Design of a GRS rockfall berm Summitville Mine site

Material Information

Design of a GRS rockfall berm Summitville Mine site Summitville, Colorado
Reiner, Mark B
Publication Date:
Physical Description:
xii, 318 leaves : illustrations ; 28 cm

Thesis/Dissertation Information

Master's ( Master of science)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
Department of Civil Engineering, CU Denver
Degree Disciplines:
Civil engineering


Subjects / Keywords:
Earthwork -- Colorado ( lcsh )
Retaining walls -- Design and construction -- Colorado ( lcsh )
Earth pressure -- Measurement ( lcsh )
Rockslides -- Prevention -- Colorado ( lcsh )
Earth pressure -- Measurement ( fast )
Earthwork ( fast )
Retaining walls -- Design and construction ( fast )
Rockslides -- Prevention ( fast )
Colorado ( fast )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


Includes bibliographical references (leaves 316-318).
General Note:
Department of Civil Engineering
Statement of Responsibility:
Mark B. Reiner.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
43926626 ( OCLC )
LD1190.E53 1999m .R45 ( lcc )

Full Text
Mark B. Reiner
B.A., Valparaiso University, 1988
B.S., Colorado School of Mines, 1996
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Master of Science
Civil Engineering

This thesis for the Master of Science
degree by
Mark B. Reiner
has been approved
Edward Nuhfer
M/vj 24-
( Date

Reiner, Mark B. (M.S., Civil Engineering)
Design of a GRS Rockfall Berm, Summitville, Colorado
Thesis directed by Professor Jonathon T. H. Wu
The highwall at the Summitville Mine is an exposed rock face of South
Mountain, the result of over a decade of open pit mining. The benched rock
slope faces northeast with an overall slope of 45 and a vertical height of
approximately 680 feet. Hydrothermally altered sulfide-bearing rocks contribute
to the complex geology that comprises the face of the highwall. Outcrops range
from blocky, irregular, competent rocks to soft rocks that weather to a clayey
soil. Rockfall at this site occurs as single events and small wedge failures.
Remediation of the site by the Bureau of Reclamation (USBR) has been under
the direction of the United States Environmental Protection Agency (USEPA)
since late 1992 when the mining company abandoned the site after declaring
bankruptcy. Mitigating the rockfall hazard at the highwall is essential to avoid
damaging the remedial strul lctures or reducing the carrying capacity of the
diversion ditches that divert acid rock drainage (ARD) from the highwall to
treatment. Due to the remoteness of the site, the possibility of even a catastrophic
failure has only limited potential for taking human life.
Due to the lack of historical rockfall data, this study gathered data to form a

comprehensive assessment of site geology, physical conditions of the rock, and
investigated the validity and coefficient sensitivity of the Colorado Rockfall
Simulation Program (CRSP). CRSP was utilized to predict the possible kinetic
energies that could be generated from rockfall at the highwall, and determine the
most effective impact face geometry. This study also presents a method for
determining an equivalent pseudo-static force for representing the dynamic
impact from rockfall. SSCOMPPC is a finite element model used to evaluate the
performance of the selected rockfall berm geometry under such impacts.
The results indicated that a geosynthetically reinforced soil (GRS) berm, with an
impact face of 0.25 : 1 (horizontal to vertical), was the most effective in
containing the rockfall from the highwalll.
This abstract accurately represents the content of the candidates thesis. I
recommend its publication.

Six years prior to the completion of this thesis I was teaching high school history
in a small school in Indiana. Many life changes have occurred during this time,
requiring much patience and humor that was provided by my wife Linda and my
little girl Kyla. I promise NO MORE CAREER CHANGES.
Also, a special thanks to Tom and Bruce at Rocky Mountain Consultants, Inc.
for acquiring the job at the Summitville Mine, Roy for his geotechnical insight,
and to Dave for his AutoCad prowess.
I also want to extend my thanks to my thesis committee: Dr. J.T. Wu, Dr. Ed
Nuhfer, and Dr. Mays. All of whom contributed to the continuing of my
education at the University of Colorado at Denver and promoted the concept of
the student/teacher interaction as a positive tool for learning.

Figures......................................................... x
Tables.......................................................... xii
1. Introduction............................................... 1
1.1 Site Background............................................ 1
1.2 Problem Statement.......................................... 2
1.3 Objectives................................................. 2
1.4 Method of Research......................................... 4
1.5 Content of Thesis.......................................... 5
2. Geography and Geology...................................... 6
2.1 Geography.................................................. 6
2.2 Geology.................................................... 8
2.2.1 Deposition of Ore......................................... 11
2.2.2 Rock Types.................................................. 13
2.3 Rock Descriptions............................................ 16
2.4 Altered Quartz-Latite of the Lower Fisher.................. 18
2.4.1 Argillaceous Zones.......................................... 19
2.5 Weathering................................................... 20
2.6 Stability Analysis........................................... 23
2.6.1 Previous Analysis........................................... 23
2.6.2 Authors Stability Analysis................................. 24
2.7 Seismicity................................................. 24
3. CRSP......................................................... 26

3.1 Description of Program.......................................... 26
3.2 CRSP Algorithm.................................................. 27
3.3 Parameters...................................................... 31
3.3.1 Coefficients of Restitution.................................... 32
3.3.2 Rock Shape..................................................... 35
3.3.3 Rock Size...................................................... 35
3.3.4 Slope Material Properties...................................... 39
3.3.5 Surface Roughness.............................................. 39
3.3.6 Rock Motion.................................................... 39
3.3.7 Site Specific Topography....................................... 44
3.4 CRSP Validation................................................. 44
4. GRS Rockfall Berm............................................ 46
4.1 GRS Design................................................... 46
4.2 Effectiveness of GRS as Rockfall Barrier........................ 49
4.3 Long-term Performance of a Sloped GRS Wall.................... 51
4.4 Degradation Properties.......................................... 55
4.4.1 Creep.......................................................... 55
4.4.2 Hydrolysis..................................................... 55
4.4.3 Sunlight (UV).................................................. 56
5. CRSP Analysis................................................. 58
5.1 Rockfall Paths.................................................. 58
5.2 Determination of the Coefficients............................... 63
5.2.1 Parametric Analysis for R..................................... 63
5.2.2 Values Used in CRSP............................................ 73
5.3 USBR Earth Berm Geometry........................................ 73
5.3.1 Proposed GRS Geometry.......................................... 75
5.4 GRS Rockfall Berm Analysis................................... 75

6. CRSP Results and Discussion of Results.........
6.1 Cross Section A-A..............................
6.2 Cross Section B-B.............................
6.3 Cross Section C-C.............................
7. Rockfall Berm Design...........................
7.1 GRS Rockfall Berm Geometry.....................
7.2 Foundation Conditions..........................
7.3 Stability Analysis.............................
7.3.1 GRS Internal Stability........................
7.3.2 GRS External Stability........................
7.4 Construction Details...........................
7.4.1 Geotextile Selection..........................
8. FEM Analysis of GRS Rockfall Berm Performance.
8.1 SSCOMPPC Description..........................
8.2 Methods of Analysis...........................
8.2.1 Pseudo-Static Equivalents.....................
8.3 FEM Model.....................................
8.4 Results and Conclusions.......................
9. Conclusions....................................
9.1 Summary.......................................
9.2 Conclusions...................................
9.2.1 CRSP..........................................
9.2.2 Geotextile Selection..........................
9.2.3 SSCOMPPC......................................
A. AASHTO Ultimate Strength analysis procedures

B. CRSP output files............................................... 126
C. Geotextile Specifications....................................... 258
D. FEM Output...................................................... 262
References......................................................... 316

1.1 Summitville Highwall Photos............................. 3
2.1 Site Location........................................... 9
2.2 Site Map..................................................... 10
2.3 Regional Geology............................................. 12
2.4 Hydrothermal Alteration Zones................................ 14
2.5 Highwall Geology............................................. 15
2.6 Oxidation Zones.............................................. 22
2.7 Stereonet of Highwall Predominant Joints..................... 25
3.1 Surface Roughness............................................ 29
3.2 R,, and R, Graphic........................................... 36
3.3 R Sensitivity............................................... 37
3.4 Rock Shape: Comer Attrition.................................. 40
3.5 Velocity and Shape Relationship.............................. 43
4.1 Apparent Cohesion and Confining Pressures.................... 48
4.2 CDOT West Rifle Wall Cross-Section........................... 50
4.3 Energy-Deformation Curve (CDOT).............................. 53
4.4 Creep Properties of Geotextiles.............................. 57
5.1 Cross-Section A-A........................................... 60
5.2 Cross-Section B-B........................................... 61
5.3 Cross-Section C-C........................................... 62
5.4 Highwall Geology with Cross-Sections......................... 64
5.5 Parametric Analysis for R: Argillic Material................ 66

5.6 Parametric Analysis for R^: Argillic Material.
5.7 Parametric Analysis for R: Argillic Material.
5.8 Parametric Analysis for R^: Vuggy Silica......
5.9 Parametric Analysis for R^: Vuggy Silica......
5.10 Parametric Analysis for R: Vuggy Silica......
5.11 Parametric Analysis for R: Vuggy Silica......
5.12 Proposed USBR and GRS Rockfall Berm...........
6.1 Rockfall Motion: Path A-A....................
6.2 Rockfall Motion: Path B-B....................
6.3 Rockfall Motion: Path C-C....................
7.1 Overall Slope Stability.......................
8.1 FEM Mesh Discretization: GRS Rockfall Berm
8.2 Factor of Safety at 0.4H (Fmax)...............
8.3 Factor of Safety at 0.4H (Favg)...............
8.4 Factor of Safety at 0.6H (Favg)...............
8.5 Factor o f S afety at 0.6H (Fmax).............

2.1 Rock Properties at Highwall........................ 22
3.1 Published R,, R, Coefficients...................... 34
3.2 Known Rock Physical Properties..................... 38
4.1 GRS Long-Term Performance.......................... 52
5.1 Rockfall Path Descriptions......................... 59
5.2 Values Used in CRSP Model.......................... 73
6.1 Cross-Section A-A CRSP Results.................... 78
6.2 Cross-Section B-B CRSP Results.................... 80
6.3 Cross-Section C-C CRSP Results.................... 83
6.4 CRSP Results....................................... 85
7.1 GRS Material Properties............................ 87
7.2 Ultimate AASHTO Stability.......................... 91
8.1 CDOT Duration of Impact............................ 95
8.2 Time Average Force (Fav)........................... 96
8.3 Time Average Force (Fmax).......................... 96
8.4 SSCOMPPC Parameter Values.......................... 98
9.1 Factor of Safety Summary........................... 108

1. Introduction
1.1 Site Background
The highwall at the Summitville Mine is an exposed rock face of South
Mountain, the result of over a decade of open pit mining. The mined profile of the
highwall consisted of fifty-foot walls mined to a slope of 70 degrees, separated by
20 foot horizontal benches, producing an overall slope of 45 degrees. The benched
rock slope has a northeastern aspect and a vertical height of approximately 680 feet,
as shown on Figure 1-1 (a). The sulfide-bearing hydrothermally altered rocks that
comprise the face of the highwall have extremely different weathering rates and
over the past decade have almost completely obscured the bench profile. Outcrops
range from blocky, irregular, competent rocks to soft rocks that weather to a clayey
soil, as shown on Figure l-l(b). Rockfall at this site is frequent, occurring as single
events and in wedge failures, as discussed in the stability analysis in Chapter 2.
Remediation of the site by the Bureau of Reclamation (USBR) has been under the
direction of the United States Environmental Protection Agency (USEPA) since
late 1992 after the mining company abandoned the site after declaring bankruptcy.
Summitville Mine was placed on the National Priority List (NPL) in 1994, due to
the threat of cyanide and metal-bearing processing fluids threatening to overtop the
heap leach pad and the high metal and low pH non-point surface discharges to
tributaries. Mitigating or eliminating rockfall hazards at the highwall is essential to
avoid damaging or reducing the carrying capacity of the remedial work, such as the
diversion ditches that divert acid mine drainage (AMD) to treatment and would
reduce operation and maintenance costs. Also, rockfall from the sulfide-bearing
altered quartz-latite highwall that comes to rest on the proposed vegetated slopes of

the regraded topography would generate ARD locally, killing the nearby
vegetation. Due to the remoteness of the site, the possibility of even a catastrophic
failure has only limited potential for taking human life.
1.2 Problem Statement
The highwall requires a durable rockfall impact berm near the toe of the slope that
will contain preferably all of the potential rockfall without launching the rocks.
Prediction and design for rockfall hazards requires a geomorphic investigation of
the slope, and site specific measurements of rockfall diameter, density, shape,
probable starting zone, and distance traveled. This information could then be used
to back calculate and calibrate the coefficients used in a rockfall program to model
rockfall behavior at the site. However, due to the mining activities and the
subsequent remedial work performed at the site, the final resting locations for
rockfall could not be determined. Therefore, this study focused on alternative
means to determine the parameters required for predicting rockfall kinetic energies
and evaluating the performance and geometry of the rockfall berm under the
dynamic impacts of rockfall.
1.3 Objectives
The objectives and tasks undertaken of this study are the following:
1. Determine the site geology and the physical characteristics of the rock mass.
2. Determine the coefficients in CRSP that represent site conditions, by
3. Determine the geometry of the impact face of the rockfall berm that would
effectively contain rockfall from the worst case scenarios,

Northeastern aspect of the Summitville highwall. View is to southwest.
Differential weathering patterns of the competent silica rocks (left) and the argillic
rocks (right).

4. Design a cost effective rockfall impact berm that can be constructed to the
required geometry,
5. Evaluate the performance and deformation of the rockfall berm under the
predicted impact force with a finite element analysis.
1.4 Method of Research
The objectives of this study will be determined in the following order:
1. A comprehensive study of the geology and remedial work at the highwall
site was essential for the prediction of expected rockfall motion, and a
thorough investigation as to the sensitivity and validity of the parameters
used in the Colorado Rockfall Simulation Program (CRSP). Also, a
parametric analysis was conducted on the parameter that was considered the
most sensitive.
2. Known rock properties and detailed topography were used to predict areas
where rockfall might achieve the highest velocities to impact the berm.
3. The use of Analysis Points (A.P.s) in CRSP allows the user to evaluate
velocity, kinetic energy, bounce height, and number of rocks passed. These
points were moved to determine the performance as the impact face was
progressively steepened.
4. Once the optimum impact face geometry was determined, the analysis point
at the base of the berm was used to determine the average and maximum
bounce heights and kinetic energies. This information was utilized to
determine the magnitude of impact and the zone of the berm that would be
receiving the impact.
5. A review of geosynthetically reinforced soil (GRS) wall performance as a
cost effective means of constructing such a rockfall impact berm was

6. In order to evaluate the deformation of the berm upon rock impact, the
kinetic energies of the rock were converted to pseudo-static forces, which
required that the average duration of impact be determined. The duration of
impact was estimated from test results from the West Rifle, Colorado site
data by comparing rock size and similar kinetic energies.
7. The performance of the rockfall impact berm was then evaluated using a
finite element program that models a non-linear (stress-strain behavior)
backfill soil and the tensile reinforcement properties of the geotextile
reinforcement. This model was used to compare results with energy-
deformation curves that were the output from the West Rifle test site for
1.5 Content of this Thesis
The components of this study of rockfall at the Summitville Mine site include a
review of the site geology and the CRSP rockfall model, analysis and design of the
GRS rockfall berm, and a prediction of performance of the rockfall berm through a
finite element model (FEM) analysis of the structure. Earthwork volumes and cost
estimates are not included in this report. Also, an analysis for a snow avalanche was
not conducted for the Summitville Mine highwall due to the steep inclination and
northeastern aspect.

2. Geography and Geology of Summitville
2.1 Geography
The Summitville Mine Superfund Site is located in the southeastern portion of the
San Juan Mountains, in the southwest comer of Rio Grande County, approximately
25 miles south of Del Norte, Colorado, as shown on Figure 2-1, The site covers the
southern one-half of Section 30 and the northern one-third of Section 31 of T37N,
R4E, of the 6Ih principal meridian. The site is located within the San Juan mountain
range of the Rocky Mountains, approximately two miles east of the Continental
Divide. Altitudes at the site range from 11,150 feet at the outlet works of the
Summitville Dam to approximately 12,300 feet at the highest extent of mine
workings. The district is remote, relatively unpopulated, and is surrounded by the
Rio Grande National Forest. Most of the surrounding peaks range between 12,300
feet to 12,800 feet, but Montezuma and Summit Peaks, located approximately 6 to
8 miles to the southwest reach altitudes of 13,131 and 13,272 feet respectively.
Due to the high altitude, Summitville has long, cold winters, and short, cool
summers. Snowfall averages more than 400 in/yr, and protected snowbanks on
northern aspect slopes can persist throughout the year. Thunderstorms are common
in the afternoon hours during the months between May and September and can be
very intense with a short duration. Many of the northern aspect slopes, and most of
the lower slopes are heavily covered with spruce and interspersed with stands of tall
aspen at the lower elevations of the mine site. Mining, lumbering, and grazing have
been the primary historic industries associated with the Summitville mine and
surrounding areas.
The permitted 1,231 acre mine contains approximately 550 acres of disturbed area,

most of which is positioned on the northeastern flank of South Mountain, as shown
on Figure 2-2. The site is located in the Rio Grande Drainage Basin near the head
waters of the Alamosa River, and is drained by two tributaries, Wightman Creek
Fork and Cropsy Creek. The confluence of the two streams is located near the
northwest comer of the Site, approximately 4.5 miles upstream of the confluence
with the Alamosa River. The site is bounded by Wightman Fork Creek and the
deserted settlement of Summitville to the north, Cropsy Creek to the south, and the
mine workings of the South Mountain highwall to the southwest. The highwall is
one of the more notable features of the mine site, as shown on Figure 1-1. The cut
is a result from nearly ten years of open pit mining by SCMCI (Summitville
Consolidation Mining Company, Incorporated) above the existing historic
underground workings. The overall vertical height of the highwall is approximately
700 feet in the center of the exposure and laterally extends approximately 3000 feet.
Colorful outcrops and scant vegetation along the limbs of the highwall indicate
areas of extremely altered rock.
Although mining ceased in the fall of 1991, heap leaching of the mined ore
continued until the spring of 1992. The mine was abandoned by SCMCI on
December 16,1992, after the company declared bankruptcy. Under the Superfund
Emergency Response authority, the USEPA took over supervision of the
remediation activities on the site the following day. Due to the high sulfide content
of the exposed rock at the highwall and stockpiles of wasterock and leached ore
rock, the primary concern was to mitigate the potential for ARD. USEPA remedial
activities included the lining of the open mine pits, at the base of the highwall, with
compacted clay and lime kiln dust (LKD) for the burial of various stockpiles of
mine wasterock. Four-foot thermal protection, low-permeability caps were placed
on the mine pits to direct surface water and limit infiltration into the buried

wasterock. During the remedial efforts at the site, Summitville was added to the
Superfund national priorities list in May 1994.
2.2 Geology
The Summitville ore deposit lies in the southeastern portion of the mid-Tertiary San
Juan volcanic field. San Juan volcanism began about 35 Ma and culminated with
the formation of over a dozen calderas from about 30 to 26 Ma (Lipman and
Steven, 1970). The Summitville caldera was formed following the quartz latite
eruption that produced the lower Fisher Formation, approximately 30 to 28 Ma
(Gray and Coolbaugh, 1994), and subsequently filled with Summitville andesite, as
shown on Figure 2-3. The quartz latite volcanic dome of South Mountain is
coarsely porphyritic with feldspar, sanidine, and quartz phenocrysts ranging in size
from a few mm to 5cm. The emplacement of the South Mountain dome developed
over six distinct events; 1) early main-stage acid-sulfate alteration, 2) subsequent
Cu-sulfide and gold mineralization, 3) widespread hydrothermal brecciation, 4)
deposition of volumetrically minor, base metal-sulfide-bearing barite veins, 5)
emplacement of small scale, local kaolinite matrix breccias, and 6) a final stage of
supergene oxidation and secondary sulfide enrichment (Gray et al, 1995).
The altered Conejos rhyodacite and the altered quartz latite, that comprises the bulk
of South Mountain, were of solfataric type alteration. Vapors and gasses that
circulated through preexisting fractures within the volcanic dome condensed into
hydrothermal fluids, rich in sulfuric acid. Along individual fractures and veins, the
alteration extended concentrically outwards from the irregular pipe-like, coarsely
porphyritic, quartz latite veins. Extensive leaching removed the feldspar

General location of the Summitville Consolidated Mining Company Inc. Facility
Figure 2.1

2000 fee*
Facility and site layout of the SCMCI Mine
Figure 2.2

phenocrysts, creating a vuggy silica zone. This was followed in order laterally by
quartz-alunite, quartz-kaolinite, argilliceous zone (containing chlorite,
montmorillonite, pyrite, and some calcite in various proportions), and peripheral
propylitic alteration, as shown on Figure 2-4. The vuggy silica and quartz-alunite
zones range from a few centimeters to several tens of meters in diameter, but can be
much wider were fractures intersect. The quartz-kaolinite zones reach a maximum
thickness of 5m and average lm. The most volumetrically significant zone is the
pervasive argillically altered rock, comprising nearly two-thirds of Summitvilles
open pit geology. The volume of the argillaceous material is significant due to the
reduced capability to consume ARD and acid mine drainage (AMD). Due to the
differing permeability of the various zones, oxidation depths range from nearly
100m in the vuggy silica to less than 10m in the argillaceous material.
2.2.1 Deposition of Ore
The Summitville epithermal acid sulfate deposit is located in the mid-level portion
of a mineralized volcanic dome (Gray and Coolbaugh, 1994). The primary ore
zones followed two northwest trends, N30W (20) and N60W. Supergene
oxidation leached copper and sulfide minerals forming ferruginous oxidized zones.
The oxidizing fluids leached the feldspar phenocrysts of the quartz latite porphyry
leaving vuggy silica until they neutralized into alunite pseudomorphs.

FROM BOR DRAWING 1556-6-082-129.
U.S.G.S. MAP I828.
Quartz latite flows and volcanic vents, vt
Commonly bleached and altered
Massive quartz latite flows. Crosses indicate area wt
large sanidine phenocrysts are abundant throughout
rock closely resembles the quartz latite dike rocks,
massive lavas are cut by a local vent complex, vt
Massive rhyodacite flows
Interleaved with pyroclastic rocks
Local rhyodacitic vitrophyre flows
Dark porphyritic andesite flows and sparse pyroclasti
* > V v
r > V V
V V v V
,'v V V

Near surface oxide zones contain the highest Au grades and decrease progressively
with depth. The vuggy silica and quartz-alunite veins, as shown on Figure 2-5,
surrounded by envelopes of argillic material, were the primary zones of
mineralization. The primary ore minerals, pyrite and enargite, deposited in the
vuggy silica and quartz alunite veins along with minor accessory minerals; native
sulfur, barite, galena, sphalerite, and gold (Steven and Ratte, 1960). The ore
deposits in Summitville occurred in the more resistant bodies of quartz and alunite
that were formed by the replacement of the original quartz latite.
2.2.2 Rock Types
The bedrock geology of South Mountain is comprised of the following rock types,
as shown on Figures 2-3 through 2-5.
The rocks that characterize the southern district of Summitville, bordered to the
south by Alamosa Creek, and extending across to Cropsy Creek, are soft, extremely
weathered, and highly colored. The yellowish to whitish clay (opal) that is visible
on Cropsy Ridge are some of the most altered rocks in the Summitville district.
Areas underlain by these rocks are typically devoid of or have scant vegetation. The
lack of silicified bodies and mineralized zones distinguishes the Conejos from the
altered quartz latite rocks of South Mountain. Kaolin, sericite, alunite, and quartz
are common constituents of the Conejos alteration products (Stevens and Ratte,

Postulated cross-section of the South Mountain hydrothermal alteration zones.
Figure 2.4

The subcircular area of Lower Fisher quartz-latite on Figure 2-3, extending from
the north base of South Mountain southward to Cropsy Creek, underwent an
intravolcanic solfataric type alteration. The alteration event was not long enough to
permit intense alteration except near the coarsely porphyritic core of the dome. The
feldspar phenocrysts were dissolved by the magmatic waters, leaving large voids
within the silicified skeleton. As the waters neutralized, the feldspar phenocrysts
were replaced concentrically around the veins by pearly pink to white alunite
pseudomorphs. Other alteration products associated with the altered rocks of the
lower Fisher Formation are the clay minerals; illite, chlorite, montmorillonite,
kaolinite, and smectite.
2.3 Rock Descriptions
(refer to Figure 2-5 and Stevens and Ratte, 1960)
Coneios flows (TF?):
The Conejos is highly altered by hydrothermal processes into strongly sheeted
rhyodacitic and andesitic flows that form blocky outcrops. Generally, the rocks
appear as dark drab to greenish-grey, with an aphanitic groundmass containing
plagioclase phenocrysts (Stevens and Ratte, 1960). The phenocrysts comprise
approximately 30% of the rock by volume, ranging in size from 0.1mm to a few
cm. Subsequent alteration replaced much of the plagioclase with calcite and
Lower Fisher, quartz-latite (Tflq):
The lower member of the Fisher is comprised of altered rhyodacitic and quartz
latite flows. The outcrops occur as a structure-less, moderately fractured, grey-blue
porphyritic rock. In places, large sanidine and quartz phenocrysts, up to 5 cm in
diameter, are irregulary distributed throughout the formation. Zones of

e> o
Quartz-alunite rock and
rock, a; or vuggy quartz
Undivided rocks of the il
and montmorillonite-chlo
Dark andesitic porphyritic
sparse pyroclastic beds

hydrothermal alteration are arranged concentrically around the irregular pipe-like
mineralized veins within the quartz latite, and laterally become progressively less
altered. The following are alteration products within the lower Fisher quartz latite.
Lower Fisher, rhvodacite (Tfld):
The lower Fisher rhyodacite occurs as massive rhyodacite flows, 5 to 50 feet thick
with very little pyroclastic or fragmental material. The flows are cut by two
intrusive bodies of dioritic and quartz monzonitic origin.
Lower Fisher, pyroclastics (TflF):
Bedded tuff and volcanic breccia comprise the limited volume of pyroclastic rocks
at the site. Generally, the bodies outcrop in small, scattered locations across the site.
The rocks range from soft friable white to earthy gray tuffs and breccias, to hard
brown and gray highly silicified tuffs.
Lower Fisher, vitrophvre ('T(1..):
Rhyodacitic vitrophyre flows of the lower Fisher represents chill phases between
intrusive dikes and the country rock. Dull-grey to black glassy matrix with tabular
plagioclase less than 1/8 in length. Locally altered to a yellowish clayey
groundmass rock containing earthy white pseudomorphs of plagioclase.
Upper Fisher, unaltered latite (T^):
Unaltered quartz latite flows and volcanic vent complexes that cap Cropsy Peak and
Cropsy Ridge. These porphyritic rocks are strong and locally as much as 400 feet
thick. The matrix is grey, pink or buff with feldspar phenocrysts, typically rounded
to rectangular glassy white crystals, 2 cm or less in major diameter.

2.4 Altered Quartz-Latite Rocks of the Lower Fisher
(Refer to Figure 2-5)
The following rock descriptions identify the rock types that comprise the face of the
highwall. Their physical properties and weathering characteristics are used to
calibrate the rockfall model.
Vuggy Silica: Leaching has removed all constituents of the original latite
other than quartz, rutile, and a few zircon grains. The original feldspar
phenocrysts have been dissolved by acidic magmatic waters and are
represented by voids within a fine-grained quartzose matrix that ranges from
porous to compact. Euhedral to subhedral pyrite crystals are commonly
disseminated through the matrix. Oxidation at the surface leaves secondary
ferruginous minerals such as limonite. The rock is generally strong and
grey-blue with heavy iron staining on joint surfaces. Joint infill is also grey-
blue and ranges from a sandy clay (SC) to a low-plasticity clay (CL). Zones
of vuggy silica range from a few cm to several 10s of m wide.
Quartz-Alunite (qa): This is a hard, silicified rock that contains alunite
pseudomorphs of feldspar phenocrysts enclosed in a matrix of interlocking,
fine-grained quartz with minor alunite. The pearly pink to white alunite
phenocryts range in size from a few mm to 5cm. The quartz-alunite is
slightly to moderately weathered, strong, and has a greyish-blue matrix.
Quartz-Kaolinite: As the magmatic waters neutralized, the original feldspar
phenocrysts weathered into argillic kaolinite. The quartz-kaolinite zone also
contains illite, montmorillonite and pyrite. The rock is relatively soft and
weathers rapidly to a low-plasticity, grey-blue clay. This is a transitional

phase of alteration between the intensely weathered argillaceous zones and
the vuggy silica/quartz-alunite zones.
2.4.1 Argillaceous Zones
(as identified by Stevens and Ratte, 1960):
illite-kaolinite zone (ik): Highly altered argillaceous, soft, whitish-
grey bleached rocks. Kaolinite and quartz are pseudomorphs of the
original feldspar phenocrysts. The matrix consists of granular
patches of quartz and illite (white) enclosed in a finer grained
aggregate of quartz and montmorillonite (gray).
montmorillonite-chlorite (me): Slightly less altered zone with a
higher montmorillonite (gray) content. Plagioclase crystals range
from virtually unaltered to slight montmorillonite alteration. While
the quartz phenocrysts are completely unaltered. Former biotite
phenocrysts are replaced by illite-sericite pseudomorphs within the
fine grained matrix.
Propylitic: Slightly altered quartz latite rock contains feldspar
phenocrysts that show chlorite or montmorillonite rims enclosed in a
grey-blue matrix of interlocking, fine-grained quartz. The cloudy
pink phenocryts range in size from a few mm to 5 cm, and contain an
assemblage of chlorite, pyrite, and some calcite. These rocks are
relatively hard and form blocky outcrops that are most commonly
found around the periphery of the mine site.

2.5 Weathering
The most severe erosion prior to the mining activities at the site occurred during the
last San Juan uplift, approximately 5 Ma, and caused deeply incised drainages near
the geologic contacts. The physical weathering of the more altered rocks
accumulated as thick deposits of colluvium and alluvium in the valley, ranging
from a few inches to more than 50 feet thick. Erosion gradually exposed more of
the deposit to weathering and increased oxidation due to contact with oxygenated
ground and surface waters. Within the mining area of Summitville, oxidation was
primarily focused on the high-permeability vuggy silica that was oxidized to 300
feet, while the surrounding argillic zones had oxidation profiles that extended only
30 feet, as shown on Figure 2-6. Sulfur compounds contained within the altered
rocks are easily oxidized by weathering processes. When pyrite is oxidized,
ferrous sulfate and sulfuric acid are formed. Subsequent minerals formed include
melanterite and jarosite which expand in volume by 536% and 115% respectively
(Bell, 1994). This expansion in volume, and subsequent exposure of more rock
material is evident due to the accelerated weathering processes on the highwall at
Summitville. Analyses of the ore bearing rock within the heap leach pad (HLP) at
Summitville have found that 21% of the volume by weight is alunite and jarosite,
and they form the bulk of the sulfates.
For silicate minerals, weathering is primarily due to hydrolysis. Deposits of kaolin
are formed when percolating acidified waters decompose the feldspar. Presently,
the original sulfides along the exposed surface have already been oxidized leaving
behind secondary weathering minerals such as; jarosite, limonite, goethite, and
chalcanthite. In parts of the open pit, resistant silicified bodies are completely
surrounded by very weak argillic rock. The benches that form the highwall are
gradually becoming increasingly obscured as the weak rock from the argillic zones

accumulates at the base of each rise. The long term weathering profile can be
expected to maintain sharp contacts between these two bodies as the argillic
material weathers to a natural angle of repose, or incised during high precipitation
events, and the silicified bodies retain their integrity. The less altered propylitic
rocks weather to slabby blocks and are subject to mass wasting that contribute
much of the talus found at the toe of the slopes of the highwall.

/ Quartz-
Depth of
4 -
Weight % 3 -
in rock
1 -
Unoxidized rock
vuggy silica,
qtz. al unite

nt-.*- -.-j

depth of
9 6 3 0 3 6 9
Meters from vuggy silica
12 15
Schematic showing approximate depth of oxidation and sulfur content in weight %
Figure 2.6

2.6 Stability Analyses
2.6.1 Previous Analyses
A preliminary pit slope stability analysis was completed by C.G. Ramos and T.V.
Rao (Ramos and Rao, 1983) for the Anaconda Mining Company. Pit bottoms were
projected to extend to an elevation of 11,680 in the North Pit and 11,800 in the
South Pit. The report listed factors favoring slope stability as: 1) silicification type
of alteration, 2) low slope heights on eastern flanks of pits, 3) absence of low
dipping structures that could cause plane-type failures, and 4) presence of
underground workings as drains. Also listed were the detrimental features of the
site: 1) argillic alteration, 2) high slope heights, 3) NE trending structures that
could form wedge failures with NW trending major faults and lineaments, and 4)
high amount of precipitation.
The report concluded that the NE trends mapped and projected on stereonets did
not appear to be detrimental to slope stability. The analysis compared several cross
sections and, in the most conservative case, modeled the entire slope as a
cohesionless clay. The results found that a 45 degree slope overall would
provide adequate stability, with the factor of safety ranging from 1.09 to 9.10.
The USBR has not completed a slope stability analysis of the highwall, but seems
to accept the results from the Ramos and Rao report. Currently, the highwall has
slopes ranging from 45 to 60 degrees.
2.6.2 Authors Stability Analysis
The orientations of nineteen joint sets and the face of the highwall (dip / dip
direction), apperatures, roughness, and texture were noted during a single day

reconnaissance of the highwall on September 2, 1998. The data was taken from
fifteen areas that extended from the southeastern limb to the northwestern limb,
following the crest of the highwall. The trends were modeled as being persistent
throughout the highwall and entered into ROCKPACKII software for a stereonet
analysis for potential modes of failure (i.e. plane, wedge, and/or topple), as shown
on Figure 2-7.
Because failure on the highwall will occur along joint sets, the shear strength
parameters of the joint infilling are the critical data required for a more thorough
analysis. However, the results of the stereonet analysis shows that the potential for
a wedge failure exists along an intersection of joint sets that trends N44E with a
plunge of 40 degrees.
2.7 Seismicity
Major extensional faulting began in the southern Rocky Mountains approximately
28 million years ago, and has continued to the present. Fault movement within the
last 28 million years has been labeled as potentially active by Kirkham and
Rogers (1981). A study by SCMCI found that three regional faults displace
Quaternary deposits and are therefore considered active. However, based on the
length and proximity of the fault to the site, the predicted peak horizontal ground
acceleration would not exceed the O.lg, horizontal as specified in the Uniform
Building Codes (UBC).

Discontinuities at the highwall indicate potential for toppling and wedge failure.
Highwall face strikes NW 43 degrees and dips 45 degrees to NE.
Figure 2.7

3.1 Description of Program
The Colorado Rockfall Simulation Program (CRSP) is one of the few empirically
calibrated computer rockfall simulation programs. The program was developed at
the Colorado School of Mines and is based on empirical data, mostly provided by
several state departments of transportation. CRSP models a random occurrence of
rockfall from a source area to the final downslope resting location. The rockfall
motion predicted is based on Newtonian laws of physics as well as slope and
material properties, as discussed below. The data that supported the recalibration of
CRSP included rocks with diameters ranging from 1.0 feet to over 5.0 feet, slope
angles from 30 to 45, and rock types ranging from competent limestone and
granite to a clayey silt slope covered with 1 to 18 rock fragments. However,
CRSP was not recalibrated for soft soil slopes or coarse talus slopes. CRSP, version
4.0, was recently changed to include angular velocity in the modeling of rockfall
motion. This was not included on older versions and as a result, CRSP was thought
to underestimate rockfall velocities.
CRSP assumptions are: 1) because the rockfall path is determined by the user,
CRSP models the rockfall path in two-dimensions; 2) the coefficients of normal
restitution and tangential frictional resistance, and R, respectively, can account
for the material properties that comprise the rock and slope; 3) the size and shape of
the rock modeled does not change by breaking apart during the analysis; and 4)
CRSP assumes a spherical shape for the rock to calculate volume and inertia. All
of these assumptions are discussed in detail in the following paragraphs.

Suggested coefficients of restitution, that account for the material properties, were
also revised to account for the new data in terms dictated by the principles of the
conservation of energy and gravitational acceleration. The hazards associated with
quantifying rockfall phenomena are the variations within the slope profile that can
launch the rock at random trajectories. These include changing slope inclination,
vegetation cover, and changing slope material along the rockfall path. CRSP has
approached this problem by requiring field measurements of the slope profile, rock
size and shape, and requiring the coefficients of restitution to account for non-
elastic properties of the rock materials.
3.2 CRSP Algorithm (Jones, 1998)
CRSP is a stochastic model that attempts to model the random variations along a
slope profile. The slope angle (4>) within each cell will change between rock
impacts up to the limit set by the maximum probable variation in the slope (0max).
The maximum allowable slope angle variation is a function of rock size and surface
roughness, with the general equation of:
0max = tan- (S/R) (3.1)
The surface roughness (S) is defined as the amplitude of surface variation,
perpendicular to the overall slope, within a slope distance equal to the radius of the
rock being modeled (R), as shown on Figure 3-1. At impact, the incoming velocity
(Vj), and impact angle (a) are used to calculate the tangential and normal
velocities, and the rate of rotation.

V = V, cos a
Vnl = V, sin a (3.3)
The new tangential velocity, after impact, is calculated from the conservation of
(Vi I w,2 + /2 MV2 ) */(F) SF = % Iw22 + y2 MVt22 (3.4)
where: M = rock mass
I = rock moment of inertia
I = 2MR2/5 (sphere)
I = MR2/2 (disk)
I = MR2/4 + ML2/12 (cylinder, L = length)
Wj = initial rotational velocity
w2 = final rotational velocity
Vtl = initial tangential velocity
= final tangential velocity
/(F) = friction function = R, + (l-RJ/tfO^, w,R)/20]2 + 1.2}
SF = scaling factor = R, / {[VnI / (250R)]2 +1}

Surface roughness (S) established as the perpendicular variation from an average
plunge line (defined by slope angle ) over a distance egual to the radius of the
rock (R). Maximum slope variation is defined by S and R (Jones, 1998)
Figure 3.1

An increase in velocity normal to the surface results in a greater normal force,
during impact, and the scaling factor adjusts for increased frictional resistance. The
energy lost during the impact is determined from the difference between rotational
and tangential velocities, the velocity normal to the slope, and the tangential
coefficient. Constants used in the friction function and the scaling factor were
determined by experiment. Solving equation 3.4 for the new tangential velocity of a
bouncing rock yields equation 3.5.
Vt2 = {[R2 (Iw,2 + MV,,2) *J[F) SF] / (I + MR2)}05 (3.5)
A new normal velocity (V^) can be determined by equation 3.6
vn2 = {(VnA) / [1 + (Vnl / 30)2]} (3.6)
However, observations from test data (Jones, 1998) show that, regardless of the
initial rotational velocity, rocks always leave the surface in the rolling mode. The
relationship between rotational velocity and tangential velocity can be determined
by equation 3.7.
Vt2 = w2R (3.7)
The new incoming +x and -y velocity components and gravitational acceleration
are input to calculate the time between impacts along the slope profile.

3.3 Parameters
CRSP suggests that the user select the upper end of suggested coefficient ranges, to
maximize velocity and kinetic energy, if the results are to be used for design
purposes. The best method for acquiring results with a good degree of confidence
requires the user to accurately measure the surface roughness along the rockfall
path. This was not possible at the Summitville highwall due to the extremely steep
topography and high frequency of rockfall. Aerial photographs of the site in 1997,
1994, 1993, and field notes regarding geology, rock shape and size, and various
geologic reconnaissances of the highwall were used to produce the input for the
CRSP model.
The current 4.0 version of CRSP (Jones, 1998) claimed that rock shape, initial
velocity (V,), and tangential coefficient of frictional resistance (R,) had only minor
influences on the output of the model. The most sensitive parameters were slope
inclination, surface roughness (related to rock size), and the coefficient of normal
restitution (RJ. Therefore, this study concentrated efforts to calibrate these latter
three parameters to simulate site specific conditions.
The reduction of error and producing a model that fairly represents actual site
specific conditions, can be accomplished by itemizing those parameters that greatly
affect the output of the model. Due to the fact that rockfall is subject to many
variable types of motion, irregular geometry and different material properties,
attention was focused on the parameters of primary importance. In the recalibration
of CRSP, Jones (1998) determined that slope angle was the most important factor
in determining the terminal and impact velocities of the rock, and therefore, the
kinetic energy imparted to the structure. Surface roughness also was sensitive in

determining rockfall velocity. An increase in roughness will generally result in a
significant decrease in velocity. The non-elastic rock material properties are
important for determining energy lost on impact, however, CRSP is not set up to
accept material coefficients. Instead, the overall coefficients of restitution within
each cell accounts for these properties by decreasing the x and y velocity
components. Of secondary importance, rock size and shape affect the afore
mentioned parameters, all of which are discussed in greater detail below.
3.3.1 Coefficients of Restitution
According to Richards (1992), the greatest effect on rockfall behavior are the
coefficients of restitution (r) which is a ratio of the rocks pre-impact and post-
impact velocities.
Spang and Rautenstrauch (1988) gave the best summary of this parameter. They
concluded that site specific empirical measurements were necessary because the
coefficients of restitution are not material constants; but rather, are dependent on
particle velocities, geometry, and materials. This can be accomplished by noting
site rockfall size, shape, and distance from the probable starting zone and back-
calculating r. This approach was not possible in the case of Summitville, because
remedial activities have totally removed historic rockfall paths and final resting

Jones (1998) states that on steep slopes, the effect of coefficients becomes
negligible as rocks impact less often and increases in importance as the slope
lessens. Although the slopes are steep at Summitville, the benched terraces are
almost completely obscured by erosion in many locations and are weathering to a
more uniform slope. Other locations on the face of the highwall consist of random
competent zones of dike rock that create launching zones. This indicates that
rolling and sliding will occur in combination with free fall motion, and an
investigation into the coefficients is necessary.
Richards (1992) cited several studies that determined that better results for rockfall
motion were obtained when separate parameters are used for the normal and
tangential components of restitution, rn and rt respectively. The coefficient of
normal restitution (rn) is the degree of a perfectly elastic collision normal to the
slope, where rn = 1 represents complete conservation of energy in the normal
direction. While the coefficient of tangential frictional resistance (rt) is the measure
of frictional resistance to movement parallel to the slope. Both are represented
graphically on Figure 3-2. The sensitivity of the coefficients was observed in tests
sited by Richards (1992) where rn and rt were varied separately. The results, as
shown on Figure 3-3, clearly show that on a benched terrace rnis much more
sensitive, which corresponds to Jones findings while recalibrating CRSP.
Published values for the coefficients are presented below on Table 3-1 (compiled
from Jones (1998) and Richards (1992)). There are no known studies correlating rn
and rt to intact rock properties; such as, uniaxial compressive strength, tensile
strength, modulus of elasticity, or durability. Direct-calculations of rn and rt are not
possible due to the remedial work at the highwall site, therefore, it is necessary to
rely on suggested coefficients from published tables of empirical tests (Table 3-1).

Table 3-1: Summary of Published Values
Description of Slope Value for r Value for rn Value for rt Remarks,'References
Smooth hard surface 0.6-1.0 0.9-1.0 Jones (1998)
Most bedrock and boulder fields 0.15-0.30 0.75-0.95 Jones (1998)
Talus and firm soil slopes 0.12-0.20 0.65-0.95 Jones (1998)
Soft soil slopes* 0.10-0.20 0.50-0.80 Jones (1998)
Vineyard slopes 0.40 Richards (1992)
Rock slopes 0.85 Richards (1992)
Rock-rock (competent) 0.75-0.80 Richards (1992)
Solid rock 0.90-0.80 0.75-0.65 Richards (1992)
Detrital material mixed with large boulders 0.80-0.50 0.65-0.45 Richards (1992)
Compact detrital material with small boulders 0.5-0.4 0.45-0.35 Richards (1992)
Grass covered slopes or meadows 0.4-0.2 0.3-0.2 Richards (1992)
Clean hard bedrock 0.53 0.99 Richards (1992)
Asphalt roadway 0.40 0.90 Richards (1992)
Description of Slope Value for r Value for rn Value for r, Remarks/References
Outcrops, large boulders 0.35 0.85 Richards (1992)
Talus cover 0.32 0.82 Richards (1992)
Talus cover with vegetation 0.32 0.80 Richards (1992)
Soft soil, some vegetation 0.30 0.80 Richards (1992)
Rock wall 0.80 0.80 Richards (1992)
compact rock scree or thin soil cover 0.50 0.70 Richards (1992)
*Slope coefficients were extrapolated due to lack of data.

Table 3-2 is a compilation of rock property information obtained from Summitville
documents purely for purposes of characterization. The substantial difference in
recommendations on Table 3-1 indicates the importance of accurately selecting
coefficients that will best represent site conditions.
3.3.2 Rock Shape
The altered rocks found near the toe of the highwall at Summitville, that had
apparently fallen some significant vertical distance showed signs of attrition at the
comers and are sub-rounded rather than an angular block. Most analytical studies
of rockfall assume a spherical shape for the rocks modeled (Spang and
Rautenstrauch, 1988). Studies quoted by Richards (1992) cited that boulders,
regardless of original shape, became progressively more rounded due to attrition at
the comers and edges as shown on Figure 3-4, and that the size and shape had little
influence on its falling or rolling characteristics. Therefore, the rock shape used for
the highwall in the CRSP model is spherical.
3.3.3 Rock Size
The average rock size at the base of the highwall at Summitville has a diameter of
1.25 feet and the maximum diameter noted was 3.5 feet. Regarding the
recalibration of CRSP (version 4.0), Jones (1998) stated that rock size in itself is
unimportant; however, the rock radius does affect the influences of slope angle and
surface roughness as in equation 3.1. The durability of a rock to sustain impact
without breaking should be considered when estimating rock size for the model.
Fragmentation of the rock will dissipate the kinetic energy as the rock loses mass.
Since CRSP does not model the fragmentation of a rock during the rockfall path, it

Impact angle (a) defined as a function of rock trajectory, slope angle, and slope
variation. Rock velocity (V) is reduced into normal (Vn) and tangential (Vt) components.
The tangential coefficient of frictional resistance (R,) and the normal coefficient of
restitution (R) act to decrease the falling rock's velocity (Jones, 1998)
Figure 3.2


50 0
50 0
Horizontal distance (metres)
The effect of rockfall trajectory of varying r
(Reproduced with permission, Richards, 1992)
The effect of rockfall trajectory of varying r,
(Reproduced with permission, Richards, 1992)
of resorarion

Rock Unit Altered State Porosity (%)1 Bulk Density (am/cm3)1 Uniaxial Compressive Strenath (PSD2 Angle of Internal Friction, d> 2 Cohesion, (PSf) 2
Andesite N/A

Quartz- Latite Unaltered 10 2.33
Vuggy Silica 5 2.77 8080
Quartz-Alunite 3 2.67
Kaolinitic 22 2.04
lllitic 18 2.11
Montmorillonite-rich 15 2.19
Chlorite-rich 4 2.50
Moderately Weathered Argillic 1480
Moderately Weathered Argillic 2000
Moderately Weathered Argillic 4150
Moderately Weathered Argillic 2590
Highly Weathered Argillic 720
Highly Weathered Argillic 1600
Completely Weathered Argillic 300 31 o'
37 oA
1 Values obtained from Klohn Leonoff report to SCMCI, "Hydrological and Geotechnical Engineering, Summitville, Colorado, August
29, 1984, Volume 1 of 2
2 Values obtained from Steven, T. A. and Ratte, J. C., 1960 "Geology and Ore Deposits of the Summitville District San Juan
Mountains Colorado" USGS Professional Paper 343

is important to input the final representative size of rock at the site based on field
notes taken at the site.
3.3.4 Slope Material Properties
Rockfall velocity will also be affected by the slope material properties in
combination with the chosen rock size. Larger rocks tend to embed in softer slopes
more than smaller rocks and will lose more rotational energy. According to Jones,
increasing rock size has a decreasing affect on velocity.
3.3.5 Surface roughness
CRSP defines surface roughness as the greatest measurement that occurs with some
frequency within each slope distance of one-rock radius, as shown on Figure 3-1.
The variability of the surface within each cell and the interaction of the
irregularities of the slope with the impacting rock has a great influence in
determining the rockfall behavior. CRSP models surface variations by randomly
varying the slope angle between limits defined by the rock size and maximum
surface roughness. As surface roughness increases, the impact angle will increase
which results in a loss of velocity and energy upon impact.
3.3.6 Rock Motion
Several important concepts need to be considered prior to evaluating the type of
rockfall motion that is to be modeled for the site. When a rock comes into contact

(E) Rock fragment cuboid
shape when freshly
(ii) Corners break off as
fragments moves dowa
(iii) If distance travelled is
long it may become
Change in bourder shape with distance traveled: a) rock fragment cuboid in
shape when freshly dislodged; b) comers break off as fragment moves down
slope; c) if distance traveled is long it may become rounded. (Reproduced
with permission, Richards, 1992)
Figure 3 .4

with the surface of the slope two important interactions occur, 1) kinetic energy is
lost due to the non-elastic components of the rock material, and 2) the frictional
resistance between the two surfaces reduces velocity components. The affects of
these interactions will result in rockfall assuming one or a combination of the
following types of motion (Richards, 1992):
Throw: an initial velocity creates a parabolic rock trajectory,
Free Fall: Falling rocks experience uniform acceleration due to gravity and
are governed by the following equations:
Where V is the resultant velocity, u is the initial velocity, a is the
gravitational acceleration constant, s is the distance, and t is the time of
travel, the affect of air friction on the velocity components is generally
Bouncing: horizontal and vertical velocity components are affected by slope
angle and coefficients of restitution,
Rolling: an angular velocity component reduces translational velocity, and
Sliding: the velocity is a function of the coefficient of kinetic friction and
slope inclination. Hoek (1987) described the velocity of a rock during
rolling or sliding as:
V = u + at
s = ut + Zi at2
V2 = u2 + 2 as
V = (V02 + 2sgK)0'5

Where K is a slope constant defined by the slope angle (0) and angle of
internal friction ():
1. Pure Sliding: K = sin0 cos0 tancj)
2. Rolling Sphere: K = 5/7 sin0
3. Rolling Cylinder: K = 2/3 sin0
Hoek also concluded that a reasonable estimate of K was K = sin0. Figure 3-5
shows that rolling, regardless of shape, produces velocities of about 50-70% of that
obtained in free fall. The energy lost during impact with the slope is a function of
the difference between tangential and rotational velocities.
In order to adjust the parameters of the rockfall model to site specific conditions, it
is necessary to determine what type of motion can be expected along the profile of
the rockfall path. Ritchie (Ritchie, 1963) produced the following general criteria for
rock motion based on slope angle from his experiments in the State of Washington:
1. Rolling 0 < 45 (where 0 is the slope angle)
2. Bouncing 45 < 0 < 75
3. Falling 0 > 75
However, slow-motion video of rockfall tests in Rifle, Colorado (1991) led Jones to
state that, Observations of bouncing rocks show that regardless of the initial
rotational velocity, rocks always leave the surface in the rolling mode (Jones,
1998). Therefore, the angular velocity component was included in the 4.0 version

1 Free falling block Velocity (m/seconds)
2 Block sliding down
45 plane; 0 = 30
3 Sphere rolling down
45 plane
4 Cylinder rolling down
45 plane
Comparison of velocities for blocks on 45 degree plane. (Reproduced with
permission, Richards, 1992)
Figure 3.5

3.3.7 Site Specific Topography
Due to the variation of slope geometry and geology at Summitville, rock
parameters and topography were of primary concern. The steep, benched slopes
have weathered to more uniform slope profiles, but still contain random zones of
differing rock properties ranging from hard and competent to soft and weak. The
motion applicable to this case will mostly involve a combination of free falling,
rolling, and sliding.
3.4 CRSP Validation
Development of CRSP took place between August of 1985 and May of 1989.
Experimental verification and calibration of CRSP was conducted in conjunction
with the testing of rock-fall fences at a site near Rifle, Colorado (Pfeiffer, 1989).
Videotapes recorded the motion of rocks traveling down a slope and impacting the
test fence. Research conducted at the Colorado School of Mines added graphical
data presentations to the program and analyzed the videotapes to verify and
calibrate the simulation program.
The CRSP algorithm was verified and calibrated from several empirical rockfall
tests. Video tapes of the full-scale test at Rifle, Colorado, were used to compare
actual rockfall velocities to those calculated by the CRSP program (Jones, 1998).
Prior to the incorporation of angular velocity, the CRSP program tended to
underestimate rockfall velocity. The experimental data from the West Rifle test site
allowed for adjustments to the friction function and scaling factors until the
simulation data fit the West Rifle values for velocity, number of bounces, and
bounce height. Since CRSP models the worst case situations, the data used to
calibrate with the Rifle data represented the fastest 50% of the rocks in simulation.

Also, CRSP data was compared to field experiments by the California Department
of Transportation (CALTRANS) (Jones, 1998). CRSP predictions tended toward a
worst case than the field results (Pfeiffer, 1989), but overall the results were

4. Geosyntheticallv Reinforced Soil (GRS) Rockfall Berm
The rockfall berm requires a separate investigation as to its geometry, construction,-
materials, and effectiveness in absorbing the kinetic energies involved in rockfall.
The geometry of the rockfall berm is of the primary importance in attenuating the
horizontal velocity component of the falling rock. Intuitively, the impact side of
the berm requires steep slopes to provide adequate reduction of lateral velocities
and attenuate energy through vertical motion produced by the rotational velocity.
This type of geometry can be accomplished with an earthen berm with structural
facing or by the use of a GRS wall. The latter provides a cost effective berm with a
steep impact side that can absorb a dynamic impact and be repaired with less
difficulty than a berm with a stiff structural facing. The following is a discussion
regarding the concept and effectiveness of the GRS wall as a rockfall barrier and
related issues concerning its long term performance.
4.1 GRS design
The design of GRS walls on competent foundations are based on the limit
equilibrium method of analysis that the entire soil mass is under similar stresses.
The GRS wall incorporates horizontal layers of geotextile reinforcements, also
known as tensile inclusions, within a granular soil backfill to resist lateral
movement of the reinforced soil mass. The soil interacts with the geotextile by
transferring a tensile force to the geotextile reinforcement through geosynthetic-soil
interactions. Studies have attempted to describe the GRS wall as a composite
material with increased shear strength characteristics. Yang (1972) suggested two
theories, as depicted on Figure 4-1: 1) an increase in the major principal stress at

failure is attributable to an increase in apparent cohesion under the same confining
stresses, and 2) the increase in major principal stress is due to an increase in
apparent confining pressure as the reinforcement provides an anisotropic restraint
to soil deformation in the direction of the reinforcement. Yang related the increases
to the Mohr-Columb failure criteria and from the geometry derived the following
equations. The apparent cohesion CR is calculated as:
For reinforcements having a tensile breaking resistance of RT, and a vertical
spacing between horizontal reinforcement layers Sv, and a backfill with an internal
friction angle (J) The latter theory of increased apparent confining pressure Ao3R is
calculated as:
The design of the GRS wall is therefore a function of the backfill and the tensile
strength and vertical spacing between the geosynthetic reinforcements. Further
discussion of the rockfall barrier design is included in the design and stability
analysis sections.
C R= -----------; where Kp = tan2(45+(j)/2)

Cn = Apparent Cohesion
4> = Internal Friction (reinforced)
s = Internal Friction (backfill)
= Confining Pressure
o, = Maximum Principal Stress
am = Max. Principal Stress (reinforced)
Apparent cohesion interpretation of strength increase due to reinforcements
(Yang, 1972)
Increased confinement concept of soil reinforcement (Ycng, 1972)
Figure 4.1

4.2 Effectiveness of GRS as Rockfall Barrier
The Colorado Department of Transportation (CDOT) conducted a series of full-
scale tests of a GRS rockfall barrier in Rifle, Colorado, in 1991 (Parsons De Leuw,
1991). The test site consisted of a slope with an overall grade of 1:1 (horizontal to
vertical) near the top and graded to nearly 3:1 (horizontal to vertical) near the toe of
the slope. The overall slope length was 610 feet and had a vertical drop of 310 feet.
The vertical GRS wall was 90 feet in length, 10 feet high, 6 feet wide, and was
constructed on a firm foundation of granular material. The geosynthetic
reinforcement used was Poly-Felt TS-700 (nonwoven, UV stabilized, spunbond,
continuous filament, needle-punched, polypropylene geotextile) with a 1 foot
vertical spacing and a nominal 6 x 6 timber facing was used for dimensional
control, as shown on Figure 4-2. The timber facing was constructed in a manner to
facilitate the construction of the vertical face and not to provide additional strength
to the rockfall berm. The backfill was a lightly compacted, well-graded roadbase.
Eighteen rocks were rolled down the same path from the top of the slope ranging in
size from 592 lbs (1.70 ft diameter) to 18,354 lbs (5.80 ft diameter). Velocity of
rock impact was measured by a video camera, calibrated at 30 ffames/second, and
painted distance markers for reference. Deformation of the rockfall berm was
measured on the front and back of the wall after each impact by simple
measurement from the original shape. The damage was not repaired after each test,
but the data were adjusted to provide net deformation of each rock impact.

Typical wall cross-section, Rifle, Colorado
Figure 4.2

The kinetic energy of each impact was calculated from the measured horizontal
velocity component and the mass of the rock. The back-wall deformation imparted
to the rockfall berm (measured at 0.4H) was correlated with the corresponding
kinetic energy to define an energy-deformation relationship that could be used for
design purposes, as shown on Figure 4-3.
Due to the difficulties associated with the prediction of rockfall design, the
empirical data from the West Rifle tests provide some validation in the design of a
GRS rockfall impact berm. It is important to note that Figure 4-3 shows a break in
the parabolic curve at 8 inches deformation (at 0.4H) and the energy-deformation
relationship becomes positive linear until discontinued at 12 inches deformation.
None of the rocks modeled breached the wall and the largest rocks were rolled last
into an already damaged wall. Therefore, the curves are likely very conservative for
estimating the breaching of the wall. Deformation along the foundation was limited
or nil, indicating that the energy of the impact was absorbed effectively without
translating into shear stress along the base that would have caused the wall to slide.
4.3 Long-term Performance of a Sloped GRS Wall
The use of geosynthetics to maintain over-steepened earth berm geometries
requires a review of long-term performance data of a GRS wall. An extensive
literature review of the long-term performance of seven GRS walls that are
currently being monitored for reinforcement strain and overall performance was
conducted in a thesis by Crouse (1996). The walls range from 15-feet to over 40-
feet in height and include reinforcements including polypropylene and polyester
geogrids and geotextiles. The GRS walls are exposed to freezing climates in
Canada and the extreme heat in the desert of Arizona. Crouse concluded that all of

the walls reviewed were performing exceptionally well. The locations of these
projects, and the principal researcher associated with the monitoring of the walls
are listed on Table 4-1.
Table 4-1: GRS Walls Monitored for Long-Term Performance
Project Date Constructed Monitoring Duration Location Principal Researcher
1-70 through Glenwood Canyon 1982 7 months Glenwood Springs, Colorado R. Barret
Tanque Verde- Wrightstown-Pantano Roads 1985 7 years Tucson, Arizona J. Collin
NGI 1987 4 years Oslo, Norway R. Fannin
Japan Railway Test Embankment 1987 2 years Tokyo, Japan F. Tatsuoka
Highbury Avenue 1989 2 years London, Ontario, Canada R. Bathhurst
Federal Highway Administration 1989 1.3 years Algonquin, Illinois M. Simac
Seattle Preload Fill 1989 1 year Seattle, Washington T. Allen
A detailed review of the typical long-term performance of a full-scale GRS wall
was conducted by Fannin and Hermann (1990). The 4.8 m high wall was
constructed with a 0.5:1 (horizontal to vertical) unreinforced face on a competent
gravelly sand foundation. The backfill soil was a medium fine sand, well graded,
constructed in 0.3m lifts and compacted to 92% maximum dry density (Standard

Ccrcnnaccn cf
Bacx raca
Rock barrier wall energy-deformation curves
Figure 4.3

Proctor). The primary and intermediate geosynthetic reinforcements were Tensar
SR55 (uniaxially oriented grid) and Tensar SSI (biaxially oriented grid)
respectively. The sloped wall was constructed in two sections; 1) the primary
reinforcements extended 3m into the backfill at 1.2m vertical spacings with
intermediate reinforcements every 0.6m, and 2) uniform reinforcements extending
2.2m into the backfill at 0.6m vertical spacings. Both sections in the wall were
monitored for a 20-month period while subjected to a period of self-weight loading,
followed by load-unload cycles of surcharge pressure using water tanks, followed
by application of permanent surcharge. Performance data were obtained for the
reinforcements using force cells, strain cells, and thermistors for the 20-month
period. The general conclusions are as follows:
horizontal displacement exceeded vertical, with outer movement in upper
parts of the wall with no significant difference between the two sections,
strain decreased from the largest values at the face of the wall to zero at the
embedded end,
the magnitude of strain was typically small with maximum values of 0.8%
for self-weight loading and 1.3% for the permanent surcharge, and
the soil strain exceeded the geogrid strain only slightly, indicating a good
interlock between the geogrid and surrounding soil.
The small strain recordings and negligible differences between the two sections of
the wall indicate that the length of reinforcement (L) to height of wall (H) ratios of
the two sections, 0.46 and 0.63, were conservative. The typical L/H ratio for design
purposes is 0.7.

4.4 Degradation Properties
Determining the performance life of a GRS wall involves investigating several site
condition issues, including; the resistance of the geosynthetic to creep, corrosion,
and sunlight exposure. The Federal Highway Administration GRS wall
(Algonquin, Illinois), referred to as Wall No. 9, was built to evaluate the
performance of continuous filament polyester geogrid reinforcements (Crouse,
1996). Wall No. 9 was constructed at a very low factor of safety value in order to
evaluate the existing design methods. After 1.3 years of monitoring, the wall
remained in excellent condition.
4.4.1 Creep
Creep of geosynthetics is a function of the stress level and polymer type.
Polypropylene and polyethylene generally exhibit larger creep deformation than
polyester and polyamide, as shown on Figure 4.4. The permissible loads for
polypropylene and polyethylene are about 20% to 25% of tensile strength, and
about 40% to 50% for polyester and polyamide (Wu, 1997). However, if the
backfill is resistant to creep, the geotextile will not be affected. When granular
backfill is employed, the problem of creep is reduced (Wu, 1997).
4.4.2 Hydrolysis
The low pH values that can be expected from the soils at Summitville can be
extremely harsh on polyamides. If polyesters are used for long-term applications, it
is important that the resin have a high molecular weight and a low carboxyl end
group (CEG) concentration. Most polyesters show little change in strength in low
pH conditions (Koemer, 1996).

4.4.3 Sunlight (UV)
Degradation of the geotextile due to exposure to sunlight will have the greatest
affects on the reduction of geotextile strength. However, the design, as discussed
in Part II, Chapter 3, will not have exposed geotextile and, if covered prior to
installation and put in place in a timely manner, the exposure to sunlight will not
pose a problem.

Strain (%)-- Strain (%)
Log time (s)
(a) Creep at 20% load
Log time (s)
(b) Creep at 60% load
Strain log time curves for various geotextiles. PE = polyethylene, PP =
polypropylene, PET = polyester, PA = polyamide
Figure 4.4

5. CRSP Analysis
The Colorado Rockfall Simulation Program (CRSP), version 4.0, was used to
analyze the kinetic energy, horizontal velocity, bounce height, and the maximum
distance of rockfall from the crest of the USBR berm with and without the
proposed GRS wall. The program models two-dimensional rockfall motion and the
results are highly dependent on the path geometry, the coefficients of normal
restitution and tangential frictional resistance (R and R, respectively), and surface
roughness (S). The slope geometry for each path chosen was broken into cells
based on uniform slope geometry and material properties, therefore, the accuracy of
the model is dependent on the number of cells required to accurately describe the
path. Each path will be modeled with the proposed USBR earth berm with impact
face of 3:1 (horizontal to vertical) and then with the modified geometry of the GRS
rockfall berm with impact faces of 0.5:1 (horizontal to vertical) and 0.25:1
(horizontal to vertical). The following is a discussion of the paths, determination of
the GRS rockfall berm geometry, and CRSP coefficients chosen for the site.
5.1 Rockfall Paths
The highwall at Summitville poses a number of considerations for determining the
rockfall paths chosen to represent the typical rockfall motion that could be
expected. As previously mentioned, the alteration zones and mined benched
geometry of the Summitville highwall produces a rock face that is comprised of
rocks varying from hard and competent to soft and easily weathered. The range of
surfaces that can be expected for a rockfall to encounter, will consist of random
competent zones of dike rock that create launching zones and soft surfaces were
rolling and sliding will occur. Rockfall motion is expected to sustain large

decreases in energy in the normal direction to the slope when in contact with the
argillic material and have minimal decreases when in contact with the more
competent vuggy silica and quartz-alunite. Therefore, the rockfall paths of most
interest are those that generate the greatest velocities and the greatest impact
energies at the proposed earth berm. The rockfall in all cross sections are presumed
to initiate near the summit of the highwall with a nominal initial velocity.
Also considered is the proximity of the proposed USBR earth berm. The rockfall
path of greatest interest intersects the proposed earth berm at a point closest to the
toe of the highwall. This path should produce the greatest number of rocks
breaching the crest and launching onto the compacted mine waste caps. CRSP was
used to model rockfall along the paths specified in Table 5-1 without the proposed
GRS rockfall berm. The cross-sections used are shown on Figures 5-1 through 5-4.
Table 5-1: Summary of Cross-Sections Investigated
Rockfall Path Overall Slope Grade Total Vertical Drop
Path with the greatest vertical drop and steepest overall grade. Figure 5-3 100% 695
Path with the greatest amount of contact with vuggy silica and quartz-alunite. Figure 5-1 90% 400
Path that intersects the USBR earth berm at the closest point to the toe of the highwall Figure 5-2 84% 645


12100 N 12100
12000 's \ \ \ 12000
11900 \ V 11900
11800 V \ s N' BOR EARTH BERM 11600
1+00 2+00 J+00 ++00 5+00 6+00 7+00 6+00 9+00
0 100

0 100 200
SCALE; r=200*

\ \ t
-j '''4 'x

\ V\ \ \
X S S \ > 's
\ \ \ \ X r EARTH BERM
4 N
1* t

1*00 2400 3400 4400 5400 0 400 7400 0400 8400 10400 11400 12400
0 100 200
SCALE: 1" =200

5.2 Determination of the Coefficients
The parameters that are readily measured, or estimated in the field include; surface
roughness (S), slope geometry, rock size and shape. Figure 5-4 shows the geology
of the highwall face and the locations of the three cross-sections chosen for
analysis. The locations of vuggy silica and quartz-alunite, as well as slope changes
were marked along the profile to be input as separate cells in the CRSP model. The
coefficients (rt and rn) were then calibrated, as discussed below, to model the
expected rolling and sliding rockfall motion along the argillic material and
projected free fall and bouncing on the competent materials.
Surface roughness (S), within each cell used in CRSP, is defined as the amplitude
of surface variation, perpendicular to the overall slope, within a distance equal to
the radius of the rock being modeled (R). The assigned slope roughness
coefficients were the highest for the rough profile of the competent vuggy silica
and quartz-alunite and lower for the smoother, weathered profile of the argillic
material. The lowest roughness coefficients were assigned to the compacted fill
material of the remedial work at the toe of the highwall. As previously discussed,
the roughness coefficient of the slope is used to determine the random variation of
impact angles within each cell of CRSP.
5.2.1 Parametric Analysis: Coefficient of Normal Restitution
The coefficients of normal restitution and tangential frictional resistance (R^ and R,
respectively) are dependent on empirical data and subjective comparisons with
published calibrations. The choice for determining the least sensitive of these two

Quartz-alunite rock and corr
rock, a; or vuggy quartz roc
Undivided rocks of the illite
and montmorillonite-chlorite,
Dark andesitic porphyritic fla
sparse pyroclastic beds
1 Sun
\ \ !\ \ \ v \ -FR^M''eOR DRAWING 155
\ \ v /\ \ \ . 2. /Gj?IQ-COORDINATES SHOV
.s \ \ \, \ \ \ *i \ \ N \ \ \ /ANACONDA LOCAL COORC
/ ' ' -REE^ENCE NC

LIPMAnV P.W., 1974, \GEC
GEOLCGr /cm. 89V 1994

coefficients, R,, was to obtain a value that was within the range suggested by the
CRSP program on Table 3-1. However, it was essential to complete a parametric
analysis to determine values of R,, for the argillic material and the vuggy
silica/quartz alunite that would obtain the expected rockfall motion for these
materials on the slopes at the Summitville highwall.
In order to determine a R^ value for the soft argillic material, the cross-section A-
A was used as a representative slope (see Figure 5-1). The slope properties were
first modeled as completely argillic, with the expected rockfall motion being rolling
and sliding. The coefficient of normal restitution was entered as a higher value and
progressively was lowered at increments of 0.05 until the motion changed from
bouncing to rolling and sliding, as depicted on Figures 5-5 through 5-7. The
rockfall motion changed from predominantly bouncing, with R^ = 0.5, to rolling
and sliding with an R = 0.3, as shown on Figure 5-7.
The R value for the vuggy silica/quartz alunite was modeled on the same A-A
cross-section, the expected rockfall motion being bouncing. The R^ was
incrementally increased as the rockfall motion changed from rolling and sliding to
predominantly bouncing, as shown on Figures 5-8 through 5-11. The motion
changed rapidly with the increasing value of R^ to a bouncing motion near the toe
of the slope where the rock experienced higher velocities. Figure 5-11 depicts the
rockfall motion as bouncing on the upper slope, indicating a relatively hard,
competent rock mass. Table 5-2 includes the values used in the CRSP model.

Parametric Analysis for Argillic Rock: R = 0.5
Figure 5-5

Predominant Motion is Bouncing
Parametric Analysis for Argillic Rock: Rn = 0.4

Parametric Analysis for Argillic Rock: R = 0.3

| AP3
II I I III i llT'ITn IT! TTI IT f if I I I ill I I I II I I I I I I'll
160 240 320 400 480 560 640 720 800 880 960

Parametric Analysis for Vuggy Silica Rock: Rn = 0.7

Parametric Analysis for Vuggy Silica Rock: Rn = 0.8

Parametric Analysis for Vuggy Silica Rock: R = 0.9

5.2.2 Values used in CRSP
The values obtained for the coefficient of normal restitution from the parametric
analysis, the estimated surface roughness, and the coefficient of tangential frictional
resistance are presented in table 5-2.
Table 5-2: Values Used in CRSP Model
Material U h Surface Roughness
Vuggy Silica/Quartz-Alunite 0.9 0.9 1.20
Argillically Altered Rocks 0.3 0.8 0.50
Compacted Fill 0.5 0.8 0.25
GRS Rockfall Berm 0.5 0.8 0.25
5.3 USBR Earth Berm Geometry
The proposed earth berm to be constructed during the Summer of 2000 at the base
of the highwall under USBR OU-4 remedial activities will extend approximately
1600 feet laterally along the base of the highwall. The toe of the berm will parallel
the toe of the highwall with an offset of 35 feet to approximately 100 feet, as shown
on Figure 5-4. A 3:1 (horizontal to vertical) impact face of the berm will extend
from the toe of the berm to the uniform crest at an elevation of approximately
11,775 feet. The crest will maintain a flat 12 foot wide service access road before
sloping to the northeast at a 20:1 (horizontal to vertical) grade, as shown on Figure

0 5 1
1 _____L_
= 10
0 2.5 5
'SCALE' I"-**' FIGURE 5-12

5.3.1 Proposed GRS Geometry
The GRS wall was modeled with the same crest elevation and overall height as the
USBR earth berm, as shown on Figure 5-12. The impact face of the GRS wall was
first modeled with a 0.5:1 (horizontal to vertical) impact face and again with a
0.25:1 (horizontal to vertical) in order to assess the effect of geometry on attenuating
the number of rocks passing the crest. Placing a partial GRS rockfall berm on the
slope of the USBR 3:1 (horizontal to vertical) impact face or at the crest was not
modeled. A geometry that includes any portion of a 3:1 (horizontal to vertical)
impact face would increase the magnitude of the vertical velocity component which
the GRS wall is not designed to resist.
5.4 GRS Rockfall Berm Analysis
The three selected rockfall paths were analyzed at three analysis points (A.P.) on the
impact face of the USBR earth berm; 1) at the toe, 2) horizontal mid-point, and 3)
the crest. These points were analyzed in order to determine the loss of energy
experienced by the rockfall, and the number of rocks that passed the crest in order to
make a comparison with the revised geometry of the GRS rockfall berm.
The three rockfall paths were run again with the new geometry of the GRS rockfall
berm. Two analysis points were considered for the GRS berm due to the steepness
of the impact face; 1) at the toe, and 2) the crest.
Each trial consisted of 100 rocks released from a starting zone (source zone) with a
nominal velocity of 1 ft/s to 10 ft/s. The rocks were modeled with uniform densities
of 150 pcf, based on an upper estimate for the density of altered latite. The rocks
were modeled as spherical with the following diameters; 1) the average size of 1.25

foot diameter, 2) 2.25 foot diameter, and 3) the maximum observed diameter of 3.5

6. CRSP Results and Discussion of Results
CRSP modeled 100 rocks on each trial, with results measured at analysis points on
the proposed earth berm. For the proposed USBR earth berm, measurements were
taken at the toe, at horizontal mid-point, and the crest. The revised geometry that
included the 0.5:1 and 0.25:1 (horizontal to vertical) impact face GRS rockfall
berm, only two analysis points were used, the toe and the crest.
Rock diameter had a great influence on the number of rocks passing the crest. For
the 1.25 foot diameter rocks, essentially none passed the crest of the proposed
USBR earth berm. However, this changes dramatically with increasing rock
diameter as nearly all of the 3.5 foot diameter rocks passed the crest of the USBR
earth berm. The steeper walls of the GRS rockfall berm converted the horizontal
velocity component of the rockfall event into a vertical component, as can be seen
from the increasing bounce height at the crest and the decreasing distance traveled
as the impact face became steeper in Tables 6-1 through 6-3.
The results for all trials run in the CRSP analysis are summarized on Table 6-4.
6.1 Cross-Section A-A
Cross-section A-A represented the typical path that would contact the most vuggy
silica and quartz-alunite, the most competent rocks at the site, and would lose less
energy due to less contact with the ground. The rockfall motion shows a
combination of rolling and sliding along the argillic rocks and projected free fall
when in contact with the competent vuggy silica and quartz-alunite, as shown on

Figure 6-1. The path was modeled from potential rockfall source zones and across
the USBR earth berm. The geometry was then steepened to model the GRS rockfall
berm in subsequent trials. Table 6-1 below summarizes the results from the crest of
each proposed berm, with a 3.5 foot diameter boulder.
Table 6-1: Cross Section A-A CRSP Results
Cross- Section A- A Max. Velocity at Crest (ft/s) Max. Kinetic Energy at Crest (ft-lb) Max. Bounce Height at Crest (ft) Number of Rocks Passing Crest Max. Distance Traveled Past Crest (ft)
1) USBR 51.91 211890 3.83 97 53 >350
1) 0.5:1 GRS 32.05 76886 7.80 61 105
1)0.25:1GRS 5.33 17434 10.32 1 10
The maximum velocity refers to the horizontal velocity. The affect of a steepened
geometry can be readily analyzed as the conversion of lateral energies into a
vertical motion caused by the rotational velocity. As the slope steepens, the
velocity, number of rocks passing the crest and maximum distance beyond the
berm crest decreases while the bounce height steadily increases. In the case of
cross-section A-A, only 1% of the rocks modeled in the worst case scenario passed
the crest.

Figure 6.1

6.2 Cross-Section B-B
Cross-section B-B represented the path that would intersect the USBR earth berm
closest to the toe of the highwall, and was expected to have the greatest number of
rocks passing the crest. The rockfall motion is mostly rolling and sliding along the
argillic rocks with a small zone of projected free fall when in contact with the
competent vuggy silica and quartz-alunite near the toe of the highwall, as shown on
Figure 6-2. The path was modeled from potential rockfall source zones and across
the USBR earth berm. The geometry was then steepened to model the GRS rockfall
berm in subsequent trials. Table 6-2 below summarizes the results from the crest of
each proposed berm, with a 3.5 foot diameter boulder.
Table 6-2: Cross Section B-B CRSP Results
Cross- Section B-B Max. Velocity at Crest (ft/s) Max. Kinetic Energy at Crest (ft-lb) Max. Bounce Height at Crest (ft) Number of Rocks Passing Crest Max. Distance Traveled Past Crest (ft)
2) USBR 57.23 260535 4.36 100 42 > 285
2) 0.5:1 GRS 107.86 737349 6.56 25 70 ft
2)0.25:1 GRS 100.10 637111 13.25 3 55 ft
Again the general conclusions apply as to those made for cross-section A-A. The
anomalous results for the increased velocity and kinetic energy at the crest for the
steepened geometry can be accounted for in the randomness of the impact angle
generated by CRSP. The average velocity was greatly reduced from the 3:1
(horizontal to vertical) geometry (34.08 ft/s) to the 0.5:1 (horizontal to vertical)
geometry (13.93 ft/s). However, the 0.25:1 (horizontal to vertical) geometry
produced an anomalous result for average velocity of 40.12 ft/s. This is possibly

explained by the randomness of the slope variation within CRSP. In cross-section
B-B\ 3% of the rocks modeled in the worst case scenario passed the crest, as
compared to 100% passing the 3:1 (horizontal to vertical) berm.

Figure 6.2

6.3 Cross-Section C-C
Cross-section C-C represented the path that had the greatest overall grade, and
would allow the gravitational acceleration to generate the greatest horizontal
velocities into the impact slope. The rockfall motion is a combination of rolling and
sliding with several zones of significant launch to free fall motion, as shown on
Figure 6-3. The path was modeled from potential rockfall source zones and across
the USBR earth berm. The geometry was then steepened to model the GRS rockfall
berm in subsequent trials. Table 6-3 below summarizes the results from the crest of
each proposed berm, with a 3.5 foot diameter boulder.
Table 6-3: Cross Section C-C CRSP Results
Cross- Section C-C Max. Velocity at Crest (ft/s) Max. Kinetic Energy at Crest (ft-lb) Max. Bounce Height at Crest (ft) Number of Rocks Passing Crest Max. Distance Traveled Past Crest (ft)
3) USBR 54.06 233207 4.53 100 44 >300
3) 0.5:1 GRS 35.03 37396 13.44 31 140
3)0.25:1GRS 31.09 80485 11.13 13 95
Conclusions for cross-section C-C are generally the same as for cross-section A-
A. This is more obvious when comparing the average values obtained for the test
included on Table 6-4. The anomalous values for cross-section C-C showed
slightly higher average values for velocity and kinetic energy of the crest of the
0.25:1 (horizontal to vertical) compared to the 0.5:1 (horizontal to vertical)
The output files for all of the CRSP runs are provided in Appendix B.

Figure 6.3

rnename M.r. i M.r. £ M.r. o M.r. i M.r. ^ M.r. o M.r. i M.r. ^ M.r. o w.
150 1.25 crspla 6.81/3.31 0.93/ 0.39 3.35/ 3.35 43.97/ 28.94 35.51/20.6 27.22/ 27.22 6873/ 3457 4197/1934 2951/2951 0
150 2.25 crsplab 8.69/1.65 3.97/1.85 3.72/1.31 112.81/45.95 51.54/34.36 49.12/28.42 210494/ 44228 53752/ 25840 50342/ 19292 45 30 12
150 3.50 crsplac 2.73/1 09 3 15/1.70 3 83/1.22 69 92/ 52.23 55.36/42.78 51.91/38.00 360218/ 205869 231315/ 144647 211890/ 118266 97 11 33
150 1.25 crspHa 1.62/1.06 NA NA 40.22/ 25.58 NA NA 5388/2676 NA NA 0
150 2.25 crspUb 3.78/1.46 7,26/ 3.88 NA 59.86/45.73 24.94/ 8.78 NA 70230/42255 12574/5686 NA 13 13 0
150 3.50 crspHc 3.18/1.11 7.8/ 3.49 NA 68.12/52.42 32.05/14.08 NA 341121/ 207158 76886/31485 NA 61 61 0
150 1.25 crsp111a 1.62/1.06 NA NA 40.22/25.58 NA NA 5388/2676 NA NA 0
150 2.25 crsp111b 4.71/1.59 NA NA 63.79/46.43 NA NA 80037/44227 NA NA 0
150 3.50 crsp111c 3.49/1.14 10.32/10.32 NA 64.44/51.92 5.33/ 5.33 NA 306037/ 203078 17434/ 17434 NA 1
150 1.25 crsp2a 2.87/1.08 1.59/0.85 NA 40.96/ 30.60 20.75/ 15.91 NA 5668/ 3546 1770/ 1238 NA 0
150 2.25 crsp2ab 39.59/ 2.22 17.32/ 2.48 5.17/1.48 98.79/ 48.72 96.38/ 33.59 48.55/ 24.75 164886/ 48911 157674/ 26797 49997/ 16079 84 64 8
150 3.50 crsp2ac 31.82/1.72 11,55/2.46 4 36/1.32 113.31/ 58 52 108.75/42.25 57.23/ 34 08 821850/ 260317 771578/ 147993 260535/ 100793 100 25 33
150 1.25 crsp22a 3.54/1.52 NA NA 41.43/30.99 NA NA 5749/ 3422 NA NA 0
150 2.25 crsp22b 35.39/1.96 12.74/5.49 NA 88.27/ 46.22 91.02/29.89 NA 130868/ 43778 137760/ 39752 NA 4 4
150 3.50 crsp22c 21.65/1.78 6.56/3.19 NA 110.06/59.49 107.86/13.93 NA 774614/ 267751 737349/ 57996 NA 25 25
150 1.25 crsp222a 2.08/0.67 NA NA 46.86/36.15 NA NA 7362/4604 NA NA 0
150 2.25 crsp222b 28.58/2.18 NA NA 103.77/47,41 NA NA NA NA NA 1
150 3.50 crsp222c 29.97/1.58 13.25/11.4 NA 118.79/56.71 100.10/40.12 NA 857515/ 242520 637111/ 230451 NA 3 3
150 1.25 crsp3a 10.23/2.20 5.10/2.05 6.14/2.64 56.31/32.04 56.52/ 23.38 30.57/14.61 10907/4116 11190/2805 3677/ 1343 9 9
150 2.25 crsp3ab 5.57/1.65 4.99/2.63 5.46/1.51 80.10/48.96 50 22/ 25.78 44.79/ 20.87 125725/ 50044 52246/ 17494 45028/ 13479 74 61 13
150 3.50 crsp3ac 4.32/1.46 3.77/2.32 4.53/1.68 104.52/63.96 59.59/ 38.98 54 06/ 32.06 776186/ 310230 266317/. 126230 233207/ 97565 100 24 32 within 300
150 1.25 crsp33a 7.10/1.50 4.93/ 4.93 NA 77.86/31.77 12.42/12.42 NA 20322/ 4350 1625/1625 NA 0
150 2.25 crsp33b 6.21/1.57 8.68/ 3.76 NA 94.49/ 50.36 24.42/13.31 NA 156894/ 52627 13921/6276 NA 7 7
150 3.50 crsp33c 3.95/1.18 13.44/5.12 NA 100.58/61.83 35.03/16.16 NA 744740/ 291039 101585/ 37396 NA 31 31
150 1.25 crsp33a 7.27/2.08 NA NA 66.98/ 36.06 NA NA 14257/5010 NA NA 0
150 2.25 crsp33b 5.65/1.56 6.58/6.58 NA 75.75/ 50.09 24.39/ 24.39 NA 112501/ 51990 14454/14454 NA 1
150 3.50 crsp33c 4.27/1.13 11.13/ 7.31 NA 93.39/62.17 31.09/18.43 NA 646213/ 292052 80485/42360 NA 13 13
ing points: 1) that the GRS rockfall berm only has two analysis points (APs), and 2) more than one rock passing an AP was required to generate data.

7. Rockfall Berm Design
7.1 GRS Rockfall Berm Geometry
The energy-deformation curves, as shown on Figure 4-3, indicate that the GRS wall
should be approximately 7 feet thick to maintain deformation measurements under
12 inches for the largest anticipated kinetic energies. The excavation of the GRS
rockfall berm into the toe of the USBR earth berm (crest and 20:1 backfill), as
shown on Figure 5-12, has three benefits; 1) this essentially creates a wall that is
several tens of feet thick, without adding to the cost of the wall, 2) the rock will
impact on a 0.25:1 (horizontal to vertical) face, and 3) the cut and fill is
approximately balanced. Also, the GRS rockfall berm would extend the length of
the proposed USBR crest elevation.
7.2 Foundation Conditions
The foundation of the GRS rockfall berm will be excavated into the compacted fill
of the proposed USBR earth berm and existing mine cap. The backfill of the GRS
rockfall berm will be of the same materials as the USBR earthberm (see Table 7.1)
and compacted to approximately uniform densities. Specifications of the
earthberm material are not yet available. However, the mine pit caps are identified
as soils with a USCS designation of SM-SC. These soils could be susceptible to
creep and settlement. Any settlement occurring during the construction of the GRS
rockfall berm can be compensated for with more backfill material. The creep
potential and estimates of shear strength characteristics are discussed in further
detail in the sections below.

7.3 Stability Analysis
The overall stability analysis for the GRS rockfall berm included determining
internal stability and external stability. The following are the estimated soil
parameters used in the analysis:
Table 7-1: Estimated Material Properties
Material Moist Unit Density (PCF) Saturated Unit Density (PCF) Effective Cohesion (PSF) Effective Friction Angle (degrees)
Compacted Foundation 122 128 250 32
GRS Reinforced Zone 122 128 4195 (apparent) 34
Compacted Backfill 122 128 0 34
7.3.1 GRS Internal Stability'
The stability of the GRS rockfall berm, as excavated into the USBR earth berm was
calculated using the AASHTO Ultimate Strength Method (Wu, 1997). This
method assumes that the earth pressure distrbution is that of the Rankine active
condition and requires that two tensile loads be determined. The first is the limit
state tensile load (T,jmjt) which is the highest load level at which the log time creep-
strain rate continues to decrease with time within the design life time without

inducing brittle or ductile failure. The second is the serviceability state tensile load
(^service) which is defined as the load level at which total strain will not exceed 5%
within design lifetime. The design lifetime for the GRS rockfall berm at
Summitville Mine is 75 years. Aging, chemical degradation, and active earth
pressures induced by the backfill are included in the factor of safety calculations.
The guidelines for the stability analysis are provided with hand-calculations (see
Appendix A). Table 7-2 is a spreadsheet calculation, prepared by the author of this
thesis, of the AASHTO Ultimate Strength factors of safety and design summary.
The factor of safety value against rupture failure of the geotextile was calculated
using the United States Forest Service Method, as described in equation 7.1.
Fsrupture = ; oh = tan2(45+/2)yH (7.1)
The rupture strength is a function of the ultimate tensile strength of the geotextile
(T^t) (see Appendix C), the vertical spacing of reinforcements (s), the lateral earth
pressure (oh), and the height of the wall (H). The factor of safety against pullout is
5.68 (Table 7-2) and against rupture is 6.01.
7.3.2 External Stability
An overall slope stability analysis of the proposed GRS rockfall berm was
performed using the computer program XSTABLE Version 5 (Interactive Software
Designs, Inc., 1997). The soil properties were estimated from gradation analyses
performed on the mine caps. The cohesion for the GRS rockfall berm is based on
the apparent cohesion concept by Yang, equation 1.9, Part I- Chapter 5.