Published in Vadose Zone Journal 4:89-100 (2005)
© 2005 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
ORIGINAL RESEARCH
Development of a Wet Plume Following Liquid Release along a Fault
Rohit Salvea,*,
David Hudsonb,
Hui-Hai Liua and
Joseph S. Y. Wanga
a MS 14R0108, Lawrence Berkeley National Laboratory, Berkeley, CA 94707
b MS 427, 1180 Town Center Drive, Las Vegas, NV 89144
* Corresponding author (R_Salve{at}lbl.gov)
Received 8 April 2004.
 |
ABSTRACT
|
|---|
To investigate unsaturated flow through a fault located within fractured welded tuff, we performed in situ field experiments in the Exploratory Studies Facility at Yucca Mountain, Nevada. This experiment involved the release of approximately 82000 L water for a period of 17 mo directly into a near-vertical fault under both constant positive head (at about 0.04 m) and decreasing fluxes. As water was released into the fault, changes in moisture content were monitored in the formation while a large cavity excavated below the test bed was visually inspected for seepage. We observed that water (introduced along the fault) maintained the fault as the primary vertical flowpath, while the adjacent fractured rock served to move water laterally and vertically. Further, unlike primary flowpaths along the fault, flow was not persistent along the secondary flowpaths. While this field experiment provided preliminary insights about the flow field associated with a fault, it also demonstrated the need to investigate the role of infill material and secondary fractures in diverting flow from gravity-driven fast flow toward flowpaths in which lateral flow may occur.
Abbreviations: ERP, electrical resistivity probe • ESF, Exploratory Studies Facility • masl, meters above sea level • UZ, unsaturated zone
|
INTRODUCTION
|
|---|
FAULTS ARE IMPORTANT geological structures for fluid, energy, and solute transport (Lopez and Smith, 1996). Typically, movement associated with the development of faults results in two major hydromechanical units, a fault core and a damage zone (Bruhn et al., 1994; Caine et al., 1996; Seront et al., 1998). The fault core is where the main slip occurs and includes subzones of breccia and cataclasis, as well as fractures of various scales. The damage zone is devoid of breccia and cataclasis and consists primarily of numerous fractures and faults running within the main fault (Gudmundsson, 2001). Faults formed in crystalline rocks have gouge filled core zones and a fracture dominated damage zone, while faults formed in well-lithified porous sandstone include a deformation band–rich damage zone (Smith et al., 1989; Aydin, 1978; Antonellini and Aydin, 1994; Caine et al., 1996, as cited in Sidga and Wilson, 2003). In the case of poorly lithified clastic sediments, the fault zone architecture is substantially different because of the presence of core, damage, and mixed zones that completely lack macroscopic fractures (Sidga and Wilson, 2003).
These structural elements and the effects of mineralization lead to fault zones with differing hydrological properties, ranging from those acting as high-permeability conduits to those acting as low-permeability barriers to fluid flow (Lopez and Smith, 1996). Caine et al. (1996) presented a framework for understanding fluid flow properties for faults formed in crystalline and lithified sedimentary rocks. The model they presented has three architectural elements: (i) the unfaulted protolith (with or without regional structures), (ii) the damage zone (with or without small faults, fractures, and veins), and (iii) the core zone composed of breccia and/or cataclastic or gouge. In Caine et al. (1996) conceptualizations of various combinations and degrees of development within damage and core zones yielded a range of possible permeability structures, including faults acting as localized conduits, distributed conduits, localized barriers, and combined conduit–barriers.
The USDOE is investigating the feasibility of using an underground facility in the unsaturated zone (UZ) at Yucca Mountain, Nevada as the proposed permanent storage site for the disposal of high-level nuclear waste. Yucca Mountain is located in the central portion of the southern Basin and Range, where mountain ranges and intervening basins are the result of late Cenozoic extensional faulting (Piety, 1996). The structural geology of Yucca Mountain is controlled by block-bounding faults spaced 1 to 4 km apart, and within these structural blocks there are intrablock faults that represent local structural adjustments in response to displacements in the block-bounding faults (Stuckless and Dudley, 2002).
A series of studies have been conducted at Yucca Mountain during the last 20 yr to develop a better understanding of flow and transport in the unsaturated fractured rocks. Recent conceptual models developed from these studies suggest that relatively faster flow in faults can lead to wetter conditions than currently observed if future climatic conditions include an increase in precipitation (e.g., Bodvarsson et al., 1999). The competence of faults to act as fast flowpaths in the UZ of Yucca Mountain is supported by the presence of bomb-pulse 36Cl and tritium primarily near faults a vertical distance of approximately 300 m from the surface (e.g, Campbell et al., 2003; Yang et al., 1998). Despite the perceived importance of faults as paths for fast flow, the role of faults in altering the hydrology in the deep UZ at Yucca Mountain is not fully understood (Flint et al., 2001). This is in part because fault zones have received less attention than rock fractures in studies of groundwater flow and also because of the inaccessibility of faults for field tests. This is compounded by the fact that insufficient data, particularly field-based data, are available to adequately characterize and compare architecture, permeability structures, fluid flow, and mechanical properties of fault zones found in different geologic environments (Caine et al., 1996).
In recent years, water-release experiments investigating the role of fast flowpaths in unsaturated rock formations, such as those at Yucca Mountain, have largely focused on fractures (e.g., Lenormand and Zarcone, 1989; Haldeman et al., 1991; Kilbury et al., 1986; Davidson et al., 1998; Faybishenko et al., 2000; Wang et al., 1999; Dahan et al., 1998; Podgorney et al., 2000; Salve et al., 2002; Trautz and Wang, 2002; Glass et al., 2002). Efforts toward identifying the role of faults in influencing flow and transport at Yucca Mountain have largely been limited to the geological and geochemical mapping of these features (Wang and Bodvarsson, 2003). An in situ experiment designed to explore unsaturated flow through faults located at Yucca Mountain was conducted by Salve et al. (2003). Their investigation of flow in the nonwelded tuffs of the Paintbrush group (PTn) suggested that the dry, porous matrix on either side of the fault was capable of attenuating episodic percolation fluxes in localized areas (such as around faults) where fast flow would be expected to dominate. However, once wetted, the matrix neighboring the fault retained moisture for a period of months. As saturation increased in the matrix, less water imbibed along the fault and more water traveled farther along the fault. It was inferred from these tests that a sequence of infiltration events separated by periods of up to a few months could allow water to travel over increasing distances along the fault.
We present the results of a field investigation of unsaturated flow through a 20-m nearly vertical section of fault located in fractured, welded rock of the Topopah Spring Tuff unit (TSw) at Yucca Mountain, Nevada. The objective of this effort was to study the evolution of the flow field as water was continuously released into the fault. Specifically, the goal was to evaluate the relative dominance of the near-vertical fault zone over the surrounding fractured welded rock in transmitting water. To our knowledge, this is the first in situ investigation of flow to include a vertical section of a fault imbedded in unsaturated fractured tuff. We also present techniques developed to conduct the in situ field experiment, observations of the developing wetted plume as water was released along the fault, and a conceptual model for water flow along faults imbedded in fractured, welded tuff.
|
MATERIALS AND METHODS
|
|---|
Site Description
Yucca Mountain, located about 160 km north of Las Vegas, NV (Fig. 1a)
is comprised of alternating layers of welded and nonwelded ash-flow and air-fall tuff (Bodvarsson et al., 1999). In the thick UZ of the mountain, which extends to a depth of 600 to 700 m, moisture flow is controlled by syndepositional processes such as welding, fracturing, and the formation of lithophysal cavities, along with postdepositional activities such as hydrothermal alteration, faulting, and additional fracturing (Bodvarsson et al., 1999). The hydrogeologic units located within the UZ consist of the welded Tiva Canyon Tuff (TCw), nonwelded rocks of the Paintbrush Group (PTn), and the welded Topopah Spring Tuff (TSw).
View larger version (128K):
[in this window]
[in a new window]
|
Fig. 1. (A) Location of Yucca Mountain in Nevada and (B) three-dimensional view of tunnels in the Exploratory Studies Facility at Yucca Mountain. The test bed for this study (Alcove 8, Niche 3) is located at the crossover point of the two tunnels.
|
|
The highly fractured TSw, which will potentially house the repository drifts, lies within a fairly homogeneous matrix (Flint, 1998). It has four units (i.e., the lower portion of the Tptpul, Tptpmn, Tptpll, and Tptpln) with compositionally similar, moderate to densely welded ash-flow tuffs (Hinds and Bodvarsson, 2003), but different fracture characteristics (Buesch et al., 1996). In this unit, which is more than 300 m thick, an underground research laboratory, the Exploratory Studies Facility (ESF), was developed to access and evaluate the potential waste emplacement environment at Yucca Mountain. It includes a 7.8-km-long, 8-m-diameter tunnel, the Main Drift, which was excavated in 1996, and a second 3-km-long, 5-m-diameter tunnel, the Cross Drift. This Cross Drift was excavated as a branch from the Main Drift, such that it crosses over the Main Drift at a vertical distance of approximately 20 m (Fig. 1b).
Test Bed
The test bed for this study is located 3 km west of the North Portal of the ESF. It lies within the TSw and extends from approximately 190 to 210 m below ground surface, with the upper and lower boundaries defined by the Cross Drift and the Main Drift, respectively (Fig. 2a)
. A near-vertical intrablock fault located along the plain at which the two drifts cross is the focus of this study. This fault has a strike/dip of 186/83W with 0.23 m of reverse offset and 0.75 m of left lateral offset.
View larger version (50K):
[in this window]
[in a new window]
|
Fig. 2. Details of the test bed: (A) Location of the test bed between the Cross Drift and the Main Drift in the Exploratory Studies Facility. Shaded plain is located along presumed location of fault. (B) Location of fault zone in Alcove 8 with the four isolated sections for releasing water. (C) Location of boreholes around Niche 3. In the figure, Tptpul refers to the Topopah Spring Tuff Upper Lithophysal Zone and Tptpmn refers to the Topopah Spring Tuff Middle Non-Lithophysal Zone.
|
|
To facilitate the controlled releases of water along the fault, a cavity was excavated into the walls of the Cross Drift. This cavity, referred to as Alcove 8, is approximately 30 m long, 6 m wide, and 4 m high. The fault is exposed around the alcove with about 5.25 m located along the floor. It is open along the sides and ceiling of the alcove, with an aperture of 0.01 to 0.002 m, and appears to be closed along the floor. A second cavity, referred to as Niche 3107, or simply as Niche 3, excavated into the walls of the Main Drift, accessed the fault approximately 20 m vertically below Alcove 8. Niche 3 extends about 10 m from the drift wall and is approximately 4 m wide and 3 m high. The fault is visible along the ceiling of Niche 3, vertically below the trace visible along the floor in Alcove 8.
Alcove 8 is located within the Topopah Spring Tuff upper lithophysal zone (Tptpul), and Niche 3 is located within the Topopah Spring Tuff middle nonlithophysal zone (Tptpmn) (stratigraphic nomenclature of Buesch et al., 1996). The floor of Alcove 8 lies approximately 1094 m above sea level (masl), and Niche 3 has a crown elevation of 1074 masl. The location of the Tptpul–Tptpmn contact is at approximately 1080 masl between Alcove 8 and Niche 3.
Liquid Release along the Fault in Alcove 8
The section of fault visible along the floor of Alcove 8 was chiseled to create a trench to facilitate the ponded release of water along the fault section (Fig. 2b). This trench was divided into four sections, with each section serving as a separate release point. Before the excavation of Sections 1 through 3, a square zone (Section 4) was excavated for liquid releases along 2.05 m of fault. Because of the low infiltration rates observed in this section of fault, we decided to utilize the maximum length of fault available for releasing water. Consequently, a 3.10-m trench (covering Sections 1 through 3, as shown in Fig. 2b) was etched such that the 5.15 m of fault extending along the alcove floor was available for ponding water. Along Sections 1 through 3, the width of the trench ranged between 0.43 and 0.46 m. In Section 4, the trench was excavated to include a square with 2.05 m of fault running through it diagonally.
While the trench was being excavated, a powdery material was observed in the section of fault located along the alcove floor. It was not present in the sections of fault visible along the walls or ceiling of the alcove. This infill was likely fine particles of rock generated during excavation of the alcove that were deposited within the fault and packed tight from the shaking associated with excavation machinery. Because this infill extended about 0.5 to 1.0 m into the fault, no attempt was made to remove it from the infiltration zone.
During a period of 10 mo beginning on 6 Mar. 2001, water was released into each of the four sections of the fault with individual permeameters. The permeameters were designed to maintain the desired height of ponded water (i.e., 
0.04 m), while continuously monitoring the rate at which water was released into the infiltration zone. Under ponded conditions, the wetted area for the first three sections was 0.40, 0.47, and 0.53 m2, respectively. In the fourth section, the wetted area was significantly larger (2.1 m2) because of the length of fault and also because of the extended boundaries of the trench.
On 29 Jan. 2002, a hole approximately 0.02 m in diameter and 0.75 m long, was inadvertently drilled through the infill material in Section 1 during an attempt to install tensiometers along the fault. This resulted in a sudden large increase in the rate at which water infiltrated through Section 1 (from <30 to 1200 L m–1 d–1). Infiltration rates remained high in this section of the infiltration plot for at least the next 2 d while efforts were made to seal the hole with material removed during drilling. Once the hole was sealed, the ponded condition prescribed on the fault was replaced with a gradually decreasing flux, dropping from approximately 170 to 30 L m–1 d–1 in the next 6 mo.
Monitoring of the Wet Plume
Sensors to monitor saturation changes were installed in 10 boreholes located in the immediate vicinity of Niche 3. Six of these boreholes, each 6 m long, originated along the walls of the niche, were oriented to horizontally fan the formation around the ceiling (BH1–6, as shown in Fig. 2c). A seventh borehole (BH7) was inclined at a 30° angle upwards above BH4. Three additional boreholes (BH8–10, as shown in Fig. 2c), each 9 m long, were drilled approximately 1 m above the ceiling to run parallel to the central axis of the niche. All 10 boreholes had an inner diameter of 0.076 m. In BH1–7, BH9, and BH10, changes in saturation were recorded with electrical resistivity probes (ERPs), while psychrometers located along BH8 measured changes in water potentials during the entire investigation. Salve et al. (2000) exploited the inverse relationship between resistance and saturation to develop the ERPs, for which the sensing element is a piece of filter paper across which changes in electrical resistance can be measured.
The ERPs (in BH1–7, BH9, and BH10) and psychrometers (BH8) were housed in trays fabricated from 0.10-m-o.d. PVC pipes cut lengthwise to produce a 0.075-m-wide curved tray. On each tray, ERPs were installed at 0.25-m intervals along the outer surface, such that when the trays were located in boreholes, the probes were in direct contact with the borehole walls. The psychrometers (in BH8) were located at 0.5-m intervals in small diameter holes (3-mm i.d.) drilled through the PVC trays.
In addition to the borehole sensors, seepage rates were continuously monitored with a water collection system connected to the trays along the niche ceiling.
|
RESULTS
|
|---|
Fault Infiltration Rates
Between 6 Mar. 2001 and 28 Feb. 2002, approximately 72000 L of water were applied to the entire trench under ponded conditions. An additional approximately 10000 L of water infiltrated the fault as water was released at gradually decreasing fluxes during the next 6 mo.
During the first 2 mo of water release, a stable water level could not be maintained along the fault due to power interruptions and equipment failure. The resulting disruptions to the daily application rate are apparent in the intake data, which show large fluctuations from early March 2001 through mid May 2001 (Fig. 3)
. During this time, about 15000 L of water were applied to the four sections of the fault. Once the supply of water to the fault was stabilized, the variability in infiltration rates was greatly reduced, ranging from approximately 25 to 100 L m–1 d–1.
View larger version (52K):
[in this window]
[in a new window]
|
Fig. 3. Infiltration rates measured along four sections located in the fault. Since sections were of different lengths the rates have been normalized to fault length of 1 m.
|
|
Within the next 9 mo, as an additional 57000 L of water were released into the fault, there was large spatial and temporal variability in infiltration rates along the fault. In Section 1, infiltration rates, which were about 100 L m–1 d–1 in early May 2001, steadily dropped to about 15 L m–1 d–1 by the end of January 2002. In Section 2, intake rates of about 25 L m–1 d–1 in early May 2001 gradually decreased to about 10 L m–1 d–1 during the next 9 mo. In Section 3, the infiltration rates decreased from about 100 L m–1 d–1 in early May 2001 to about 70 L m–1 d–1 by late January 2002, while the rate in Section 4 fell from approximately 60 to 20 L m–1 d–1 during this time.
The observed infiltration rates reflect the permeability of the infill material that clogged the near-surface sections of the fault rather than an "open fault." The variability measured in the four sections was most likely determined by the depth to which this infill material had penetrated into each of the sections. The magnitude of the reduction of permeability along the fault, caused by the presence of the infill material, was demonstrated when a hole was drilled vertically through approximately 0.75 m of infill material within Trench 1 in early February 2002, and infiltration rates immediately jumped from approximately 30 to 1200 L m–1 d–1 (Fig. 3).
Wetting-Front Movement along the Fault under Ponded Conditions
The three horizontal boreholes located immediately above the ceiling of Niche 3 (i.e., BH8–10) intercept the fault at a distance of 1.27, 1.93, and 2.08 m, respectively, from the collars. The vertical distance from the liquid release zone in Alcove 8 to these boreholes is about 19 m.
Arrival of the wetting front above Niche 3 was first detected on 6 Apr. 2001, 31 d after the start of water release along the fault in Alcove 8, by ERPs located along the fault 1.9 m from the collar in BH9 (Fig. 4)
. The wetting front was first detected in BH10, also at the location of the fault, 3 d later than at BH9, at a similar distance from the collar. The ERPs located close to the fault in BH9 suggested a rapid increase in saturation (as indicated by decreasing resistance values) for a period of 24 h. This initial sharp increase was followed by gradual increases in saturation that continued until late June 2001. The response of the ERPs located close to the fault in BH10 shows that wetting was quick, with saturations reaching a constant value within 2 d.
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 4. Wetting-front arrival above the ceiling of Niche 3 as detected by psychrometers (BH8) and electrical resistance probes (BH9 and BH10).
|
|
The psychrometer located closest to the fault in BH8 (at a distance of 1.4 m from the collar) was the first sensor in that borehole to detect the arrival of the wetting plume. At this location, the data show large fluctuations in water-potential measurements (i.e., between –190 and –250 m) that occurred before mid May 2001, which are indicative of the changes in relative humidity of the ambient tunnel environment. (Given that the fault was about 1 m away from the excavated niche, it is likely that atmospheric perturbations in the drift were transmitted along sections of the fault). Once wetting began at this location, the water potential steadily increased toward zero in the next 6 wk.
Development of the Wetting Plume
After water was first detected along the fault in Niche 3, the wetting front gradually expanded perpendicular to the fault, extending horizontally approximately 6 m in 200 d (Fig. 5)
. In BH1, located almost parallel to the fault, wetting occurred fairly uniformly along 4.0 m of borehole close to the niche cavity, while the remaining 2 m remained dry. In BH2–7, surrounding Niche 3, no measurable changes in saturation occurred during the ponded release of water between March 2001 and late February 2002.
View larger version (44K):
[in this window]
[in a new window]
|
Fig. 5. Resistance changes measured along Boreholes 1, 9, and 10 show progression of the wetting front at 0.25-m intervals during the first 10 mo of liquid release along fault.
|
|
Figure 6
shows how quickly the wetting front of the plume reached various locations surrounding Niche 3. The apparent travel times of water in the test bed were determined by tracking changes in electrical resistivity in the monitoring boreholes located around and above Niche 3. The response from ERPs located along the length of BH9 and BH10 suggests that the fastest velocities (0.65 m d–1) were observed along an approximately 0.50-m-wide zone centered along the fault (i.e., 1.25 m from the borehole collar). The velocity of the wetting front continued to decrease with depth (i.e., further into the niche). At distances between 4 and 6 m from the fault, the wetting-front velocity was relatively constant at approximately 0.1 to 0.3 m d–1. Changes in saturation were not detected at greater distances from the fault.
View larger version (37K):
[in this window]
[in a new window]
|
Fig. 6. Velocity of wetting front as detected by electrical resistivity probes located in boreholes surrounding Niche 3. Note that the fault intercepts BH9 and BH10 at a distance of 1.9 and 2.1 m, respectively, from their collars. The fault does not intercept BH1 and is approximately perpendicular to the borehole.
|
|
In the sections of BH9 and BH10 close to the Main Drift (i.e., at depths from 0 to 1.0 m from the collars), wetting was detected between 75 and 125 d after the start of release of water along the fault. It is possible that this late signal is the result of evaporation effects closer to the drift.
Wetting and Drying within the Plume during Ponded Infiltration
Spatial and temporal variability in resistance measurements at discrete locations along the horizontal plain about 19 m below the point of liquid release show that water flowed toward Niche 3 at a range of velocities, while exhibiting significant temporal variability. Along the immediate fault zone, saturation levels remained consistently high following the arrival of the wetting front for the entire duration of liquid release, even as the application rate of water along the fault in Alcove 8 was gradually reduced. At a short distance from the fault, there was discernable temporal variability that occurred at the scale of days. At increasing distances from the fault, these high-frequency fluctuations appeared to dampen, with saturation changes observed at significantly larger temporal scales. For example, the ERPs located 3.9 and 5.9 m from the collar of BH10 (Fig. 7)
showed saturations gradually building up at these locations in a period of about 5 and 8 mo, respectively, and then gradually decreasing (despite continued release of water along the fault) without the high-frequency temporal variability observed closer to the fault. At a depth of 7.9 m, where there was no clear signal of saturations increasing following the ponding of the fault in Alcove 8, small fluctuations persisted during the entire monitoring period.
View larger version (27K):
[in this window]
[in a new window]
|
Fig. 7. Temporal patterns of saturation changes at 1-m intervals along Borehole 10 during the period when water was released along the fault under constant head (March 2001–January 2002) and then decreasing flux conditions (February–August 2002).
|
|
Seepage into Niche 3
Water began to drip along the fault trace in Niche 3 after 3300 L of water had been released along the fault. As more water was released, measurable seepage was observed in 10 locations close to the fault trace, while further into the niche (i.e., up to a distance of 2–3 m from the fault) the ceiling remained visibly damp without seeping.
The first seeps were observed immediately below the location where the wetting front was first detected in BH10. At this single location, a consistent seepage rate (i.e., 6 L d–1) persisted for the duration that water was released under ponded conditions (Fig. 8)
. In other locations, the seepage rates initially climbed rapidly to about 2 to 8 L d–1 in a period of 2 to 3 wk before dropping sharply to rates below 2 L d–1 by early August 2001. At a few other locations along the fault trace, seeps were observed up to 9 mo after the ponded infiltration began. At these locations, the seeps were sporadic and occurred at rates consistently <3 L d–1.
View larger version (44K):
[in this window]
[in a new window]
|
Fig. 8. Seepage rates measured in 5 of 10 locations where seeps were observed along the fault trace in Niche 3. The plot shows the variability in temporal response at different locations. A few seepage points responded almost immediately to high infiltration rates during a brief period in the first week of February. Legend shows the chronological order in which these seeps appeared.
|
|
The amount of water recovered as seepage varied significantly during the time water was ponded along the fault. Between mid April and mid May 2001 the percentage of injected water recovered sharply increased to about 8%. During the next 5 mo, the recovery percentage fluctuated between 5 and 11%. Seepage into niche was not observed after the constant head release was replaced by a decreasing flux at the upper boundary.
Preferential Flow Following Perturbations to the Upper Fault Boundary
When a hole was created through the infill material in Section 1 on 29 Jan. 2002, the entire length of BH6 detected a wetting front within a period of 25 to 28 h. The velocity of this wetting front (0.7–0.8 m h–1) was significantly higher than the early travel times measured along the ceiling of Niche 3 (Fig. 6). BH6 had not shown any signs of wetting during the previous 11 mo when water was ponded along the fault. Sensors along the borehole showed a sharp drying trend (beginning 4 Feb. 2002) along the entire length of borehole. Other boreholes located around the niche (i.e., BH1–5, BH7) did not detect any changes in saturation following this release into Section 1.
In addition to the response along BH6, at some of the seepage points there was a brief period when there were significant increases in the flux rates, while at other locations the seepage rates remained unaffected (Fig. 8).
|
DISCUSSION
|
|---|
Observations from this field test suggest that the upper boundary condition (i.e., permeability of the infiltration zone) dominated the ensuing hydrologic response along the formation below. During the ponded infiltration experiment, the rate at which water flowed into the fault was controlled by the infill that plugged the fault to a depth of about 1 m along the floor of the Alcove 8. The effectiveness of this infill in reducing the infiltration flux and influencing downstream flow was apparent when a small hole drilled through it resulted in significantly higher infiltration rates (Fig. 2), larger volumes of flow along some of the existing flowpaths (Fig. 8), and the development of new flowpaths (Fig. 9) .
View larger version (50K):
[in this window]
[in a new window]
|
Fig. 9. Changes in saturation along BH6 following perturbations to Trench 1 on 29 Jan. 2002. The shaded section of the plot indicates the duration of the perturbation.
|
|
The infiltration rates measured along four sections of the fault (before perturbations to the infill) suggest that water entered the fault at a steady flux with some spatial variability (Fig. 3). As water moved through the infill toward the open fault below, the permeability contrast likely initiated gravity driven fingers (Nicholl et al., 1992, 1993a, 1993b, 1994; Glass and Nicholl, 1996), which then persisted for the duration of the infiltration experiment. Given that the fault is a near-vertical planar break with <1 m of transitional movement, it is likely that some of the flowpaths initiated at the lower infill boundary conveyed much of the flow observed as discrete seeps in Niche 3.
From the velocity profile of the wetting front, three distinct zones of wetting can be identified (Fig. 9). In the immediate vicinity of the fault is the "primary flow zone" identified by the relatively higher flow velocity measurements in BH9 and BH10 at the approximate location of the fault above the niche ceiling. The velocity measurements indicate that the fast flowpaths associated with the fault were not limited to the 2- to 4-cm "fault gash" visible along the ceiling, but extended for a distance of about 0.5 m on both sides of the fault. The velocity profile also shows a second flow zone that extended approximately 4 m from the fault plane. In this zone, there was an approximately linear decrease in velocity with increasing (perpendicular) distance from the fault. A third flow zone that extended approximately 5 to 7 m from the fault had low velocities that were relatively constant along the zone, despite the increasing distance from the fault.
The relatively small amounts of water recovered as seepage and the length to which secondary flowpaths extended perpendicular to the fault plane suggest that there was a significant horizontal component to flow as water moved vertically through the formation. The absence of saturation changes 3 to 4 mo after the start of infiltration, at distances >6 m from the fault, show this secondary flow regime associated with the fault had a limited width (Fig. 5). Thus, once the flow field was established during the ponded release of water, flow remained within the wetted area until the upper boundary was perturbed.
Despite the evidence of lateral movement of water, it is not immediately apparent where this lateral diversion was initiated within the test bed. As gravity-dominated flow proceeds in a generally downward direction, lateral broadening occurs because of capillary effects, which are exacerbated by dispersion effects from random medium heterogeneities (Pruess, 1999). Contention over where this lateral movement was initiated is provided by two important features of the test bed—the change in lithology from the upper lithophysal zone (Tptpul) to the middle nonlithophysal zone (Tptpmn) (Fig. 2) and infill located within the upper fault boundary. In the absence of these features, we could assume that as water flowed through the fast flowpaths along the fault, some of it was diverted into fractures located adjacent to the fault. However, the limited seepage points observed in Niche 3 suggest that there were few flowpaths along the fault. Therefore, few fractures intercepting the fault could have also intercepted these flowpaths to divert water into the neighboring fractured rock. The extent of wetting observed in BH1, BH9, and BH10 (Fig. 5) indicated that the wetted area was more uniformly distributed and more pervasive than would be expected under such conditions.
There is the possibility that lateral diversion occurred along the Tptpul–Tptpmn contact a few meters above the niche. Conceivably, water entering the fault along the floor of Alcove 8 could move rapidly through the vertical fault plane and the relatively high permeability fractures of the Tptpul, before encountering the lower-permeability fractures of the Tptpmn. This permeability contrast could force a large volume of the infiltrating water to flow along this sloping contact, which dipped toward the back of Niche 3. However, because all water that seeped into the niche was focused along the fault trace and because secondary flowpaths remained confined to a perpendicular distance of <7 m from the fault, it is unlikely that there was significant lateral diversion associated with this permeability contrast along the test bed.
It is more likely that the observed lateral diversion was initiated by infill located along the upper boundary of the fault. As water was introduced into the fault, some likely flowed vertically through the infill into the "open" fault, where a few flowpaths served as conduits for quick flow to the niche below. However, the bulk of water injected that was not recovered as seepage (i.e., 90%) was likely distributed via the infill to adjacent fractures (in immediate contact with the infill) before migrating along connected fractures toward the ceiling of Niche 3 (Fig. 10) .
View larger version (100K):
[in this window]
[in a new window]
|
Fig. 10. (A) Observations of the flow field as water was released into the fault; (B) conceptual model of the flow field.
|
|
Besides lateral flowpaths, some of the water not recovered as seepage was likely held by potential capillary barrier effects above the niche ceiling (Philip et al., 1989; Birkholzer et al., 1999). An indication of this is the timing of the wetting front, which arrived at a fault location about 1 m above the ceiling in 31 d, and the observation of the first seeps in Niche 3 in 35 d after the initial release of water. If a constant wetting-front travel velocity (0.65 m d–1) is used, it should have taken 1.6 d for the wetting front to travel the 1-m distance from the monitoring borehole to the niche ceiling.
The existence of capillary barrier effects (corresponding to underground openings) in unsaturated fractured rock was previously demonstrated by Trautz and Wang (2002). They released water from boreholes about 1 m above the ceiling of a drift and observed the spreading of a wetting front across the drift ceiling and water movement up fractures exposed in the ceiling before seepage began. While both this study and the field observations of Trautz and Wang (2002) support the existence of capillary barriers in unsaturated fractured rock, the detailed mechanisms of the capillary effects are not entirely clear, nor is it possible to quantify the amount of water that could have been diverted around the cavity during the course of the experiment.
|
CONCLUSIONS
|
|---|
We made following observations and inferences:- The fault along which water was introduced served as the primary vertical flowpath. However, while seeps were observed along an extended section of fault, each seepage point occupied a small area. Further, the amount of water recovered as seepage (<10%) was relatively small.
- In the adjacent fractured formation where the bulk of injected water migrated, water moved laterally and vertically. However, unlike the primary flowpaths along the fault, which remained active for the duration of the ponded release of water, flow was not persistent along the secondary flowpaths.
- There were significant transient effects, such as decreasing infiltration rates, fluctuating saturations, and intermittently flowing paths, even under a steady (ponded) upper-boundary condition.
- The infill plugging the fault at the infiltration zone significantly influenced the hydrologic response by limiting the amount of water that entered the fault as fast flow and by increasing the amount of water that was introduced into the adjacent fractures in immediate contact with the infill.
- Within the context of waste isolation, the capacity of a potential fast flowpath can be significantly dampened by infill deposited during construction. This may reduce the amount of water and dissolved solutes impacting an isolated waste package and also disperse the infiltrating water over a larger area.
While this field experiment showed some of the fundamental features of the flow field associated with a fault, it has also demonstrated the complexity of flow in this unsaturated environment. In particular, it has demonstrated that flow in the immediate vicinity of a fault embedded in unsaturated fractured rocks is consistently unsteady, as seen in the spatial and temporal variability of infiltration and seepage rates and in the wetting patterns of the fractures (despite a long-term steady supply of water).
This persistence in unstable behavior has been demonstrated before in laboratory experiments (e.g., Glass et al., 2002) and field experiments (e.g., Dahan et al., 1998; Faybishenko et al., 2000; Podgorney et al., 2000) in unsaturated fractured rock environments. The effects of this temporally unstable behavior on overall flow and transport processes at different scales are not fully understood at this time. Obviously, more experimental and theoretical studies are needed to resolve this issue.
|
ACKNOWLEDGMENTS
|
|---|
We thank Diana Swantek (LBNL) for her contributions in the preparation of graphics for this paper. Thanks are also due to Atlantis Czarnomski, Phil Rizzo, Alex Morales, Paul Cook, and John (Tom) Kostalek for assistance in developing various aspects of the field experiments. Reviews of the manuscript by Chin-Fu Tsang, Raymond Torres, Pradeep Talwani, Dan Hawkes, and an anonymous reviewer are gratefully acknowledged. This work was supported by the Director, Office of Civilian Radioactive Waste Management, U.S. Department of Energy, through Memorandum Purchase Order EA9013MC5X between Bechtel SAIC Company, LLC, and the Ernest Orlando Lawrence Berkeley National Laboratory (Berkeley Lab). The support is provided to Berkeley Lab through the U.S. Department of Energy Contract no. DE-AC03-76SF00098.
|
REFERENCES
|
|---|
- Antonellini, M., and A. Aydin. 1994. Effect of faulting on fluid flow in porous sandstones: Petrophysical properties. AAPG Bull. 78:355–377.[Abstract]
- Aydin, A. 1978. Small faults formed as deformation bands in sandstones. Pure Appl. Geophys. 116:913–930.
- Birkholzer, J., G. Li, C.-F. Tsang, and Y. Tsang. 1999. Modeling studies and analysis of seepage into drifts at Yucca Mountain. J. Contam. Hydrol. 38:349–384.
- Bodvarsson, G.S., W. Boyle, R. Patterson, and D. Williams. 1999. Overview of scientific investigations at Yucca Mountain—The potential repository for high-level nuclear waste. J. Contam. Hydrol. 38:3–24.
- Bruhn, R.L., W.T. Parry, W.A. Yonkee, and T. Thompson. 1994. Fracturing and hydrothermal alteration in normal fault zones. Pure Appl. Geophys. 142:609–644.
- Buesch, D.C., R.W. Spengler, T.C. Moyer, and J.K. Geslin. 1996. Proposed stratigraphic nomenclature and macroscopic identification of lithostratigraphic units of the Paintbrush Group exposed at Yucca Mountain, Nevada.. USGS Open-File Rep. 94-469. USGS, Denver, CO.
- Caine, J.S., J.P. Evans, and C.B. Forster. 1996. Fault zone architecture and permeability structure. Geology 24:1025–1028.[Abstract/Free Full Text]
- Campbell, K., A. Wolfsberg, J. Fabryka-Martin, and D. Sweetkind. 2003. Chlorine-36 data at Yucca Mountain: Statistical tests of conceptual models for unsaturated-zone flow. J. Contam. Hydrol. 62–3:43–61.
- Dahan, O., R. Nativ, E. Adar, and B. Berkowitz. 1998. Measurement system to determine water flux and solute transport through fractures in the unsaturated zone. Ground Water 36:444–449.[ISI]
- Davidson, G.R., R.L. Bassett, E.L. Hardin, and D.L. Thompson. 1998. Geochemical evidence of preferential flow of water through fractures in unsaturated tuff, Apache Leap, Arizona. Appl. Geochem. 13:185–195.
- Faybishenko, B., C. Doughty, M. Steiger, J.C.S. Long, T.R. Wood, J.S. Jacobsen, J. Lore, and P.T. Zawislanski. 2000. Conceptual model of the geometry and physics of water flow a fractured basalt vadose zone. Water Resour. Res. 36:3499–3520.
- Flint, A.L., L.E. Flint, G.S. Bodvarsson, E.M. Kwicklis, and J. Fabryka-Martin. 2001. Evolution of the conceptual model of unsaturated zone hydrology at Yucca Mountain, Nevada. J. Hydrol. (Amsterdam) 247:1–30.
- Flint, L.E. 1998. Characterization of hydrogeologic units using matrix properties, Yucca Mountain, Nevada. Water Resour. Investigations Rep. 97-4243. USGS, Denver, CO.
- Glass, R.J., and M.J. Nicholl. 1996. Physics of gravity-driven fingering of immiscible fluids within porous media: An overview of current understanding and selected complicating factors. Geoderma 70:133–163.[ISI]
- Glass, R.J., M.J. Nicholl, A.L. Ramirez, and W.D. Daily. 2002. Liquid phase structure within an unsaturated fracture network beneath a surface infiltration experiment: Field experiment. Water Resour. Res. 38(10):1199. doi:10.1029/2000WR000167
- Gudmundsson, A. 2001. Fluid overpressure and flow in fault zones: Field measurements and models. Tectonophysics 336:183–197.[ISI]
- Haldeman, W.R., Y. Chuang, T.C. Rasmussen, and D.D. Evans. 1991. Laboratory analysis of fluid flow and solute transport through a fracture embedded in porous tuff. Water Resour. Res. 27:53–65.
- Hinds, J.J., G.S. Bodvarsson, and G.H. Nieder-Westermann. 2003. Conceptual evaluation of the potential role of fractures in unsaturated processes at Yucca Mountain. J. Contam. Hydrol. 62–63:111–132.
- Kilbury, R.K., T.C. Rasmussen, D.D. Evans, and A.W. Warrick. 1986. Water and air intake of surface-exposed rock fractures in situ. Water Resour. Res. 22:1431–1443.
- Lenormand, R., and C. Zarcone. 1989. Capillary fingering: Percolation and fractal dimension. Transp. Porous Media 4:599–612.
- Lopez, D.L., and L. Smith. 1995. Fluid flow in fault zones-analysis of the interplay of convective circulation and topographically driven groundwater flow. Water Resour. Res. 31:1489–1503.
- Lopez, D.L., and L. Smith. 1996. Fluid flow in fault zones—Influence of hydraulic anisotropy and heterogeneity on the fluid flow and heat transfer regime. Water Resour. Res. 32:3227–3235.
- Nicholl, M.J., R.J. Glass, and H.A. Nguyen. 1992. Gravity-driven fingering in unsaturated fractures. p. 321–331. In Proceedings of the Third Annual International Conference on High Level Radioactive Waste Management, American Nuclear Society, Las Vegas, NV. 12–16 Apr. 1992. SNL paper SAND91-1985C. Sandia Natl. Lab., Livermore, CA.
- Nicholl, M.J., R.J. Glass, and H.A. Nguyen. 1993a. Small-scale behavior of single gravity-driven fingers in an initially dry fracture. p. 2023–2032. In Proceedings of the Fourth Annual International Conference on High Level Radioactive Waste Management, American Nuclear Society, Las Vegas, NV. 26–30 Apr. 1993. SNL paper SAND92-2790C. Sandia Natl. Lab., Livermore, CA.
- Nicholl, M.J., R.J. Glass, and H.A. Nguyen. 1993b. Wetting front instability in an initially wet unsaturated fracture. p. 2061–2070. In Proceedings of the Fourth Annual International Conference on High Level Radioactive Waste Management, American Nuclear Society, Las Vegas, NV. 26–30 Apr. 1993. SNL paper SAND92-2791C. Sandia Natl. Lab., Livermore, CA.
- Nicholl, M.J., R.J. Glass, and S.W. Wheatcraft. 1994. Gravity-driven infiltration instability in initially dry nonhorizontal fractures. Water Resour. Res. 30:2533–2546.
- Philip, J.R., J.H. Knight, and R.T. Waechter. 1989. Unsaturated seepage and subterranean holes: Conspectus, and exclusion problem for cylindrical cavities. Water Resour. Res. 25:16–28.
- Piety, L.A. 1996. Compilation of known or suspected Quaternary faults within 100 km of Yucca Mountain. OF 94-0112. USGS, Denver, CO.
- Podgorney, R.K., T.R. Wood, B. Faybishenko, and T.M. Stoops. 2000. Spatial and temporal instabilities in water flow through variably saturated fractured basalt on a one-meter field scale. p. 129–146. In B. Faybishenko et al. (ed.) Dynamics of fluids in fractured rock. Geophysical Monogr. 122. AGU, Washington, DC.
- Pruess, K. 1999. A mechanistic model for water seepage through thick unsaturated zones in fractured rocks of low matrix permeability. Water Resour. Res. 35:1039–1051.
- Salve, R., C.M. Oldenburg, and J.S.Y. Wang. 2003. In situ flow experiments in nonwelded tuff: Conceptual models and implications for flow in the PTn. J. Contam. Hydrol. 62–3:269–286.
- Salve, R., J.S.Y. Wang, and C. Doughty. 2002. Liquid flow in unsaturated fractured welded tuffs: I. Field investigations. J. Hydrol. (Amsterdam) 256:60–79.
- Salve, R., J.S.Y. Wang, and T.K. Tokunaga. 2000. A probe for measuring wetting front migration in rocks. Water Resour. Res. 36:1359–1367.
- Seront, B., T.F. Wong, J.S. Caine, C.B. Forster, and R.L. Bruhn. 1998. Laboratory characterisation of hydromechanical properties of a seismogenic normal fault system, J. Struct. Geology 20:865–881.
- Sidga, J.M., and J.L. Wilson. 2003. Are faults preferential flowpaths through semiarid and arid vadose zones? Water Resour. Res. 39(8):1225. doi:10.1029/2002WR001406
- Smith, L., C. Forster, and J. Evans. 1989. Interaction of fault zones, fluid flow, and heat transfer at the basin scale. p. 41–67. In S.P. Neuman and I. Neretnicks (ed.) Hydrogeology of low permeability environments. Verlag Heinz Heise, Hannover, Germany.
- Stuckless, J.S., and W.W. Dudley. 2002. The geohydrologic setting of Yucca Mountain, Nevada. Appl. Chem. 17:659–682.
- Trautz, R.C., and J.S.Y. Wang. 2002. Seepage into an underground opening constructed in unsaturated rock under evaporative conditions. Water Resour. Res. 38:1188.
- Wang, J.S.Y., and G.S. Bodvarsson. 2003. Evolution of unsaturated zone testing at Yucca Mountain. J. Contam. Hydrol. 62–3:337–360.
- Wang, J.S.Y., R.C. Trautz, P.J. Cook, S. Finsterle, A.L. James, and J. Birkholzer. 1999. Field tests and model analyses of seepage into drift. J. Contam. Hydrol. 38:323–347.
- Yang, I.C., P. Yu, G.W. Rattray, J.S. Ferarese, and J.N. Ryan. 1998. Hydrochemical investigations in characterizing the unsaturated zone at Yucca Mountain, Nevada. Water Resour. Investigation Rep. 98-4132. USGS, Denver, CO.
TOP
ABSTRACT
INTRODUCTION
