Geotechnical_Final_Report_No._02-17.pdf

Geotechnical Report No. 02‐17

Many Glacier Road

Milepost 6.2 Landslide

Final Geotechnical Report

MT PRA GLAC 14(3)

Rehabilitate Many Glacier Road Slide

Glacier National Park

Montana March 2017

GR02-17 Page i

Table of Contents

1. INTRODUCTION 1

1.1 PROJECT LOCATION 1

1.2 SCOPE OF PROJECT 1

1.3 PROJECT BACKGROUND 1

1.3.1. LAKE SHERBURNE DAM CONSTRUCTION AND LANDSLIDING 2

1.3.2. MANY GLACIER ROAD CONSTRUCTION AND LANDSLIDING 3

2. GEOLOGY AND SITE CONDITIONS 4

2.1 REGIONAL GEOLOGY 4

2.2 LOCAL GEOLOGY 5

2.3 SITE CONDITIONS 6

2.4 CLIMATE 7

2.5 LAKE SHERBURNE 8

3. SUBSURFACE INVESTIGATIONS 10

3.1 1956 DRILLING AND HORIZONTAL DRAIN PROGRAM 10

3.2 1994 SUBSURFACE INVESTIGATION 11

3.3 2011 SUBSURFACE INVESTIGATION 12

3.4 2015 SUBSURFACE INVESTIGATION 12

3.4.1. SUBSURFACE PROFILE 13

3.4.2. GROUNDWATER OBSERVATIONS 14

3.4.3. INSTRUMENTATION 14

3.4.4. LABORATORY TESTING 15

4. ANALYSIS AND RECOMMENDATIONS 17

4.1 DISCUSSION 17

4.2 ANALYSIS 18

4.3 RECOMMENDATIONS 19

4.4 CONSTRUCTION CONSIDERATIONS 20

4.5 LIMITATIONS OF THIS REPORT 21

5. REFERENCES 23

List of Figures

Figure 1. Project Location Maps. ................................................................................................ iii Figure 2. Aerial Photograph of Many Glacier Valley. .................................................................. iv Figure 3. Aerial Photo of MP 6.2 Landsliding with Site Observations Figure 4. Lake Sherburne Reservoir Forebay Elevations, 2006-2016 Figure 5. Cross-section of interpreted slide surfaces from Schwegler (1958)

List of Tables

Table 1. Climate Summary from WRCC Station 245361, Many Glacier Table 2. Climate Summary from WRCC Station 247292, St. Mary Table 3. 2015 Boring Locations, Drilling Depths, and VWP Installations Table 4. Summary of Torsional Ring Shear Test Results

GR02-17 Page ii

Appendices

APPENDIX A – SITE PHOTOGRAPHS

APPENDIX B – BORING LOCATION MAP, BORING LOGS, CORE PHOTOS

Boring Location Map, page B-1 Unified Soil Classification System and Descriptive Terminology, page B-2 to B-4 2015 Boring Logs, page B-5 to B-23 2015 Core Photographs, page B-24 to B-28 2011 Boring Logs, page B-29 to B-30 1994 Boring Log, page B-31 to B-32

APPENDIX C – 2015 INSTRUMENTATION INSTALLATION RECORDS

APPENDIX D - LABORATORY INDEX TEST REPORTS

2015 Investigation Laboratory Test Reports, page D-1 to D-110 2011 Investigation Laboratory Test Reports, page D-111 to D-126

APPENDIX E - LABORATORY RING SHEAR TEST REPORTS

APPENDIX F – INSTRUMENTATION PLOTS

2015 Investigation Instrumentation Plots, page F-1 to F-15 2011 Investigation Instrumentation Plots, page F-16 to F-18

GR02-17 Page iii

Figure 1. Project Location Maps.

Milepost 6.2 landslides

GR02-17 Page iv

Figure 2. Aerial Photograph of Many Glacier Valley.

Source: 48.8269° N, 113.5523° W, Google Earth, July 26, 2014, accessed February 7, 2017.

GR02-17 Page 1

1. INTRODUCTION

In 2013 Western Federal Lands Highway Division (WFLHD) identified six active landslides (outlined in green in Figure 2) causing damage to Many Glacier Road within the Park boundary. In 2015 the National Park Service made the decision to complete landslide mitigation at one of the landslide locations. This report presents the results of field investigations and modeling of the landslides at MP 6.2 on Many Glacier Road. Based on site observations along the road corridor, this landslide is the fastest moving and therefore is also causing the most roadway damage relative to the other slides along the road within the Park boundary. Recommendations for using horizontal drains to locally lower water pressures and slow the landslide movement are presented.

The intent is that this project will serve as a trial to test the effectiveness of using horizontal drains to mitigate the landslides along Many Glacier Road.

Landslides affect the road outside the Park boundary also; however, they are outside the scope of our evaluations at this time.

1.1 PROJECT LOCATION

This project is located at milepost (MP) 6.2 on Many Glacier Road. MP 0 is located at the Many Glacier Road intersection with US Hwy 89 in Babb, Montana. MP 6.2 is also known as Cassidy Curve. The project location is shown on the maps in Figure 1 and Figure 2 on page iii at the beginning of this report. The GPS coordinates for the project site are latitude 48.8269° North, 113.5523° West (WGS84).

1.2 SCOPE OF PROJECT

The scope of this project is to design and construct horizontal drains. Our work includes conducting a subsurface investigation, installing instrumentation, monitoring instrumentation, geotechnical modeling, development of design plans, specifications and cost estimate (PSE) package for construction, and construction administration.

The construction will include constructing access roads and drill pads, installing horizontal drains, erosion control, riprap revetment, grading, and construction of a maintenance access road.

WFLHD will continue to monitor the geotechnical instrumentation before, during, and after construction in order to assess the effectiveness of the horizontal drains on slowing landslide movement.

1.3 PROJECT BACKGROUND

Many Glacier Road has a long history of roadway distress and closure as a result of recurrent landslide activity. According to historic files provided by GNP, some landslide areas have been active since the original road construction in 1927. In addition, the U.S. Bureau of Reclamation

GR02-17 Page 2

(USBR), the maintaining agency for the Lake Sherburne Dam, provided several historic geotechnical reports. These reports indicate that landslide activity has been ongoing in the hillsides above the Lake Sherburne Dam and was first documented when ground movement caused damage to the spillway during the original construction in 1917.

1.3.1. LAKE SHERBURNE DAM CONSTRUCTION AND LANDSLIDING

The MP 6.2 project site is located approximately 1 ¼ miles west of Lake Sherburne Dam and the history of the dam provides insight to the history of landsliding in the Many Glacier Valley. The information below is a summary of the construction history of the Lake Sherburne Dam obtained from Geotechnical Reports provided to us by the US Bureau of Reclamation (USBR). The USBR has been the owner and operating agency of the Lake Sherburne Dam since the original construction. A list of the reports provided by the USBR is contained in Section 5.

Lake Sherburne Dam is a compacted earthfill structure built between 1914 and 1921 to alleviate downstream flooding and divert water for irrigation along the Milk River Valley.

An attempt to construct a spillway in 1917 was halted when movement and cracking were observed in the north hillside above the dam and upward movement and bulging were observed in the spillway during its construction. Movement was attributed to removal of material (~150,000 cu yds) along the spillway channel, the saturated condition of the hillside, and the pre-existence of some deep planes of movement. Mitigation to address the landslide in the north hillside above the dam was completed in 1920. The partially completed spillway was repaired and a flexible timber flume structure was constructed to connect the upper and lower concrete structures. The brush on the hillside above the dam was removed to promote spring snowmelt. A surface drainage system was installed in the north hillside. Lake Bessie and other small sag ponds within the landslide mass above the dam were drained.

The landslide continued to move and eventually the spillway became damaged beyond repair. In 1960 the spillway was diked off and the original outlet works was modified into a glory hole type spillway. In 1973 a revised design flood required modifications to the dam to prevent overtopping during the design flood event. Alternate spillway locations proved to be just as unstable which led to the raising of the dam in 1983 to elevation 4816.5. The north hillside appeared to remain stable from 1983 to 1996. In 1995, an inclinometer on the north hillside could no longer be read below a depth of 136 feet, presumably because the casing has sheared at that depth. Highway damage was also observed immediately downstream of the dam in 1995.

The following conclusions were made in the 2006 (most recent of the provided reports) USBR geotechnical report:

• Investigations and instrumentation installed since construction have determined that the entire dam is situated on one or more active landslides.

• The dam impounded reservoir now tends to serve as a buttress to the valley walls.

Sporadic movement still occurs in the left [north] valley wall (mildly exacerbated by periods of high precipitation).

• The slide surface(s) targeted to be intersected by the USBR drill holes are believed to be secondary failure planes within the landslide mass and the primary (main) slide surface occurs at depth and was not encountered.

• The trace of the main slide surface at the slide toe is believed to occur at about elevation

4603.1 (USBR local datum, 4617.0 NAVD88) feet or lower.

GR02-17 Page 3

1.3.2. MANY GLACIER ROAD CONSTRUCTION AND LANDSLIDING

Glacier National Park (GNP) provided WFLHD with historical reports and plans related to construction on Many Glacier Road dating back to 1927. The 1927 construction plans show an existing road at that time, but it is understood that the 1927 construction was the first modern road extending from Babb to the Many Glacier Hotel. In 1953 the road was widened, raised and reconstructed up to top of base course and asphalt pavement was first placed in 1958.

The 1927-1928 construction is well documented in a report titled “Final Construction Report (1927-1928), Babb-Many Glacier, Babb Section, Grading and Surfacing” prepared by A.V. Emery, Asst. Highway Engineer. The project was 12 miles in length extending from the Many Glacier Hotel on Lake McDermott (now known as Swiftcurrent Lake) to Babb.

The construction report documents unfavorable weather in 1927 season. It rained every other day on average between August 13th and 31st. It rained 25 days and snowed 7 days of the 43 days between September 1st and October 13th. There were three days of rain the remainder of October. Gas shovels occasionally sank and became stuck in through cuts for hours or days.

Unstable slopes between MP 0.8 to 2.8 had to be flattened from 1:1 to 1V:1.5H.

A mild winter in 1927/28 allowed spring rains to infiltrate into the ground and resulted in several slides. Two slides crossed the road at MP 6.2 and MP 6.55, both about 500’ long x 300’ wide.

The slide at MP 6.55 moved 3’ horizontally during 1928 season. The slide at MP 6.2 moved 10’ horizontally & 4’ vertically during 1928 season. The third slide was located between Swiftcurrent Creek and road at MP 2.3. As mitigation, lakes and swamps uphill of slides were drained and many fills were brought up 3 or 4 times to grade “before any stability was realized”.

The crushed aggregate surfacing was placed between August 3 and November 2, 1928.

Settlement at some slide areas was observed at completion of the project. The construction report draws the conclusion: “It is quite probable that there will always be movements at some points along the grade.”

In 1953 the road was reconstructed with minor realignments, widening, grade raises, cutting and filling, pipe extensions, and finished up to top of crushed stone base course. No work was done at three locations identified as slide areas (MP 5.25, MP 6.2, and MP 6.55). Geotechnical work was performed during 1957 at those three locations and is discussed below in Section 3.1. A 1958 construction project included placing 2” bituminous surfacing on mainline and paving parking areas. Pavement was not placed in the three slide areas.

GR02-17 Page 4

2. GEOLOGY AND SITE CONDITIONS

2.1 REGIONAL GEOLOGY

The project is located in a region of varied geologic processes and forces. The principal natural architects of the current geologic scene have been the geologic history that created the rock types in the area, the mountain building forces that created the tremendous vertical relief, especially near the west end of Many Glacier Road, and the glaciation that carved the huge U-shaped valley and left lakes and extensive deposits of silty and clayey gravel in the bottom and sides of the Many Glacier Valley.

Thrust faulting was a major factor in creating the mountainous terrain and was directly responsible for reversing the normal positions of younger rocks overlying older rocks. The west-to-east upward angle of thrust faulting in the eastern part of Glacier National Park has brought much older, relatively hard metasedimentary rocks from deep within the earth’s crust and placed them on top of younger, relatively soft sedimentary rocks. Geologists have interpreted the age difference between the two rock masses to be over 500 million years. This feature extends for tens of miles along the entire east side of Glacier National Park and is known as the Lewis Thrust Fault.

Glaciers have played a major role in creating the landscape and depositing the soils that help affect the stability of the Many Glacier Road. A major impact of glaciation in this region is interpreted to have occurred during the Bull Lake glaciation 70,000 to 130,000 years ago. During this period individual alpine glaciers that formed high in the Rockies moved close to the eastern front of the range, coalesced into piedmont glaciers, and continued flowing eastward to their termini in the flatlands. One of those piedmont glaciers was called the St. Mary glacier, a large branch of which flowed into the plains from the Swiftcurrent/Many Glacier valley. The many branches joined into one large moving ice mass that combined with another large glacial ice mass at Babb that was moving north from what is now St. Mary. The glaciers continued north to what is now the U.S./Canada border where they met a continental glacier moving south from interior Canada. The rocks in the moraines of the alpine and piedmont glaciers (such as those found along the Many Glacier Road) are composed largely of limestones and red and green mudstones of the Altyn, Grinnell, and Appekunny formations. Once the glaciers cut through these harder rocks, the underlying Cretaceous shales would have presented little resistance to erosion by the ice masses. The valley contained smaller lakes behind terminal moraine dams before Sherburne Dam was constructed. Withdrawal of the glacier and great inflows of sediments into the valley as the climate warmed could have left several hundred feet of outwash sediments in the valley floor in addition to extensive deposits of morainal till on the valley floor and walls. The fine grain portion of the riverine and morainal deposits was produced as the glaciers pulverized rocks into glacial “flour”. Only portions of the hardest rock types survived the milling and were left as the gravel, cobble, and boulder component of the soils.

Swiftcurrent Creek flows from its headwaters at the continental divide through a series of alpine lakes, over a falls in the limestones of the Altyn formation at Many Glacier Lodge, across the Lewis Thrust Fault, and eastward to the flatlands and the Saint Mary River. The Sherburne Dam, completed in 1921, contains the creek’s flow for water storage in Sherburne Reservoir. The reservoir stretches about 6 miles from the Park Boundary to about a mile east of the Many Glacier Lodge.

GR02-17 Page 5

Distinct differences in vegetation and mass movement are evident in the slopes that angle into Sherburne Reservoir. The Many Glacier Road is located on the south facing slope on the north side of the reservoir. Looking across the water to the south side of the valley, a distinct break in the vegetation color and size is apparent on the unroaded north facing slope. The slope is heavily treed except for some places along the reservoir shore where landsliding has occurred from the erosive impact of wind-generated waves. For most of the length of the impoundment the lower third or so of the north facing slope supports a dense stand of mostly medium size, moderately dark green conifers. The trees at higher elevations on that slope appear darker green and noticeably taller. This is interpreted to reflect differences in the soil, nutrients, and possibly permeability of the glacier moraine soils on the lower slope and the soil types developed by weathering of bedrock on the upper slope.

Vegetation and landforms are very different on the north side of the valley where the road is located. Scattered spruce and fir populate the slope but deciduous trees, primarily aspen, and various woody shrubs are the most common vegetation. The slope is also marked by irregular, open gulches devoid of woody plants. The other striking difference is the presence of irregular, hummocky topography and numerous sag ponds, classic evidence of large scale landsliding.

These features are readily visible, both on the ground and on aerial photos, from the lake shore to the top of the north ridge paralleling the reservoir. Most of the larger landslides that are affecting the road occur along the reservoir between Babb and the entrance kiosk at MP 7.4, which is approximately where the road passes over the Lewis Thrust Fault.

2.2 LOCAL GEOLOGY

Geologic maps of Glacier National Park (Carrara, 1990, and Wipple, 1992) both map most of the Many Glacier Valley as landslide deposits. Carrara and Wipple both note that landslides are common on the east side of the park and that some of the larger landslides exceed 50 m (165 feet) in thickness and cover several square kilometers.

Swiftcurrent Ridge along the north side of the Many Glacier and Boulder Ridge along the south side are mapped by Wipple as till. The till is described as unsorted, subrounded to subangular boulder rubble, consisting mainly of Belt Supergroup rocks, and lesser amounts of sand, silt and clay. He goes on to note that on Swiftcurrent and Boulder ridges the till unit also includes pre- Wisconsin glacial drift which may be up to 60 m (200 feet) thick.

Based on our review of the samples obtained from the MP 6.2 borings performed by WFLHD in 2011 and 2015 (drilled to depths up to 204 feet below the ground surface) the landslide deposits at this site are disturbed glacial sediments that consist of a mixture of clay, silt, sand, gravel, and cobbles with some boulders. Bedrock was not encountered in any of the borings drilled at MP 6.2.

GR02-17 Page 6

2.3 SITE CONDITIONS

Figure 1 and Figure 2 at the beginning of this report are topographic maps and aerial photographs, respectively, of the Many Glacier Valley. Appendix A contains multiple photographs of the site.

Figure 3 is a Google Earth aerial photograph of the MP 6.2 landslide project site. The green line in Figure 3 outlines the interpreted limits of active landsliding at MP 6.2. These limits were verified in the field by using a GPS to record the locations of observed landslide scarps during a site visit in October 2016.

The orange lines in Figure 3 delineate two active individual landslides that have been delineated at MP 6.2. These will be hereafter referred to as the west and east landslides at MP 6.2.

Although the movement of these two areas appears to be somewhat independent and at different rates we believe that due to their proximity the movement of each is related to the other, especially since the inclinometers installed in 1994 and 2001 in the area between these two landslides both showed movement. In addition, movement was detected in boring MG15-05 which was installed outside the limits of the east and west landslides, albeit had much less relative movement.

Figure 3. Aerial Photo of MP 6.2 Landsliding with Site Observations.

Source: 48.8269° N, 113.5523° W, Google Earth, July 26, 2014, accessed February 7, 2017.

The adjacent terrain is mountainous, but the road is generally level and it parallels the north shoreline of Lake Sherburne. At the MP 6.2 site the road abruptly curves downslope about 20 feet due to landslide displacement (see report cover photo). A buried layer of pavement was observed about 70 feet downslope from centerline in a cut made for an access road during the

GR02-17 Page 7

2015 subsurface investigation. We estimate that the road translates downslope at an average rate of about 1 foot per year.

The landslide terrain above the road is hummocky with inclinations ranging from 1V:2H to 1V:10H and an average slope of about 1V:4H. The terrain above the road is vegetated with grasses, woody shrubs and aspen.

The landslide terrain between the road and the lake is also hummocky with slopes ranging from 1:1 to 1V:4H and an average slope of about 1V:3H. The surface is sparsely vegetated with grasses. The surficial soils on much of the slope, especially closer to the road, consist of sandy gravel that is interpreted to be roadway material that has been translated downslope by the landslide. In the bank above the beach the native glacial sediments are exposed consisting of a mixture of clay, sand, and gravel with some cobbles and boulders. Numerous seeps were observed where the toe of the slope intersections the beach (See Figure 3).

Below the high water elevation the beach is sloped between 1V:2H and 1V:8H with an average slope of about 1V:4H in the area where the horizontal drains will be installed. The surficial soils exposed on the beach consist of gravel and cobbles with some sand and boulders. It is interpreted that the finer grained sand to clay sized particles in the glacial sediments have been eroded from the beach and left behind a thin cap of coarser material on the surface.

There are not any known underground utilities at the site. There is an overhead power line that parallels the road but it is located about 500 feet north of the road at the project location.

2.4 CLIMATE

The local climate is typical of the higher elevations in the Rocky Mountain Region – long cold winters with moderately warm summers, moderate amounts of precipitation, and moderate to high amount of snowfall. The Western Regional Climate Center (WRCC) has two weather Stations in the area, one at Many Glacier (with a limited amount of data) and another in the town of St. Mary.

The Many Glacier Station is located near the Many Glacier Lodge which is about 5 miles further up the valley. The St. Mary Station is located in the lower reach of the St. Mary valley. There is a significant difference in the average annual precipitation and snowfall between the two Stations.

Weather at the project site is likely to be similar to either these two stations depending on the local climatic conditions at a specific time. Data from the stations is summarized in Table 1 and Table 2.

GR02-17 Page 8

Table 1. Climate Summary from WRCC Station 245361, Many Glacier.

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Annual

Average Max.

Temperature (F)

22.8 34 37.1 46.6 56.5 64.8 72.5 70.8 61.1 46.7 36.4 28 48.1

Average Min.

Temperature (F)

2.8 17.2 20.4 29.2 36.5 42.1 45.1 44.3 39.5 31.3 21.2 11.1 28.4

Average Total Precipitation (in.)

5.47 2.71 3.66 1.82 3.57 4.9 1.82 2.09 3 3.3 3.11 5.35 40.8

Average Total SnowFall (in.)

63 21.1 36.6 21 5 1.5 0 0 1.5 8.8 18.9 33.1 210.5

NOTES:

Data from Western Regional Climate Center website, www.wrcc.dri.edu, Many Glacier, Montana (245361), accessed 1/30/2017.

Period of record monthly climate summary from 8/9/1967 to 10/31/1980; Percent of possible observations for period of record:

Max. Temp.: 39.9% Min. Temp.: 39.4% Precipitation: 41.4% Snowfall: 27.1% Snow Depth: 25.5%

Table 2. Climate Summary from WRCC Station 247292, St. Mary.

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Annual

Average Max.

Temperature (F)

32.3 36.6 42.2 51.6 60.7 68.7 78.1 78.7 67.4 53.7 38.9 31.5 53.4

Average Min.

Temperature (F)

13.8 16.5 21.9 28.1 35.6 42.2 46.9 45.4 38.8 32.9 23.8 14.9 30.1

Average Total Precipitation (in.)

2 1.7 1.79 1.93 2.95 3.58 1.82 1.65 2 2.17 2.44 2.07 26.11

Average Total SnowFall (in.)

19.9 18.3 17.9 13.4 4.7 0.9 0 0.4 1.5 7.5 16.8 18.9 120.3

Average Snow Depth (in.)

6 7 6 2 0 0 0 0 0 0 2 4 2

NOTES:

Data from Western Regional Climate Center website, www.wrcc.dri.edu, St. Mary, Montana (247292), accessed 1/30/2017.

Period of record monthly climate summary from 5/1/1981 to 5/25/2016; Percent of possible observations for period of record:

Max. Temp.: 98.8% Min. Temp.: 98.3% Precipitation: 97.4% Snowfall: 98% Snow Depth: 92.9%.

2.5 LAKE SHERBURNE

The US Bureau of Reclamation (USBR) is the operating agency of the Lake Sherburne dam.

Water retained in the reservoir is used primarily for irrigation but also provides habitat for fish and other wildlife and recreation opportunities.

The vertical datum used for the original construction of the dam is not known but the USBR continues to use this datum for all reported elevations. In 2016, a land surveyor for the Montana Area Office of USBR determined that the elevation differential for the Lake Sherburne Dam vertical datum to the North American Vertical Datum of 1988 (NAVD88) is +13.915 feet. This correction is applied to all data from the USBR in order to relate data to the project datum. This report presents all vertical datum elevations in NAVD88.

Lake Sherburne reservoir forebay elevations for the last 10 years are presented in Figure 4. The reservoir spillway elevation is 4801.9 feet (NAVD88). From the figure it can be seen that the reservoir elevations rise and fall relatively fast, with elevation rises averaging 25 feet, but as much as 45 feet, during the months of May and June and elevation drops averaging 32 feet, but as much as 45 feet, during the months of July and August.

GR02-17 Page 9

Figure 4. Lake Sherburne Reservoir Forebay Elevations, 2006‐2016.

GR02-17 Page 10

3. SUBSURFACE INVESTIGATIONS

MP 6.2 has experienced instability since landsliding occurred during the road construction in 1927. Multiple subsurface investigations have been performed at this location. Prior to our investigation in 2015, the investigations were generally limited in scope and did not include any geotechnical instrumentation to collect data on subsurface groundwater and movement of subsurface slide surfaces. This section summarizes the results of the investigations that have been performed at MP 6.2 to date.

3.1 1956 DRILLING AND HORIZONTAL DRAIN PROGRAM

In 1956, MP 6.2 was one of three locations along the road that were identified for an experimental drilling program to provide drainage as a means of stabilization. The project included a total of 7647 feet of drilling at the three sites in the spring of 1957, including 8 vertical and 31 horizontal holes (Schwegler, 1958). Vertical holes were drilled up to depths of 200 feet and horizontal holes to depths of 430 feet. The horizontal holes were drilled at inclinations between +6% to +15% and 1.5” slotted steel water pipe was installed as drain pipe. The total water flow from the horizontal holes in October of 1957 was greater than 5000 gallons per day. Schwegler interpreted that the slide surfaces were relatively shallow with slip planes at depths between 22 and 43 feet below the ground surface in most cases.

As part of the experimental program, two vertical holes and 8 horizontal drains were installed at MP 6.2, which was referred to as Station 350 or Location No. 2 at that time. Two of the horizontal drains were installed adjacent to the road and 8 horizontal drains were installed along the lake shore. The locations of the drains were not reported but the report did include cross-sections with interpreted subsurface stratigraphy for the sites where the drains were installed (See Figure 5).

One of the horizontal drains installed adjacent to the road is still visible and sections of the steel pipe that were installed can be found along the beach below the slide. The drains were installed at inclinations of 6 and 15 degrees and had lengths between 180 and 325 feet. Water was reported out of 6 of the 8 drains and ranged from “dripping” up to 390 gallons per day.

Although many of the horizontal drains produced several hundred gallons per day, many others were only dripping which led to the conclusion that they were not effective in lowering the water tables in the slide areas. This was attributed to the “gumbo” type soil encountered in the slide areas being very slow draining.

GR02-17 Page 11

Figure 5. Cross‐section of interpreted slide surfaces from Schwegler (1958).

The cross-section for MP 6.2 (Location No. 2) shows a layer of light brown clay and boulders up to 20 feet thick overlying shale and argillite was encountered at depths between 94 and 109 feet.

We believe that the shale and argillite bedrock are misinterpretations of the subsurface conditions likely due to the drilling methods that were used. Schwegler indicates that gear type rotary oil well bits were used, and since no sampling is mentioned we believe that soil descriptions were based on observations of cuttings only. The glacial sediments that were encountered at similar depths in recent investigations contained shale and argillite particles at those elevations that could have been interpreted as bedrock if relying on cuttings as the only means to determine subsurface stratigraphy.

3.2 1994 SUBSURFACE INVESTIGATION

In the Fall of 1994, two investigation borings were drilled by the USBR on behalf of Western Federal Lands Highway Division (WFLHD) as a limited effort to collect some initial subsurface information at some of the significant landsliding locations along the road. One of the borings was off the shoulder at MP 6.2 (boring DH1-94-6.1). The boring location is shown on the Boring Location Map in Appendix B and the boring log is attached in Appendix B. The boring was drilled to a depth of 101 feet and slope inclinometer casing was installed. The slope inclinometer casing sheared between depths of 40-44 feet the following spring.

Boring DH1-94-6.1 encountered 8 feet of roadway fill overlying glacial sediments that extended to the termination depth of 101 feet. The soil layers consisted of various mixtures of clay, sand and gravel described as clayey sand with gravel, sandy lean clay with gravel, sandy lean to fat clay, clayey gravel with sand, poorly graded gravel with clay and sand, and sandy lean to fat clay with gravel. Water levels were recorded at depths of 42.1 and 48.2 feet during drilling. Penetration resistances using the Standard Penetration Test (SPT) ranged from 6 to 19 blows per foot in the upper 35 feet, then from 23 to 46 blows per foot at depths of 35 to 68 feet, then to values greater than 50 blows per foot from 68 feet to bottom of hole.

GR02-17 Page 12

3.3 2011 SUBSURFACE INVESTIGATION

In 2011 WFLHD drilled borings at 9 locations along Many Glacier Road between the Park Boundary at the dam (MP 4.7) and the Park Entrance Kiosk (MP 7.5). The intent of that study was to determine the general size and depth of the landslides and develop conceptual alternatives for landslide stabilization. Field observations, site studies, instrumentation results, evaluation of alternatives, and concepts for landslide mitigation are presented in WFLHD Geotechnical Memorandum GM03-13 to Mike Traffalis, WFLHD Project Manager, prepared by Brian Collins, WFLHD Geotechnical Engineer, dated March 29, 2013.

One of the 2011 borings was off the shoulder at MP 6.2 (boring MG11-01, also referred to as B-1). The boring location is shown on the Boring Location Map in Appendix B and the boring log is contained in Appendix B. Boring MG11-01 encountered 14 feet of fill at the surface described as cobbly gravelly silty sand. Beneath the fill the boring encountered glacial sediments consisting of a mixture of clay, silt, sand and gravel with some cobbles to the termination depth of the boring at 150.3 feet. The glacial sediments were described as: gravelly sandy clay, sandy clay, clayey sand, lean clay with sand, sandy gravelly clay, silty clayey sand with gravel, sandy lean clay with gravel, silty clay, and gravelly silty clayey sand.

Penetration resistances in the upper 32 feet were generally less than 20 blows per foot (bpf).

From 30 to 55 feet penetration resistances were generally between 20 and 50 bpf. Below 55 feet penetration resistances were generally greater than 40 bpf. The reported penetration resistances were obtained by alternating use of the SPT (2-inch), modified California (2.5-inch), and Dames and Moore split barrel sampler (3-inch).

Water was recorded at a depth of 32 feet while drilling. A vibrating wire piezometer was grouted into the borehole and data was recorded every 12 hours from July 13, 2011 to October 13, 2015.

A plot of the vibrating wire piezometer data from Boring MG11-01 is contained in Appendix F.

Slope inclinometer (SI) casing was also installed in Boring MG11-01. The SI casing extended to a depth of 150 feet. SI data was collected on 7/19, 8/16, 9/16, 10/18, 11/16, and 11/29 in 2011.

When we attempted to perform a reading in June 2012 we discovered that the casing had sheared off at a depth of 37 feet. Plots of the SI data are contained in Appendix F.

3.4 2015 SUBSURFACE INVESTIGATION

In October 2015 five borings were drilled and instrumentation was installed in each borehole.

This effort was performed by WSP Parsons Brinkerhoff (PB) on behalf of WFLHD. PB subcontracted the drilling services to Haztech Drilling, Inc., of Meridian, ID and the instrumentation installation and setup to Geocomp’s New York, NY office.

Boring locations are shown on the Boring Location Map in Appendix B. Boring locations, coordinates, ground surface elevations, depths drilled, and depths and elevations of vibrating wire piezometers are summarized in Table 3.

GR02-17 Page 13

Table 3. 2015 Boring Locations, Drilling Depths, and VWP Installations.

Boring No.

GPS Coordinates

(WGS84)

Ground Surface El. (NAD88)

(feet)

Total Depth Drilled (feet)

VWP Depth (feet)

VWP

Elevation

(feet)

MG15-01

48.8268536° N

4836.0 76.5

40.5 4795.5

113.5518539° W 69.5 4766.5

MG15-02

48.8265933° N

4803.4 64.0

30.3 4773.1

113.5518269° W 60.3 4743.1

MG15-03

48.8268019° N

4848.8 75.5

40.2 4808.6

113.5527419° W 70.2 4778.6

MG15-04

48.8266808° N

4828.7 66.5

29.8 4798.9

113.5527781° W 59.8 4768.9

MG15-05

48.8271581° N

4855.3 204.0

46.0 4809.3

113.5512656° W 179.5 4675.8

Logs for all five borings (MG15-01 through MG15-05) are contained in Appendix B.

The borings were completed between October 6 and October 14, 2015. A track-mounted CME 850 drill operated by Haztech Drilling, Inc. was used to advance the borings. 8 1/4 -inch outside diameter hollow-stem augers were used to advance boring MG15-01 through MG15-04. In boring MG15-05, hollow-stem augers were used to advance the hole to a depth of 55 feet, then 3-7/8-inch tricone mud rotary and HQ3 wireline coring methods were used until the bottom of hole.

Penetration tests were completed from continuous to 5-foot intervals in the borings. Penetration testing used both the Standard Penetration Test (SPT) sampler (2-inch O.D.) and the Dames & Moore sampler (3-inch O.D.). The penetration test consists of using a 140 pound hammer with a 30 inch drop to drive a two inch outside diameter split-barrel sampler 18 inches below the bottom of the bore hole and recording the number of blows for each 6-inch increment of penetration. The last two numbers (for the final 12 inches of penetration) are added together to determine the blows per foot (bpf), which is known as the N-value. In some cases, the split-barrel sampler was driven 24 inches. In those cases the numbers for the middle two 6-inch increments were added together to determine the N-value. Note that corrections for sampler size, overburden pressure, and hammer energy have not been applied to the N-values shown on the boring logs.

Borings were logged in the field by a Parsons Brinkerhoff Geotechnical Engineer in general accordance with the Federal Lands Highway Soil Description and Identification Guidelines (2015).

Samples were collected and sent to WFLHD in Vancouver, WA. The samples and draft boring logs were reviewed by the WFLHD Geotechnical Engineer assigned to the project.

3.4.1. SUBSURFACE PROFILE

As discussed previously in the LOCAL GEOLOGY Section it is our interpretation that the landslide deposits at this site within the depths penetrated by our borings are disturbed glacial sediments

GR02-17 Page 14 that consist of a mixture of clay, silt, sand, gravel, and cobbles with some boulders. Boring logs are contained in Appendix B.

On average the split barrel samples that were tested contained about 20 percent gravel, 30 percent sand, and 50 percent fines (silt and clay). Based on the lab test results most of the soils at the site classify as sandy lean clay, sandy lean clay with gravel, clayey sand, clayey sand with gravel, and silty clayey sand with gravel according to the Unified Soil Classification System

(ASTM D 2487).

Generally, SPT N-values increased with depth in all the borings. The SPT N-values were generally less than 20 bpf in about the upper 30 feet of the borings. Between the depths of about 30 to 50 feet the N-values generally ranged from 20 to 50 bpf. Below about 50 feet the N-values were generally greater than 50 bpf.

Cobbles and boulders are present in the subsurface profile, but are not represented in the lab test results since the samples that were tested were retrieved from 2-inch and 3-inch O.D. split barrel samplers. Boring MG15-05 used HQ wireline coring methods from depths of

70.5 feet to 204 feet. A layer of cobbles in a clay and gravelly sandy clay matrix was encountered from 70.5 feet to 85 feet. From 85 feet to 110 feet MG15-05 encountered gravelly lean clay with cobbles. From 140 to 142.5 feet MG15-05 encountered gravelly clay with cobbles (See core photos in Appendix B). The present of cobbles and boulders in the profile is also evident by the boulders observed on the shoreline. Most of the boulders observed on the shoreline are 1 to 2 feet in diameter with a few up to 3 to 4 feet in diameter.

3.4.2. GROUNDWATER OBSERVATIONS

Groundwater was observed at depths between 8 and 27 feet either during drilling or at drilling completion in boring MG15-02, MG15-03, and MG15-05. In the other two holes groundwater was not observed during drilling. Two vibrating wire piezometers (VWP) were installed in each borehole. The VWPs were grouted in the boreholes following the procedures recommended by the manufacturer (DGSI, 2017). The installation depths of the VWPs are summarized in Table 3 on page 13.

Plots of the recorded groundwater levels over time are contained in Appendix F. The VWPs in MG15-03 and MG15-04 stopped working in June 2016. We believe the cables likely failed due to landslide displacement.

3.4.3. INSTRUMENTATION

Instrumentation consisting of vibrating wire piezometers (VWP) and ShapeAccelArrays (SAA) was installed in boring MG15-01 through MG15-05. A tipping bucket rain gauge was also installed at the site. VWP installation is discussed in Section 3.4.2. Geocomp was subcontracted by WSP Parsons Brinkerhoff to order, setup, test, and install the geotechnical instrumentation at the site.

Two Geocomp technicians were on site during the drilling in October 2015 to install and setup the instrumentation. Installation records for all the instruments installed in the boreholes are contained in Appendix C.

GR02-17 Page 15

A Measurand ShapeAccelArray (SAA) was installed in each borehole to approximately the bottom depth shown on Table 3. The SAAs used on this project consist of arrays of 1.64 feet-long (500 mm-long) segments placed vertically in the boreholes. Each segment contains 3 orthogonal microelectromechanical sensors (MEMS) accelerometers. The SAAs were installed into 1.049 inch (27 mm) ID PVC casing prior to being inserted into the borehole. The VWPs and grout tubes were attached to the PVC casing and all the instruments were grouted into the boreholes.

All instrumentation was connected to Campbell Scientific CR1000 data loggers installed inside weather resistant enclosures. One data logger station was installed east of the slide area (remote station) and other to the west (base station). MG15-01, MG15-02, and MG15-05 were connected to the remote station and MG15-03 and MG15-04 were connected to the base station. A Campbell Scientific RF401 Radio is used to send data from the remote station to the base station.

Data is transmitted from the base station using a satellite modem. Rechargeable batteries connected to solar panels are used to power the instrumentation and data loggers. Data that is collected from the instrumentation is viewed on a website hosted and maintained by Geocomp https://gccgis.geocomp.com/iscgis/project/?Glacier (accessed 2/21/2017). Data is recorded and transmitted from all the instruments once per day.

Gravity measurements by the MEMS accelerometers in the SAA allow the 3D shape of the borehole to be determined. Comparing data sets from the SAA over time shows locations where subsurface movements occur. SAA plots are contained in Appendix F. One of the sensors in MG15-04 stopped working in June 2015, at about the same time the VWPs stopped working in MG15-03 and MG15-04. The timing of this malfunction coincides with an episode of rapid landslide movement on the order of 6 inches. The satellite connection was used to turn the sensor off in August 2015. The damaged sensor affected the entire SAA data set so there is a gap in the data for MG15-04 during this period.

Plots of data from the tipping bucket rain gauge and of Lake Sherburne reservoir elevations for the same period of record are contained in Appendix F. Daily data for the Lake Sherburne reservoir elevation are available from the US Bureau of Reclamation. We have collected approximately 16 months of data at the time this report was prepared.

3.4.4. LABORATORY TESTING

Selected split barrel samples were submitted to the WFLHD Materials lab for soil index tests.

Laboratory test reports are contained in Appendix D. Note that data from boring MG11-01 is listed as boring B-1 on the lab test reports. The following is a list of tests that were completed on the selected soil samples and the standards that were followed:

Particle Size Analysis – AASHTO T 88 Liquid Limit, Plastic Limit, and Plasticity Index – AASHTO T 89/90 Natural moisture – AASHTO T 265 Apparent Specific Gravity – AASHTO T 100

The results of the soil index tests were used to classify the samples according the Unified Soil Classification System (USCS). The USCS classification is reported on the boring logs.

Based on the results of 48 samples tested by particle size analysis, the soil has from 0 to 58 percent gravel-sized particles (minus 3 inch, retained on No. 4 sieve) and an average of 17

GR02-17 Page 16 percent gravel. The soil has from 1 to 58 percent sand-sized particles (passing No. 4 sieve, retained on No. 200 sieve) and an average of 34 percent sand. The soil has have from 13 to 99 percent silt- and clay-sized particles (passing No. 200 sieve) and an average of 49 percent silt and clay.

From the Atterberg limit test results the liquid limit ranged from 19 to 56 and averaged 31, the plastic limit ranged from 14 to 28 and averaged 17, and the plasticity index ranged from 4 to 28 and averaged 14.

Seven split barrel and Shelby tube samples were submitted to the Washington State DOT Materials Laboratory for ring shear tests. Laboratory ring shear test reports are contained in Appendix E. A summary of the Torsional Ring Shear Test results is presented in Table 4.

Table 4. Summary of Torsional Ring Shear Test Results.

Boring No. MG15-01 MG15-01 MG15-01 MG15-03 MG15-03 MG15-03 MG15-04

Sample Depth (feet) 25-26.5 28.5-30 35-36 36.5-37.5 37.5-38.5 40-42 12.5-14

Sample No. S-9 S-11 S-15A S-16A S-16B S-18 S-5

USCS

SC,

clayey

SAND with gravel

SC,

clayey

SAND with gravel

CL,

sandy lean

CLAY

CL,

sandy lean

CLAY

CH,

fat CLAY

CL,

lean CLAY

SC,

clayey

SAND

Water Content (%) 12 10 16 21 31 29 12

Liquid Limit 31 28 37 35 56 45 30

Plastic Limit 18 18 21 20 28 23 19

Plasticity index 13 10 16 15 28 22 11

Percent Passing No. 200 Sieve

43 24 64 51 99 87 49

Clay Fraction (%) 12 6 23 16 57 42 13 r' best fit (deg) 18.3 17.1 17.4 11.5 9.2 19.9 20.4 r' @58 psi (deg) 21.0 19.1 19.4 12.8 9.1 19.5 20.3 r' @29 psi (deg) 19.9 24.1 19.3 17.1 9.1 20.0 21.6

GR02-17 Page 17

4. ANALYSIS AND RECOMMENDATIONS

4.1 DISCUSSION

It is our interpretation that the extremely active slide area at MP 6.2 is limited to the area outlined as “Limits of landsliding” outlined in green in Figure 3. These limits are very small in comparison to the larger landslide complex delineated by the red scarp lines in Figure 2. Although the larger landslide complex has features that indicate recent activity along the entire north side of the valley, it is our interpretation that these much larger landslides have much less movement compared to the smaller more active landside masses that cause the recurring roadway damage.

Charts comparing precipitation, vibrating wire piezometer water levels, SAA displacement, and Lake Sherburne elevation at each borehole location are contained in Appendix F. From these charts it can be seen that landslide movement at this site is influenced by a combination of precipitation, snow melt and reservoir fluctuations. In addition, based on our site observations, beach erosion is also a contributing factor to landslide movement at the MP 6.2 site.

The VWPs generally all showed a response following heavy precipitation events, with groundwater levels spiking 2 to 5 feet following those events. Several of the VWPs also had a general rising trend from late February through June which was a wetter period with about 10 inches of accumulated precipitation. The VWPs in MG15-02, which is closest to the beach, responded to fluctuations in the lake water level. The water level indicated by the lower VWP at this location generally followed within 1 to 2 feet of the elevation of the reservoir. The water level indicated by the upper VWP at this location generally followed the reservoir elevation at higher pool levels but it had a much slower response and never dropped below about elevation 4783.

During the same observation period the reservoir elevations ranged from 4765 to 4801. Elevation

4801.9 is the Lake Sherburne spillway elevation. Further away from the reservoir there appears to be two water levels: a lower water level that is generally stable near the reservoir spillway elevation of 4801.9, and higher water levels that fluctuate in response to precipitation and snowmelt. The higher water levels that are indicated by the VWPs range from elevation 4810 to 4850.

The SAA installed in MG15-01, MG15-02, MG15-03, and MG15-04 all show that accelerated movement started in mid-late February 2016 and displacements continued at increasing rates until early April 2016. The landsliding continued at these relatively fast rates until mid-August 2016, at which time the movement decelerated to a relatively slow rate for the remainder of the monitoring period. The on-site rain gauge recorded over 7 inches of rain between mid-February and mid-August 2016. The two spikes in movement in the MG15-02 SAA correlate well with drops in reservoir level that occurred in late March and July 2016. Total displacements of these four SAA ranged from 4 to 42 inches over the monitoring period.

The MG15-05 SAA was installed with the intention of determining if a larger, deep seated landslide is present at the site. The SAA extends to a depth of 200 feet below the ground surface.

The SAA indicates that displacement is occurring at elevation 4802 at a fairly steady rate of about

1.5 inches per year. This is approximately the same elevation as the lower slide plane detected in boring MG15-01, which had a similar magnitude of movement at this elevation. Although it appears to be moving at a relatively slower rate, this area is interpreted to be part of the active landslide mass as shown in Figure 3 on page 6.

GR02-17 Page 18

Unfortunately the drilling program was limited to the slide area between the road and the beach due to environmental and cultural concerns. Therefore our interpretations of the subsurface movements in the slide area above the road are based solely on our site reconnaissance.

The data indicates that at MP 6.2 the extremely active sliding is occurring in the upper 15 to 40 feet of the soil profile in the glacial sediment layers that had N-values between 3 and 20 bpf.

Slide movement in this extremely active zone averages about 14 inches per year. The deeper seated relatively slower sliding is occurring at depths between 35 to 55 feet in the glacial sediments layers that had N-values between 10 and 50 bpf. Slide movement in this less active zone averages about 1.5 inches per year. These findings are consistent with the results of the previous investigation (documented in Geotechnical Memorandum GM03-13). GM03-13 discussed that ancient massive landslides are active between MP 4.8-7.4 with relatively slow and variable rates of movement such that the effects of the movement have not caused significant damage to the roadway other than generally hummocky terrain and occasional cracking where the road crosses smaller faster moving slides and along the flanks of the massive landslides. Six locations were identified, including MP 6.2, with more localized landslides that are significantly impacting the performance and maintenance requirements of the roadway. It was interpreted that these landslides are generally much shallower and moving at a much faster rate relative to the more massive landslides within the complex. These interpretations were confirmed by the data collected from the instrumentation installed at MP 6.2 in 2015.

4.2 ANALYSIS

The software program SLIDE 7.0 by Rocscience Inc. was used to perform global stability analyses of the landslides. SLIDE 7.0 is a two-dimensional limit equilibrium program that also has the capability to perform a finite element groundwater seepage analysis. The landslides were evaluated by reviewing the slide conditions, groundwater conditions, subsurface materials, surface topography, and aerial photographs to produce generalized cross-sections. A single subsurface layer was modeled since the glacial sediments within the depth of sliding were relatively consistent in soil type and SPT penetration resistances (N-values).