Bulletin: West Salt Creek Report

7/17/2019 Bulletin: West Salt Creek Report

1/55

The West Salt Creek Landslide:

A Catastrophic Rockslide and Rock/Debris Avalanche

in Mesa County, Colorado

By Jonathan L. White, Matthew L. Morgan, and Karen A. Berry

Bulletin 55

2015

Colorado Geological Survey

COLORADOSCHOOLOFMINES


2/55


3/55

Te West Salt Creek Landslide

i Colorado Geological Survey

Bulletin 55

The West Salt Creek Landslide:

A Catastrophic Rockslide and Rock/Debris Avalanche

in Mesa County, Colorado

by

Jonathan L. White, Mahew L. Morgan, and Karen A. Berry

Colorado Geological Survey

COLORADOSCHOOLOFMINES

2015

GEO

LOGICAL

SU

RV

EY

COLO

RADO


4/55


iiColorado Geological Survey

Bulletin 55

Cover Photo: Oblique aerial image o the May 25th, 2014 West Salt Creek landslide near Collbran, Colorado. Te landslidecomplex, which includes rock slide, rock avalanche, and debris-flow components, traveled approximately 3 miles down the WestSalt Creek valley. Te landslide covers almost one square mile. Aerial view is looking south toward the Grand Mesa. Image wastaken on May 26th, 2014.

Tis report is dedicated to the memory o the three men that perished in the landslide:Wes Hawkins, Clancy Nichols, and Danny Nichols.


5/55


iii Colorado Geological Survey

Bulletin 55

TABLE OF CONTENTS

Table of Contents. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii

List of Figures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v

Acknowledgements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii

Introducon. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Regional Seng. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

Evidence of Past Landslide Acvty. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

Landslide Modes of Failure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Data Used in Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Timeline. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Descripon. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

Headscarp. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

Upper Block. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Rock Avalanche. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Esmated Velocity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Landslide Mobility. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Rock Avalanche Morphology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

Timing of Major Deposional Events. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

Early Surge. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Middle Surge. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Late Surge. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

Triggering Mechanisms. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

Rotaonal Failure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Rock Avalanche and Debris Flow Failure(s). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

Monitoring. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

Future Concerns and Long-term Hazard Assessment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 Long-term Stability of the Upper Block. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

Reacvaon and Retrogressive Failure from Above the Landslide Headscarp. . . . . . . . . 33

Threat of Mud/Debris Flows. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

Creeping Earth-ow Spread and Sediment Loading. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

Burial and Shearing of Oil and Gas Well Heads. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

Land-use Recommendaons and Future Study Consideraons. . . . . . . . . . . . . . . . . . . . . . . . . . 35

Public Use of U.S. Forest Service (USFS) Lands within the Landslide Area. . . . . . . . . . . . . 35

Future Development and Land Usage of the Landslide Vicinity. . . . . . . . . . . . . . . . . . . . . . . 35

Addional Instrumentaon. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

Regional Recommendaons. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

Appendix A - Police Reports. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

Appendix B - Eyewitness Interview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45


6/55


ivColorado Geological Survey

Bulletin 55

TABLE OF CONTENTS con't

Map Plate 1 - Geologic Map

Map Plate 2 - Cross Secon A-A


7/55


v Colorado Geological Survey

Bulletin 55

LIST OF FIGURES

1. Locaon map of the West Salt Creek landslide and surrounding area. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

2. Topographic map and oblique views of original topography and prehistoric landslides prior to the WSCL. . . 4_5

3. Pre- and post-WSCL oblique images with locaon of the prehistoric bench and landslides. . . . . . . . . . . . . . . . . . . 6

4. Seismogram of the WSCL event. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

5. Oblique photo of the WSCL taken May 26th, 2014 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

6 . Photos of the rockmass at WSCL headscarp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9_11

7. Aerial photo of the WSCL Upper Block and hypothesized cross secon. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

8. Photos of disturbed sediments on the Upper Block. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

9. False-color material dierence model with bulleted photo locaons. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

10. Photos of rock avalanche deposits showing variaons in clast size, matrix, and color. . . . . . . . . . . . . . . . . . . 14_16

11. Hillshaded LiDAR of the WSCL with data point locaons used in velocity calculaons. . . . . . . . . . . . . . . . . . . . . . 17

12. Hillshaded LiDAR of the WSCL analyzed for velocity esmates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

13. Cross-secon proles created from pre- and post-event DEMs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

14. Graphic representaons of volume, mobility index, and runout lengths of catastrophic landslides . . . . . . . . . . 20

15. Low-sun-angle aerial photograph showing crosscung relaonship of discrete surges in avalanche debris

eld. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

16. Photos demonstrang rapid slaking of Green River Formaon claystone and marlstone clasts . . . . . . . . . . . . . 24

17. Photos of middle surge red bouldery deposits showing dierences in composion, color, and cross-cung

relaonships with early surge rock-avalanche deposits. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

18. Photo of well-dened side furrow and lateral shear zone on west side of the debris eld. . . . . . . . . . . . . . . . . . 26

19. Instrumentaon and property ownership map. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

20. GPS sensor locaons with horizontal vectors and vercal osets measured through October 1, 2014. . . . . . . 30

21. Photo of large ponds in the hummocky late-surge oor of the WSCL rock avalanche debris eld taken

June 5, 2015. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

22. Photos of a natural pipe with owing water near the base of the Upper Block taken June 6, 2015. . . . . . . . . . 32

23. Image of the WSCL sag pond near spill over taken June 6th, 2015. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

24. Map of the greater Plateau Creek basin showing locaon of oil and gas well pads and CGS landslide

inventory. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36


8/55


viColorado Geological Survey

Bulletin 55

Table 1. Summary of measured variables from cross-secons A'-C' . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18


9/55


vii Colorado Geological Survey

Bulletin 55

ACKNOWLEDGEMENTS

Data and observations included in this report were used rom the ollowing sources:

Field mapping and geologic characterization by Jon White and Matt Morgan o the Colorado

Geological Survey (CGS), with collaborative observations by Karen Berry (CGS). Geologists andhydrologists involved in emergency response activities in the two week period immediately ater

the landslide were Jon White and Karen Berry; Je Coe, Jason Kean, and David Brown with the

U.S. Geological Survey (USGS), and Ben Stratton with the U.S. Forest Service (USFS). Je Coe,

Rex Baum, and Matt Morgan completed detailed ield mapping or the U.S. Geological Survey

(USGS) and provided key observations and discussions. Rex Cole and Verner Johnson o Colorado

Mesa University also conducted ield reconnaissance o the landslide and provided inormation on

the bedrock units. im Hayashi o the Mesa County Public Works Engineering section managed

overall site coordination. He and Frank Kochevar also provided and installed instrumentation, as

did Ben Stratton, Je Coe, and David Brown. he USFS provided and installed web-based camera

stations. Immediate post-landslide UAV imagery was provided by Mesa County Sheri s Oicecourtesy o Ben Miller. Frank Kochevar provided additional UAV imagery in the summer o 2014.

he CGS acquired high-resolution LiDAR (Light Detection and Ranging) elevation data. Doug

Bausch at FEMA supplied a pre-slide 4-m-resolution ISAR (Intererometric Synthetic Aperature

Radar) DEM. Other imagery was made available rom Mesa County Public Works and GIS/I

departments. Seismograms are rom the Colorado Mesa University Seismic Network, courtesy o

David Wolny. Mesa County, USGS, and Colorado School o Mines (CSM) also shared other data

sets. Kevin McCoy and Dr. Paul Santi, CSM Department o Geology and Geological Engineering,

conducted a stability analysis o the landslide Upper Block. Deputies Jim Fogg and erry Bridge

wrote Mesa County Sheri s emergency response/eyewitness reports. he West Salt Creek

Emergency Action Plan was written by Andy Martsol, Mesa County Emergency Manager. his

paper greatly beneited rom reviews by Rex Baum and Je Coe (USGS).

Jay Melosh (Purdue University) and Brandon Johnson (MI) provided valuable discussion

on velocity estimations and reviewed the velocity estimation section o this report. Francis Scot

Fitzgerald (CGS) generated the material dierence model o the landslide and provided GIS

support during the project. Karen Morgan (CGS) drated the cross-section proiles used in the

velocity calculations and also provided GIS support. Larry Scott (CGS) drated several igures,

compiled the geologic map plate and created the design and layout o this report.


10/55


viiiColorado Geological Survey

Bulletin 55


11/55


1 Colorado Geological Survey

Bulletin 55

On May 25th, 2014 the longest landslide in Coloradoshistorical record occurred in west-central Colorado, 6 mi(10 km) southeast o the small town o Collbran in MesaCounty. Tree local men perished during the catastrophicevent and subsequent recovery efforts ailed to locate theirremains or vehicles. Te toe o the landslide came within200 (61 m) o active gas-production wellheads. Loss oirrigation ditches and water impacted local ranches andresidents. Continuing threats to nearby residents remain.

Te landslide complex is reerred to as the West SaltCreek landslide (WSCL) by local authorities (see ront-piece photo). Te ast-moving, high-mobility landslidewas caused by an initial rotational slide o a hal-mile-wideblock o Eocene Green River Formation. Te resultantrock ailures, rockmass disaggregation, and mostly valley-constrained rock avalanche, dropped approximately 2,100 (640 m) in elevation while extending 2.8 mi (4.5 km) romhead scarp to toe, 2.2 mi (3.5 km) as a rapid series ocascading rock avalanche surges o chaotic rubble composedo ragmented to pulverized rock, vegetation, topsoil, and

lesser amounts o colluvium and mud. Local seismometersrecorded a magnitude 2.8 earthquake rom the event witha seismic wave train duration o slightly less than 3 minutes.Within minutes, the WSCL covered almost a square mileo the West Salt Creek valley and the net volume displace-ment, calculated by digital elevation model (DEM) differ-ence modeling, was 38 million yd3(29 million m3).Tough its very high mobility and surace morphologynormally indicate a flow-like mechanism, the rock avalanchedeposits were surprisingly dry with no visible mud orwater seeps when examined 24 hours aer the event.

Calculated thicknesses o the rock avalanche depositswere as high as 123 (~38 m) at the valley floor. Althoughprecise pre-event conditions are largely unknown, geomor-phic analyses o aerial photography, digital terrain models,and review o earlier landslide hazard mapping (Soule,

1988) indicate the WSCL occurred at the same location oa similar prehistoric landslide rom the recent geologic past(Holocene,


12/55


2Colorado Geological Survey

Bulletin 55

Te landslide lies in west-central Colorado within thewest-ork valley o Salt Creek, which is a tributary streamwithin the greater Plateau Creek basin between BattlementMesa to the north and the northern flank o Grand Mesato the south. Te closest town, Collbran (pop. 708, 2010census) is 6 mi (10 km) downstream on Plateau Creek.County Road 330 and State Highway 65 ollow PlateauCreek to its confluence with the Colorado River in DeBequeCanyon where State Highway 65 links with Interstate 70(see Figure 1).

Te bedrock in the upper Plateau Creek basin area isearly Cenozoic in age. Te units dip very gently andthicken to the north into the Piceance Basin, a syntectonicLaramide structural basin and an important oil and gasproducer with one o the largest reserves o oil shale in theworld. Te Paleocene to Eocene Wasatch and Green RiverFormations underlie most o the Plateau Creek basinterrain. Above the Green River Formation at the rim oGrand Mesa are strata o the Uinta Formation, which isoverlain by the Goodenough ormation (inormal name) oprobable Miocene age (Cole and others, 2013). A geologicmap, cross section, and list o map units adjacent to theWSCL are shown in Map Plates 1 and 2. Te mudstonesand sandstones o the Wasatch Formation occupy thePlateau River valley floor. Te shales, marlstones, andminor oil shales o the Green River Formation and shaleand sandstone beds o Uinta Formation at the top, orm

REGIONAL SETTING

Figure 1. West Salt Creek Landslide (WSCL) map area shown in eastern Mesa County.

Map area withinColorado (inset)

WSCL Map Area inPlate 1


13/55



Bulletin 55

the upper 2000 (610 m) o slope and a low escarpmenton the north flank o Grand Mesa where the landslideheadscarp is located. Grand Mesa is a nearly flat-lyingmesa that covers roughly 600 mi2(1,550 km2) and isapproximately 10,500 (3,200 m) above sea level. Directlyupslope (southward) rom this lower escarpment is thewide alpine plateau o Grand Mesa. Above the landslidearea, at an elevation o about 10,000 (3,050 m), the

plateau is locally named Sheep Flats on the 1:24,000-scaletopographic map. Te plateau and edge o the escarpmentis underlain by the Goodenough ormation. Te topographyo this plateau is hummocky and pockmarked with manysag ponds and lakes that reflect localized ground distur-bances related to Pleistocene glacial morphology and tilldeposition, ground movements within the Goodenoughormation, and colluvial sediment deposition. Further tothe south and southwest, Grand Mesa is capped by 10-million-year-old basalt flows. Te high point o the mesais 6 mi (10 m) due south o the WSCL, at Leon Peak, wherethe elevation is 11,236 (3,425 m).

Te climate o the area is considered semi-arid but variesconsiderably with elevation. Te average high and lowtemperatures or the month o May in Collbran are 70 and40 F (21 to 4 C), respectively, and the average monthlyprecipitation or May is 1.6 in (4.1 cm) (Western RegionalClimate Center, 2014). Vegetation types within the WestSalt Creek drainage basin include Gambel oak, spruce, andsubalpine fir and aspen orests. Average annual precipita-

tion at the valley floor at Collbran is only 15 in (38 cm).o the south on the mesa above the headscarp o thelandslide, some 2,500 (762 m) higher in elevation, thereis annual winter snowpack and the yearly precipitationaverage is over 30 in (76 cm). Te landslide is on the southside o the greater Plateau Creek valley. Te north-acingslopes in this area retain higher moisture and are moreheavily vegetated compared to south-acing slopes on theopposite side o Plateau Creek valley.


14/55



Bulletin 55

EVIDENCE OF PAST LANDSLIDE

ACTIVITY

Landslides have been mapped in this area by previousinvestigators and as part o this study (see geologic map oWSCL vicinity in Plate 1). Colton and others (1975)mapped landslides at a scale o 1:250,000 over the entireSheep Flats plateau region (discussed previous chapter)and small portions o the lower escarpment where theWSCL headscarp is located. Ellis and Freeman (1984)described the colluvium (Qc) covering the plateau above asbeing ormed by landslides, slumps, earthflow, mudflow,soil creep, and solifluction. Soule (1988), in his regional1:24,000-scale landslide study o the Vega Reservoir area,mapped three landslides within the WSCL extent. For ourstudy, an evaluation o pre-landslide topography and stereoaerial photography revealed an amphitheater at the head othe West Salt Creek valley below the plateau rim escarpment.

Tis bowl was the location o earlier significant,post-glacial landslide(s) (Figure 2) that will be reerred toas the prehistoric landslide in subsequent discussions.

Many subparallel intermediary scarps and ground-sagdepressions above the amphitheater rim, shown in thePlate 1 geologic map, also indicate the potential or utureretrogressive ailures. Te topographic bench shown inFigure 2b was ormed rom the prehistoric landslide andwas the location o the landowners down-valley huntingcamp. Te atal May 25th event occurred at the samelocation o these earlier ground movements. Subsequentdiscussion will illustrate that the WSCL upper block ailuresurace approximated the prehistoric ailure surace(Figures 3a and 3b). Recent shallow ailures were alsoidentified along the east flank o the prehistoric scarp.Post-event interviews with landowners revealed that thishistoric landslide occurred in 1984 when a cattle trail waslost.

Figure 2. Outline o the WSCL is shown by solid red line. Sectiongrid lines on USGS 1:24,000-scale topographic map are 1-mile

square. Approximate boundary o pre-historic landslide is shownby black dashed line.

a


15/55



Bulletin 55

b

Figure 2(c). South-acing view o the WSCL with the approxi-mate location o May 25th precursor landslides (black dashedline) that extended to the beaver ponds. Te approximate

extent o the precursor landslides were outlined by an eye-witness who saw them on morning o May 25th. Pre-event

terrain models were created with ArcScene 10.2 using 2012,9-in resolution color aerial photography rom Mesa County

GIS Department draped on a FEMA 4-m DEM. Red star is theassumed location o the missing pick-up truck at the end o theranch road.

c

Figure 2(b). Approximate location o a prehistoric (Holocene)landslide (white dashed lines) and an ancient scarp (hachured

solid white line) superimposed on a 2012 aerial photograph.


16/55



Bulletin 55

Figure 3. Pre- (top) and post-WSCL (bottom) images created rom draping high resolution aerial photography onto DEMs.op uses 2009 NAIP image and FEMA 4-m DEM as elevation base. Red line shows the WSCL extent. Dashed white linedelineates runout o prehistoric landslide. Yellow dashed line shows prominent intermediate scarp block that has controlled

the paleochannel o West Salt Creek (dashed blue line) that doglegs against it to ollow a prominent shear zone on the paleo-bench beore cascading down an earlier escarpment. Bottom image uses 1-m DEM rom CGS LiDAR data and early post-

WSCL aerial photo rom Google Earth where the sag pond (showing as muddy water) was small and still filling. Prehistoriclandslide is shown by dashed white line. Yellow dashed lines mark steep scars in block ace that were likely avalanche detach-

ment zones. White arrows show location o small 1984 landslide on both images.

Prehistoric

Bench


17/55



Bulletin 55

LANDSLIDE MODES OF FAILURE

Data Used in AnalysisGeologists rom the CGS and the USGS mapped the

landslide in the summer and all o 2014. Te timeline andflow mechanisms o the landslide provided here are basedon:

Seismic records of the event;

Eyewitness accounts;

Comparing pre- and post-landslide surface elevationswith high-resolution DEMs and stereo aerial photo-graphy;

Evaluation of cross-cutting relationships of individualcascading surges preserved in the landslide morphologyand the material characteristics o the landslidedeposits;

LiDAR elevation dataset of the landslide, acquiredon June 1-3, 2014, and a 1-m DEM created rom thedataset;

IfSAR DEM (4 m) of the terrain prior to the landslide;

High-resolution orthorectied and georeferencedaerial photography

According to provisional SNOEL data, the spring 2014snowpack was near to slightly above average on Grand Mesa(NRCS, 2014). However, residents reported significantrainall prior to the landslide event. Coe and others (2014)reported the WSCL occurred 3.25 hours aer a 30-minuterainall event with an intensity o 21 mm/h (3/4 in/h).National Oceanic and Atmospheric Administration(NOAA) provisional data indicates that precipitation or

the period o May 12th to June 10th was almost 200% oaverage (NOAA, 2014). Tis departure rom average wasin part due to a wet snowstorm the area experienced May10-11, two weeks beore the landslide occurred. Localresidents reported over a oot o heavy wet snow ell on theplateau above the landslide scarp.

TimelineTe timing o the landslide activity on May 25th, 2014 is

best constrained by two seismic events that were recordedon local seismometers. Te first and smaller event occurred

at 7:18 AM on May 25th and registered as a small seismicdisturbance on the North Fork Valley Network (D. Wolny,Colorado Mesa University, written communication, 2014).Te seismometer is located 20 mi southeast o the landslide.Tis seismograph may have recorded an initial precursorlandslide that was noted by local ranch owner Melvin SlugHawkins. Early that morning, he observed landslide activityat the prehistoric landslide on the east side o the head oWest Salt Creek. Mr. Hawkins recalled that a second smaller

precursor landslide also occurred on the west side o thecreek later that morning. Te approximate sizes o thesemorning precursor landslides are shown in Figure 2c.Te traces are based on Mr. Hawkins recollections whileexamining oblique 3-D block diagrams with draped aerialphotos o the terrain taken beore May 25th, 2014.

Following these precursor landslides, water in irrigationditches originating rom West Salt Creek stopped flowing.Mr. Hawkins stated that he went up the valley and heardtrees snapping and toppling, possibly as a result o a mud/debris flow occurring high on the valley floor in the areao the beaver ponds (Figure 2c). In the aernoon, WesHawkins, Clancy Nichols, and Danny Nichols drove up theWest Salt Creek Road to assess damages to the irrigationdiversion structure and whether there was a landslide riskto Salt Creek Road (CR 60 Rd.). Te assumed last loca-tion o their pick-up truck is shown in Figure 2c where theWest Salt Creek ranch road ends.

On May 25th at 5:45 PM, a magnitude 2.8 earthquakewas recorded approximately 8 mi southeast o Collbran (DWolny, Colorado Mesa University, written communication2014) with a duration o approximately 3 minutes (Figure 4).Te overall appearance o the wave orm was indicative o

a surace event created by landslide-related movement inthe West Salt Creek area (USGS, 2014).

Te catastrophic ailures o the prehistoric landslideblock at the valley head maniested itsel as a rock ava-lanche that buried the three men with over 123 (~38 m)o landslide debris. Te rock avalanche flowed down thevalley and the valley-constrained toe encroached upon anOccidental Petroleum Corporation well pad where threeproducing wells are located (Figure 5). Te shut off valvesat the wellheads were not impacted but there was slight


18/55



Bulletin 55

64.0

64.0

32.0

16.0

16.0

80.0

80.0

48.0

48.0

2 3: 43 :4 2 2 3: 44 :0 2 2 3: 44: 22 2 3: 44 :4 2 2 3: 45 :0 2 2 3: 45 :2 2 2 3: 45 :4 2 2 3: 46: 02 23 :4 6: 22 2 3: 46 :42 2 3: 47 :0 2

Z

32.0

mv/s

Figure 4. Colorado Mesa University Seismic Network seismogram o the WSCL event. Seismometer was 24.7 miles away attheir Rapid Creek Station (courtesy o Dave Wolny, Colorado Mesa University). Note the spikes in the wave train that may

indicate episodic ailures o the rockmass.

Figure 5. Oblique photo o the WSCL taken May 26th, 2014. Te red star is the assumed truck location o the three men killed

in the landslide. It is approximately located at the end o the West Salt Creek ranch road, which orks off at the ront o thewell pad.

1,700ft

Oil and GasWellheads

(520 m)

2.8 mi(4.5 km)

660

(200 m)


19/55



Bulletin 55

damage to surrounding inrastructure. Soon aer theevent, Mesa County received 911 calls rom local residentsbut there are no living eyewitnesses to the actual rockavalanche. According to sheriffs reports (reproduced intheir entirety in Appendix A), local residents first on thescene indicated that the landslide was essentially in placebut material was creeping to a stop rom the time the 911call was made until the arrival o Sheriff s deputies about

1 hours later.Limited search and rescue efforts were made that

evening. A helicopter was called on May 26th and rescueoperations resumed that morning. Representatives o theMesa County Road and Bridge Department, NOAA, andCGS made a cursory inspection and assessment o thelandside by air later that day while search and rescueoperations continued. Investigators noted that springrunoff was flowing at an estimated 50 cs (1.4 m3/s) intothe landslide as wateralls over the head scarp into thelarge depression ormed by the rotated block and that a sag

pond was beginning to orm. Many seeps rom the rockace exposed in the headscarp were also noted. Te MesaCounty Sheriff s office used UAV flights to search or themissing men and their vehicle, and acquire low-altitudeaerial photography o the landslide. At 6:00 PM on May26th, as the sag pond was enlarging and stability o theoversteepened ace o the rotated block was questionable,search efforts by oot on the landslide were suspended orsaety reasons.

DescriponTe entire landslide covers an area o 0.95 mi2(2.46 km2).

Tis includes the exposures o bedrock at the flanks andhead scarps. Te landslide is 2.80 mi (4.51 km) long(horizontal distance) and 2,200 (670 m) in height, roman elevation o 9,600 (2,926 m) at the landslide crown(undisplaced zone above the headscarp) to 7,400 (2,256 m)

at the toe. Te maximum width is 3,113 (949 m) at thesurace o rupture. In the series o events, approximately39 million yd3(30 million m3) o rock and debris weremobilized at the WSCL, calculated using a comparativeGIS analyses o a pre-slide, 4-m resolution DEM, and a1-m resolution DEM created rom post-event LiDAR.Tis volumetric estimate is based solely on changes in thesurace topography and does not account or the geometry

o the basal shear surace and overall thickness o theUpper Block since the depth to the ailure plane is unknown.Using an assumed circular-slip ailure plane shown in themap plate 1 cross section, the total displacement volumemay be as high as 82 million yd3(63 million m3).

Te landslide includes two undamental zones: a zoneo depletion and a zone o accumulation (Varnes, 1978).Tese zones define the transition o the WSCL rom a rota-tional rockslide to a rock avalanche. Te zone o depletionis within the morphologic bowl or amphitheater wherethe landslide slid rom, which includes the depression below

the headscarp and the rotated, partially disaggregated,Upper Block. Te zone o accumulation is the resultingrock avalanche debris field that partially filled the WestSalt Creek valley.

HeadscarpTe near-linear headscarp is 2,070 (631 m) long, 450

(137 m) high, and exposes strata o the Parachute Membero the Green River Formation below interfingered sand-stone beds o the Uinta Formation. Te lateral side scarpscurve outward to a point 1,500 horizontal (457 m) romthe headscarp crown, where the maximum rupture widtho the Upper Block is reached (see geologic map in Plate1). A major slickenside o the landslide shear plane wasobserved on the headscarp rock ace (Figure 6a). Teheadscarp extends upward and above the WSCL crown, toa heavily treed steep slope; this was also the location o the

aFigure 6(a). View o the

east flank o the WSCLheadscarp rom the Upper

Block. Black arrow showlocation o slickensideremnants o the landslid

ailure surace. Tis slipsurace remnant has sub-

sequently allen into thesag pond. Note downed

tree orientations towardthe scarp on rupturedsurace o the back tilted

Upper Block.


20/55



Bulletin 55

prehistoric landslide scarp (Figure 2). Vegetation andglacial/colluvial sediments obscure the Goodenoughormation but ormational clay shale is exposed in the eastravine above the headscarp. Isolated zones o chertylimestone rocks rom the ormation are also poorlyexposed near the plateau rim above the headscarp. Testrata exposed in the wall o the headscarp show evidenceo disturbance, including: localized tilted beds, structural

offsets along subparallel intermediary scarps, and manysub-vertical to steeply-dipping to-the-south shear suraces.Tese orm an overall sheared abric in the rock mass bestseen in the exposed strata o the east flank o the heads-carp (see Figures 6b and 6c). Te exposed strata along thelateral scarps were very thinly bedded to laminated, brittle,

and highly ractured with prominent blue-gray stainedsuraces (Figure 6d). Active seeps in the rock aces o thehead scarp and flank scarps continued or several daysaer the landslide.

Upper BlockTe Upper Block is the remnant o the prehistoric

landslide o the same area (Figure 3). On May 25th, 2014,the Upper Block rotationally slid down this pre-existinglistric (concave upwards) surace o rupture related to theprehistoric landslide. Te rotation and back tilt o the blockdisrupted the block surace into many fissured domainswith varying declinations, typically between 12 and 15toward the headscarp. Concurrently, the near vertical

c

bFigure 6(b). Photo o theeast side scarp. Disturbed

rockmass is the GreenRiver Formation. Note

sheared rock abric andlandslide displacement a

side scarp shown by redarrow. Colluvial red soilis also seen in the scarp

along lef edge o photo.Red dashed line approxi-

mates major near-verticashear zone. Dashed oval

marks tightly spaced,near vertical shearing inrockmass.

Figure 6(c). Photo is aclose up o east side o

headscarp (courtesy oRex Cole, Colorado Mesa

University). Note steepshear suraces shown

by arrows. Dashed redline and oval is in samelocation as shown in 6b.

Seepage flow on rock aceto lef is occurring rom

another shear zone. Tesame location is shown b

black arrow in 6b.


21/55



Bulletin 55

shear planes in the disturbed rock mass (Figure 3, top),which were dipping about 75 aer the prehistoric land-slide, were urther rotated an additional 12 to 15 andestimated to be dipping steeply 55 to 70 to the north.

Te large closed depression behind the Upper Block had

a maximum depth o 110 (34 m) (Figure 7). Te joltand resultant ground disturbance rom down droppingand back tilting, as the block slid down >450 vertical (137 m) along the scarp, caused extensive ground-suraceruptures, heaves, and subparallel fissures. Most o the treeswere tilted or downed and thrown back towards the scarpace. Much o the broken ground on the eastern hal o theblock exposed reworked and transported reddish-orangecolluvial/landslide soil with abundant Grand Mesa basaltboulders that were originally deposited along the east rimo the amphitheater (Figure 8a).

Tis ruddy-colored material became an importantmarker or distinguishing the different source areas andrelative timing o the rock avalanche deposits. Heaved soilexposures also revealed alluvial sediments (Figure 8b)deposited on the original bench aer the prehistoriclandside occurred. Te highly disturbed, steeply-slopingdownslope ace o the Upper Block, also known as thearea o evacuation is where the rock avalanche originat-ed. Several intermediary, steeply-dipping aces o the ronto the Upper Block likely reflect deeper shear-slip ordetachment surace remnants (see Figure 3 bottom) along

this escarpment ace. Some o the later slip ailures wererelatively thin, evidenced by shallow slumps o soil withallen trees.

Rock AvalancheTe extent, approximate thickness, and morphology o

the rock avalanche is shown in Figure 9. Te volume orock avalanche debris that filled West Salt Creek valley, cal-culated rom the base o the Upper Block is approximately17 million yds3(13 million m3) using comparative analyti-cal tools in Global Mapper GIS soware. Te GISanalysis also indicates rock avalanche deposit thicknessesup to 123 (~38 m) along the thalweg o the originalstream channel. Tis event was extremely rapid, witheyewitness accounts indicating the majority o the rockavalanche deposit was in place over two miles down thevalley in a matter o minutes (Melvin Slug Hawkins, oralcommunication, 2014). Based on the seismic data shown inFigure 4, the rock avalanche occurred rom a 3-minutesequence o ailure and rotation o the Upper Block andepisodic rock mass ailures rom the zone o evacuationalong the ront o the Upper Block as it rotated, ragment-ed, and large blocks slipped down the very steep shearsuraces. Initial pulses o disaggregated and pulverizedrock and related soil and vegetal debris had the volumeand velocity to overtop and spillover two ridges o theWest Salt Creek valley where the creek thalweg curves.Tese ridgelines, shown on the geologic map in Plate 1, arereerred to as the West Ridge and the East Ridge.

Figure 6(d). Exposed Green River Formation on west flank oscarp. Note blue-gray staining and degree o racturing in the

rockmass. Photos taken summer o 2014.

All Figure 6 photo locations are shown on Figure 9 map.

d


22/55



Bulletin 55

Figure 8. Upper Block sediments ruptured and heaved by May 2014 landslide movements. a) red colluvium with basalt clasts,b) Alluvium deposited on original Upper Block bench when prehistoric sag eature in-filled afer ailure in the recent

geologic past. Photos taken summer o 2014. Photo locations are shown on Figure 9 map.

Area ofExcavaon

~2,900 f

(885 m)

~450 f

(137 m)

1,2501,0007505002500

Meters

Elevation(M)

2,850

2,950

2,700

2,650

2,800

2,750

2,600

No vertical exaggeration

2,900

?

?

Figure 7. Photo at top is an oblique east-acing aerial photo o the Upper Block taken May 26th the day afer the landslide

beore the sag pond began to fill. Lower image is a cross section at the center o the block at the approximate linear scaleo the photo showing the assumed rotational ailure surace as a dashed line. Dark green area is the post-ailure surace

afer pond began to fill and material sloughed into depression rom the scarp. Light green area is the pre-ailure surace.

ba


23/55



Bulletin 55

Figure 9. Material difference model. LiDAR hillshade is used as the basemap. Location o photo figures inthis report are shown by points or view-direction arrows.

d

Map of the West Salt Creek Landslide ShowingAreas of Excavation and Accumulation

Fig 5

Fig 10i

Fig 10f

Fig 10h

Fig 15

Fig 22

Fig 10b

Fig 10e

Fig 10a

Fig 6aFig 6d

Fig 6b

Fig 7

Fig 8Fig 6c

Fig 17a

Fig 17b,c

Fig 18

Fig 10d

Fig 23

Fig 21

Fig 10c


24/55



Bulletin 55

Te rock avalanche deposit is chaotically mixed, com-posed predominantly o disaggregated and pulverizedmudstones and marlstones o the Green River Formationwith clast sizes that grade rom very large boulders, somein excess o 8 , down to rock flour. Te much slower, latersurge o the rock avalanche along the thalweg o the creekwas much coarser, containing very large, disturbed blocks(car to cabin sizes) o Green River Formation that ormed

a hummocky topography that can be seen in the LiDARhillshade base maps in Figures 9 and 11. Tere was noevidence o a dust cloud generated by the rock avalanche. Aqualitative assessment o the surace exposures o the rockavalanche debris shown in Figure 10images seems toindicate the majority o the blocky, angular to subangular,rock clasts in the debris are generally less than smallboulder size with a high percentage o the rock ragmentsalling within a fist to ootball size (310 in; 825 cm)range. However, there are many areas o the debris fieldthat does have segregations o very large, boulder-sizedclasts. Many clast suraces have the original staining seen

in outcrop near the headscarp (Figure 6d); while specula-tive, this range in clast size may be indicative o the racturenetwork and preerential gradation in the disaggregationo the rock mass. Accessory clasts types include raresandstone and limestone, chert, and basalt bouldersderived rom alluvium, colluvium, and glacial till depositsthat were originally deposited on the plateau above.

Depending on the original source area, the rock avalanche

debris color may range rom shades o light gray to gray todark red and there may be compositional differences inaccessory soils, orest duff, vegetation, trees, and mud. Inmany locations, there are segregations o downed trees andshattered woody material that may reflect a deposit rich innear-surace material. Other surace areas o the debrisfield are 100% Green River Formation clasts without afine-grained matrix. Deposits rom the early surges tendto be gray in color, more rich in rock ragments andcontain much less clay in the matrix.

Figure 10. Rock avalanche debris locations: (a) typical matrix o avalanche deposit. Note pen or scale; (b) typical rocky debris fieldon west ridgeline. Note red-soil zebra-stripes related to spread o surficial soils on east flank; (c) subsidence fissures in deposit; (d)

bouldery segregation within late-surge hummocky rock-debris flow; (e) debris on rise o west ridge spill over, red circle enclosesperson or scale; () east edge o debris field near toe. Photos taken summer o 2014. Photo locations are shown on Figure 9 map.

dc

ba


25/55



Bulletin 55

fe

Figure 10 (con't). Rock avalanche debris locations: (g) debris field, photo taken rom West Ridge towards East Ridge. Late surgehummocky rock-debris flow can be seen in center; (h) view up slope rom flank o East Ridge, note spill over at West Ridge,

dog-leg in East Elk Creek thalweg, and late surge hummocky rock-debris flow on floor; (i) view up slope in toe area, noteproximity to oil and gas well inrastructure.

g

h


26/55



Bulletin 55

Figure 10 (i) View up slope in toe area, note proximity to oil and gas well inrastructure.

i

In later surges derived rom the east side o the amphi-theater, the deposit contains remnants o the distinctivered-brown, basalt-rich colluvial soil. Also, late detachmentand slips o large rock blocks, likely rom deeper and lowerin the excavation zone o the Upper Block, did not havethe energy to completely disaggregate. Tese very largeblocks o rock, some the size o small homes, while highlydisturbed and broken, remained intact within the lastslower surge o the rock avalanche flow and ormed thehummocky topography along the valley floor.

Variations in velocity, deposit abric, and soil remnant/clast composition can be assigned to differing surges o therock avalanche. Tose variations and the sequentialtiming o the discrete cascading pulses o rock avalancheand debris flows are discussed in the Rock AvalancheMorphology and iming o Major Deposition Eventssections.

Overall, the rock avalanche was a markedly dry deposit;during the field investigation, no water or mud was seenseeping or anning out rom the steeply sloped edge o theavalanche toe. Te deposit did contain some moist muddyzones and pockets o damp, fine-grained matrix. Alongthe flanks, at shear ridges, and at the toe o the avalanche,

the 20- to 35-thick rocky deposit was able to supportslopes up to 45, which gave the deposit a morphologysimilar to a lava flow at the toe. Te toe o the rockavalanche remained dry until it was observed that groundwater, percolating through the entire length o the depositand orming a series o ponds in the debris field, finallyreached the toe sometime in December 2014. During thespring 2015 run-off, a re-established stream now flowsrom these larger ponds, and is incising the landslide toe asthe stream re-enters the original West Salt Creek bedbelow.

Esmated VelocityTe velocity estimates presented here are preliminary andare provided or discussion purposes with the hope they canbe used to better refine uture estimates. Early estimates,using the standard velocity equation v=d/t, were based upona run-out distance o 2.8 mi (4.5 km) (the complete extent othe landslide rom headscarp to toe) and duration o almost3 minutes. However, aer completion o mapping and fieldreconnaissance, the CGS reconsidered these variables,specifically the sources and extents o the mapped deposits.Te precise area o origin, within the evacuation zone o theUpper Block, or the rock avalanche debris is unknown and

the run-out distance or individual surges is complex. Tus,the CGS chose an initiation point, at the evacuation zonealong the ace o the rotated block (Figure 11), at theapproximate location o a scarp o more intact rock relativeto the majority o ragmented rock below the scarp.

wo end points in the rock avalanche debris field wereconsidered (Figure 11). Te first (labeled End Point 1) isthe urthest mapped extent o the early surge deposit. Tetotal travel distance rom the initiation point to this EndPoint 1 is 2.36 mi (3.80 km) and the total vertical drop is1,362 (415 m). Te second possible end point (labeled

End Point 2 on Figure 11) is where the gray brecciaprobably impacted the valley side as part o it came downthe first surmounted hill and this is the approximate lastoccurrence o the middle surge red basaltic breccia. Tetravel distance to End Point 2 is 2.12 mi (3.41 km) and thetotal vertical drop is 1,267 (386 m).

Te event as recorded on the seismograph, began at17:43:45 with a duration o approximately 3 minutes. Arudimentary velocity calculation using v=d/t where d isdistance and t is time, assuming the shortest distance o2.12 mi (3.41 km), yields


27/55



Bulletin 55

2.12 mi/0.05 hrs = 42.40 mph (18.95 m/s)

Alternatively, using the last known undisturbed locationo the gray breccia at 2.36 mi (2.80 km) rom the potentialsource area, yields

2.36 mi/0.05 hrs = 47.20 mph (21.10 m/s)

Tese velocity estimates should be considered mini-mum values or the early surge o rock material, basedupon current knowledge o the seismic wave duration andbeginning and end points o the debris discussed above.

Another velocity estimate can be calculatedrom the height o the first surmounted hill,which was measured to be approximately 280 (85 m) high, by using the velocity potentialenergy equation (Chow, 1959):

v = 2gh

wheregis the acceleration due to gravity(9.8 m/s) and his the height gained, giving:

v = 2 x 9.8 x 55 = 40 m/s or 89 mph

A second hill with a height o 180 (55 m)

was also surmounted:

v = 2 x 9.8 x 85 = 33 m/s or 74 mph

However, this equation assumes that energywas completely kinetic and converted topotential energy and the surmounted objectmust be perpendicular to the flow(Jakob, 2005). It should be noted that i thesurge approaches the hill obliquely, theestimate o the velocity to top the hill is a strictminimum, because the velocity measured by

the velocity potential energy equation is only the compo-nent perpendicular to the hill. Furthermore, it could alsohave a component o undetermined size parallel to the hillcrest, so the flow velocity must be even higher, which isnecessary to supply the gravitational energy to overtop it.Because the calculation does not take into account bendsin the stream or the trim lines o correlative deposits, theestimated velocities are approximate and only give thevelocity needed to run up the given height.

Te preerred method to account or the channel bendsand the banking angle o the deposit at the channel bend is

Figure 11. Hillshade o the WSCL with data

points used in simplified estimated velocitycalculations. Te Initiation Point (white star)

or the rock avalanche deposits is at the approxi-mate location o a scarp consisting o more intact

bedrock relative to the majority o brecciateddebris below the scarp. End Point 1 (red circle)corresponds to urthest mapped deposit o a gray

polymict breccia (part o early Surge ) and End

Point 2 (yellow circle) is where early surge mayhave impacted the valley side as it came off thespillover and where the red basalt-rich deposits

end.

End Point 1

End Point 2

Initiation Point


28/55



Bulletin 55

the orced vortex equation (Chow, 1959; Henderson, 1966;Hungr and others, 1984; and Johnson, 1984). Discussiono this method is provided in Prochaska and others (2008)and it has been used to back calculate rock avalanchevelocities i these actors are known (Plafer and Erikson,1975; Evans and others, 1989, 2001; Boultbee and others,2004). Tis equation is as ollows:

v = Rcg tan

Where Rcis the channels radius o curvature,gis theacceleration due to gravity and is the banking angle othe deposit rom horizontal.

Te radius o curvature (Rc) wasdetermined by drawing circles that ap-proximated the channel curvature throughat least 3 points along the stream courseusing 1:24,000-scale topographic mapsand a pre-slide 4-m DEM (Figure 12).Tree cross-section profiles (locationsalso shown in Figure 11) were placedperpendicular to the channel bend andwere plotted or each analyzed bend usingthe pre-slide 4-m DEM and the post-slide1-m LiDAR DEM. Te three generatedcross sections are shown in Figure 13.Te banking angle () was determined byapproximating the trimline o correlative

deposits and measuring the angle rom horizontal.

able 1summarizes the measured variables. Te calcu-lated velocity o 89 mph or the first surmounted hill usingthe velocity potential energy equation compares avor-ably with the velocity o 83 mph or the same hill usingthe orced vortex equation. However, the orced vortexcalculation or cross section A-A' likely underestimatesthe velocity because the inner trimline o the Early Surge

is either lower in the valley or obliterated by later surgesthat were higher than the first surge. Similarly, the lowertrimline o the first rock avalanche in cross section B-B'

Figure 12. Hillshaded LiDAR DEM o thearea o the WSCL analyzed or velocity

estimates. Best-fit circles were used todetermine the radius o curvature (Rc) orchannel beds. Alignment o cross-section

profiles are labeled A-A', B-B', and C-C'.

Section Rc(m) Banking angle () tan Maximum velocity in(degrees from m/s (mph)

horizontal)

A-A' 739 11 (?) 0.19 37 (83)

B-B' 512 4 to 9.5 (?) 0.07 19 (42) to 29 (65)

C-C' 432 1 0.02 9 (21)

able 1.


29/55



Bulletin 55

(Figure 13) is also unknown and the calculated velocitiescan vary widely based on the banking angle.

Te calculated velocities rom the velocity potential en-ergy equation (based on run-up height) and orced vortexequation (based on channel bend and trimlines) all in therange o velocities given or other rock/debris avalanches:105 mph (47 m/s) at Nevados Huascaran in 1962 (Plaferand others, 1971), 86 mph (38 m/s) at Huancavelica, Peru(Kojan and Hutchinson, 1978), 77 mph (34 m/s) at Mt.St. Helens (Schuster, 1983; Voight and others, 1983), and

76 mph (33 m/s) at Zymoetz River (Boultbee and others,2004). While it is likely that the velocity o the materialwould have varied within the flow zone and the physicalcharacteristics o each surge differ, velocities between 40mph (18 m/s) and 85 mph (38 m/s) are plausible or therock avalanche.

We acknowledge that velocity estimations are variableand additional details about this event may refine the ve-locities provided here. Factors such as the precise point oorigin or the measured surges, variations in channel bed

8,500

7,750

7,500

8,250

8,000

2,300

2,500

2,400

C C'

8,500

ELEV IN

FEET

7,750

8,250

8,000

2,500

2,400

B B'

8,500

7,750

8,250

8,000

2,500

2,400

A A'

ELEV IN

METERS

Figure 13. Cross-section profiles created rom pre- and post-event DEMs. Location o the cross-section alignments are shown

on Figure 12. Te original topography is indicated by the green line and the landslide debris topography is shown by the redline. Dotted black line is the banking angle as determined rom trim lines o probable correlative deposits. wo banking angleswere proposed or cross section B-B'. Te dashed black line is a very conservative estimate rom hilltop to the widest extent o

deposits. Te dotted line was determined rom the angle o the first surmounted hill and the approximate location where a par-allel line would have intersected the valley below.


30/55



Bulletin 55

sinuosity, channel depth, precise superelevation o correla-tive deposits, and deposit engineering properties all actorinto the velocity calculations and with more investigationwill help determine a more precise velocity value.

Landslide MobilityWhile the actual range o velocities is not precisely

known, the WSCL was catastrophic in nature and was wellabove the minimum speed or the highest, extremely rapid,classification (>5 m/sec or about 11 mph, the approximatespeed o a person running) (Cruden and Varnes (1996).Te run-out extent (2.8 mi, 4.5 km) o the deposit also re-flects an extremely high mobility. A simple mobility index

is the H/L ratio: the height o the landslide by the lengtho the landslide, also reerred to as the riction coefficient.Tis value is also the tangent o the run-out angle o reachrom the crown to the toe o the landslide. Coe and others(2014), in an early preliminary approximation, stated anH/L ratio o 0.10.

Based on detailed GIS-based measurements, using aheight o 2,100 (640 m) and length o 14,750 (4,496 m)

along the thalweg o the stream, the H/L ratio is nowcalculated at 0.14 and the run-out angle o reach(ahrbschung) at 8. Figure 14contains two graphs with

points plotted o the WSCL data. In these graphs, we used

0.02

0.01

Volume (km3)

10-1 100 101 102 103 104 105 106 107 108 109 1010 1011 1012 1013

0.03

0.040.05

0.2

0.1

0.3

0.40.5

2.0

1.0

3.0

4.05.0

tan 40

tan 30

MobilityIndex(H/Lortan(angleofreach))

tan 8

WSCL

3X107m3Oso, WA

8X106m3

Extraterrestrial(Melosh 1987; Lucchita 1978)

Kaolinised granite(Hutchinson 1988)

Chalk(Hutchinson 1988)

Coal mine waste rock(Golder Associates 1992;

Hutchinson 1988)

Rock avalanches(Sassa 1988; Cruden 1976,

Li Tianchi 1983: Lucchita 1978)

Retaining walls

Fills

Cuts

Corominas (1996) datamean

95% Confidence limit ofCorominas (1996) data

0.03 Km3

4.5 Km

Oso,

WA

WSCL

Figure 14. Graphic representation o volume

angle o reach, and runout lengths o largecatastrophic landslides and rock/debris ava-

lanches. op graph is a Log mobility indexvs. Volume (V) graph modified rom Finlay

and others (1999) compilation. Lowergraph is a runout length versus volume chartrom the GEER Association report (Keaton

and others, 2014) o the Oso landslide thatshows rock avalanches and debris flows

compiled by Legros (2002). WSCL pointis the star. For comparison, the Oso land-

slide point (Iverson, 2014) is shown as blackcircle.


31/55



Bulletin 55

the volume o the landslide deposit calculated rom vol-ume differences between a pre- and post-landslide DEM- 39 million yd3(30 million m3). However, using the rockavalanche deposit alone, approximately 17 million yds3(13 million m3), would result in a slight shi o the plottedpoints on the logarithmic scale.

Te top graph, with compiled data rom Finlay andothers (1999), illustrates the low mobility index and theflat angle o reach o the WSCL. Conversely, this indi-cates an extremely high mobility; one o the highest orits comparative volume o material when plotted within

Finlay and others (1999) compilation. Te second lowergraph o Figure 14 contains a compilation o data romLegros (2002) that shows the relationship between volumeand runout length. Tis graph does not take into accountlandslide height and the WSCL plots within the generaltrend o rock avalanches. Also included in Figure 14 is acomparison point rom the March 22, 2014 Oso landslidethat caused 43 atalities. Tis smaller but extremely rapid,

high mobility, catastrophic mud and debris flow occurredin Washington State (Iverson and George, 2014; Keatonand others, 2014).


32/55



Bulletin 55

Mapping by the CGS and USGS using hillshadedLiDAR data, and both oblique and orthorectified aerialphotography, documented many different landslide ormsthat indicate overlapping, crosscutting, and episodicdeposition. Tese eatures include:

discrete thrust and overlap eatures wheresubsequent flows surged over earlier deposits or thevelocity allowed later surges to overtake earlier oneswhile they were still moving;

discrete run-up and surge overlaps onto and overexisting topographic ridgelines;

zebra-striped soil deposits o alternating redand gray material rom the surace spreadingo mobilized material;

v-shaped depositional ridges that point uphill on theeast, overtopped ridgeline, which mark where latermiddle-surge flows stopped on the ridgetop butcontinued to flow around and down the ridgeline;

parallel-to-flow depositional lateral levees where

flow edges slowed and stopped;

steep parallel-to-flow-direction urrows and ridgelines at lateral shears along cross-cutting surgeboundaries (Figure 15); and

pressure-ridge flow banding at the toe.

Te morphology o the rock avalanche deposits indicatethat the event occurred as a series o cascading surges thatreflect rapid, but episodic ailures o the rock mass withinthe evacuation zone o the lower Upper Block. Teseepisodic ailures resulted in detachment and disaggregation

o the rock mass, causing stacked surges o rock and debrismaterial to course down the West Salt Creek valley. Teepisodic spikes within the 3-minute seismic wave train inFigure 4 support this conclusion. It was also apparent thatthere are zones in the rock avalanche debris field that weremodified immediately aerwards by slumping and settle-ment. We observed slumps and steep scarps whereoversteepened debris had slid, as well as settlement fissuresand depressions (Figure 10c) that we attribute to consoli-dation or packing o the loose, lower-density debris oncemovement o the rock avalanche ceased. Again, we oundno evidence that the rock avalanche created a dust cloud.

Tere was also no evidence o high heat or used rock(rictionite) that is reported to occur with high-velocitysturzstrom-type deposits (Erismann, 1979; Legros andothers, 2000; Lin and others, 2001).

During the summer and all o 2014, CGS observed thatclasts o the Green River Formation, which comprisealmost all o the landslide deposit, are highly susceptible toslaking when exposed to physical weathering. Te clastson the surace o the debris field rapidly disintegrated intofiner and finer material (Figure 16) that will be easier oruture debris flows to entrain and may increase theirbulking actor.

Timing of Major Deposional EventsTe precursor movements and ailure mechanisms o

the Upper Block ormed the conditions or the rock massailure, disaggregation, pulverization, and cascading flow orock, soil, and orest cover. Tese latter catastrophic move-ments at 5:45 PM completely obliterated and/or buried theprecursor landslides earlier that morning so we must relyon eyewitnesses and the early seismic record o 7:18 AM

(D. Wolny, Colorado Mesa University, written communi-cation, 2014). We also have no account o any precursoractivity at the main landslide scarp.

We have subdivided the run out o the rock avalancheinto three surge periods (see the geologic map in Figure 2)Deposits consistent with smaller sub-events within thesemajor surges were observed and mapped but discussion ocomplex sub-events is beyond the scope o this report. Tethree major surges are based on the morphology, clast, andmatrix variations o the rock avalanche deposits.

In addition, peaks in the seismic wave train (Figure 4)

may indicate larger rock mass ailures that correspondto the discrete surge episodes. O significant importanceis the pocket o red-clay colluvium with abundant basaltrocks that occurs on the east flank o the valley above theUpper Block, and as disturbed prehistoric landslide de-posits on the eastern side o the Upper Block bench in theoriginal pre-May 25th landscape (see Figures 2, 3, and 6b).Te first two surges were the result o high-velocity rockavalanches, while the last deposit occupies only the valleyfloor. Tis last surge was lower velocity and has morpho-logical characteristics o a rocky earth flow.

ROCK AVALANCHE MORPHOLOGY


33/55



Bulletin 55

Early SurgeTe early surge had the velocity to surmount the West

Ridge, overwhelm and cover the East Ridge, and flow downclose to the existing WSCL toe. Tis first widespreadrock avalanche deposit was light-gray, yellowish-gray, andbrownish-gray, poorly sorted, matrix supported, rockydebris derived rom the Green River Formation. Terocky deposit consists o angular to sub-angular granule-,cobble- and boulder-sized clasts o similar composition

to the matrix. Also present are rounded basalt clasts, treeragments, orest duff, and dark-gray to brownish-grayorganic-rich soil remnants. Te fine-grained matrix con-sists o sand- to rock flour-sized pulverized rock ragmentsand wood particles, duff, and dark-gray to brownish-grayorganic-rich soil remnants; some is charred by previousorest fires. Te deposit was derived rom the disintegra-tion and pulverization o disturbed rock mass o the UpperBlock and probably previously existing landslide depositsalong with wet orest slope colluvium, topsoil, and woodydebris.

Tis surge is believed to reflect an initial ailure o theace o the Upper Block where there was little red soil, andhad the volume and velocity to surge over the ridgelineswhere the stream thalweg doglegs down the valley. Mosto this deposit was subsequently buried or incorporatedinto later surges. However, there are remnants o this firstsurge along the east margin in the lower parts o the land-slide, the toe, and where it overtopped the west and eastridgelines.

Middle SurgeTe middle surge is more complex and included dis-

aggregation and rock avalanche flows along the entirewidth o the evacuation zone o the Upper Block. Itreflects ailure(s) o the entire width o the block becauseit includes surges similar to the early surge rom the westside that is gray in color and almost entirely derived romthe Green River Formation, but also includes depositsrom the east side that consist o significant percentageso red-clay soil with abundant basalt clasts. Tis episodealso attained substantial velocity to surge over the two

Figure 15. Aerial photograph at low sunangle showing complex crosscutting

relationship o discrete surges in theavalanche debris field. Note parallel to-flow linear shears and ridgelines at the

boundary o the late hummocky surge.Photo was stitched rom UAV images

courtesy o Ben Miller, Mesa CountySheriffs Office taken 24 hours afer

the landslide occurred. Red dot showsapproximate location o photo in Figure18. Extent o photo is shown by boxed

area in Figure 9.


34/55



Bulletin 55

ridgelines. In the debris field, there is a rough segregationo rock avalanche deposits rom the east side with red soiland west side without.

Middle surge deposits with the red soil can range com-positionally. Tey can be the typical gray rock avalanchebreccia composed predominantly o Green River clastsdescribed above but with scattered surace clumps o redsoil and basalt clasts. Where the surge source includedhigher percentages o the original red colluvial soil, the

deposit can be a reddish-brown to red, clay-matrix supported,and basalt-rich debris; with lesser percentages o shaleand marlstone, rare sandstone, and scattered Miocene (?)Goodenough ormation chert. Basalt boulders can exceed8 in diameter but they are rare. Where there are highpercentages o the red-clay soil, there are correspondingincreases in orest cover material typically incorporatedinto the deposit, such as tree debris, shrubs, duff andorganic-rich soil remnants, and ragments o matrix-supported rock and soil debris that resemble prior land-slide deposits (Figure 17).

Late SurgeTe late surge deposit is a narrower unit (rom 400

to 500 wide) that approximately ollows the originalthalweg o West Salt Creek rom below the Upper Block tonearly the same level as the oil and gas well pad up val-ley rom the toe. Its channelized landslide morphology istypical o rapid-moving earth flows or debris flows and ischaracterized by chaotic hummocky topography that isvisible in the LiDAR hillshade image in Figures 9, 11 andin Figure 15 aerial photo. Te flow is comprised o a mix-ture o the brecciated units described above, smaller local-ized muddy debris flows, plant and tree debris, red soilswith basalt boulders, and most importantly, many very largedisturbed blocks o Green River Formation that have raedon and within the hummocky rock-debris flow. Te local-ized debris flows are reddish-brown in color and appear tohave moved as a thin (less than 3 thick) viscous slurry osurace soil, rock and plant debris deposited in and aroundthe large disturbed blocks.

Figure 16. Rapid slaking o Green River Formation claystone and marlstone clasts in the WSCL debris field. Note surace stain-

ing o clasts and progression, clockwise rom upper lef, o rock disintegration in less than 4 months.

September 2014

June 2014

August 2014

July 2014


35/55



Bulletin 55

Te rock blocks vary in size, approximating the size ocars to small houses. Te blocks are in various stages oragmentation, rom nearly intact with only minor dis-ruption o bedding planes to highly disturbed, almostdisintegrated, where the original bedding is only slightlyapparent. Te blocks commonly orm hummocks in thecenter o the landslide mass and appear to have bulldozedthe underlying deposits into low ridges at the ront and

sides o the blocks. Remains o the original orest floor,with living trees (predominantly conierous pines) andplants in growth position, is preserved on some o thesuraces o the hummock-orming blocks. Tis degree oblock preservation in conjunction with the great distancesthe blocks traveled rom the evacuation zone sourceindicates that near the end o its rotational movements,the Upper Block was detaching blocks o Green RiverFormation rom the evacuation-zone without sufficientmomentum or speed to completely disaggregate the rockmass as what occurred in the earlier surges. It is clear thiswas the last significant movement o the landslide because

the lateral boundaries o its entire length is marked bysharp urrows (Figure 18) that define shear zones as thismaterial flowed through, over, or past the pre-existingearly to middle surge deposit runouts. Te late surgerock-debris flow ground to a halt as it bulldozed earliersurges at the toe near and below the oil and gas well pad.Te late surge boundary was discerned near the toe by theormation o pressure-ridges and flow bands normal to theflow direction and the down-slope cessation o red-soilremnants.

Figure 17. Red bouldery deposits: a) photo taken rom the

east side o the disturbed bench o the Upper Block at red soildeposit; b and c) photos on boundary, upslope and downslope,where middle surge material (with red soil remnants, basalt

boulders, and tree ragments) has overrun light gray earlysurge deposit that is nearly completely composed o finer

Green River Formation clasts. Photos taken summer o 2014.Photo locations are shown on Figure 9 map.

a

b

c


36/55



Bulletin 55

Figure 18. Well-defined side urrow (dashed line) on west side o debris field marks a lateral shear o subsequent movements o

a later, lower level, debris surge. West edge o late, hummocky surge along stream thalweg is shown by dotted white line.Photo taken summer o 2014. Tis photo location is also on the aerial photo in Figure 15 and is shown on Figure 9 map.


37/55



Bulletin 55

Te triggering mechanisms or the WSCL includes thedriving orces or two modes o ailure: 1) the rotationalailure o the Upper Block, and 2) the mechanics thatallowed the disturbed Green River Formation rock mass torapidly disaggregate and create a condition where avalley-constrained flow o rock and debris could run outover two miles on a relatively low gradient.

Rotaonal Failure

Te rotational movement o the Upper Block is themore easily understood ailure at the WSCL rom alimit-equilibrium and kinematic ramework. As discussedearlier, the Upper Block was the remnant o a pre-existingprehistoric landslide and the WSCL movement appears tohave occurred along the same, pre-existing, deeply-curvedshear surace (Figures 2 and 3). Prior to the May 25th WSCLailure, a prehistoric shear plane had already experiencedover 270 (82 m) o vertical movement in the geologicpast and was at a much weakened, residual-shear strengthcondition.

Tree drainage networks also converge onto the UpperBlock rom the plateau above, insuring that ground waterlevels are high in the spring. Te plateau escarpmentreceives a precipitation average o about 30 in per year.Data rom the NOAA regional climate center indicates thataverage precipitation was over 200% o normal or the periodrom 5/12/2014 to 6/10/2014. According to local residents,significant heavy wet snow ell during a Mothers Dayweekend storm. High ground water was also indicated byseeps in the scarp seen or several weeks aer the land-slide. While the precise triggering threshold is not known,there were precursor ailures and slumps along the ront o

the Upper Block recorded by both a seismic station andeyewitness account (Figure 2), and muddy debris flowswere occurring at the toe o the block where beaver pondsexisted. High ground-water levels in the disturbed rockmass, an already weakened shear strength o the slip planematerial, high pore pressure reducing the effective normalstress, and loss o lateral support o the Upper Block rontby precursor landslides earlier that morning all contributedto triggering the slip along the pre-existing, deeply-curved,shear surace and causing the rotational ailure o theUpper Block. Tis 3-minute-duration rotational ailure

occurred at 5:45 PM and was the source o materials thatormed the rock avalanche/debris flow(s) down the valley.

Rock Avalanche and Debris-Flow

Failure(s)Te ailure and run out o the rock avalanche is complex,

and detailed discussion o the mechanical transition o arock slide to a rock avalanche is beyond the scope o thisreport. However, based on field observations, the CGS

believes a number o conclusions and assumptions can bestated:

1. Te disturbed rock mass was thrust out rom thetoe o the Upper Block along an assumedrotational slip plane shown in Figure 7.

2. Te rock mass was composed o brittle, ractured,and thinly bedded shale and marlstone that wasalready disturbed by shearing and earlier landslidingTe original near-vertical joints and shears, nowtilted by rotation o the Upper Block by both the

prehistoric landslide and the WSCL, dip steeply (55- 70) toward the valley below.

3. Very large segments o mobilized rock mass, with alateral velocity component, sheared off the ace othe Upper Block along these high-angle shear zones.While speculative, the seismic waves (Figure 4) seemto indicate that there was a time-relative spacing othe ailure o the rock mass segments and develop-ment o the cascade o rock avalanche debris surgesover the 3-minute period. Te crosscuttingmorphology o the rock avalanche deposits supportsthe theory o discrete avalanche surges.

4. Te sliding disturbed rock mass very rapidly disaggre-gated into the granular material seen in the debrisfield.

5. Tis material, including soil and orest cover, had theability to mobilize rom the evacuation zone o theace o the Upper Block and flow approximately2.2 mi (3.5 km) at an extremely high velocity. Figure13 contains superimposed cross sections that werecut normal to the flow direction o the rock ava-lanche using pre- and post-landslide topography.

TRIGGERING MECHANISMS


38/55



Bulletin 55

Te sections show the comparative elevation differ-ences o the opposing lateral deposit trim lines atbends o the stream thalweg. Te extremely highvelocity o the rock avalanche ront caused the debristo surge up and overtop both ridgelines; the westridge trim line has a differential rise o a minimum206 (63 m), compared to the inside-curve trimlineelevation on the opposing east side.

6. No mud or water flowed rom the steep-walled toeo the landslide. Te dry nature o the deposit wouldimply that, while ground water and pore pressurelikely played a major contributing triggering actoror the rotational ailure o the Upper Block, it likelyplayed a lesser role in the flow o the granular rockavalanche deposit and did not provide the upliingstress (i.e., riction reduction) to allow the avalanche

to achieve such high mobility.


39/55



Bulletin 55

Initial monitoring o thelandslide included water levelgages o the sag pond and thefloor o West Salt Creek 750 (228 m) below the landslidetoe, 5 GPS stations, and 2web-based cameras. Telocations o these tools areshown in Figure 19which also

shows property ownership inthe WSCL area. Followingpeak runoff (late June 2014)the pond level remainedrelatively consistent, indicatingthe lower summer inflow rateinto the sag pond was aboutthe same as the infiltrationrate o water percolating intothe landslide mass and rockavalanche debris field. Duringspring 2015, sag-pond levels

rapidly rose 13 (4 m) beoreslowly dropping in mid June.

MONITORING

Figure 19. WSCL instrumenta-tion and property ownershipmap. Map courtesy o the Mesa

County GIS/I Department.


40/55



Bulletin 55

GPS stations were installed by Mesa County at five loca-tions; three are on the Upper Block within the landslidewhile two are located on the flanks above the headscarp.Vectors o movements through October 1, 2014 are shownon Figure 20(B. F. Kochevar, written communication,2014). Most o the movements outside the landslide appearto be within instrument tolerances. Vertical movementis occurring in the center o the Upper Block, which are

attributed to settlement o the dilated and disturbed rockmass. As o early June 2015, no discernible movements othe Upper Block have occurred.

Te USGS and Mesa County replaced wireless-basedcamera stations in early spring 2015 to conduct remotevisual inspections o the sag-pond and evacuation zone othe Upper Block where the new piping outlet flows water.In videos created by Jeff Coe (USGS), other than minorlocalized sloughing o material in very steep areas o theUpper Block ace, no gross movements o the Upper Bockhave been visually detected.

Figure 20. GPS sensor locations with horizontal vectors through October 1, 2014. Vertical displacements are also shown inyellow text. Lengths are in eet. Image courtesy o Frank Kochevar, Mesa County.

-0.11' vercal

-0.77' vercal

-0.27' vercal

+0.03' vercal

+0.15' vercal


41/55



Bulletin 55

Fortunately, the WSCL occurred in a rural setting anddid not directly damage the town o Collbran or destroynearby residences. However, the community, and localbusinesses were affected personally and economically. Teloss o lie, pasture, access to grazing land, irrigation water,revenues rom reduced tourism, and living with the threato additional flooding and landslides were devastating tothe local community. Oil and gas operations had to behalted and mitigation measures were completed to helpprevent uture damage to the well pad.

Troughout the summer and all o 2014, staff rom theCGS, USGS, and Mesa County observed ponds ormingbelow the sag pond and along the length o the landslide,indicating saturation o the rock avalanche deposit.However, no water was seen seeping rom, or flowing over,the toe o the landslide until December 2014. Te land-

owner reported that surace flow rom the debris-fieldponds had extended to a 4-wheeler track Mesa Countybuilt across the landslide toe just above the oil and gas wellpad (. Hayashi, Mesa County Road and Bridge Dept.,written communication. 2014). In January 2015, 7 monthsaer the landslide event, surace waters began to flow overthe landslide toe into the original channel o West SaltCreek. However, this surace flow is still an accumulationo ground-water seepage in the rock avalanche deposit. Inspring 2015, debris-field ponds on the hummocky floorhave increased in size, some up to 300 (91.5 m) across(Figure 21), which have urther ed the new creek thatflows down the lower hal o the debris field to the land-slide toe where the debris material has slumped andgullied to create a small 50- (15.2 m) long alluvial annear the original West Salt Creek bed.

FUTURE CONCERNS and LONG-

TERM HAZARD ASSESSMENT

Figure 21. A year afer the WSCL, large ponds ormed on the hummocky late-surge rock/debris avalanche deposits. View is to

the northeast rom the West Ridge at the same location o Figure 11g. Note finer-grained debris field in oreground comparedto Figure 11(g), as a result o year-long slaking o Green River Formation clasts. Photo date: June 5, 2015. Photo location isshown on Figure 9 map.

East Ridge


42/55



Bulletin 55

In April 2015, as this report was being completed, MesaCounty Office o Emergency Management prepared a WestSalt Creek Landslide Emergency Action Plan. Tis planincludes an emergency action plan, inundation study, sensorlocation map, a trail map, and communication directory.Te emergency manager also prepared a notificationsystem so the appropriate response personnel would beautomatically called i any o the threat scenarios discussed

below were to occur. In late May and early June 2015, thesag-pond water level rapidly rose 13 (4 m) and, moreominously, a flowing spring had begun to pipe rom thebase o the Upper Block (Figure 22). Saety concernscaused the Emergency Manager o Mesa County to issueLevel I and II alerts. West Salt Creek still flows into the sagpond at the landslide scarp. In early June o 2015, the sagpond level is approximately 2_3 (0.6_0.9 m) belowspill-over elevation. By mid-June the pond level began toslowly drop.

Long-term Stability of the Upper BlockTe disturbed and rotated Upper Block is now resting at

a temporary equilibrium and the existing driving and

resisting orces are at parity. Colorado School o MinesDepartment o Geology and Geological Engineeringconducted limit-equilibrium slope stability analyses o theUpper Block. Since it wasnt easible to complete investiga-tive borings, ailure models were back-calculated rombeore and aer topography.

Sensitivity analyses were completed by varying crosssections through the block, ailure suraces, and groundwater levels in the model. Teir executive summarysuggests the global stability o the rotated block is good,with a 26-53% higher saety actor (SF) than the pre-May25th condition. However, they caution that there is strongsensitivity to water levels, and as the sag pond fills toovertop, the SF could drop as much as 30%, and slopeailures could breach the pond (McCoy and Santi, 2014).Tis became a significant concern in the spring o 2015 asthe sag pond filled to almost the spill-over elevation andthe active pipe began to flow water rom the base o theblock. At present, the global stability o the Upper Block is

unknown, as well as whether any potential ailure wouldbe the entire block, or only a portion as an embankment-type ailure.

Figure 22. spring 2015 opening o natural pipe that has begunto slowly incise into the base o the Upper Block. Inset close-

up photo shows extent o gullying. Photo taken June 6, 2015.Photo location is shown on Figure 9 map.

Head Scarp

Upper Block


43/55



Bulletin 55

Reacvaon and Retrogressive Failure

from above the Landslide HeadscarpTere is a level o uncertainty in the extent o disturbance

and overall stability o the bedrock exposed above theescarpment o the May 25th landslide. Considerableevidence suggests that an extensive disturbed bedrockzone lies around the perimeter o the headscarp where

many geologically recent subparallel escarpments ovarying steepness and vertical slip extent were mapped(see Map Plates 1 and 2). Some are prominent: a narrowremnant o one prehistoric intermediary block that sliddown 110 (33.5 m) remains above the east end o theMay 2014 headscarp, portions o pre-existing eastern flanklandslide scarps were urther ruptured by the May 25thevent, and the newly exposed rock shows evidence o shearin the downhill direction (Figures 6b and 6c). Tesesecondary scarps that outline the perimeter o the currenthead- and lateral-scarps have had the most recent move-ment prior to the May 25th event. At these locations,

exposed bedrock is disturbed and should be consideredunstable. During inspections in June 2015, new smallscarps and widening o existing tension cracks wereobserved at or near the headscarp. One would expectsome dilation o the rock mass at the scarp cliff with theloss o lateral support rom the 450- (137-m) downwardslip o the Upper Block. While there has been smallshallow slips rom the scarp into the sag pond, other thanvisual inspections, there is currently no instrumentedmonitoring o the rock mass exposed along the 2014 scarp.

It is also unknown whether old scarps above the escarp-ment, with slips less than 10 (~3 m), reflect precursordeep-seated shear zones within the Uinta and Green RiverFormations, or are only near-surace ailures within soclaystones o the Goodenough ormation. Because o theorientation o the subparallel scarps to the perimeter o thevalley landslide amphitheater shown on Plate 1, it is ourjudgment that these smaller eatures are tied to disturbance

(rock dilation and shear) within the deeper bedrock. TeCGS considers those areas o the mapped disturbedbedrock zone to be potentiall

Bulletin: West Salt Creek Report

Documents

Transcript of Bulletin: West Salt Creek Report