VZJ sign up for etocs
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 8 March 2006
Published in Vadose Zone J 5:480-492 (2006)
DOI: 10.2136/vzj2004.0180
© 2006 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Flint, L. E.
Right arrow Articles by Flint, A. L.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Flint, L. E.
Right arrow Articles by Flint, A. L.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Flint, L. E.
Right arrow Articles by Flint, A. L.
Related Collections
Right arrow Field-Scale Studies
Right arrow Inverse Procedures/Parameter Estimation
Right arrow Fractured Rock

ORIGINAL RESEARCH

Characterization of Unsaturated Zone Hydrogeologic Units using Matrix Properties and Depositional History in a Complex Volcanic Environment

Lorraine E. Flinta,*, David C. Bueschb and Alan L. Flinta

a U.S. Geological Survey, Placer Hall, 6000 J Street, Sacramento, CA 95819-6129
b U.S. Geological Survey, Las Vegas, NV
* Corresponding author (lflint{at}usgs.gov)

Received 18 December 2004.



   ABSTRACT
TOP
 EXECUTIVE SUMMARY
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Characterization of the physical and unsaturated hydrologic properties of subsurface materials is necessary to calculate flow and transport for land use practices and to evaluate subsurface processes such as perched water or lateral diversion of water, which are influenced by features such as faults, fractures, and abrupt changes in lithology. Input for numerical flow models typically includes parameters that describe hydrologic properties and the initial and boundary conditions for all materials in the unsaturated zone, such as bulk density, porosity, and particle density, saturated hydraulic conductivity, moisture-retention characteristics, and field water content. We describe an approach for systematically evaluating the site features that contribute to water flow, using physical and hydraulic data collected at the laboratory scale, to provide a representative set of physical and hydraulic parameters for numerically calculating flow of water through the materials at a site. An example case study from analyses done for the heterogeneous, layered, volcanic rocks at Yucca Mountain is presented, but the general approach for parameterization could be applied at any site where depositional processes follow deterministic patterns. Hydrogeologic units at this site were defined using (i) a database developed from 5320 rock samples collected from the coring of 23 shallow (<100 m) and 10 deep (500–1000 m) boreholes, (ii) lithostratigraphic boundaries and corresponding relations to porosity, (iii) transition zones with pronounced changes in properties over short vertical distances, (iv) characterization of the influence of mineral alteration on hydrologic properties such as permeability and moisture-retention characteristics, and (v) a statistical analysis to evaluate where boundaries should be adjusted to minimize the variance within layers. Model parameters developed in this study, and the relation of flow properties to porosity, can be used to produce detailed and accurate representations of the core-scale hydrologic processes ongoing at Yucca Mountain.


   INTRODUCTION
TOP
 EXECUTIVE SUMMARY
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
CHARACTERIZING the physical and hydrologic properties of subsurface materials is critical to local land-use practices as well as regional-scale hydrologic analyses. Subsurface material properties are needed for artificial and natural recharge measurements and models, three-dimensional geohydrologic framework development and modeling, and estimating or modeling unsaturated zone travel time in recharge analyses (Flint et al., 2000). More detailed analyses are required to evaluate site-specific land-use practices such as landfills, radioactive or chemically contaminated waste burial, and nuclear weapons testing. Predictions of water flow and contaminant transport pathways and flux rates, major issues facing many land managers and environmental researchers, often require the use of numerical models. These models generally require input of parameters describing physical and flow properties of the media, geometry of the modeling domain, and initial and boundary conditions. In addition, subsurface processes such as perched water or lateral diversion of water, which are influenced by features such as faults, fractures, or abrupt changes in lithology, also can be clarified by characterizing the properties of the media.

The approaches for site characterization developed in this study are based on the heterogeneous and complex volcanic tuffs located at Yucca Mountain, Nevada, which is a potential location for a geologic repository for high-level radioactive waste (Fig. 1 ). An extensive and detailed data set of rock properties described in Flint (2003) was used for the development of the approach and supporting rationale. Similar approaches have been successfully applied to less complex alluvial settings where the units are stratified by a deterministic depositional history (Fogg et al., 1998). The Yucca Mountain volcanic site also exemplifies deterministic depositional characteristics that encompass the range of features influencing pore-scale and bulk-scale hydrologic processes. These features include (i) varying pore sizes and structures resulting from the initial deposition of the volcanic material, or from welding, crystallization, and alteration; (ii) different types of minerals deposited in pores from both high- and low-temperature processes; (iii) a suite of fracture types and apertures, and faults; and (iv) lithostratigraphic layers that induce locally saturated conditions. Numerous researchers have measured hydrologic properties of the tuffs at Yucca Mountain (e.g., Rush et al., 1983; Weeks and Wilson, 1984; Klavetter and Peters, 1987; Peters et al., 1984; Flint and Flint, 1990; Loskot and Hammermeister, 1992; Rautman et al., 1995; Flint et al., 1996), but those studies did not provide a comprehensive unsaturated zone parameterization corresponding to detailed hydrogeologic units that can be spatially distributed on a site scale.


Figure 1
View larger version (112K):
[in this window]
[in a new window]
  		Fig. 1. Map of study site with locations of boreholes used to develop matrix properties data set and hydrogeologic units.

 
In this paper we describe an approach for systematically evaluating the features of a site that contribute to water flow to provide a deterministically based set of hydrogeologic units for site-scale model development. Using data collected on the laboratory scale, a representative set of physical and hydraulic parameters are derived for numerically calculating flow of water through the materials at a site or, at a minimum, providing the initial estimates and range of uncertainty for the inversion methods that use additional field data to better constrain parameters (Bodvarsson et al., 2001). The matrix properties relate implicitly to large-scale features such as welding, fractures, and alteration, which are captured in the hydrogeologic unit designations. For any site characterization project, the deterministic processes that control the distribution of features and properties should be considered to determine adequate sample collection intervals to represent those processes. A data set of properties is compiled, and then the data set is simplified to distribute properties throughout the modeling domain. This is done deterministically or stochastically, depending on the processes dominant at the site.

At Yucca Mountain, the volcanic rocks present in a collection of layered strata represent relatively large- and small-scale deterministic processes of deposition, cooling, crystallization, and alteration. Hydrogeologic unit designations at this site historically have been confined to large unit divisions primarily based on a simplified representation of the amount of welding and implicitly whether the rocks are vitric or crystallized (Scott et al., 1983; Montazer and Wilson, 1984). Montazer and Wilson (1984) grouped rocks into five hydrogeologic units (Table 1): Tiva Canyon welded (TCw), Paintbrush nonwelded (PTn), Topopah Spring welded (TSw), Calico Hills nonwelded (CHn) and (upper) Crater Flat variably to nonwelded units (CFu). In this study, these vertical strata were divided into detailed, distinct hydrogeologic units based on these deterministic features, and hydraulic modeling parameters were developed for each unit. Being able to define most of the hydrogeologic boundaries from the lithostratigraphic boundaries was needed to spatially distribute the properties for any modeling effort with the use of the three-dimensional lithostratigraphic framework model produced by the USGS (Buesch et al., 1996b). The framework model then provides the lateral changes in unit type and thickness to which properties are applied. To provide reasonable detail and predictive ability in the magnitude and heterogeneity of the Yucca Mountain site and to allow for detailed process modeling and large-scale three-dimensional modeling, the number of distinct vertical layers was limited to a total of 31. Relation of modeling parameters to porosity also are presented to provide the means by which even more detailed process models can be constructed (e.g., through the scaling of parameters in transitional zones where properties change dramatically over short distances or in clay-altered or vapor-phase corroded zones).


View this table:
[in this window]
[in a new window]
  		Table 1. Generalized lithostratigraphy (Buesch et al., 1996b), corresponding major units (Montazer and Wilson, 1984) and detailed hydrogeologic units at Yucca Mountain, Nevada. (Flint, 2003).

 
The Yucca Mountain site provides several challenges to understanding the distribution and modeling of physical properties and the hydrologic parameters caused by the depositional, cooling, and alteration history of the rocks. This history results in differences in porosity, connectivity and tortuosity of flow paths, water-retention character, vertical heterogeneities, and various scales of features that influence the resulting hydrogeology at the site. These characteristics are discussed in detail in this paper. At Yucca Mountain, nine characteristic rock types result from the combination of different types of deposition, amount of welding and crystallization, and composition of grains, which can be glass (congealed magma during eruption), crystallized at high temperature (feldspar, quartz, and vapor-phase minerals), or diagenetically altered (zeolites and clays). It is the vertical sequence, and possibly the lateral lack of continuity, of these rock types that results in the lithostratigraphic and hydrogeologic complexity of the site. Data presented are in the context of these lithologic rock types. In several locations, the near horizontal boundaries between lithostratigraphic and hydrogeologic units are distinctly abrupt. These contacts are characterized and modeled as distinct boundaries to promote the appropriate modeled hydrologic responses such as lateral diversion due to capillary barrier or permeability barrier effects (Montazer and Wilson, 1984). In other locations, the transitional zones are characterized and modeled appropriately, so, for example, abrupt discontinuities arising from discretization in modeling matrix properties do not result in unrealistic diversion or ponding of water where the actual changes in properties are gradual (Moyer et al., 1996). Some glass has been altered to clays and zeolites in several locations where lithologic transitions have caused high liquid saturations, there was an ancient water table or perched water zone, weathering following deposition has produced alteration, or locally in welded tuffs where moderate-temperature alteration occurred during cooling (Levy, 1984; Broxton et al., 1987; Bish and Aronson, 1993; Buesch et al., 1995; Levy et al., 1996; Buesch and Spengler, 1998). Various amounts of vapor-phase corrosion and mineralization have occurred in numerous lithostratigraphic units (Buesch and Spengler, 1999a, 1999b), and these processes (and products) can change properties with depth.

Consideration of Large-Scale Features
Any field site consists of a complex set of physical and hydrologic characteristics that overlie each other and function at various scales, affecting the retention, storage, and flow of water. At Yucca Mountain, dominant features controlling the distribution and flow of water are the pore-scale rock matrix and the bulk-scale fractures. An even larger scale feature is faults, but this is less typically associated with the lithostratigraphy at the site scale. A common limitation to the characterization of a large, heterogeneous site is the scale at which these features can be measured. Laboratory measurements are useful when describing heterogeneity and distribution of properties, but are limited to samples of relatively small volume that generally can be collected by hand at the surface or from the drilling of boreholes. The direct measurement of matrix properties for this study was performed using 7-cm-diameter cores, whereas the required description of hydrogeologic units must capture much larger scale features. Large-scale features, such as the bulk properties of lithophysal cavities, fractures, and faults, are not explicitly represented. This study does not include characterization of the lateral distributions of rock properties, which are being much more rigorously addressed by other researchers using such approaches as the relation of hydrologic properties to porosity (Rautman, 1995; Flint and Selker, 2003).

The approach of vertically discretizing detailed hydrogeologic units on the basis of characteristics introduced by large-scale processes such as pyroclastic flow deposits, welding, and alteration, allows for the implicit incorporation of features, such as fractures, that influence hydrologic character. The development of hydrogeologic units incorporates the fractures from lithostratigraphic descriptions into the unit criteria, and correlates the frequency of fractures with the properties of the units. At Yucca Mountain, the fracture density is generally correlated to the bulk density and porosity (Fig. 2 ). While not providing a good characterization of the fracture properties, this characteristic allows the porosity to be a correlative property that helps to incorporate the heterogeneity of the fractures at this site into the development of hydrogeologic units, which particularly emphasizes the contrasts in bulk properties at transitional or contrasting layers. Lateral heterogeneity is described through the development of distinct, detailed units, which are then spatially distributed throughout the site on the basis of the three-dimensional lithostratigraphic framework model. This approach to describing the hydrologic properties of the unsaturated zone using matrix properties is reasonable at this site, where water is dominant in the matrix, fractures are generally drained, and inverse modeling of bulk properties to measured water content can be used for site-scale model calibration.


Figure 2
View larger version (16K):
[in this window]
[in a new window]
  		Fig. 2. The relation of fracture density to matrix properties. Data from video logs of the north ramp geologic boreholes (NRG).

 
Site Description and Depositional Features
Location and Geologic Setting
Yucca Mountain is 145 km northwest of Las Vegas, NV (Fig. 1), in a climatic transitional zone between the Mojave and Great Basin Deserts in the Basin and Range physiographic province (Grayson, 1993). It is in the rain shadow of the Sierra Nevada and has an arid climate and an average annual precipitation of approximately 165 mm yr–1 (Hevesi et al., 1992). In the general vicinity of Yucca Mountain, middle Miocene sedimentary deposits are overlain by a thick sequence of rhyolitic lavas and pyroclastic deposits from 14 to 9.5 Ma (Sawyer et al., 1994) that encompass an extensive area of about 13 000 km2, known as the southern Nevada volcanic field (Byers et al., 1976). Many of the volcanic deposits in this field do not have identified source areas, but several came from source areas within a group of calderas known as the Timber Mountain–Oasis Valley caldera complex (Byers et al., 1976; Sawyer et al., 1994). The Tertiary volcanic sequence varies in thickness from 1 to 3 km, with an unsaturated zone approximately 500 to 750 m thick. The formations in the unsaturated zone at Yucca Mountain consist primarily of pyroclastic flow deposits. The lithostratigraphic and hydrogeologic unit symbols (Table 1) are used throughout this paper.

Lithologic Features
Characteristics of the pyroclastic flow deposits result from the sequential development of depositional features, zones of welding, and zones of crystallization, although in some rocks welding can be coincident with deposition, and crystallization can occur synchronously with welding to inhibit development of welding (Riehle et al., 1995; Buesch et al., 1996b). Welding of a simple cooling unit typically includes nonwelded rocks at the top and bottom, with increased welding toward the center of the deposit and inward from the lateral margins (Smith, 1960). Thick deposits can have the complete range of welding, but thin deposits can lack the more welded rocks at the center of the unit (Smith, 1960). The overlap of depositional and zonal features at Yucca Mountain has resulted in a complex, detailed lithostratigraphy with zones of welding and crystallization throughout the site in the different tuffs (Buesch et al., 1996b). Many rocks are vitric and have not undergone crystallization at high temperature, which forms mostly feldspar and quartz (the process is also commonly referred to as devitrification), or alteration at low temperature, which forms clays and zeolites. Alteration of glass to clays and zeolites typically occurs where porosity is relatively high and water is present for long enough periods of time to provide the necessary reactions. The low-temperature alteration potential of any lithologic unit is dependent on (i) rock composition, (ii) rock texture, (iii) proximity to structure, and (iv) water–rock interaction (Broxton et al., 1987). High porosity results in large surface area of glass shards in tuffaceous rocks or in highly fractured glass of lava flows. Therefore, the most likely candidates for alteration are high porosity, vitric rocks associated with faults, perched water or lateral flow horizons, and ancient or present saturated zones. The vitric–zeolitic boundary of the ancient water table at Yucca Mountain (Broxton et al., 1987; Buesch et al., 1995) can be a distinct boundary or a transitional zone where the geometry appears to depend, at least in part, on the hydrologic and lithologic character of the rocks. On the basis of macroscopic textural and mineralogic evidence (Rautman and Engstrom, 1996), quantitative mineralogy (Broxton et al., 1987; Bish and Chipera, 1989; Bish and Aronson, 1993), and data presented herein, a relatively abrupt boundary probably occurs where an ancient water table was confined below the rocks in the vitric, densely welded subzone of the crystal-poor Topopah Spring Tuff (Tptpv3 in Table 1), or as a gradational transition across a vertical span of 30 to 40 m in the nonwelded tuffs in the Calico Hills Formation, as observed from geophysical logs and core samples from Borehole SD7.

Textural and particle-size distribution of fallout tephras and bedded tuffs generally relate to several lithostratigraphic characteristics and hydrogeologic properties (Moyer et al., 1996) that determine the moisture-retention character and permeability of the rocks. In some cases, however, there are localized regions of argillic and zeolitic alteration in the bedded units that contribute to their hydrologic character (Moyer et al., 1996). The nonwelded zones of the pyroclastic flow deposits generally follow the same lithologic constraints, with their hydrologic character being a function of pore-size distribution and subsequent alteration products.

In thick pyroclastic flow deposits, various features and processes influence the hydrologic character of the rock by changing the pathways through which water must pass to flow through the rocks. An increase in welding, which is described on the basis of the amount of deformation of shards and pumice clasts, results in a reduction of porosity, pore size, and permeability. Crystallization can affect hydrologic characteristics by (i) the reordering of atoms in amorphous glass that increases porosity; (ii) the production of "secondary porosity" that results from vapor-phase corrosion of glass particles, increasing the pore volume and permeability; and (iii) vapor-phase mineralization, where the deposition of crystals from the vapor phase occurs in the walls of pore space and across narrow pore throats, reducing the porosity and permeability. Typically, vapor-phase corrosion occurs in the crystal-rich (upper), nonlithophysal zones of the major pyroclastic flow units (Buesch et al., 1996b; Buesch and Spengler, 1999a, 1999b). Vapor-phase corrosion also commonly occurs in the lower part of the columnar subzone and the upper part of the vitric, moderately welded subzone of the crystal-poor Tiva Canyon Tuff, and in the lower crystallized lithofacies of the Prow Pass, Bullfrog, and Tram Tuffs; it is particularly well developed where the vapor-phase corrosion has propagated down into the increasingly porous parts of the moderately welded rocks (Buesch et al., 1996b; Moyer et al., 1996; Buesch and Spengler, 1999a, 1999b). This secondary porosity can result in measurements that are anomalously high compared with those which would be estimated on the basis of the textures of welding (Buesch et al., 1996b; Moyer et al., 1996). Vapor-phase mineralization is typically associated with areas of vapor-phase corrosion and in lithophysal cavities, and may affect the porosity, depending on the amount of crystallization. Properties of lithophysal cavities, while mostly inferred, cannot be directly measured using core-scale analyses. On a larger scale, lithophysae may act to concentrate volumes of water. As water moves through lithophysal zones, the volume of rock through which the water passes is reduced by the occurrence of cavities, but the water doesn't move into the cavities under unsaturated conditions. This results in an increase in the saturation in the matrix in the lithophysal rocks.

There are wide ranges for the particle density and porosity of the rocks at Yucca Mountain (particle density varies from 1.2 to 2.8 g cm–3, and porosity varies from 2 to 60%; however, several systematic and deterministic relations are based on the type of material (vitric, crystalline, vapor-phase minerals, and zeolite and clay minerals) and the associated particle density. The particle density for most data described herein is calculated from the bulk density and porosity, both of which are measured. The bulk density, particle density, and porosity typically follow the expected relation where the porosity equals 1 minus the bulk density divided by the particle density. In vitric rocks, bulk density varies from 1.2 to 2.4 g cm–3, porosity varies from 2 to 60% (v/v), and the particle density is typically 2.35 g cm–3. Variations in the particle density and porosity help identify the processes that formed the rocks for the lithostratigraphic identification. The typical particle density is 2.55 g cm–3 for crystalline rock, 2.30 g cm–3 for vapor-phase minerals, and 2.10 to 2.40 g cm–3 for zeolite and clay minerals. Samples with particle densities that differ from these typical values commonly are a mixture of different types of materials. These relations are invaluable as a context for interpreting how and why hydrogeologic properties change near contacts and for determining the boundaries for hydrogeologic units.

Many rocks contain fractures, some of which result from cooling of the deposit and others are related to the bending or faulting of the rocks. The distribution and characteristics of the fractures can be distinctive for individual lithostratigraphic units (Buesch et al., 1996b) and, therefore, hydrogeologic units. Fractures are prevalent in the crystallized and vitric, moderately to densely welded rocks, and are frequent to infrequent in the vitric and altered, nonwelded rocks. Crystallized, densely welded rocks typically have a variety of fractures, the most common of which are many meters long (5–50 m) that formed during cooling of the deposit; however, microfractures also are occasionally sampled in cores and typically do not contain alteration minerals. These microfractures will probably not contribute to flow under unsaturated conditions. The densely welded, vitric rocks with very little porosity and extremely low permeability have a high density of microfractures that transmit water under high saturations. The dimensions of the microfractures, and whether they contain alteration products, determine whether they contribute to unsaturated flow; very small fracture apertures, or those containing alteration materials, may support unsaturated flow. Because the matrix permeability of these rocks is so low, it is common to find high saturations due to concentration of water flowing from higher porosity rocks to lower porosity rocks, as well as the effects of permeability barriers.

The distribution and properties of fractures have been shown to dominate the bulk of the unsaturated flow regime at Yucca Mountain (Flint et al., 2001b). However, fracture densities and apertures are not well characterized for all lithostratigraphic units, although estimates have been made from borehole core logs. Properties of fractures depend on fracture aperture and whether the fractures are open or filled with calcium carbonate or siliceous materials. Bulk permeability data of fractures are scarce, but calculations can be made of porosity and saturated hydraulic conductivity of fractures with assumed or estimated density and aperture (Kwicklis and Healy, 1993), and measurements of saturated hydraulic conductivity have been made on fractures filled with calcium carbonate.

Hydrologic Implications of Lithologic Stratification
Hydrologic modeling using a detailed vertical distribution of properties displays the importance of the vertical sequence of lithologic features in controlling the distribution of water (Flint et al., 1993; Moyer et al., 1996). Several of these distinctive lithologic features are critical in the characterization of water flow at Yucca Mountain, and the characterization of their physical and hydrologic properties is discussed in detail.

Transition Zones
Transition zones have a pronounced change in matrix properties with depth, typically caused by a change in porosity due to the amount of welding; contacts between vitric, crystallized, or altered rocks; or vapor-phase corrosion and mineralization. Transition zones occur at the top and bottom of the Tiva Canyon and Topopah Spring Tuffs, where the rocks grade from nonwelded to densely welded and locally incorporate changes in porosity due to vapor-phase corrosion and mineralization.

Corresponding to the amount of welding, the fracture density, geometry, and continuity are typically poorly developed in vitric and altered nonwelded tuffs and well developed in the vitric and crystallized, moderately to densely welded rocks. The transition zone from fracture-dominated flow in welded rocks, where saturation is high, to matrix-dominated flow in nonwelded rocks can significantly affect flow. The rapid decrease or increase in porosity (that also corresponds to changes in the degree of fracturing) in these transition zones may lead to the localized concentration or depletion of water.

Capillary and Permeability Barriers
The concept of a natural capillary barrier is summarized by Montazer and Wilson (1984) as a fine-grained layer overlying a coarse-grained layer, where water cannot flow from the smaller pores into the larger pores until the height of water in the overlying layer exceeds a critical height, equivalent to the difference in the capillary rise of the two pore sizes. At Yucca Mountain, the potential for this condition theoretically occurs at several stratigraphic contacts (Flint et al., 2003). The contact of the columnar subzone and the moderately welded subzone of the Tiva Canyon Tuff (Tpcplnc and Tpcpv2, respectively) has fine-grained crystallized rocks with small porosity in contact with vitric, moderately welded rocks. A capillary barrier also could occur at the base of the nonwelded rocks where the coarse-grained rocks overlie the larger aperture fractures in the densely welded rocks of the Topopah Spring Tuff (units Tptrv3, Tptrv2, and Tptrv1 or Tptrn). If these conditions exist in the layered, dipping beds at Yucca Mountain, they could potentially result in the lateral diversion of water (Flint et al., 2003). The reverse condition, when coarse-grained materials overlie fine-grained materials, can also result in lateral diversion due to the lower layer acting as a permeability barrier. This condition occurs under less than saturated conditions in the transition between the coarse-grained PTn and the densely welded, fractured, crystal-rich rocks near the top of the Topopah Spring Tuff (units Tptrv3, Tptrv2, and Tptrv1 or Tptrn).


   MATERIALS AND METHODS
TOP
 EXECUTIVE SUMMARY
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Data collection, analysis procedures, and a methodical approach to identify and characterize discrete hydrogeologic strata are described below. To assess the character of this large, heterogeneous, three-dimensional site, analyses of physical and hydrologic properties for 5320 rock-core samples from 33 boreholes were compiled in a database. This database and all the field sample collection and laboratory methods are described in Flint (2003).

Laboratory Sample Processing
Laboratory measurements include porosity, bulk density, particle density, water content, saturated hydraulic conductivity, and moisture-retention data from which curve-fit param eters were calculated (van Genuchten, 1980). To provide initial conditions and site calibration data, present-day field moisture conditions were included, along with the associated lithostratigraphic units and hydrogeologic unit for all samples. The distribution of boreholes is indicated on Fig. 1. Details of borehole locations and descriptions, and drilling information, can be found in Flint (2003).

Mean water content and saturation have been calculated (Table 2), but several factors should be kept in mind. Assumption of a constant value of water content, saturation, and water potential should not be used across the site for individual units because of the spatial distribution of infiltration that results from the distribution of precipitation, the varying thickness of alluvial cover, the topographic positions of boreholes, and the variable thickness of shallow rock units with different properties (Flint and Flint, 1995; Flint et al., 2001a). Variances also are a function of spatial distribution of alteration features or sampling of microfractures. Large variances occur in the hydrogeologic unit CMW because the argillic alteration is likely to be variable among samples vertically and laterally between boreholes and in TC owing to the random sampling of microfractures.


View this table:
[in this window]
[in a new window]
  		Table 2. Mean values of physical and hydraulic properties for each hydrogeologic unit. A thick line separates adjacent units that differ by more than two standard deviations, and a thin line separates adjacent units that differ by one standard deviation. Adjacent units of moisture-retention curve-fit parameters, {alpha} or n, that differ by more than the 95% confidence limits are separated by a line.{dagger}

 
Geometric means of fracture density data from four surface-based boreholes along the north border of the potential repository were calculated for the units for which there were data (Table 2). This data is for illustration purposes only, representing the general relation of fracture characteristics with rock matrix properties, and constitutes only a small part of the suite of fracture characterization studies that have been done at this site (see Sweetkind et al., 1997).

Determination of Porosity
Alteration can occlude the interconnected pores of the rock matrix as clays, zeolites, opal, and calcite form in place or are deposited in pore channels and throats. The clay or zeolite zone only slightly reduces the measured porosity because water can be stored in clay and zeolite mineral structures rather than occupy space in the pores, which would reduce the measured porosity. This is true unless water is held in smectites that swell with the incorporation of water to as much as 300 times their original size, thus reducing the porosity. Clays and zeolites also influence the pore-size distribution and, thus, the character of the moisture-retention curve and the permeability. Zones with clay or zeolites may be an important factor contributing to perched water. Borehole SD7 has a saturation profile (Fig. 3b ) in the rocks below the Topopah Spring Tuff (Tpt) that strongly indicates by high saturations the presence of alteration minerals and their ability to store water and influence its downward transmission. In zones where there is little or no alteration, such as in the vitric rocks of the Calico Hills Formation and some of the devitrified and vapor-phase corroded rocks of the Prow Pass Tuff, the saturations are extremely low because the pores are larger and drain more easily. In the crystallized and minimally vapor-phase-corroded Prow Pass Tuff, the saturation is high because the rocks are welded with small pores.


Figure 3
View larger version (46K):
[in this window]
[in a new window]
  		Fig. 3. Profiles of total porosity calculated from 105°C oven drying (105 porosity) and effective porosity calculated as total porosity minus porosity calculated from elevated relative humidity drying (relative humidity porosity) for core from two boreholes, (a) N31, and (b) SD7. A difference between 105 and relative humidity porosity of >5% is indicated by black bars (adapted from Flint, 2003).

 
The influence of alteration, whether clays, zeolites, or vapor-phase minerals, on pore structure and water-holding capacity can be examined more closely by varying the relative humidity and temperature of the drying oven during the processing of the samples for the determination of porosity. Bush and Jenkins (1970) determined that drying rock samples in an oven at a temperature of 60°C and 40% relative humidity maintained water in the clay structures, as well as one to two molecular layers of water adsorbed to their surface, while removing it from the pore channels. Studies on soils indicate that residual water content can be defined by the amount of water left in the pores after equilibrating in an environment of approximately –70 MPa water potential, or about 65% relative humidity at 60°C (Rose, 1963; Jackson, 1964). This value subtracted from a measured total porosity can be assumed to be an effective porosity (Flint and Selker, 2003). The method of determining effective porosity from relative humidity drying and comparing it to porosity determined from standard oven drying has been shown to be useful in classifying rocks into zones that have similar properties such as permeability and moisture-retention characteristics for the development of distinct model layers. It also may provide parameters useful in modeling because the measurements of all the physical properties using relative humidity drying better reflect field conditions and could be used to interpret geophysical logs. Effective porosity (Flint and Selker, 2003) was determined from samples dried for 48 h in a relative humidity oven at 60°C and 65% relative humidity, and total porosity was determined from standard oven drying for 24 to 48 h at 105°C.

Characterization of Hydrogeologic Units
The intent of developing hydrogeologic units as discrete layers with associated parameters and properties for flow models was twofold. First, layers could be related spatially to existing three-dimensional lithostratigraphic or porosity models. Second, layers would be distinct enough, with minimal vertical and lateral heterogeneity, that when used in a numerical model would reasonably predict measured field conditions in a vertical borehole at any location on the mountain, while not being so detailed as to be numerically cumbersome or limiting. With these criteria in mind, the following procedure was implemented to identify and characterize discrete hydrogeologic strata:

Phase 1.
A useful first step in developing layers for hydrologic flow models uses lithology and its corresponding relationship with porosity. Porosity has been shown to be a reasonable surrogate for flow properties when relationships are developed according to the lithologic and alteration character of the rocks (Flint and Selker, 2003; Istok et al., 1994). Keeping in mind that all layer divisions have to reflect the vertical distribution of rock units (i.e., units with similar properties, if separated in space vertically, need to be separate layers), the first separation is by lithostratigraphic unit (Fig. 4 ). In zones where vapor-phase corrosion is prevalent, such as in Tpcrn, Tcpuc, Tpcplnc, and Tptrn (Table 1), porosity is used as the determinant for layer boundaries to reduce the variation in properties in a unit. In locations where several lithostratigraphic units have similar hydrologic properties, units are combined, such as the middle and lower parts in the crystallized rocks of the Tiva Canyon Tuff, where the Tpcpll, Tpcpln, and Tpcplnc have low variability in porosity and comprise the hydrogeologic unit CW (Table 1, Fig. 4a). Layers were refined in the transition zones at the top and bottom of the PTn to minimize the range in porosity in a single layer.


Figure 4
View larger version (38K):
[in this window]
[in a new window]
  		Fig. 4. Porosity, saturation, and particle density with depth for borehole SD7 (adapted from Flint, 2003).

 
Phase 2.
To accommodate the influences of mineral alteration, layer properties were evaluated on the basis of the effective porosity profiles, which indicate where water is being held in the pores, and the relationship of Ks to effective porosity, which indicates how the altered zones influence water flow. Unique zones were refined where altered materials influence flow such as at the base of the Tiva Canyon Tuff and in the zeolitized rocks below the Topopah Spring Tuff.

Phase 3.
Differences in moisture-retention curves are chiefly a function of the pore-size distributions of the rock types. These different distributions are primarily due to amount of welding and secondarily to the amount of alteration. Moisture-retention characteristics of layers were evaluated, and predicted water potentials were compared with measured water-potential profiles in selected boreholes. If moisture-retention characteristics differed appreciably in layers, then adjustments of layers were made accordingly.

Phase 4.
Once all the measured data were compiled, along with the estimated Ks from the regression analysis, mean values and standard deviations were calculated for each property for each hydrogeologic unit. Mean values for physical and hydraulic properties for each hydrogeologic unit are listed in Table 2. If significant variances that existed within hydrogeologic units could be reduced by the adjustment of layer boundaries, then the appropriate adjustments were made, particularly where porosity was the factor determining the location of the boundary, such as the upper units and near the base of the Tiva Canyon Tuff.


   RESULTS AND DISCUSSION
TOP
 EXECUTIVE SUMMARY
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
As outlined above, division of the rocks into hydrogeologic units is an iterative process. Hydrogeologic unit divisions were initially based on lithostratigraphic unit boundaries, refined on the basis of changes in porosity and other properties such as moisture-retention characteristics, evaluated as to statistical similarity, and examined in the context of lithostratigraphic features. This section describes the nomenclature (Table 1) and character of the hydrogeologic units and is followed by details and rationale behind the influences of the various features, alteration and permeability, and moisture-retention character.

Porosity, saturation, and particle density generally are good indicators of lithostratigraphic boundaries (Tables 1 and 2; Flint, 2003). This is shown, along with lithostratigraphic units, for samples from a deep borehole, SD9, in Fig. 4. This figure exemplifies the vertically stratified nature of the unsaturated zone at Yucca Mountain. In general, porosity is high and saturation is low in the vitric nonwelded rocks, except for those close to the water table, and porosity is low and saturation is high in the vitric and crystallized, densely welded rocks (Tables 1 and 2; Fig. 4). Particle density is generally high in the devitrified rocks and low in the vitric rocks (Tables 1 and 2). General information regarding the individual units is discussed below. Details on units and specific data for each hydrogeologic unit designation are discussed in Flint (2003).

Phase 1: Porosity Profiles
To capture the influences of lithologic transitions and contrasts in hydrologic properties, we present several examples of the consideration of porosity. Sharp contrasts in porosity across lithologic contacts influence the lateral movement of water. Several of these contacts occur in the lithostratigraphic equivalents of the Paintbrush nonwelded (PTn) hydrogeologic unit (Fig. 3a). The lithostratigraphic units of the PTn commonly are thin, but distinct enough in properties to delineate as separate hydrogeologic units. A numerical modeling exercise by Moyer et al. (1996) assessed the hydrologic impact of these individual units to determine whether the properties were different enough to maintain the individual layers as separate units. It was determined that abrupt and linear contacts, along with the contrasts in properties, were instrumental in creating lateral diversion along the sloping contacts. As an exercise, this indicated that the property contrasts were different enough, in most cases, to maintain separate units. Additional sharp contrasts in porosity occur above (Tptrv1) and below (Tptrv3) the vitric, densely welded rocks of the Topopah Spring Tuff. These units typically have porosity <5% The sharp change in porosity at the top of Tptrv1 is from a sharp increase in welding, and the sharp change at the base is the contact with the crystalline rocks in Tptrn. The sharp change in porosity at the top of Tptpv3 is the contact of the crystalline rocks in Tptpln, and the gradational contact at the base is determined by the character of the Tptpv2 or alteration associated with the vitric–zeolitic boundary. Other large changes in porosity between adjacent units are gradational with depth and have less of an influence on the distribution of water in the profile.

Most of the crystallized, moderately to densely welded Tiva Canyon Tuff and Topopah Spring Tuff appear similar in porosity, but most of these units are divided into hydrogeologic units that closely correspond to lithostratigraphic units, such as lithophysal zones and nonlithophysal zones. Mean values of porosity in Table 2 indicate differences between adjacent units in these formations of at least one standard deviation, which provides a rationale for these divisions partially due to the variation in porosity that results from lithophysal cavities. Primarily, however, the differences are due to pore size, which is reflected in moisture-retention characteristics, and are discussed in a later section.

Large vertical variation in porosity does not always indicate a hydrogeologic boundary, and the mechanism for the distribution of water is not always obvious. For example, consider the profiles in Borehole SD7 (Fig. 4). The locations of high saturation are not well correlated with the several notable changes in porosity. For example, the transition from low to high saturation in the unit described lithostratigraphically as Tac, zeol. (Calico Hills Formation, zeolitized; see Table 1) has a relatively uniform porosity, but the pore sizes become smaller with depth because of a transition in amount of alteration of glass to zeolites.

Phase 2: Alteration, Microfractures, and Permeability
The total porosity profile is illustrated in Fig. 3, as well as the degree of alteration calculated as the total porosity minus the effective porosity. This approach indicates the units that consist of altered rocks, such as Tacbt, Tacbs, Tcp unit 4, and Tcp unit 1. Values of the percentages of measured zeolite (in the form of clinoptilolite; Chipera et al., 1994) are indicated (Fig. 3b). In locations in SD7 where there are no measured zeolites, but where there is a 5% difference between total porosity and effective porosity (indicated by the black bar), there are elevated quantities of the vapor-phase minerals tridymite and cristobalite (Chipera et al., 1994). Cristobalite is present from depths of 502.3 to 569.4 m (2–35%), and tridymite is present from 502.3 to 538.4 m (1–6%). The 5% differential porosity typically occurs in rocks where zeolites were observed. In rocks with no zeolites, it does not occur where only cristobalite was present (538.4–569.4 m), but does occur at depths where there were both tridymite and cristobalite, indicating a possible influence of certain vapor-phase minerals, possibly tridymite or the combination of both, on the structure or geometry of the pore spaces. The increase in tortuosity of the flow channels due to the presence of clay, zeolites, or vapor-phase minerals also reduces the permeability, which is discussed below. (Bish and Chipera [1989] showed that very high percentages of Opal-CT, which is a disordered silica phase containing both cristobalite- and tridymite-like structural units and are often found in the pre-Calico Hills bedded tuff (Tacbt), may influence the hydrologic character of rocks by reducing the permeability even more than the presence of zeolites.)

Hydrogeologic properties similar to those in SD7 that result in the 5% differential porosity are apparent in rocks near the base of the Tiva Canyon Tuff (Tpcpv), as exemplified in N31 (Fig. 3a). This zone of alteration, on the basis of quantitative mineralogy from similar stratigraphic position, contains up to 35% smectite and pervades the site (Bish and Chipera, 1989). The extent of the alteration appears to be related to the topographic location of the borehole; for example, boreholes in narrow up-wash channels receive more frequent runoff and, therefore, maintain higher water contents for longer periods of time (Flint and Flint, 1995). This zone is an important hydrologic feature because it may accentuate the capillary barrier effect caused by small pores above large pores at this transition into the nonwelded tuffs. The alteration appears through the base of the Tpbt3, but the combination of very large porosities and the particular pore-size distributions of the units below the moderately welded and altered rocks, despite the presence of clay, results in the drainage and relatively low saturation of these units.

The vitric–zeolitic boundary in the Calico Hills Formation and Prow Pass Tuff is distinguished not by change in porosity, but by a contrast in pore structure and pore size, which results in corresponding saturation profiles (Fig. 4). The lithostratigraphic units at the vitric–zeolitic boundary vary spatially over the mountain. In the southern part of the site where the boundary is as much as 140 m below the Tptpv3, the upper parts of the Calico Hills Tuff (Tpbt1, Tptpv1, and Tptpv2) are unaltered. This condition exists in three boreholes sampled (SD6, SD7 and SD12). Remaining boreholes are to the north and east, where the vitric–zeolitic boundary is typically defined at the base of Tptpv3, with zeolitization along the fractures in that unit. Properties for the Tpbt1 units are, therefore, divided into altered and unaltered hydrogeologic units distributed on the basis of lateral location. The vitric–zeolitic boundary is based on the degree of hydrologic response rather than using lithologic descriptions that typically include estimates of mineral percentages. In this case, it is defined as the 5% difference in total porosity and effective porosity (see SD7, Fig. 3).

The Prow Pass Tuff, a variable among the five boreholes sampled, represents a compound cooling unit. The alteration of the initially vitric rocks to zeolites pervades the Prow Pass Tuff, with the exception of the devitrified and vapor-phase crystallized Tcp3 where welding is greater. As a result, the Prow Pass Tuff is hydrogeologically divided primarily on the basis of lithostratigraphic unit boundaries and secondarily on the >5% difference between total porosity and effective porosity (Fig. 4b).

The effects of alteration materials on the hydrologic character of the rocks were evaluated by comparing the effective porosity to the measured Ks for 604 samples from all lithostratigraphic units (Fig. 5 ). The figure shows the relation of the log of Ks to effective porosity for samples grouped by alteration, either vitric or crystallized, and those units that have microfractures. Samples from the crystal-rich and crystal-poor vitrophyres of the Topopah Spring Tuff (Tptrv1, TC and Tptpv3, PV3) have very low matrix porosities, but have several samples with high conductivities resulting from microfractures. Three crystallized and welded samples from Tptrn, Tptpul, and Tptpln were included in this group because they contained visible fractures. The altered category includes the lithostratigraphic units not devitrified below the vitric–zeolitic boundary and the altered moderately welded rocks at the base of the Tiva Canyon Tuff; the vitric–crystallized category includes all remaining hydrogeologic units. A more detailed study of what defines the altered rocks is discussed in Flint and Selker (2003). Relations between these simplified, lithostratigraphic, feature-based categories are represented by simple regression models predicting conductivity from relative humidity porosity (Fig. 5).


Figure 5
View larger version (39K):
[in this window]
[in a new window]
  		Fig. 5. Relation of effective (relative humidity) porosity and saturated hydraulic conductivity for samples grouped by vitric/crystallized, altered and microfractured units (adapted from Flint, 2003).

 
Phase 3: Moisture Retention Characterization
An example of the variation apparent in moisture retention characteristics of samples from the Tiva Canyon Tuff is shown in Fig. 6a and 6b. The samples in Fig. 6a are from the welded rocks of this pyroclastic flow unit, while those in Fig. 6b are from the moderately welded and altered base of the unit. The alteration is apparent as high air-entry values at high saturation, representing small pores. The welded tuff units all have low {alpha} values (Table 2), or high air-entry pressures, approximately equivalent to 1/{alpha} (describing the water potential at which the pores initially drain), and, thus, include the largest pores. Rocks in the welded hydrogeologic unit TR are vapor-phase corroded, which increases the size of the pores and results in pore structures that drain at lower water potential than other noncorroded welded rocks, thus, the higher {alpha} (see Table 2). Another notable feature is the residual saturation, or the total porosity minus the effective porosity, which is approximately represented by the saturation at which the dry end of the curves becomes asymptotic. Altered rocks, such as CHZ, or vitrophyres, such as PV3, have the largest residual saturations. The differences between the lithophysal and nonlithophysal units is also apparent in Table 2 in the difference between {alpha} values for lithophysal units TUL and TLL, and nonlithophysal units TMN, TM2, and TM1 because nonlithophysal zones have a greater abundance of small pores and, thus, smaller {alpha} values. In addition, the primary reason for dividing the lower nonlithophysal of the Topopah Spring Tuff, TM, into TM2 and TM1 was the difference in the moisture retention parameter {alpha} (Table 2), which indicates smaller pores.


Figure 6
View larger version (22K):
[in this window]
[in a new window]
  		Fig. 6. Moisture-retention curves for core samples from hydrogeologic units (a) CW and (b) CMW (adapted from Flint, 2003).

 
Phase 4: Statistical Analysis to Produce Mean Values for Modeling Parameters
In general, the variation in porosity in each hydrogeologic unit was relatively small, with the exception of BT4, BT2, and PV2. BT4 is very thin, and few samples were collected. The relatively variable properties in BT2 result from being composed of three units (Tpbt2, Tptrv3, and Tptrv2), where Tbpt2 and Tprv3 vary in the amount of argillization, and Tptrv2 is typically thin (<2 m) and grades sharply into low porosity rocks of Tptrv1 (Moyer et al., 1996). The same mixture of properties is true for PV2 that grades relatively sharply from very low to relatively high porosity (Tptpv3 and Tptpv1, respectively).

Several trends emerge from comparing the mean values of porosity for the hydrogeologic units. In the Tiva Canyon Tuff, the lithophysal CUL is twice as high in mean porosity (16%) as the mostly nonlithophysal CW (8%). The Topopah Spring Tuff (Tpt) does not differ as much, with the upper (TUL) and lower (TLL) lithophysal units having mean porosities of 15 and 13% respectively, whereas the upper and lower nonlithophysal units (TMN, TM2, and TM1) are 11, 11, and 9%, respectively. The vitric bedded tuffs and nonwelded rocks in the PTn, included in CNW, BT4, BT3, and BT2, vary in mean porosity from 39 to 49%. These porosities are significantly higher than similarly deposited vitric tuffaceous rocks in BT1 and CHV that have mean porosities of 27 and 34%, respectively. All units below the Topopah Spring Tuff have moderate to high mean porosity (26–35%) except for BF3 (12%). Frequency distributions of porosity for hydrogeologic units indicate normal distributions for most units (Flint 2003).


   CONCLUSIONS
TOP
 EXECUTIVE SUMMARY
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
A logical sequential approach was presented for the assignment of hydrologic parameters for numerical modeling of flow and transport in a large and complex unsaturated zone. Measured values of rock core porosity, bulk density, particle density, saturated hydraulic conductivity, moisture-retention characteristics, saturation, and lithostratigraphic descriptions were used to divide the unsaturated zone at Yucca Mountain, Nevada, into 31 discrete hydrogeologic units that can be related to, and spatially distributed using, the existing three-dimensional lithostratigraphic framework models. These units, with some exceptions in the Prow Pass and Bullfrog Tuffs, are intended to be hydrologically similar enough to be used successfully to design large- and small-scale numerical models to describe water flow at Yucca Mountain. Mean values and standard deviations of bulk density, porosity, particle density, saturation, and saturated hydraulic conductivity are provided for each hydrogeologic unit, as well as representative values of moisture-retention curve-fit parameters {alpha} and n.

Porosity was the most useful property in characterizing individual hydrogeologic units because it was measured on the most samples and is well correlated to saturated hydraulic conductivity. Several additional features played an important part in the hydrogeologic unit development. Zones of mineral alteration, especially at the base of the Tiva Canyon Tuff and the zeolitized rocks below the Topopah Spring Tuff, defined on the basis of differences between total and effective porosity, have a substantial influence on the storage and transmission of water. The close spatial sampling through many units helped define two types of boundaries between units: transition zones where properties change dramatically but smoothly over short vertical distances, such as at the top of the PTn, and abrupt changes in properties, such as at the bottom of the PTn. These two types of boundaries are of particular interest because of the possibility of capillary or permeability barriers. Zones of variably developed vapor-phase corrosion are characterized individually on the basis of changes in porosity. To illustrate the differences in flow properties of various rock types, the relation of porosity to saturated hydraulic conductivity and {alpha} was analyzed. Hydrologic properties specific to the volcanic rocks at Yucca Mountain were obtained that provide modeling parameters representing all significant hydrogeologic units in the unsaturated zone.

Representing hydrologic properties of a spatially heterogeneous site, whether saturated or unsaturated, based on core-scale sized samples from one-dimensional boreholes sparsely distributed over the study area involves uncertainty. An understanding of the depositional environment provides a basis for approaching the characterization of hydrologic properties. Vertical variability of matrix properties in hydrogeologic units will often follow deterministic patterns on the basis of depositional and (or) cooling processes, depending on the particular environment. Problematic units in this pyroclastic flow deposit and fallout tephra environment are (i) very thin units or units with multiple layers, (ii) units with features larger than core-size samples that might dominate the flow of water and cannot be implicitly captured in the matrix properties, or (iii) units with lateral variability represented by very few samples. The lateral distribution of core-scale properties might adequately be represented for large-scale models by the correlation of properties with surrogates such as porosity that can be modeled using lithostratigraphic distributions and has been shown to be related to flow properties in this report and others. The correlation of porosity with lithology provides a much larger database with which to calculate spatial distributions.


   ACKNOWLEDGMENTS
 
This study was funded by the U.S. Department of Energy, Office of Civilian Radioactive Waste Management, Yucca Mountain Site Characterization Project Office under contract DE-AIO8-92NV10874.


   REFERENCES
TOP
 EXECUTIVE SUMMARY
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow