See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/233905967

hillslope processes

Chapter · January 2010

CITATIONS READS
0 3,218

5 authors, including:

Marten Geertsema James W. Schwab

Ministry of Forests, Lands, Natural Resource Operations and Rural Development J.W. Schwab P.Geo. Eng.L Geomorphologist
149 PUBLICATIONS   2,531 CITATIONS    36 PUBLICATIONS   1,035 CITATIONS   

SEE PROFILE SEE PROFILE

Terry Rollerson
Stantec
17 PUBLICATIONS   252 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Natural dams and floods View project

Peace River area radiocarbon database View project

All content following this page was uploaded by Marten Geertsema on 22 May 2014.

The user has requested enhancement of the downloaded file.

 Chapter 8

Hillslope Processes
Marten Geertsema, James W. Schwab,
Peter Jordan, Thomas H. Millard, and
Terrence P. Rollerson

Introduction

Understanding and managing landslides and soil ciers, water, wind, and mass movement. Some sedi-
erosion can be challenging in British Columbia. ments are deposited directly by glaciers (till), some
Hillslope processes occur on a complex template, settle out in running water (fluvial), in lakes (lacus-
composed of a wide array of topographies, climates, trine/glaciolacustrine), or in seas (marine/glacioma-
geologies, and ecosystems (Chapter 2, “Physiography rine), and others are deposited by wind (eolian) or by
of British Columbia”). landslides (colluvium).
British Columbia is a diverse province with This chapter is divided into three parts. The first
mountain ranges and incised plateaus. The moun- describes landslides and landslide processes, the sec-
tainous topography is responsible for creating a ond explains soil erosion and related processes, and
variety of climates. On the Coast, mountain slopes the third examines the reading and interpretation of
support rainforests under a maritime climate, and the landscape. However, several important geo-
at higher elevations have an extensive cover of snow morphic processes are not included in this chapter.
and ice. The interior mountains are drier and colder Although snow avalanches are important hazards
in winter. In the southern Interior, some valleys in western Canada, with thousands occurring each
are semi-arid grasslands. In the northern Interior, year and claiming more than 10 lives annually, a
mountain ranges and plateaus have sporadic perma- treatment of these avalanche processes is beyond the
frost. scope of this chapter. Snow avalanches are further
Underlying these climates and topographies is a discussed in Chapter 9 (“Forest Management Effects
diverse geology. Bedrock types and surficial materi- on Hillslope Processes”), and an excellent discussion
als vary greatly throughout the province (Chapter of snow avalanche processes is given in the Ministry
2, “Physiography of British Columbia”). In general of Forests and Range’s Land Management Handbook
though, bedrock types range from flat-lying sedi- 55 (Weir 2002). Periglacial processes, such as niva-
mentary rock in the northeast, to faulted and folded tion and solifluction, are also ubiquitous in British
sedimentary rocks in the Rocky Mountains, extru- Columbia’s mountains. Although these processes
sive volcanics in the central Interior, and igneous may play a role in the priming of alpine areas for
intrusive rock on the Coast. Surficial sediments are debris flows and rock slides, a large body of literature
eroded from this bedrock and re-deposited by gla- already exists on periglacial geomorphology.

213
Landslides

What is a Landslide?

Landslide is a generic term used to describe the decomposed organic soil. Debris is usually non-
downward movement of soil, rock, or other earth plastic to weakly plastic (Hungr et al. 2001).
material under the influence of gravity. Landslides Earth refers to generally unsorted, cohesive, plas-
occur in several material types (earth, debris, rock, tic materials with a low percentage of sand-sized and
organic materials), move at varying rates (milli- larger fragments. Cruden and Varnes (1996) stated
metres per year to tens of metres per second), and that 80% of the particles in a material referred to as
can involve different styles of movement (fall, flow, “earth” should be smaller than 2 mm. Hungr et al.
slide, spread). Landslides can evolve through various (2001) proposed that the term “earth” should apply
stages of activity ranging from relict to dormant to materials with a consistency closer to the plastic
to active. They can be retrogressive or progres- limit than the liquid limit but did not specify limits
sive, advancing or enlarging, move along planar or for coarse fragment content.
curved surfaces, and be shallow or deep. In addition, Mud, a term favoured by Hungr et al. (2001),
landslides are often complex, involving more than refers to liquid or semi-liquid, clay-rich materials
one type of material and style of movement. (or “earth” with high water content). Such materials
The various types of landslides all behave dif- include clay-rich, sensitive, marine and glaciomarine
ferently and thus have different associated hazards. sediments occurring in some coastal areas, lake sedi-
The management of landslides and landslide-prone ments, and some volcanic soils.
terrain necessitated a classification scheme to enable Organic materials include both saturated lowland
intelligent and efficient communication. Several peats and thick upland forest humus forms known as
classifications are in use today. The main classifica- “Folisols” (Soil Classification Working Group 1998).
tions used in British Columbia are those of Cruden These are mainly non-mineral materials resulting
and Varnes (1996) and Hungr et al. (2001). It is useful from biogenic accumulation.
to indicate which classification system is used when
describing a landslide. British Columbia landslide Movement Type
types are also discussed in Chatwin et al. (1991) and
in B.C. Ministry of Transportation and Highways Topple refers to the forward rotational movement of
(1996). a mass of soil or rock outward from a slope about a
point or axis below the centre of gravity of the dis-
Material Types placed mass. Toppling can result from forces exerted
by surficial materials or bedrock upslope of the dis-
The terms used for landslide materials should placed mass and sometimes by ice or water in cracks
describe the displaced material as it was before the in the mass (Cruden and Varnes 1996). Because the
landslide occurred. For example, the term “rock” in moving mass is still attached to its base, toppling is
“rock slide” describes the in-place, intact bedrock considered by some to be a precursor to a landslide,
before displacement and not the rubble or debris in rather than a true landslide movement. Topples are
the landslide’s accumulation (deposition) zone. usually precursors to falls (Figure 8.1).
Rock describes landslides resulting from the fail- Fall refers to detachment of soil or rock with lit-
ure of intact, in-place bedrock. tle or no shear displacement (frictional movement
Debris refers to generally loose, unsorted material, between two surfaces) and descent mainly through
typically derived from colluvium, till, glaciofluvial the air by falling, bouncing, and rolling (Cruden and
sands and gravels, and anthropogenic materials such Varnes 1996) (Figure 8.2).
as mine waste and embankment fills. Debris typical- Slide involves movement of a relatively intact soil
ly consists of a mix of pebbles, cobbles, and boulders or rock mass along a surface of rupture or along one
in a matrix of sand, silt, and clay. Debris can contain or several discrete shear surfaces. These surfaces
significant volumes of woody material, including often form characteristic slickensides (Figure 8.3), or
branches, tree stems, stumps, and finely ground or polished shear surfaces akin to those in fault zones.

214
FIGURE 8.1 Flexural topple in limestone in the Rocky Mountains northeast of Prince George. (Photo:
M. Geertsema)

a b c

FIGURE 8.2 Examples of (a) rock fall, (b) sand fall, and (c) earth topple and fall near the British Columbia towns of
Prince Rupert, Williams Lake, and Terrace, respectively. (Photos: M. Geertsema)

215
FIGURE 8.3 Slides, such as this earth slide near Fort St. John, have discrete shear surfaces. The rupture surface here is
polished and is known as a slickenside. (Photos: M. Geertsema)

Sliding can be translational (the surface of rupture is like behaviour often depends on water entrainment
sub-planar or undulating) or rotational (the surface after the initial failure. Hungr et al. (2001) restricted
of rupture is curved and concave and is sometimes use of the term “flow” to channelized movement.
referred to as a “slump”), or intermediate or a com- Avalanche is often used to describe an unconfined
bination of the two (rotational-translational). The flow, and is especially used for rapid landslides that
displaced mass initially remains intact, but as the have long run-out distances (Figure 8.5). Debris
detached material slides further, the displaced mass avalanches and rock avalanches are commonly oc-
tends to break up. When the disrupted mass begins curring types.
to flow, the landslide is no longer termed a “slide” Spread is extension of a cohesive soil or rock mass
(Cruden and Varnes 1996) but is instead referred to and subsidence of the fractured mass of cohesive
as a “complex slide-flow.” material into a softer, underlying material (Cruden
Flow involves the movement of a mass with signif- and Varnes 1996). Spreads may result from the lique-
icant internal distortion or disruption (Figure 8.4). faction or flow and extrusion of soft or weak materi-
Flows have multiple dissipated shear surfaces distrib- als underlying competent materials. An example of
uted throughout the mass. A flow-like landslide will bedrock spread is shown in Figure 8.6.
often begin as a slide moving along a rupture sur- Complex landslides involve more than one type
face, but then continues as a flow down unconfined of material, and consequently involve more than
surfaces (e.g., debris avalanche) and (or) confined one type of movement. Large rock slides can change
channels (e.g., debris flow), and can travel for long behaviour when the displaced rock mass moves
distances. In granular materials, the initial sliding downslope and impacts other materials. Three com-
movement leads very quickly to complete disinte- mon scenarios include rock slides that transform
gration, producing flow-like motion characterized into earth flows, debris flows, or debris avalanches
by nearly complete remoulding of the moving mass (Figure 8.7).
(Cruden and Varnes 1996; Hungr et al. 2001). Flow-

216
FIGURE 8.4 Mud flows following a rainstorm in the Peace River area. Sliding in the source areas
gives way to flowing in the narrow transport zones. The landslides shown are about
100 m long. (Photo: M. Geertsema)

FIGURE 8.5 Rock avalanche near Chisca River west of Fort Nelson. (Photo: M. Geertsema)

217
FIGURE 8.6 Rock spread west of Fort Nelson. The ridges are transverse to the direction of movement. (Photo:
M. Geertsema)

FIGURE 8.7 Illustrations showing various rock slide interactions upon impact with soil. (Modified from Geertsema et al. 2006a)

What Causes Landslides?

It is important to distinguish between the causes and construction. The trigger may be an earthquake,
triggers of landslides. In the broad sense, cause is the increase in surface loading (surcharge), intense or
sum of factors that renders a slope unstable; trigger is prolonged rainfall, or river undercutting of a slope.
the final, possibly small, event that leads to the ulti- If an increase in landslides in a specific area is
mate point of slope failure. Causal factors of instabil- evident, it is useful to look for changes in potential
ity include preconditions such as weathering, glacial causal factors. Some disturbances result in large-
erosion, river erosion, geologic structure, tectonics, scale changes to forested watersheds. Wildfires (Can-
changes in climate, wildfire, deforestation, and road non 2001; Sanborn et al. 2006) and logging practices

218
(Schwab 1983; Rood 1984; Jakob 2000; Guthrie 2002) The overall effect of ground motion depends on the
are known to cause debris slides and flows. Changes topographic and geological setting. Generally, an
in climate can also be important (Chapter 19, earthquake greater than Magnitude 6 is required
“Climate Change Effects on Watershed Processes in to trigger landslides (G. Rogers, Pacific Geoscience
British Columbia”). Global climate warming may be Centre, Sidney B.C., pers. comm., 2001). The Mag-
affecting landslide rates in mountain ranges around nitude 9.2 earthquake that occurred March 28, 1964,
the world (see “The Influence of Climate on Slope in Alaska’s Prince William Sound caused landslides
Stability” below). in marine sediments and rock slides–avalanches
on mountain slopes. Near Fairbanks, Alaska, the
Landslide Triggers large inland Magnitude 7.9 Denali Fault earthquake
of November 3, 2002, preceded by a Magnitude 6.7
Numerous inherent (internal) and variable (external) event on October 23, 2002, triggered many rock falls,
forces interact to control or modify hillslope stabil- snow avalanches, and a large rock avalanche. British
ity. Principal among these factors are changes in Columbia’s short recorded history and sparsely pop-
strength–stress relationships on a slope over geologic ulated area provide only a few anecdotal reports of
time. The importance or effectiveness of a specific earthquake-triggered landslides. Vancouver Island’s
factor is determined by the local geologic and hydro- largest historic earthquake (Magnitude 7.3) occurred
logic conditions. The factors modifying slope stabil- on June 23, 1946, in the vicinity of Forbidden Plateau,
ity are discussed in numerous publications. Jakob central Vancouver Island, and caused numerous
and Hungr (editors, 2005) have provided an in-depth landslides including submarine landslides (Hodg-
presentation on debris-flow phenomena. Landslide son 1946; Mathews 1979; Rogers 1980). The largest
Investigation and Mitigation (Turner and Schuster historic earthquake recorded in Canada occurred
[editors] 1996) provides a detailed discussion on on August 22, 1949, off the west coast of Graham
large and complex landslides. In British Columbia, Island, Haida Gwaii (formerly known as the Queen
Chatwin et al. (1991) have discussed landslides in a Charlotte Islands). This Magnitude 8.1 earthquake
forest management context. Sidle and Ochiai (2006) was believed to have caused extensive rock fall,
have also provided an overview of management ef- debris slides, and debris avalanches. Geologists with
fects on landslides within a global context. the Geological Survey of Canada working on the
A landslide trigger is the external force that acts northern end of Graham Island recorded that they
on a slope to increase shear stress, which results in could not stand up during the event, which lasted
a landslide. The trigger attributed to the landslide is 5 minutes. The Prince Rupert Daily News, on August
the driving force that has set off the landslide on a 28 and 30, 1949, provided several eyewitness ac-
predisposed unstable slope. The triggers discussed counts: a fisheries patrol boat anchored at Lockeport,
within a British Columbia context are earthquakes Kulunkwoi Bay, reported the crashing of rocks and
and climate (meteorological conditions, intense rain- timber downslope, and the next day many new slide
fall and rapid snowmelt, long-term climate trends, scars were visible; a logging camp on the north shore
frost wedging, and thawing permafrost). Vallance of Cumshewa Inlet received heavy damage, and new
(2005) offered an in-depth presentation of direct and slides were reported all along Cumshewa Inlet. Me-
indirect effects of volcanic eruptions, a trigger of teorological conditions before the 1949 earthquake
large debris flows (lahars), and hence these are not likely aggravated the occurrence of slope failures.
discussed here. However, massive landslides have The field observations of Alley and Thomson (1978)
occurred in the vicinity of eruptive centres, particu- noted that many large landslide areas on the north-
larly within the Garibaldi volcanic belt (e.g., Devas- ern part of Graham Island were characterized by
tation Glacier landslide [Patton 1976]; Rubble Creek forest cover of similar age and species, suggesting
landslide [Moore and Mathews 1978]; and Dusty that rock slides and debris avalanches may have been
Creek, Mount Cayley [Clague and Souther 1982]). triggered by the 1949 earthquake or a large prehis-
toric earthquake. Schwab (1997, 1998) and Martin et
Earthquakes al. (2002) indicated that the majority of the debris
Ground shaking caused by earthquakes has trig- avalanches and debris flows on the north coast and
gered catastrophic landslides around the world. Haida Gwaii were triggered by large precipitation
Earthquakes increase stress, increase temporary events. Nevertheless, large 100- to 300-year-old land-
pore-water pressures, and decrease soil strength. slides are visible on Haida Gwaii. Likewise, many

219
pre-historic landslides were identified on Vancou- rainfall. Unfortunately, rainfall at higher elevation
ver Island (VanDine and Evans 1992). These large was probably underestimated and the time of the
landslides on Haida Gwaii, Vancouver Island, and many slope failures can only be approximated. The
the mainland were likely caused by a combination of climatological station located at Rennell Sound was
factors with a probable earthquake trigger. relatively close to the debris flow activity; however,
the general problem with landslide studies through-
The influence of climate on slope stability out British Columbia is a sparse climatological
network located at valley-bottom sites that are not
Precipitation representative of weather at the location of landslide
In a simple sense, the occurrence of landslides will activity. Satellite images, combined with weather
increase in relation to the amount and duration of radar and synoptic radiosonde,1 provide a record of
precipitation; however, slope factors may be more upper air masses and synoptic climatology and may
complicated and thus preclude a simple analysis. aid future studies in areas of sparse climatic stations
Some landslides respond rapidly to rainfall, oth- (Egginton 2005).
ers have delayed responses. In any case, antecedent Following devastating debris flows in the early
conditions have been shown to be very important. In 1980s, Church and Miles (1987) undertook a detailed
general, soils must become saturated, allowing the analysis of meteorological antecedents for case
build-up of pore pressures. Threshold pore pressures studies in the Howe Sound area and eastern Fraser
develop more rapidly in shallow materials than in Valley. A remarkably wet period in the early 1980s
deeper materials. Landslides that respond rapidly was the best correlate found for the increased debris
to changes in precipitation include shallow debris flow activity. The described meteorological condi-
slides, debris flows, and some rock falls. Landslides tions for debris-flow–triggering storms based on
that have delayed responses to changes in precipita- regional climatological stations show no exceptional
tion and temperature are usually the larger, deeper- events. Eyewitness reports, however, suggest that
seated rock slides, earth slides, and earth flows. locally intense cells of precipitation, not recorded in
the regional rain gauge network, are likely impor-
Rapid response landslides tant. Such locally intense precipitation is sometimes
In mountainous areas of British Columbia, shallow generated by convective cells embedded in a frontal
debris slides, debris avalanches, and debris flows storm. Free convection of warm air is enhanced by
are typically associated with heavy rainfall or rapid the impingement of the air stream and rapid forced
snowmelt, provided threshold conditions are met. lifting on mountain slopes. Church and Miles (1987)
Precipitation attributes that influence pore water concluded that the traditional hydrometeorological
pressure are antecedent storm precipitation, storm indices, based on routine meteorological measure-
duration, total rainfall, and precipitation intensity. ments, are unlikely to provide consistent indications
Studies from many parts of the world document the of the likelihood for debris flows to occur.
response of rainstorms of different intensity and On Haida Gwaii, Hogan and Schwab (1991) stud-
duration for various geologic settings (Wieczorek ied rainfall characteristics before and during verified
and Glade 2005). In British Columbia, only a few debris flow events. They compared the temporal
studies attempt to link type of storms, antecedent frequency of slope failures to antecedent precipita-
meteorological conditions, and precipitation in- tion trends for time scales ranging from years to
tensity and duration as triggers of shallow hillslope days. Over the longer time scale, a positive correla-
failures. However, most landside investigations use tion was found between annual moisture conditions
probable climatic triggers to describe the event. For and the reported frequency of hillslope failure;
example, the 1978 storm in Rennell Sound, Haida however, annual moisture was less significant than
Gwaii, triggered 264 debris avalanches and debris precipitation over shorter periods (months or days)
flows (Schwab 1983). A meteorological station located preceding a slope failure. On the shorter time scale,
at sea level in the centre of a study area recorded the only the months immediately preceding slope failure
cumulative precipitation over a 3-day period with were important in conditioning the hillslope to fail-
two intense 12-hour periods of 120 and 110 mm of ure. The most important situation leading to a high

1 Upper Air Observing System measures temperature, pressure, humidity, wind speed, and wind direction.

220
 frequency of slope failures was continuously wet the 6-hour rainfall intensity during a storm to be
weather before the landslide-triggering storm. Pre- significant because it causes already-wet soils to gen-
cipitation values required to trigger slope failures are erate positive pore-water pressures, which further
regularly exceeded during fall and winter months on decrease the effective strength of the soil.
Haida Gwaii and British Columbia’s north coast. Schwab (1997, 1998) used dendrochronology
Jakob and Weatherly (2003) studied landslide- techniques to determine the ages (back to the early
triggering storms on the North Shore Mountains of 1800s) of large debris slides, debris avalanches, and
British Columbia’s south coast. Data were collected debris flows in selected areas along British Colum-
for 18 storms that triggered landslides and 18 storms bia’s north coast and on Haida Gwaii. A catalogue of
that did not. Data included antecedent rainfall, storm information compiled by Septer and Schwab
rainfall intensity-duration during storm events, and (1995) as part of the study was used to verify dates
discharge exceedance of a small watershed. Through determined through tree-ring analysis of increment
discriminant function analysis, the authors sepa- cores and impact scars on trees. A conclusion from
rated landslide-triggering storms from storms that this work is that most shallow landslides occur dur-
did not trigger landslides. This analysis indicated ing a few major storms. Data from Graham Island,
that antecedent conditions are more important in Haida Gwaii, and from the Prince Rupert area on the
developing a threshold for landslide initiation than British Columbia north coast (Figure 8.8), indicate
short-term rainfall intensities alone, which fail to that six storms over the past 150 years transported
include hydrologic effects such as snowmelt and 76% of the volume of landslide debris: 9.5% (1875),
antecedent rainfall. The 28-day antecedent precipita- 14% (1891), 30.9% (1917), 6.5% (1935), 6.4% (1957), and
tion amounts are critical in bringing soil moisture 9.1% (1978). Interestingly, different watersheds (study
to saturation levels conducive to landslides. Jakob areas) did not experience the same magnitude of
and Weatherly (2003) concluded that considerable landslide activity for the various storms; for exam-
amounts of precipitation are required to saturate for- ple, at Beresford Creek on the northwest corner of
est soils to the point where a landslide is likely, once Graham Island, 36.2% of the volume transported by
an intensity threshold is exceeded. They also found landslides occurred in 1917, but no landslides were

45

40

35
Landslide volume (%)

30

25

20

15

10

0
1850 1875 1891 1917 1923 1935 1945 1949 1957 1964 1972 1975 1984 1989 1991
Years with landslides (1850–1991)

FIGURE 8.8 Percent volume transported by debris slides, debris avalanches, and debris flows on for-
ested terrain at Rennell Sound and Pivot Mountain, Haida Gwaii. Most of the sediment
was mobilized during a major storm in 1917. Landslides occurring before 1875 are prob-
ably under-represented because only large events were identified on aerial photographs.

221
 experienced in the watershed in 1978. Overall, the Prince Rupert alone (see Chapter 3, “Weather and
data suggest that British Columbia’s north coast has Climate,” Figure 3.16, for satellite photograph show-
yet to re-experience meteorological conditions simi- ing moisture transport during the October 2003
lar to the event of October 28 to November 19, 1917. event). Landslides on November 4, 2004 severed
Most precipitation on the British Columbia north a natural gas pipeline serving Prince Rupert and
coast is associated with frontal systems that pass closed Highway 16, the only highway link to the city
across the region from the Pacific Ocean (Chapter 3, (Geertsema et al. 2006a).
“Weather and Climate”). Loci of warm-wet condi-
tions shift north or south depending on the prevail- Delayed response landslides
ing storm track in the Pacific (Karanka 1986). Most Larger, and especially deeper, landslides tend to
landslide-triggering storms involve a warm front have delayed responses to precipitation. It takes time
followed by a cold front (Jakob et al. 2006). The for water to infiltrate and saturate potential slide
events appear to involve a flow in the lower part of masses. Geertsema et al. (2006a, 2007) studied large
the atmosphere, extending south to incorporate sub- landslides in northern British Columbia. They found
tropical moisture. Jakob et al. (2006) found that the that some large
following combination of meteorological conditions landslides in soil occurred after long periods of
is associated with the occurrence of debris slides, above-average precipitation. The large earth flow
debris avalanches, and debris flows: spread at Mink Creek (see “Flows and Slides in
Sensitive Clays” below) occurred after a 10-year
• warm fronts with strong SE–SW winds at the wet period (Figure 8.9). Many flow slides in glaci-
850-mb level omarine sediments occurred in the Terrace area
• warm fronts that extend south over the Pacific 2000–3000 years ago when the climate was wetter
Ocean, bringing much additional moisture to (Geertsema and Schwab 1997). This suggests that
north coastal British Columbia long-term climate trends may influence these deeper
• warm fronts with high freezing levels landslides.
• a strong jet stream with airflow arcing north of Individual rainstorms, rain-snowmelt events, and
the region and W–SW winds exceeding 90 knots outburst floods can indirectly trigger landslides by
at the 250-mb level increasing peak flows in streams. High water flows
are known to increase bank erosion. Many land-
These meteorological conditions prevailed during slides are triggered by bank erosion (Figure 8.10). An
landslide-triggering storms in 2003 and again in example is the June 2002 rain-on-snow event that
2004. The event on October 25, 2003 triggered many resulted in a flood flow from the Gillis Mountain
large landslides, with 12 reported in the vicinity of area in central British Columbia. The flood triggered

40
Cumulative deviation from mean (%)

20

0
1953 1957 1961 1965 1969 1973 1977 1981 1985 1989 1993 1997 2001
-20

-40 landslide

-60

-80

-100

-120
Year

FIGURE 8.9 Percent cumulative deviation from mean precipitation at Terrace airport from 1953 to
2002. (Modified from Geertsema et al. 2006b). The Mink Creek landslide occurred
after a decade of increased precipitation.

222
FIGURE 8.10 Small flows within a larger landslide along the Prophet River shows partial reactivation of a landslide.
The movements were triggered during a rainstorm in July 2001. The larger (deeper) landslide appears to
be influenced by bank erosion. (Photo: M. Geertsema)

numerous shallow debris slides, destabilized an al- a significant impact on the stability of slopes. For
luvial fan, and washed out a bridge. example, the sudden and delayed melt of above-aver-
age snowpacks can lead to increased occurrence of
Temperature  Landslides in mountainous regions landslide events caused by increases in pore pressure.
may be especially responsive to increases in tem- Whatever the cause, large natural landslides
perature. Two important factors may be the debut- appear to be increasing in frequency in northern
tressing of valley walls as a result of glacier melting British Columbia. When comparing the 1970s and
and retreat, and the thawing of permafrost. Recent 1980s to the last two decades, Geertsema et al.
research has shown that 20th-century retreat of (2007) reported a near tripling of rock slides and a
glaciers in British Columbia has debuttressed rock 2.5-fold increase in large soil landslides. Rock slides
slopes, causing deep-seated slope deformation, in particular have occurred in years of above-aver-
joint expansion, and catastrophic failure (Evans age temperature and at least one of the rock slides
and Clague 1994; Holm et al. 2004; Geertsema et al. was confirmed to have permafrost in its main scarp.
2006a). Mountain permafrost appears to be degrad- Thawing permafrost could contribute to the destabi-
ing, especially in the last decade, decreasing the sta- lization of slopes.
bility of slopes (Davies et al. 2001; Harris et al. 2001: Large rock slides have happened in recently degla-
Gruber and Haeberli 2007). Recent large rock slides ciated areas as well as in areas that have been ice-free
in the European Alps have been attributed to thaw for most of the Holocene. Half of the rock slides in
of mountain permafrost (Dramis et al. 1995; Bottino northern British Columbia reported by Geertsema
et al. 2002; Noetzli et al. 2003), and permafrost thaw et al. (2006a) occurred on steep rock walls above
may also play a role in initiating large rock slides in glaciers. They suggested that glacier thinning under
northern British Columbia. a warming climate following the Little Ice Age had
In addition to debuttressing and permafrost destabilized many slopes through the debuttressing
degradation, changes to the dynamics, amount, and of steep rock walls.
seasonal distribution of snow and rain can have In British Columbia, the Pacific Decadal Oscilla-

223
tion (PDO) strongly influences temperature and, to frequent rock slides. We need to extend our com-
a lesser extent, precipitation patterns in 20–30 year parison of landslide frequency and the PDO beyond
cycles (Egginton 2005; Chapter 3, “Weather and the current phase into the past and also to investi-
Climate”). The last decade of the current warm phase gate the link, if any, between landslides events and El
of the PDO, which started in 1977, was a period of Niño–Southern Oscillation.

Common Landslide Types in British Columbia

Landslides are common in mountain regions around On natural slopes, the initiation points are often
the world, and British Columbia is no exception. seepage zones (see “Indicators of Instability” in Part
The varied geology, topography, and climate result III) or concave hollows where seepage water and
in several landslide types. Shallow debris slides and hydrostatic pressures can build up. Debris slides also
flows are especially common on the coast. Rapid initiate where water is concentrated by improper
flows and spreads (flow slides) are common in some road drainage, and in unstable fillslopes.
lake sediments in northeastern British Columbia
and in glaciomarine sediments near the coast. Large Case study: Debris slides and debris avalanches in
rotational rock slides (slumps) are common in glacial organic soils
sediments and volcanic bedrock in the Interior. Rock The Wet Hypermaritime and Maritime subzones of
avalanches occur in various settings in mountainous the Coastal Western Hemlock biogeoclimatic zone
regions of the province. Many of the larger land- extend from northern Vancouver Island along the
slides are complex, involving more than one style of mainland of British Columbia’s mid- and north
movement and often more than one type of material.

Debris Slides

Debris slides are the most simple and perhaps most

common landslide type in British Columbia. The
landslides are shallow, occur on steep slopes, and
move rapidly (Figure 8.11). Shallow debris slides typi-
cally have rapid response times to rainfall, and it is
possible to establish hydrometeorological thresholds
for their initiation.
Shallow debris slides occur in various earth ma-
terials including till, colluvium, and water-deposited
sediments. On the north and central coast, debris
slides also initiate in upland organic soils called
Folisols (see “Debris slides and debris avalanches in
organic soils,” below). Rupture surfaces may develop
along a bedrock interface, and along textural change
boundaries or soil horizons, including cemented
horizons, common in Podzols, and clay-enriched Bt
horizons in Luvisols caused by soil-forming proc-
esses.
Natural debris slides tend to initiate at a point on
a slope and flare distally downslope. Debris slides
are often the feeders for other types of landslides. In
some cases, debris slides will transform into debris
avalanches, which are essentially open-slope debris
flows. If debris slides enter gullies, they tend to be- FIGURE 8.11 Shallow debris slides near Prince Rupert. (Photo:
come channelized debris flows. J.W. Schwab)

224
coast. The cool, wet climate results in extensive for- passes over the site. Landslides of this type were
ested landscapes with folic soils that develop on the documented at Kennedy Island and Kumealon Inlet
nutrient-poor plutonic rock. In the south, Folisols on the province’s north coast and at Security Inlet
(Soil Classification Working Group 1988) have also and Mackenzie Sound on the mid-coast.
developed on deep, compact, impermeable basal till. These sites generally appear relatively stable
Debris slides and debris avalanches occur in these with no apparent landslides evident on the land-
shallow organic soils (Figure 8.12). scape. At some sites, signs of periodic natural creep,
Nagle (2000) described the physical and chemi- single-tree slump blocks, or small, scallop-shaped
cal characteristics of Folisols and the terrain in depressions of an old head scarp may or may not
which they occur in the Prince Rupert area (Figure be present. The organic soils are 10–100 cm deep
8.13). Steep slopes (30–60°) with an irregular slope over bedrock (sometimes over compact till), and
configuration had the greatest influence on slope are commonly found in isolated concave pockets
stability. This geologic/hydrologic condition cre- where surface and subsurface seepage concentrates
ates many seepage exit points or locations of high (i.e., irregular or convergent slopes), but they can
hydrostatic pressure that in turn create the potential also occur on convex slopes and across larger, more
for slope failure through a reduction in soil strength. homogeneous slopes. The most obvious sites will be
These sites are often concave in shape and distinc- the poor to very poorly drained areas where seepage
tively recognized by water-tolerant vegetation and or water flow is visible at the surface. Less obvious
wet organic soil. are the moderately well to imperfectly drained sites
Slope stability problems with organic soils appear where subsurface seepage dominates (Figure 8.14).
during road construction at slope angles of 20–30°. Poor to very poorly drained sites can be identified
Problems arise when vibrating heavy machinery by the presence of skunk cabbage, sphagnum moss,
both overloads and liquefies the saturated organic goldthread, and devil’s club.
materials, triggering a landslide as the machine

FIGURE 8.12 Close-up of a moderately well-drained Folisol at a landslide head scarp (organic soil
situated on bedrock). (Photo: J.W. Schwab)

225
FIGURE 8.13 Debris avalanche in a Folisol near Prince Rupert. (Photo: J.W Schwab)

FIGURE 8.14  Debris flow initiated at a low slope angle in a poorly drained folic soil. (Photo: J.W. Schwab)

226
Debris Flows and clay to large boulders, and it often includes a
substantial proportion of organic material, from
Debris flow is a form of mass movement in which a finely shredded wood to entire trees. A debris flow
saturated slurry of soil, stones, and vegetation flows in motion resembles, and has similar fluid proper-
rapidly down a slope, usually in a confined channel. ties to, wet concrete pouring down a flume from a
The terminology used to describe debris flows and cement truck. The debris is saturated with water and
related phenomena in North America has varied flows under a fairly limited range of water contents
over the past several decades, and has therefore (sediment concentration is typically greater than
caused some confusion. In British Columbia and the 80% by weight). If the water content is too low, the
U.S. Pacific Northwest, the term “debris torrent” was debris flow will stop; if it is too high, it will separate
widely used (Swanston and Swanson 1976; VanDine into two phases of muddy water and bedload. Large
1985; Chatwin et al. 1991) to describe coarse-textured stones tend to accumulate on top of the flow and at
debris flow in steep mountain channels, and was also the front (Figure 8.15). Where the channel widens,
sometimes loosely used to refer to debris floods. The levees of boulders are often deposited along the edges
term “debris torrent” has fallen out of use and is no of the flow (Figure 8.16). Debris flows frequently
longer recommended. Other sources have applied occur as a series of surges, possibly caused by the
the term “debris flow” to all flow-like mass move- coarse front jamming behind obstructions and then
ment processes, regardless of speed or degree of releasing. In many steep mountain channels, the
confinement (after Varnes 1978), including uncon- peak discharge of debris flows is typically 10–100
fined, open-slope events that are more properly clas- times the peak discharge of normal streamflow
sified as “debris avalanches” (Fannin and Rollerson floods (Jakob and Jordan 2001).
1993). We recommend the use of Hungr et al.’s (2001) In steep channels, large debris flows can reach
definition: “Debris flow is a very rapid to extremely high speeds—5–20 m/s is typical—and have great
rapid flow of saturated non-plastic debris in a steep erosive power (Figure 8.17); however, on losing
channel.” The distinguishing attributes of debris confinement on a fan, they move much more slowly,
flows, according to this definition, are: rapid speed covering or engulfing objects, often without damag-
(usually greater than 1 m/s), saturation, confine- ing them. They can surround automobiles without
ment in a channel, and relatively coarse texture (low breaking windows or scratching the paint (Figure
clay content). The term “mud flow” is used for rapid 8.18).
flows of fine-textured, cohesive (clay-rich) material Coarse-textured debris flows tend to slow down
(defined by Hungr et al. 2001 as having a plasticity and stop when the slope reaches about 15% (9°),
index > 5%). although this critical slope varies somewhat, de-
Debris flows are probably one of the most im- pending on the magnitude of the event and its water
portant mass movement processes in the amount of content. Fans that are formed primarily by debris
geomorphic work done. In some areas of the prov- flows tend to approximate this gradient (Figures 8.19,
ince, rock avalanches and earth slides overwhelm 8.20). Fine-textured debris flows (those containing
sediment budgets, but overall it is the debris flow more clay) can traverse more gentle slopes, as low
that moves most of the sediment produced by weath- as 2% for large volcanic mud flows (Figure 8.21). The
ering, rock fall, and small landslides in steep terrain deposits of debris flows on alluvial fans can often be
into valley bottoms. Debris flows occur at all scales, recognized by levees and irregular lobes of bouldery
ranging from tiny flows a few centimetres deep that debris (Figures 8.22, 8.23), and by unsorted layers
can be observed on any road cut during a spring of debris that are matrix-supported (large clasts not
thaw, to rare events that can carry a cubic kilometre touching), and that contain more silt, fine sand, and
or more of debris from a collapsing volcano over a wood fragments than streamflow deposits (Costa
hundred kilometres down a low-gradient river valley 1984; Figure 8.24).
(Costa 1984; such volcanic-source debris flows are Debris flows can originate in a number of ways
often called “lahars”). Most of the debris flows of in- (Hungr et al. 2005). The most common in the for-
terest in a forestry context range from a few hundred ested mountain landscapes of British Columbia is
to a few tens of thousands of cubic metres. for a small landslide to enter a steep stream chan-
Debris flows have a number of distinctive prop- nel (Millard 1999). The resulting debris flow can be
erties (e.g., VanDine 1985; Pierson and Costa 1987). many times larger than the originating landslide,
The debris is very poorly sorted, ranging from silt as debris is eroded from the channel bed and banks

227
FIGURE 8.15  Bouldery debris flow deposits at Lower Arrow Lake near Burton. (Photo: P. Jordan)

when the debris flow moves down the channel. Most

open-slope debris slides lack sufficient water content
to flow under their own weight. A landslide into a
channel may dam the stream, and the dam soon
breaks, producing a slurry with the proper water
content to flow as a debris flow. Alternatively, a
rapidly moving landslide may overtake and consume
water flowing in the channel, transforming into
a debris flow. Another common mechanism that
enhances transport is that unusually high stream-
flow in a small, steep channel may undercut the
streambanks, causing a temporary dam, which when
failure occurs, results in a debris flow. Yet another
mechanism (well described in the literature from
Japan, where higher rainfall intensities occur than
in Canada) is that unusually high streamflow in a
small, steep channel may cause the channel bed itself
to fail, reaching a sufficiently high sediment concen-
tration through entrainment of streambed and bank
material to transform into a debris flow (Takahashi
1981).
In western Canada, debris flows tend to be re-
stricted to channels with drainage basins of about
1–10 km2, which contain channels steeper than about
FIGURE 8.16 Steep debris flow fan at Salal Creek near 50% in their upper reaches, and which are nowhere
Pemberton. (Photo: P. Jordan) less steep than about 15% (VanDine 1985). A simple

228
FIGURE 8.17 Channel stripped of vegetation and alluvial sediment by a large debris flow in the Monashee Mountains
near Edgewood. (Photo: P. Jordan)

FIGURE 8.18 Vehicle engulfed by slow-moving, distal deposit of a debris flow, which damaged a logging camp at
Meager Creek near Pemberton. (Photo: P. Jordan)

229
FIGURE 8.19 Aggrading alluvial fan formed by debris flows and fluvial sedimentation in the Monashee Mountains
near Edgewood; note former channel and avulsion caused by recent debris flow. (Photo: P. Jordan)

index of debris flow susceptibility is that the “Melton

ruggedness number” (the ratio of elevation range
to the square root of watershed area, in consistent
units) be greater than 0.6 (Wilford et al. 2005b).
The time of occurrence of debris flows is quite un-
predictable, although the susceptibility of a channel
to debris flow hazard can be assessed fairly reliably
from evidence on the fan and in the channel, and
from geomorphic conditions in the watershed (Jakob
and Jordan 2001). For a debris flow to be initiated,
enough sediment and enough water must arrive at
the same place and time in a channel that is suffi-
ciently steep and confined to support debris flow. Al-
though heavy rainstorms and peak snowmelt events
are responsible for triggering most debris flows,
many debris flow events are triggered by rainfall or
snowmelt that is not unusually heavy; extreme rain-
storms fail to trigger debris flows most of the time in
most susceptible channels. Perhaps some random oc-
currence, such as a small debris slide or bank failure
caused by the peak flow, is necessary to add enough
sediment to trigger a debris flow. Alternatively,
when released, a transient flow blockage (a fallen
FIGURE 8.20 Alluvial fan formed by frequent debris flows into tree, woody debris, or snow) may increase the peak
the Ryan River near Pemberton. (Photo: discharge to a level sufficient to entrain sediment by
P. Jordan) scouring the channel. The volume of loose, entrain-

230
FIGURE 8.21  Repeated debris flows have deposited sediments near the Zymoetz River. (Photo: M. Geertsema)

FIGURE 8.22 Levees formed by debris flows originating in fractured sedimentary rocks at Fountain Ridge near Lillooet.
(Photo: P. Jordan)

231
FIGURE 8.23 Levee formed by debris flow originating in granitic rocks at the Upper Lillooet River near Pemberton.
Largest boulders are 5 m in diameter. (Photo: P. Jordan)

FIGURE 8.24 Debris flow deposit showing inverse grading (larger stones toward top) and matrix support (stones not
touching, surrounded by fine material) at Meager Creek near Pemberton. (Photo: P. Jordan)

232
able material stored in a steep channel provides some as Na+ and Ca2+ in seawater. The resulting sediment
indication of the possible volume of a future debris has an open structure with high water content. The
flow, assuming all the material is mobilized (Hungr positive charges of the salts maintain the interparti-
et al. 2005). cle bonds.
An almost unlimited supply of debris is possible As the glaciers melted, the land began to rebound
in watersheds dominated by active slope processes isostatically, rising as much as 230 m above present-
such as rock fall (Bovis and Jakob 1999). In such wa- day sea level (McCuaig 2000). This exposed the
tersheds, debris flows may occur annually, or more glaciomarine sediments to rainfall and groundwater.
often if rainstorm events generate enough stream- Salts in the clays were gradually leached out of the
flow; however, in most forested watersheds or alpine sediments. Salt content would decrease from an ini-
areas with competent rock, debris flows are infre- tial 30 g/L to below 1 g/L. With a lower salt content,
quent. On many steep stream channels, debris flows repulsive forces between particles increased, leaving
occur at average intervals of several decades on the the saturated, porous sediment prone to collapse.
Coast, to several centuries in the Interior. Follow- An imposed load, vibration, or bank erosion
ing debris flow events, stream channels go through a can trigger collapse of the sedimentary structure in
cycle in which debris in the channel is slowly replen- sensitive glaciomarine deposits, causing liquefaction.
ished by rock fall, tree fall, and soil creep (Jakob et During liquefaction, the weight of the soil is trans-
al. 2005). After several decades or centuries, a critical ferred from the solids to the pore water.
hydrologic event occurs that triggers a debris flow,
and the accumulated debris in the channel is carried Examples of landslides in sensitive clays
away. Disturbances in a watershed, such as logging, In preparation for widening the British Columbia
road construction, or wildfire, can cause hydrologic highway between Kitimat and Terrace in 1962, earth
changes that lead to greater peak streamflow, or they fills were placed beside the highway on sensitive
can cause small slope failures that enter the channel. clays soils. When the loads failed, two enormous
Both these changes can greatly increase the prob- landslides occurred May and June. These landslides,
ability of a debris flow in a channel. which both flowed into Lakelse Lake, destroyed sec-
tions of the highway (Figure 8.25).
Flows and Spreads in Sensitive Clays In winter 1993/94, a similar landslide flowed into
Mink Creek (Figure 8.26), a tributary of Lakelse
Rapid landslides may occur on extremely low gradi- River, moving 2.5 million m3 of mud. The land-
ents in some low-lying areas of the British Columbia slide failed on a nearly flat slope of just over 2°. The
coast. They occur in fine-textured sediments that landslide likely began with a small failure along the
may be prone to sudden liquefaction. The sediments creek, which exposed sensitive clays further in the
can be relatively strong in the undisturbed state, but slope. Once the sensitive materials were exposed, the
a small vibration or load can make them flow like landslide retrogressed rapidly, eating its way into the
wet porridge. Such materials are referred to as “sen- slope.
sitive clays.” In the extreme states, when the undis- Landslides of this nature often move as spreads
turbed strength is more than 30 times greater than or flows. Flows tend to occur in materials with
the remoulded strength, such sediments are called lower shear strengths than spreads. Flows may also
“quick clays.” occur where a thick zone is sensitive. Spreads are
The British Columbia coastline was submerged more likely to occur where a thin zone is sensitive.
by the weight of a large ice sheet during the last In flows, ridges (if they occur) are oriented paral-
glaciation (Clague 1989). As the ice retreated, the lel to movement direction. In spreads, ridges are
sea migrated inland with the retreating ice fronts. transverse to movement. The Mink Creek landslide
Glaciomarine sediments composed of rock flour, silt, was complex, displaying both spreading and flowing
and clay minerals were deposited in the sea. In fresh- behaviour (Figure 8.27).
water, clay particles tend to settle much more slowly In 2003 a large flow in sensitive clays ruptured
than the larger silt particles. In salt water, clays and a natural gas pipeline between Terrace and Prince
silts aggregate together forming floccules and settle Rupert (Schwab et al. 2004). The landslide cut off the
in a random orientation. Negative, repulsive charges natural gas supply to Prince Rupert for 10 days. The
on the clay particles are neutralized by cations such landslide also dammed Khyex River.

233
 a

FIGURE 8.25 One of two destructive, sensitive clay landslides at Lakelse Lake,

1962 showing a) fluid mud in the zone of depletion, b) debris
accumulation in Lakelse Lake, and c) damage to vehicles.
(Photos: Terrace Library)

234
FIGURE 8.26 The Mink Creek earth flow spread illustrates a rapid failure on a nearly flat gradient. (Photo:
M. Geertsema)

FIGURE 8.27 Transverse ridges of spreading (S) give portions of the Mink Creek landslide a ribbed appearance; the
central part of the slide experienced more complete liquefaction and became a flow (F). (Modified from
Geertsema and Torrance 2005)

235
Large Earth Flows water fluctuations, and have been more active in the
past, when climate was wetter, than at present.
Large, slow-moving earth flows, involving millions
of cubic metres at individual sites, are common Rock Avalanches
along the incised valleys of interior British Colum-
bia’s dissected plateaus (Bovis 1985). Some gigantic Rock avalanches involve the initial failure and
landslides, such as the Sheslay slide in northwestern subsequent disintegration of large rock masses on
British Columbia (Souther 1971), cover tens of square mountain slopes and the rapid downslope movement
kilometres and may involve billions of cubic metres of this debris to a lower slope. Documented cases
of material. These large landslides occur in weak, involve volumes of more than 10 million m3. These
fine-textured sedimentary and volcanic bedrock, or landslides often attain velocities in excess of 100 
in the softened materials derived from the bedrock. km/hr, sometimes exceeding 300 km/hr. The veloc-
Typically, the landslides have lateral ridges and ity of rock avalanches depends not only on slope
lobate tongues and often have transverse ridges gradient, but also on the thickness of the debris
(Figure 8.28). Although these landslides resemble wave; thus, greater velocities may be attained where
flows, movement generally occurs along discrete the debris is channelized, and lower velocities where
shear surfaces, with the body of the landslide mov- it is able to spread and thin.
ing as a rigid plug. For this reason, the movements Catastrophic movements in rock avalanches last
may more accurately be described as “slides” rather for a matter of minutes. The conditions that weaken
than flows, and some researchers refer to these types a rock mass for catastrophic sudden failure can take
of landslides as “mudslides” (e.g., Hutchinson 2004; thousands of years. Globally, a common triggering
Picarelli et al. 2005). Most of these landslides display mechanism for rock avalanches is seismic activ-
very slow seasonal movement in response to ground- ity (Keefer 1984). The historic rock avalanches in

FIGURE 8.28  Pavilion earth flow; note the transverse ridges in the zone of accumulation. (Photo: J. Ryder)

236
northern British Columbia have not been associ- Downslope dipping bedrock
ated with earthquakes (Geertsema et al. 2006a). For Long run-out rock slides are sometimes initiated
some, there have been no apparent triggers. Other on dipping sedimentary rock slopes, particularly
movements have been associated with intense and in fault zones. West of Fort Nelson, two rock ava-
prolonged rainfall. lanches had detachment zones on 27–36° west-dip-
An important consideration is not only the trig- ping strata (Geertsema et al. 2006a). Exposed rock
ger, but also the preconditions that make the slopes on these dip slopes tended to be weak shale but
susceptible to movement. These preconditions occasionally sandstone would be exposed (Figure
include: 8.31). On steeper slopes, failures would occur more
continually, preventing the conditions for a single
• development of fractures or joints dipping movement of a large volume of rock. If the slopes are
downslope; too shallow, the rock masses will not gain enough
• progressive reduction of cohesion along shear energy to travel at a high velocity.
planes;
• development of low friction angles and low re- Deforming mountain slopes
sidual strength along weak planes; and Mountain slope deformation is common in northern
• vertical relief and steep slopes. British Columbia. These sagging slopes, or sackun-
gen, are often the result of glacial debuttressing fol-
In British Columbia, rock avalanches have oc- lowing deglaciation (see “Sackungen” below). These
curred in the following major settings (Figure 8.29): features are especially common in weak volcanic and
sedimentary bedrock. Slope sagging may also be at-
• cirque walls above glaciers tributed to differential zones of weakness in bedrock,
• on sedimentary dip slopes weathering, and tectonic activity.
• in association with slow mountain slope deforma- During the last glaciation, glaciers gouged out
tion valleys, putting stress on rock masses. When the
glaciers melted, isostatic adjustments of the Earth’s
Cirque wall failure crust and lateral adjustments from valley walls
Cirque walls tend to be steep and are often frac- caused some mountain slopes to sag. Such sagging is
tured. Fracturing probably occurs as a result of stress often a precursor to rock avalanching (Figure 8.32).
release such as the removal of glacial ice buttressing, Most rock avalanches are complex landslides
or from exposure to modern weathering processes that frequently start out as frictional rock slides,
such as freeze–thaw cycling. A 1999 rock avalanche but then transform into debris avalanches or debris
involving rock fall from a cirque wall occurred in flows when impacting and entraining soil lower on
the Rocky Mountains near McBride (Figure 8.30). the slopes. Run-out models that incorporate these
The landslide was associated with heavy, prolonged various phases of rock slides have been developed
rainfall (Geertsema et al. 2006a). in British Columbia by Oldrich Hungr and his PhD

FIGURE 8.29 Illustration showing the settings for rock avalanches in British Columbia. (Modified from Geertsema et al. 2006a)

237
FIGURE 8.30 Rock fall from this cirque wall above the Kendall Glacier, 45 km northwest of McBride,
disintegrated and transformed into a rock avalanche with more than 1 km of run-out
in July 1999. The landslide was associated with prolonged, heavy rainfall. (Photo:
C. Erickson)

FIGURE 8.31 The Tetsa rock avalanche was triggered by disintegrating rock masses on dip slopes of
approximately 30° in the Rocky Mountain foothills; note the run-up on the valley wall
opposite the landslide. (Photo: M. Geertsema)

238
 a

FIGURE 8.32 The 2002 Pink Mountain rock slide–debris avalanche was associated with mountain
slope deformation (Geertsema et al. 2006c). (a) Note ongoing deformation on the
mountain, and (b) an older scarp and tension fractures above the 2002 landslide.
(Photos: M. Geertsema)

239
students, such as Scott McDougall (McDougall and slope scarps shows signs of movement, and little
Hungr 2004). growth occurs at the base of the troughs along the
fractures.
Sackungen  Sackungen is a German term defined Dead, fallen trees stretch across the troughs, some
as a deep-seated rock creep that produces ridge-top showing signs of excessive bow before falling (Figure
trenches through the settling of a stable mass into an 8.33). Converging and diverging ridges are found
adjacent valley (Zischinsky 1969). It denotes “slope where troughs meet or separate, creating a maze on
sagging,” “ridge spreading,” gravitational spreading, the hillslope. Along ridges, bowed trees are leaning
or deep-seated gravitational slope deformation. It is both ways in an effort to grow upright (Figure 8.33).
used to describe anomalous linear features, includ- Split trees and trees with healed cracks occur in
ing upslope-facing (antislope) scarps and troughs some troughs, indicating recent, abrupt movement.
trending parallel to slope contours. These linear Discerning sackungen features in a forested land-
features are found in mountainous landscapes and scape from aerial photographs or by flying over the
are often associated with the retreat of glacial ice terrain is difficult because many of the features are
from valleys leaving slopes over-steepened and un- visible only on the ground. Nevertheless, the follow-
supported (Bovis 1982, 1990; Bovis and Evans 1995, ing larger features are evident on aerial photographs.
1996). Sackungen may also be the precursors of cata-
strophic rock avalanches (Geertsema et al. 2006a). • Lineations in the forest cover extend parallel to
Sackungen features were identified on a forested the contours of the slope (Figure 8.34). These are
slope in the Kitnayakwa River drainage of west-cen- breaks in the forest cover where tree growth has
tral British Columbia (Schwab and Kirk 2003). The been prevented by the fractures in the ground, or
sackungen features at Kitnayakwa create giant steps where scarp slopes are too steep for tree growth.
in the landscape, with moss-covered downslope • Sharp elevation breaks in the hillslope are some-
scarps up to 30 m high dropping down to meet anti- times accompanied by a rise in elevation at the
slope scarps up to 10 m high (Figure 8.33). bottom of the break. These breaks indicate larger
The scarps and troughs between scarps trend sackungen, and the accompanying antislope
roughly parallel to slope contours. A distinct frac- scarp.
ture line runs along the trough in most features. • An over-steepened slope at the base of the hill
Larger ridges are situated approximately 150 m apart breaks sharply onto the valley flat where no
with a trough up to 25 m wide. Trees are growing evidence of this type of slope is seen in the sur-
in the trough as well as on the steeper ridge crests. rounding terrain.
Elongate pools of standing water are also found in • A small pond present at the base of a slope may
some troughs. Smaller ridges in-between the larger show a dissimilarity from neighbouring slopes.
ridges are spaced roughly 25 m apart. Trenches up • Ridge crest depressions are often parallel to the
to 7 m wide and 3 m deep cut across the hillslope at slope.
regular intervals. Vegetation on the smaller, anti-

FIGURE 8.33  Antislope scarp and ridges indicative of slope sagging near the Kitnayakwa River. (Photos: J.W. Schwab)

240
 The sackungen features at Kitnayakwa signify
an unstable, sagging slope. Another landslide
could occur at any time in this deeply weathered,
over-steepened, unsupported glacial debuttressed
slope through seasonal and climatic cycle effects on
groundwater and (or) from earthquake shock.

Complex Landslides

If one examines any landslide closely enough,

complexity will be found; however, some landslides
clearly involve distinct materials and separate move-
ment types. For example, rock slides will often trig-
ger much larger secondary movements in soil, which
would not have been possible without the initial
slide. Two examples are presented below.

Rock slide – earth flow

Sometime around 1979, a rotational rock slide
(slump) involving an estimated 2–3 million m3 of
material triggered an earth flow of 15 million m3 in
a clay-rich diamict near Muskwa River, west of Fort
Nelson (Figure 8.35). The landslide travelled more
FIGURE 8.34 Old landslides (blue outline) and transverse than 3 km on a 3.5° slope (Geertsema and Cruden
sackungen ridges (red) at the Kitnayakwa 2008). Such a long, low-gradient movement in this
River. The landslides are about 220+ and 460+ type of soil is impossible without invoking some type
years old. The largest landslide involved about of dynamic loading (Geertsema and Schwab 2006).
9.5 million m3 of material. The bedrock is a Hutchinson and Bhandari (1971) explained such
highly fractured and crumbling, well-bedded, behaviour by a process called “undrained loading.”
fine-grained, crystal-lithic tuff. (Photo interpreta- In this case, the rock slide became a dynamic load
tion: J.W. Schwab) for the clay soil. The load transfer caused pore-water

FIGURE 8.35  Muskwa rock slide–earth flow, west of Fort Nelson. (Photo: M. Geertsema)

241
pressures in the soil to increase, and because the Most of the rock rubble was left upslope of this. The
groundwater could not readily escape from clay-rich thinly spread debris avalanche entered a gully, which
soil, liquefaction within the clay occurred, mobiliz- confined the moving mass into a debris flow.
ing the mass to instability in what would otherwise
have been relatively stable ground. Landslide-generated Tsunamis

Rock slide – debris avalanche – debris flow In British Columbia, landslide-generated tsunamis
In late June 2002, a rock slide west of Smithers im- are not well recorded and the majority likely remain
pacted glacial till above Harold Price Creek (Figure unknown. Two main types have been documented:
8.36). After travelling 1.3 km, the rock slide trans- (1) displacement waves generated by subaqueous
formed into a debris avalanche, and after 2.2 km the (underwater) landslides, and (2) those created by
landslide entered a gully and changed into a debris subaerial (on land) landslides.
flow, travelling a total of 4 km in three different
phases. The initial frictional rock slide impacted, Tsunamis from subaqueous landslides
partially liquefied, and entrained glacial sediments, Active fans and deltas are high-energy fluvial systems
transforming the movement into a debris avalanche. subject to continuous and progressive sediment
loading. Where these deposits are underlain by soft
muds, they may be prone to large-scale slumps,
a earth flows, or flow slides, as a result of site loading
(natural or human-caused), vibration from heavy
machinery, undercutting, interruption of intertidal
drainage, and high artesian pressures.
Fan-deltas are common along the sidewalls of
coastal fjords and also on steep-sided, long, narrow,
fjord-type lakes in mountainous areas of British
Columbia. The geomorphology of fan-deltas has
been studied and described in coastal fjords (Prior
and Bornhold 1988, 1990; Bornhold and Harper
1998). The processes are similar in freshwater lakes.
Fan-deltas have developed over the last 8 000–
12 000 years, commencing with the high-energy,
sediment-laden meltwater during deglaciation of
fjord-side drainage basins. Rivers continue to supply
b sediment during spring freshets and floods with the
aggradation of sediments over the delta through
multiple distributary channels on its surface with
deposition in various layers of cobbles, gravels, and
sands. Materials at depth generally include complex
layers of interstratified muds, sands, and gravels. A
fan-delta at first glance often appears as a relatively
flat, unassumingly stable landform, but evidence
of a historic failure is sometimes revealed as a
steep, exposed head scarp along the fan-delta face.
Large portions of fan-delta surfaces are submerged.
Submerged materials perched on the steeply sloping
faces are generally highly unstable. The evidence of
landslide activity on fan-deltas is often concealed
FIGURE 8.36 A complex landslide (2002) at Harold Price because landslide features are under water; however,
Creek, near Smithers: (a) a rock slide trans- the delta front can be viewed through side-scan
formed into a debris avalanche, which in turn sonar imagery. Debris slides and debris flows are
transformed into (b) a debris flow, travelling a displayed as a slight trough and debris lobe. The
total distance of 4 km. (Photos: M. Geertsema) presence of crescent-shaped slide scarps and blocky

242
debris ridges suggests the occurrence of a retrogres- development, a detailed investigation of the fan or
sive rotational type of landslide movement (Figure fan-delta must be carried out to ensure stability and
8.37). safety of the sites.
Large fan-deltas are tempting sites for waterfront In part because of the remote location, little docu-
development because of their relatively flat terrain. mentation exists of landslides that have occurred on
The possibility of landslide activity and hazard is fan-deltas; however, a few destructive landslide-tsu-
often not considered because the evidence is covered nami events within British Columbia coastal fjords
by water. However, working with equipment or the and on fjord-type freshwater lakes associated with
placement of structures close to the fan-delta water fan-deltas have been recorded. For example, in 1975,
edge may cause disastrous slope failures. Before a large submarine flow slide occurred on the front
of a fjord-head delta in Kitimat Arm. The failure
(volume estimated at up to 55 million m3) happened
about 1 hour after an extreme low tide (Prior et al.
1984). According to Murty (1979), the slide generated
a tsunami up to 8.2 m high, causing considerable
property damage. Site loading (construction activi-
ties) under conditions of high pore-water pressures,
created by the extreme low tide, was thought to have
triggered the slide.
Damaging landslides are not restricted to marine
coastlines. On October 6, 1998, a fan-delta collapse
occurred in Troitsa Lake, near Houston. Schwab
(1999) estimated the volume of the largely underwa-
ter slide to be 3 million m3. The resultant displace-
ment wave was 1.5 m high and it damaged dock
facilities. It travelled 1 km to the opposite side of
the lake where it caused trees to topple. A backwash
wave about 2 m high travelled back to the head scarp
area carrying debris up to 150 m inland. Waves oscil-
lated on the lake for 4 hours.
The cause of sediment failure along the delta front
is unknown; however, human activity and seismic-
ity were ruled out. Interestingly, a row of horizon-
tally lying trees, approximately 650 years old, were
exposed at the head scarp, indicating a past subaque-
ous slope failure event that was likely a landslide-
generated tsunami.

Tsunamis from subaerial landslides

Some landslides discharge into water bodies and
generate displacement waves (a type of tsunami).
The most spectacular events occurred in Lituya Bay,
Alaska, in the last century. The most recent event
(1958) occurred when a large rock avalanche entered
the bay. The landslide generated a displacement wave
of more than 500 m high that stripped the forest
from the opposite mountain slope and the lateral
shorelines as it travelled down the bay (Pararas-
Carayannis 1999).
FIGURE 8.37 Illustration showing subaqueous landslide A large rock slide (4 million m3) entered Knight
associated with fan-delta collapse. (Modified Inlet sometime in the mid-1500s, dropping more
from Schwab 1999) than 800 m vertically (Bornhold et al. 2007). The

243
resulting tsunami destroyed the village of Kwalate, A large, jointed mass of quartz diorite estimated at
and about 100 (or possibly more) of its First Nations 3 million m3, slid about 450–600 vertical metres into
inhabitants. The wave height was estimated at 2–6 m. the lake. When the rock mass entered Chehalis Lake,
In 1946, a Magnitude 7.2 earthquake triggered a it caused at least one wave that eroded the shoreline
1.5 million m3 rock avalanche from the north face in some places to a height of more than 18 m. At the
of Mount Colonel Foster on Vancouver Island that south end of the lake, 8 km from the slide location,
entered Landslide Lake (Evans 1989). The resultant the maximum wave height was 6.5 m. Trim lines
tsunami wave ran up a maximum height of 51 m, were observed in the Chehalis River downstream
washing away the forest in its path. of Chehalis Lake, indicating that a wave contin-
The most recent landslide-generated tsunami ued down the river. The displaced water continued
occurred in 2007 at Chehalis Lake, about 40 km downriver at least 14 km to the Chehalis alluvial
north of Chilliwack (Figure 8.38). The event results fan, where large volumes of fresh woody debris were
were first observed December 6, 2007, with the event evident. Landslides entering rivers can also create
likely occurring December 3–4 during heavy rains. displacement waves.

FIGURE 8.38 The 2007 Chehalis rock slide triggered a large displacement wave. (Photo: B.C. Ministry of Forests and Range,
Recreation, Sites and Trails Branch)

244
Erosion

Types of Erosion

Erosion is generally considered the gradual wearing dislodged, the soil particles may be easily washed
away of the Earth’s surface by water, wind, glaciers, downslope by sheet erosion.
and frost action. Erosion is a generic term for a Sheet erosion involves the removal of thin layers of
group of processes in which materials are loosened surficial material by broad, continuous, uniform flow
or worn away and simultaneously moved from one over a slope. Rill erosion is the development of small,
place to another. Erosion is intimately related to closely spaced channels resulting from surface soil
chemical and physical weathering of rock and other removal by running water. Rill erosion is intermedi-
materials on and below the surface. Forest fires, ate between sheet and gully erosion (Figure 8.40).
logging, grazing, tillage, mining, road building, and Erosion is concentrated at these sites and may result
other phenomena that remove surface vegetation can in both downward and headward erosion. When
increase erosion rates. Wave action is an important a non-erodible base is encountered, lateral erosion
component of shoreline erosion, and both ice jams tends to occur. With time, rill erosion may develop
and high streamflow erode riverbanks. Erosion is ac- into gully erosion (Figure 8.40) (see “Gully Proc-
companied by transport and sedimentation. esses,” below).
The geomorphic processes of soil erosion and The erosion of an area depends on several fac-
landsliding exist along a continuum. In this section, tors relating to soil, slope, vegetation, climate, and
we consider various forms of water erosion ranging erosion control practices. The Universal Soil Loss
from splash erosion to rill erosion. We also consider Equation (Wischmeier and Smith 1960) may be used
piping (a type of subsurface erosion), washouts, gul- to calculate erosion (soil loss):
lies, and fire-generated erosion leading to landslides.
X=R×K×L×S×C×P
Rain splash erosion occurs when soil particles
are dislodged and lifted into the air upon impact by where: X = soil loss, R = rainfall factor, K = soil
raindrops (Figure 8.39). The level of erosion depends erodibility index, L = slope length factor, S = slope
on the intensity and kinetic energy of the rainfall gradient factor, C = cropping or vegetation factor,
and the cohesion of the soil. Splash erosion results and P = erosion control practices factor.
in crusting, detachment, and transport of soil. Once Rainfall factor (R) is based on the sum of storm

FIGURE 8.39 Splash erosion: the raindrop is a few millimetres wide. (Photo: M. van Hees)

245
FIGURE 8.40 Rill and gully erosion along the Chilcotin River. Gullies are eroded into weakly cemented
glaciolacustrine deposits (right) and into shallow rills in less cohesive colluvial deposits (left). Both
gullies and rills on the lower 20 m of the slope were washed away after inundation from a landslide
dam. (Photo: M. Geertsema)

energy and intensity and will generally be specific control practices factor (P) depends on the applica-
to a given region or area. Soil erodibility index (K) tion of non-vegetative erosion control practices,
provides an estimate of how erodible the local soil is such as installation of check dams or silt fencing and
and, therefore, will be site-specific. In general, soils surface roughening. Any such treatments can reduce
with low cohesion (e.g., sandy soils derived from this factor.
fluvial or glaciofluvial deposits) will have a high in- Hydraulic connectivity refers to the “connection”
dex, and soils with high cohesion (e.g., silty or clayey of hillslope areas to larger creeks or streams beyond
soils derived from basal tills) will have a lower index. the toe of the hillslope. In areas of direct hydraulic
Slope length factor (L) is determined by the length of connectivity, any sediment that reaches a hillslope
the slope at a specific slope gradient. This slope gra- creek is transported directly downstream at a sig-
dient defines the slope gradient factor (S). Together, nificant gradient (i.e., greater than 5%) to locations
these factors (LS) may be considered the “topograph- where it may result in adverse effects to water quality
ic” factor. Long, steep slopes will be more susceptible or aquatic resources. For areas with no connectiv-
to erosion than short, gentle slopes because of the ity, the hillslope stream must flow into a swamp or
greater volume of water running off the slope and lake and trap sediment. Indirect connectivity may
a higher flow energy. The topographic factor will occur where the hillslope stream flows through a
depend on site conditions; however, it may be pos- lower-gradient reach (typically less than 5% gradient
sible to alter the topographic factor through treat- for a minimum length of 100 m) before connecting
ments such as re-sloping or terracing. The cropping with any stream reach with water quality or resource
or vegetation factor (C) reflects the cover provided values.
by vegetation or other material (e.g., mulch, erosion Surface erosion from rain splashing and sheet
control matting), with higher cover yielding a lower wash delivers fine sediment to streams. Landslides
cropping factor. Appropriate revegetation treatment such as debris flows tend to deliver coarser materials.
can lead to a reduced cropping factor. The erosion

246
Gully Processes A fan is often located at the mouth of the gully
unless the gully discharges into a larger stream or
Gullies are small, steep, and incised drainages on lake. The fan is a conical-shaped deposit formed by
hillslopes. Gullied terrain may indicate a geomor- sediment and woody debris discharged from the
phically active area that is sensitive to forest man- gully. Where the gully channel discharges onto the
agement activities. Landslides common in gullies fan, it may be incised into the fan deposits or it may
include debris slides and debris flows. With increas- flow across the fan surface. Channels that are incised
ing water concentration, debris floods and floods into the fan surface may emerge onto the fan surface
also occur. at a location further downslope. Where a channel
flows across the surface of a fan with little incision, it
Gully Morphology may avulse (a complete or partial change in channel
In general, two types of gullies occur in British location). On many fans, multiple channel locations
Columbia. The first type is incised into deep surficial are the result of avulsions. Many fans below gullies
materials, typically glacial till, and glaciofluvial or are subject to debris flow deposition. Many gullies
glaciolacustrine sediments. The second type is in- are relict features of deglaciation, particularly in the
cised into bedrock. Often gullies formed in bedrock Interior, and are no longer active. Relict fans (“pa-
are the result of differing bedrock types, competency raglacial fans”) below these gullies may have had
and structure, and groundwater conditions. In some little or no sedimentation for millennia; however,
locations, smaller gullies lie parallel to each other development can re-activate gully processes. In some
across an open slope (dissected terrain), and other cases, fans are a combination of relict surfaces and
gullies are tributary to one another, forming den- active surfaces. Additional information on fans is
dritic gully systems. available in Wilford et al. (2005b) and Millard et al.
Gullies are divided into a number of zones: the (2006).
headwall, sidewall, transport zone (or channel base),
and mouth. Often a depositional fan is present Landslides in Gullies
downslope of the gully mouth. The headwall is the Within gullies, the steep slopes and drainage con-
uppermost part of the gully system, and is typically centrations result in a range of possible hazards.
a concave-shaped area shallowly incised into the Steep headwalls and sidewalls are often sites of
hillslope. Headwalls in unconsolidated materials fre- slope failures (usually debris slides). Debris slides
quently exhibit soils with moderate or poorer drain- that enter the gully channel may continue down the
age and are often the site of slope failures. Headwalls channel as a debris flow, a result of the contribution
can be subtle features and are not always recognized. of water from the channel.
The transport zone of the gully is what many Almost all gully debris flows in British Columbia
people recognize as “the gully.” It is typically V- or initiate from a debris slide or other slope failure from
U-shaped, with steep sidewalls forming the sides a headwall or sidewall. In one study of debris flows
of the V or U. The transport zone usually carries in coastal British Columbia, only 2% of debris flows
surface discharge along the channel at the base of initiated from high water discharges in the channel
the V or U. Sidewalls can range from a metre to tens (Millard 1999). A debris slide from headwalls is most
of metres in height. The Forest Practices Code (B.C. likely to initiate channelized debris flows because it
Ministry of Forests 2001) defined gullies as those maintains a directional momentum on entering a
with sidewalls greater than 3 m in height. Similarly, gully channel to which it is already closely aligned.
sidewall gradients are generally steeper than 50% In contrast, most debris slides from gully sidewalls
(27°) but in some cases are less steep. The size of the need to turn almost 90° to continue down the gully
channel in the gully will depend on the amount of channel. This loss of momentum results in more
surface water draining into the gully as well as any debris slides depositing in the gully channel. In the
groundwater that emerges from the gully sidewalls Interior of British Columbia, in-channel initiation
or from the base of the gully. Defining drainage area of debris flows as a result of high-discharge events
for gullies can be difficult, since groundwater move- appears to be more common.
ment on adjacent open slopes may or may not deliver Since confined debris flows travel much farther
subsurface flows towards the gully. than unconfined debris flows, most gully debris

247
flows reach the fan or valley bottom. Debris floods face erosion. Classic examples are strings of collapse
and flows from gullies can result in significant features forming beaded gullies common in silt
aggradation of streams and fans. fans derived from glaciolacustrine deposits in the
High water discharge rates in gullies can typically Thompson-Okanagan area (Evans and Buchanan
result in water floods or debris floods. Water floods 1976).
are the typical type of flood, transporting coarse These features, combined with an accompanying
sediment as bedload and with woody debris floating cave or pipe, are generally observed in arid climates
on the surface of the flow. Large discharge events in landforms composed of silt and clay. The term
supplied with large volumes of sediment can achieve “piping” is also used to describe soil pipes and (or)
sediment concentrations sufficient to support smaller water movement in forest soils, particularly along
clasts in a slurry, resulting in a debris flood. Debris root channels and small erosion conduits in surficial
floods may be four times larger than water floods materials. In some cases, piping may contribute to
expected in the same channel and are more destruc- landslides and washouts.
tive. Washouts (rapid erosion of seepage faces) have
been observed in glaciofluvial sands, silts, and
Piping gravels throughout British Columbia. Some ex-
amples include: the grand campus washout of 1935
Piping is subsurface soil erosion (Parker and Higgins at the University of British Columbia, Vancouver
1990). Pipes may be exposed as cavernous openings (Williams 1966); the Maryhill gravel pit washout,
in eroding banks (Figure 8.41). Terrain undergoing Coquitlam River valley (Allen 1957); caving erosion
subsurface erosion often exhibits hollows or col- in the Lower Mainland (Hungr and Smith 1985);
lapsed depressions aligned along routes of subsur- gully erosion / caving erosion found in the Coquit-
lam River valley (Siebert 1987); and seepage-face
erosion (Parker and Higgins 1990). Catastrophic
seepage-face erosion was also observed at Bowser
River in 1995, a result of the diversion of a small
stream by beaver onto a glaciofluvial terrace of sands
and silts—no surface water flow reached the cata-
strophically formed gully. The most spectacular is
the Donna Creek wash-out flow of 1992 described by
Schwab (1997, 2000, 2001).
A schematic representation of catastrophic seep-
age-face erosion is presented in Figure 8.42. Under
normal conditions, seepage exits the slope without
causing erosion as the ground slope of the seepage
face has adjusted to the highest recurrent seepage
exit gradient2 and resulting seepage forces. A change
in the discharge conditions results in an increase
in seepage exit gradient above a critical value. This
increase in exit gradient may result from a rapid in-
crease in groundwater discharge or sudden removal
of material from the seepage face by a landslide or
by running water. If the resultant erosion is of suf-
ficient intensity, the retreating seepage face uncovers
material with greater groundwater pressure, the exit
gradient increases, and a continuing chain reaction
results (O. Hungr, University of British Columbia,
FIGURE 8.41 Pipe exposed on a bank of the Chilcotin River. Earth and Ocean Science, Vancouver, B.C., pers.
(Photo: M. Geertsema) comm., 1997).

2 Seepage exit gradient is the slope of the piezometric surface adjacent to the drainage discharge point.

248
Stable seepage discharge Surface flow

Outwash sand/gravel

Seepage
Sand/gravel lenses
Groundwater flow

Till/bedrock
Lacustrine silt/clay

Increase in seepage exit gradient

– increase in seepage discharge
– increase due to retreat of seepage face
from landslide or surface erosion

Debris slide exposes Perched water table

saturated layers

Cavities/layers filled
with water

Material collapses, carried away

by debris flow processes

Debris flow

Process stabilizes
– excess water inflow stops
(seepage exit gradient returns to normal)
– or all available material is removed

Eroded material

FIGURE 8.42 Development of catastrophic seepage-face erosion concept. (Adapted from Siebert 1987)

249
 Knowledge of the processes occurring during the mineral soil from raindrop impacts. Loss of
such a catastrophic event enables the event to be clas- the forest floor can therefore increase the likeli-
sified. The outcome is the formation of a large gully, hood of overland flow and splash erosion of the
eroded largely by running water, hence a “washout.” mineral soil.
Landslide processes are also involved: the collapse, • The structure of the surface soil can be altered by
dilation of material, and extremely rapid debris flow fire, and by ash, which can clog soil pores or form
of liquefied material, hence a “flow” or “debris flow.” a crust. This can further reduce the permeability
The term “washout-flow” thus describes the overall of the surface soil.
complex landslide process. Catastrophic seepage-face • Death of trees results in increased snow accu-
erosion describes the sudden change in the discharge mulation, increased snowmelt rates, and reduced
conditions that result in an increase in seepage exit evapotranspiration. This can cause increased
gradient, the subsequent convergence of surface and runoff from both snowmelt and long-duration
subsurface flows, the retreating seepage face, caving, rainfall, and higher groundwater levels, which
collapse, and debris flow surges. can reduce slope stability. This change may be
long-term until the forest regenerates, and may be
Fire-generated Erosion and Landslides equated to similar effects of clearcutting or moun-
tain pine beetle infestation.
Effects of wildfire on slope stability
An increased incidence of landslides, erosion, and Debris flows are the most common mass move-
flooding is common following wildfire in many parts ment process observed following fires, although the
of the world (Shakesby and Doerr 2005). Most litera- processes described above can affect many types of
ture on the subject has focussed on surface erosion, slope instability. In areas of stable terrain, debris
but more recently the risks associated with large floods caused by a combination of overland flow
mass movement events, especially debris flows, have during rainstorms and an accumulation of eroded
been appreciated (Cannon and Gartner 2005). In the sediment in stream channels are common.
western United States, considerable resources are ex- In northern regions underlain by permafrost,
pended on assessing the risk of post-wildfire erosion shallow debris slides caused by detachment of the ac-
and mass movement, and on applying treatments to tive layer are common following wildfires (Lipovsky
burned areas to reduce these risks (Robichaud et al. et al. 2005). The main contributing factor for these
2000). failures is probably loss of the insulating organic sur-
Wildfire typically causes several changes in soil face layer, as well as loss of shade and the decreased
and vegetation, which can affect hydrology and slope albedo of the burned area.
stability (Scott and Pike 2003; Shakesby and Doerr In 2003, many large wildfires occurred in the
2005; Curran et al. 2006). These effects include: southern Interior of British Columbia, including
several in population interface areas. In the follow-
• A water-repellent layer can be formed just below ing 2 years, flooding and mass movement events
the surface of the mineral soil, caused by the occurred in several of the burns, and included the
deposition of hydrophobic compounds that are following notable examples (Jordan et al. 2004; Jor-
volatilized by burning of the forest floor. Water dan and Covert 2009).
repellency can cause the generation of overland
flow across large areas, especially during high- • Okanagan Mountain Park fire near Kelowna: In
intensity rainfall under dry soil conditions. The October 2003, debris floods and flooding oc-
effect typically persists for 3–5 years after the fire. curred in several creeks, caused by a short-dura-
Many post-wildfire debris flows have been caused tion rainstorm with a return period estimated
by runoff and eroded sediment generated from at 10–100 years. Municipal roads, culverts, and
areas of water-repellent soils during intense rain- private property were affected.
fall, which accumulate in gullies or steep stream • Cedar Hills fire near Falkland: In June 2004, sev-
channels in or below the burn. eral small debris flows and flooding occurred in
• Combustion of the forest floor and understorey gullies below the burn, caused by a short-duration
vegetation results in reduced water storage and in- rainstorm (Figure 8.43). A highway and private
terception capacity. Also, the forest floor protects property were affected.

250
 a b

FIGURE 8.43 Evidence of flooding and mass movement events (June 2004) following the 2003 Cedar Hills fire, near Falkland:
(a) overland flow, (b) concentration of flow and eroded sediment in gully, and (c) debris flow. (Photos: B. Grainger)

• Kuskonook fire near Creston: In August 2004, The number and severity of events following the
debris flows destroyed two houses and blocked 2003 fires was unprecedented in British Columbia.
a highway for several days (see below). A possible reason is that the summer of 2003 was
• Lamb Creek fire near Moyie Lake: In August exceptionally hot and dry, which may have resulted
2004, a large debris flood damaged a forest road in unusually high soil burn severity due to a dry
and affected fish habitat. forest floor. An additional factor may be the unusual
• Ingersoll fire near Burton: In October 2005, number of short, high-intensity rainstorms in fall
during a 3-day, low-intensity rainstorm, about 15 2003 and summer 2004.
debris flows and debris slides occurred in gullies Until the 2003 fire season, landslides, debris flows,
in and below the burn (Figure 8.44). Several of and other major erosion events following fires were
these were very large and damaged forest roads rarely reported in British Columbia. Some explana-
and forested private property below. tions may include:

Smaller flooding or debris flow events that af- • Rainfall intensities are lower than in more south-
fected highways or private property occurred in ern latitudes.
several other 2003 burns, including the McClure, • Soil burn severity may be lower at most British
Vaseaux Lake, McGillivray, and Strawberry Hill Columbia locations compared with fires further
fires. A debris flow also occurred in the 2004 Botanie south, because of soil and forest floor differences
fire near Lytton. and a relatively cooler climate.

251
• The hydrology in the British Columbia interior is and the 1998 Silver Fire, both near Salmon Arm. No
dominated by snowmelt, and even with a substan- systematic documentation of mass movement events,
tial increase in peak flow on small streams from rainfall events, or burn severity has taken place. It
summer rainstorms, these flows still may not appears likely that, although localized erosion may
exceed typical spring snowmelt peaks. be common, most large wildfires in British Columbia
• Most wildfires in British Columbia are in relative- do not produce significant mass movement events,
ly remote areas, and landslide or erosion events erosion, or flooding. Nevertheless, the probability
may have occurred but not been observed. of such events is increased over pre-fire conditions,
especially where soil burn severity is high. The oc-
Despite the well-documented events that followed currence of major erosion events may depend on
several of the 2003 fires, most of the burnt-over areas whether unusually intense rainstorms fall in the
had no significant erosion or mass movement events. critical 3–5 years after a fire, when water repellent
Several areas of large wildfires, which occurred conditions may still persist (see “Hydrophobicity,”
earlier near populated areas (e.g., near Kimberley below).
and Canal Flats in 1983), have been observed for
many years by Forest Service staff and others, and Case study: Kuskonook Creek debris flow (2004)
no major events have been noted. However, debris near Creston  The Kuskonook fire, east of the south
flow events were reported anecdotally following arm of Kootenay Lake, burned 4800 ha of forest in
several earlier fires, including the 1973 Eden Fire the late summer of 2003 (Figure 8.45). This was one
of several large wildfires in the southern Interior of
British Columbia in the severe fire season of 2003.
In the following 2 years, unusual landslide or ero-
sion events occurred in or below at least six of these
burns.
Around midnight on August 6, 2004, debris flows
occurred in both Kuskonook and Jansen Creeks dur-
ing a short, intense rainstorm. Both of these blocked
Highway 3A, which was closed for several days.
The larger flow caused severe damage in the small
community of Kuskonook, destroying two houses
and damaging several other buildings. Fortunately,
no fatalities or injuries were inflicted, as the houses
were unoccupied that evening.
The investigation of the Kuskonook event showed
that the debris flow was caused by water-repellent
soils in the burned area, which generated an unusual
volume of overland flow in the headwaters of the
drainage (Jordan et al. 2004; VanDine et al. 2005;
Curran et al. 2006) (Figure 8.46). This runoff was
concentrated in steep gullies below (Figure 8.47),
and became a debris flow as the exceptionally high
discharge entrained sediment from the channel bed
and sidewalls. An estimated 20 000–30 000 m3 of
debris was deposited on the Kuskonook Creek fan
(Figure 8.48).
This event is a good example of a debris flow
caused by changes in soil hydrology following a
severe wildfire. Although commonly reported in the
United States (Kalendovsky and Cannon 1997), such
FIGURE 8.44 Evidence of debris flows (October 2005) events were believed to be unusual in Canada, until
following the 2003 Mt. Ingersoll fire, near the events following the 2003 fires.
Burton. (Photo: P. Jordan)

252
FIGURE 8.45  View over Kuskonook and Jansen Creek watersheds, burned in 2003 fire. (Photo: D. Nichol)

Hydrophobicity  Soils that repel water are hydropho- ate hydrophobic layers in soil. Thick acidic humus,
bic. Normally, soils have a high absorptive capacity coarse-textured soils, high-intensity fires, and pro-
for water such that water applied to coarse-textured longed periods of intense heat are important factors
soils will readily infiltrate the soil (DeBano et al. for the creation of water-repellent soil. Hydropho-
1967); however, some soils repel water. These soils are bic substances are released from humus forms and
referred to as “non-wettable,” “water repellent,” or condense at depth. Extremely high surface tempera-
“hydrophobic.” In wettable soil, a strong attraction tures may actually destroy surface hydrophobicity,
exists between dry soil particles and water. In hydro- but cause a subsurface water-repellent layer to form
phobic soil, mineral particles are usually coated with deeper in the soil (perhaps 1–10 cm in depth).
non-polar organic materials that repel water. In wa- With a hydrophobic layer at depth restricting
ter-repellent soils, drops of water form surface beads. infiltration, surface soil may become saturated fol-
Measuring the slope of the wetting angle between a lowing precipitation, leading to overland flow, soil
droplet of water and the ground surface indicates the detachment, and erosion. The eroded material may
degree of hydrophobicity (Figure 8.49). coalesce into shallow landslides, or fill gullies and
In British Columbia, forest fires sometimes cre- transform into debris flows.

253
FIGURE 8.46  Soil erosion caused by overland flow on burned soils. (Photo: M. Curran)

FIGURE 8.47  Gully erosion in headwaters of Kuskonook Creek. (Photo: P. Jordan)

254
FIGURE 8.48  Debris flow deposit on Kuskonook Creek fan. (Photo: P. Jordan)

Wetting angle

Hydrophobic material

Wetting angle

Hydrophilic material

FIGURE 8.49 The wetting angle between a droplet of water

and a solid surface is an indication of the degree
of hydrophobicity: the steeper the angle, the
more hydrophobic the soil. (Modified from
DeBano et al. 1967.

255
Reading and Interpreting the Landscape

Indicators of Instability

Field terrain stability analysis requires an eye for Tension fractures, spoon-shaped depressions,
detail to read subtle changes in the landscape and on ridges, scarps, antislope scarps, seepage, and ponds
aerial photographs. It not only requires an under- on hillslopes may all be indicative of unstable ter-
standing of geomorphic processes, but also of local rain. Lower down the slope, one might find ridges,
ecosystems. hummocks, lobes, or rubble. Rotational slides may
daylight (emerge at the surface) in rivers or streams,
Aerial Photograph Interpretation pushing up the river bed (Figure 8.50).
Water-loving vegetation on slopes indicates seep-
Field investigations are usually preceded by aerial age and is often associated with unstable terrain.
photograph interpretation. Sometimes unstable ter- The linear tracks of vegetation seen on aerial photo-
rain is easy to recognize. Fresh, unvegetated scars graphs may also be recognizable on the ground. Old-
often indicate recent landslide activity. Unstable er trees adjacent to the younger vegetation may bear
terrain and old landslides that are overgrown with scars or have persistent mud lines on their trunks. In
vegetation are more difficult to discern. Unvegetated addition to scars, trees may display pronounced in-
and vegetated linear tracks are usually indicative clination or curvature. Trees tend to lean upslope at
of debris slides or flows. Debris flows usually have the head of landslides, and downslope on lobes, but
moderately steep fans, but the debris may also enter this can be highly variable. In spreads and rotational
directly into powerful streams that remove the mate- landslides, trees tend to lean in the direction of rota-
rial, preventing fans from developing. tion of the ridges and blocks formed beneath them.
Flows tend to have ridges aligned in the direction In some cases, trees may be split when they strad-
of movement, whereas spreads and rotational slides dle a tension crack. Although trees typically display
(slumps) have ridges transverse to movement. Some- swelling at their bases, debris on fans may bury the
times spreads and slumps transform into flows as the lower parts of trees. Absence of buttswell should
slope over which they move becomes steeper, or as alert an observer to a burial event (Figure 8.51). Logs
material becomes progressively weaker. entrained in debris flows tend to be severely abraded
Deep-seated landslides tend to have arcuate main (Figure 8.52).
scarps, often with wet depressions, referred to as “sag Deposits formed by landslides may display a wide
ponds.” Over time, the ponds may drain or become range of textural and morphological characteristics
filled with peat. The negative surface expression of depending on the type of landslide. Deposits range
zones of depletion is to be contrasted with the posi- from coarse angular blocks resulting from rock
tive surface expression of zones of accumulation. avalanches to fine-textured mud flow deposits. In
Lobate and hummocky surfaces are also characteris- general, landslide deposits display perturbation,
tic of landslide deposits. but some translational landslides transport mate-
rial with little to no internal disturbance. Typical
Field Interpretation colluvial soils are loose, with angular clasts. Debris
flow clasts are often matrix-supported. Buried soils
Many diagnostic features of past and potential land- indicate repeated deposition events.
slides can be noted in the field. Indicators of instabil-
ity are manifested in surface expression, vegetation,
rock, and soil.

256
FIGURE 8.50 Upturned rotational landslide toe exposed (arrow) in Kiskatinaw River, near Dawson Creek.; this
landslide is sensitive to toe erosion by the river. (Photo: M. Geertsema)

257
FIGURE 8.51 Balsam poplars buried in debris-flow deposits
near Chetwynd; note the absence of buttswell.
(Photo: M. Geertsema)

FIGURE 8.52 Severely abraded log entrained in the 2002 Zymoetz rock slide–debris flow, near
Terrace; the log is oriented in the direction of flow. (Photo: M. Geertsema)

258
Dating Hillslope Processes

An important component of hazard and risk analy- have occurred more frequently under certain past
sis is the determination of recurrence intervals and climates (e.g., Geertsema and Schwab 1997) or be
magnitude or frequency relationships. Both the spa- clustered around strong earthquake events. Knowl-
tial and temporal distributions of hazardous events edge from the past can be extrapolated to help with
are required for such analyses. This section discusses predictions for the future.
the determination of the distribution of hazards in Even though landslides may be dated in numer-
time. ous ways, the data are often scarce and difficult to
Landslides occur every year in British Columbia, obtain. One must choose an appropriate sampling
and have probably been happening since deglacia- location and interpret the data correctly. In Brit-
tion. They leave their marks as records on the land- ish Columbia, the most common landslide-dating
scape—as visible scars and lobes, and through burial methods are dendrochronology (tree ring dating)
of soil and plant materials. Over time, the visible and radiocarbon dating. Recent events, or events
records diminish in clarity and buried records may that recur on decadal time scales, require high-reso-
be eroded. Nonetheless, numerous techniques are lution dating techniques. Tree rings and varves3 can
available to date landslides. provide excellent annual resolution. Radiocarbon
dating is best left for century-scale resolution.
Human Records
Tree ring dating
The most reliable sources of information are well- In forested landscapes, trees often provide the
documented, eye-witness accounts. Newspaper best records of past slope processes (Shroder 1980;
accounts may provide an accurate date of some Braam et al. 1987a, 1987b). Several properties of trees
event, but spatial descriptions are not always ac- make them particularly useful for landslide dat-
curate. Aerial photographs and satellite images help ing: (1) they produce annual growth rings; (2) their
constrain the times of events, provided the events rings are sensitive to environmental conditions and
are sufficiently large to be recognizable at the image traumatic stresses; and (3) trees usually attempt to
scale. Old maps, reports, Hudson’s Bay Company grow vertically—if something causes them to lean,
records, and even landscape paintings (e.g., Nilson they attempt to correct the lean by resuming vertical
2005) have been recorders of landslides. Some large growth. Leaning trees will produce dark eccentric
rock slides generate their own seismic signatures, rings called “reaction wood.”
and landslides that dam rivers may be recorded Landslides affect forests in different ways. Shal-
indirectly by hydrometric stations as dips and spikes low debris slides and avalanches tend to remove trees
on hydrographs, corresponding to the formation and from their paths (nudation) and accumulate woody
rupture of the landslide dams. Some catastrophic debris along their lower margins. Standing trees be-
landslides are recorded in the oral traditions of First yond the limit of the slides are often pushed off their
Nations people; however, the dates of the events vertical axes. Slow-moving landslides may result in
are uncertain and are usually tested with radio- the production of curved tree trunks. Debris flows
carbon dating methods. For example, stories of a and rock falls tend to scar trees in their paths. A sin-
large rock slide that buried a village near Hazelton gle tree may bear scars from multiple events. Rota-
some 3000 years ago (Gottesfeld et al. 1991) and of a tional slides, spreads, and low-gradient translational
rockslide–generated tsunami that wiped out a village slides tilt trees. In some cases, blocks of ground are
in Knight Inlet about 500 years ago (Bornhold et al. rafted horizontally, or dropped vertically without
2007) are told to this day. tilting trees. Flows may bury standing trees in zones
of accumulation. Exposed surfaces on soil landslides
Natural Records in forested areas tend to be rapidly recruited by trees
and shrubs; however, large rock slide rubble may
It is important for researchers to establish the oc- remain unvegetated for many decades. A fan-delta
currence of landslides over time. Landslides may collapse west of Houston caused trees weighed down

3 Annual lake sediments usually separated by winter clay and summer silt layers.

259
by sediment in root wads to float vertically in a Dating landslides from new colonizing trees is not
lake for some weeks before sinking. Landslides that as simple as it might seem, as many potential sources
impound streams may drown trees upstream of the of error exist. Trees may not appear within the first
landslide dam. Over time, sediments accumulating year following the landslide and may colonize at
in the temporary lakes may bury the trees. varying times. It is not always sufficient to date the
Master chronologies can extend tree ring records oldest trees either, as many translational landslides
back in time over a considerably longer span than transport trees in the vertical position (Figure 8.53).
living trees. The technique involves the statistical These trees simply continue to grow. Even if large
matching of ring sequences from living and fossil trees are tilted during landslide transport, seedlings
trees. Such cross-dating is far superior to radiocar- and small saplings may be transported without tilt-
bon dating of fossil trees, but established chron- ing, and these may be difficult to distinguish from
ologies are region-specific and time consuming post–landslide-colonizing plants. Another source of
to obtain. Brian Luckman (University of Western error lies in the difficulty of finding the first growth
Ontario) has developed a master chronology for the ring. Minor subsequent movements may result in
Rocky Mountains, and Dan Smith (University of staggered colonization.
Victoria) has developed one for the Coast Moun- If the colonizing vegetation is significantly differ-
tains. ent from that of the surrounding forests (e.g., Salix
in a Picea matrix), one can be more confident in
Nudation the trees providing a reasonable minimum date of
Landslides tend to remove vegetation from slopes, the landslide. The technique is best used on debris
resetting the plant successional clock. New recruit- slides, and rock and debris avalanches where little of
ment of trees and shrubs on landslides facilitates the original surface is preserved. In landslides that
landslide dating by providing a minimum age for an carry rafts of surface material, large sample sizes are
event’s deposit. Some species are rapid invaders, such recommended. More precise and accurate dates can
as Alnus, Betula, Salix, and Picea, and these are use- be obtained from the analysis of reaction wood in
ful for obtaining minimum dates of a landslide. leaning trees and from tree scars.

FIGURE 8.53 Rafted red alder in the Khyex landslide near Prince Rupert. Ring counts of these trees would
over­estimate the age of the landslide. (Photo: M. Geertsema)

260
Inclination landslide movement (Figure 8.55). Trees often tilt in
The inclination (or leaning) of trees produces reac- random directions in hummocky, permafrost ter-
tion wood, which in cross-section shows dark eccen- rain. They may tilt towards openings in the forest,
tric rings rather than normal light concentric rings exhibiting phototrophism, or growing towards the
(Figure 8.54d). Reaction wood can be used to directly light. Trees may lean due to wind, particularly in
date landslide events. In sudden movements, the shallow soils. Trees may be curved due to snowpack
change can be abrupt, often with the tree curving up creep.
to grow vertically again. Sometimes, a lateral branch
takes over vertical dominance (Figure 8.54b). Corrasion (Scarring)
Gradual or episodic movement can also be re- Corrasion of tree bark can result from the impact of
corded in trees as can a change in the direction of rocks, debris, or other trees (Figure 8.56). Scars yield
tilting. Trees tend to be tilted upslope on rotational direct dates of landslide events. A single tree can
slide blocks and downslope on lobes. Trees can be record multiple events by displaying multiple scars
tilted in multiple directions (“jackstrawed”) in hum- of varying ages. However, scars can also result from
mocky deposits or in areas with many small failures numerous non-landslide events, including the fall
within a larger landslide body. of neighbouring trees, animal gnawing, and frost
Trees can lean for many reasons other than from damage.

a b

c d

FIGURE 8.54 Trees displaying variable responses to mass movement: (a), (c), and (d) show slow,
episodic mass movement; (b) shows sudden, rapid movement that topples a tree and
causes a lateral branch to assume apical dominance. (d) Note the various times and
directions of reaction wood (arrows), which indicates leaning in different directions.
(Photos: M. Geertsema)

261
 Burial
a Flows in the lower, distal portions of landslides
may partially bury trees without causing mortality.
Evidence of burial may be lack of buttswell, and the
sprouting of adventitious roots in the newly formed
ground surface (Wilford et al. 2005a). Trees may also
respond to the stress of burial by undergoing growth
suppression. Response is often not immediate, and
thus less reliable than scars and reaction wood for
landslide dating.

Drowning
Many landslides form lakes when streams are
dammed, or when closed depressions (e.g., sag
ponds) are formed on the landslide surface. Trees
b inundated by standing water may die as a result
of drowning (Figure 8.57). Careful cross-dating of
drowned tree ring sequences to those of pre-slide
living trees can yield a landslide date; however, trees
may not die in the first season of inundation, and

FIGURE 8.55 Trees can lean for reasons other than slope

instability: (a) trembling aspen near Dawson
Creek display growth towards the light; (b) black
spruce leaning due to ice dynamics in hummocky
permafrost terrain west of Fort Nelson. (Photos:
M. Geertsema)

a b

FIGURE 8.56 Tree scars: (a) scarred western hemlock on debris flow levee; (b) four separate debris
flow scars shown on one tree cookie. (Photos: (a) M. Geertsema; (b) J.W. Schwab)

262
 a

FIGURE 8.57 Landslides often dam rivers, inundating their floodplains: (a) in 2007 a rock slide–
debris avalanche in northwestern British Columbia impounded the Todagin River,
drowning spruce trees; (b) spruce trees drowned by a landslide that occurred in the
mid-1600s at Halden Creek in northeastern British Columbia (Geertsema and Clague
2006). The lake filled with sediment, preserving the trees. After the dam broke, and
the lake drained, bank erosion exhumed the tree. (Photos: [a] M. Geertsema; [b]. J.
Clague)

263
large sample sizes are needed. In cases where the Conventional radiocarbon dating involves the
drowned trees do not overlap in age with living trees, counting of beta-rays given off as a decay product.
radiocarbon dating of organic layers in lake sedi- The newer Accelerator Mass Spectrometry (AMS)
ments behind the landslide dam can be used. dating method actually counts atoms, and requires
much smaller sample sizes.
Varve chronologies Radiocarbon dates can be improved by statisti-
Varves, or annual rhythmic sedimentary couplets cal methods such as wiggle matching (Figure 8.58).
that occur in certain types of lakes, typically include Using a technique described by Bronk Ramsey et al.
a silty summer layer and a clayey winter layer. Clay- (2001), Geertsema and Clague (2006) radiocarbon
size particles take longer to settle out of suspension dated both the pith and outer rings of trees buried
than larger grain sizes, and are thus the only sedi- by landslide debris near Fort Nelson. Counting the
ment left to settle out of lake water in winter. Epi- rings between the inner and outer sampled wood al-
sodes of increased sedimentation point to periods of lowed for wiggle matching of the radiocarbon dates
increased instability in watersheds (e.g., Menounos to improve the statistical precision (Figure 8.58).
et al. 2006). Sometimes, the distal portions of large Buried wood, charcoal, and basal peat are com-
debris flows enter lakes. Varve counting above and monly radiocarbon dated to determine landslide
below such event beds, often in combination with ages. Where datable material overlies a landslide, a
other marker beds such as tephras, can provide minimum age is given, whereas a landslide overlying
precise and accurate dates of the events, but poten- a datable organic layer yields a maximum age for the
tial sources of error exist. For example, varves can be event.
difficult to count because of ambiguities in the sedi-
ment record. The spatial distribution of the sources Lichenometry
of the sediment and the landslide deposits usually
remain unknown. The data collection is also ex- Lichenometry is the dating of lichens by measuring
pensive and time consuming. The method has been their nominal diameter (Figure 8.59). This tech-
applied to the dating of large submarine mud flows nique has been applied to the dating of landslides
in Saanich Inlet (Blais-Stevens and Clague 2001). that expose rock faces and produce boulders. It has
been especially useful in Europe, where calibration
Radiocarbon Dating of Rhizocarpon spp. growth on gravestones is pos-
sible. Although tried in British Columbia, the lack of
Radiocarbon dating is one of the most common growth curves and rigorous statistical sampling has
and useful methods of dating prehistoric landslide thus far precluded its widespread use. The combina-
deposits. It is useful for century-scale resolution, tion of lichenometry with dendrochronology, AMS
but can also be used for decadal timescales by using radiocarbon dating, and surface exposure dating
statistical techniques on more than one date. could help establish growth curves for localized
Living plants take up CO2 from the atmosphere. areas of the province.
Animals also take up this carbon because they either
eat plants or eat other animals that have eaten plants. Simple Comparative and Observational Dating
Two stable isotopes of carbon occur: 12C and 13C. The
approximate relative ratios of 12C and 13C are 99% and In many cases, the age of landslides exceeds the age
1%, respectively. The unstable 14C isotope undergoes of the trees on their surfaces. Datable elements may
radioactive decay and has a half-life of 5730 years. exist, but in routine terrain stability field assess-
14
C occurs in trace amounts (1.2 x 10–10%) that vary ments and mapping it is impractical and prohibi-
slightly over time. Because the atmospheric produc- tively expensive to use expensive dating techniques.
tion of 14C is not uniform, tree rings have been used Fortunately, landscape indicators provide clues about
to produce absolute calibration of radiocarbon data. the age of landslides. These include soil develop-
The calibration curve plots radiocarbon age against ment, degree of acidification, thickness of peat in
calendar age. depressions, rounding and degradation of ridges and
When an organism dies, it ceases to take up CO2, scarps, rock colour, and lichen cover. A few examples
and thus sets the radiocarbon clock. By measuring are given below.
the declining proportion of 14C relative to total C, When a landslide removes surface soil, it sets back
the approximate age of the item can be determined. the pedogenic clock (Geertsema and Pojar 2007).

264
 a Soil-forming processes begin anew in the unweath-
ered soil or rock. For example, a Podzol (a well-de-
veloped soil; Soil Classification Working Group 1998)
may be removed, restarting the cycle from Regosol
to Brunisol to Podzol. Although soil development
rates vary with climate and texture, a landslide
with a Brunisol should be younger than that of a
neighbouring landslide with a Podzol. It needs to be
stressed that pre-slide soil horizons can be preserved
in rafted blocks and, therefore, a careful evaluation
of the context is essential.
Surface soils in forested areas are often more
acidic than deeper soils. Testing pH, or simply test-
ing for effervescence with hydrochloric acid, can dis-
tinguish between younger and older surfaces in areas
where carbonate-rich parent material is evident.
For example, exposed depth material at the Mink
Creek landslide near Terrace had a pH of 8 in sharp
contrast to acidic surface soil with a pH of less than
5. Over time, the high-pH soil would be expected to
become acidic.
Many landslides contain water-filled depressions,
both in zones of depletion (usually sag ponds) and
in zones of accumulation. Organic sediments often
accumulate in these ponds, eventually forming
b
HC12-1 kill date : 644 ± 46 peat. Although peat compacts over time, variable
thickness of peat in landslide depressions may give
some indication of relative age (e.g., Geertsema and
Schwab 1997).
Sharp edges of landslide scarps and ridges de-
Relative probability

0.8 grade, becoming more rounded with time (Keaton

0.6 and Degraff 1996). This is especially true for soil
0.4 landslides. In some landslides, ridges degrade rap-
0.2 idly, whereas in others ridges persist for millennia,
0.0 especially in the drier interior of the province. The
relative ages of fractures in bedrock can be distin-
guished, in part on the basis of physical weathering
1000 1100 1200 1300 1400 1500 of the rock. In general, older landslides will have a
Calendar date more established drainage network and more sub-
dued relief than fresh landslides.
FIGURE 8.58 Wiggle matching improves the precision of Rock slide and rock fall scars expose fresh, un-
radiocarbon dating: (a) the rings of this cot- weathered rock faces. These often have a lighter tone
tonwood log buried by landslide debris near Fort than the greyish, weathered (often lichen-covered)
Nelson were counted, and radiocarbon dates areas surrounding unweathered rock. On aerial pho-
obtained from the pith and outer wood; (b) the tographs, fresh rock slides have light tones. Without
dates were then wiggle matched with the ring resorting to full lichenometric analysis, a compari-
count, improving the precision of the calibrated son of lichen absence to extensive lichen cover can
radiocarbon dates from 110 and 130 year errors
indicate the relative ages of two surfaces.
to 25 and 45 year errors at 68 and 95% prob-
abilities, respectively. This suggests that the kill
date of the tree occurred sometime around 1388
AD ±12.5 years at the 1 sigma confidence level.
(Modified from Geertsema and Clague 2006)

265
 a

FIGURE 8.59 (a) Lichen-covered rubble in a bedrock spread southeast of Tumbler Ridge;

(b) close-up of Rhizocarpon species. In lichenometry, several measurements are made
from the centre to the outside of the thallus to provide an average measurement of
radial growth. (Photos: M. Geertsema)

266
Summary

British Columbia’s diverse physiography, geology, The landscape holds records and provides clues
soils, and climate zones provide a unique template about past and potential hillslope processes. Veg-
that hosts a wide variety of landslide types and ero- etation, slope geometry, and microtopography can
sion processes. Our mountainous topography results reveal important information. For example, wet-site
in steep slopes, great local relief, and high precipita- indicators on slopes, tilted trees, spoon-shaped de-
tion, all of which predispose much of the province to pressions, and hummocky lobes are all signs of slope
landslide activity. Because landslide types can have instability. This evidence can be seen on air photos,
drastically different consequences, it is important to and also in the field.
distinguish between them. For instance, a slow-mov- Determining when a landslide occurred is an
ing earth slide has different impacts than a rock ava- important component of establishing return periods
lanche, although both are classified as “landslides.” or landslide frequencies. Dating landslides is accom-
Preconditioning factors, such as weathering, glacial plished in many ways, including the examination of
erosion, river erosion, geologic structure, earth- air photos and written records. A common method
quakes, permafrost thaw, wildfire, deforestation, and of dating landslides is through dendrochronol-
road construction, can all lead to landsliding. The ogy. Trees may bear scars, or reaction wood, from
ultimate trigger may be heavy rainfall, snowmelt, multiple landslide events. Interpreting the tree rings
undercutting, an increase in surface loading, or an permits the establishment of landslide frequencies.
earthquake. Other dating methods used in British Columbia
Erosion is often a precursor of landsliding, and include lichenometry, varve counting, and radiocar-
plays an important role in landscape evolution. bon dating.
For example, rain splash, and sheet, rill, and chan- Land use changes, such as road construction, tim-
nel erosion contribute to the formation of gullies. ber harvesting, and urbanization, can significantly
Erosion can also occur beneath the ground surface increase the likelihood of landslides by changing
in pipes. Wildfires can cause surface erosion, partly the course of surface and subsurface drainage or by
through the creation of water-repellent soils, as well loading or undercutting slopes. Chapter 9 (“Forest
as through the loss of the protective forest floor. Ero- Management Effects on Hillslope Processes”) dis-
sion can lead to the “charging” of gullies and stream cusses some of the effects of land use, in particular
channels with sediment, which can ultimately lead forest management, on landslide hazard.
to damaging debris flows.

REFERENCES

Allen, J.F. 1957. Landslides, washouts and mud flows B.C. Ministry of Transportation and Highways. 1996.
in the Lower Fraser Valley, B.C. BASc thesis. Natural hazards in British Columbia. Geotech.
Univ. British Columbia, Vancouver, B.C. Materials Br., Victoria, B.C.
Alley, N.F. and B. Thomson. 1978. Aspects of envi- Blais-Stevens, A. and J.J. Clague. 2001. Paleoseismic
ronmental geology, parts of Graham Island, signature in late Holocene sediment cores from
Queen Charlotte Islands. B.C. Min. Environ., Saanich Inlet, British Columbia. Mar. Geol.
Resour. Anal. Br. Victoria, B.C. Bull. No. 2. 175:131–148.
B.C. Ministry of Forests. 2001. Gully assessment pro- Bornhold, B.D. and J.R. Harper. 1998. Engineering
cedure guidebook, 4th edition, version 4.1. For. geology of the coastal and nearshore Canadian
Pract. Br., Victoria, B.C. For. Pract. Code B.C. Cordillera. In: Engineering geology: a global
Guideb. www.for.gov.bc.ca/tasb/legsregs/fpc/ view from the Pacific Rim. Proc., 8th Int. Con-
fpcguide/GULLY/GAPGdbk-Web.pdf. gr., Int. Assoc. Eng. Geol. Environ., Vancouver,
B.C. D.P. Moore and O. Hungr (editors). A.A.
Balkema, Rotterdam, Netherlands, 1:63–75.

267
Bornhold, B.D., J.R. Harper, D. McLaren, and R.E. Chatwin, S.C., D.E. Howes, J.W. Schwab, and D.N.
Thomson. 2007. Destruction of the First Na- Swanston. 1994. A guide for management of
tions village of Kwalate by a rock avalanche- landslide-prone terrain in the Pacific North-
generated tsunami. Atmos. Ocean 45:123–128. west. 2nd ed. B.C. Min. For., Res. Br., Victoria,
B.C. Land Manag. Handb. No. 18. www.for.gov.
Bottino, G., M. Chiarle, A. Joly, and G. Mortara.
bc.ca/hfd/pubs/docs/Lmh/Lmh18.pdf (Ac-
2002. Modelling rock avalanches and their
cessed May 2010).
relation to permafrost degradation in glacial
environments. Permafrost Periglac. Process. Church, M. and M.J. Miles. 1987. Meteorological
13:283–288. antecedents to debris flow in southwestern
British Columbia: some case histories. In:
Bovis, M.J. 1982. Uphill facing antislope scarps in the
Debris flows/avalanches: process, recognition,
Coast Mountains southwest British Columbia.
and mitigation. J.E. Costa and G.F. Wieczoreck
Geol. Soc. Am. Bull. 93:804–812.
(editors). Geol. Soc. Am., Boulder, Colo. Rev.
_______. 1985. Earthflows in the Interior Plateau, Eng. Geol. 7:63–79.
southwest British Columbia. Can. Geotech. J.
Clague, J.J. (compiler). 1989. The Quaternary geology
22:313–334.
of the Canadian Cordillera. In: Quaternary
_______. 1990. Rockslope deformation at Affliction geology of Canada and Greenland. R.J. Fulton
Creek, southern Coast Mountains, British Co- (editor). Geol. Surv. Can., Ottawa, Ont. Geol-
lumbia, Canada. Can. J. Earth Sci. 27:243–254. ogy of Canada, Vol. 1. (Also Geological Society
of America. The Geology of North America,
Bovis, M.J. and S.G. Evans. 1995. Rock slope move-
Vol. K-1), pp. 17–96.
ments along the Mount Currie “fault scarp,”
southern Coast Mountains, British Columbia. Clague, J.J. and J.G. Souther. 1982. The Dusty Creek
Can. J. Earth Sci. 32:2015–2020. landslide on mount Cayley, British Columbia.
Can. J. Earth Sci. 19:524–539.
_______. 1996. Extensive deformations of rock slopes
in southern Coast Mountains, southwest Brit- Costa, J.E. 1984. Physical geomorphology of debris
ish Columbia, Canada. Eng. Geol. 44:163–182. flows. In: Developments and applications of
geomorphology. J.E. Costa and P.J. Fleisher
Bovis, M.J. and M. Jakob. 1999. The role of debris
(editors). Springer-Verlag, Berlin, Germany,
supply conditions in predicting debris flow ac-
pp. 268–317.
tivity. Earth Surf. Process. Land. 24:1039–1054.
Cruden, D.M. and D.J. Varnes. 1996. Landslide types
Braam, R.R., E.E.J. Weiss, and P.A. Burrough. 1987a.
and processes. In: Landslides investigation
Dendrogeomorphological analysis of mass
and mitigation. A.K. Turner and R.L. Schuster
movement: a technical note on the research
(editors). Natl. Res. Counc., Transport. Res.
method. Catena 14:585–589.
Brd., Washington, D.C. Spec. Rep. No. 247, pp.
_______. 1987b. Spatial and temporal analysis of 36–75.
mass movement using dendrochronology.
Curran, M.P., B. Chapman, G.D. Hope, and D. Scott.
Catena 14:573–584.
2006. Large scale erosion and flood after wild-
Bronk Ramsey, C., J. van der Plicht, and B. fires: understanding the soil conditions. B.C.
Weninger. 2001. ‘Wiggle matching’ radiocar- Min. For. Range, Res. Br., Victoria, B.C. Tech.
bon dates. Radiocarbon 43:381–389. Rep. No. 030. www.for.gov.bc.ca/hfd/pubs/
Docs/Tr/Tr030.htm (Accessed March 2010).
Cannon, S.H. 2001. Debris-flow generation from
recently burned watersheds. Environ. Eng. Davies, M.C.R., O. Hamza, and C. Harris. 2001. The
Geosci. 7:321–341. effect of rise in mean annual temperature on
the stability of rock slopes containing ice-filled
Cannon, S.H. and J.E. Gartner. 2005. Wildfire
discontinuities. Permafrost Periglac. Process.
related debris flow from a hazards perspective.
12:137–144.
In: Debris-flow hazards and related phenome-
na. M. Jakob and O. Hungr (editors). Springer- DeBano, L.F., J.F. Osborne, J.F. Krammes, and J.
Praxis, Berlin, Germany, pp. 321–344. Letey. 1967. Soil wettability and wetting agents:

268
 our current knowledge of the problem. U.S. and climate change: challenges and solutions.
Dep. Agric. For. Serv., Pac. S.W. For. Range R. McInnes, J. Jakeways, H. Fairbank, and E.
Exp. Stn., Berkeley, Calif. Res. Pap. PSW 43. Mathie (editors). Taylor and Francis, London,
U.K., pp. 9–16.
Dramis, F., M. Govi, M. Guglielmin, and G. Mor-
tara. 1995. Mountain permafrost and slope sta- Geertsema, M., O. Hungr, S.G. Evans, and J.W.
bility in the Italian Alps: the Val Pola landslide. Schwab. 2006c. A large rock slide–debris ava-
Permafrost Periglac. Process. 6:73–82. lanche at Pink Mountain, northeastern British
Columbia, Canada. Eng. Geol. 83:64–75.
Egginton, V.N. 2005. Historical climate variability
from the instrumental record in Northern Brit- Geertsema, M. and J.J. Pojar. 2007. Influence of
ish Columbia and its influence on slope stabil- landslides on biophysical diversity: a perspec-
ity. MSc thesis. Simon Fraser Univ., Burnaby, tive from British Columbia. Geomorphology
B.C. 89:55–69.
Evans, S.G. 1989. The 1946 Mount Colonel Foster Geertsema, M. and J.W. Schwab. 1997. Retrogres-
rock avalanche and associated displacement sive flowslides in the Terrace-Kitimat, British
wave, Vancouver Island, British Columbia. Columbia area: from early post-deglaciation
Can. Geotech. J. 26:447–452. to present and implications for future slides.
Proc., 11th Vancouver Geotech. Soc. Symp.,
Evans, S.G. and R.G. Buchanan. 1976. Geotechnical
Vancouver, B.C., pp. 115–133.
problems associated with silt deposits in the
South Thompson Valley, British Columbia. _______. 2006. Challenges with terrain stabil-
Dep. Highways, Geotech. Materials Br., Victo- ity mapping in northern British Columbia.
ria, B.C. Interim Rep. Streamline Watershed Manag. Bull. 10:18–26.
www.forrex.org/publications/streamline/
Evans, S.G. and J.J. Clague. 1994. Recent climatic
ISS34/Streamline_Vol10_No1_art4.pdf
change and catastrophic geomorphic processes
(Accessed May 2010).
in mountain environments. Geomorphology
10:107–128. Geertsema, M. and J.K. Torrance. 2005. Quick clay
from the Mink Creek landslide near Terrace,
Fannin, R.J. and T.P. Rollerson. 1993. Debris flows:
British Columbia: geotechnical properties,
some physical characteristics and behaviour.
mineralogy, and geochemistry. Can. Geotech.
Can. Geotech. J. 30:71–81.
J. 42:907–918.
Geertsema, M. and J.J. Clague. 2006. 1000-year re-
Gottesfeld, A.S., R.W. Mathewes, and L.M. Johnson-
cord of landslide dams at Halden Creek, north-
Gottesfeld. 1991. Holocene debris flows and
eastern British Columbia. Landslides 3:217–227.
environmental history, Hazelton area, British
Geertsema, M., J.J. Clague, J.W. Schwab, and S.G. Ev- Columbia. Can. J. Earth Sci. 8:1583–1593.
ans. 2006a. An overview of recent large land-
Gruber, S. and W. Haeberli. 2007. Permafrost in
slides in northern British Columbia, Canada.
steep bedrock slopes and its temperature-re-
Eng. Geol. 83:120–143.
lated destabilization following climate change.
Geertsema, M., D.M. Cruden, and J.W. Schwab. 2006b. J. Geophys. Res. 112, F02S18, DOI:10.1029/
A large, rapid landslide in sensitive glacioma- 2006JF000547.
rine sediments at Mink Creek, northwestern
Guthrie, R.H. 2002. The effects of logging on fre-
British Columbia, Canada. Eng. Geol. 83:36–63.
quency and distribution of landslides in three
Geertsema, M. and D.M. Cruden. 2008. Travels in watersheds on Vancouver Island, British Co-
the Canadian Cordillera. In: 4th Can. Conf. lumbia. Geomorphology 43(3–4):273–292.
Geohazards. May 20–24, 2008. Quebec City,
Harris, C., M.C.R. Davies, and B. Etzelmüller. 2001.
Que.
The assessment of potential geotechnical haz-
Geertsema, M., V.N. Egginton, J.W. Schwab, and J.J. ards associated with mountain permafrost in a
Clague. 2007. Landslides and historic climate warming global climate. Permafrost Periglac.
in northern British Columbia. In: Landslides Process. 12:145–156.

269
Hodgson, E.A. 1946. British Columbia earthquake. J. initiation and forest operation shutdowns on
Roy. Astron. Soc. Can. 40:285–319. the north coast of British Columbia. Landslides
3:228–238.
Hogan, D.L. and J.W. Schwab. 1991. Meteorological
conditions associated with hillslope failures on Jakob, M. and O. Hungr (editors). 2005. Debris flow
the Queen Charlotte Islands. B.C. Min. For., hazards and related phenomena. Springer-
Victoria, B.C. Land Manag. Rep. No. 73. www. Praxis, Berlin, Germany.
for.gov.bc.ca/hfd/pubs/Docs/Mr/Lmr/Lmr073.
Jakob, M. and P. Jordan. 2001. Design flood estimates
pdf (Accessed May 2010).
in mountain streams—the need for a geomor-
Holm, K., M.J. Bovis, and M. Jakob. 2004. The land- phic approach. Can. J. Civil Engin. 28: 425–439.
slide response of alpine basins to post-Little Jakob, M. and H. Weatherly. 2003. A hydrom-
Ice Age glacial thinning and retreat in south- eterological threshold for landslide initiation
western British Columbia. Geomorphology on the north shore mountains of Vancouver
57:201–216. British Columbia. Geomorphology 54:137–156.
Hungr, O., S.G. Evans, M.J. Bovis, and J.N. Hutchin- Jakob, M. and H. Weatherly. 2003. A hydroclimatic
son. 2001. Review of the classification of land- threshold for landslide initiation on the North
slides of the flow type. Environ. Eng. Geosci. Shore Mountains of Vancouver, British Colum-
7:221–238. bia. Geomorphology 54:137–156.

Hungr, O., S. McDougall, and M. Bovis. 2005. En- Jordan, P. and S.A. Covert. 2009. Debris flows and
trainment of material by debris flows. In: De- floods following the 2003 wildfires in south-
bris flow hazards and related phenomena. M. ern British Columbia. Environ. Eng. Geosci.
Jakob and O. Hungr (editors). Springer-Praxis, 15:217–234.
Berlin, Germany, pp. 135–158. Jordan, P., M. Curran, and D. Nicol. 2004. Debris
Hungr, O. and D. Smith. 1985. Landslides and flows caused by water repellent soils in recent
development in the Lower Mainland, British burns in the Kootenays. Div. Eng. Geosci. For.
Columbia. B.C. Prof. Eng. 36:11–14. Sector, Assoc. Prof. Eng. Geosci. B.C. Aspect
9(3):4–9.
Hutchinson, J.N. 2004. Review of flow-like mass
movements in granular and fine-grained Kalendovsky, M.A. and S.H. Cannon. 1997. Fire-
materials. In: Proc., International workshop on induced water-repellent soils: an annotated
occurrence and mechanism of flow-like land- bibliography. U.S. Geol. Surv., Golden, Colo.
slides in natural slopes and earthfills. Sorrento, Open-file Rep. 97-720. http://pubs.usgs.
May 14–16, 2003. L. Picarelli (editor). Patron gov/of/1997/720/pdf/OFR-97-720.pdf (Accessed
Editore, Bologna, Italy. May 2010).

Hutchinson, J.N. and R. Bhandhari. 1971. Undrained Karanka, E.J. 1986. Trends and fluctuations in
loading: a fundamental mechanism of mud precipitation and stream runoff in the Queen
flows and other mass movements. Geotech- Charlotte Islands. B.C. Min. For., Res. Br., Vic-
nique 21:353–358. toria, B.C. Land Manag. Rep. No. 40. www.for.
gov.bc.ca/hfd/pubs/Docs/Mr/Lmr/Lmr040.pdf
Jakob, M. 2000. The impacts of logging on landslide (Accessed May 2010).
activity at Clayoquot Sound, British Columbia.
Keaton, J.R. and J.V. DeGraff. 1996. Surface obser-
Catena 38:279–300.
vation and geologic mapping. In: Landslides
Jakob, M., M. Bovis, and M. Oden. 2005. The signifi- investigation and mitigation. A.K. Turner
cance of channel recharge rates for estimating and R.L. Schuster (editors). Natl. Res. Counc.,
debris-flow magnitude and frequency. Earth Transport. Res. Brd. Washington, D.C. Spec.
Surf. Process. Land. 30:755–766. Pap. No. 247, pp. 178–230.
Jakob, M., K. Holm, O. Lange, and J. Schwab. 2006. Keefer, D.K. 1984. Landslides caused by earthquakes.
Hydrometeorological thresholds for landslide Geol. Soc. Am. Bull. 95:406–421.

270
Lang, A., J. Moya, J. Corominas, L. Schrott, and R. Nagle, H.K. 2000. Folic debris slides near Prince
Dikau. 1999. Classic and new dating methods Rupert, British Columbia. MSc thesis. Univ.
for assessing the temporal occurrence of mass Alberta, Edmonton, Alta.
movements. Geomorphology 30:33–52.
Nilson, E. 2005. From landscape painting to remote
Lipovsky, P.S., J. Coates, A.G. Lewkowicz, and E. sensing: reconstructing 200 years of land-cover
Trochim. 2005. Active-layer detachments fol- change in Central Europe. Gottinger Geogra-
lowing the summer 2004 forest fires near Daw- phische Abhandlungen 113:331–339.
son City, Yukon. In: Yukon exploration and
Noetzli, J., C. Huggel, M. Hoelzle, and W. Haeberli.
geology 2005. D.S. Emond, L.H. Weston, G.D.
2003. GIS-based modelling of rock/ice ava-
Bradshaw, and L.L. Lewis (editors). Yukon
lanches from Alpine permafrost areas. Com-
Geol. Surv., Whitehorse, Y.T., pp. 175–194.
put. Geosci. 10:161–178.
Martin, Y., K. Rood, J.W. Schwab, and M. Church.
Pararas-Carayannis, G. 1999. Analysis of mechanism
2002. Sediment transfer by shallow landsliding
of the giant tsunami generation in Lituya Bay
in the Queen Charlotte Islands, British Colum-
on July 9, 1958. In: Abstracts, Tsunami Symp.
bia. Can. J. Earth Sci. 39(2):189–205.
1999, May 25–27, 1999, Honolulu, Hawaii.
Mathews, W.H. 1979. Landslides of central Vancou- http://users.tpg.com.au/users/tps-seti/tsym.
ver Island during the 1946 earthquake. Bull. html (Accessed May 2010).
Seismol. Soc. Am. 69(2):445–450.
Parker, G.G., Jr. and C.G. Higgins. 1990. Piping and
McCuaig, S.J. 2000. Glacial history of the Nass pseudokarst in drylands, with case studies by
River region. PhD thesis. Simon Fraser Univ., G.G. Parker, Sr., W.W. Wood. In: Groundwater
Burnaby, B.C. geomorphology: the role of subsurface water
in earth surface processes and landforms. C.G.
McDougall, S. and O. Hungr. 2004. A model for
Higgins and D.R. Coates (editors). Geol. Soc.
the analysis of rapid landslide motion across
Am., Boulder, Colo. Spec. Pap. No. 252, pp.
three-dimensional terrain. Can. Geotech. J.
77–110.
41:1084–1097.
Patton, F.D. 1976. The Devastation Glacier slide,
Menounos, B., E. Schiefer, and O. Slaymaker. 2006.
Pemberton, B.C. In: Program and abstracts,
Nested temporal-scale sediment yield est-
Geol. Assoc. Can., Cordillera Sect., pp. 26–27.
imates, Coast Mountains, British Columbia,
Canada. Geomorphology 79:114–129. Picarelli, L., G. Urcioli, M. Ramodini, and L.
Comegna. 2005. Main features of mudslides in
Millard, T.H. 1999. Debris flow initiation in coastal
tectonized highly fissured clay shales. Land-
British Columbia gullies. B.C. Min. For.,
slides 1:15–30.
Vancouver For. Reg., Nanaimo, B.C. Tech. Rep.
No. TR-002. www.for.gov.bc.ca/RCO/research/ Pierson, T.C. and J.E. Costa. 1987. A rheologic clas-
georeports%5Ctr002.pdf (Accessed May 2010). sification of subaerial sediment-water flows. In:
Debris flows/avalanches: process, recognition,
Millard, T.H., D.J. Wilford, and M.E. Oden. 2006.
and mitigation. J.E. Costa and G.F. Wieczoreck
Coastal fan destabilization and forest manage-
(editors). Geol. Soc. Am., Boulder, Colo. Rev.
ment. B.C. Min. For. Range, Coast For. Reg.,
Eng. Geol. 7:1–12.
Nanaimo, B.C. Tech. Rep. No. TR-034.
www.for.gov.bc.ca/rco/research/georeports/ Prior, D.B. and B.D. Bornhold. 1988. Submarine
tr-034.pdf (Accessed May 2010). morphology and processes of fjord fan deltas
and related high-gradient systems: modern
Moore, D.P. and D.J. Mathews. 1978. The rubble
examples from British Columbia. In: Fan
Creek landslide, southwestern British Colum-
deltas: sedimentology and tectonic settings. W.
bia. Can. J. Earth Sci. 15:1039–1052.
Nemee and R.J. Steel (editors). Blackie and Son,
Murty, T.S. 1979. Submarine slide-generated water London, U.K., pp. 125–143.
waves in Kitimat Inlet, British Columbia. J.
Geophys. Res. 84:7777–7779.

271
_______. 1990. The underwater development of _______. 1999. Tsunamis on Troitsa Lake, Brit-
Holocene fan deltas. In: Coarse-grained deltas. ish Columbia. B.C. Min. For., For. Sci. Sect.,
A. Colella and D.B. Prior (editors). Int. Assoc. Smithers, B.C. Exten. Note No. 35. www.for.
Sedimentol. Spec. Publ. No. 10, pp. 75–90. gov.bc.ca/rni/Research/Extension_notes/
Enote35.pdf (Accessed May 2010).
Prior, D.B., B.D. Bornhold, and M.W. Johns. 1984.
Depositional characteristics of a submarine _______. 2000. Donna Creek washout-flow. B.C.
debris flow. J. Geol. 92:707–727. Min. For., Prince Rupert For. Reg., For. Sci.
Sect., Exten. Note No. 42. www.for.gov.bc.ca/
Robichaud, P.R., J.L. Beyers, and D.G. Neary. 2000.
rni/Research/Extension_notes/Enote42.pdf
Evaluating the effectiveness of post-fire re-
(Accessed May 2010).
habilitation treatments. U.S. Dep. Agric. For.
Serv., Rocky Mtn. Res. Stn., Fort Collins, Colo. _______. 2001. Donna Creek washout-flow: what
Gen. Tech. Rep. RMRS-GTR-63. did we learn? In: Terrain stability and forest
management in the interior of British Colum-
Rogers, G.C. 1980. A documentation of soil failure
bia: Workshop proceedings. May 23–25, 2001.
during the British Columbia earthquake of 23
Nelson, B.C. P. Jordan and J. Orban (editors).
June, 1946. Can. Geotech. J. 17:122–127.
B.C. Min. For., For. Sci. Sect., place Tech. Rep.
Rood, K.M. 1984. An aerial photograph inventory No. 003. pp. 1–13. www.for.gov.bc.ca/hfd/pubs/
of the frequency and yield of mass wasting on Docs/Tr/Tr003/Schwab.pdf (Accessed May
the Queen Charlotte Islands, British Columbia. 2010).
B.C. Min. For., Victoria, B.C. Land Manag.
Schwab, J.W., M. Geertsema, and A. Blais-Stevens.
Rep. No. 34. www.for.gov.bc.ca/hfd/pubs/Docs/
2004. The Khyex River landslide of Novem-
Mr/Lmr/Lmr034.pdf (Accessed May 2010).
ber 28, 2003, Prince Rupert British Columbia,
Sanborn, P., M. Geertsema, A.J.T. Jull, and B. Canada. Landslides 1:243–246.
Hawkes. 2006. Soil charcoal evidence for Ho-
Schwab, J.W. and M. Kirk. 2003. Sackungen on a
locene fire-related sedimentation in an inland
forested slope, Kitnayakwa River. B.C. Min.
temperate rain forest, east central British Co-
For., Prince Rupert For. Reg., Prince Rupert,
lumbia, Canada. The Holocene 16:415–427.
B.C. Exten. Note No. 47. www.for.gov.bc.ca/rni/
Schwab, J.W. 1983. Mass wasting: October–November research/Extension_Notes/Enote47.pdf (Ac-
1978 storm, Rennell Sound, Queen Charlotte cessed May 2010).
Islands, British Columbia. B.C. Min. For.,
Scott, D. and R. Pike. 2003. Wildfires and watershed
Victoria, B.C. Res. Note No. 91. www.for.gov.
effects in the southern B.C. interior. Streamline
bc.ca/hfd/pubs/Docs/Mr/Scanned-Rn/Rn067-
Watershed Manag. Bull. 7(3):1–4. www.forrex.
Rn100/Rn091.pdf (Accessed May 2010).
org/publications/streamline/ISS26/streamline_
_______. 1997. Historical debris flows, British Co- vol7_no3_art1.pdf (Accessed March 2010).
lumbia north coast. In: Proc., Forestry geotech-
Septer, D. and J.W. Schwab. 1995. Rainstorm and
nique and resource engineering, Richmond,
flood damage: northwest British Columbia
B.C. BiTech Publishers, Vancouver, B.C.
1891–1991. B.C. Min. For., Res. Br., Victoria,
_______. 1998. Landslides on the Queen Charlotte B.C. Land Manag. Handb. No. 31. www.for.gov.
Islands: processes, rates, and climatic events. bc.ca/hfd/pubs/docs/Lmh/Lmh31.pdf (Ac-
In: Carnation Creek and Queen Charlotte Is- cessed May 2010).
lands fish/forestry workshop: applying 20 years
Shakesby, R.A. and S.H. Doerr. 2005. Wildfire as
of coastal research to management solutions.
a hydrological and geomorphological agent.
D.L. Hogan, P.J. Tschaplinski, and S. Chatwin
Earth Sci. Rev. 74:269–307.
(editors). B.C. Min. For. Res. Br., Victoria, B.C.
Land Manag. Handb. No. 41, pp. 41–47. www. Shroder, J.F. 1980. Dendrogeomorphology: review
for.gov.bc.ca/hfd/pubs/docs/Lmh/Lmh41-1.pdf and new techniques of tree-ring dating. Prog.
(Accessed May 2010). Phys. Geogr. 4:161–188.

272
Sidle, R.C. and H. Ochiai. 2006. Landslides: pro- Society, May 6–9, 1992. BiTech Publishers,
cesses, prediction, and land use. Water Resour. Vancouver B.C. pp. 193–197.
Monogr. Vol. 18. Am. Geophys. Union, Wash-
VanDine, D.F., R.F. Rodman, P. Jordan, and J. Du-
ington, D.C.
pas. 2005. Kuskonook Creek: an example of a
Siebert, S.A. 1987. A case history of a catastrophic debris flow analysis. Landslides 2:257–265.
erosion in the Coquitlam valley, B.C. BSc the-
Varnes D.J. 1978. Slope movement types and process-
sis. Geol. Eng. Program, Univ. British Colum-
es. In: Landslides, analysis and control. R.L.
bia, Vancouver, B.C.
Schuster and R.J. Krizek (editors). Transport.
Soil Classification Working Group. 1998. The Cana- Res. Brd., Natl. Acad. Sci. Spec. Rep. No. 176,
dian system of soil classification. 3rd edition. pp. 11–33.
Natl. Res. Counc. Press, Ottawa, Ont. Agric.
Weir, P. 2002. Snow avalanche management in for-
Agri-Food Can. Publ. No. 1646.
ested terrain. B.C. Min. For., Res. Br., Victoria,
Souther, J.G. 1971. Geology and mineral deposits of B.C. Land Manag. Handb. No. 55. www.for.gov.
Tulsequah map-area, British Columbia (104K). bc.ca/hfd/pubs/Docs/Lmh/Lmh55.pdf (Ac-
Geol. Surv. Can., Dep. Energy, Mines, and cessed May 2010).
Resources, Ottawa, Ont. Memoir No. 362.
Wieczorek, G.F. and T. Glade. 2005. Climatic factors
Swanston, D.N. and F.J. Swanson. 1976. Timber influencing occurrence of debris flows. In: De-
harvesting, mass erosion, and steepland forest bris flow hazards and related phenomena. M.
geomorphology in the Pacific Northwest. In: Jakob and O. Hungr (editors). Springer-Praxis,
Geomorphology and engineering. D.R. Coates Berlin, Germany, pp. 325–362.
(editor). Dowden, Hutchinson, and Ross,
Wilford, D.J., P. Cherubini, and M.E. Sakals. 2005a.
Stroudsburg, Pa., pp. 199–221.
Dendroecology: a guide for using trees to date
Takahashi, T. 1981. Debris flow. Ann. Rev. Fluid geomorphic and hydrologic events. B.C. Min.
Mech. 13:57–77. For., Res. Br., Victoria, B.C. Land Manag.
Handb. No. 58. www.for.gov.bc.ca/hfd/pubs/
Turner, A.K. and R.L. Schuster (editors). 1996.
Docs/Lmh/Lmh58.pdf (Accessed May 2010).
Landslides investigation and mitigation. Natl.
Res. Counc., Transport. Res. Brd., Washington, Wilford, D.J., M.E. Sakals, and J.L Innes. 2005b.
D.C. Spec. Rep. No. 247. Forest management on fans: hydrogeomor-
phic hazards and general prescriptions. B.C.
Vallance, J.W. 2005. Volcanic debris flows. In: De-
Min. For., Res. Br., Victoria, B.C. Land Manag.
bris-flow hazards and related phenomena. M.
Handb. No. 57. www.for.gov.bc.ca/hfd/pubs/
Jakob and O. Hungr (editors). Springer-Praxis,
Docs/Lmh/Lmh57.pdf (Accessed March 2010).
Berlin, Germany, pp. 247–274.
Williams, M. 1966. The grand campus washout.
VanDine, D.F. 1985. Debris flows and debris torrents
Univ. British Columbia Alumni Chron.
in the southern Canadian cordillera. Can. Geo-
20:9–11.
tech. J. 22:44–68.
Wischmeier, W.H. and D.D. Smith, 1960. A universal
VanDine, D.F. and S.G. Evans. 1992. Large landslides
soil-loss equation to guide conservation farm
on Vancouver Island, British Columbia. In:
planning. Trans. Int. Congr. Soil Sci. 7:418–425.
Geotechnique and natural hazards: a sympo-
sium sponsored by the Vancouver Geotechni- Zischinsky, U. 1969. Uber sackungen. Rock Mechan.
cal Society and the Canadian Geotechnical 1:30–52.

273
View publication stats