2014-038

2015
research-articleResearch article10.1144/qjegh2014-038Mapping landslides at different scalesJames S. Griffiths, Anne E. Mather, Martin StokesXXX10.1144/qjegh2014-038J. S. Griffiths et al.Mapping landslides

Research article Quarterly Journal of Engineering Geology and Hydrogeology

Published online January 28, 2015 doi:10.1144/qjegh2014-038 | Vol. 48 | 2015 | pp. 29­–40

Mapping landslides at different scales

James S. Griffiths*, Anne E. Mather & Martin Stokes
School of Geography, Earth and Environmental Sciences, Plymouth University, Drake Circus, Plymouth PL4 8AA, UK
* Correspondence: jim.griffiths@plymouth.ac.uk

Abstract:  Landslides can be mapped for purely academic reasons to understand their nature and properties, or for engi-
neering purposes because they constitute a hazard to an existing human activity or a proposed development. The scale of
the mapping to be carried out will be influenced by the requirements of the investigation. For engineering investigations
large-scale mapping (>1:10000) is generally the most appropriate whereas for academic studies smaller-scale mapping can
be just as valuable as it can reveal much about the role landsliding has in overall landscape development. In this paper three
examples of mapping landslides at different scales are presented and the implications for the mapping procedures employed
are explored. These examples are: (1) large-scale mapping of an active landslide affecting a road and mobile home site on
the SW coast of the Isle of Wight, UK; (2) medium-scale mapping of landslides in the proximity of a river capture site in SE
Spain; (3) small-scale mapping of large coastal landslides in northern Chile and the implications for mass sediment transfer
along an active subduction margin. All three mapping programmes had a clear focus and intent but the method employed
varied with the purpose of the investigation and the scale of the mapping that was carried out. It was concluded that it is
important for any landslide mapping programme to have clearly identified objectives from the outset if time, and money, is
not be wasted. However, even in site-specific investigations it is imperative for landslides to be seen in the wider context of
their geomorphological situation with their role in landscape development assessed so as to understand the likelihood and
potential magnitude of any hazards.
Received 5 January 2014; revised 21 November 2014; accepted 21 November 2014

Dramis et al. (2011, p. 43) stated that ‘scale is one of the main categories are presented in Table 1, and although the boundaries to
issues in geomorphological mapping’, and this comment is partic- the categories should be regarded as flexible, there are differences
ularly relevant to the geomorphological and engineering geologi- in the way the maps are compiled, the techniques employed in their
cal investigation of landslides. The investigation of landslides is a construction, and their intended end-use.
longstanding pursuit given these are one of the most deadly natural There are many instances of large-scale mapping of landslides;
hazards and each year are responsible for hundreds of millions of excellent examples used for landslide hazard assessment in Hong
dollars of damage and the loss of more than 1000 lives (Clague & Kong have been given by Parry (2011). Medium-scale engineering
Roberts 2012). Indeed, CRED (2012) suggested that landslides geology maps are probably the most widely used and examples can
were responsible for 10000 deaths and left 2.5 million people be found throughout the world; for example, the 1:25000 scale
homeless over the period 2001–2010. ZERMOS maps in France (Humbert 1977), and the 1:20000 scale
One of the most important techniques for determining the nature GASP series in Hong Kong and notably its use in the compilation
and extent of landsliding is through a systematic process of geo- of a landslide inventory (Burnett et al. 1985). Small-scale maps
morphological and engineering geological mapping. However, generally are best exemplified by the PUCE (Pattern–Unit–
before any mapping is undertaken it is necessary to establish the Component–Evaluation) system developed by CSIRO in Australia
purpose of the investigation; fundamentally the question is whether (Finlayson 1984). This programme produced maps at a scale of
the investigation is going to concentrate on a discrete landslide or 1:250000 that defined broad terrain patterns, within which particu-
whether the work is a much broader evaluation of landslide hazard lar assemblages of landforms, pedological soils and vegetation
and risk (Lee & Jones 2014). Also, it must be accepted that map- sequences occur. However, a small-scale (1:100000) version of the
ping to solve an engineering issue, such as the potential for a land- French ZERMOS system was used in some areas for the specific
slide to damage an existing or proposed development, is distinct purpose of investigating landslide hazard (Landry 1980).
from more academic geomorphological investigations related to It is not just the maps themselves where ‘scale’ is an issue, and
long-term landscape development. In this paper these issues are Figure 1 provides a standard geomorphological nomenclature for
explored through three case studies at different scales and for dis- landform size, illustrating this with a landslide example. There
tinctly different purposes: (1) large-scale mapping of an active may even be difficulty when investigating landslides in actually
landslide on the Isle of Wight, UK, that was affecting a main road defining what constitutes a ‘landslide’ (Griffiths 1999). The scale
and mobile home site; (2) medium-scale mapping initially under- and purpose of the mapping will determine what size of geomor-
taken as part of a hazard and risk assessment within a single river phological feature is recorded. For example, even large-scale map-
catchment in SE Spain; (3) small-scale mapping of the coastal cor- ping for engineering purposes may not record micro-features, such
dillera in northern Chile to evaluate the role of landsliding in the as spalling or small rockfalls, as they are below the threshold of
mass transfer of sediment in an active subduction margin. what can actually be recorded on a map sheet, whereas single large
landslide blocks and the nature of the landslide movement can be
mapped in detail. In contrast, typical medium-scale maps of
The question of scale
national geological surveys (1:10000 to 1:50000) will record a
In mapping landslides it is possible to envisage three broad landslide only as a feature with its spatial extent delimited rather
scale categories (large, medium, and small) being utilized. These than identifying details of its internal structure. Small-scale

© 2015 The Author(s). Published by The Geological Society of London. All rights reserved. For permissions: http://www.geolsoc.org.uk/permissions.
Publishing disclaimer: www.geolsoc.org.uk/pub_ethics
30 J. S. Griffiths et al.

>ĂŶĚĨŽƌŵ^ŝnjĞĂƚĞŐŽƌŝĞƐ
>ĂŶĚĨŽƌŵ ƐŝnjĞ >ŝŶĞĂƌĐŽŵƉŽŶĞŶƚ ;ŬŵͿ ^ƉĂƚŝĂůĐŽŵƉŽŶĞŶƚ ;ŬŵϮͿ
ĐĂƚĞŐŽƌLJ
DĞŐĂ хϭϬϯ ;ϭϬϬϬ͛ƐŬŵͿ хϭϬϲ

DĂĐƌŽ ϭϬͲϭϬϯ ;ϭϬ͛ƐƚŽ ϭϬϬϬ͛ƐŬŵͿ ϭϬϮͲϭϬϲ

DĞƐŽ Ϭ͘ϭͲϭϬ;ϭŵƚŽ ϭϬŵͿ Ϭ͘ϬϬϭͲϭϬϮ

DŝĐƌŽ фϬ͘ϭ;фϭŵͿ ϭϬͲϲͲϭϬͲϮ

>ĂŶĚƐůŝĚĞĞdžĂŵƉůĞ

DĞŐĂ DĂĐƌŽ DĞƐŽ DŝĐƌŽ

ŽŶǀĞƌŐĞŶƚƉůĂƚĞ /ŶĐŝƐĞĚƐĞĚŝŵĞŶƚĂƌLJ /ŶĚŝǀŝĚƵĂůůĂŶĚƐůŝĚĞ ^ŵĂůůƌŽĐŬĨĂůůƐ ĂŶĚ

ŵĂƌŐŝŶ ďĂƐŝŶ ƐƉĂůůŝŶŐ

Î >E&KZD ,/ZZ,z Î ƌĞůĂƚŝŽŶƐŚŝƉďĞƚǁĞĞŶ͚ƐƉĂĐĞ͛ĂŶĚƚŝŵĞ

Fig. 1. The size of geomorphological features (after Summerfield 1991).

­ apping is one further step removed from this and is likely to

m Kong originally outlined by Brand et al. (1982) and Burnett &
record only the location of a landslide, normally as the first stage Styles (1982). For an overall approach to the identification of land-
in compiling an inventory. slide hazard and risk reference should be made to the book by Lee
Therefore, it emerges that some of the critical decisions that & Jones (2014). A summary of the techniques to be employed in
need to be made before embarking on any programme of landslide general terrain assessment has been given by Lawrance et al.
mapping are as follows. (1993) and Griffiths (2001), and the continuing relevance of the
terrain evaluation approach for engineering is demonstrated by the
• What is the purpose of the investigation? work of Hearn (2011a,b) and Shilston et al. (2012).
• What scale of mapping is most appropriate for the required In the following sections three case studies of the investigation
purpose, as this will affect the techniques that have to be of landslides are presented. In each case the original reason for and
employed (Table 1)? purpose of the investigation is defined, followed by a description
• What is the minimum size of landslide feature that can be of the geology and geomorphology of the area investigated and
recorded? then some conclusions from the work.
• What level of detail is needed about any single landslide to
meet the required purpose? Large-scale mapping of an active landslide
(Binnel Bay, Isle of Wight, UK)
Mapping approaches
The techniques of mapping landslides have been widely docu- In January 2001 the Binnel Bay landslide on the Isle of Wight,
mented and do not need to be repeated here. A detailed review of located off the south coast of the UK, caused damage to the main
the engineering geomorphological investigation of landslides has road in the south of the island and virtually destroyed a mobile home
been given by Griffiths & Whitworth (2012). An account of gen- site. Large-scale geomorphological mapping was used to identify
eral geomorphological mapping for applied purposes, along with a the nature of the landslide problem, and to provide a framework for
collection of case studies at a range of mapping scales, has been subsequent ground investigations and design of remedial measures.
given by Smith et al. (2011). For engineering geological mapping The Binnel Bay landslide lies within the feature known as the
the book by Dearman (1991) remains the definitive volume. The Undercliff landslide complex, which is one of the largest areas of
use of medium- and large-scale terrain classification techniques in landsliding in Europe (Chandler 1984). The Undercliff is a dip
landslide investigations is best exemplified by the work in Hong slope landslide complex that has developed off the south coast of
 Mapping landslides
31
Fig. 2. Basic geology map of the South Downs of the Isle of Wight (after BGS 2013).
32 J. S. Griffiths et al.

Fig. 3. Geomorphological map of the Binnel Bay landslide.

the island (Fig. 2) with a terrestrial component up to 1 km wide and Undercliff landslide complex have occurred along these Sandrock
some 10 km long. There is extensive offshore evidence to indicate Formation clay layers, but movements both within and in the upper
that the landslide has been part of a retrogressive landslide system parts of the main landslide occur at the boundary between the plas-
that has migrated shoreward in response to rising sea levels since tic clay and silty clay lithologies in the Gault Formation.
melting of terrestrial glacial ice at the end of the Younger Dryas The Binnel Bay landslide occurred towards the western end of the
Period, c. 11500 years ago (Lee et al. 1991). Undercliff landslide complex, close to the St. Catherine’s Point land-
The geology of the South Downs of the Isle of Wight comprises slide investigated by Hutchinson et al. (1991). Investigations at the
a conformable series of Cretaceous strata dipping c. 2° to the south, site comprised geomorphological field mapping at 1:2500 scale by a
forming an escarpment (Fig. 2). The highest part of the sequence two person team over 1½ days using global positioning system
exposed on the southern coast is the fossiliferous calcarenites and (GPS), compass–clinometer, Suunto clinometer, and 50 m tape. The
calcareous mudstones of the Chalk, comprising the Holywell main landslide movements were essentially north to south with the
Nodular Chalk Formation, the Zig Zag Chalk Formation and the backscar some 2 m downslope of the road, although tension cracks
West Melbury Marly Chalk Formation. Of most relevance to the were appearing in the road at the time of the site visit. By reference
landsliding at Binnel Bay is the geological sequence below the to base maps from 1995 it was established that the backscar had
Chalk, starting with the ≥40 m of the Upper Greensand Formation. retrogressed 5 m over a 5 year period. Within the landslide complex
The weak sandstones of the Upper Greensand Formation form a itself there was extensive evidence of active movement, notably
significant aquifer developed on the 45 m thick Gault Formation exposed shear surfaces, stretched roots and vegetation, debris runout
aquiclude (Fig. 3). The Upper Greensand Formation conformably over beach gravels, disrupted drainage, fallen trees, Reidel shears on
overlies the Gault Formation, which can be divided into two dis- the slide margins, out of place geological units, and damage to prop-
tinct lithological units, an upper plastic clay and a lower, much erty. The upper part of the landslide contained detached boulders
more silty, less plastic clay. The base of the Gault Formation is from the Upper Greensand Formation, but the main part of the land-
identified by the 3.7 m thick Monk’s Bay Sandstone Formation. slide constituted debris of disrupted and remoulded Gault Formation.
The underlying Sandrock Formation sequence is mainly a weak Geomorphological mapping identified four distinct landslide
sandstone containing clay layers up to 1 m thick. The most sub- components, labelled A, B, C and D in Figure 3. There was no
stantial movements on the basal shear surface of the main indication that these landslides were part of a large-scale reactiva-
 Mapping landslides 33

Fig. 4. Cross-section through the Binnel Bay landslide.

Fig. 5. Simplified geology map of the

study area in SE Spain (from Mather &
Harvey 1995).
 34
J. S. Griffiths et al.
Fig. 6. Isopleth map of the landslide distribution in the Río Aguas catchment (from Hart 2004).
 Mapping landslides 35

tion of the overall Undercliff landslide complex but they repre- The region has been affected by epeirogenic uplift and the area
sented continuing degradational movements within the debris remains seismically active; it is these processes that have had a
apron. Landslides A and C were part of an upper landslide move- significant effect on the nature and distribution of landsliding in
ment that appeared to be failing along the lithological boundary the area. The uplift since the Late Pliocene is calculated as
within the Gault Formation. The main movement was landslide A, 80 m Ma−1 in the central areas of the Sorbas Basin and 15 m Ma−1 in
with landslide C occurring as a result of the loss of support along the Vera Basin (Mather 1991; Mather et al. 2001). The differential
the eastern boundary caused by the landslide reactivation. uplift rates between the sedimentary basins increased regional gra-
Landslide A appeared to be spilling over an in situ bench in the dients, which, when combined with the variable lithologies of the
Sandrock Formation. Landslide D was a discrete part of this in situ basin fill (generally comprising weak calcareous mudstones–
bench that had also started to retrogress. marls, weak gypsum, medium strong and strong sandstones and
Based on the mapping, the cross-section shown in Figure 4 was limestones) led to fluvial incision, the creation of steep hillside
drawn up as a basis for designing and interpreting the subsequent slopes, and river capture. The sediment and water routing within
ground investigations. Given there was no indication of basal ero- the Río Aguas catchment was severely altered by a river capture
sion by the sea, it was hypothesized that the cause of the reactiva- event estimated to have occurred at about 100 ka, based on the U/
tion of the landslide was high porewater pressures in the landslide Th isochron dating of carbonate soils from river terrace landforms
system resulting from the previous very wet autumn. It was sug- at c. 88 ka (68–104 ka) (Kelly et al. 2000). The river capture low-
gested that the landslide movements would continue if rainfall ered base levels at the point of capture (Fig. 5) by c. 90 m (Harvey
remained high. The mapping established that to determine the et al. 1995), and resulted in c. 73% of the Sorbas Basin drainage
exact nature of the landslide would require ground investigation being re-routed from its original southerly route through the Níjar–
and monitoring, and it was recommended that any remedial meas- Carboneras Basin via the Rio Feos, into an easterly one through the
ures would have to be based on effective drainage rather than tech- Vera Basin (Mather & Harvey 1995; Mather 2000). It has been
niques such as toe-loading, anchoring or retaining walls. estimated that one consequence of the base-level change was a
What emerged from this study was that to obtain the detail five-fold increase in incision rates over most of the river network,
needed for the design of an engineering investigation and solution and close to the point of capture incision rates may have increased
to the consequences of the landslide movement necessitated map- 10-fold (Mather 2000; Mather et al. 2002; Stokes et al. 2002).
ping by somebody experienced in engineering requirements. This The first phase of the landslide research was to compile an
project utilized large-scale field mapping techniques as it was nec- inventory of landslide occurrence within a single river catchment,
essay to identify all the various components of the landslide com- the 425 km2 Río Aguas. This was completed through a combina-
plex to gain an understanding of the landslide mechanisms and tion of medium-scale terrain evaluation and selected medium- to
their likely cause (Figs 3 and 4). It was recognized, however, that large-scale geomorphological mapping from aerial photograph
the investigation of the site was aided by the extensive previous interpretation and in the field. This work revealed an unexpected
work that had been undertaken on the whole Undercliff landslide landslide distribution, illustrated by the isopleth map (Fig. 6) with
complex (Chandler 1984; Hutchinson et al. 1991; Lee et al. 1991) a concentration immediately upstream of the identified main river
and the data from these investigations were compiled during desk capture site rather than in the higher and steeper ground in the
studies carried out prior to going on site. mountains that formed the boundaries of the watershed in the north
and south. Clearly, the river capture had had a profound effect on
Medium-scale mapping to establish the the geomorphology (Mather et al. 2002). In the Sorbas Basin, deep
canyons had developed in the lithological units that were more
geomorphological controls on a landslide
resistant (the strong sandstones and limestones), and large
cluster (Río Aguas Basin, SE Spain)
V-shaped valleys had formed in the weaker lithologies (notably
The requirement for an investigation of landslides in the Río Aguas the weak calcareous mudstones–marls). In the formation of this
Basin, which is located in the semi-arid SE part of Spain, emerged terrain sediment had been liberated from the slopes to supply both
from research into the Cenozoic evolution of the drainage system suspended and bedload material to the fluvial system through a
(Mather 1991, 2000; Mather & Harvey 1995; Mather et al. 2002; variety of hillslope processes. The topography generated by the
Stokes et al. 2002). It was recognized that landsliding was a signifi- enhanced incision provided the ideal situation for landslides to
cant process in the development of the contemporary landscape (e.g. become a significant mechanism for this sediment transfer. Also,
Mather et al. 2003) and an important contributor to the sediment flux given the rapid rates of incision, the landscape had little time to
(Mather et al. 2002; Stokes et al. 2002). In March 1998 it was decided adjust and hence many hillside slopes were over-steepened, leav-
to develop a research programme to investigate the scale and extent of ing them in a state of marginal stability. This had led to a concen-
landsliding with the specific objectives of evaluating landslide causes tration of landslides around the river capture site with a ‘wave of
and mechanisms, ascertaining what had initiated landslide activity, aggression’ (Bjerrum & Jorstad 1968) making its way through the
identifying the controls on susceptibility and hazard, and establishing catchment, initially created by the differential uplift between the
the role of landslide processes in landscape development. This work Vera and Sorbas basins, and then continuing as a result of the
was based on field mapping, remote sensing interpretation, data colla- increase in stream power following the river capture (Griffiths
tion and analyses (Hart 2004; Griffiths et al. 2005). et al. 2005). Because the river capture event had occurred rela-
The general geology of the study area is presented in Figure 5. In tively recently in the geological record the wave of headward ero-
broad geomorphological terms the area can be described as a ‘basin sion has yet to extend more than 25 km upstream of the capture
and range’ with predominantly Neogene rocks infilling fault- point, hence the concentration of landslide activity (Mather et al.
bounded sedimentary basins surrounded by mountains (sierras) 2002; Mather et al. 2003). The landsliding at the river capture site
formed in Mesozoic and older basement rocks that make up a com- is active to the present day, with massive tension cracks visible in
plex nappe sequence (Sanz de Galdeano 1990). The Sorbas and Vera the area close to the original elbow of capture. Elsewhere within
basins (Fig. 5) are part of the Trans-Alboran shear zone of Larouziére the zone of the migrating wave of aggression there are a number of
et al. (1988), a zone dominated by sinstral movement within the large landslides, including Maleguica (Fig. 7), located 8  km
internal zone of the Betics. A detailed description of the develop- upstream from the capture site with a volume of >1 million m3.
ment of these Neogene basins and their subsequent Quaternary ero- Large-scale mapping of the landslides in the capture col area
sional history has been given by Mather et al. (2001). was subsequently undertaken to establish the local controls on the
36 J. S. Griffiths et al.

Fig. 7. The Maleguica landslide located

8 km upstream of the river capture site.

Table 1. Scale of landslide mapping necessary to meet different requirements (developed from Griffiths & Whitworth 2012; Shilston et al. 2012)
Large-scale
• 1:10000 or greater. Plans and sections up to 1:500 scale can be used in particular circumstances
• Based on detailed field mapping and ground investigation data. These data will be supplemented by: interpretation of large-scale aerial photographs
and high-resolution airborne and satellite imagery; development of satellite-derived digital elevation models (DEMs); analysis of airborne and ground
LIDAR. Data should be compiled into a suitable spatial database (geographical information system; GIS) to allow the creation of a ground model that can
form the basis for subsequent ground investigations (Parry et al. 2014)
• Uses: investigating the type, nature and causes of single landslides; designing and interpreting ground investigations of a landslide; preparing remedial
measure designs
• End-users: applied and academic geomorphologists; civil engineers; engineering geologists; geotechnical engineers; quantity surveyors
Medium-scale
• 1:10000 to 1:100000 scale
• Based on: terrain evaluation and analysis; remote sensing interpretation (aerial photographs and satellite imagery); satellite-derived DEMs; airborne
LIDAR; ground-truth field mapping of selected areas. These data can readily be compiled into a GIS to facilitate interpretation and a range of high-
quality outputs including natural hazard and risk studies (Lee & Jones 2014)
• Uses: investigating the types and causes of a landslide and placing it in its ‘geomorphological situation’; landslide hazard and risk evaluations at local to
regional level; locating landslides for route alignment planning; disaster relief planning; landscape development studies
• End-users: academic geomorphologists and geologists; civil engineers; engineering geologists; local and regional planners; local and regional
government; insurers; emergency services
Small-scale
• 1:100000 or smaller
• Based on: terrain evaluation and analysis; remote sensing interpretation mainly of ASTER imagery and Landsat 7 ETM data; limited reconnaissance of
field area usually by vehicle or helicopter. Remote sensing data collected may be suitable for compilation in a GIS to help create hazard and risk maps
(e.g. Teeuw 2007)
• Uses: generalized landslide hazard and risk evaluations; development planning at the regional or national level; investigations of long-term landscape
development
• End-users: academic geomorphologists and geologists; local and regional planners; regional and central government; insurers

failures. This demonstrated that the local lithology was the major surveys covering some 450 km of the seaward scarp of the coastal
control on the nature of the landslides, which ranged in type from cordillera (Fig. 8a and b) a series of arcuate embayments 1–10 km
large rock falls to translational slides and rock avalanches. across were observed (Fig. 9). The geometry of the embayments
However, it was the original medium-scale mapping that identified was typical of those reported from the backscarps of giant land-
the geomorphologically anomalous landslide concentration and slides associated with volcanic island terrains (e.g. Masson et al.
provoked a more detailed examination of the causes of landsliding 2002). This raised a number of questions. First, were these features
within the catchment. Essentially, this study led to the identifica- landslides or just examples of ‘landform convergence’ or ‘equifi-
tion of a unique geomorphological control on the landslide hazard nality’, where the processes of hillslope, fluvial and coastal ero-
for this area; namely, a river capture event. sion combine to create features that resemble landslides? Second,
if they were landslides, what had caused them? Third, what was
Small-scale mapping: coastal landslide the role of landsliding in the evolution of the contemporary land-
occurrence with an active subduction plate scape of the coastal cordillera?
The study area is developed within Jurassic volcanic rocks, lies
boundary (northern Chile)
between the Pacific coast and western edge of the Coastal
A very different landscape development question was raised dur- Cordillera of northern Chile, and forms part of the Andean fore-arc
ing geological studies of the Coastal Cordillera in a hyper-arid (Von Huene et al. 1999; Taylor et al. 2007). The presence of
region of northern Chile. During vehicular ground reconnaissance Oligocene and Miocene marine sediments along the coastline
 Mapping landslides 37

Fig. 8. Main terrain units of northern

Chile: (a) location map showing
identified coastal landslides plus the
inland Latagulla landslide (LL), the
Lluta collapse (LC), and landslide prone
areas within a drainage basin (LPD); (b)
representative cross-section.

(Ibaraki 2001) indicated that uplift had dominated and that subsid- (1) A coastal plain. This is a narrow and discontinuous plain
ence had been minimal during the Neogene. Application of a geo- present at the foot of the coastal scarp and is defined as the area
morphological index (Stream Gradient Index; Hack 1957) to the beneath the 200 m contour line. It comprises alluvial fan material,
region revealed that the study site was one of the most actively landslide debris and some marine terraces.
uplifting areas of the coastline, with postulated uplift rates of the (2) The coastal Cordillera. This is up to 2.2 km high and 20–
order of 480–1152 mm ka−1 (Mather & Hartley 2006). Although 50 km wide. Delineation of the cordillera as a separate morphotec-
the area lies within seismically active terrain (Spence et al. 1999), tonic unit commenced in the Oligocene and continues to the
historically it is part of a well-known seismic gap with recurrence present day (Hartley et al. 2000). It largely comprises basaltic
intervals for major seismic events in excess of 100 years. The last andesites and granodiorites cut by north–south- to NE–SW-
major seismic event was in 1877 and generated a 25 m run-up tsu- trending, landward-dipping, normal faults that bound a series of
nami in Arica (Kulikov et al. 2005). half-grabens. The eastern flank of the Cordillera is bounded by the
Initial studies of the site had concentrated on understanding the 900 km long Atacama Fault Zone. The western margin of the
drainage systems, regional tectonics and seismicity, and the rela- Coastal Cordillera is represented by a coastal scarp, which forms a
tionship between this fore-arc system and the offshore subduction prominent break in slope 600–1500 m high.
trench (Hartley et al. 2000; Mather & Hartley 2006). However, Where the coastal cordillera scarp is separated from active
once the gross morphology of the region came under scrutiny the coastal erosion by the coastal plain it is considered to represent a
mechanisms for onshore to offshore sediment transfer emerged as weathered palaeo-cliffline (Mortimer & Saric 1972). It is on these
a specific research question and raised the issue of the role of land- areas that large terrestrial landslides have been observed (Fig. 9).
sliding in this process. The age of the scarp is unknown but stratigraphical relationships
The geomorphology of the coastal region is dominated by two suggest that it is composite and extends back to at least the mid-
main terrain units within the Andean land system (Fig. 8), as fol- Miocene. Where the scarp is located on the coast it is suggested
lows. that the large embayments identified represent the degradational
38 J. S. Griffiths et al.

Fig. 9. A landslide generated on the

1 km high western scarp of the Coastal
Cordillera (see small settlement on left of
image for scale).

Fig. 10. Conceptual model of the

relationship between landsliding and
regional tectonic activity, northern Chile
(from Mather et al. 2014).

zones of large landslides that have run out offshore (Mather et al. For this study the investigation of the landslides was carried out
2014). The interpretation of the coastal embayments as landslides as part of the research into mass sediment transfer. As it was neces-
was based on the combination of field reconnaissance and remote sary to identify the nature and extent of landslide activity over
sensing interpretation (Soeters & van Westen 1996) set within a 450 km of coastline it was not feasible, or essential, to map each
process of terrain evaluation (Lawrance et al. 1993). If this inter- landslide in detail, so rapid reconnaissance methods were used.
pretation is correct then over the length of the 450 km long study These included the interpretation of remote sensing and ground
area there are 60 landslides each with volumes up to 9 km3, giving truthing within an overall terrain evaluation framework. The chro-
a landslide density of 1 per 7.5 km of coastline. These are very nology of the landslide model that has been developed (Fig. 10)
large landslides but not as large as the collapsing flanks of hotspot remains speculative and further research is under way to check
volcanic islands such as Hawaii (Moore 1964) and the Canary both its validity within the study area and its applicability to other
Islands (Carracedo et al. 1999). For example, McMurtry et al. convergent plate boundaries (Mather et al. 2014).
(2004) estimated that the Alia landslide that occurred on the south-
western flanks of Mauna Loa, Hawaii, c. 120000 years ago had a
minimum volume of 400 km3. There is no evidence that the embay- Conclusion
ments were created by single massive events and it is possible that Three very different case studies have been provided, all of which
the observed morphology represents the combination of a number fall under the umbrella of landslide mapping. It is apparent that for
of smaller mass movement features that have coalesced. Large engineering measures to be designed on a discrete landslide, large-
coastal and submarine landslides have previously been associated scale mapping by experienced practitioners is required (Case
with mid-ocean volcanic islands formed over mantle hotspots Study 1). These maps should form the basis for the design and
(Keating & McGuire 2000; Masson et al. 2002; Mitchell et al. interpretation of the ground investigations and must remain live
2002). Collapse of these volcanic edifices is believed to have been documents that are updated as new information emerges. For land-
responsible for mega-tsunamis (Ward 2001; Whelan & Kelletat slide hazard investigations, commonly required for engineering
2003). It has, therefore, been suggested that these Chilean land- feasibility studies and planning, medium-scale mapping is the
slides may be one cause for tsunamis emanating from this coastal most appropriate. At this scale landslide inventories can be com-
region (Mather et al. 2014). Furthermore, in this hyper-arid region piled that can form the basis for hazard and risk maps, although
it is hypothesized that intermittent and large-scale mass movement they also have a role in understanding aspects of the geomorpho-
failures of the coastal escarpment, in response to tectonic uplift logical development of the landscape (Case Study 2). Small-scale
and seismic activity, are the main processes of erosion of the mapping (Case Study 3) lends itself only to long-term landscape
coastal scarp and transfer sediment from the land surface into the development investigations to answer the ‘mega-landscape’ issues.
subduction trench (Fig. 10). In all the case studies the scale of mapping undertaken and the
 Mapping landslides 39

techniques that were employed were determined by the require- Spain. In: Lewin, J., Macklin, M.G. & Woodward, J.C. (eds) Mediterranean
ments of the investigation. It is important that any landslide map- Quaternary River Environments. Balkema, Rotterdam, 263–281.
Hearn, G.J. 2011a. Desk studies. In: Hearn, G.J. (ed.) Slope Engineering for
ping programme has clearly identified objectives from the outset if Mountain Roads. Geological Society, London, Engineering Geology Special
time, and money, is not to be wasted. However, even in site-spe- Publications, 24, 71–101.
cific investigations (Case Study 1) it is imperative for landslides to Hearn, G.J. 2011b. Field mapping. In: Hearn, G.J. (ed.) Slope Engineering for
Mountain Roads. Geological Society, London, Engineering Geology Special
be seen in the wider context of their geomorphological situation Publications, 24, 103–116.
and their role in landscape development to be assessed so as to Humbert, M. 1977. La cartographie en France des zones exposées à des risques
understand the likelihood and potential magnitude of any hazard. lies aux mouvements du sol—cartes ZERMOS. Bulletin of the International
Association of Engineering Geology, 16, 80–82.
Hutchinson, J.N., Bromhead, E.N. & Chandler, M.P. 1991. Investigations of
Acknowledgements and Funding landslides at St Catherine’s Point, Isle of Wight. In: Chandler, R.J. (ed.)
The authors wish to thank The GeoMapping Unit, Plymouth University for Slope Stability Engineering, Developments and Applications. Thomas
Figures 2, 3 and 4; our former PhD student A. Hart, now at Atkins, for his field Telford, London, 169–179.
mapping in SE Spain and for Figure 6; Google Earth for Figure 9; and L. Walsh Ibaraki, M. 2001. Neogene planktonic foraminifera of the Caleta Herrdura
for many years of support at the Urra field centre, SE Spain. J.S.G. would like de Mejillones section in northern Chile; biostratigraphy and palaeoceano-
to acknowledge the support of his wife, Sarah, as his field assistant during map- graphic implications. Micropalaeontology, 47, 257–267.
ping on the Isle of Wight. A.E.M. would like to acknowledge her co-worker in Keating, B.H. & McGuire, W.J. 2000. Island edifice failures and associated tsu-
Chile, A. Hartley from Aberdeen University, and support from Royal Society nami hazards. Pure and Applied Geophysics, 157, 899–955.
Research Grant 24407 for herself and A. Hartley to undertake fieldwork in Kelly, M., Black, S. & Rowan, J.S. 2000. A calcrete-based U/Th chronology for
Chile. J.S.G. acknowledges the support of the Hong Kong Regional Group of landform evolution in the Sorbas basin, southeast Spain. Pleistocene Science
the Geological Society for funding support when initially presenting this paper Reviews, 19, 995–1010.
in Hong Kong in February 2014. Kulikov, E.A., Rabinovich, A.B. & Thomson, R.E. 2005. Estimation of
tsunami risk for the coasts of Peru and northern Chile. Natural Hazards, 35,
Scientific editing by Nicholas Koor; David Entwisle 185–209.
Landry, J. 1980. Notice explicative de la Carte des Zones Exposée à des
Mouvements du Sol et du Sous-sol a 1:100,000. Région de Pontarlier (Jura).
References Bureau de Recherches Géologiques et Minières, Orléans.
BGS. 2013. Geological Survey of England and Wales 1:50,000 Scale Larouziére, F.D., Bolze, J., Bordet, P., Hernandez, J., Montenat, C. &
Geological Special Sheet for the Isle of Wight. British Geological Survey, Ottd’Estevou, P. 1988. The Betic segment of the lithospheric Trans-Alboran
Keyworth, Nottingham. http://www.largeimages.bgs.ac.uk/iip/mapsportal. shear zone during the late Miocene. Tectonophysics, 152, 41–52.
htm?id=1004974. Lawrance, C.J., Byard, R.J. & Bevan, P.J. 1993. Terrain evaluation manual.
Bjerrum, L. & Jorstad, F. 1968. Stability of rock slopes in Norway. Norwegian HMSO Transportation Research Laboratory, State of the Art Review, 7.
Geotechnical Institute Publication, 79. Lee, E.M. & Jones, D.K.C. 2014. Landslide Risk Assessment, 2nd edn. Thomas
Brand, E.W., Burnett, A.D. & Styles, K.A. 1982. The geotechnical area stud- Telford, London.
ies programme in Hong Kong. In: Proceedings of the 7th Southeast Asian Lee, E.M., Moore, R., Brunsden, D. & Siddle, H.J. 1991. The assessment of
Geotechnical Conference, 22–26 November 1982, Hong Kong, 107–123. ground behaviour at Ventnor, Isle of Wight. In: Chandler, R.J. (ed.) Slope
Burnett, A.D. & Styles, K.A. 1982. An approach to urban engineering geological Stability Engineering, Developments and Applications. Thomas Telford,
mapping used in Hong Kong. In: Proceedings of the Fourth Congress of the London, 207–212.
International Association for Engineering Geology, New Delhi, 1, 167–176. Masson, D.G., Watts, A.B., Gee, M.J.R., Urgeles, R., Mitchell, N.C., Le Bas,
Burnett, A.D., Brand, E.W. & Styles, K.A. 1985. Terrain classification map- T.P. & Canals, M. 2002. Slope failures on the flanks of the western Canary
ping for a landslide inventory in Hong Kong. In: Proceedings of the Islands. Earth-Science Reviews, 57, 1–35.
Fourth Congress of the International Association and Field Workshop on Mather, A.E. 1991. Cenozoic drainage evolution of the Sorbas Basin, SE Spain.
Landslides, Tokyo, 63–68. PhD thesis, University of Liverpool.
Carracedo, J.C., Day, S.J., Guillou, H. & PerezTorrado, F.J. 1999. Giant Mather, A.E. 2000. Adjustment of a drainage network to capture induced base-
Quaternary landslides in the evolution of La Palma and El Hierro. Journal of level change. Geomorphology, 34, 271–289.
Volcanology and Geothermal Research, 94, 169–190. Mather, A.E. & Hartley, A.J. 2006. The application of drainage system anal-
Chandler, M.P. 1984. The coastal landslides forming the Undercliff of the Isle of ysis in constraining spatial patterns of uplift in the Coastal Cordillera of
Wight. PhD thesis, Imperial College, University of London. Northern Chile. In: Willett, S.D., Hovius, N., Brandon, M.T. & Fisher, D.M.
Clague, J.J. & Roberts, N.J. 2012. Landslide hazard and risk. In: Clague, J.J. & (eds) Tectonics, Climate and Landscape Evolution. Geological Society of
Stead, D. (eds) Landslides: Types, Mechanisms and Modeling. Cambridge America, Special Papers, 398, 87–99.
University Press, Cambridge, 1–9. Mather, A.E. & Harvey, A.M. 1995. Controls on drainage evolution in the
CRED. 2012. EM-DAT: The OFDA/CRED International Disaster Database. Sorbas basin, southeast Spain. In: Lewin, J., Macklin, M.G. & Woodward, J.
Université Catholique de Louvain, Centre for Research on the Epidemiology (eds) Mediterranean Quaternary River Environments. Balkema, Rotterdam,
of Disasters. http://www.emdat.be/ (accessed 8 December 2013). 65–76.
Dearman, W.L. 1991. Engineering Geological Mapping. Butterworth– Mather, A.E., Martin, J.M., Harvey, A.M. & Braga, J.C. 2001. A Field Guide to
Heinemann, London. the Neogene Sedimentary Basins of the Almería Province, South-East Spain.
Dramis, F., Guida, D. & Cestari, A. 2011. Nature and aims of geomorphological Blackwell Science, Oxford.
mapping. In: Smith, M.J., Paron, P. & Griffiths, J.S. (eds) Geomorphological Mather, A.E., Stokes, M. & Griffiths, J.S. 2002. Quaternary landscape evolu-
Mapping: Methods and Applications. Developments in Earth Sciences, 15, tion: A framework for understanding contemporary erosion, Southeast Spain.
39–73. Land Degradation and Development, 13, 89–109.
Finlayson, A.A. 1984. Land surface evaluation for engineering practice: Mather, A.E., Griffiths, J.S. & Stokes, M. 2003. Anatomy of a fossil landslide
Applications of the Australian PUCE system for terrain analysis. Quarterly from the Pleistocene of SE Spain. Geomorphology, 50, 135–149.
Journal of Engineering Geology, 17, 149–158. Mather, A.E., Hartley, A.J. & Griffiths, J.S. 2014. The giant coastal landslides
Griffiths, J.S. 1999. Proving the occurrence and cause of a landslide in a legal of Northern Chile: Tectonic and climate interactions on a classic convergent
context. Bulletin of the International Association for Engineering Geology plate margin. Earth and Planetary Science Letters, 388, 249–256.
and the Environment, 58, 75–85. McMurtry, G.M., Watts, P., Fryer, G.J., Smith, J.R. & Imamura, F. 2004. Giant
Griffiths, J.S. (ed.) 2001. Land Surface Evaluation for Engineering Practice. landslides, mega-tsunamis, and paleo-sea levels in the Hawaiian Islands.
Geological Society, London, Engineering Geology Special Publications, 18. Marine Geology, 203, 219–233.
Griffiths, J.S. & Whitworth, M.Z. 2012. Engineering geomorphology of land- Mitchell, N.C., Masson, D.G., Watts, A.B., Gee, M.J.R. & Urgeles, R. 2002.
slides. In: Clague, J.J. & Stead, D. (eds) Landslides: Types, Mechanisms and The morphology of the submarine flanks of volcanic ocean islands—A
Modeling. Cambridge University Press, Cambridge, 172–186. comparative study of the Canary and Hawaiian hotspot islands. Journal of
Griffiths, J.S., Hart, A.B., Mather, A.E. & Stokes, M. 2005. Assessment of some Volcanology and Geothermal Research, 115, 83–107.
spatial and temporal issues in landslide initiation within the Rio Aguas catch- Moore, J.G. 1964. Giant submarine landslides in the Hawaiian ridge. US
ment, Southeast Spain. Landslides, 2, 183–192. Geological Survey Professional Paper, 501-D, 95–98.
Hack, J.T. 1957. Studies of longitudinal stream profiles in Virginia and Mortimer, C. & Saric, N. 1972. Landform evolution in the coastal region of
Maryland. US Geological Survey Professional Papers, 294B. Tarapacà Province, Chile. Revue de Géomorphologie Dynamique, 21, 162–170.
Hart, A.B. 2004. Landslide investigations in the Río Aguas catchment, southeast Parry, S. 2011. The application of geomorphological mapping in the assess-
Spain. PhD thesis, University of Plymouth. ment of landslide hazard in Hong Kong. In: Smith, M.J., Paron, P. &
Hartley, A.J., May, G., Chong, G., Turner, P., Kape, S.J. & Jolley, E.J. 2000. Griffiths, J.S. (eds) Geomorphological Mapping: Methods and Applications.
Development of a continental forearc: A Neogene example from the Central Developments in Earth Sciences, 15, 413–441.
Andes, northern Chile. Geology, 28, 331–334. Parry, S., Baynes, F.J., et al. 2014. Engineering geological models: An intro-
Harvey, A.M., Miller, S.Y. & Wells, S.G. 1995. Quaternary soil and river ter- duction: IAEG Commission 25. Bulletin of Engineering Geology and the
race sequences in the Aguas/Feos river systems: Sorbas basin, southeast Environment, 73, 689–706.
40 J. S. Griffiths et al.

Sanz de galdeano, C. 1990. Geologic evolution of the Betic Cordilleras in SE Spain. In: Jones, S.J. & Frostick, L.E. (eds) Sediment Flux to Basins:
the Western Mediterranean, Miocene to the present. Tectonophysics, 172, Causes, Controls and Consequences. Geological Society, London, Special
107–119. Publications, 191, 23–35.
Shilston, D.T., Teeuw, R.M. & West, G. & The Engineering Group Working Summerfield, M.A. 1991. Global Geomorphology. Longman, Harlow.
Party. 2012. Desk study, remote sensing, geographical information systems Taylor, G.K., Grocott, J., Dashwood, B., Gipson, M. & Arévalo, C.
and field evaluation. In: Walker, M.J. (ed.) Hot Desert: Engineering Geology 2007. Implications for crustal rotation and tectonic evolution in the
and Geomorphology. Geological Society, London, Engineering Geology central Andes fore arc: New paleomagnetic results from the Copiapó
Special Publications, 25, 159–200. region of northern Chile, 26°–28°S. Journal of Geophysical Research, 112,
Smith, M.J., Paron, P. & Griffiths, J.S. (eds) 2011. Geomorphological Mapping: B01102.
Methods and Applications. Developments in Earth Sciences, 15. Teeuw, R.M. (ed.) 2007. Mapping Hazardous Terrain using Remote Sensing.
Soeters, R. & van Westen, C.J. 1996. Slope instability recognition. In: Turner, Geological Society, London, Special Publications, 283.
A.K. & Schuster, R.L. (eds) Landslides Investigation and Mitigation. Von Huene, R., Weinrebe, W. & Heeren, F. 1999. Subduction erosion along the
Transportation Research Board, Special Report, 247, 129–177. north Chile margin. Journal of Geodynamics, 27, 345–358.
Spence, W., Mendoza, C., Engdahi, E.R., Choy, G.L. & Norabuena, E. 1999. Ward, S.N. 2001. Landslide tsunami. Journal of Geophysical Research—Solid
Seismic subduction of the Nazca Ridge as shown by the 1996–97 earth- Earth, 106, 11201–11215.
quakes. Pure and Applied Geophysics, 154, 753–776. Whelan, F. & Kelletat, D. 2003. Submarine slides on volcanic islands—a source
Stokes, M., Mather, A.E. & Harvey, A.M. 2002. Quantification of river cap- for mega-tsunamis in the Quaternary. Progress in Physical Geography, 27,
ture induced base-level changes and landscape development, Sorbas Basin, 198–216.