Missouri University of Science and Technology Scholars' Mine

International Conference on Case Histories in (1988) - Second International Conference on Geotechnical Engineering Case Histories in Geotechnical Engineering

02 Jun 1988, 10:30 am - 3:00 pm

A Finite Element Analysis of the Utah "Thistle" Failure

Blaine D. Leonard LTR Associates, Salt Lake City, Utah

Joseph M. Olsen University of South Alabama, Mobile, Alabama

Follow this and additional works at: https://scholarsmine.mst.edu/icchge

Part of the Geotechnical Engineering Commons

Recommended Citation Leonard, Blaine D. and Olsen, Joseph M., "A Finite Element Analysis of the Utah "Thistle" Failure" (1988). International Conference on Case Histories in Geotechnical Engineering. 2. https://scholarsmine.mst.edu/icchge/2icchge/2icchge-session3/2

This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.

This Article - Conference proceedings is brought to you for free and open access by Scholars' Mine. It has been accepted for inclusion in International Conference on Case Histories in Geotechnical Engineering by an authorized administrator of Scholars' Mine. This work is protected by U. S. Copyright Law. Unauthorized use including reproduction for redistribution requires the permission of the copyright holder. For more information, please contact [email protected]. Proceedings: Second International Conference on Case Histories In Geotechnlcel Engineering, June 1-5, 1988, St. Louis, Mo., Paper No. 3.48 A Finite Element Analysis of the Utah "Thistle" Failure Blaine D. Leonard Joseph M. Olsen President, LTR Associates, Salt Lake City, Utah Chairman, Civil Engineering Department, University of South Alabama, Mobile, Alabama

SYNOPSIS: In the Spring of 1983, a large landslide occurred near the town of Thistle, Utah which blocked major transportation routes and impounded the Spanish Fork River, inundating the town with 200 feet of water. While much attention has been given to the slide and its impact, very little has been directed toward a quantitative understanding of its causes. An analysis was performed of the Thistle landslide using the SEEPSLOPE finite element system in order to evaluate the mechanisms, factors, and causes of the failure. An elastic, perfectly-plastic stress-strain curve was employed in the analysis to model the behavior of the overconsolidated clay soils. It is concluded that the landslide was a compound, progressive failure which initiated at the toe and progressed uphill. Seep­ age forces played a significant role in the failure. INTRODUCTION Early in April 1983, motorists driving through This study was undertaken to quantitatively Spanish Fork Canyon south of Provo, Utah, began address the geotechnical mechanisms involved in noticing cracks in the road near the small town the occurance of the Thistle Landslide. The of Thistle. During the next few days, the objective was to use a finite element stability cracks enlarged, the road began to heave, and analysis to draw conclusions about the physical the nearby railroad tracks began to be distorted. causes and mode of the failure which occurred By the 14th of April, the Thistle landslide, a there. Evidence is presented which demonstrates Quaternary earthflow deposit lying in a small that the slide was a progressive failure in over­ canyon essentially perpendicular to the road, consolidated clays which started at the toe of had moved sufficiently to lift the highway, the slide and progressed uphill. High seepage sever the railroad, and block the Spanish Fork forces played a significant role. in the failure. River. Within 30 days, this blockage had filled the canyon to a height of over 200 feet, butres­ BACKGROUND sing against a large sandstone formation known locally as Billies Mountain. The small town of The area of the slide is near the easternmost Thistle was buried with 62000 acre feet of water. edge of the Middle Rocky Mountain Province which is characterized by generally high mountain Property damage triggered a Presidential Disas­ ranges and plateaus transected by deeply incised ter declaration. Opening a passageway for the erosional valleys. The toe of the slide in ·the Spanish Fork River and rebuilding transportation bottom of the canyon was at an elevation of lines through the canyon subsequently cost local about 5030. The slide extended west approxi­ government and private entities in excess of mately perpendicular to Spanish Fork Canyon a 200 million dollars. The question of what to do horizontal distance of 1200 feet, and then with the slide mass is still unanswered. In southwest another 4500 feet, reaching an eleva­ addition, severe hardships were imposed on the tion of about 5900 at the top. A twenty-foot displaced residents of the small town who as of scarp at the top is noticable, and several re­ early 1987 had not been compensated for the loss lated slides adjacent to the main slide mass are of their property. also evident. Total slide widths vary from about 850 feet at the top to 1200 feet just This mass of overconsolidated clay, over a mile above the bend. Below the bend, widths are long and several hundred feet wide, has moved slightly less than 1000 feet (Duncan et al., several times previously in geologic time, and 1985). The volume of landslide material in the has plagued the railroad lines for most of this canyon has been estimated to be between 3 and century with occassional track movements due to 6.5 million cubic yards (FEMA, 1983; Dames and creep. Deformation had been noted by some Moore, 1985; Kaliser and Fleming, 1986). Total during the early Spring of 1983 which was attri­ volumes involved in the landslide are generally buted to the wet weather conditions. Precipi­ considered to be about 25 to 30 million cubic tation during the winter of 1982-1983 had yards. reached record levels, resulting in abnormally high antecedent moisture conditions. However, a Descriptions of the geology of the Thistle area movement of this magnitude had not been experi­ are given by Witkind and Page (1983), Duncan, et enced in recorded time. al. (1985), and Kaliser and Fleming (1986). According to these sources, three formations The geotechnical aspects of the Thistle land­ underlie or are present as outcrops adjacent to slide are complex and varied, and have been the landslide. The Triassic Ankareh Formation, treated only qualitatively in previous studies. which is a weak, reddish, shaly siltstone and

Second International Conference on Case Histories in Geotechnical Engineering 593 Missouri University of Science and Technology http://ICCHGE1984-2013.mst.edu sandstone underlies the Triassic-Jurassic April 14, vertical deformation of the highway Nugget sandstone, a strong, light colored sand­ surface was so severe that the road was closed stone which in turn underlies the Tertiary to traffic, and an the following day, the road North Horn Formation, a weak, partly alluvial had displaced approximately 10 vertical feet partly lacustrine deposit consisting of mud­ (Dames and Moore, 1985). On the evening of stone, claystone, sandstone, conglomerate, and April 15, the last train used the tracks liNestone. The valley in which the Thistle (Kaliser and Fleming, 1986). landslide rests was cut from the Ankareh Forma­ tion which completely underlies the landslide From the first signs of movement considerable and is exposed an the north boundary of the effort was applied to prevent the Spanish Fork landslide. The. Nugget Formation farms the River from being dammed and to keep the canyon prominent ridge that delineates the southeast open. By April 17, it was clear that these flank of the landslide and underlies the land­ efforts were failing and the residents of slide in the canyon bottom. Duncan, et al. Thistle were evacuated. At this time efforts (1985) concluded that "all movement of the had already turned to unloading what was thought Thistle landslide apparently was above this to be the "head" of the slide, in the area imme­ bedrock unit." The North Horn Formation is diately upslope from the railroad cut. Attention exposed along ridges an both the southeast and also turned to preventing overtopping of the dam northwest borders of the landslide. The by the new Lake Thistle, and evaluating new majority of the landslide is composed of debris transportation routes. On April 22, Utah Gover­ and earthflaw material derived from the North nor Scott Matheson declared the area a state Horn Formation. disaster area, and on April 30, President Ron­ ald Reagan made the Thistle slide area Utah's The Thistle landslide is an early Holcene first National Disaster Area. (Kaliser and Fleming, 1986) landslide mass that has moved on several accassions in geologic Movement of the slide was measured by Railroad time (Schroder, 1971 and Duncan, et al., 1985). and county crews during the first crucial weeks However, Duncan, et al. (1985) argue that there of sliding. According to Duncan et al. (1985), was "no evidence that these alder, deep-seated the Railroad reported that the landslide was landslides, should they be present, were active moving at about 0.75 feet per hour on April 14. during 1983 or later." This average rate increased to a maximum of 2.5 to 2.8 feet per hour during the period of April In approximately 115 years of historic records, 17 to 19, and declined to 0.80 feet per hour by there is no indication of massive movement of April 25. Total horizontal displacement for the the Thistle slide, and no available written bottom of the slide during this period is esti­ accounts of small movement (Kaliser and Fleming, mated to be about 500 feet. Vertical displace­ 1986). However, the slide has caused repeated ments Of up to 1.5 feet per hour were noted problems to the rail lines located at. its toe (Dames and Moore, 1985). Peak sliding rates (FEMA, 1983; Sumsion, 1983). The most recent measured by Utah County were on the order of report of troublesome movements dates to just 6.6 feet per hour on April 19. two months before the catastrophic failure of 1983. Ry the beginning of May, slide movement had largely halted because of the accumulation of PHYSICAL DETAILS OF THE FAILURE slide debris in the valley and the buttressing effects of Billies Mountain. Opinions differ about when the 1983 movement of the Thistle Landslide began. Kaliser &'7.Fieming, ANALYSIS PROCEDURES (1986) report that in January of 1983 an offi­ cial of the Denver and Rio Grande Western Rail­ An analysis of the Thistle failure was performed road reported cracks an the cut slope immedi­ using the SEEPSLOPE finite element system which ately west of the tracks which were "of size consists of two component codes, CFLOW and WOOD­ and depth far in excess of normal." These LUND. CFLOW is a code which calculates steady cracks apparently did nat extend upslope of the state flow through nonhomogeneous media (Taylor cut slope. In early March (Kaliser and Fleming, and Brown, 1967). WOODLUND is a stress analysis 1986) , an inspection revealed that a set of program far plane strain conditions originally cracks had advanced upslope a distance of about developed to analyze underground openings in 100 feet, but no material was noted on the in­ rock (Chang and Nair, 1973). It employs an side track. Sumsion (1983, p. 12), showed a elastic-perfectly plastic stress-strain formu­ photograph taken April 2 of an active slump on lation using a Drucker-Prager yield criterion. the surface of the railroad cut at the toe of the landslide. Tension cracks parallel to and upslope from the railroad cut can be clearly These twa programs were combined and enhanced by seen in the photograph. the Bureau Of Mines (Corp, Schuster, and McDon­ ald, 1975} to analyze mine tailing impoundments. The earliest official records of the Thistle SEEPSLOPE computations for the Thistle analysis Landslide are dated April 13, 1983. At 7:30 were performed on a Gould 9080 minicomputer at a.m. on that day, Denver & Rio Grande western the University of Utah College of Engineering. Railroad personnel reported that their tracks Details of the implementation of SEEPSLOPE are were out of alignment (Sumsian, 1983; Duncan given by Leonard (1987). et al., 1985). As the tracks continued to heave, the railroad attempted to keep the tracks An initial finite element mesh was fashioned open. Late that evening, heave was noted along which consisted of 1160 nodes and 1048 elements the road surface of u.s. Highway 6 and 89, which along an axis approximating the centerline of lay about 200 feet east of the tracks and across the slide. The mesh is 5800 feet long and rela· the Spanish Fork River. By the afternoon of tively shallow, as is necessary to model the

Second International Conference on Case Histories in Geotechnical Engineering 594 Missouri University of Science and Technology http://ICCHGE1984-2013.mst.edu actual failure. Thicknesses vary from 180 feet which are more appropriate for slopes in stiff at the top to 265 feet at the middle, to 480 fissured materials (Lambe.. & Whitman, 1969), feet near the bottom. Topography is based on USGS Quadrangle maps and post-slide topographic The strength parameters for the materials which maps prepared by the Utah County Engineer's Of­ participated in the slide are shown in Table 1. fice. vfuile there was considerable variation in the measured values of friction angle and cohesion Elements at the head of the slide were initially (Leonard, 1987), there were data which suggested 100 feet long by about 30 feet high. Elements that the upper 50 feet was slightly weaker than in the middle of the slide were reduced to 50 the underlying materials. Accordingly, the top feet long and about 35 feet high. In the lower row of elements was designated as soil 1 while portion of the slide, elements were further re­ the remainder of the material above the clay­ duced to 33 feet long by about 30 feet high. stone and sandston~ bedrock was designated as The majority of the elements are rectangular, soil 2. although some triangular elements were included as needed to accommodate the physical boundaries TABLE 1. soil Strength Parameters for Soils 1 and mesh transitions. The CFLOW program uses and 2 for Each Case of WOODLUND this initial mesh together with an estimate of Stress/Stability Analysis the phreatic surface to perform seepage calcu­ lations. As CFLOW moves the nodes which lie on the phreatic surface to satisfy continuity of Soil 1 Soil 2 flow, the remainder of the mesh is distorted accordingly and element sizes vary. phi c phi c

The finite element mesh was divided into four (deg) (psi) (deg) (psi) soil types, representing the primary strata en­ countered on the site. Material parameters used in the analysis were based on data from inves­ Case I-14 11.0 0.0 9.0 0.0 tigations performed at the site of the slide Case I-15 27.4 0.0 27.4 o.o under the direction of the Utah State Division Case I-16 11.0 106.0 9.0 106.0 of Water Rights and the Utah County Engineer Case I-17 11.0 122.0 9.0 122,0 (Duncan, et.al., 1985). While four principal Case I-18 20.0 33.3 20.0 47.8 soil types were identified for analysis pur­ Case I-19 20.0 122.0 20.0 122.0 poses, there was significant spacial variability Case J-1 11.0 122.0 9.0 122.0 in the materials encountered at the site. Case K-1 11.0 122.0 9.0 122.0

Very little is known about the precise pre-slide hydraulic conditions in the slide mass. Perme­ The first two cases represent the pure residual abilities were estimated from post-slide tests. shear strength condition. Duncan et al. (1985) Various inflow and outflow nodes were selected theorized that "antecedent movements had proba­ in the CFLOW analysis to model infiltration and bly reduced the strength along the sides and known springs within the slide area. The SEEP­ base of the slide to residual frictional values~ SLOPE program calculated a phreatic surface In such a case, the stress strain curve would location and computed the corresponding seepage have no peak, but would level off at the resi­ forces. dual value. In addition, the cohesion intercept would be zero. Case I~l4 uses the lowest resi­ For the stress analysis portion of the study, dual friction angle measured on Thistle samples eight cases were developed, each with a slightly in the laboratory. Case I-15 represents the different set of soil parameters. The cases highest angles measured. were designed to provide insight into the Thistle failure, since it was unlikely that any Even in circumstances where residual shear stre­ specific set of parameters would exactly model ngths generally govern stress-strain behavior, the actual conditions. Six of these cases used small peak strengths are sometimes developed the full seepage forces computed by earlier por­ over time due to thixotropy. In addition, por­ tions of the SEEPSLOPE system. For the other tions of a slide which do not contain. previous two cases, seepage forces were reduced by one­ failure surfaces will exhibit peak strengths half (J-1) or one-quarter (k-1) for all nodes before dropping to residual levels. In order to within or above the zone of failure. This was model these conditions the soil parameters used done in an attempt to compensate for the assump­ in case I-16 were modified to include a cohesion tions of steady-state seepage and fully satura­ intercept. An unconfined compression test value ted conditions which the SEEPSLOPE system made. of 106 psi was selected for use as a cohesion :The transient conditions 1.and '.partially satura­ value, representing the second highest such ted soil zones above the phreatic furface be­ value noted in the test data. A relatively high lieved to exist prior to the slide would yield value was selected in order to insure stable lower seepage forces than computed by the results. Preliminary analyses had indicated steady-s~ate analysis performed by SEEPSLOPE. that lower cohesion values would yield a totally unstable slide mass. The materials in the slide were predominantly stiff overconsolidated clays with many fissures Case I-17 uses sligh.tly higher cohesion values and zones of high permeability. The strength in an attempt to bracket the true failure conai­ of these materials also exhibited significant tion~~t. spatial variation. While there was undoubtably some minor excess pore pressure buildup during Case. I-18 represents the average frictional and the failure, the analysis was performed using cohesion shear l!!trength parameters from the effective stress methods and drained conditions laboratory data. 1!\bt a true residual strength.

Second International Conference on Case Histories in Geotechnical Engineering 595 Missouri University of Science and Technology http://ICCHGE1984-2013.mst.edu case, b'lis set of data reflects a buildup of sone peak TABLE 2. Factors of Safety for selected Failure Surfaces strength due to thi.JUted stress levels were so surface is a sinple rrethod for app=ximating the tran­ high relative to soil strengths that nearly every ele­ sient a:>nditions. seepage forces were cut to one-half rrent yielded in early load incrercents. In successive and one-quarter their original levels and the effects of COI!pUter iterations, stresses were transferred back and those lowered forces studied in the stress analysis por­ forth between yielded elerrents and convergence toward tion of this research. The effectiveness of this approach a solution could not occur. is derronstrated in two sets of relationships. First, it was evident in the distribution of failed elenents in In all of the other cases a solution was P=duoed with cases I-19, J-1, and K-1 that the number of failed ele­ elerrents at or near failure clustered at or near the rrents in the analyses with reduced forces is less than in bottom of the slope. Only case I-18 with lower cohesion I -19, particularly in the shallower elerrents. Since the failed elerrents higher than app=xirrately one-third of seepage forces were reduced in these cases for elenents the way up the slope. In this case only the upper above the p=posed failure zone, this result is expected. approxirrately one-fourth of the rresh was stable. second, sone trends can be seen in the factors of safety for the trial failure surfaces. Conparing case J-1 with In order to better oonpare the various analyzed cases I -19, the decreased seepage forces yield higher or es­ and nore fully evaluate the stability COIIputations for sentially equal factors of safety for five of the seven the 'lhistle landslide, average factors of safety were failure surfaces. The average increase in factor of carputed for seven trial failure surfaces using data safety is 27 percent. COnparing case K-1 with J-1, higher f=m each analysis case. The seven failure surfaces or equivalent factors of safety are yielded for four of were selected specifically to aid in the understanding five surfaces. The factors of safety for Surfaces 6 and of the nature of the failure. Weighted average factors 7 rerrain =ughly unchanged in all three cases. This would of safety were COI!pUted for each of the seven failure have relatively low seepage forces • .surfaces for each of the WCODLUND cases described above. Results of these COI!putations are shown in Table 2. SEEPSIDPE was not able to carpute realistic factors of Although the total length of the landslide is known to safety for cases I-14 and I-15 which did not converge. be about 5700 feet, the depth and shape of the failure 'Ihe specific results of these two cases I!RlSt be generally surface is still in question, and it is tmelear whether disregarded, except as they relate to the overall failure the slide is Itade up of one long failure mass or many parameters of the Thistle slide. The results of these Sllall oonpound failures. two cases indicate that 'Ihistle was not strictly a re­ currance of an old slide along a previously failed sur­ Failure surface 1 represents the full length failure, face. Elerrent factors of safety were so unifonnly low based on the theo:cy that the slide was a long, shallow, that transfer of excess stresses could not be accan­ a:>ntinuous failure mass. Surface 2 also represents a plished. Even the high friction angle used in case I-15 relatively long, shallow landslide mass, but extends was not sufficient to bring about stability. Based on only one-third the length of the total slide, and does this information, it is likely that sone peak strengths not conpletely c=ss the canyon floor. Failure Sur- were involved in the failure. face 3 is much shorter, and approaches the shape of a circular failure. Failure Surface 4 was developed to All of the cases analyzed clearly indicate that the slide evaluate the validity of a deep failure. Sone have initiated at the toe. Even when large numbers of elerrents postulated that the 'lhistle landslide involved a deep are yielded, as in case I-18, factors of safety near the failure which lifted the highway in the canyon floor. head are consistently greater than one. As the number of Surface 5 is a small, shallow, nearly circular failure. yielded elerrents decreases, they tend to be located pri­ Surfaces 6 and 7 represent small, localized failures of marily at the toe. 'Ibis observation is st=ngly supported the landslide. A srrall failure of this sort could be by the failure surface factors of safety. Disregarding triggered by weather conditions or construction activity, cases I-14 and I-15, Failure Surfaces 1 and 2 never have and rray lead to a larger, p=gressive failure event. the lowest factors of safety for a given case. In addi-

Second International Conference on Case Histories in Geotechnical Engineering 596 Missouri University of Science and Technology http://ICCHGE1984-2013.mst.edu tion, in five of the six cases, the factors of safety for progress up the hill. For instance, in case I-16, the these tw:l surfaces are considerable higher than those for minimum factor of safety is for the short, shallow Fail­ :the other surfaces. ure Surface 5. With a factor of safety of 1.09, this surface is near failure. Once this mass of elements Tension cracking noted in a photograph taken ten days fails and the support is renoved from the mass above before the slide began supports the argurrent that sliding Failure Surface 3, it is likely that Surface 3 will began at the toe. It has been suggested (Duncan, et al., approach instability. With the re.roval of elements n~ 1985) that it was unclear whether events at the toe or the canyon bottan, excess stresses would be passed to the head were rore inp:>rtant in the initiation of sli­ elements higher on the slide, and lower factors of safety ding. Evidence from this study clearly indicates other­ would result. SUch a scenario supports the corrpound wise; the slide began at the toe. failure theo:cy.

With full seepage loads applied, the rost stable result CX>NCLUSIONS AND RECXM1ENDATICNS is yielded in case I -19. The distribution of yielded elements clearly shows fewer yielded elements in this Several conclusions can be drawn fran this study of the case than in the other cases using a· full seepage load. 'Ihistle Landslide based on the results presented and '!he failure surface factor of safety results also support considerations discussed. 'l'hey are as follows: this observation. Fbr ~y given failure surface, case I-19 yields a higher factor of safety then the others. 1. The 'lhistle Slide was not a sinple recurrance of a pre-existing slide acting 100lely at residual shear Two observations can be drawn about the shape of the strength levels. It appears to be a ca:cp:>und pro­ failure surface based on results of trial failure sur­ gressive failure with peak strengths nobilizing in face COIIputations. First, the failure was not a deep many of the soil zones. While ancient slides have seated circular failure. Am:mg the six converging cases, occurred at this location and soma of those failure the minimuin factor of safety was associated with Failure surfaces may have been reactivated in 1983, the ma­ Surface 5 in three of the cases and with Failure Surface jority of the failure surface appears to have been in 7 in tw:l of the cases. Surface 5 is a very shallow, previously llllfailed material. If this were not so, circular surface, and Surface 7 is a superficial stability could be seen in those cases involving sloughing failure of the toe. Failure Surface 4, the residual. strength parazreters. deep seated failure, never yields the lowest factor of safety. Second, the failure did not involve a long, 2. Analyses perforned in this study indicate that the continuous surface. Failure Surface 1 yields the highest slide started at the toe and progressed ll};tli.ll. aggregate factor of safety in each of the six cases. 3. '!he failure did not involve a deep-seated, circular Relatively high cohesion values were required to achieve failure surface or a long, continuous failure sur­ convergence and stability in the various analyses. Cases face. It is rore likely that relatively shallow involving low cohesion values, such as I-18, resulted in masses of soil failed, progressively triggering nunerous yielded elerrents and low factors of safety on failures in adjacent, uphill zones. the trial failure surfaces. ·Fbur of the seven surfaces had factors of safety near or below one in case I -18. 4. Seepage forces played a significant role in the fail­ None of the failure surfaces analyzed yielded factors of ure at 'lhistle. As suggested by previous investi­ safety less than one in cases I -17 and I-19. '!he co­ gators of the Thistle. Slide, water in the soil re­ hesion value used in these tw:l cases was 122 psi, repre­ sulting from high precipitation levels was definitely senting the highest value obtained in the laboratory a factor contributing to this failure. £rem unconfined COitpression tests. Since significant cohesion was required to achieve stability, it seems REFERENCES likely that soil strengths nobilized during the slide were higher than residual levels. Chang, C-Y. 1 and K. Nair, (1973). DevelEV~t and Ageli­ cations of 'lheoretical Methods for ~ting Stab~lity Only small increases in stability were obtained by in­ of openings in Rock, Final Report. ~-Lundgren & creasing the friction angle. Cortparison of Cases I-17 .Assoc., u.s. Bureau of Mines Contract H0220038. and I -19 de!ronstrates that for an 82 pereent increase in friction angle, only a 9 percent increase in factor of Corp, E.L., R.L. Schuster, and M.M. ~, (1975). safety resulted. The conpa.rison of Cases I-16 and I-17 Elastic-Plastic Stability Analysis of Mine-Waste and Cases I-18 and I-19, however, derronstrate the sensi­ 'Eil6arikiierits, Bureau of Mines Report of Investigation tivity of the analysis to the cohesion parazreter. Case #8069. I-17 used a 15 percent higher cohesion value than I-17, with a resulting increase in factor of safety of 11 per­ Danes and M:lore, (1985). En~ing Geology and Geo­ cent. case I-19 used a cohesion value averaging 310 per­ technical Engineering Studies: 'Ihistle Landslide cent higher than case I-18, with a resulting increase in Alternatives Project, unpublished report. factor of safety of 340 percent. Duncan, J.M., R.W. Fleming, andF.D. Patton (1985). Definite trends relating to the developnent of a conp:nmd :Report of the 'lhistle Slide COmnittee, unpublished failure surface are not evident in the results of this report to the Utah State Engineer. study. H<:ll\'ever, the observations that yielded elements are always canoentrated at the toe of the slide, and FEM!I., (1983). FEMA Hazard Mitigation Team Intergovem­ long, continuous failure surfaces always have high fac­ :rrental Hazard Report for the State of Utah in Response tors of safety indicates that the failure as a cxxrpound to the Apnl 30, 1983 D~saster Declarat~on, FEMA report pmgressive failure. Conputed failure surfaces with low #FEMA-680-D~~. factors of safety are generally smrt, but the actual failure was quite long. It seems reasonable that small Kaliser, B. and R.W. Fleming, (1986). "The 1983 Landslide zones could fail near the toe of the slide, resulting in Dam at 'lhistle, Utah," Proc. , ASCE Geoteclmical Spe­ the rencval of the resisting forces supporting higher ciality COnference: Landslide Dams: Processes, Risk elements, ult:inately causing similar small failures of and Mitigation, p 59-83. uphill areas. In this way, the failure would gradually

Second International Conference on Case Histories in Geotechnical Engineering 597 Missouri University of Science and Technology http://ICCHGE1984-2013.mst.edu Leonard, B.D. (1987). A Finite Elerrent Analysis of the 'Ihistle Landslide, unpublished thesis presented to the graduate school of the University of Utah in partial fulfillm:nt of the requirerrents for the degree of ' Master of Science.

Lambe, T.W. and R.V. Whitmm, (1969). Soil Mechanics, John Wiley & Sons, New York, N.Y.

Schroder, J.F. (1971). "Landslides of Utah," Utah Geological and Mineralogical SUrvey, Bulletin 90.

Taylor, R.L. and C.B. Brown, (1967). "Darcy Flow Solutions with a Free Surface," ASCE Hydraulics Division Journal, vol 93, HY2, 25-33.

Witkind, I.J. and W.R. Page, (1983). "Geologic M3.p of the 'Ihistle Area, Utah co., Utah," Utah Geological and Mineral Survey, Map 69.

598 Second International Conference on Case Histories in Geotechnical Engineering Missouri University of Science and Technology http://ICCHGE1984-2013.mst.edu