Réf. Flageollet 1994 - R: EPOCH

Référence bibliographique complète
FLAGEOLLET J.C. Introduction & Final remarks. Temporal occurrence and forecasting of landslides in the European Community (EPOCH Programme) - Final Report. / ed. by. Casale, Fantechi and Flageollet. European Commission, Bruxelles, 1994, 1, p. 1-5 & 935-940.

Abstract: This research programme has been set up, to investigate the possibility of establishing the temporal patterns of landsliding in European regions, with four main objectives:
to describe the frequency and magnitude of events and of different types of landslide;
to describe the current statistical distribution of European landslides in the context of historical records and, as far as possible, in the context of the time scales of global environmental change;
to assess the external causal factors of landsliding (human, climatic and tectonic), their temporal properties and the degree to which they explain the temporal variations in landslide activity. As human impact through changes of landuse, tourist activity and industrial development will very probably increase, we must clarify the raie of this human factor in the complex combination of factors triggering landslides;
to use the empirical data to establish predictive models of landslide occurrence.

Landslides, temporal occurrence, European Community.

Organismes / Contact
  European Commission - Science Reserch Development

Principaux rapports scientifiques sur lesquels s'est appuyé le rapport
This synthesis is based on many studies done in the framework of the programme EPOCH.

(1) - Paramètre(s) atmosphérique(s) modifié(s)
(2) - Elément(s) du milieu impacté(s)
(3) - Type(s) d'aléa impacté(s)
(3) - Sous-type(s) d'aléa
Precipitation   Mass movements Landslides

Pays / Zone
Massif / Secteur
Site(s) d'étude
Période(s) d'observation
Great Britain, Germany, France, Italy and Spain Kent
Normandy, Baronnies, Drac and Durance basins, Vars Commune, Trieves, Barcelonnette basin, Bachelard Valley
Dolomites, Marche region
Santander, Nalon, Eastern Pyrenees
      Holocene - Historic

(1) - Modifications des paramètres atmosphériques

Informations complémentaires (données utilisées, méthode, scénarios, etc.)

(2) - Effets du changement climatique sur le milieu naturel

Sensibilité du milieu à des paramètres climatiques
Informations complémentaires (données utilisées, méthode, scénarios, etc.)


(3) - Effets du changement climatique sur l'aléa
In England, cold periglacial environments are presumed to be periods of landslide activity or accelerated landslide activity. This may also be the case in Cantabria, but we think it is during a deglaciation period, before the disappearance of the glaciers, around 11,000 BP. For the French Alps, the opinion expressed here is that the period favourable to mass movement was the time of climatic change, during intermediate stages between cold periods and temperate periods, when the morphodynamic activity was more intense. According to earlier works, in Great Britain the Pleistocene-Holocene passage (11,000 - 10,000) is also a period of cold periglacial conditions, of discontinuous permafrost, i.e. a period of widespread slope instability throughout the country. There is confirmation in the Cortina d'Ampezzo basin, where several movements are dated between 10,000 and 9,000, and where it is thought that numerous and extensive landslides occurred after the retreat of the glacial masses, partly due to the unloading of the glacier on the slopes. However, the example of the Eibsee rockfall calls into question the supporting role of glaciers in the movement and thus its presumed age.

It is assumed, since Starkel (1966) that late Boreal and early Atlantic (7,500 - 6,000 BP), periods of climate slightly warmer and wetter than at present, with rising sea levels, are phases of frequent landsliding in Europe. Several landslides in the Cortina d'Ampezzo basin are around 4,000 - 5,000 years old, which is assumed, in this area, to be linked with the not-humid "climatic optimum" of the Upper Atlantic. The results obtained in Cantabria show geographical limits of this warmer and wetter Upper Atlantic environment and its rate in triggering mass movements. In these mountains of NO in Spain average temperatures were approximately two to three degrees centigrade higher than they are today, but the climate was also drier, approximately 300 to 400 mm less annual rainfall. This period was also one of great expansion for Neolithic populations, of deforestation and pasturing. These, therefore, are the natural and anthropic factors which were not favourable to mass movements, but which, on the contrary, promoted intense fluvial and pluvial erosion.

On the other hand, the transition between the Subboreal and the Sub-Atlantic (3,000 - 2,000) coïncides with an increase in rainfall and deforestation and sets in motion a recrudescence of earth movements. It should also be remembered in this work that during the Little Ice Age (1550 - 1850), Britain suffered an appreciable climatic deterioration, characterized by cold winters and wet summers, probably with abnormally high ground water levels and a large number of major landslides have been reported as occurring during this period. In Italy, in the Marche region, active or dormant phenomena represent reactivations of ancient movements, which in several cases have their origin in the beginning of the Little Ice Age.

With regard to the last two centuries, another important result is the demonstration of the significance of human factors and their frequent implication in the triggering or activation of earth movements. The abandonment of cultivation and hay fields, deforestation, the opening up of rural local and main roads, the mechanical ski-lifts, the remaking of ski-runs in the mountain communes, now transformed by winter tourism and the urbanisation of peninsular Italy are striking examples of this. The part played by earthquakes is demonstrated in the Dolomites but, taking account of the seismic distribution in the regions studied, it appears to be a subordinate role by comparison with anthropic and above all climatic causes. In most regions these are brought together, as far as possible over a period of one to two centuries, in spite of the distortion, firstly, between the movements, not all of which are known, or flot always known accurately and secondly, the climatic data, generally complete, accurate and day-to-day.
France - Normandy: The juxtaposition of annual rainfall diagrams and the annual number of cases recorded shows a close relationship between years of heavy rainfall and the number of movements. The best adjustment is provided by a mobile average calculated over three years. The pluviometric structures and their combinations are decisive: heavy or very heavy rainfall in the preceding month or years. A factorial analysis of multiple correspondence shows that if the pluviometric conditions preceding earth movements are varied, disturbance can occur even after relatively dry months, if they were preceded by abundant annual rainfall.

France - Bachelard valley:
The mean threshold intensity for the occurrence of overland flows as determined from rainfall simulations is about 42 - 47 mm/h over at least 5 minutes. If the regolith is very wet the threshold rainfall intensity is much lower: 3 mm/h. In dry regolith mean sorptivity values measured were 0.14 cm/min.

Southern Germany - Rheinhessen:
In Southern Germany the geomorphological, geological and climatic context are still akin to the regions above but there is no longer a destabilizing marine action and the land use and its development have doubtless been a very strong factor. Nevertheless we still observe the same essential rate of rainfall in landslide occurrence and activity.

- Eastern Pyrenees: Heavy rain is responsible for triggering landslides, as are wet periods, and a threshold of 140 mm in 2 days may be accepted as an essential condition for triggering mudslides, though not sufficient in itself. Several rainfall events have exceeded this threshold without causing reactivation, for example in Clos d'Esquer. The results of tree examination cannot all be accepted as a sign of landslide activity, and when rainfall data are not available landslide activity deduced exclusively from dendrochronology must be based on a representative number of trees affected (at least three).

Italy - Sicily:
Movement phases are found to be linked with the highest cumulative rainfall values over 50 days. Shallow slides are more likely to occur after short heavy showers which do not infiltrate easily. This aspect is part of a proposed double-threshold model for forecasting future landslides when rainfall reaches a given intensity/time configuration.

Paramètres de l'aléa
Sensibilité du paramètre de l'aléa à des paramètres climatiques
Informations complémentaires (données utilisées, méthode, scénarios, etc.)
Mass movements controlling factors

Heavy rainfalls

France - Normandy: A systematic examination of newspapers dating from 1885 to 1985 in and around the Pays d-Auge, enabled us to list 320 cases of earth movements, 90% of which related to the populated coastline.

Southern Germany - Rheinhessen:
There have been over 750 earth movements over the last 150 years in the 1,000 square kilometres studied. They enable us to establish a return frequency of approximately 25 - 30 years over the last 100 years. For the average annual effective rainfall calculations, the missing pluviometric data having been completed by linear regression from the most central station (Alzey).

- Santander: Analyses of pollen were taken from a bibliographical review of pollen analyses over the whole of Northern Spain and served to establish an interpretation of climatic changes since the end of the Pleistocene period and their effects on vegetation at different altitudes. The chronology of earth movements is based on spatial relationships between deposits and formations and on a chronostratigraphy based on dated fossil peat.

- Eastern Pyrenees: Several dozen landslides were visited in the search for datable elements, which could sometimes be observed on the landslide surface (tilted trees, peat bogs) but in many cases buried or hidden by the slide debris, colluvium or alluvial deposits (charcoal, stems etc.). Six places provided elements for carbon dating, and eight further locations were chosen for dendrochronology analysis and to establish the eccentricity value.

France - Baronnies, Drac and Durance basins: Geomorphological cartography have enabled us to establish relative chronology from the packing and superimposition of formes and formations arising from morpho-climatic altemations, in particular the cutting or recutting of several generations of periglaciary glacis in the marl and limestone and criteria such as cementation and rubefaction of the breaches.

In the Cortina d'Ampezzo basin geomorphological and stratigraphical analysis identified 30 earth movements, one of more that 2 million km², though over a relatively limited spread: 5 complexes, 4 falls, 9 slides and 11 flows. They were dated by C14 radiometric analysis of organic remains.
In the Alpago basin work relates to post-1960 records, which indicate 28 movements spread over 7 valleys, all high or very high frequency. The development of the large landslip at Tessina (3.5 million m3) was analyzed by aerial photography for 33 years. It was a rotational movement at the out set and developed into a flow which threatens Funes. An installation upstream and in the flow will give the alarm if there is a rapid rise in the refilling and a risk of overflow.

France - Trieves; Barcelonnette basin: In the Trieves and Barcelonnette basins, several shallow and deep-seated landslides respectively in varved clays and black earth were investigated with a view to understanding their temporal behaviour in recent times.

France - Vars Commune: A communal data base relating to hazards and land use covers all the 9,000 ha of the Commune's territory; it records events over 200 years. Three hundred earth movements are of various dimensions and types: mudflows, rotational landslips; rockfalls, composite and complex movements.

Italy - Sicily:
Information on landslide movements was collected by means of field interviews and technical literature. Less information was obtained from newspapers owing to incomplete consultation and the fact that some landslides involved uninhabited areas and were not therefore regarded as newsworthy. There is no significant landsliding, most landslides involving only the surface ground. Dates which include the day of occurrence are very scarce.

The researchers engaged in the programme used the same methodologies for the collection, storage and analysis of data.
Raw pluviometric data was converted into effective rainfall by calculating the evapotranspiration by different models, sometimes Thornthwaite, sometimes Turc, and we used various methods of relating rain to movements, including graphic correlations, factorial analysis and multiple correspondance, or mathematical formulae defining alert threshold values from the ratio between rainfall intensity and duration.

Finally a significant part of the research was devoted to a short period, the last five to ten years, in order to confirm the processes involved, particularly in debris flow, to measure movements, both on the surface and in depth, to analyse their seasonal rhythm and to establish hydrological prediction models.

(4) - Remarques générales
If we must express a regret, it is that comparisons would have been easier and more fruitful if the nine participants had all been able to work in one or more of the predefined time-scales; unfortunately this was impossible, and recognized as being so from the outset.

(5) - Préconisations et recomandations
Destinataires et portée du rapport Scientific community
Types de recommandations et / ou préconisations
The operation should continue, giving priority to medium-term prediction. To this end it would be useful to augment regional research into the inventory of movements during the last two centuries and in-depth research into the factors or the combination of the natural and anthropic factors involved in their appearance or reactivation.