Villavil rockslides, Catamarca Province, Argentina
Published:January 01, 2002
Large-scale catastrophic landslides have occurred at a number of locations in central and northern Argentina. The landslide hazard has recently been studied at one of these locations, in the vicinity of Villavil, Catamarca Province, in the Sierras Pampeanas of northwest Argentina. Villavil is a little village of ~350 inhabitants under serious risk of a landslide, which would destroy the village and surroundings. Seven large block-glide rockslides have developed in a Tertiary volcaniclastic-sedimentary sequence exposed in hogback ridges near the town. In the hogbacks, northwestward-dipping rocks have sheared across bedding and the detached slabs have moved downdip and overridden younger rocks on the lower slopes. Distal flows, resulting from the fragmentation of the slabs, have moved 1800 m into the ancient valley of Río Bolsón-Villavil. The movement of the detached slabs has been controlled by bedding and transgressive faults. These structures are already developed in the hogback behind Villavil, which at present has not undergone large displacements, suggesting that the sliding process may have begun and is in the quasistatic stage. Field evidence indicates conclusively that the movement mechanism involved frictional sliding on a basal slip surface, followed by a dynamic stage, which could have been a catastrophic event. Rockslide velocity may have exceeded 2 m/s. The transition from the quasistatic to the dynamic stage is thought to be abrupt, possibly resulting from dynamic (earthquake induced) loading. Geomorphic evidence indicates that the landslides have occurred intermittently from perhaps a few thousand years ago to 25 ka.
LARGE-SCALE LANDSLIDES IN ARGENTINA
The relief of Argentina is dominated by the Andes Mountains, a continuous belt of high mountains extending north-south along the Argentina-Chile border. Numerous peaks exceed 6000 m; the highest is Aconcagua (6959 m). To the east, these pass through lower foreland ranges into a low extended plain (La Pampa plains), with isolated higher elevations (Fig. 1).
Settlement is mainly developed in the cattle-agricultural area of the Pampa Húmeda, along the La Plata and Paraná Rivers. This population distribution, the fact that major landslides occur in areas of high relief, and historical patterns of occurrence mean that mass movements are not a natural hazard that commonly affects human life in Argentina.
However, abundant mass movement processes occur in the uninhabited cordillera and other sparsely populated hilly areas. Comparatively few, however, have been described or studied in detail (Fig. 1).
Groeber (1916) described the effects of the outburst of Lake Carri-Lauquén, a landslide-dammed lake in Nequén Province (8 in Fig. 1) in 1914. A massive rockslide had dammed the Río Barrancas, a tributary of the Río Colorado, in prehistoric time. The failure of the landslide dam, and the draining of a large part of the lake that formed behind it, caused a destructive flood in the Río Colorado. The lake emptied almost completely in one night when the surface of the lake dropped ~95 m and 2 × 109 m3 of water suddenly drained into the Río Barranca. The onrush of water scoured the valleys of the Río Barranca and the Río Colorado, resulting in important economic losses because of the destruction of crops and farmland downstream.
Harrington (1946) described and analyzed debris flows in the El Volcan area (1 in Fig. 1), Quebrada de Humahuaca (Jujuy Province). In addition to other studies, Polanski (1961, 1966) established the differences between supposed piedmont glacial diamicton and those diamicts derived from debris flows in Mendoza Province (10 in Fig. 1). Polanski described these deposits, and alerted the geological community to the pernicious consequence of debris-flow activity, with respect to potential loss of human lives and economic and infrastructure damage.
Rock avalanches are the mass-wasting phenomena that are the most dangerous, due to their enormous volume, great velocity, and wide area covered by debris. Examples in Argentina were described in Fauqué (1987) and Fauqué and Strecker (1988). The first work concerns the rock avalanches that occurred in Quebrada de Segovia (6 in Fig. 1), on the southern slope of the Sierra de Famatina (La Ríoja Province). These rock avalanches are compared with those that started from the mountain front and traveled over the piedmont. The different morphology of the latter deposits results from their travel over the piedmont area. The second work made reference to rock avalanches that occurred in the western slope of the Sierra de Aconquija, northern Sierras Pampeanas, Catamarca Province (3 in Fig. 1) and discussed their morphology, mechanics of movement, and genesis.
González Díaz (1972) noted rock-avalanche deposits at the foot of the abrupt fault scarp that defines the eastern limit of Sierra La Palca range, Catamarca Province (5 in Fig. 1). Other examples are the landslide of Vavarco-Campos (Neuquén Province, 9 in Fig. 1), which produced a landslide dam forming Lake Vavarco-Campos in the high sierras of Neuquén Province, and the landslide dam that forms Lago Atuel (Mendoza Province, 11 in Fig. 1), described by González (1979).
Rock avalanches are being studied in Potrero de Leyes (7 in Fig. 1), near San Francisco de Oro (San Luis Province; González Díaz et al., 1997), and in the Yala area (2 in Fig. 1), near San Salvador de Jujuy city (Jujuy Province), where González Díaz and Mon (1996) interpreted the Lagunas de Yala as water-filled depressions of a massive rock-avalanche deposit. In this chapter we discuss the Villavil rockslides in Catamarca Province (4 in Fig. 1).
During field work in the Belén Quadrangle of Catamarca Province in the Sierra Pampeanas of northwestern Argentina (Cortés et al., 1991) it was discovered that the village of Villavil (Fig. 2), situated along the Villavil River in the Belen Department, is practically surrounded by landslide deposits (Fig. 3). The movements have involved Tertiary sedimentary and volcaniclastic rocks that crop out in the Villavil area, forming a landscape of sharp crested ridges and cuestas. The village of Villavil was built in one of the strike valleys between the cuestas and its population of nearly 350 residents is distributed in 75 houses scattered over an area of 82 hectares. The settlers are principally shepherds and farmers; there is also regional textile production.
The little town was settled on a section of alluvial plain not covered by landslides debris; this remnant of alluvial plain exists because the slope located at the southeast of the town has not slid, despite the factors that have favored landslide movement on nearby slopes. As documented in the following, evidence of small and very slow movements that may be the beginning of, or preparation for, later fast movements have been found on the slopes located behind Villavil, raising the question of the vulnerability of the site to future landslides. In parallel with the geological mapping of the Belén Quadrangle, therefore, the threat of landslides at Villavil was studied in order to assess the risk of movements that could affect the town if they occurred.
In this work the geological framework of the Villavil rockslides is presented, the mechanics of their movement are investigated, and the stability of the ridge located southeast of the town is analyzed. We analyze the mechanics of prior movements along the ridge, examine the possible triggers, and discuss landslide velocity and age.
The Sierras Pampeanas forms part of Andean foreland located above an area of the crust where the subducting Nazca plate is nearly horizontal (Jordan et al., 1983). The structural style is that of a folded and thrusted belt showing characteristics of thick-skinned tectonic deformation (Allmendinger, 1986; Jordan and Allmendinger, 1986). The structure is characterized by inverse listric faults of moderate to high angle in the surface that decrease in dip with depth. These faults define the boundary of irregular crystalline basement blocks that form the basin and range morphology of the Sierras Pampeanas. Between these crystalline blocks, Tertiary strata have been tilted, folded, and thrusted. In the Villavil area they form asymmetric cuestas with gentle dip slopes and steep reverse slopes.
The complex deformation history of the area has continued into the Holocene. The occurrence of small earthquakes in historical time and the presence of numerous tectonic scarps that cut young Quaternary deposits are evidence of active neotectonic deformation in the Sierras Pampeanas (Allmendinger, 1986; Costa and Vita-Finzi, 1996).
A belt of Tertiary sediments and volcaniclastics (Fig. 4) crops out between two Precambrian-Paleozoic crystalline basement blocks forming the Sierra de Hualfín to the southeast and the Sierra de Papachacra to the northwest. The Tertiary sequence forms a homocline that dips to the northwest. Villavil is located in a sedimentary basin between en echelon structural blocks that form the eastern margin of the Puna Plateau.
1. A lower unit, the Calchaquense, is 500–700 m thick, and is considered to be Miocene (e.g., Turner, 1973). It is composed principally of gray-red sandstone; intercalations of red pelites and conglomerates are more common toward the top.
2. A middle unit termed the Complejo Volcanico (Volcanic Complex) is 100–200 m thick, and is composed of volcanic breccias, tuff breccia, tuffs of andesitic to basaltic composition, tuffaceous psammite, tuffaceous psephite, and tuffaceous pelite.
3. An upper unit, the Araucanense, is nearly 4000 m thick (Fig. 5). This unit is considered Pliocene-Pleistocene and is composed of tuffaceous sandstones with conglomeratic beds at various levels, and intercalations of tuffs and tuffites. The resistant beds form low scarps on the dip slope, which we refer to as the Araucanense crests.
Although subsequent studies assigned different divisions to the Tertiary rocks in the area (e.g., Turner, 1973; Bossi et al., 1987), we apply González Bornino's initial division of the Calchaquense, Volcanic Complex, and Araucanense to the Villavil slopes. The initial subdivision is simple and the limits of the units are useful in explaining the mechanics of the Villavil landslides.
The eastern side of the Villavil valley consists principally of volcanic breccias, tuffaceous pelites, tuffaceous psammites, and tuffaceous psephites of the Volcanic Complex that dip at the same angle (or steeper) as the slope. These rocks overlie red pelitic beds that form the greater part of the Calchaquense sediments.
The lithological-structural situation described here has resulted in several massive landslides (Fig. 6) on the west side of the ridge, in which part or all of the Volcanic Complex have moved downhill, sliding over the red pelites or the tuffaceous pelite (Figs. 3, 4, and 6), shearing across bedding, and overriding the younger Araucanense. These landslides are similar in form and mechanism to the block-glide landslides described by Braddock and Eicher (1962) and Braddock (1978) from the Front Range foothills of Colorado in the western United States.
Local structure and geomorphological evolution
Villavil is located in the complex structural framework of the northwest Sierras Pampeanas at the southeastern border of the Puna Plateau. It is surrounded by important faults, many of which are still active.
Allmendinger (1986, his Fig. 4) mapped a group of en echelon faults immediately west of Villavil; the faults are 15–30 km long and dip ~17°–55° northwest, and have displacements of ~5 km. These end laterally with plunging drape folds in Tertiary strata.
One of these en echelon faults (Fig. 6) has resulted in a sequence of anomalous morphological features in the landscape, suggesting Quaternary reactivation of this fault as indicated in Figure 7. These features are (1) the existence of the El Jarillal (2200 m above sea level), which is a wide valley but is not occupied by an major stream (Figs. 6 and 7); (2) the headwaters of the El Jarillal valley are in low hillocks of Araucanense that are roughly interposed, closing the depression; and (3) the Villavil River rises at the northwest of the homoclinal ridge of the Volcanic Complex (Fig. 6), fed by the Bolsón and Aguas Calientes Rivers. Immediately to the southeast of Villavil, after passing landslide VI, the river cuts through the homoclinal ridge and drains through a subsequent valley located southeast of it (Figs. 6 and 7).
The question is raised as to why the Villavil River does not cross the low and friable Araucanense hillocks into El Jarillal instead of cutting through the high homoclinal ridge, 400–500 m in height, which consists of the much more resistant Volcanic Complex. The logical answer is that the Villavil River crosses the homoclinal crest, carving a water gap controlled by a fault line (Figs. 6 and 7). A further deduction is that the Río Bolsón valley, which is under a landslide, earlier drained to the Corral Quemado River through the El Jarillal valley (Fig. 7A). It is supposed that the Río Bolsón was captured by the Villavil River, which broke through the homoclinal ridge along the fault.
The differential movements on either side of the fault therefore produced the following effects (Fig. 7): (1) the cutting of the homoclinal ridge; (2) the production of a lateral offset of several thousand meters; and (3) gave rise to the Araucanense hillocks of the El Jarillal headwaters (Fig. 7).
Taking in account these three points, we suggest that the fault has had a vertical component and a strike-slip component.
Almost all of the Catamarca Province is located in a seismic zone where the maximum intensities expected to occur in 50–100 yr are between I and VII (Institute Nacional de Prevecion Sismica [INPRES], 1978). According to INPRES (1993), Villavil is located in an area of moderate seismic activity compared to other areas of the country; e.g., there was a destructive earthquake on October 21, 1966, in the Catamarca Province at Belén, 90 km southwest of Villavil (Fig. 2).
On June 23, 1986, an earthquake with a magnitude of ~4 occurred near Hualfín (Fig. 2); ~100 aftershocks were recorded in the form of microearthquakes (Assumpçao and Araujo, 1993). In Villavil the strongest seismicity during this episode reached intensities IV–V on the Modified Mercalli scale. All the aftershock events had a left-lateral strike-slip motion and occurred in the same fault plane as the main shock (Assumpçao and Araujo, 1993). Their epicenters were quite shallow and were located exactly in the northwest-southeast fault that cuts the Tertiary volcanics and the Paleozoic-Precambrian of Sierra de Hualfín, as mapped by Allmendinger (1986, his Fig. 4) and shown in Figure 6.
DESCRIPTION OF THE VILLAVIL LANDSLIDES
The landslides have been numbered from I to VII (Figs. 3, 6, and 8). We begin the description with those landslides exhibiting the simplest mechanics of movement. This allows a better understanding of the other more complex landslides.
Figure 9 has been modified from Braddock (1978, his Fig. 3). It shows an idealized profile of a block-glide landslide, illustrating the principal terms used in the following descriptions. In these landslides, which consist of a sliding slab that has moved downdip on a composite sliding surface, we can identify four different areas: (1) the updip slab overlying the bedding-plane fault; (2) the folded slab (Braddock's  “riser”), overlying the transgressive fault; (3) the overriding slab, which covers the Araucanense crests; and (4) distal flows.
During movement, the sliding slab becomes folded, forming a syncline and anticline (Fig. 9). Because the deformation takes place in the surface, the rocks have very brittle behavior.
These folds do not result in plastic deformation, but in intense internal fracturing that allows the overriding slab to adapt to the relief over which it moves. The intense internal fracturing also assists in the transformation of the slab into an incoherent and chaotic deposit, which, with further downslope movement, becomes a distal flow.
Landslide VII is relatively small (Figs. 3 and 6); its estimated volume is 15 × 106 m. Like the others, it involves Volcanic Complex rocks that form the ridge crest. It is located ~8 km southwest of Villavil and extends ~ 1 km along strike (Figs. 3 and 6). It is very well preserved, allowing us to appreciate the features that indicate the mechanics of the movement. The geology and structure of the landslide are shown in Figure 10. The Villavil River is separated from El Jarillal valley by a homoclinal ridge, nearly 300 m in height, that dips to the northwest (Fig. 6). The beds that constitute the ridge have a northeast strike and a dip that progressively decreases from southeast to northwest.
The greater part of the steep southeast-facing reverse slope is composed of Calchaquense sediments (Fig. 10) that strike between N12°E and N15°E and dip 29°–30° northwest. The clastic rocks making up the Calchaquense sequence consist of coarse sandstone beds, 2 m in thickness. The top of this sequence, which directly underlies the Volcanic Complex, consists of conglomeratic and sandy beds and alternating red pelitic units.
The crest of the ridge, and almost the entire dip slope, is formed by Volcanic Complex rocks 70–80 m thick (Fig. 10). The strike is similar to that of the Calchaquense and dips range between 251 and 281 northwest. The base of the Volcanic Complex has coarse beds of conglomerate that consist of large granite and andesite blocks enclosed in a pyroclastic matrix. Above these beds, the succession is dominated by welded volcanic breccias in thick (8–10 m) massive beds separated by 1–2-m-thick tuffaceous pelite beds.
In the lower part of the dip slope, the Araucanense overlies the Volcanic Complex, cropping out in a belt nearly 800 m wide (Fig. 10). Morphologically it consists of low crests and ridges of sandstones with alternate fine conglomerate beds striking N18°E and dipping 18° northwest.
The sliding surface is in the top Calchaquense unit (Fig. 10), where the total thickness of the Volcanic Complex has slid on the red pelite beds. The greatest displacement of the landslide, measured parallel to the base of the updip slab, is nearly 290 m, decreasing gradually to zero 1 km northeast in the strike direction.
The folded slab shows intense fracturing. The synform in the riser has limbs that get smaller toward the northeast, i.e., in the direction of decreasing slip. This slip reduction indicates that the downdip slope movement has been very small in this area, and it is probable that the sliding did not produce a transgressive fault, passing instead into a flexure, or buckle, where the movement decreases below a certain threshold (cf. Braddock and Eicher, 1962).
The limbs of the synform have an unequal dip; the southeast (upper) limb dips, as the rest of the sequence, nearly 28° northwest and the west (lower) dips from 5° to 10° southeast, resulting in an asymmetrical fold (Fig. 11). The folding results because the transgressive fault and the bedding fault form a concave curve in profile. The extreme southwest slab that has had greater slip exhibits fracturing along the fold axis. Shattered zones and fractured beds that correspond to the shear zone of the transgressive fault are also exposed.
The front part of the overriding slab that moved across the transgressive fault remained intact after passing over it. As it descended the dip slope, however, it became highly fractured and the disintegrated slab became the source of a distal flow; the flow moved over the irregular topography formed by Araucanense crests that form the lower part of the ridge (Fig. 10). In this way, the movement is transformed in a downslope direction from block sliding to debris flow.
Landslide II (Figs. 3, 6, 8, 10, and 12) is much larger than landslide VII, but it has had similar behavior. The northeast extreme of the landslide has slipped nearly 720 m in two stages (Fig. 10). The slip gradually reduces to 0 at 1.5 km in a southwest direction (Figs. 3, 6, and 8).
An interesting feature observed in this landslide is that sliding has occurred on at least two distinct sliding surfaces located at different stratigraphic levels (Fig. 10), suggesting that the landslide was produced in two stages. In the first, the whole thickness of the Volcanic Complex slid over the upper Calchaquense. After the initial landslide a new slip surface was formed, located higher in the stratigraphic sequence in the Volcanic Complex (Fig. 10), and favored by the presence of thin, fine-grained pyroclastic deposits located between the stronger pyroclastic breccia beds. This landslide characteristic is very important because it indicates that new slip surfaces could be produced after an initial landslide, and in this way its mobility could be increased (Fig. 10). This process is particularly favored by the Volcanic Complex lithology, wherein dense, massive, breccias (welded pyroclastic flow deposits), 10–15 m thick, alternate with thin (1.5–2 m) beds of tuffaceous pelites. The latter beds act as Volcanic Complex slip surfaces.
This situation makes it difficult to estimate the mobilized volume; however, taking into account all variables, we estimate a slide volume of nearly 184 × 106 m3 for landslide II.
Another interesting feature in this landslide is the different morphology adopted by the sliding slab as a function of its slipping distance. The extreme northeast part of the sliding slab moved nearly 720 m on the sliding surface (Figs. 3 and 10), but the slide mass moved 1250 m into the valley because of bulking of the debris.
This landslide mass obstructed the valley, building a thin ridge across it. Later, the Río Bolsón opened a gap in the obstruction that is now a narrow, 200-m-wide passage known as La Angostura (Figs. 3 and 12). In both sides of La Angostura, the Volcanic Complex breccias contained in the debris are very fractured. After the landslide, the river made its way through them rather than the Araucanense sandstones that form the eastern end of the interposed ridge in the valley. Observing La Angostura from the northeast, we can see how the Volcanic Complex that slid overlies the younger Tertiary Araucanense, generating a very smooth angular discordance between both of them. The overthrusted beds have a 45° northwest dip with a N251E strike; the high dip is due to the landslide.
The northeast extremity of the slide slab is characterized by a very smooth synform-antiform pair, the axes of which are nearly perpendicular to the movement direction (Fig. 10).
As the slip distance decreases toward the southwest (Figs. 3 and 8), the slide mass constitutes only a synform fold with northwestern flanks that rise and form a slightly east dipping crest (Fig. 10). This synform fold gradually disappears while the slip decreases toward the southwest. Movement has formed a lateral scarp >200 m high (Figs. 3 and 8) exposing Volcanic Complex slabs that slid and became folded while overriding the lower Araucanense outcrops. These slabs have a dip of nearly 40°.
The movement of the sliding slab in landslide II reduces to zero before arriving at the lateral breakage wall of landslide III; this may explain why transgressive faults are not observed in this cliff.
Landslides III, IV, and V
These landslides (Figs. 3, 10, 13, and 14) form a continuous segment of slab-sliding that extends for 5 km parallel to strike along the dip slope of the hogback (Figs. 6 and 8). They have several similarities, and for this reason we describe them as a group. In these landslides, the total thickness of the Volcanic Complex (80 m) has slipped over the top of the Calchaquense.
The width of the slid areas, multiplied by this thickness, determines great mobilized volumes: landslide III, 247 × 106 m3; landslide IV, 375 × 106 m3; landslide V, 243 × 106 m3; producing a total of 865 × 106 m3 for all three landslides.
Nature of the sliding surface.
The sliding surface in these landslides (Figs. 9 and 15) coincides with a planar bedding-plane fault. Gullies have been cut into the sliding surface parallel to it, extending to the high part of the ridge crest. The gully intensity is notable in landslide III (Fig. 8), the oldest landslide of the three, where it has practically resulted in the destruction of the sliding surface. Good exposures of Calchaquense pelites, located below the bedding fault, are found in the sides of the gullies that cut through the sliding surface.
The sliding surface of landslide V, the youngest of the three, shows almost no gullying. There are few exposures of the lower Calchaquense pelites; in its place the zone is littered by great Volcanic Complex blocks, which stand out through a colluvium and soil mantle. It is inferred that these blocks were fragmented during movement.
The sliding surface of landslide IV shows intermediate characteristics between those of landslides III and V. Gully development is of less intensity than in landslide III, but more intensive than in landslide V. The sliding surface is partially eroded and the mantle of superficial blocks is less compared to landslide V; this is because gully development undermines the support of many blocks, resulting in their disintegration.
The relative age of these landslides can be estimated by the degree of erosion of the sliding surface, the riser, and the distal flow deposits. According to these criteria, the most ancient is landslide III, next is landslide IV, and the youngest is landslide V. This age sequence suggests that the movements are younger toward Villavil.
With reference to the nature of the sliding surface, it is important to note that in most cases it consists of Calchaquense red pelites. However, in some cases the sliding surface consisted of very compact, thin beds of tuffaceous pelite. These beds form the boundary between the Calchaquense pelite and the overlying resistant breccia beds of the Volcanic Complex.
Folded slabs (riser).
The folded slabs that remain over the bedding and transgressive fault form a synform. The location and the topographic expression of these change according to the specific landslide and depend on several factors, i.e., the dip of the beds and the bedding fault, the dip of the transgressive fault, the position of the transgressive fault on the slope (i.e., the height above the valley floor), and the slip distance of the sliding slabs.
The trace of the synform marks the approximate position of the transgressive fault. Another characteristic of the folded slab is intense fracturing produced during the slipping and folding in superficial conditions. This response to movement is consistent with a brittle material behavior. During slope movement, differential displacement occurs on vertical joints in the Volcanic Complex, forming numerous blocks.
In landslides IV and V the synform is located in the lower half of the slope, but near the middle part of it. In landslide III, the syncline is located lower. This situation, together with the slip distance (minor in landslide III) and the transgressive fault angle, results in the fact that in landslides IV and V, rather tight and symmetrical folding has occurred. In landslide III, the folding is more open and is asymmetrical; the southeast limb overlying the bedding fault is very long, and the northwest limb, forming the raised ridge over the transgressive fault, is very short and tilts less.
Another notable characteristic in this folded slab region is the different degree of surficial erosion that the slabs exhibit as a function of their age. For example, if we compare the folded slabs of landslides IV and V, the first one shows a drainage network, developed transverse to the direction of movement, in which the principal watercourse is parallel to the synform axis and its tributaries run over the flanks parallel to the direction of movement. The synform of landslide V, the youngest landslide, is crossed by some water courses running parallel to the movement.
Another superficial feature is related to the topography of the folded slab. In landslide V it shows a great quantity of small lineaments transverse to movement resulting from the activation of vertical joints by compression during folding. This explains the presence of several superficial blocks on this slab. In landslide IV, because of weathering and superficial erosion, these features have been attenuated, the surface is gentler, and blocks are not present.
With respect to the overriding slab, the comparison between landslides is very difficult, because common to all of them is the intense fragmentation and deformation that the overriding slab generally undergoes when displaced over an irregular topography formed by the Araucanense crests. The morphology is largely determined by the characteristics of the relief of the Araucanense overridden by the landslides. When the slab slides (or their flows) totally cover the Araucanense crests, the apparent landslide thickness is greater than the real thickness, because the deposits form high, ~300 m, promontories.
The part of the folded slab that forms the synform over the bedding and transgressive faults keeps its continuity and preserves a rough stratification despite intense fracturing. The overthrusted slab, however, shows more deformation and a higher degree of fragmentation.
The boundary between the overthrusted slab and the folded slab that forms the synform is well marked by topographic depressions. These areas represent zones of extension generated during the movement leading to the development of a graben, which disconnects both sides of the slabs.
The location of these topographic lows would coincide with the antiform position if the synform-antiform pair had formed. The mobilized slab first suffers a superficial compression and a deep tension when sliding on the transgressive fault and then a superficial tension and a deep compression in the area of the antiform location area (Fig. 9). The brittle material behavior supports the first synform folding effect, but when the second antiform folding is produced, there is already too much fracturing to support the continuity of the slab. A graben is generated by traction, represented by the topographic lows that disconnect the folded slab (riser) from the overthrusted slab. In both slabs a great quantity of small faults, that in many cases are similar to slumps, may be seen near the lows. The topographic lows have importance as aquifer-recharge areas that discharge at the surface near to the water intake of Villavil.
The deformation of the Volcanic Complex sliding slab also results from shearing stress parallel to the surface that produces the shear planes that facilitate slab delamination along its bedding plane.
The distal flows are voluminous (Fig. 16). Overriding the Araucanense crests, the material mobilized in the overthrusted slabs disintegrates and moves toward the valley bottom like a debris flow. The flow consists of a great quantity of separate blocks defined by original structural elements, represented mainly by vertical joints and bedding planes. The debris-flow matrix consists of rock ground by basal shearing during movement, together with some Araucanense clastic debris stripped off and incorporated by the overriding landslide.
It is interpreted that the distal flows correspond to the final stage of the movement. The surfaces of the distal flows consist of a chaotic distribution of blocks, some of which have volumes >10 m3. The gullies at the distal parts of the flows undermine support of the blocks, resulting in them becoming mobile. This constitutes a serious risk for some residences located near Provincial Road 43 at the foot of the flows (Fig. 16).
The distal flows are best exemplified in landslide V; they are more difficult to observe in landslides III and IV, although they are present in these landslides. Landslide IV completely obstructs the Río Bolsón valley (Figs. 6 and 8). In Cuesta de Indalecio, it can be seen how the distal flows of this movement, ~50 m thick, overlie the Araucanense strata on the northwestern slope of the valley.
This landslide is relatively small (Figs. 3, 6, 8, and 17). It has mobilized nearly 34 × 106 m3, which has generated a distal flow that advanced 620 m over the valley side, covering Araucanense outcrops that overlie the Volcanic Complex in the low part of the slope.
The landslide has the following characteristics that distinguish it from all of the other landslides.
1. Not all the thickness of the Volcanic Complex has slid, but only the upper 30–35 m. This confirms that there are sliding planes within the Volcanic Complex. The sliding surface corresponds to a thin, very compact purple tuffaceous pelite bed that strikes N12°E and dips 30° NW.
2. The lateral breakage walls curve inward downslope because the slid material goes through a bottleneck, narrower than the crown or breakage zone.
3. This bottleneck is probably the result of the transgressive fault that allows the movement; it has a smaller overrun than the amplitude of the breakage zone.
4. The slide material undergoes a great compression when passing the bottleneck, pulverizing the mass and causing a distal flow downslope from the bottleneck.
5. It begins with a block-glide landslide and ends as a debris flow. Folded and overthrusted slabs are absent, in contrast to landslides III, IV, and V.
6. Next to the lateral breakage wall located to the northeast, there is a great glide block, supported over the slide surface, that stays because it was only mobilized 200 m.
This landslide, nearly 500 m north of La Angostura, is small, impinging on the alluvial plain by no more than 100 m (Figs. 3 and 6). It is a very ancient movement, and erosion has removed the distal deposit and badly drawn its breakage zone.
Braddock (1978) suggested that block-glide type of landslide has two stages in its evolution; the quasistatic stage, in which the transgressive fault and the bedding fault develop as a consequence of downdip creep, and the dynamic stage, which corresponds to the movement of the sliding slab along the faults. The occurrence of the dynamic stage is therefore conditioned by the presence of the quasistatic stage and predynamic stage-limited displacements. Hazard assessment of unfailed slopes at Villavil initially involved the detection of creep movements associated with the quasistatic stage.
In the unfailed Volcanic Complex slopes located behind Villavil, the transgressive fault and the bedding fault have been detected. A similar set of structures could be seen in the lateral breakage wall located in the south lateral scarp of landslide V (Figs. 8 and 18). The transgressive fault has a throw of no less than 5 m. A fault breccia resulting from the displacement along the fault plane is present. Other evidence of movement on this fault is in the degree of fracturing present in the upthrown block, as compared with that of the downfaulted block (Fig. 18). This results from the slow and imperceptible movement that the slab has undergone.
An important feature associated with these faults is that the intersection of the fault plane with the dip-slope surface appears as a small shoulder produced by the transgressive fault throw. Three of these morphologic shoulders have been detected in the slope behind Villavil (Fig. 14), where they interrupt the dipslope surface. It is also probable that a third morphologic shoulder exists in the upper part of the slope, but the enlargement of the gulch in this region has obscured the feature due to erosion (Fig. 14). In exposures on the side of the gulch that dissects the homoclinal ridge, it was observed that the morphologic shoulder coincided with brecciated Volcanic Complex rocks, where subvertical beds deformed by trangressive fault drag can be seen.
Here it is noted that the throw of a few meters observed in the transgressive fault implies that the transgressive fault passes under the riser to become a bedding fault beneath the updip slab.
Two morphological shoulders that result from the intersection of the transgressive fault with the dip slope have also been located in the ridge between landslide VI and the Villavil River gorge (Fig. 6). A landslide in this area could dam the river and cause upstream flooding inundating Villavil.
The presence of these morphological shoulders is evidence that limited block-glide movement has begun (i.e., quasistatic stage) in the Volcanic Complex slab located behind Villavil and that the conditions are set for the total rupture of the dip slope associated with the dynamic stage.
The dynamic state, according to Braddock (1978), could develop in two ways. One of them corresponds to a viscous creep, in which case the landslide could be more or less slow, mobilizing at velocities of millimeters or centimeters per year, depending on the viscosity of the slip-surface materials and mobilized mass thickness. In this case the transition from the static state to the dynamic would be imperceptible. In contrast, the other way in which the dynamic state can occur would be by frictional sliding. In this case the movement velocities could be very high, 5–10 m/s (Braddock, 1978), and the event would be catastrophic. In this second case the transition from the static state to dynamic may require a trigger.
In this context, it is important to note that INPRES (1993) defined the Villavil area as one of moderate seismic activity compared to other regions of Argentina. As noted here, Assumpçao and Araujo (1993) located the epicenter of the 1986 Hualfín earthquake exactly at the northwest-southeast geologic fault trace that cuts the homoclinal ridge near Villavil (Fig. 8).
Mobility of the Villavil landslides
Landslide mobility, or travel distance, is essential information in the planning of disaster prevention. It is important in the analysis of the slide mobility in Villavil.
In the Villavil landslides, the vertical distance of the mass fall (H), although not the same, is comparable in all of the movements (Table 2). In spite of this similarity, landslides II and VII have different travel distances (L) that depend on the following characteristics.
1. The direction of sliding of the sliding slabs, whether parallel or oblique to the dip, has a major influence on its mobility. Landslides III, IV, and V that slid parallel to dip exhibit higher mobility (Table 2). However, movements that show an oblique slip over the sliding surface (landslides II and VII) had much less mobility (Table 2). This difference in the direction of sliding appears to depend on the plunge of the transgressive fault.
2. The angular value of the dip of the transgressive fault is also very important; when nearer to the horizontal, movement is much more favored and mobility is increased. For example, the low mobility of landslide VII probably reflects two factors; a small mass volume and the high dip angle of the transgressive fault, which obstructs the overriding sliding slab, limiting its downslope movement and decreasing its mobility. In this landslide the greater part of the sliding slab remains on the transgressive and bedding faults and the distal flows are therefore very limited.
3. Mobility is generally in direct relation to slide mass volume (Fig. 19). Landslides VI and VII, of small volume, also have low mobility (Table 2). However, the largest landslide (landslide IV) is also the one that has the highest mobility (L = 2300 m). It totally buried the valley at the latitude of Cuesta de Indalecio (Fig. 8).
|Landslide .||Volume (x 106m3) .||H (m) .||L (m) .|
|Landslide .||Volume (x 106m3) .||H (m) .||L (m) .|
Note: L is length; H is height.
If we represent the volume of a possible failure in the Villavil slopes (156 × 106 m3) using a plot of H/L versus volume, the travel distance (based on the geometry of the other landslides and given H = 500 m) reached by the landslide could be 1800 m (Fig. 19). With this runout distance, the village of Villavil will be overridden.
AGE OF THE VILLAVIL LANDSLIDES
Knowledge of the age of landslides can be useful in assessing landslide risk. We have analyzed the relative age of one movement with respect to the others and have obtained an absolute age by radiocarbon dating.
The relative age of the landslides is obtained by comparing surface morphological characteristics such as erosion, or degree of preservation, of the debris and the sliding surfaces of the individual landslides. In landslides III, IV, and V, which are contiguous with one another, this relative age indicates that the most modern movement is landslide V and the most ancient is landslide III. This conclusion is of concern because it indicates that through time the movements are getting closer to the site of Villavil.
|Landslide||Thickness (m)||Elevation at crest (m)||Elevation at base (m)||Runout into the valley (m)||Displacement on slip surface (m)|
|Landslide||Thickness (m)||Elevation at crest (m)||Elevation at base (m)||Runout into the valley (m)||Displacement on slip surface (m)|
The absolute age is difficult to obtain because no organic material suitable for radiocarbon age dating has been found in the intensely disturbed coarse deposits of the distal flows. However, organic material was recovered from lacustrine-fluvial deposits resulting from the damming of the Río Bolsón valley. Its dating provides the basis for an approximate estimation of ages for landslides IV and V.
Landslide IV totally blocked the Río Bolsón valley (Fig. 8) in the Cuesta de Indalecio area. After this closure, a body of water was formed that began aggradation in the Río Bolsón valley. The stream found an outlet through Araucanense deposits at the right margin of the valley, and the lake drained. The aggradation formed a flat flood plain, 2 km wide, crossed by a sinuous stream (Figs. 6 and 8). The change in slope decreased the river competence and, for that reason, only fine sand and silts were deposited in the backwater zone in Cuesta de Indalecio.
Later, downstream of this closure, landslide V also produced a dam, now cut by subsequent erosion. As a result of the last river damming, lacustrine clays and silts were deposited, grading up to a fluvial sandy-silty deposit. The temporary dam formed a body of water that, after being filled by deposits, became an alluvial plain consisting of fine material resulting from the decrease in river competence.
After the cutting through of the closure by erosion, these deposits were incised and now form levels of river terraces downstream of the Cuesta de Indalecio (Fig. 20). Carbonaceous material, coming from indigenous ceramic and bone remains and buried dry walls, were found in the terrace. Radiocarbon dating gives an age of 1432 ± 132 yr.
Taking this age in account we may try to obtain an approximated age of the landslides IV and V. Figure 21 is a schematic profile that shows the location of the dated carbonaceous material in the backwater deposit. The carbonaceous remains are at a level of 4.8 m below the surface of the deposit, which is 12 m thick. The 1432 ± 132 yr then represents the time required for the deposition of the 4.8 m of sediments and the time for erosion down to the actual position of the fluvial stream.
If we assign arbitrarily a period of 1 k.y. for the deposition of 4.8 m of sediments and 432 yr for the riverbed erosion, then to deposit the total 12 m thickness would require roughly 2.5 k.y. If we add to this the 432 yr that we supposed for erosion, we obtain a very tentative age for landslide V of nearly 3 k.y.
With this single radiometric date, we try to evaluate the age of landslide IV. The difference in elevation between the alluvial plain of Villavil River and the Río Bolsón is ~130 m; the greater part of it must be the result of the aggradation deposited after the damming by landslide IV.
We are going to consider arbitrarily the following.
1. In the 130 m difference in elevation, 100 m is due to aggradation and 30 m would reflect the previous relief of the valley.
2. The sedimentation velocity is the same as previously assumed with reference to landslide V, i.e., 12 m of backwater sediments accumulated in 2.5 k.y.
Then with this sedimentation rate, 100 m of sediments requires 25 k.y. to accumulate. This will approximate the hypothetical age of landslide IV.
Very rough estimates of the landslide ages indicate that the recurrence of the movements along the homoclinal ridge could span great lapses of time, perhaps thousands of years.
The principal conclusion herein is that the Volcanic Complex slab located on the dip slope behind the village of Villavil presents features that strongly indicate its potential instability. This could result in a landslide that would bury the village. This conclusion is supported by the following field observations. (1) In the vicinity of Villavil there are seven large-scale block-glide landslides, indicating conditions very favorable for the development of this type of slope movement. (2) The lithological and structural conditions that favored the occurrence of the seven movements located of both sides of Villavil are the same that are present in the Volcanic Complex slab located behind the town. (3) In the Volcanic Complex slab located behind Villavil, transgressive and bedding faults have been detected, indicating that prefailure movements have already taken place. The lateral breakage wall of landslide V allows the observation of these structures that, according to Braddock (1978), are evidence that the movement has concluded its quasistatic stage. (4) In the lateral breakage wall it can also be observed that rock fracturing and deformation has occurred in the riser. This is further evidence that small movements have already taken place.
In the landslides that have occurred previously, the dynamic state has involved frictional sliding on weak beds within the Tertiary sequence. Movement was probably rapid, and seismic shaking is suggested as the initiating trigger. The Villavil area has been subject to considerable historical seismicity. Landslides do not occur along the entire length of the homoclinal crest, but are concentrated in the vicinity of the fault that cuts and displaces the crest. The epicenter of the June 1986 Hualfin earthquake, that was felt in Villavil and reached IV to V on the Modified Mercalli Scale, has been located on this fault.
If a landslide similar in magnitude to those that have already occurred took place, it could have a velocity of ~2 m/s and its distance of travel could be 1800 m, which would bury Villavil.
In spite of the fact that the age dating is not conclusive, it is very probable that the Villavil rockslides are ancient (pre-Colombian). However, because the geological-structural and seismic conditions do not appear to have changed, movements may recur.
Provincial authorities are studying these results and the possibility of land-use restrictions being implemented in the vicinity of Villavil is being considered.
We thank Néstor Alsina for pleasant collaboration and consideration in several matters, and the Servicio Geológico Minero Argentino for permission to publish this work. We are grateful to the people of Villavil who contributed to this work directly or indirectly. We also sincerely thank the editors and reviewers of this volume for their extraordinary patience and courtesy during the preparation of this work.
Figures & Tables
|Landslide||Thickness (m)||Elevation at crest (m)||Elevation at base (m)||Runout into the valley (m)||Displacement on slip surface (m)|
|Landslide||Thickness (m)||Elevation at crest (m)||Elevation at base (m)||Runout into the valley (m)||Displacement on slip surface (m)|
>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.