Lessons from early site investigations at the Waste Isolation Pilot Plant
Published:January 01, 2008
Site-specific investigations of bedded evaporites began at the Waste Isolation Pilot Plant site in New Mexico (USA) in 1976, and the first waste was accepted in 1999. Here, we describe and discuss some lessons learned from personal experience.
“Fatal flaws” may not be fatal. Features, events, or processes are sometimes useful exclusionary factors, especially during site selection. Solution chimneys discovered northwest of the site in 1975 were possible vertical pathways for radionuclide transport. Intensive field studies since then have indicated no solution chimneys at the Waste Isolation Pilot Plant site. Known chimneys are related to a geologic unit not found at the Waste Isolation Pilot Plant, and chimney fill is not very permeable. Normal fluid flow should be downward if subevaporite formations are connected to near-surface units. If a solution chimney had been found early at the pilot plant, there might have been pressure to relocate it.
Timing is important. Potash resources were assessed in 1976 by drilling 21 boreholes; four were completed as shallow hydrology observation wells. Data from all boreholes would have provided a comprehensive picture of the hydrology early in the project history. Resource conflicts were considered more important at the time than hydraulic parameters.
Critics will always be with you. the Waste Isolation Pilot Plant is recertified every five years, offering multiple opportunities for outside review and comment. Repeated comments about dissolution of some halite beds, for example, rely on conclusions reached before site-specific studies. Intensive studies since 1984 of shafts, cores, and geophysical logs have shown that halite is distributed mainly by depositional processes. Some critics remain well behind the curve of technical work; we still must respond.
Brief Background of the Waste Isolation Pilot Plant
Location and Status
The Waste Isolation Pilot Plant (WIPP) has been in operation in southeastern New Mexico since 1999 as a disposal site for radioactive waste from U.S. defense programs. The U.S. Environmental Protection Agency (EPA) certified WIPP in 1999 for transuranic waste, and the New Mexico Environment Department issued a permit to WIPP for mixed waste in 2000. Opening the site for waste came 25 yr after geotechnical and engineering characteristic studies began of the site and surrounding areas. Here, we discuss a few areas of investigation where we believe there are lessons to be learned or lessons have been learned; they come from our experiences and opinions and are not institutional lessons.
Outline of Lessons Learned
Three threads are common to the specific areas or points being discussed. First, the Waste Isolation Pilot Plant had neither an outside regulator nor a specific regulatory objective during its early years. This affected activities for several years, as objectives were internally generated and were mainly deterministic. Second, there will always be conflicts over resources (money, expertise, and time) for a project that cannot be reduced to a competitive environment or “Wal-Martized.” Decisions made about resources have long-term consequences, even if the decisions meet clear short-term objectives. Third, technical issues in the charged atmosphere surrounding a “nuclear” project are not always resolved at the idealized level of scientific discourse of review comment and response. Some erroneous statements or misinterpretations have persisted for years, despite having been resolved or corrected. The public comment and resolution system in place that regulates WIPP and other projects is at times a rather messy affair, but it must be approached with patience.
In the early days of the WIPP (1970s to early 1980s), the project was essentially self-regulated by the Atomic Energy Commission (AEC), Energy Research and Development Administration (ERDA), and the Department of Energy (DOE), successively. (These organizations are designated simply by “DOE” in the rest of this paper.) Program reviewers included the New Mexico Governor's Task Force, review panels overseen by the National Academy of Sciences, and the New Mexico Environmental Evaluation Group (EEG). Project activities were carried out under DOE administration by Sandia National Laboratories, the U.S. Geological Survey (USGS), Bechtel Corporation (architect-engineer), and contractors providing technical assistance to DOE. Geotechnical programs were commonly developed and managed collaboratively during the early years by organizations from Sandia and the USGS in consultation with DOE. General objectives for isolation of radioactive waste at WIPP during these years were developed in part from this collaboration, taking into account background such as National Academy of Sciences and Oak Ridge National Laboratory guidelines and recommendations. Site characteristics expected to be positive were explicitly stated, and the site was judged to meet these expectations (e.g., Weart, 1983). The objective most similar to modern probabilistic assessments of isolation was the notion of estimating the possible uptake of radionuclides over 250,000 yr (∼10 half-lives of 239Pu) at a point on the Pecos River where it was assumed releases would discharge to the accessible biosphere. Some features, events, or processes were examined as possible exclusionary factors or “fatal flaws” as well as necessary (inclusionary) factors for locating WIPP at its current site. One factor is discussed here as an illustration of possible fatal flaws.
The Land Withdrawal Act of 1984 established the EPA as the regulator for radioactive waste disposal at WIPP. The Resource Conservation and Recovery Act established responsibility for hazardous waste disposal with the State of New Mexico. Regulations promulgated by the agencies established the objectives for isolation, and site investigations were altered to obtain information adequate for demonstrating compliance. The regulatory lesson is that it is much more appropriate to have a well-specified regulatory objective that can be evaluated through careful work than to use fatal-flaw factors for processes and events.
For a well-defined manufacturing process or service, competitive bidding can provide a quality product, and necessary resources are reasonably predictable. Establishing a disposal site for radioactive waste is anything but well-defined. The regulatory “risk” (i.e., standard) to be achieved by the site may be well defined now, but the process for establishing and defending that a site achieves such a standard in the United States is still being defined. At any stage, available resources have to be assigned based on priorities as they are understood at the time. It is an imperfect system, and the points we discuss represent hindsight in the allocation of resources at a time when a standard did not exist. An example is drawn from early work at WIPP, where several holes to obtain potash resource information were drilled through a hydrologic unit of interest, the Culebra Dolomite Member of the Permian Rustler Formation, and most were plugged and abandoned without hydrologic testing of the unit or attempting to establish monitoring wells. The decision to plug and abandon the drilholes was made on the basis of available funding at the time and the higher priority to identify potential conflicts between locating the site and developing mineral resources.
The resource allocation lesson is that an early, small investment was likely to have yielded information and data that have still not been entirely duplicated.
From the beginning, WIPP, like similar projects, has had its critics or objectors. We focus on an example of misconceptions and erroneous or outdated information that remains in the mainstream of WIPP criticism through review of the application for EPA recertification of compliance.
The notion of dissolution of halite from formations at WIPP existed before the site was located in southeastern New Mexico. It became part of the official literature, including the geological characterization report (Powers et al., 1978). The possibility of karst in evaporite units at WIPP became a contentious issue early in the 1980s. By the end of 1984, we had personally mapped two large-diameter WIPP shafts in detail and recognized that previous assertions of halite dissolution had rested on assumptions not tested by looking at the rocks. The assumptions were not warranted, at least to the extent that dissolution had been proposed. Even now, some of the early erroneous work is cited as evidence of extensive dissolution of halite at WIPP, despite cycles of comment and response over years.
The review lesson here is that some scientific issues do not get resolved by the normal process of comment and response; the burden is on us to continue to respond to criticism patiently, clearly, and with good science.
General Comments on “Lessons Learned”
There is no doubt that any list of “lessons learned” is going to be rather personal, in that the selections and prominence given to any “lesson” are going to be based on the experiences of the “learners.” Here, we have selected what we think are significant “lessons” in three different areas that cover a variety of experiences, but they are still colored by our own experiences and perceptions. With more than 50 yr of experience with WIPP between us, we have been personally involved in activities that cross these boundaries.
Fatal Flaws Versus Regulatory Objectives
Basic Concept and Background
An initial site in southeastern New Mexico was investigated for the location of the Waste Isolation Pilot Plant in 1974 and 1975 (Fig. 1). The evaporite rocks at depth were significantly deformed, and the site was rejected in 1975 as not meeting necessary criteria for generally horizontal strata for mining (Powers et al., 1978). From late 1975 to early 1976, several areas of southeastern New Mexico were considered that had various characteristics required or to be avoided, before the current WIPP site was settled on as a likely location. Although not defined at the time (late 1975–1976), a “fatal flaw” would simply have been a feature, an event, or a process that, if found (or in some cases not found) at WIPP, would render it unacceptable. Inclusionary criteria required certain factors (e.g., salt depth) or a site could be eliminated. A site exclusionary factor was boreholes through evaporites in the belief that they might create pathways and dissolution that could affect performance. This factor was lessened to 1 mile based on research that indicated the helpful effects of borehole plugging, and the current WIPP site was located to meet this criterion. The factor was based on presumed hydraulic parameters. At a later time, it was determined that hydraulic units above the repository had higher potentiometric heads than those immediately below the evaporite section, reducing concern that the interconnection could cause upward flow.
At that time, there was no specific regulatory objective governing the evaluation of the WIPP site as an acceptable site for disposing of radioactive waste. The mission of WIPP at the time also included the possibility of disposal of “high-level” radioactive waste, which would require more active protection measures during handling and disposal, and planning included a separate disposal level for “high-level” waste. The larger mission certainly had an effect at the time on perceptions of the characteristics that would be acceptable.
In July 1975, ERDA-6 was drilled as part of the exploration/characterization of a potential location for WIPP in southeastern New Mexico. Deeper evaporite beds were deformed, and at 2711 ft (826 m) depth, the drill hole encountered high-pressure brine with H2S. That location was abandoned because of the deformed beds, and a new site in southeastern New Mexico was sought.
In 1975, a potash mine in the region encountered a collapse chimney underlying a surface domal feature (Figs. 2A, 2B, and 3). (These were dubbed “breccia pipes” by participants who had long experience with collapse features in volcanic rocks in southwestern United States.) It was clear from the exposures that the process began below the evaporites (or lower in the evaporite section) and progressed to the surface. Although little more was known about the characteristics of these solution collapse chimneys, there was a sense that such a feature at the site, or the feasibility of the process to generate such a feature, might render the site unacceptable. Breccia pipes and the processes responsible for them were considered a serious, if not fatal, flaw.
Field work directed at detecting and understanding breccia pipes from 1976 to 1982 represented a large part of the characterization studies of the site and surroundings (Powers, 1996). Mainly electrical geophysical techniques were applied to an area of ∼30 square miles (∼78 km2), several holes were drilled to test anomalies, two known breccia pipes located elsewhere were drilled and cored from the surface, a slug test was conducted in one of these holes, the underground exposures of breccia in the potash mine were cored and mapped, and local and regional geology was studied to provide clues regarding timing and process. Snyder and Gard (1982) provided concluding details and a summary of geologic processes; they concluded that these features northwest of WIPP developed over a buried reef (Fig. 1) before ca. 0.5 Ma. Fresher water in the reef had removed salt above over fractures, and the collapse propagated upward when pressure in the reef system was reduced updip by the Pecos River (Bachman, 1980). Shallow dissolution along the upper surface of halite rocks later lowered the surrounding surface, producing a domal structure around the collapse chimney (Figs. 2A, 2B, and 3). The process, therefore, was believed to be associated with a geological feature (the Capitan reef) that does not exist at the WIPP site (Fig. 3). No such feature has been found at WIPP.
The drill-stem test in hole WIPP 31 (Fig. 2A) and the characteristics of the chimney fill are often overlooked as reasons that breccia pipes are not considered to be a significant threat to WIPP. Mercer (in Snyder and Gard, 1982) reported several drill-stem tests in WIPP 31. One test was interpreted to indicate 0.90 mD permeability. Two tests in the lower part of the hole did not yield enough fluid for a permeability calculation. In addition, mapping in the underground, and cores from surface holes, showed halite still present in parts of the collapse chimneys. Fluid is not circulating at any measurable rate through these breccia pipes, and they are not active.
As the hydrologic system at WIPP became better known, it also became apparent that fluid pressures in units above the disposal horizon were greater than pressures in rocks below the evaporites. For a vertical connection such as a breccia pipe, flow should be downward, into units that pose no threat. Spiegler (1982) was one of the first to incorporate this into an estimate of the consequence of a breccia pipe at WIPP.
Intensive searching of the WIPP site and immediate surrounding areas for surface geophysical anomalies that might signify the presence of a breccia pipe led to direct investigations by drill holes (Powers, 1996). None was found. The consequences of finding a breccia pipe at WIPP during early site investigations remain untested. It would have created a large controversy, at best, to have attempted to evaluate the real impact of such a feature in the vicinity of WIPP at the time in the absence of a specific regulatory goal such as now exists.
Allocation of Resources
Basic Concept and Background
No Worthwhile Project Receives or Allocates Funding to Match All of the Perceived (Or Even Real) Needs. Also, Understanding the Difference Between What Is Urgent and What Is Important Often Requires Hindsight. Here, We Use An Early Example of the Allocation of Resources at Wipp to Illustrate That A Different Perception of What Was Important (For the Long Term) and A Modest Change in the Allocation of Funding Could Have Represented A Significant Early Advance in Understanding the Hydraulic Regime of the Wipp Site and Immediate Surroundings.
The first half of 1976 was a time of major events for WIPP: the new site was selected for preliminary characterization, surface geophysical techniques were employed to examine the larger features at depth, and ERDA-9 was drilled very near the center of the site to obtain the first direct evidence of the strata and characteristics that would be used to decide which stratigraphic intervals would be used for disposal (Griswold, 1977).
In 1976, the U.S. federal government changed the beginning of its fiscal year calendar from July 1 to October 1 and designated the transitional quarter year as 1976T. One consequence was that additional funding became available after July 1 that needed to be expended or committed before October 1. Few data were available at the time to evaluate potash resources within the area considered for the WIPP site, and it was decided that a program to drill and core the potash zone could be undertaken with these funds in the time allowed. Twenty-one locations were selected around the WIPP site area (Figs. 4A and 4B), and they were duly drilled, cored, and analyzed (e.g., Jones, 1978). Four of these holes were used for hydraulic testing of the Rustler Formation and were then completed as monitoring wells in the Culebra Dolomite Member of the Rustler Formation (Fig. 4B). The rest of the holes were plugged and abandoned.
Little was known at the time about the hydrology of the units overlying the disposal horizon in the Permian Salado Formation. The Rustler was expected to have three more significant water-bearing units or intervals: the contact between the Rustler and Salado (thought to be a horizon of dissolution of halite from the upper Salado), the Culebra Dolomite Member, and the Magenta Dolomite Member, from bottom to top, respectively. Regional data suggested that none of them would be prolific in the site area. Developing ideas of failure scenarios for WIPP included transport of radionuclides through one or more of these intervals toward Nash Draw and eventually to the Pecos River, the nearest point where water or brine from these units naturally reached the surface (U.S. Department of Energy, 1980).
As data accumulated, it became apparent that the Culebra Dolomite was the more transmissive of the continuous water-bearing units above the Salado. A particular failure scenario, involving a deeper drill hole connecting pressurized brine reservoirs to the Culebra, emerged as the leading means by which radionuclides might escape the disposal site (see Bingham and Barr [1979, 1980] for early analysis of scenarios). The EPA later assumed regulatory responsibility for WIPP radionuclide disposal and developed a probabilistic standard for isolation at WIPP; it is tied to releases at the current WIPP administrative boundary. Through time, the hydrology of the Rustler has been perceived differently (always important), but more detailed data are needed to evaluate performance in light of a specific standard for performance. Here, we examine differences in knowledge that might have been achieved much earlier in the history of the studies of the WIPP site if all of the potash exploratory holes drilled in 1976 had been converted to testing and monitoring wells at the time.
The assessment here is that uncertainty in the spatial distribution of Culebra hydraulic properties and water levels would be less than the current level if all the original potash exploratory holes had been completed and tested in the Culebra. We use a geostatistical approach to illustrate the reduction of uncertainty in Culebra transmissivity.
Forty-two wells in the Culebra have been tested, and they provide a data set of transmissivity (T) that can be further evaluated. It became apparent more recently that there were good correlations between the value of T and three factors: depth, presence or absence of dissolution of upper Salado halite, and spatially interconnected fractures in the Rustler (Holt and Yarbrough, 2002; Powers et al., 2003; Holt et al., 2005). Holt and Yarbrough (2002) developed a predictive linear regression model for Culebra T using these factors; the multiple correlation co-efficient for this regression model was greater than 0.94, and the regression was significant above the 99.9% confidence level.
We used the regression model of Holt and Yarbrough to detrend Culebra T values in the WIPP region, and we fitted an isotropic variogram to the residual T values using the GSLIB subroutine gam2 (Deutsch and Journel, 1998). This variogram is fit with an exponential variogram model where the correlated variance of transmissivity is 0.116, the nugget variance is 0.008, and the correlation length is 1800 m.
Using the GSLIB subroutine kb2d (Deutsch and Journel, 1998), we calculated the kriging variance for two distributions of WIPP wells—the current well distribution (Figs. 4A and 5A) and a distribution including potash (P) holes (treated as completed Culebra wells) (Figs. 5A and 5B). The kriging variance can be used as a surrogate measure of the spatial uncertainty due to data location. It increases to a maximum value in regions beyond the correlation length of Culebra T. Ideally, we would like to have wells spaced close enough that the kriging variance is low in the WIPP 16 square mile region.
Results and Discussion
Figure 5A shows the normalized kriging variance for the current well distribution. Large regions in the western and eastern parts of the WIPP area show high kriging variances. The normalized kriging variance in these areas is greatly reduced when the P holes are included (Fig. 5B).
The potash holes (P holes) appear to be relatively well distributed around the WIPP site (Fig. 4B) in comparison to the distribution of wells that have been tested within the Culebra (Fig. 4A). One reason for this is that some holes drilled for other purposes were later converted to monitoring wells, and their locations were not distributed as widely and uniformly as the P holes.
Many of the hydrologic testing and monitoring wells have been targets of opportunity; they have been completed following a drilling program (and location) dictated by geological or resource concerns. Their distribution would have been greatly enhanced if the potash evaluation wells had all been completed for testing and monitoring in 1976.
Basic Concept and Background
Long-held or often-stated ideas yield slowly to detailed research that contradicts those ideas. For some, it is convenient to ignore or disparage detailed research that contradicts cherished notions. The lesson we take away from this is that one should not expect detailed research to resolve easily those kinds of issues. Here, we explore the evidence for dissolution of halite from the Rustler as an example. We have contributed numerous detailed studies of this topic.
One of the major concerns about radioactive waste disposal in evaporite rocks is the higher solubility compared to many rock types. Many aspects of dissolution have been picked over for WIPP (including “breccia pipes” discussed previously), and some of these are still topics of great interest, especially to critics of WIPP. Comments and assertions continue to be made about the alleged presence of (evaporite) karst at WIPP and attendant claims of the negative effects on probabilistic assessment of performance (e.g., Greenwald and McMullen, 2005; Hill, 2003). Responses have been made many times, and Lorenz (2006a, 2006b), for example, has systematically examined a large part of the argument (as a nonparticipant in the studies) and the evidence available bearing on the notion of karst at WIPP. The study of Rustler dissolution has been interwoven into negative comments regarding karst, as in Greenwald and McMullen (2005). We focus on the more restricted issue of Rustler halite dissolution to illustrate how such issues are (or are not) resolved through technical studies.
Rustler Halite Dissolution
Where has halite been dissolved from the Rustler? This question has a long historical trail, and the answer has been considered important to understanding of the hydrology of the units at WIPP, in particular the Culebra Dolomite.
The Rustler includes several units that consist of relatively pure halite (H) east of WIPP (Holt and Powers, 1988; Powers and Holt, 1990, 2000). Near WIPP, the stratigraphic equivalent of each of these units is mostly siliciclastic mudstone (M), with some gypsum. Halitic portions are thicker than equivalent mud-stones, and one halite unit is about an order of magnitude thicker than its correlative mudstone. These facts are not in dispute, and the approximate limits of halite in the units (e.g., Snyder, 1985) have been known for many years.
Prior to 1984, there was “consensus” that halite had been dissolved from Rustler units to leave a mudstone residue. Although this persistent claim is clear from the literature up to that date, the background that led to it is often obscured in the simple citing of authors (including Powers et al., 1978) who concurred with, or simply repeated, this opinion. The origins of this notion (that the mudstones are residues after dissolution of halite) are probably in the long experience of C.L. Jones of the USGS, expressed first in Jones et al. (1960). There is a thick (and well-established) residue of clay and sulfate at the top of the Salado under Nash Draw and to the west after dissolution of upper Salado halite beds. Jones et al. (1973, p. 20) argued correctly that this residue is stratigraphically part of the Salado, although the predominantly siliciclastic content after dissolution is more like the lower Rustler, especially in geophysical log characteristics. This experience may have contributed to the attribution of the mudstones of the Rustler to dissolution residues without close examination. Somewhat later, we see the inference that Rustler mudstones are dissolution residues clearly stated in various references (e.g., Jones, 1978; Powers et al., 1978) with respect to WIPP and areas to the west. There is little reporting of features within the mud-stone units, although there are occasional references to breccias. For the WIPP area, there is no evidence that any cores or exposures of these mudstones were available for study before 1974. We find virtually no direct evidence from this period of reported core or rock features that support interpreting Rustler mudstones as dissolution residues.
The principal support for the interpretation of mudstones as insoluble residue was the lateral relationship of thin mudstone to thick halite. All units of the Rustler were assumed to be laterally continuous, implying that the halite beds must also have been deposited across the area like dolomite and sulfate beds. Nevertheless, lateral facies changes should be expected, and the Rustler provides a good case for caution.
Why is this issue significant? Through much of the early history of WIPP hydrogeology studies, it was believed that dissolution of halite from the Rustler was the principal cause of the significant (orders of magnitude) east to west increase in Culebra T (e.g., Chaturvedi and Channel, 1985). The notion that little halite had been dissolved from the Rustler left no clear hypothesis to explain lateral variation in Culebra T; that has been rectified (e.g., Holt and Yarbrough, 2002; Holt et al., 2005), with factors as explained in the previous section.
Rustler Sedimentological Studies
We were responsible in 1984 for detailed mapping of WIPP shafts through the Rustler because the waste shaft was being enlarged and the exhaust shaft was being constructed (Holt and Powers, 1984, 1986). As the waste shaft was being deepened, the observable details of bedding, channeling, and sedimentary features in the uppermost mudstone (Fig. 6) of the Rustler were pronounced and not consistent with the notion of significant dissolution of thick halite to produce a residue. This was the beginning of significant modern sedimentological studies of the Rustler in this area, although we acknowledge with respect the contributions of Gard (1968) and Eager (1983).
Our sedimentological and geohydrological studies of the Rustler and documentation from 1984 through 2006 include sixteen major reports and articles, seven basic data reports of core and log findings in newer drill holes, and several analysis reports on geological factors and their hydrological significance. A succinct version (Powers and Holt, 2000) summarized many of the findings: bedding, cross-bedding, channeling, graded clasts, textures indicating synsedimentary dissolution of halite, and pedogenic cutans reveal the depositional nature of the Rustler mudstones. Twenty-two years after the initial report on sedimentary structures in the Rustler and questions about the prevailing doctrine of mud-stones as dissolution residues, formal comments on the ability of WIPP to isolate waste still include canards that have been rebutted before. Here, we show a brief evolution regarding our studies.
Use, Misuse, and Ignorance of Study Data
The initial report (Holt and Powers, 1984) generated considerable debate within the project, and the EEG convened a meeting in March 1985 to hear presentations on dissolution of halite in the Rustler and discuss the results. of the presentations, three were more directed at geological observations about dissolution of halite in the Rustler.
Chaturvedi (1987) summarized some points of his presentation that were considerably expanded and published after the meeting as Chaturvedi and Channel (1985). There is confusion between karst conditions in Nash Draw and halite in the Rustler at the WIPP site. Nevertheless, the sense of the approach clearly is to continue to think of the Rustler mudstones as dissolution residues; a chapter is devoted to “Dissolution of Salt from the Rustler Formation” with subchapters for “The Upper Residue,” “The Middle Residue,” and “The Lower Residue.” Following a quote from Holt and Powers (1984) about the significance of the finding of depositional features, Chaturvedi and Channel (1985, p. 47) concluded that “in as much as this statement is based only on the mapping of one shaft, it requires no further discussion unless the results of detailed sedimentological studies of the rock cores…point to the depositional mode for the absence of salt in Rustler as a more logical explanation.”
We undertook a major sedimentological study of the Rustler (Holt and Powers, 1988). All WIPP Rustler cores were redescribed; select cores were slabbed, analyzed, and examined petrographically; shaft details were compared; and hundreds of geophysical logs from industry wells were compared to logs from cored wells at WIPP. From this, we extended stratigraphic details and drew inferences about basinwide processes. Overlapping this project, the EEG contracted T.K. Lowenstein (1987) to perform a similar, though much more limited, study. Lowenstein inferred that some Rustler halite had been dissolved, but he clearly stated his assumptions, including lateral continuity of Rustler lithofacies. The Lowenstein study has many useful observations, most of which we believe is consistent with the Holt and Powers (1988) study. It is the only other extensive sedimentology-based study of these rocks around WIPP.
The conclusions of the Holt and Powers (1988) study were not adopted immediately by the community of WIPP project investigators. Lappin et al. (1989) began in early 1989 to summarize understanding of the long-term behavior of WIPP, acknowledging differences in interpretation of Rustler halite dissolution (e.g., p. 3–25). Lappin et al. (1989) did not feel the need to accept one or the other approach, basing this decision on the lack of threat to the WIPP repository within 10,000 yr due to dissolution. Lappin et al. (1989) vaguely related Rustler halite dissolution to possible changes in hydraulic properties. “Expected” and “degraded” properties of the Culebra were used in estimates of long-term performance for WIPP; “degraded” properties included reduced porosity and tortuosity and increased apertures of fractures. These are possible consequences of dissolution of evaporites. We appreciate that Lappin was one of the strong supporters of our study of Rustler facies variability.
Regional groundwater modeling in support of the original application to EPA for WIPP certification (e.g., Corbet and Knupp, 1996) did fully adopt the concept that the distribution of halite in Rustler members is largely a product of depositional processes, where possible dissolution at WIPP is mainly limited to the margins of halite in these members (Beauheim and Holt, 1990).
Recent comments to the EPA about karst and complaints about Rustler halite studies that we have conducted (Greenwald and McMullen, 2005) exemplify the failure of outside critics to keep up with these studies. (We do not attempt to respond here to the numerous allegations about karst at WIPP in the comment.) the complaints are based on statements attributed to R.Y. Anderson (http://www.unm.edu/ryand/otherwipp/otherwipp3.html) and allege falsification of information denying dissolution of halite from the Rustler Formation over the WIPP site. It is asserted that “the results of previous geologic investigations, reviewed and published long before the WIPP site was selected, demonstrated that a regional front of dissolution had passed over the area and removed substantial thickness of salt from the Rustler Formation. To head off this evidence WIPP science managers asked in-house geologists to prepare a new report. The new report dismissed the previous evidence, claimed that little or no salt had been removed by dissolution, and offered the opinion that the site had been stable for millions of years.” the quoted material attributed to Anderson by Greenwald and McMullen (2005) only refers to “in-house geologists,” but there are no other studies fitting this description than Holt and Powers (1988), and possible other reports of our studies. Also without specific citation, the study of Lowenstein (1987) is offered as a rebuttal.
Within the Web site attributed to Anderson, there is a photograph purporting to be a fracture in the Rustler Formation that shows evidence of young (perhaps continuing?) dissolution of salt from the formation. The import of using the photograph is to attempt to demonstrate that our evaluation of Rustler halite is not only wrong (and falsified?), but that the analyses of Rustler hydrology and WIPP performance will be incorrect. The photograph is not directly attributed; it is Plate 1 of Chaturvedi and Channel (1985), and the original (Fig. 7) is captioned “An open fracture in the unnamed lower member of the Rustler Formation.”
Some background is in order to unfold this story. The photograph shows a label reading “∼800′ v.s. 9/30/82.” the photograph was taken at ∼800 ft (∼244 m) depth in the shaft then known as the ventilation shaft (v.s.). This shaft was enlarged in 1984 to be used as the waste shaft. The original shaft was 1.8 m (6 ft) in diameter, and it was drilled with circulating fluid. Some sections were covered temporarily with steel plate for safety, but the zone photographed was open to water flow down the walls from the Culebra for more than 6 months between the end of drilling and the date of the photograph. The shaft was enlarged to a minimum of 6 m (20 ft) in diameter through this zone in early 1984, and drilling and blasting were used (Holt and Powers, 1984). In the enlarged waste shaft, the Culebra occurs from 215.3 to 221.9 m (706.5–728.5) ft depth, above the area of the photograph. There was no circulating drilling fluid used in enlarging the shaft, and the zone shown in the photograph was mapped within hours of being exposed. Concrete lining of the shaft had already reached below the Culebra to shut off water inflow. Detailed mapping through the zone where the photograph was taken shows numerous small fractures filled with halite (Fig. 8).
The fracture in the photograph is part of the system of halitefilled fractures in the lower Rustler (Los Medaños Member) revealed by mapping in the absence of water running down the shaft wall. They are not open fractures in the rock, the halite has not been dissolved by predrilling natural causes, and the fracture did not “carry a lot of water,” as alleged in the Anderson Web site. Anderson does correctly infer that something must have occurred “not too long ago.” It did—between the time of drilling and the time of the photograph, Culebra water dripping down the shaft wall dissolved halite from the fracture.
This photograph and its alleged relationship to ongoing (or recent) dissolution is an example of not keeping up, or not wanting to keep up, with the technical work on the issue. The waste shaft mapping report has been available since 1984; there is no excuse that it was unknown to Anderson or to Greenwald and McMullen (2005) because the report (Holt and Powers, 1988) referred to so circumspectly also refers to the shaft mapping work and results and provides a clear citation. The technical information presented in the reports (and many others) is not referred to, and there is no recognition that (again unnamed) geologists working in the WIPP area prior to WIPP exploration had virtually no direct evidence on which to base an opinion, other than general thickness changes. Despite years of detailed studies and numerous detailed reports on the geology of the Rustler, as observed in large-diameter shafts, cores, and geophysical logs, some critics continue to cite assessments made without such data and to conclude that later detailed studies are false because they differ from these early reports.
What Have We Learned?
Exclusion factors (or fatal flaws) can serve useful purposes at an early stage of a project like WIPP, but they need to be carefully considered. An example at WIPP is the bedding, which was required to be nearly horizontal to accommodate normal mining without mining through different lithologies. The mining equipment could deal with one general lithology (salt), and mechanical instabilities would not be more complicated by crossing lithologic boundaries. This is the technical reason the original site was abandoned in 1975 after ERDA-6 was drilled (Powers et al., 1978). Such a factor may not affect the long-term performance of a repository, but it is convenient to have such exclusion (or inclusion) factors when there are areas in which the condition can be avoided.
Features such as breccia pipes were of concern because they indicated possible vertical pathways, and it was unknown whether they might have been in a nascent state under the WIPP site and undetectable from surface features. A great deal of resources went into establishing the processes and distribution of breccia pipes to determine that none is known to exist, or is likely to form, at WIPP. Performance assessment is a complicated, time-consuming, and lengthy process, and one may avoid some unnecessary complications by applying some exclusion factors. Nevertheless, clear regulatory objectives were not yet established at the time of most of these investigations, and it might have been difficult to let the process unfold for a thorough evaluation of the threat of a breccia pipe had one been found at the site at that time. Perhaps the earliest clear expression of consequences associated with a breccia pipe is by Spiegler (1982). He analyzed two scenarios of breccia pipe occurrence below a repository, one with flow to the surface and one with flow downward into formations underlying the evaporites and thence to the Capitan reef. Spiegler implicitly recognized the downward flow of interconnected water-bearing units, but analyzed upward flow as well without finding significant radiological impact relative to maximum permissible concentrations under federal regulations.
Exclusionary or inclusionary factors (“fatal flaws”) are probably best applied for specific features, while processes or events with recurrence rates should be analyzed in terms of performance.
We have shown an example of the possible differences in the understanding of the hydrology as a result of modest reallocation of resources early in the project. The problem, of course, is recognizing which areas of investigation can most benefit at the time from such a reallocation. Resource analysis was a high priority item at the time, and virtually nothing was known of the hydrology of the site, other than there probably was not much water in the various units above the Salado. Until scenario analysis for the site and surroundings proceeded, with an understanding of different consequences for events and processes, resources were allocated on our best judgment at the time of important issues. Work by Bingham and Barr (1979, 1980) provided the beginnings of understanding for scenarios that could be used to direct the allocation of resources. Now, it seems more likely that such analysis could more quickly be applied, even to a new location.
There is a regular mechanism for resolution of issues that derive from the application to EPA for certification or recertification of WIPP and the review by EPA regarding the completeness and technical content of the application. RCRA permitting, by the New Mexico Environment Department, also is governed by procedures for resolving issues brought up by the state. Both have means of accepting and resolving issues brought by other parties as well.
Comments by other parties often follow the same path over and over—comment and response, same comment and additional response. We illustrated an example where comments (and pejorative remarks) seem frozen from long ago, with no acknowledgment of detailed studies regarding the issue. Nevertheless, this is a part of the process established by law, and we continue to respond to similar comments. It is the ultimate responsibility of the regulatory agency, such as EPA, to determine when an issue has been resolved and determine that comments need not be addressed further. We derive great satisfaction from continuing to conduct studies of various interesting aspects of geology for WIPP, even if the studies seem to be ignored.
It may appear that we are criticizing the decisions and decision-makers from another time. Tools and data now available enable us to look back and see some possible different courses of action or wonder at the possibilities. We do recognize that decisions were made in the framework of the time and that the early geologists did not have the benefit of shafts, cores, and logs from holes that did not exist at the beginning of WIPP. It is a testament to the combined efforts of many that WIPP is operating and accepting waste. The most significant lesson that permeates our discussion is the difference that having a regulatory objective or target makes in the approach to site selection and characterization. A similar project, starting in a new location, would be able to take considerable advantage of that early target in evaluating processes and events as well as allocating resources to the studies that contribute to that evaluation.
We thank Leif Eriksson, Art Moss, Norbert Rempe, and William Roggenthen for reviews and absolve them of responsibility where we didn't take their counsel.
Figures & Tables
Deep Geologic Repositories
Deep Geologic Repositories reviews the success stories of underground waste isolation. It focuses on repositories that did, do, and will permanently and safely isolate dangerous materials from the near-surface biosphere. Complementary topics address the isolation capability of average crustal rock, investigations at one representative underground research laboratory, and the geologic preservation of fission products from Precambrian nuclear reactors. An international cast of contributors presents proven practical solutions to a formerly confounding issue in environmental and engineering geology: What do we do with wastes that retain their dangerous characteristics in human terms forever? The principal answer: Recycling into the lithosphere by “reverse” mining.
- chemically precipitated rocks
- cross sections
- Eddy County New Mexico
- government agencies
- hydraulic conductivity
- New Mexico
- public policy
- sedimentary rocks
- site exploration
- statistical analysis
- U. S. Environmental Protection Agency
- underground installations
- underground storage
- United States
- waste disposal
- Waste Isolation Pilot Plant