The temporal stability and activity of landslides in Europe with ž ...

Ž .
Geomorphology 30 1999 1–12
The temporal stability and activity of landslides in Europe with
ž /
respect to climatic change TESLEC : main objectives and results
Abstract
Richard Dikau ) , Lothar Schrott
Department of Geography, UniÕersity of Bonn, Meckenheimer Allee 166, D-53115 Bonn, Germany
Received 23 December 1998; received in revised form 24 March 1999; accepted 7 May 1999
The major aim of the European project ‘‘The temporal stability and activity of landslides in Europe with respect to
climatic change Ž TESLEC . ’’ was to investigate the interrelationship between landslides, climate and time. The research was
focused on three main objectives: Ž. 1 developing criteria for the recognition of landslides, Ž. 2 reconstructing past
distributions of landslide incidents and their relationship to climatic change parameters, and Ž. 3 developing a hydrological
and slope stability modelling framework using different test sites. The results of the project are related to these major
objectives and include: Ž. 1 a technical manual for landslide recognition, Ž. 2 records of landslide activity, and Ž. 3 an
evaluation of different hydrological and slope stability models. Landslide activity since 1950 has been generally high at all
test sites. In some areas, there has almost been a continuous activity observed since the beginning of the monitoring. The
records before 1950 are incomplete and probably indicate a lack of data rather than a lack of landslide activity. Whether the
observed active landslides are carrying a climate signal cannot be stated for all test sites with high confidence, since some
relationships between climate and landslides are uncertain. Thus, for the present, the complexity of the relationships between
climate and landsliding seems to make it not feasible to establish ‘‘universal laws’’ all over Europe. On the other hand, it
was possible to establish for some areas a cumulative rainfall-duration threshold for the reactivation of landslides. Future
scenarios of regional precipitation were derived from downscaled general circulation model Ž GCM. experiments and used
within simple slope hydrological and slope stability models. The evaluation of hydrological and slope stability models shows
that physically based models are not always the best solution due to the model complexity and data requirements. For
shallow landslides, more simple tank models are sometimes the better alternative. Future model development should
strengthen considerations of fissure flow, sudden changes in permeability, larger landslide volumes and complex landslide
topography. q 1999 Elsevier Science B.V. All rights reserved.
Keywords: landslides; climatic signal; GCM; hydrological and slope stability modelling
1. Introduction
The assessment of the temporal stability and activity
conditions of existing landslides is a difficult
) Corresponding author.
0169-555Xr99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved.
Ž .
PII: S0169-555X 99 00040-9
area of study in geomorphology, geology and
geotechnics. In many landslide classification systems,
little recognition is given to the present-day or
potential activity of landslides. These classifications
use terms, like ‘‘fossil’’, ‘‘dormant’’ or ‘‘active’’,
which are too vague to give a reliable picture of the
temporal stability and activity conditions of the slides.

2
Moreover, they do not give the quantitative characteristics
and physical background of the temporal
moving pattern in the past nor they reliably indicate
future patterns of activity. Methods are available to
distinguish between active and inactive landslides,
but a larger number of mutually supportive and
complementary methods are required to assess the
past, present and possible future patterns of behaviour.
A detailed diagnosis for activity requires
more detailed geological–stratigraphical data and
field observations and especially more elaborate
methods of modelling and analysis. Earth sciences,
hydrological sciences and geotechnics have methods
and models suitable to carry out these tasks.
One major task of the temporal stability and
activity of landslides in Europe with respect to climatic
change Ž TESLEC. project was to combine the
resources of these disciplines by focusing them on
the solution to this problem. A further task of the
TESLEC project was the identification of the temporal
stability condition of landslides which serves as a
necessary calibration to assess the future slide behaviour
under a changed climate. The results are
supposed to improve the predictability of future
landslide activities by means of a better understanding,
firstly, of the climatically caused fluctuations in
slope stability and, secondly, of the dominant parameters
which control the activity of landslides.
The purpose of this paper is to summarise the
objectives, methodologies and achievements of the
TESLEC project. It will identify the constraints and
opportunities that have arisen in the course of the
research. In addition, the value of the results will be
discussed and directions for further research will be
identified. Different landslide test sites in Europe
have been selected in order to achieve the aims of
the project. The main criteria for the selection of the
test sites were related to: Ž. 1 the existence of ongoing
research with landslide monitoring andror historical
data, Ž. 2 site specific hazard potential, Ž. 3
high recent landslide activity, and Ž. 4 the requirement
that different landslide types be included. Thus,
research was focused on site specific scales, which
means that the representativeness of each particular
site was not the major objective of site selection.
Depending on the test site, the temporal scale of the
research covered a wide range including historical
events as well as present activity.
( )
R. Dikau, L. SchrottrGeomorphology 30 1999 1–12
Key questions related to climatic signals in landslide
processes were discussed on the European Science
Foundation workshop ‘‘Rapid mass movement
and climatic variation during the Holocene’’ at the
Academy of Sciences, Mainz, Germany in 1993
Ž Matthews et al., 1997 . . The discussion was focused
on the climatic controls on rapid mass movements
during the Holocene. The general conclusions show
that despite many landslide data sets available, too
few high quality records are able to give a continuous
picture even of single sites. It has been further
shown that while the data records contain a climatic
signal, non-climatic factors Žgeology,
geomorphology,
human impact. significantly influence the climatic
signal. It was stated that more work has to be
done in correlating precipitation proxy data with
rapid mass movement records including rainfall
threshold models for regions with similar climate. A
further aspect concerning the character of the climatic
signal itself and its change in time has been
stressed by Thornes Ž 1987. and Crozier Ž 1986, 1997 . .
Thornes Ž 1987. discussed the problem of changing
behaviour of state variables in dynamical systems.
Applied to landslide response and rainfall triggers
the different types of behaviour that could help in
understanding the character of change as follows:
– damped behaviour, when landslide activity declines,
e.g., because of decreasing material
availability Ž Crozier and Preston, 1998 . ;
– gradual behaviour, when landslide activity increases,
e.g., because of a decrease of the factor
of safety in time Ž Popescu, 1996 . ;
– explosive behaviour, when a progressive increase
of landslide activity occurs Že.g.,
due to
deforestation, Crozier and Preston, 1998 . ;
– periodic behaviour; and
– unsystematic behaviour.
Crozier Ž 1997. discussed different climatic signal
types responsible for the triggering of landslides
including Ž. 1 frequency, Ž. 2 magnitude, and Ž. 3 duration
of rainfall and emphasised different changing
climatic conditions responsible for a change in these
rainfall attributes. This means that at least for periods
for which rainfall records are available Žapproxi-
mately 1850 to the present . , trend analysis of precipitation
data Ž Schonwiese ¨
and Birrong, 1990. is a

fundamental prerequisite to evaluate changes of
landslide activity for different regions in Europe.
This synoptic work is not completed and should be a
high priority research objective in future landslide
projects.
The temporal activity of landslides can be evaluated
with respect to the impacts of a changing climate
in Europe. Climatic change experiments represent
the scientific base for the design of plausible
scenarios of future precipitation. These experiments
comprise ‘‘transient’’ climatic change experiments
simulating the effect of a continuous increase of
greenhouse gas concentration. Models of climatic
change must be treated with caution. There are,
however, sufficient indications that the climatic impact
will increase the intra-annual variability of the
climate. Climate research indicates an increase in
precipitation in northern and central Europe in winter
Ž Schonwiese ¨ and Birrong, 1990 . . Therefore, the study
of the effect of a possible intra-annual shifting of
climatic states on landslide activities is a current
research task.
2. Research objectives and project framework
The TESLEC project was established to investigate
the interrelationship between landslide, climate
( )
R. Dikau, L. SchrottrGeomorphology 30 1999 1–12 3
and time. The research interest was focused on three
Ž . Ž .
main objectives Dikau et al., 1996a, b see Fig. 1 .
2.1. ObjectiÕe 1: deÕelopment of criteria for the
recognition of landslides
This includes the publication of a technical manual.
2.2. ObjectiÕe 2: the reconstruction of past distributions
of landslide incidence related to the change of
Õarious climate parameters
Objective 2 assumes that it is possible to identify
occurrences of landslides and landslide remnants that
are not related to human activity. This can only be
stated if existing landslides were reactivated due to
natural causes such as an increase in precipitation,
earthquakes or long-term climatic change. Techniques
for the detection of mass movements are well
known and form part of existing good practice. The
prediction of the reactivating of landslides, however,
is a difficult task to tackle and requires a thorough
study of past activity using a complete range of
investigation methods in order to recognise the
causative factors especially those related to climatic
Fig. 1. Major objectives, methods and deliverable products of the TESLEC project.

4
change. The basic research questions of objective 2
are:
– Does the dated landslide event carry a climate
signal?
– Are there other signals carried by the event?
– Are there any regional coincidences in dated landslide
events in Europe?
– Are there palaeoclimatic reconstructions in the
region under investigation?
– Are we able to relate the dated event with general
circulation models Ž GCM. predictions?
– Is there any information in terms of the lifetime
of landslides?
( )
R. Dikau, L. SchrottrGeomorphology 30 1999 1–12
– Are we able to draw first conclusions to continue
the European programme on climate and natural
hazards Ž EPOCH. synthesis Ž see Table 1 . ?
For the second objective, the methodology included
geomorphological mapping, a variety of suitable
dating techniques, especially lichenometrical and
radiometrical methods and standard chronostratigraphic
techniques in order to establish time series of
landslide incidents. A further, special technique was
the use of trees as a tool to study the history of slope
activity Ž dendrochronology . . During movement of a
slide or parts of the slide, the tree is bent. The
redressing of the trees after the incident can be
Table 1
The importance of different objectives of the TESLEC project in relation to the previous Ž EPOCH. and subsequent project Ž NEWTECH.

detected by asymmetric ring growth. Using carefully
applied statistical filtering techniques, the rate of
strain can be dated. A motion activity time series
related to different parts of the landslide complex
can be obtained ŽCorominas
and Moya, 1999; Fantucci
and Sorriso-Valvo, 1999 . .
2.3. ObjectiÕe 3: the deÕelopment of a hydrological
and slope stability modelling framework for the prediction
of landslide actiÕity in a changing climate
Landslides are often considered as rigid blocks
with a constant safety factor along the slope. However,
this approach gives only a general view of the
stability of the slope. For a better understanding of
complex landslides, the spatial distribution of stability
within the landslide complex must be known.
Objective 3 linked climatic change model outputs to
hillslope hydrology and groundwater models used
within slope stability analysis ŽBuma
and Dehn,
1998; Bonomi and Cavallin, 1999; Dehn and Buma,
1999; Angeli et al., 1999 . . The location of the
‘‘weakest parts of the chain’’ gives an idea of future
distribution on unstable and effectively stable parts
of the landslide. In some cases, the hydrological part
in the stability model forms, in fact, the trigger
determining the activity of the slope. The water
balance must be analysed by both simple hydrological
grey box models or tank models and the new
models of climatic change. Calibration of the hydrological
part of the stability model was done with
existing piezometric data or back analysis. An important
tool can be the use of probability stability
models. These models calculate the transition probabilities
of the different sections within the landslide.
Other models, showing the internal stability distribution
of landslides, have been developed. These
probability models show the distribution of the safety
factor Ž stability coefficient along the slip surface.
and may give a direct indication of internal stability.
Once the stability model procedures have been
made by means of operation and calibration, they
will be manipulated to simulate the effect of climatic
change. The most important task was to recognise
the relationship between changing effective precipitation
and groundwater level or piezometric conditions.
A search was be made for groundwater stations
where this relationship could be developed.
This modelling is important because it proves or
( )
R. Dikau, L. SchrottrGeomorphology 30 1999 1–12 5
refuses the use of our existing knowledge of global
climatic change in order to predict the spatial and
temporal occurrence of landsliding. A further research
task was to evaluate various climatic change
models with respect to the changing probability of
landsliding. Today, many global models of climatic
change exist. With regard to questions about environmental
changes, one of the most important research
tasks was to find the key to how these models
can be regionalised to smaller areas or how the
generalities can be applied to the specific problems
in landslide areas. Because the confidence in the
regional changes simulated by GCMs is low and
simple interpolation of the coarse grid output is
inadequate, downscaling techniques are applied.
These objectives are in close connection with two
other EC-funded research projects related with landsliderclimate
relations Ž Table 1 . . The TESLEC project
shows a shift in research focus. While the previously
project running from 1991–1993 within the
EPOCH Ž Soldati, 1996. was mainly concerned with
the development and definition of time classification
standards and the reconstruction of past landslide
events, the TESLEC project was additionally focused
on modelling the current and future climatic impact
on single landslides. The continuation of the research
tasks in the NEWTECH project was related to a
stronger emphasis on single landslide monitoring and
modelling and future climate impact scenarios
Ž .
Corominas et al., 1998 .
3. Study areas
The study areas of the TESLEC project were
selected in terms of having a variety of landslide
types showing different sensitivities to climatic impacts.
The areas are located in five European countries
Ž Fig. 2. including four different landslide types
Ž fall, slide, flow, complex . .
The different time scales of investigation are
shown in Fig. 3. Most of the studies are related to
landslide events during the last 50 years. The study
in southern Italy goes back to 1850 using dendrogeomorphological
analysis ŽFantucci
and Sorriso-Valvo,
1999 . , and for two areas in northern Spain ŽCantabria
and Asturias. a first attempt has been made towards
chronologies of landslide activities during the last
120,000 years Ž Gonzales-Dıez ´ ´
et al., 1999. and mid

6
( )
R. Dikau, L. SchrottrGeomorphology 30 1999 1–12
Fig. 2. European study areas of the project. Each study site refers to one or more papers of this issue: Ž. 1 Cortina d’Ampezzo ŽBonomi
and
Cavallin, 1999; Dehn and Buma, 1999; Pasuto and Soldati, 1999; Angeli et al., 1999 . ; Ž 2. Lago ŽFantucci
and Sorriso-Valvo, 1999;
Sorriso-Valvo et al., 1999 . ; Ž 3. Barcelonnette and Vars basin Ž Dehn and Buma, 1999; Flageollet et al., 1999; van Asch et al., 1999 . ; Ž 4.
Llobregat basin Ž Moya and Corominas, 1999 . ; Ž. 5 Cantabrian Range Ž Gonzalez-Dıez ´ ´ et al., 1999 . ; Ž. 6 Asturian coastal valley ŽJimenez
´
Sanchez ´ et al., 1999 . ; Ž 7. North of Lisbon Ž Zezere ˆ et al., 1999 . ; Ž 8. Roughs Ž Brunsden 1999 . .
Holocene Ž Jimenez ´ Sanchez ´ et al., 1999 . , respectively.
Special attention with respect to monitoring
and modelling has been drawn to the Alvera ` mudslide
in Cortina d’Ampezzo Ž Dolomites, Italy . ,
mainly because of the availability of several time
series of data with a relatively good resolution
Ž meteorological, hydrological, geotechnical . . This
particular mudslide is considered under different top-
ics Že.g.,
hydrological modelling and slope stability
analysis. in several papers of this issue ŽBonomi
and
Cavallin, 1999; Dehn and Buma, 1999; Pasuto and
Soldati, 1999; Angeli et al.,1999 . . The relationship
between climate and landslide activity played a major
role in the Vars and Barcelonnette basin, the
Llobregat basin, the Cantabrian range, the Asturian
coastal valley, and in the Lisbon area ŽCorominas

( )
R. Dikau, L. SchrottrGeomorphology 30 1999 1–12 7
Fig. 3. Landslide events and periods of activity at the different study sites of the TESLEC project. Note: The Alvera ` mudslide ŽDolomites,
Northern Italy. was not only active throughout the defined time period. This area is probably affected by landslides since the early Holocene
and several phases of major activity are proposed Ž see Panizza et al., 1996 . . The record of landsliding in the Cantabrian range and in the
Asturian coastal valley are not included because they cover much longer time periods Ž HolocenerPleistocene. and display generally
landslide activity in a different resolution Ž see Gonzalez-Dıez ´ ´ et al., 1999; Jimenez ´ Sanchez ´ et al., 1999 . .
and Moya, 1999; Cuesta et al., 1999; Flageollet et
al., 1999; Gonzales-Dıez ´ ´ et al., 1999; Jimenez ´
Sanchez ´ et al., 1999; Zezere ˆ et al., 1999 . .
4. Results
The TESLEC project delivered several products
including:
– a technical manual for landslide recognition;
– a slide collection on CD-ROM showing different
types of mass movements and one of the
most prominent landslides in Europe ŽCD-ROM
enclosed . ;
– an evaluation of new techniques and dating
methods;
– past distribution of landslide occurrences;
– evaluations and applications of new hydrological
and slope stability models;
– general evaluations of climatic impacts on landslide
processes including scenarios from downscaled
GCM experiments.
4.1. Landslide recognition
The preparation and editing of the manual
‘‘Landslide Recognition’’ was the first deliverable of
the project which presents the main characteristics of
different mass movement types in Europe. The book
assists in the education of landslide recognition with
the aim of helping the reader to distinguish different
mass movement types in the field. The manual emphasises
the description of diagnostic features of
each landslide category for potential and relict circumstances
and also summarises different classification
systems. The book is a considerable international
achievement and a contribution to a European
landslide standard Ž see Dikau et al., 1996c . .
4.2. New techniques and dating methods
One prerequisite to reconstruct landslide distributions
in time is the direct or indirect dating of
material exposed or removed by a landslide process.
In a previous European project, EPOCH ŽThe
temporal
occurrence and forecasting of landslides in the

8
European Comunity, 1991–1993. ŽCasale
et al.,
1994; Soldati, 1996. classical dating methods like
14
C, dendrochronology or lichenometry were described
and applied. In the TESLEC project, the new
techniques, like Accelerator Mass Spectrometry
Ž AMS . , Thermally Ionised Mass Spectrometry
Ž TIMS . , laser fusion, and methods, like Optically
Stimulated Luminescence Ž OSL . , alpha recoil track
dating, were discussed in their potential to date
landslides Ž Lang et al., 1999 . . Especially, surface
exposure dating of in situ produced cosmogenic nuclides
and OSL-dating are recommended as new
methods for dating past landslide events. The application
of cosmogenic isotopes in geomorphology
was also discussed in a special session of the Association
of American Geographers in April 1997 and
the outcomes are published recently Ž Harbour, 1999 . .
However, the project also states that most of the new
techniques and methods described are in an early
stage of development and are currently not used
widely in landslide research.
4.3. Past distribution of landslide incidents
The second objective of the TESLEC project was
the investigation of past distributions of landslide
incidence and their relationship to climatic change
parameters. This included the understanding of the
nature of climatically controlled landslide distribution
in space and time and the behaviour of individual
landslides. If the past landslide event has been
dated, there is, however, no complete certainty as to
whether this landslide has been triggered by the
climatic conditions present at the time of failure or
by other causes. Therefore, the project delivered a
critical discussion of the frequency and activity of
Holocene landslides in western European countries
in terms of the relationship between landslide events
and climate. The establishment of a Holocene slope
instability history is difficult because of the low
number of dated landslides available. Landslide series,
therefore, should be taken with caution as they
may be only partly representative of the climatic
conditions of the studied region. Caused by the lack
of data and by the fact that often a landslide can be
only indirectly dated as, for example, fluvial terraces
developed on the landslide mass or peat bogs located
( )
R. Dikau, L. SchrottrGeomorphology 30 1999 1–12
at the landslide surface. Despite these uncertainties,
the new results of the TESLEC project show that it
appears that climatically generated landslide activities
have not been homogenously distributed all over
Europe. Only cool phases, such as the Younger
Dryas and the Little Ice Age, are characterised by an
increase in landslide activity in most European countries.
This is consistent with what was found by other
researchers in European mountain areas. However,
wet and warm periods seem not to affect Europe
uniformly. During the early Atlantic and Sub-Atlantic
Ž wet periods. in Northern and Central Europe landslide
reactivation is very significant. An increase in
landslide activity in the late Atlantic and Sub-Atlantic
occurred in the Mediterranean region but no reactivations
in Northern Europe were observed. Based on
the problems described, it is clear that the increase of
landslide activity is not uniform across Europe and
can be explained by local factors such as the increase
in mean annual precipitation Ž Eastern Pyrenees . , an
increase in the mean annual temperature that causes
the melt of permafrost ice Ž Alps . , the effect of sea
level rise and storms responsible for coastal erosion
and retreat Ž Southern Britain . .
Landslide activity since 1950 has been generally
high at all test sites. In some areas, there has almost
been a continuous activity observed since the beginning
of the measurements. However, this information
should be interpreted with caution, because the time
series before 1950 are incomplete and more characterised
by a lack of information and data rather than
a lack of landslide activity. Therefore, Fig. 3 should
be mainly seen as a documentation of available
landslide information through time and should not be
used in order to distinguish periods of landslide
activity through time over Europe.
A further task of the project was distinguishing
the climatic and dynamic influences from dated landslides.
The central question to be asked was whether
or not a dated landslide carries a climatic signal.
There may be many possible interpretations of each
landslide occurrence not all of which involve climate.
It is important to consider that the time of
climate influence may range from the nature of a
climate experienced by the landslide, e.g., the effective
precipitation available in a humid tropical or a
cold periglacial climate, to an alteration of the
weathering, erosion, deposition, rainfall or water

egimes by climate variability. Therefore, it is usually
not possible to assess accurately the climatic
cause of an individual slide unless there is independant
colaborative information, e.g., a direct observation
of a wet winter or a high magnitude storm.
4.4. Rainfall triggering thresholds
The causal factors of landslides include climate,
geomorphological processes, and ground conditions
which are considered as preparatory and triggering
effects. Special emphasis lies on the temporal change
of intrinsic characteristics of the slope by weathering
or hydration. Rainfall plays the essential role in
triggering landslides by changing the ground conditions
of a slope. At the present level of knowledge,
no agreement has yet been reached about the identification
of pluviometric thresholds above which
landslides are triggered. Despite clear uncertainties
on the relationships of the process-trigger-ground
conditions, it can be stated that beside the mean
annual rainfall, the cumulative effective rainfall over
weeks, months and years as well as the antecedent
maximum rainfalls, especially 1 to 3 days, should be
considered in more detail for prediction modelling of
landslides. In this context, it has to be taken into
account that the useful time span of antecedent rainfall
conditions for analysing landslide triggers depends
on the size Ž depth . , type and geological–geomorphological
characteristics of the landslide itself
Ž see Van Asch et al., 1999 . .
A further task of the TESLEC project was related
to the differences between a first time trigger of a
landslide and a landslide reactivation. It became
evident that there is a high complexity of combinations
of long-term and short-term climatic factors
which trigger a first time landslide or reactivate a
landslide. There are situations where no clear climatic
signal can be found because events may occur
after dry months preceded or not by abundant annual
rainfall. The complexity of the relationships between
climate and landsliding seems to make it difficult to
establish ‘‘universal laws’’ for Europe.
It seems possible to establish rainfall trigger indices
for some European regions including the landslide
type, the status of reactivation, the seasonality
and the initial degree of the slope stability, however,
this implies the use of high resolution data, which
( )
R. Dikau, L. SchrottrGeomorphology 30 1999 1–12 9
are only available for a few regions. Nevertheless, in
some cases, it was possible to reconstruct even unnoticed
reactivations of landslides by means of dendrogeomorphological
analysis ŽCorominas
and Moya,
1999 . . This methodology allows the reconstruction
of phases of landslide activities with an accuracy of
one year and should applied to other sites in which
alternative sources of landslide records are incomplete
or not available.
4.5. Climatic change scenarios and hydrological and
slope stability modelling
The combination of climatic change modelling by
GCMs and slope hydrologyrstability models for an
assessment of climatic change impacts on landslide
activity has shown that GCM results have to be
downscaled by appropriate techniques in order to
derive local scale climatic scenarios. With empirical–statistical
downscaling techniques, it became
possible to derive from GCMs’ climatic change impact
scenarios for modelling landslide activity on a
small scale. A first critical evaluation of different
climatic change impact scenarios has been examined.
The evaluation of downscaling techniques shows that
empirical–statistical approaches are more appropriate
than nested dynamical climate models. These
techniques include a linear regression model with a
canonical correlation analysis, an analog approach
and an analog approach with rainfall generator which
is suitable in landslide research. In the TESLEC
project, an empirical–statistical downscaling technique
was applied for two study sites ŽCortina
d’AmpezzorDolomites, Italy and Barcelonnette,
French Alps. using monthly winter precipitation Žsee
Fig. 2 . .
While for the Alvera` landslide in Cortina
d’Ampezzo, the statistical relationship is too weak
for a predictive task, the case study of the Riou
Bourdoux landslide in Barcelonnette revealed many
problems concerning the performance of the different
seasons and decreasing amplitudes due to the
downscaling procedure. This shows that the links
between climate modelling and impact modelling are
still in a developing stage. One of the results is that
hydrological and slope stability models using downscaled
GCM data produce only probabilistic state-

10
ments. As a first effort, the presented approach for
coupled modelling of future landslide activity is
promising because it shows that a quantification of
climatic change impact is feasible ŽDehn
and Buma,
1999 . . However, the case study carried out shows
useful results only for monthly time steps. Although
most slope hydrologyrstability models require daily
data resolution, the deeper clayey landslides show a
good prediction with monthly precipitation data Že.g.,
Riou Bourdoux, French Alps . . The analysis of rainfall–landslide
relationships, which are based on historical
landslide data, shows clear correlations between
antecedent precipitation and landslide events
if enough landslide and rainfall data are available. In
numerous European regions, however, historical
records are difficult to obtain and are too incomplete
to carry out detailed statistical analysis. A further
problem is that different landslide types show different
movement patterns under the same climatic condition.
Therefore, in order to reconstruct in a reliable
way changes in landslide frequencies as a response
to changes in precipitation patterns, more detailed
investigations pertaining to one region are needed.
These investigations must be focused on soil mechanical
and hydrological factors of different landslide
types. Combined hydrologicalrequilibrium
models can be assessed for different landslide types
which can deliver critical precipitation thresholds.
From these results, two research concepts can be
drawn. Firstly, these models can be used to obtain
climatic signals from landslide frequencies in the
past and climatic scenarios can be constructed which
explain the change in landslide frequencies in the
past. Secondly, deterministic hydrological and slope
stability models can be used to evaluate the stability
of landslide test sites and to assess future climatic
change impacts which was the third objective of the
TESLEC project. To achieve this, the project had to
evaluate the available hydrological and slope stability
models and climatic change scenarios based on
GCMs. With empirical–statistical downscaling techniques,
it became possible to derive from GCMs’
climatic change impact scenarios for modelling of
landslide activity on a small scale. A first critical
evaluation of different climatic change impact scenarios
has been examined. Taking into consideration
several sources of uncertainties such as GCM quality
or shortcoming of downscaling techniques, it is fea-
( )
R. Dikau, L. SchrottrGeomorphology 30 1999 1–12
sible to simulate future landslide activitity ŽDehn
and
Buma, 1999.
The hydrological models evaluated included
GWFLUCT, HILLFLOW, HYWASOR, MOD-
FLOW, SEEPrW, SWMS-2D and TANK ŽIbsen
and Collison, 1996 . . The evaluation of the slope
stability models included static and dynamic models.
The static models imply a close relationship between
the onset of the landslide movement and rising
groundwater level. The dynamic models predict different
landslide behaviours as groundwater levels
continue to rise after the initial reactivation.
Based on these different models, the modelling
framework of the TESLEC project has been carried
out. There are various ways of modelling slope
behaviour according to the different types of landslide
events described in objective one. The project
decided to use a simple form of landslide events, the
translational slide with a planar shear surface for
analysis with the available models. For this the test
sites Alvera ` in Cortina d’Ampezzo ŽDolomites,
Italian
Alps . , Riou Bourdoux in the Barcelonnette basin
Ž French Alps. and the Roughs landslide complex
Ž Southern Britain. were selected. These test sites had
relevant data which were available for the TESLEC
project to calibrate and validate the modelling framework.
The sites were chosen to represent different
climates, locations and landslide types across Europe.
Additionally, the Lago Sackung in southern
Italy was discussed in terms of the serious problems
to model complex and deep seated landslide types
Ž Sorriso-Valvo et al., 1999. Ž see Fig. 2 . .
From the hydrological and slope stability modelling,
the following conclusions can be drawn:
Physically based models are operative and are clearly
capable of producing useful results. However, these
models had been developed for simple shallow landslides
where the data requirements were quite small.
For large landslides with multiple layers or permeability
differences, or for locations with multiple
hydrological regimes, it is difficult to replicate the
observed field data. These results could be simulated
by reducing the complexity of the model from two to
one dimension. Therefore, simple conceptual models
currently seem to be the best modelling alternative.
Fissure flow and vegetational influences have been
neglected in many studies, although they may have
an important influence in the landslide process. Fu-

ture studies should draw more attention to these
aspects.
5. Conclusions and future needs
Our knowledge is still too incomplete to draw a
complete picture of landslide activity in Europe during
the Holocene. Progress in analytical instruments
and new dating methods, however, allows a broader
appreciation of Late Quaternary mass movements
and may lead to considerable improvements of landslide
activity records of different regions. For some
areas, it is possible to establish rainfall thresholds.
But, generally, this remains limited to a local scale
and cannot be upscaled to larger areas. Due to the
heterogenity of the available data sets and due to a
considerable lack of information for the early
Holocene, it is still not feasible to establish ‘‘universal
laws’’ for the landslide activity in Europe. The
TESLEC project and related activities showed that
there is an enormous amount of widely distributed
landslide data in Europe. Future work should concentrate
on the collection and continuation of accurately
dated events and on the establishment of a European
landslide database.
Currently available general hydrological models
have two disadvantages in modelling the activity of
landslides: they require data in a spatial resolution
that often cannot be provided, and they fail to cope
with landslide specific processes like fissure flow
Ž Van Beek and Van Asch, 1998 . . Future models
must be able to consider not only these effects but
also incorporate sudden changes in permeability,
complex topography and large landslide volumes.
In relation to the three-dimensional nature of the
landslide phenomenon, it is clear that insufficient
attention has been given to the meaning of retrogressive
processes and to the interaction effect of landslide
units locked together in a complex area. In the
future, as three-dimensional slope stability models
are constructed, it will be necessary to take the
following conclusions into consideration Žsee
also
Brunsden, 1999 . :
– the three-dimensional shape of the natural topography,
the structural ground water and the
shear surface control, the application of
stress, the mobilisation of resistance and the
( )
R. Dikau, L. SchrottrGeomorphology 30 1999 1–12 11
availability of three-dimensional weakness
patterns;
– the internal structural behaviour of the system in
which complex self-loading and unloading
effects occur such as undrained loading, etc.;
and
– the existing pattern of shear surfaces and landslide
debris which will control the way a slope
will unravel when the landslide is reactivated.
It is therefore very important to determine whether
the three-dimensional pattern and analysis is for a
single first-time slide, first-time retrogression at the
head of an existing slide or reactivation of a whole
complex. Based on these conclusions, it is felt that
the TESLEC project has provided a clear research
direction for the future.
Acknowledgements
The TESLEC project was possible through the
contributions and intensive cooperations between
different European teams. This cooperation and the
stimulating and open discussions with numerous colleagues
during the workshops and fieldtrips throughout
Europe is gratefully appreciated. The participants
of all the groups are too numerous to name separately
but they are thanked for their advice and skill.
References
Angeli, M.-J., Pasuto, A., Silvano, S., 1999. Towards the definition
of the slope instability behaviour in the Alvera ` mudslide
Ž Cortina d’Ampezzo, Italy . . Geomorphology 30, 201–211.
Bonomi, T., Cavallin, A., 1999. Three dimensional hydrogeological
modelling application to the Alvera ` mudslide ŽCortina
d’Ampezzo Italy . . Geomorphology 30, 189–199.
Brunsden, D., 1999. Some geomorphological considerations for
the future development of landslide models. Geomorphology
30, 13–24.
Buma, J., Dehn, M., 1998. A method for predicting the impact of
climate change on slope stability. Environmental Geology 35
Ž 2–3 . , 190–196.
Casale, R., Fantechi, R., Flageollet, J.-C. Ž Eds. . , 1994. The temporal
occurrence and forecasting of landslides in the European
Community — final report. European Community, Programme
EPOCH Ž Contract 90 0025 . , 957 pp.
Corominas, J., Moya, J., 1999. Reconstructing recent landslide
activity in relation to rainfall in the Llobregat River basin.
Eastern Pyrenees, Spain. Geomorphology 30, 79–93.

12
Corominas, J., Moya, J., Ledesma, A., Gili, J.A., Lloret, A., Rius,
J., 1998. New technologies for landslide hazard assessment
and management in Europe — final report. CEC Environment
Programme Ž Contract ENV-CT966-0248 . , 431 pp.
Crozier, M.J., 1986. Landslides. Causes, Consequences and Environment.
Croom Helm, London, 252 pp.
Crozier, M.J., 1997. The climatic–landslide couple: a Southern
Hemisphere perspective. In: Matthews, J.A., Brunsden, D.,
Frenzel, B., Glaser, ¨ B., Weiß, M.M. Ž Eds. . , Rapid mass
movement as a source of climatic evidence for the Holocene.
Palaeoclimate Research 19, pp. 333–354.
Crozier, M.J., Preston, N.J., 1998. Modelling changes in terrain
resistance as a component of landform evolution in unstable
hill country. In: Hergarten, S., Neugebauer, H.J. Ž Eds. . , Process
modelling and landform evolution. Lecture Notes in Earth
Sciences 78, pp. 267–284.
Cuesta, M.J., Sanchez, M.J., Garcia, A.R., 1999. Press archives as
temporal records of landslides in the North of Spain: relationships
between rainfall and instability slope events. Geomorphology
30, 125–132.
Dehn, M., Buma, J., 1999. Modelling future landslide activity
based on general circulation models. Geomorphology 30,
175–187.
Dikau, R., Schrott, L., Dehn, M., Hennrich, K., Ibsen, M.-L.,
Rasemann, S., Ž Eds. . , 1996a. The temporal stability and activity
of landslides in Europe with respect to climatic change
Ž TESLEC . . Final Report, Part I, Summary Report. European
Community CEC Environment Program, Contract No. EV5V-
CT940454, Brussels, 208 pp.
Dikau, R., Schrott, L., Dehn, M., Hennrich, K., Rasemann, S.,
1996b. The temporal stability and activity of landslides in
Europe with respect to climatic change Ž TESLEC . . Final
Report, Part II, National Reports. European Community CEC
Environment Program. Contract No. EV5V-CT940454, Brussels,
408 pp.
Dikau, R., Brunsden, D., Schrott, L., Ibsen, M.L. Ž Eds. . , 1996c.
Landslide Recognition. Identification, Movement and Causes.
Wiley, Chichester, 251 pp.
Fantucci, R., Sorriso-Valvo, M., 1999. Dendrogeomorphological
analysis of a slope near lago. Calabria, Italy. Geomorphology
30, 165–174.
Flageollet, J.-C., Maquaire, O., Martin, P., Weber, D., 1999.
Landslides and climatic conditions in the Barcelonnette and
Vars Basins Ž Southern Alps; France . . Geomorphology 30,
65–78.
Gonzales-Dıez, ´ ´ A., Remondo, J., Diaz de Teran, J.-R., Cendrero,
A., 1999. A methodological approach for the analysis of the
temporal occurrence and triggering factors of landslides. Geomorphology
30, 95–113.
Harbour, J., 1999. Cosmogenic isotopes in geomorphology. Geomorphology
27 Ž 1–2 . .
( )
R. Dikau, L. SchrottrGeomorphology 30 1999 1–12
Ibsen, M.L., Collison, A., 1996. Hydrological hillslope models.
In: R. Dikau, L. Schrott, M. Dehn, K. Hennrich and S.
Rasemann Ž Eds. . , The temporal stability and activity of landslides
in Europe with respect to climatic change Ž TESLEC . .
Final report, Part II, National Reports. European Community.
CEC. Environment Programme, Brussels. EV5V-CT94-0454,
pp. 77–86.
Jimenez ´ Sanchez, ´ M., Farias, P., Rodrıguez, ´ A., Menendez ´ Duarte,
R.A., 1999. Landslide development in a coastal valley in
northern Spain: conditioning factors and temporal occurrence.
Geomorphology 30, 115–123.
Lang, A., Moya, J., Corominas, J., Schrott, L., Dikau, R., 1999.
Classic and new dating techniques for assessing the temporal
dynamic of mass movements. Geomorphology 30, 33–52.
Matthews, A.J., Brunsden, D., Frenzel, B., Glaser, ¨ B., Weiß,
M.M., 1997. Rapid mass movement as a source of climatic
evidence for the Holocene. Palaeoclimate Research 19,
Stuttgart, 444 pp.
Panizza, M., Pasuto, A., Silvano, S., Soldati, M., 1996. Temporal
occurrence and activity of landslides in the area of Cortina
d’Ampezzo Ž Dolomites Italy . . Geomorphology 15, 311–326.
Pasuto, A., Soldati, M., 1999. The use of landslide units in
geomorphological mapping. Geomorphology 30, 53–64.
Popescu, M.E., 1996. From landslide causes to landslide remediation.
In: Senneset, K., Ž Ed. . , Landslides. Proceedings of the
7th International Symposium of Landslides, Trondheim,
Balkema, Rotterdam, pp. 75–96.
Schonwiese, ¨ C.-D., Birrong, W., 1990. European precipitation
trend statistics 1851–1980 including multivariate assessments
of the anthropogenic CO2 signal. Z. Meteorol. 40, 92–98.
Soldati, M. Ž Ed. . , 1996. Landslides in the European Union. Geomorphology
15 Ž 3–4 . .
Sorriso-Valvo, M., Gulla, G., Antronico, L., Tansi, C., Amelio,
M., 1999. Mass movement, geologic structure and morphologic
evolution of the Pizotto-Greci slope Ž Calabria, Italy . .
Geomorphology 30, 147–163.
Thornes, J., 1987. Environmental systems — patterns, processes
and evolution. In: Clark, M.J., Gregory, K.J., Gurnell, A.M.
Ž Eds. . , Horizons in Physical Geography, Totowa. pp. 27–46.
Van Asch, Th.W.J., Buma, J.T., Van Beek, L.P.H., 1999. A
review of some hydrological triggering systems in landslides.
Geomorphology 30, 25–32.
Van Beek, L.P.H., Van Asch, Th.W.J., 1998. A combined conceptual
model for the effects of fissure-induced infiltration on
slope stability. In: Hergarten, S., Neugebauer, H.J. Ž Eds. . ,
Process Modelling and Landform Evolution. Lecture Notes in
Earth Sciences 78, pp. 147–167.
Zezere, ˆ J.L., de Brum Ferreira, A., Rodrigues, M.L., 1999. The
role of conditioning and triggering factors in the occurrence of
landslides: a case study in the area north of Lisbon Ž Portugal . .
Geomorphology 30, 133–146.