The previous section dealt with the stability of rock faces and the critical slope along which rock falls may occur. Rock slides may be of any size and are caused by ruptures located deep in the rock mass. The statics and dynamics of a very large slide may be completely different from those of a smaller one and sometimes they are most difficult to analyse.
There are many varieties of rock slides, both man-made and from natural causes. Cuttings for roads or railways is one cause. It was the cause at the construction site of the Lotschberg railway in the centre of the Alps. Here the south ramp to the main terminal was too sharply inclined and rock slides were a constant hazard. They also occur as a result of excavations for dam
208 Rock slopes and rock slides
foundations (Bort dam in France) and open-cast mines (Hoek, 1970, 1972, 1973; Londe et al.9 1969). But by far the most common man-made cause is the varying water levels in artificial storage reservoirs.
The result of water and ice percolating through rock will be dealt with in the next section, and the section following that will deal with water percolating from a hydraulic pressure tunnel.
9.2.1 The formation of rock slides, shape of sliding surfaces and the progress of slides
The rupture of a rock mass and the progress of rock slides is very different to, and more complex than, slides occurring in loose soils, for which rela- tively simple theories have been established. As early as 1882, Heim had stressed the difference between slides occurring in loose soils, rock slides and mixed slides, a classification which he maintained in his main paper, Berg- sturz und Menschenleben (1932), in which he discussed various types of large and very large slides, as they occurred in the Alps up to 1932. He analysed conditions in the area where the ruptures first occurred and followed the progress of the slides along the mountain slopes until the movement ceased.
Figure 14.1, from Miiller on Vajont, illustrates perfectly a type of rock slide with downturning movement of the rock tops which could not occur in loose material.
According to Mencl (1967), confirming Terzaghi's remarks on the stability of rock faces, the pattern of the main sets of joints determines the rupture of rock masses and the shape of most rock slides. Mencl shows how the move- ment of the whole mass of rocks sometimes depends on the inclination of a near-vertical set of joints which may be 'positive' (inclined towards the valley) or 'negative' (inclined away from the valley).
Figure 9.3 shows the different stages of a rock slide starting with tensile fractures on the top of the cliff and progressive macrofractures within the rock masses. According to this theory, final rupture would occur at the base of the slope.
r . fractures
fractures
(a)w
Fig. 9.3 Progressive deep fracture of the rock mass causing a deep rock slide.
(a) Tensile stresses cause fissures at the top of the cliff; (b) progressive rock fractures in the deeper layers; (c) rupture at the basis of the slope causing rock slide (after Miiller, 1963a).
Some recent tests on photoelastic models show that tensile stresses develop at the base of the slope - at depth inside the rock. If these findings are substantiated then it may be that tensile stresses activate the rupture. Two- dimensional models of a problematic rock bank at Kurobe IV dam (Japan) were constructed of highly elastic material with a gelatinous base. First the isotropic case was investigated: then a model featuring a system of joints was tested. The results provided valuable data within the complex of stress distribution (Muller & John, 1963).
Stability analysis of rock masses in doubtful equilibrium along an inclined plane follows the simple rules for statics. In fig. 9.4 the weight W of the rock
w\\ bedding plane or fault
II III
Fig. 9.4 Rock slide along an inclined
bedding plane or fault (<£ < a). Fig. 9.5 Rock fall, possible lines of rupture: I or II.
mass above a possible plane sliding surface (bedding plane or rock fault) is decomposed in a triangle of forces. When the weight component, T — W sin a, is larger than the friction force, N tan </>, there is danger of a slide unless there is high rock cohesion along the sliding surface which compensates for the differences between positive and negative forces. However, the stress analysis discussed in section 7.3.2 usually applies. In fig. 9.5 rupture is supposed to occur within the rock mass, probably following the bedding planes and sur- faces of fractures (lines I or II). Rupture would also happen along slip line III, fracturing the rock and cutting through the bedding planes (Vajont).
In a high mountainous area bisected by deep valleys, the rock underlying the valley slopes may be the seat of shearing stresses. If these stresses are already close to the shearing resistance of the rock, any increase in the valley depth, caused perhaps by accelerated erosion, would result in a general rock slide.
In fig. 9.5, the curved line III—III is supposed to develop progressively in the mass of fissured rock, but cases have also occurred where a curved smooth rock stratum, lying deep below the rock surface, offered an easy start to a rock slide. The Vajont rock slide followed such geological stratifications in the upper, steeply inclined, parts of the slope, but cut through other strati- fications on the lower part.
Heim has described several major Alpine rock slides which could be classified under either the case illustrated in fig. 9.4 or in fig. 9.5.
210 Rock slopes and rock slides
Bernaix (1975) suggests that instability of a rock mass could be due to progressive shearing of a rock zone well below the surface. In a high moun- tainous area bisected by deep valleys, the rock underlying the valley slopes may be the seat of shearing stresses. If these stresses are already close to the shearing strength of the rock, any increase in the valley depth, caused perhaps by accelerated erosion, would result in a general rock slide.
Many problems concerning stability of rock slopes (open-cast mines) are three-dimensional problems. Such problems were systematically analysed by Londe et al. (1969), John (1970) and Hoek (1972).
After rupture of the rock masses, the slide progresses downwards. Heim made a distinction between slowly progressing slides and rapidly accelerating slides. Modern examples of slow and rapid slides are the Pontesi (slow) and the Vajont (rapid) slides (see section 9.6).
It has been observed that major rock slides along deep-seated surfaces occur with very small relative movements between the rocks which constitute the mass. Harrison & Falcon (1937), describing the Saidmarreh landslide in south-west Iran, noted this fact. Miiller (1961) on p. 203 of his detailed first report on the Vajont slide, compared this rock slide to snow avalanches.
He writes of the 'thixotropic behaviour of masses'. Photographs of the rock cliffs taken before and after the slide, when the rocks had shifted 300 to 400 m across the gorge and lifted 140 m upon the opposite side, show hardly any change in the stratifications.
Moving rock masses absorb energy. Mencl (19666) analysing the Vajont rock slide, used models to simulate slides moving along sharp bends of 'chair-shaped' sliding surfaces. Some authors estimate that considerable energy losses occur locally at such bends and suggest that the equivalent friction loss at Vajont could have been as high as tan <f>' = 0-05 (<£' ^ 3°).
Model tests (Fumagalli & Camponuovo, 1975) and in situ measurements have contributed to our knowledge of statics and dynamics of steep rock slopes. Different types of ruptures were detected, analysed and even model- tested; it was found that the time factor is vital (see chapter 14 on Vajont).
Extrapolation of precise measurements of displacements permits the accurate prediction of the final rupture and of the start of the slide (Jaeger, 1969a;
Kennedy, 1970). The method has been used systematically for checks on the stability of the steep high slopes of open-cast mines.
9.2.2 Deep-seated sliding surfaces
It is now evident that some of the geomorphological features of the Alps, formerly ascribed to ice action, could be surface manifestations of deep-seated rock slides. At several dam sites in exceptionally narrow sections of deep valleys it was found that the rock forming one of the slopes had advanced to- wards the valley in a more or less horizontal direction, whereby the rock involved in the movement remained relatively intact (Ampferer, 1939; Stini,
1942). At one of these sites ground moraine was encountered in a boring at a depth of about 300 ft below the surface of what appeared to be rock in situ (Stini, 19526). It is known that several very large slides with deep lines of rupture occurred in Switzerland in historical times.
Heim (1932) and Muller (1964) have produced lists of all the known major rock slides which have occurred throughout history. The most tragic was the Vajont rock slide (Italy) which happened during the night of 9 October 1963. A very high dam had been built across the narrow canyon of the Vajont River, with the dam crest at a level of 722-50 m. The fissured dolomite rock had to be reinforced on both sides of the gorge to withstand the thrust of the dam abutments. In I960, rock masses on the left bank started to move very slowly. At the same time the water level rose in the larger reservoir.
The slide reached as high as 1250 m on the slopes of Mount Toe. In 1963 it was observed that there was some connection between the different water levels in the reservoir and the rock movements. Geologists and rock experts predicted a progressive rock slide which after gradually filling the reservoir would come to a natural standstill. But tragically, they were wrong. The rock slide was unexpectedly violent and it reached extremely high velocities. The resulting water wave was so tremendous that it spilled over the dam crest and flooded the little town of Longarone (province of Venice) causing 2400 deaths. (See chapter 14 for a more detailed account of this disaster.)
It is worth while quoting some of Terzaghi's remarks:
Practically nothing is known concerning the mechanism of these deep-seated large- scale rock slides. It is not known whether the slides took place rapidly or slowly, and it is doubtful whether they are preceded by important deep deformation of the rocks located within the shear zone. However, it is known that the rock located above the surface of sliding has been damaged at least to a moderate extent. Existing joints have opened and new ones have been formed. Hence the compressibility and secondary permeability of the rocks has increased. Furthermore, in the immediate proximity of the surface of sliding, the rock is completely broken or crushed. Hence a site for a high concrete dam should not be considered suitable unless there is positive evidence that the underlying rock has never been subject to displacement by a deep-seated rock slide.
Terzaghi wrote this about a year before the Vajont-Longarone tragedy.