6(1) specifies that the regulatory standard is "reasonable assurance" of stability of the tailings disposal for the 200 to 1,000 year period. Moab is located within the interior of the Colorado Plateau which has been generally considered to be seismically inactive and devoid of large earthquakes. Seismological studies performed in the past decade, however, indicate that seismicity is fairly widespread throughout the Plateau interior, albeit at a low to moderate level, and that earthquakes up to M 6 have occurred in historical times (Wong and Humphrey 1989). Although detailed fault studies have not been performed to date within the Colorado Plateau. the available geologic data suggests that only a few significant late-Quaternary faults may exist in the Plateau interior (Hecker 1993). Thus there appears to be at least a low level of earthquake hazard within the Plateau.
In our seismic hazard evaluation of the Moab site, potentially seismogenic faults and seismic source zones (areal sources) significant to the site were identified, characterized, and considered in the analysis. These seismic sources included 11 faults, a zone of microseismicity along the Colorado River southwest of Moab, and a seismic source zone for the Colorado Plateau which represents unknown earthquake sources having no geologic surficial expression (Figure 2). The closest fault to the site is the Moab fault which trends beneath the northeastern corner of the site. Available geologic and geophysical evidence, however, indicates that the fault is not capable of producing significant earthquakes (Olig et al. 1996). In fact, 10 of the 11 faults considered in our evaluation are associated with salt structures and are probably not seismogenic (Wong et al. 1996).
Based on an Appendix A approach, ground motions, as characterized by peak horizontal acceleration, were estimated for three potential earthquake scenarios: (1) a M 5.0 earthquake at a source-to-distance of 30 km, our proposed largest event along the Colorado River seismicity trend; (2) a M 6 1/2 earthquake along this same zone at a distance of 5 km from the site as proposed by the NRC; and (3) a "floating" earthquake of M 6 1/4 at a distance of 15 km. In the absence of any nearby capable faults, the NRC's policy requires that the MCE be represented by a floating (random) earthquake. For the second scenario, the NRC assumed that half of the seismicity zone along the Colorado River could rupture in a single large earthquake. Based on geological and seismological arguments presented in Woodward-Clyde Federal Services (1996), we consider this scenario to be extremely unlikely.
Given a maximum magnitude and source-to-site distance, empirically-based attenuation relationships can be used to estimate median (50th percentile) and median plus one standard deviation (84th percentile) ground motions for a site. The NRC stipulated 84th percentile peak horizontal accelerations at the Moab site were 0.06 g, 0.63 g, and 0.29 g, respectively for the above earthquake scenarios. Based on this analysis, the MCE for the site would be the NRC's M 6 1/2, earthquake occurring along the Colorado River seismicity trend at a source-to-site distance of 5 km.

As an alternative approach, we evaluated the earthquake hazard at the Moab site probabilistically similar to, but in a more rigorous manner than was done by LLNL. In a probabilistic seismic hazard analysis, levels of ground motions associated with a probability or likelihood of being exceeded in a specified time period (or inversely, return period) can be calculated. This approach also allows for the explicit inclusion of the range of possible interpretations and uncertainties in components of the model including seismic source characterization and ground motion estimation. The probabilistic seismic hazard model used in our study is similar to the hazard model originally developed by Comell (1968) and refined by McGuire (1974).
All seismic sources within a distance of about 150 km from the site were characterized and input into the analysis (Wong et al. 1936). This included the 11 faults such as the Moab fault, the Colorado River seismicity trend, and the Colorado Plateau source zone. Ten of the 1 l faults were assigned low probabilities of being seismogenic because they show no evidence for Quaternary activity except deformation related to shallow salt dissolution and flowage (Wong et al. 1996). The attenuation of ground motions was addressed through the use of state-of-the-art empirical relationships for peak horizontal acceleration and stiff soil conditions.

(Source: Wong et al. 1996)

Figure 2. Seismicity (1953 to 1994) and selected Cenozoic faults (after Hecker 1953) in the Moab study area. Stippled areas represent areas of distributed deformation due to salt dissolution. Ball on normal faults is on downthrown side.

Please click on the image for a larger better quality version of it!

The probabilistic seismic hazard analysis resulted in peak horizontal accelerations at the Moab site of 0.05 to 0.18 g for return periods ranging from 500 to 10,000 years (Figure 3). The MCE 84th percentile peak horizontal acceleration of 0.63 g has a return period of about 750,000 years (Figure 3) or 750 times greater than the 1000-year design life stipulated in 40 CFR 192.02 and Appendix A Criterion 6(1). The major contributor to peak: acceleration hazard at 10,000 years is the background earthquake in the Colorado Plateau source zone. The Colorado River seismicity trend and the Moab fault contribute little to the hazard at the Moab site at this return period (Wong et al. 1996).

4 SEISMIC HAZARD ISSUES IN THE INTERMOUNTAIN U.S.

In the seismic hazard evaluation of the Moab site, three significant issues were raised due to NRC regulations governing Title II sites. The first issue stems from the NRC's current

(Source: Wong et al. 1996)

Figure 3. Probabilistic seismic hazard curves for the Moab site. The fractile curves give the range of uncertainty about the mean or median (50th percentile) values. The peak horizontal acceleration of 0.18 g at a 10.000 year return period, our recommended seismic design value, can be read from the mean hazard curve.

Please click on the image for a larger better quality version of it!

position of requiring the seismic design of Title II sites be based on a deterministic Appendix A approach incorporating the concept of the MCE. In such an approach, the 84th percentile ground motions generated by the MCE provide the basis for the Design Basis Earthquake. Intertwined in this issue is also the issue of the reasonableness of the 15 km source-to-site for the floating earthquake in areas of low seismicity.
We believe the MCE peak horizontal acceleration for the Moab site (0.63 g) and even the value estimated for the floating earthquake (0.29 g) are overly conservative for seismic design purposes given the low seismic potential that exists within the interior of the Colorado Plateau. This latter observation is supported by the available seismological and geological data. In particular, the location of the Moab site in the Canyonlands region where many precariously balanced rocks occur throughout the area, some very delicately, suggests that this portion of the Colorado Plateau interior has not been subjected to strong earthquake ground shaking for at least several thousands of years (Wong et al. 1996).
As described earlier, the NRC's policy specifies the 15 km source-to-site distance for the floating earthquake. This distance is rather arbitrary because it is independent of the seismic potential of the region being considered. Thus whether a site is located along the more seismically active Wasatch Front in central Utah or the much less active Moab area, the 15-km distance is fixed. In general, deterministic approaches such as dictated in the NRC's Appendix A can result in overly-conservative seismic design criteria in areas of low earthquake potential. Even for sites in more seismically active areas of the Intermountain U.S., deterministically-based ground motions can also be too high for seismic design because the majority of late-Quaternary faults are characterized by long recurrence intervals far exceeding the lifetimes of engineered structures.
The second issue is whether probabilistic seismic hazard analysis is acceptable to the NRC as an alternative to their Appendix A deterministic approach for developing seismic design criteria at Title II sites. The NRC has endorsed the use of probabilistic risk assessment in nuclear regulatory matters as specified in their final policy statement in the Federal Register (16 August 1995). At this time, however, the NRC has not officially established a policy for Title II sites. Probabilistic analysis has become increasingly used in seismic hazard analysis for a wide range of facilities and structures. It provides the basis for the Uniform Building Code and is now become acceptable for evaluating the potential seismic hazards to nuclear reactors.
Given the uncertainties in seismic source characterization and ground motion estimation in the Intermountain U.S., probabilistic seismic hazard analysis is well suited to addressing these uncertainties. For example, given the observation that the largest known earthquake along the Colorado River is less than M 3, there is considerable uncertainty in the assumption that the maximum earthquake for this zone is M 5 relevant to the Moab site. As previously discussed, the NRC's position that a maximum earthquake of M 6 1/2, could occur within this zone is even more uncertain. Additionally, because the acceptable risk of Title II sites has been defined in terms of time (200 to 1000 years), it is best evaluated through probabilistic analysis which incorporates the recurrence of earthquake sources.
If probabilistic analysis is acceptable for Title II sites, a significant issue is at what return period (or alternatively a probability of nonexceedance) is deemed appropriate by the NRC. It was our recommendation that the seismic design criteria for the Moab site be based on a return period of 10,000 years (corresponds to a 10% chance of exceedance in 1000 years). We selected and recommended this very conservative return period based on the fact that the Moab site is located adjacent to the Colorado River and that radionuclide release into the major water source, if possible, might be considered higher risk than other Title II sites. In the probabilistic seismic hazard analysis performed by Bernreuter et al.(1995) for Title II sites, they calculated peak horizontal accelerations assuming a return period of 10,000 years. They adopted this value because, in their opinion, it satisfied the criteria cited in Appendix A. Furthermore, they stated that such a probability of exceedance may be too conservative for design because of the "relatively low risk posed by the tailings piles." For comparison, the current design life for the proposed underground nuclear waste repository at Yucca Mountain, Nevada is 10,000 years.
Because we considered a 10,000 return period to be very conservative compared to the required 1,000 years cited in 40 CFR 192.02 and Appendix A and because both EPA and NRC considered but explicitly rejected a 10,000 year control period for uranium mill tailings, our recommended seismic design value of 0.18 g for the Moab site provides "reasonable assurance" of a level of protection "equivalent to, to the extent practicable" stipulated in Appendix A. We believe that selection of longer return periods, which correspond to lower probabilities of exceedance, would certainly result in overly conservative seismic design criteria not consistent with the available geologic, seismologic, and geophysical data pertinent to earthquake hazards in the vicinity of the Moab site and the interior of the Colorado Plateau.

5 CONCLUSIONS

Probabilistic seismic hazard analysis has been increasingly accepted as an approach often superior to deterministic methods alone for evaluating seismic hazards for a wide variety of facilities and structures. The probabilistic methodology is particularly well suited in applications for uranium mill tailings sites because of their generally lower risk and locations in the Intermountain U.S. In this region, large damaging earthquakes are possible but relatively infrequent There are also considerable uncertainties in characterizing seismic sources and estimating ground motions which can be explicitly incorporated into probabilistic seismic hazard analysis. Finally, because the level of acceptable risk for Title II sites has been expressed in a time frame of 200 to 1000 years (40 CFR 190.02), probabilistic seismic hazard analysis is better suited to providing the basis for seismic design criteria than deterministic approaches, which are time independent.

6 ACKNOWLEDGMENTS

The seismic hazard evaluation of the Atlas site was performed by Woodward-Clyde Federal Services for Smith Environmental Technologies on behalf of Atlas Corporation. Our thanks to Jacqueline Bott. Andy Gorton. Ann Becker Clark Fenton. Bob Green, Jeff McCleary, Sue Penn, Doug Wright, Richard Ely, Jim Cooksley, Fumiko Goss. and Sadako NcInerney for assistance in this study. We also benefitted greatly from interactions with NRC staff particularly Phil Justus, Bakr Ibrahim, Joe Holonich, Myron Fliegel, Ted Johnson and Dan Rom We greatly appreciate review of this paper by Tony Thompson.

REFERENCES

Bernreuter D, E. McDermott & J. Wagoner 1995. Seismic hazard analysis of Title II Reclamation Plans. Unpublished report prepared by Lawrence Livermore National Laboratory for the U. S. Nuclear Regulatory Commission.

Buil. W.B. and P.A Pearthree 1988. Frequency and size of Quaternary surface ruptures of the Pitaycachi fault, northeastern Sonora Mexico, Bull Seismol. Soc. Am., 78: 956-978.

Cornell, C.A 1968. Engineering seismic risk analysis, Bull Seismol. Soc Am.. 58:1583-1606.

Hecker, S 1993. Quaternary tectonics of Utah with emphasis on earthquake-hazard characterization. Utah Geol. Survey Bull. 127:157 p

McGuire, RK. 1974. Seismic structural response risk analysis incorporating peak response regressions on earthquake magnitude and distance, Massachusetts Institute of Technology Dept. of Civil Engineering Research Report R74-51.

Olig, S.S., C.H. Fenton, J. McCleary & I.G. Wong 1996. The earthquake potential of the Moab fault and its relation to salt tectonics in the Paradox Basin, Utah. In C. Huffman (ed.) 1996 Symposium and Field Conference on the Geology and Resources of the Paradox Basin (in press).

Sanford, A.R., L.H. Jaksha & D.J. Cash 1991. Seismicity of the Rio Grande rift in New Mexico. In D.B. Slemmons, E.R Engdahl, M.D. Zoback & D.D. Blackwell (eds.), Neotectonics of North America, Geol. Soc. Am. Decade Map 1 :229-244

Smith, RB. & W.J. Arabasz 1991. Seismicity of the Intermountain seismic belt. In D.B. Slemmons, E.R. Engdahl, M.D. Zoback, M.L. Zoback & D. Blackwell (eds.), Neotectonics of North America, Geol. Soc. Am. Decade Map 1:185-228.

Smith, RB. & M.L. Sbar 1974. Contemporary tectonics of the interior of the western United States with the emphasis of the Intermountain seismic belt. GeoL Soc. Am. Bull. 85:1205-1918.

U.S. Nuclear Regulatory Commission 1994. Final safety evaluation report to license the construction and operation of a facility to receive, store and dispose of 11e(2) byproduct material near Clive, Utah. NUREG-1486.

Wong, I.G. & J.R. Humphrey 1989. Contemporary seismicity, faulting and the state of stress in the Colorado Plateau. Geol. Soc. Am. Bull 101:1127-1146.

Wong, I.G., S.S. Olig, & J.D. Bott 1996. Earthquake potential and seismic hazards in the Paradox Basin, southeastern Utah. In C. Huffman (ed.) 1996 Symposium and Field Conference on the Geology and Resources of the Paradox Basin (in press).

Wong, I.G., F. Swan m & L.S. Cluff 1982. Seismicity and tectonics of the Basin and Range and Colorado Plateau provinces: Implications to microzonation. In Third International Earthquake Microzonation Conference Proceedings 1:53-69.

Woodward-Clyde Federal Services 1996. Evaluation of potential seismic and salt dissolution hazards at the Atlas uranium mill tailings site, Moab. Utah. Unpublished final report prepared for Smith Environmental Technologies and Atlas Corporation.