Large rockslides in the Alps: A commentary on the contribution of G. Abele (1937–1994) and a review of some recent developments
Published:January 01, 2002
- PDF LinkChapter PDF
Andreas von Poschinger, 2002. "Large rockslides in the Alps: A commentary on the contribution of G. Abele (1937–1994) and a review of some recent developments", Catastrophic Landslides, Stephen G. Evans, Jerome V. Degraff
Download citation file:
Herein I describe the status of rockslide studies in the Alps and present a brief history of the studies. Special attention is paid to the outstanding part in landslide studies in the past 25 years taken by G. Abele, and I introduce his main work to those English-speaking readers who may not be familiar with this scientist. After extensive morphological studies on many rockslide deposits in the Alps, Abele studied huge debris flows that were possibly triggered by rockslides. In addition, new aspects in rockslide studies and the latest rockslides in the Alps are described.
In the geoscientific literature on the Alps, there are hundreds of rockslides described to a degree of detail, and about the same amount that have not been investigated in any depth. Nevertheless, a great number of rockslide deposits are still unknown today, wrongly interpreted, or only locally known. Even after intensive campaigns of geological mapping in the Alps, it is still possible to detect rockslide deposits of hundreds of million cubic meters in volume (e.g., Poschinger, 1986).
When the slope of Monte Toc failed in October 1963, and 250 × 106 m3 of rock slid into the new Vajont reservoir, the displaced water caused the largest landslide disaster known in recent European history. The question of responsibilty led to many detailed studies of the Vajont site (e.g., Müller, 1964, 1968; Weiss, 1964) and of rockslides in the Alps and elsewhere in the world.
Dangers posed by rockslides in the Alps are currently increased by establishment of settlements in hazardous areas and by possible changes in environmental factors influencing slope stability. Research budgets to support study of rockslide hazards are typically small. However, increased consciousness of slope movements, at least among the more environmentally aware segments of the population, may facilitate future efforts in rockslide research, mitigation, and avoidance. In order to reduce losses by future rockslide disasters as far as possible, more knowledge of the characteristics and effects of rockslides is necessary. Because of current small budgets, the potential of all the branches of science concerned with rockslides must be focused in an interdisciplinary way in order to find new answers to the outstanding questions of the triggering factors, the slide mechanisms, and the potential reach of rockslides.
ROCKSLIDES IN THE ALPS
The Alpine population has had to deal with rockslide problems for many centuries. The Greek historian Polybios (201−120 B.C.) reported that even the famous passage of Hannibal through the Alps with elephants was hindered by a rockslide (Seitz, 1987, p. 23). More detailed descriptions of historic rockslide events exist; e.g., in the seventeenth century, the prosperous city of Plurs was entirely destroyed by a rockslide (Figs. 1 and 2) (Presser, 1963), and in the nineteenth century (1806) the village of Arth-Goldau in Switzerland (Figs. 1, 3, and 4) was annihilated by a huge rockslide, resulting in 457 deaths (Heim, 1932, p. 198; Zehnder, 1974). Scientific interest in rockslides developed rapidly in the middle of the nineteenth century after some disastrous events, and was given further impulse with the catastrophe of Elm (Figs. 5–7) in 1881, when 115 people died (Heim, 1882a).
The Swiss geologist Albert Heim was one of the first to describe not only the destructive results, but also the dynamics, the rheology, and the causes of rockslides. He created a classification of rockslides and debris slides with 20 different types, and gave detailed descriptions of examples for each type. Heim paid most attention to his types 14 and 15, the proper rockfalls and rockslides. There he used the terms “Trümmerstrom” (debris stream) and in an analogue sense “Sturzstrom” (Heim, 1932); the latter is still in use in English terminology, but not in German. In describing these terms he compared the motion of large, rapidly moving rock masses with a liquid motion. He proposed, together with E. Müller-Bernet (Heim, 1932), the energy line or slide model (Fig. 8) to estimate the overall coefficient of friction, and thus the potential reach of rockslide debris (Heim, 1882b, 1932). The energy line model describes the path of the center of gravity of a sliding mass, following the central line of the stream with all its curves. In practical field work the energy line model is often difficult to apply and the generalized model of the “Fahrböschungswinkel” (overall inclination of the slide slope) can give a good approximation of mobility. The “Fahrböschungswinkel” (Fig. 8) describes the inclination of the line connecting the highest top of the scar and the extreme limit of the deposit, also regarding curving of the main stream line.
Historic rockslides were described by Escher (1807) and Baltzer (1875), and by Montandon (1933), in a more detailed study. A more recent and very valuable attempt to describe the main features of historic rockslides was made by Eisbacher and Clague (1984). They collected information on ~137 events in the Alps, more than two-thirds of which are debris flows, and nearly a third rockslides, rockfalls, and landslides.
After the Vajont disaster in 1963, no significant rockslide occurred in the Alps until the Val Pola rockslide in 1987 in the Valtellina Valley, Italy. People seemed surprised that alpine morphogenesis was still active, because 24 yr had passed since the previous major rockslide disaster in the Alps. There was a tendency to blame human activity for having triggered the rockslide; however, a combination of natural conditions resulted in the failure, as discussed herein and by Govi et al. (this volume).
The historical examples of landslide and rockslide catastrophes in the Alps indicate that damage has often been caused by secondary effects. In many cases, rockslides have been transformed into debris flows, devastating dwellings far away from the original rockslide site. One of the numerous examples is the Schrofen mountain at Brannenburg, Germany (Eisbacher and Clague, 1984; Gümbel, 1861). In addition, the damming of a valley by rockslide masses and the resulting outburst of the lake have been common phenomena. In 1513, after a rockslide, a debris mass of ~10–20 × 106 m3 blocked the Val Blenio at Biasca, Switzerland (Eisbacher and Clague, 1984). An outburst of the rockslide-dammed lake more than 1 yr later was responsible for more than 600 deaths. This type of problem became a matter of concern again in 1987, after the Valtellina rock avalanche.
ABELE'S BERGSTÜRZE IN DEN ALPEN
One of the most important treatments of alpine rockslides since Albert Heim's (1932) fundamental work, Bergsturz und Menschenleben (“Rockslide and human life”), is G. Abele's (1974) Bergstürze in den Alpen (“Large Rockslides in the Alps”) (Bergstürze is plural of Bergsturz). In most respects the work is still up to date, even ~25 yr after its publication; Abele died in 1994, and this has remained his main work since he had only begun to write the book Dynamics of Rockslides and Rockfalls with T.H. Erismann (Erismann and Abele, 2001) shortly before his death. (For an obituary of G. Abele and a list of all his publications, see Leidlmair, 1994.)
In Bergstürze in den Alpen, Abele's first intention was to compile a map showing the distribution of large rockslides in the Alps and to indicate the possible existence of geographical, geological, or morphological zones with high rockslide densities. His second aim was to describe the geomorphological features of large rockslides, and to present correlations between various geomorphological, geological, and mechanical parameters. He reported on the phenomena and processes occurring during, or subsequent to, rockslide events.
Abele never considered the term “Bergstürze” to be satisfactory. Like the English term “landslide,” it has been used for too many different phenomena. Furthermore, the literal meaning (translated as “fall of a mountain”) is prone to exaggeration. Nevertheless, the term was commonly accepted in scientific literature and a new definition was required. Abele (1974, p. 4) postulated two criteria as a definition for the term “Bergsturz.” First, a quantitative threshold has to be chosen arbitrarily to indicate a minimum size. Abele fixed this minimum debris volume as 1 × 106 m3. Because the debris volume is often not easily ascertained, as an alternative he fixed the minimum area covered by the rockslide deposits as 0.1 km2 (Abele, 1974, p. 4). If one of these two criteria is met, the dimension is large enough to call the rockslide a “Bergsturz.” The second characteristic factor is the velocity of the moving mass. Because the word “Sturz” means a rapid movement, Abele only accepted fast slope movements, coming down a slope “in some seconds or a few minutes” (Abele, 1974, p. 5), to be a Bergsturz. In this sense there is an analogy to Heim's term “Sturzstrom,” meaning a large volume and a rapid movement. The term “Bergsturz” is not restricted to streamlike movements, but includes in the same way en bloc displacements (e.g., the rockslide of Vajont).
Following from this definition, Abele examined 279 rockslide deposits all over the Alps, among them 186 “Bergstürze.” He examined or mapped nearly all of them (e.g., Figs. 9 and 10). In this way Abele gathered an enormous body of data on large rockslides. In order to acquire these data, he had to amass a sophisticated collection of literature, and as a secondary result his work is one of the best bibliographical references on rockslides in the Alps.
In accordance with these aims, he did not describe each site one by one, as the majority of previous authors had; he wanted to point out common features in order to set up patterns of behaviors and to support them statistically.
The incidence of rockslides is not equally distributed all over the Alps from Marseille to Vienna (Fig. 1). Abele (1974) pointed out some areas of relatively low rockslide intensity, e.g., in the southern part of the Western Alps (Fig. 1). However, he found concentrations of large rockslides in some regions. The most important of these are the surrounding areas of the Aarmassif in Switzerland, west of Chur (Fig. 1), the line from Köfels in the Ötz Valley to the Fernpass, west of Innsbruck in the Austrian Tyrol, and in the Italian Southern Alps east of Bozen (Fig. 1). The Rhine-Rhone line just south of the Aar massif is a region of high rockslide intensity. This line is also the area of the most intensive relative uplift in Switzerland (Gubler and Kahle, 1984), so the question of a possible correlation between neotectonics, uplift, and rockslides has to be posed.
The intensity of large rockslides is relatively high in areas of sedimentary rock, especially in the Northern and in the Southern Calcareous Alps (Fig. 1). Abele (1974) calculated the proportion of areas covered by rockslides in the Southern Alps as 0.38% and as 0.43% in the Northern Calcareous Alps, where, compared to the Southern Alps, fewer but larger rockslides have occurred. In the crystalline zones of the central Alps, the rockslide intensity is only 0.08%, comparatively low. It reaches a minimum in the old crystalline cores, where 0.02% area covered by rockslides. Abele (1974) explained the clear domination of areas with carbonate sediments as being due to the massive rigid carbonate units that often dip valleyward and have only few widely spaced joint systems. The large amounts of potential energy stored in these masses can be released instantaneously. In contrast to the sedimentary rocks, crystalline rocks generally show a marked highdensity jointing. Consequently, crystalline rocks produce more frequent, but smaller mass movements, thus rarely fulfilling the conditions of a Bergsturz as described herein. Abele (1974) also remarked on the frequent spatial clustering of separate rockslides. This is due to the similarity of geotechnical conditions in adjacent areas. He pointed out that this observation could serve as a hint that future rockslides might usually occur in the vicinity of old rockslide scars.
Abele's analysis of the morphological dimensions at the 186 sites he investigated shows up some regularities, in that it shows the extent and the scale of the problem (Abele 1974). He compared the slope morphology and the inclination of the slope before failure with the relative dimensions and the volumes of the detachment zones. The biggest rockslide in the Alps is by far that of Flims (20 km west of Chur, Figs. 1, 10, and 11), with an estimated volume of 7–15 km3. The next largest are the rockslides at Sierre (>2 km3), Köfels (>2 km3), Engelberg (1.5–2 km3), Tamins-Säsagit (1.6 km3), Fernpaß (1 km3), and Kandersteg (0.9 km3). With the exception of Köfels, all of these occurred in carbonate rocks.
Abele (1974) classified the deposition areas of large rockslides and analyzed their spreading, length, slope angle, surface area, and thickness. A clear correlation between the material and the form of the deposit was recognized: rockslides in crystalline rock produced relatively thick deposits with little spreading. The commonly observed phenomenon, that rockslide masses evacuate the scar entirely and produce a proximal depression on the foot of the slope, has been found as occurring in ~60% of the sites.
Special emphasis has been put on the influence of the path of the moving masses, especially on their length, their spreading, and on the value of the “Fahrböschungswinkel” introduced by Heim (1932). As has been discussed by many since (e.g., Scheidegger, 1973; Erismann, 1979), a clear influence of the volume of the sliding mass on the “Fahrböschungswinkel” is evident as huge masses are associated with significantly smaller angles.
Abele (1974, p. 61) gave the traditional explanation for the causes of large rockslides, i.e., that the retreat of the glaciers after the last glaciation is the main cause of most of them. Support for his hypothesis was the observation that more than half of the rockslides for which at least a relative dating was possible had been thought to be late glacial in age (Abele, 1974, p. 89–90). Abele later changed his opinion.
Abele paid special attention to the interpretation of the morphological forms of rockslide deposits. Many phenomena indicate distinct differential movements of the masses, in both vertical and horizontal directions. Abele (1974) interpreted the common phenomena of longitudinal ridges within rockslide masses, often assumed to be morainic deposits, as levees produced by internal differential shearing during the movement. Another phenomenon is the frequent preservation of distinct geological units in rockslide deposits in vertical and/or horizontal directions reflecting the geological succession in the source area (Abele, 1974). The movement in these cases cannot have been turbulent, but must have been laminar, the differential movement being concentrated in some critical zones. Such movements may explain how even huge brittle blocks are preserved. They must have been transported on top of the debris mass or within zones of less differential movement.
Characteristic lateral levees are commonly found along the rims of the deposits parallel to the direction of the movement. In contrast to the main debris streams, which destroy any obstacle in their path, the lateral levees in many cases were emplaced gently, in some cases flowing around trees without harming them (Abele, 1974, p. 78).
1987 rockslide in the Valtellina, Val Pola, Italy
On July 28, 1987, at 7:23 a.m. local time, a mass of ~35 × 106 m3 of crystalline rock (gneiss and micaschists) swept down from the eastern flank of the Monte Zandila, in the Italian valley of Valtellina, and blocked the Adda River (Figs. 1 and 12) (see also Govi et al., this volume). The rockslide source area was already known to be an area of old, deep-seated creep exhibiting characteristic sacküng morphology. After intense precipitation at relatively high temperatures that pushed the 0°C line above the altitude of 3000 m, on July 19 a debris flow originated from the Val Pola Valley, a small steep tributary of the Adda River. The fan of this debris flow dammed the Adda River to create a small lake of only ~5 m depth. It was this little lake that was responsible to a considerable degree for the disastrous consequences of the rockslide. On July 25, large active cracks at an altitude of 1200 m above the valley floor were detected by regional geologists. This, together with the increasing rockfall activity, was warning enough for the authorities to evacuate the villages beneath Monte Zandila.
When the entire rockmass came down on the morning of July 28, it killed seven workers on the small Val Pola debris fan. The mass entered the valley and ran up the opposite side to an altitude of ~300 m above the valley floor. The lake water of the little Val Pola debris-flow dam was displaced in an upstream direction. The flood wave had a height up to 95 m and traveled 2.1 km to the north (Costa, 1991), devastating the village of Aquilone, where another 20 people were killed. This village had not been evacuated because it was thought to be too far away from the rockslide area.
An explanation for the unexpectedly high energy of the flood wave was given by Costa (1991, p. 21). He tried to determine the velocity of the sliding mass, and arrived at the unusually high value of 310–390 km/h. which would be “one of the fastest large mass movements in historic time” (Costa, 1991, p. 25). Erismann (1993) considered this estimate of the velocity to be exaggerated and made new calculations taking into account the true position of the center of gravity. The result was an estimated maximum velocity of ~248 km/h (Erismann, 1993), which is nevertheless an extremely high value.
The slide mass blocked the Adda Valley with a rockslide dam at least 33 m high, and caused the formation of a new lake. A rupture of this dam during overtopping would have caused an even greater catastrophe in the densely populated lower Valtellina Valley. Through an enormous financial and technical effort, the lake was brought to a controlled overtopping and then to a consequent lowering of the water level by pumping (Costa, 1991).
A further rock slab of some million cubic meters adjacent to the rockslide scar to the north was also considered to be rather unstable. To avoid a second damming of the Adda River, two bypass tunnels were excavated. A giant spillway was constructed on the artificially flattened southside of the slidemass, in order to avoid their erosion in case of the potential overtopping of a second rockslide dam. According to Costa (1991, p. 36), the hazard management activities cost ~$400 million U.S.
Following M. Govi (1988, oral commun.), the precipitation preceding the rockslide event had not been as high as in adjacent regions, e.g., in Ticino. Because the Monte Zandila area is a comparatively dry region, it meant nevertheless an event with a statistical recurrence of 250 yr, whereas in the usually wetter Ticino, the recurrence interval would have been only ~80 yr. This suggests that it is not the average values of precipitations that are important, but the local occurrence of unusually high values.
1991 Randa rockfall
On the morning of April 18, 1991, a rockfall occurred on the west side of the Matter Valley near the village of Randa, in Switzerland (Fig. 1) (Noverraz and Bonnard, 1992; Schindler et al., 1993). The mass of ~20 × 106 m3, mostly of gneiss, failed without clear warning signals, except for an increase in rockfall activity that began just before the event. It blocked the valley floor only halfway (Fig. 13), but nevertheless interrupted the railway and the only road to Zermatt. The Vispa River was dammed by the rock mass, and the possible rupture of this dam threatened the dwellings and the chemical industry facilities downstream. The interest of this first Randa event is in the fact that the failure did not occur in one moment, but was obviously a continuing rockfall lasting some hours (Schindler et al., 1993) The rock masses did not travel far, but formed a steep cone at the foot of the slope (Fig. 13). They destroyed only some stables and holiday chalets without harming any people.
After this first failure, warning signals of a further failure were recognized. With the help of high-tech instrumentation, it was possible to predict the timing of this second event. The movements showed an exponential acceleration before the final movement of another 10 × 106 m3 rockfall on May 5, that also lasted from the afternoon until midnight, and could be described as a continued blockfall with a very small reach. In that sense it was not a real rockslide.
The second failure enlarged the steep cone (Figs. 13 and 14), and increased the problems in the valley (Schindler et al., 1993; Bonnard et al., 1995). The lower parts of the village of Randa were flooded by the lake dammed by the rockslide, before an artificial channel (Fig. 15) through the rock debris restored the runoff of the Vispa River. To avoid future problems caused by potential further rockfalls, a 3.6 km bypass tunnel was built (C. Schindler, 1993, oral commun.).
The Randa rockfall scar involves two geological formations (Fig. 13): massive orthogneisses at the bottom and schistose paragneisses with amphibolites on the top (Schindler et al., 1993). It is assumed that water entered into the upper parts of the series, which show deep stress-relief joints parallel to the surface due to stress-relaxation movements on the steep valley slope (Schindler et al., 1993). Moreover, morphological phenomena on the upper slope such as horizontal crests and ridges indicate that creeping movements had already taken place on this slope long before (Schindler et al., 1993). Schindler (1992) brought up the hypotheses that also the loss of former permafrost in the adjacent higher slopes might have caused the breakdown (see also Schindler et al., 1993). The loss of permafrost enabled the infiltration of surface water, and consequently raised the joint water pressures and caused the erosion of cohesive joint fills.
NEW ASPECTS OF ROCKSLIDE INVESTIGATIONS
The 1987 Valtellina event gave new impetus to rockslide investigations. Abele reentered the field, first at the university of Mainz, Germany, then since 1990 leading the Geographical Institute of Innsbruck, Austria. Reducing the morphological studies in the Andes with which he had been concerned, he paid new attention to rockslides in the Alps. His main research activities in his last years concerned the mobilization of sediments by rockslides. Other new aspects of rockslide investigation were included in his later works, partly in cooperation with me.
Redating of rockslide events
A large number of rockslide deposits in the Alps have been qualitatively dated to the late Pleistocene era, without support from quantitative radiocarbon data or other absolute dating methods. Many events of unknown age were simply said to be late glacial (e.g., Sierre, Fernpass, Eibsee) (Abele, 1974, p. 89), because it was generally believed (Abele, 1974, p. 89; Heim, 1932, p. 185) that this was the most favorable time for triggering rockslides. According to this rather simple view, the melting of the glaciers at the end of the Pleistocene exposed slopes that had been undercut by the ice during glaciation and that had, until the retreat of the ice tongues, been supported by the ice: the phenomenon is probably more complex than this simple scenario.
Another suggested proof for the late glacial age of many rockslide deposits was the presence of morainal deposits on top of the slide masses. Some of these are true moraines with typical striated pebbles, but some are only moraine-like forms in which the material has the typical block composition of rockslides. Typical examples for such an interpretation are the deposits of the Tschirgant rockslide (Fig. 1) in the Inn Valley (Tyrol) (Heuberger, 1975) and the Flims (Fig. 1) rockslide (Gsell, 1918; Nabholz, 1987). Such debris masses were often interpreted to be the deposits of rockfalls on the surfaces of active glaciers that later were transported by the ice downvalley. A famous example for a recent large landslide coming down on a glacier is that of the Sherman Glacier (Bolt et al., 1975).
In the past few years several rockslide deposits in the Tyrolian (Austria), the Bavarian (Germany), and the Swiss Alps have been dated using radiocarbon methods. Some examples are as follows.
The Köfels Bergsturz (Heuberger, 1966), which had some morainal deposits on it, was dated to ca. 8700 14C yr B.P.
The Tschirgant Bergsturz (Patzelt and Poscher, 1993), previously dated by terminal moraines to be late glacial, was dated to ca. 2900 14C yr B.P.
The Hintersee Bergsturz (Poschinger and Thorn, 1995), previously dated by glacial deposits to be late glacial, was redated to ca. 3500 14C yr B.P.
The Eibsee Bergsturz (Jerz and Poschinger, 1995), thought to be late glacial because of a partly glacial morphology, was dated to ca. 3700 14C yr B.P.
The Flims rockslide (Poschinger and Haas, 1997), until now dated by glacial deposits to be late glacial, has been radiocarbon dated to ca. 8300 14C yr B.P.
The results illustrate that even rockslides dated previously by glacial deposits have yielded young, clearly postglacial ages; the glaciation in this region ended ca. 15 ka. Morainal deposits on top of the rockslide masses are, therefore, no longer considered a reliable tool for dating. These deposits may have been transported by the rockslide to their present position. An example of this is at the Köfels Bergsturz (Heuberger, 1994), where a moraine cover can be seen, in addition to a former rock surface with glacial striatum, the latter rotated by 180° in plan to indicate an unrealistic direction of ice movement upvalley. In the same way, the qualitative dating of the huge rockslide of Sierre must be called into question. Because of morainal gravels on top of the rockslide mass, it has been dated as late glacial (Abele, 1974, p. 89).
Recent tunneling excavations revealed masses of till within rockslide deposits (Fig. 16), showing that the rockslide was not deposited by a glacier overriding an even older rockslide mass, as had been suggested. It is much more likely that a rockslide picked up preexisting moraine cover and incorporated it into the sliding masses. At Flims(Fig. 10), Abele (1991) described a similar interfingering of morainic sediments with rockslide masses, and suggested that some of the morainal material at Flims had been deposited after the rockslide by local glaciers. In the meantime, I successfully redated the Flims rockslide by radiocarbon methods as being a postglacial event (Poschinger and Haas, 1997).
Principal causes of rockslides
The redating of several important rockslides in the Alps prompts reflection on the principal causes of rockslides. In probably all of the cases mentioned here, the climatic circumstances were similar to those of the present, perhaps even warmer: 10 k.y. after the retreat of the glaciers, the glacial influence can therefore not have been responsible, at least not directly. Glacial influence always was a good deus ex machina for many unexplainable phenomena of rockslides, but now other solutions must be considered.
It seems that until now the long-term factors of stress relaxation and weathering have not been taken into proper consideration. Schindler et al. (1993) gave great importance to this factor for the Randa rockfall. Clear evidence of the influence of stress relaxation is the common occurrence of joints parallel to slopes, without tectonic influence and 200–250 m beneath the slope surface. Such joints were found in the Randa rockfall scar, in the bypass tunnel of Randa, and in many other alpine tunnels (Schindler et al., 1993). Whereas the common hypothesis is that most of the rockslides came down just after the melting of the glaciers, another speculation is that weathering and stress relaxation is responsible for increasing rockslide activity following the erosion of the slopes by the glaciers. The truth may be between these two extremes; i.e., after an elevated rockslide activity following the melting of the glaciers, their incidence reduced to a moderate level, and due to time-dependent stress relaxation, loosening, weathering, and similar processes might grow slowly over time (Fig. 17). This means that the statistical probability of a future rockslide event may be much higher (H3 in Fig. 17) than according to the traditional hypothesis (H1 in Fig. 17). Events of the magnitude of Flims cannot be ruled out in the present environment. Both on these grounds, and because of the effect of possible changes in environmental conditions described in the following, an even higher future rockslide activity must be anticipated.
Valley sediments and rockslide reach
It is the merit of G. Abele's work that in discussing rockslide mechanics, considerable attention is paid to valley sediments in the path of a moving mass. Starting from observations in the vicinity of the Flims site (Fig. 10), he postulated the mobilization of gravels on the valley floor by the impact of a sliding mass (Abele, 1991; 1997). He analyzed the Bonaduzer Schotter (or Bonaduz gravels) deposits, and came to the conclusion that these sediments in the vicinity of the Flims rockslide must have originated through such a mobilization. The Bonaduz sediments are gravels within a sandy matrix that lack any layering (Fig. 18). The sediment pile is as thick as 50 m and shows a clear upward grading with sandy coarse gravel at the bottom and pure fine sands on top. The gravel components show a preferred sub-horizontal orientation in the center of the valley, an orientation parallel to the adjacent slopes at the margins. Characteristic are clods not only of rockslide material, but also of layered silt (Fig. 18) “swimming” in these sediments; the layering is arbitrarily, not horizontally, oriented. Obviously, lacustrine sediments have been ripped up and integrated into the gravel mass. Pavoni (1968) discovered vertical pipes centimeters thick and meters high, lacking fine material (Fig. 18). He interpreted these as dewatering pipes of a quickly consolidating water-rich sediment. I found some unusual long vertical joints in a gravel sediment, filled with some centimeters of vertically banded sand and silt (Fig. 19). Another curiosity of these sediments is the direction of sedimentation, which is up the Hinterrhein Valley. Pavoni (1968) remarked on the similarity of these sediments to those of a “wet debris flow”; Zimmermann (1971, p. 168) compared them to an “oversized graded bedding.” The deposits are reminiscent, in analogy to Zimmermann's observation, of the sediments of an oversized turbidite. It seems that the transport mechanisms of the gravels have been unknown, and there is no recent natural example for a similar process.
Abele discovered that the Bonaduz gravels are not unique in the Alps. Similar sediments, with characteristic lack of layering and the typical graded bedding, are in front of and beneath the rockslides of Tschirgant (Fig. 1) and in the Alm Valley (Figs. 1 and 9) (Abele, 1991). An interfingering of gravels and rockslide masses is exposed at the Tschirgant site (Fig. 20). Here, gravels seem to have “intruded” upward into the overriding rockslide debris. Abele searched for and discovered such gravels in front of the Köfels site (Abele, 1991, p. 33).
Abele believed that these sediments had been mobilized by fast rockslide masses. The valley-floor sediments are assumed to have been saturated in water when the sliding mass arrived. The sudden overloading generated pore-water pressures in the sediments, reduced friction to nearly zero, and mobilized the sediments. This reduction of friction explains the high travel distance of the rockslide. Typically, some smaller parts of the slide mass have traveled within and on top of the mobilized sediments to form separate little hills, far from the main rockslide deposit. For example, the hills of Rodels at Flims (Fig. 10), consisting of typical rockslide material embedded in the Bonaduz gravels, are more than 12 km upvalley, south of the front of the Flims rockslide mass.
These examples illustrate that the phenomenon of undrained loading by slide masses, resulting in the development of excess pore-water pressures in overridden sediments, is not uncommon. Even during slow movement it can prolong the travel distance of a sliding mass (Poschinger, 1994). Abele concluded (1993, oral commun.) that excess pore-water pressures as a factor in reducing friction must be considered when making predictions about runout distances of large rockslides. Sassa (1988) investigated the influence of pore pressure at the slip plane on the internal friction in the laboratory, and found very low values of friction when considering pore pressure.
Human activity and ecological factors triggering rockslides
Ecological damage in the Alps brought about by human activity started very early through the exploitation of natural resources. For example, salt mining activities are known from the “Hallstatt” time (800−400 B.C.), and the “ferrum noricum,” an iron ore from the Tauern region, was an important strategic material for the Roman armament industry. In smelting ore, or carrying out salt panning, it was necessary to burn enormous masses of wood from the surrounding forests. So even in early time, deforestation of whole mountains took place. It was not before the sixteenth century that the growing devastations of the forests caused the first foresting regulations, e.g., for Bavaria in 1586 (Bayerisches Staatsministerium für Ernährung, Landwirtschaft und Forsten, 1987, p. 5). Furthermore, forests of the Southern Alps fell victim to the ship-building acivities of the Mediterranean countries. Even where there is no documentation, it must be assumed that many landslides and rockslides occurred as a consequence of this massive deforestation. Surely slate mining directly caused the famous 1881 rockslide of Elm (Heim, 1882a), and the Plurs disaster (Fig. 2) has been explained as due to the mining of serpentinite (Presser, 1963).
Today, unlike probably any other mountain range, the Alps region has increasingly become a playground for sport and leisure. The coincidence of a high density population in the Alps and their surroundings, and an elevated standard of prosperity, high mobility, and increased leisure time, are responsible for growing ecological problems. Often it is not the sports activities that cause problems, but the necessary infrastructure, such as access roads, water drainage, and canalization or leveling of slopes. Until now these measures had no direct influence on large rockslides. Damage has been restricted to local erosion, including smaller debris flows and shallow landslides.
Another growing ecological problem in the Alps is the destruction of forests by air pollution. The proportion of trees with severe damage in the eastern and central Alps is between ~10% and 35% (Bayerisches Staatsministerium für Landesentwicklung und Umweltfragen, 1992). This means the loss of a good part of the water-retaining function of alpine forests, with the consequent effect on the problems of slope stability. Although no large-scale mass movement has yet been clearly attributed to the damage of the forests, some erosion features, debris flows, and smaller landslides have undoubtedly occurred for this reason. Most of the forests in the Alps are concentrated on the lower slopes and their destruction will not affect the upper slopes, where most of the large rockslides start. Nevertheless the possibility of important rockslides of the dimension of a Bergsturz being triggered by these phenomena cannot be ruled out.
Any change in long-term environmental conditions might influence the stability of slopes. As described by leading climatologists, climatic change as a consequence of greenhouse effect and air pollution is probable (Houghton et al., 1995): most agree that a change has started or will take place. The possible influence on slope stabilities has been the subject of a study in Switzerland (Bonnard et al., 1995). Two main effects have to be taken into consideration concerning slope stability; i.e., temperature rise and increase in heavy rainstorms (Houghton et al., 1995).
As the example of the Valtellina rockslide shows, extraordinary precipitation can be responsible for triggering rockslides. The slopes became accustomed over hundreds of years to the load changes resulting from subsurface water-table changes, and adopted a stability or inclination to conform to these conditions. According to climatologists, an increase in heavy rainstorms might be the consequence of intensified airstreams due to an increase of temperature gradients. In a similar way to the Val Pola event, exceptionally heavy and warm precipitation might come down as rain at higher altitudes than usual, the snow limit climbing to more than 3000 or 4000 m, thus making unusual amounts of water available to surface and subsurface runoff. This brings exceptional load conditions to the slopes, which will tend to overcome the generally small threshold to instability.
A rise in temperature will have significant influence on the slope stability in glaciated and permafrost areas (Haeberli, 1992). Haeberli (1990) demonstrated that the limit of mountain permafrost probably rose during the years since the Little Ice Age, and since 1920 (Kerschner, 1991) by ~150–200 m. A further accelerated rise of the permafrost limit as a consequence of the present temperature increases can be expected. The loss of permafrost in mountain areas causes deformations and weakening of the strata. The most important effect is that in permafrost areas most runoff will be on the surface, because the ice has sealed the subsurface waterways. The opening of this ice sealing facilitates the infiltration and so changes severely the hydrological conditions. The loss of permafrost can also cause a rockslide. Such an influence has been suggested for the Randa (Schindler, 1992) and Valtellina (Govi et al., this volume) events.
Summary and Conclusions
Since the Vajont catastrophe much rockslide research has been carried out in the mountain areas of the world. Gerhard Abele contributed a significant part to its progess in the Alps, particularly with his interpretations of field examples to explain high runout distances. His interpretations of the Bonaduz gravels as a debris flow triggered by the Flims rockslide and traveling ~15 km up the Hinterrhein Valley may be important for future hazard evaluations and risk assessments. Those responsible will have to consider if the specific situation might be comparable and allow for the generation of excess pore-water pressures in the valley sediments. If so, the endangered zones may become extremely large, as will the associated socioeconomic hazards.
Further investigation of the possible risk areas, especially those of old sagging slopes and the so-called dormant rockslides, is necessary. It is clear that analyzing the morphological warnings of reactivation in old, supposedly dormant, rockslides must become a routine part in the prediction of present and future hazard areas. This becomes even more urgent if expected environmental changes take place.
In reviewing rockslides in the Alps, the contribution of G. Abele is outstanding. He studied the geomorphology of rockslide deposits, and from this drew conclusions regarding the mechanics of large rockslides. To evaluate the present risk of rockslides, changing environmental influences must be taken into account. Radiocarbon dating shows a growing number of rockslides to be younger than previously thought, and so in many cases to suggest a direct glacial cause is inappropriate. These combined factors mean that the hazard of large rockslides in the Alps is ever present and potentially on the increase.
Figures & Tables
>This volume documents further advances in our knowledge of catastrophic landslides since the pioneering compilations of the late 1970s by Barry Voight. It provides a worldwide survey of catastrophic landslide events written by leading authorities. Catastrophic Landslides begins by drawing upon South America to dramatically illustrate the impact of these phenomena on human populations. The occurrence of catastrophic landslides, including site-specific insights, is shown through six events of the past 20 years. Several other chapters focus on the mechanisms involved with catastrophic landsides both in relation to geologic factors in a particular geographic area as well as to specific geologic processes.