The following manuscript is being submitted for consideration to the 2012 Canadian Geotechnical Conference in Winnipeg, Manitoba.
Ashcroft Thompson River Landslides: Spatial and Temporal Scale of Controls
I. Hall1, M. Porter1, P. Quinn2, and K.W. Savigny1
BGC Engineering Inc., 1Vancouver, British Columbia, Canada; 2Victoria, British Columbia, Canada
A 10 km reach of Thompson River south of Ashcroft, British Columbia is affected by a number of large landslides which have been intermittently active for more than 100 years, since the first rail line was constructed through the corridor. Both national railways, CN and CPR, traverse several landslides, some of which are known to be moving. A two phase program of study was initiated to better understand and manage the landslide risk, and this paper describes some findings of the Data Gap Analysis in the first phase. Landslide activity is shown to be strongly correlated to peak flood levels, which correlate strongly with peak snow pack in the previous winter. Thus snow pack monitoring can serve as a form of early warning, one potential component of risk management efforts.
BGC Engineering Inc. (BGC) was engaged by Public Works and Government Services Canada on behalf of Transport Canada, with additional support from CN Rail (CN) and Canadian Pacific Railway (CPR), to study the Ashcroft Thompson River Landslides (ATRL) south of Ashcroft in British Columbia. Both the CN and CPR main lines traverse several landslides that move intermittently in this corridor. As both railways are on the same side of the river for several kilometres crossing these landslides, sudden reactivation of one of the landslides could sever the national rail network, creating significant disruption to the Canadian economy. BGC was engaged to complete three major tasks in the first phase of a two phase program: Task 1 – Data Gap analysis; Task 2 – Stakeholder Workshop; and, Task 3 – Phase 2 Plan. The second phase of the program is expected to be a multi-year effort, to be developed subsequent to the completion of this work.
This paper presents an overview of the findings of the Task 1 – Gap Analysis of the Phase 1 ATRL work, with emphasis on new findings related to watershed-scale controls on landslide activity.
An area of prehistoric, historic and active landslides extends along a 10 km reach of the Thompson River valley from Ashcroft to Basque, B.C. These landslides are known collectively as the ATRL, and their locations are illustrated in Figure 1. The location of these landslides within the overall Thompson River watershed is shown in Figure 2, and the location of the watershed within the province of British Columbia is shown in Figure 3.
The two major Canadian railways, CN and CPR, pass through this reach and are vital to the North American economy and its shipping network, as they provide a strategic connection between the Canadian west coast and the majority of the continent’s resources and population to the east. The landslides have resulted in train derailments and numerous service disruptions since the original CP grade construction in the 1880s. In several places the rail lines rest directly on active landslides where gradual, continuous slope movements (and occasional rapid failures) affect their safe and reliable operation. This has resulted in significant recurring disruptions in rail service and the need for periodic track re-alignment.
More than twenty individual landslides can be identified on aerial photographs as shown on Figure 1. Of these, several significant landslides identified by CP, CN and others, have been active in the last thirty years. These include, for example, the Goddard, North, South, Ripley, and CN Mile 50.9 landslides. Most recently, the CN Mile 50.9 landslide affected rail traffic in the early 2000’s. In 1982, the Goddard landslide resulted in extensive damage to CP, rendering the track out of service for seven days. The initial failure of the North Slide occurred in 1880, resulting in a landslide debris dam that temporarily blocked Thompson River and flooded parts of the Town of Ashcroft. Significant slope movements have been noted here as recently as August 2008. The South Slide moves at up to 10 mm/yr, influencing the CP track alignment and to a lesser degree the CN tracks. The Ripley Slide is currently the most active landslide and affects the CP tracks, with observed movement rates as high 80 mm/yr.
On average, a major landslide has occurred within Thompson Valley about once every 50 years. Significant slope movements, which force the rail companies to issue traffic slow orders, occur approximately once every 5 to 10 years. Despite their frequency, the factors contributing to these landslide movements are not fully understood.
Porter et al. (2002), and Clague and Evans (2003), suggest that the low shear strength of a glaciolacustrine silt and clay unit located near river elevation predispose the slopes to failure. Porter et al. (2002) and Eshraghian et al. (2006) identified river drawdown and erosion as contributing factors. Modern irrigation of cropland within the valley has been tentatively identified by other authors as a potential contributing factor, but this has not been supported by recent modelling conducted by Bishop (2008), who demonstrated that regional groundwater flow dynamics are likely responsible for the elevated pore pressures in the basal glaciolacustrine unit that contribute to slope instability.
3 LANDSLIDE HAZARD AND RISK
3.1 Risk Context
A preliminary risk framework was developed to support the rational evaluation of relative importance of various factors associated with landslide risk in the ATRL study area. Risk terminology is defined in general accordance with CSA standards (CSA 1997) and other reference standards on landslide risk (e.g. AGS 2000, BC MOF 2004). A mathematical formulation for risk has been proposed, involving the multiplication of Hazard Probability (PH), Spatial Probability (PS:H), Temporal Probability (PT:S), Vulnerability (V) and value of Elements at Risk (E). An understanding of risk requires consideration of specific hazard scenarios. The following have been proposed:
– Acceleration of slow creep movement of existing landslides leading to excessive ground movement:
– Sudden reactivation and large rapid to extremely rapid movement of existing landslide(s):
– Loss of ground (subsidence, disruption of existing ground);
– Excessive deformation (damage or destruction of railway tracks, roads or other infrastructure);
– Initiation of new large rapid to extremely rapid landslide(s):
– Landslide dam and associated upstream flooding; and
– Outbreak flood downstream associated with breach of landslide dam.
Risk may affect different elements, each with their own intrinsic value, in different ways. It is suggested that the following elements could be at risk due to the identified hazard scenarios:
– Residents of, or visitors to, First Nations reserves, Ashcroft, Spences Bridge and the rural properties between the two towns, and potentially further upstream of Ashcroft;
– Buildings and businesses in these same areas;
– Railways through the corridor;
– Roads and highways through the corridor;
– Hydro lines and other linear infrastructure;
– Productive cropland;
– Fish and fish habitat;
– Cultural heritage features; and
– Freshwater intakes along the corridor.
It is expected that work in Phase 2 will be planned in order to reduce landslide risk, therefore the preceding overview of risk provides a general context for framing future action. The following section discusses the findings of an analysis of available data to identify gaps in existing knowledge. These gaps should be filled through further study to enhance the understanding of landslide risk.
3.2 DATA GAP ANALYSIS
3.2.1 Key Issues
The landslide problem in the ATRL is believed to be governed by a complex set of factors, some of which may be related in space and time, and others of which are likely unrelated. The following section presents a context for organizing thoughts about these various factors, as summarized in the following nine primary issues.
Issue Number 1 – Scale Context
The main factor that distinguishes work completed to date from the proposed ATRL phase 1 and 2 studies is scale.
The ATRL shown in Figure 1 range up to nearly 1 km2 in surface area (plan view). They extend up slope as far as almost 1 km from the river’s edge where the rail lines are located, and as much as 1350 m along the tracks. The headscarps are as much as 145 m above the river and movements have been detected in slope indicators at depths well below the Thompson River thalweg elevation. Natural exposures of basal slide surfaces for geological mapping are rare and limited to the toe areas along the Thompson River banks where in situ materials are highly disturbed.
Site investigations for the purpose of risk control at active slide locations along the rail lines are typically near the toes of the ATRL. Hence, current understanding of the landslides is based on relatively narrow, discontinuous swaths of geotechnical data that have been acquired along the two rail lines, and is not fully representative of the areal or vertical extent of the actual slides, or the mechanisms that control them. In short, these landslides are little understood at their full scale.
Issue Number 2 – The Thompson Preglacial Valley
The Thompson Valley is a Tertiary age bedrock valley that formed through prolonged downcutting over many millions of years. Immediately prior to Pleistocene glaciation, the valley bottom was tens of metres deeper than the current river thalweg and the floodplain was wider. Floodplain and terrace gravels rested on weathered bedrock, comprising well-bedded clastic sedimentary and volcanic rocks with minor carbonates.
The preglacial geology described above has been observed in outcrop (Clague and Evans 2003) and in several engineering drill holes. However, the continuity of the gravels between outcrop and at depth below both Thompson River and the ATRL remains to be verified. As well, the configuration of the bedrock surface (subcrop) beneath the river and landslides is unknown.
Issue Number 3 – Significance of Initial Pleistocene Glaciation
The basal Pleistocene unit is rhythmically laminated glaciolacustrine clay and silt with high plastic clay couplets. It was deposited in a glacial lake impounded by ice flowing up the Thompson Valley from the Coast Range. The unit acts as an aquitard, confining artesian water pressures in the underlying Tertiary age alluvial deposits and bedrock (Porter et al. 2002). As most, and possibly all of the ATRL are seated in this basal Pleistocene unit, its sedimentology, stratigraphy, engineering properties and role in creating artesian groundwater conditions are crucial to the ATRL project’s risk management objectives.
As described previously, the continuity of the basal Pleistocene section has not been demonstrated. The many tens of metres of vertical offset of what has traditionally been viewed as the same units, coupled with somewhat different sedimentology between outcrop and depth below river may indicate two or more distinct units. This is crucial to verify in order to correctly bound operative shear strengths; constrain lateral extension of the ATRL or any new landslides; and, interpret the upward hydraulic gradient as well as its relationship to local and regional groundwater regimes and surface hydrology.
Issue Number 4 – Record of Other Glaciations
Clague and Evans (2003) and Johnsen and Brennand (2004) describe the stratigraphic record of other Pleistocene glaciations that inundated the Ashcroft area. The modern Thompson River is eroded into the “drift package” left by these glaciations. The package comprises most of the materials covering the slopes of the valley, particularly at lower elevations. During late glacial and Holocene downcutting, numerous terraces were deposited on top of the drift. At lower elevations, colluvium derived from the drift and alluvial terraces are testament to prehistoric landslide activity. The ATRL are confined within this drift package and do not extend into bedrock. However, the stratigraphic complexity of the drift package coupled with its disturbance by slide movements makes it challenging to represent in hydrogeological and slope stability models. Currently very little is known about the drift package at distance from the river banks.
Issue Number 5 – Landslide Configurations
Many of the ATRL show stacked failure surfaces and multiple slide blocks. In general, movement rates increase with proximity to the river bank and at shallower depths (Porter et al. 2002, Clague and Evans 2003, Eshraghian et al. 2006 and 2007). Stability analyses completed to date depict the landslides as having a layer-cake stratigraphy that is horizontally disposed and hundreds of metres across in cross section (e.g. Eshraghian et al. 2006 and 2007). This approach is adequate for short-term risk control considerations in proximity to the river banks, but its suitability decreases quickly as a function of distance away from the river’s edge. As indicated under issues 1, 2 and 4 above, very little is known about subcrop, the drift package, perched water tables in the drift package, how landslide failure surfaces propagate upwards through the drift package, and the ATRL configurations at distance from the river. What can be surmised, however, is that the geology will not be simply horizontal and layer-cake.
Issue Number 6 – Mobilized Shear Strength and Hydraulic Conductivity of Basal Unit
To date, most shear strengths used in stability analyses have been determined through back analysis of moving slide blocks with the results calibrated against Stark and Eid (1994) empirical correlations. Although these compare favourably with the few direct shear tests reported by Eshraghian et al. (2007), much more detailed consideration is needed in order to make the most of improved understanding of the landslide configurations. In particular, if the basal glaciolacustrine unit is actually two distinct layers, then it is reasonable to speculate that the lower of the two, derived from residual soils and a landscape millions of years old, will be of lower strength and lower hydraulic conductivity.
Issue Number 7 – Groundwater Regime
Regional groundwater flow is controlled by the secondary permeability of the rock mass in keeping with the models considered by Hodge and Freeze (1977), but also influenced by pervasive north-trending regional faults that occasionally intersect the natural valley slopes (based on observations by one of the authors). Little is known about this regional groundwater system with the exception of the vicinity of Cache Creek where publically available documents related to active and proposed landfill schemes provide a starting point for the gap analysis.
Most groundwater wells in the Ashcroft area are developed in the Pleistocene drift package and Holocene terrace deposits. These wells provide valuable base-line information but the data have been little used as part of landslide studies completed to date.
Groundwater monitoring completed as part of landslide investigations has traditionally been the most widely used source of groundwater information. Most stability analyses depict the groundwater regime very simplistically, related in large part to the horizontal layer-cake assumptions described previously.
Little is known about the landslides at a distance from the railway alignments, however it is unlikely that stratigraphy is flat-lying and continuous. One of the main contributors to slide movement may be lateral flow of groundwater along silt and sand portions of many rhythmite couplets. Infiltration is at the valley sides where the texture of these specific couplets is the coarsest and flow is lateral and downward following the draped configuration of the laminae and beds. When a slide severs the lateral continuity of the beds, porewater pressure could build up to where the failure surface cuts the bed off. BGC has found evidence of this scenario during the South Slide drilling in previous work. Confirming the interpreted stratigraphy is therefore critically important for correct hydrogeological modelling as well as for risk management planning.
Issue Number 8 – Surface Hydrology and Climate Change
Surface water runoff regimes approaching and downstream of the Thompson Valley rims have played an important role in the ATRL. These provide primary sources of water that infiltrate into the glacial drift package. Since the onset of European settlement, these patterns have been altered, most notably in association with early irrigation practices, which are well known for landslide causation. Sensitivity of the ATRL to surface runoff regimes is an important issue that was addressed in the present gap analysis but much will remain for further study in phase 2. How precipitation patterns can be expected to adjust to various climate change scenarios is an important component of this effort.
Issue Number 9 – Influence of Thompson River
Surface water regimes are also affected by risk control measures taken by the railways (e.g. channel bank armouring, river control structures and local inflow control). These impose hydrological cause-effect impacts with regard to river and slope hazards. For example: is the river scour or erosion exposing remnants of landslide material in a manner that will re-activate the ATRL? Has the intrusion of rail grades into the channel resulted in increased scour depth, thereby exacerbating landslide toe erosion? Joint work by BGC and Northwest Hydraulic Consultants Ltd. (NHC) over the last decade has determined that this is indeed the case because the toes of several ATRL have been confirmed to up-thrust into the active Thompson River floodplain where they are steadily eroded by the river. Has erosion protection along one bank increased the bank erosion potential on an opposite bank that lies along the toe of an old slide; or, has the river responded to a constricted channel by deeply scouring the bed and eroding the opposite bank? Also, the river flood level can drop quickly, thereby increasing relative bank pore pressures and potentially disturbing or removing stabilizing toe loads (Eshraghian et al. 2006).
3.2.2 Controls on Landslide Activity and Implications for Risk Management
The ATRL have been the subject of several previous investigations in the literature, beginning with Stanton (1898), and including Clague and Evans (2003), Eshraghian (2007), Eshraghian et al. (2005, 2006, 2007 and 2008) and Porter et al. (2002). The geology has also been described by Fulton (1969), Fulton and Smith (1978), Johnsen and Brennand (2004), Ryder (1970 and 1976) and Ryder et al. (1991). These references provide very good background, and establish a general context for beginning to understand the ATRL issues.
The ATRL occur in a complex geological setting, occurring within a complex sequence of glacial and interglacial deposits confined along a pre-existing narrow bedrock valley. Failure is believed to involve primarily thinly laminated glaciolacustrine silt and clay units. A typical exposure is shown in Figure 4. The ATRL are relatively flat features, with low travel angles ranging from about 7 to 18 degrees, with a mean travel angle of 10 to 11 degrees. This geometry suggests either low shear strength, high pore pressures, or a combination of both. It is often assumed that the high plastic clay laminae have very low residual shear strength, with values as low as perhaps 10 to 15 degrees possible, according to relationships proposed by Stark and Eid (1994), with values depending on the Atterberg limits, clay fraction and overburden pressure. While such low friction angles are theoretically possible along pre-existing shear surfaces, the limited available strength test data do not fully support this interpretation. Ring shear test data reported by Bishop (2008) suggest residual strengths in the clay potentially much higher than expected from Stark and Eid (1994), with a mean value of 20 degrees, low values of 10 and 11 degrees, and one result of 33 degrees on a clay-rich slickenside. Such strengths on their own do not fully explain the range of observed travel angles. It is therefore expected that seepage pressures must also play a significant role, and it is further suggested that static liquefaction of the silt may also be a factor. This possibility is supported by observations of flowing silt reported by Stanton (1898) in the earliest investigations of the ATRL.
Eshraghian (2007) has demonstrated a strong relationship between river behaviour and landslide activity. This temporal coincidence of significant landslide movement and river flood levels is illustrated in Figure 5. Significant landslide movements have only been observed during or immediately following years with peak flood levels significantly higher than average. Furthermore, almost every year with a significant peak flood level has seen significant landslide movement.
The temporal coincidence of landslide activity and elevated peak flood may be expected to be due either to significant erosion, fluctuation of pore pressures, or a combination of both. The available bathymetric data are not sufficiently extensive to permit an examination of the magnitude of erosional episodes; however, groundwater data are available for a number of piezometers installed along the river, with nearly continuous data available for a period of about four years. Selected data from a set of three nested piezometers is shown in Figure 6, which also shows water level in the adjacent Thompson River for comparison. This graph shows that groundwater levels have the same general trend as river levels, with shallow piezometers showing a more direct hydraulic connection. The deeper piezometers have higher water levels, suggesting upward gradients and associated seepage pressures.
The seasonal variation in upward gradient is illustrated in Figure 7. It is evident that a rapid increase in upward seepage pressure occurs in late summer as river levels drop from the peak flood. These upward seepage pressures are sustained through the winter, dropping again the following year with the return of rising river levels. Similar plots can be made for horizontal gradients, showing that inward flow away from the river occurs close to the river during peak flood, and outward flow, toward the river, occurs after drawdown, resulting in destabilizing seepage pressures in late summer or early fall.
Landslides are often controlled by local weather, which can have a governing effect on local hydrology. Therefore it is relatively common to look for local climactic controls on river behaviour when investigating landslide controls. Thompson River drains a fairly large watershed, and the ATRL are located in the lower part of the watershed, very close to its confluence with the Fraser River, as shown in Figure 2. One might therefore expect that river levels will not be affected strongly by local weather, and instead be more strongly related to conditions extending across the watershed. Figure 8 compares mean annual temperature and precipitation with peak flow in Thompson River near the ATRL. This graph shows that local river levels have no relationship with local rainfall or temperature, at least when examined on the scale of annual mean values. Since landslide activity has a clear relationship with peak flood, it may be deduced that local weather has little or no role in landslide activity.
Figure 9 shows mean daily flows for Spences Bridge, located approximately 25 km downstream of the ATRL, past the Nicola River, and Savona, located approximately 35 km upstream, at the outlet of Kamloops Lake. This graph shows very little difference between the two stations, located approximately 70 km apart along the river. Therefore local tributaries, including the Nicola and Bonaparte Rivers and numerous small creeks, carry very little flow in comparison with Thompson River. These local tributaries would be expected to respond to local weather conditions, but since their effect on flow in the Thompson is small, the associated effect of local weather is correspondingly small.
It is hypothesized that summer peak flood levels may correlate with the previous winter’s snowfall across the watershed. A small number of snowpack monitoring stations exist in the uplands in the upper part of the watershed, as shown in Figure 10. Peak snow pack data from these six stations have been obtained and averaged for the period of record, which extends from 1984 to present. These mean peak snow pack thickness data are plotted against peak flood levels in Thompson River at Spences Bridge in Figure 11. It can be seen that a strong temporal correlation exists between the mid-winter peak snow pack and mid-summer peak flood. Snow pack data may therefore be used as an early indicator of the potential for a higher than average peak flood, which would further indicate the potential for landslide movement. Peak snow pack may therefore be a good early warning indicator of the potential for significant landslide activity in a given year.
3.2.3 Identified Data Gaps
A major focus of the work was to examine existing information for the purpose of identifying weaknesses, or gaps, in the current state of knowledge. An analysis of the existing data in relation to the ATRL landslide problem suggests the following important data gaps:
1. Uncertainty of subsurface conditions outside and between existing landslides, particularly from about 100 m to 500 m from the river:
b. configuration of remnant pre and interglacial landslides, if any;
c. groundwater levels and hydraulic conductivity(ies);
d. depth to and shape of bedrock surface, and determination of possible involvement of weak bedrock in the landslides.
e. stress-deformation behavior of the materials involved in failure:
f. peak and large deformation strength-deformation behavior of glaciolacustrine sediments involved in failure;
g. time-dependent behavior and brittleness of fine sediments; and,
h. time-strength-deformation behavior of involved rock, if applicable.
2. Imperfect understanding of the complex mechanical model for development and reactivation of the landslides, including large rapid landslides that could potentially dam Thompson River. This would include understanding whether the most significant known movements in the late 1800s were first time failures versus major reactivation events.
3. Relative importance of river drawdown (and associated seepage gradients), river erosion, and infiltration as key triggering factors.
4. Controls on pore pressure fluctuations in the river bank and beneath the river bed.
5. Evaluation of erosive potential of river.
6. Development of a clear understanding of linkages between watershed climate data, seasonal flood levels and landslide movement (in terms of timing, magnitude and location).
7. Detailed topographic and bathymetric models.
8. Establishment of movement limits defining serviceability issues for railways.
9. Establishment of movement limits indicative of concern for sudden large movements or significant detrimental creep movements.
10. Development of a local water balance model integrating climate, infiltration, runoff with river behavior and landslide activity.
11. Inventory of fish and fish habitat distribution and the effects of past and future stabilization efforts on fish.
12. Inventory of past stabilization efforts, their relative successes and capital and operating costs.
3.2.4 Priorities for Further Study
A consensus on priority of the preceding data gaps could not be reached at this stage of study among the project stakeholders. In order to develop a reasonable plan for further work, one that balances costs of further study against probable value in risk reduction, it is necessary to frame the gaps within a qualitative framework that seeks to answer priority questions first, along the path to more confident risk assessment.
The railways appear to be able to tolerate the typical slow creep observed in several of the known landslides during normal years. The most significant risk to railway operation is likely associated with significant sudden movement, a form of which occurred at the Goddard slide in 1982, and as occurred in the late 1800s as either reactivation or first time movement with the large landslides reported by Stanton (1898), which will be denoted the “Stanton Slides.”
The “Stanton Slides” occurred in the vicinity of Black Canyon, North Slide immediately upstream and South Slide immediately downstream. Portions of the head scarps of these two historical slides converge to a ridge situated immediately east of Black Canyon. Black Canyon itself is a deep bedrock canyon incised by Thompson River since the last glaciation. Previously, Thompson River flowed to the east of the canyon, hence the ridge where the two “Stanton Slides” converge represents a complete Quaternary Section that has not been influenced by landslides since the last glaciation. This location is a prime target for geotechnical and hydrological investigations in Phase 2 of this project for the purpose of understanding the large-scale, high velocity “Stanton Slides” – and potential future events like them – that represent the greatest industrial and environmental landslides risks. Some of the key questions these investigations would focus on include:
1. What is the undisturbed stratigraphy of the Quaternary Section and the geomorphic history that accounts for it? In particular, what are possible mechanisms for pre-shearing in what is considered an undisturbed Quaternary Section (e.g. preglacial and/or interglacial landslides; compaction shearing, etc.)?
2. What are the spatial distribution and engineering properties of basal sediments (currently believed to be of glaciolacustrine origin) in the Quaternary Section that are believed to control the “Stanton Slides”?
3. What are the groundwater regimes in the Quaternary Section and in bedrock beneath it?
4. How are these groundwater regimes influenced by the following: past and present irrigation practices; past (Holocene and historical) wetting periods; and, potential future wetting periods that may accompany climate change forecasts?
5. How will the Thompson River hydrology respond to potential wetting periods that may accompany climate change forecasts?
6. Based on the knowledge gained from the foregoing studies, where are the high risk areas along the affected reach of the Thompson River valley? Are they associated with retrogression of existing large landslides, or with reaches along which valley slopes comprise thick undisturbed portions of the Quaternary Section? The answers here may of course involve a broader investigative program than for the “Stanton Slides” understanding.
Large landslides in a 10 km reach of Thompson River pose a risk to rail infrastructure and traffic. Many of these landslides move intermittently, sometimes suddenly and with significant deformations. The potential for significant movement is shown to correlate strongly with peak summer flood levels in the adjacent Thompson River, and flood levels are shown to correlate with peak snow pack depth in the previous winter. Snow pack monitoring thus provides early warning of the likelihood of late summer or early fall landslide activity, and may therefore serve as a useful component of risk mitigation. The body of existing information has been examined to identify gaps in the current understanding of these landslides. These gaps have been examined in the context of a risk framework, and priorities for future work have been developed.
The writers would like to acknowledge CN, CP Rail and Transport Canada for sponsorship of the work that led to development of this paper.
Australian Geomechanics Society (AGS) Sub-committee on Landslide Risk Management. 2000. Landslide Risk Management Concepts and Guidelines. Australian Geomechanics: 49-92.
Bishop, N.F. (2008) Geotechnics and hydrology of landslides in Thompson River Valley, near Ashcroft, British Columbia. Masters Thesis. University of Waterloo, Waterloo, Ontario, Canada.
British Columbia Ministry of Forests (BC MoF). 2004. Land Management Handbook 56: Landslide Risk Case Studies in Forest Development Planning and Operations.
Canadian Standards Association. 1997. CAN/CSA Q850-97. Risk Management: Guidelines for Decision Makers. Prepared by Canadian Standards Association.
Clague, J. J., & Evans, S. G. (2003). Geological framework of large historic landlsides in Thompson River Valley, British Columbia. Environmental Engineering Geoscience Vol. 9, 201 – 212.
Eshraghian, A., (2007). Hazard analysis of reactivated earth slides in the Thompson River Valley, Ashcroft, British Columbia. Edmonton, Alberta, Canada: Ph.D. thesis, Department of Civil and Environmental Engineering, University of Alberta.
Eshraghian, A., Martin, C. D., and Morgenstern, N. R. (2008) Movement triggers and mechanisms of two earth slides in the Thompson River Valley, British Columbia, Canada. Canadian Geotechnical Journal. Volume 45: 1189-1209 (2008).
Eshraghian, A., Martin, C. D., & Cruden, D. M. (2005). Earth slide movements in the Thompson River Valley, Ashcroft, British Columbia. The 58th Canadian Geotechnical Conference. Saskatoon, Saskatchewan, Canada: Canadian Geotechnical Society, Saskatoon.
Eshraghian, A., Martin, C. D., & Cruden, D. M. (2007). Complex earth slides in the Thompson River valley, Ashscroft, British Columbia. Environmental & Engineering Geoscience Vol. 13 , 161 – 181.
Eshraghian, A., Martin, C. D., Morgenstern, N. R., & Cruden, D. M. (2006). Groundwater movements of earth slides in the Thompson River Valley. Vancouver, British Columbia, Canada: Sea to Sky Geotechnique , 59th Canadian Geotechnical conference.
Fulton, R. J. (1969). Glacial lake History, Southern Interior Plateau, British Columbia. (pp. 69 – 37). Geological Survey of Canada.
Fulton, R.J. and Smith, G. W. (1978). Late Pleistocene Stratigraphy of South-Central British Columbia. Canadian Journal of Earth Science. Vol. 15, 971 – 980.
Hodge, R.A.L. and Freeze, R.A. (1977). Groundwater Flow Systems and Slope Stability. Canadian Geotechnical Journal. Vol. 14, 1977, 466 – 476.
Johnsen, T. F., & Brennand, T. A. (2004). Late-glacial lakes in the Thompson Basin, British Columbia. Paleogeolography and Evolution Vol. 41 , 1363 – 1383.
Porter, M., Savigny, K. W., Keegan, T. R., Bunce, C. M., and MacKay, C. (2002). Controls on stability of the Thompson River landslides. Proceedings of the 55th Canadian Geotechnical Conference (pp. 1393 – 1400). Niagara Falls, Ontario, Canada, October 20–23, 2002.
Ryder, J. M. (1970). The Stratigraphy and Morphology of Para-Glacial Alluvial Fans in South-Central British Colombia. Canadian Journal of Earth Sciences. Vol. 8, 279.
Ryder, J. M. (1976). Terrain Inventory and Quaternary Geology, Ashcroft, British Columbia (92 I NW). Geological Survey of Canada Paper 74-49.
Ryder, J. M., Fulton, R.J., Clague, J. J. (1991). The Cordilleran Ice Sheet and the Galcial Geomorphology of Southern and Central British Colombia. Geographie Physique et Quaternaire. Vol. 45, 365 – 377.
Stanton, R. B. (1898). The great land-slides on the Canadian Pacific Railway in British Columbia. Proceedings Civil Engineers Vol. 132, (pp. 1 – 48).
Stark, T. D., & Eid, H. T. (1994). Drained residual strength of cohesive soils. Journal Geotechnical Engineering Vol. 120 , 856 – 871.
Tribe T., (2005). Eocene Paleo-Physiography and Drainage Directions, Southern Interior Plateau, British Columbia. Canadian Journal of Earth Sciences. Vol. 42, 215 – 230.