The Pico de Navas slump (Burgos, Spain): a large rocky landslide caused by underlying clayey sand

The Pico de Navas slump (Burgos, Spain): a large rocky landslide caused by underlying clayey sand E. Sanz-Pérez1, I. Menéndez-Pidal1, A. Lomoschitz2, R. Galindo-Aires1,* 1 Escuela Técnica Superior de Ingenieros de Caminos, Canales y Puertos. Universidad Politécnica de Madrid, Spain. 2 Escuela de Ingenierías Industriales y Civiles. Universidad de Las Palmas de Gran Canaria, Spain. e-mail addresses: [email protected] (R.G-A, corresponding author), [email protected] (E.S-P), [email protected] (I.M-P), [email protected] (A.L). Received: 31 July 2015 / Accepted: 1 March 2016 / Available online: 30 April 2016 Abstract The Pico de Navas landslide was a large-magnitude rotational movement, affecting 50x106m3 of hard to soft rocks. The objectives of this study were: (1) to characterize the landslide in terms of geology, geomorphological features and geotechnical parameters; and (2) to obtain an adequate geomechanical model to comprehensively explain its rupture, considering topographic, hydro-geological and geomechanical conditions. The rupture surface crossed, from top to bottom: (a) more than 200 m of limestone and clay units of the Upper Cretaceous, affected by faults; and (b) the Albian unit of Utrillas facies composed of silty sand with clay (Kaolinite) of the Lower Cretaceous. This sand played an important role in the basal failure of the slide due to the influence of fine particles (silt and clay), which comprised on average more than 70% of the sand, and the high content presence of kaolinite (>40%) in some beds. Its microstructure consists of accumulations of kaolinite crystals stuck to terrigenous grains, making clayey peds. We hypothesize that the presence of these aggregates was the internal cause of fluidification of this layer once wet. Besides the faulted structure of the massif, was an important factor for the occurred landslide. Other conditioning factors of the movement were: the large load of the upper limestone layers; high water table levels; high water pore pressure; and the loss of strength due to wet conditions. The numerical simulation of the stability conditions concurs with our hypothesis. The landslide occurred in the Recent or Middle Holocene, certainly before at least 500 BC and possibly during a wet climate period. Today, it appears to be inactive. Due to mineralogical features of involved material, facies Utrillas, in the landslide, the study helps to understand the frequent slope instabilities all along the Iberian Range where this facies is present. Keywords: deslizamiento rotacional, Cordillera Ibérica, facies Utrillas, Cretácico, simulación numérica Resumen El deslizamiento de Pico de Navas fue un movimiento rotacional de gran magnitud, que involucró 50x106 m3 de rocas. Los objetivos de este trabajo han sido: (1) caracterizar el deslizamiento a partir de la Geología, los elementos geomorfológicos y los parámetros geotécnicos; y (2) obtener un modelo geomecánico que permita explicar adecuadamente la rotura, teniendo en cuenta la topografía, previa y posterior al movimiento, y las condiciones hidrogeológicas y geomecánicas. La superficie de rotura atravesó, de arriba abajo: a) 200 m de calizas y arcillas del Cretácico superior, afectadas a su vez por fallas, y b) la unidad del Albense de facies Utrillas, compuesta de arenas limosas con arcilla (caolinita) del Cretácico Inferior. Las facies Utrillas tuvieron un papel importante en la rotura basal del deslizamiento debido a la influencia de las partículas finas (limos y arcillas) que alcanzan el 70% de media y el alto contenido en caolinita, >40% en algunos niveles. Tienen una microestructura particular: al microscopio electrónico se ven paquetes de cristales de caolinita pegados sobre los granos de terrígenos, formando agregados porosos. Se propone que la formación de estos agregados sea la causa intrínseca de la fluidificación de los niveles de arenas caoliníticas del Albiense. Entre los factores condicionantes del movimiento destacan: la estructura fallada del macizo, la gran carga vertical trasmitida por las capas de calizas superiores; los elevados niveles freáticos, la alta presión de poros y, en consecuencia, la disminución de los parámetros resistentes. Journal of Iberian Geology 42 (1) 2016: 55-68 http://dx.doi.org/10.5209/rev_JIGE.2016.v42.n1.49120 www.ucm.es /info/estratig/journal.htm ISSN (print): 1698-6180. ISSN (online): 1886-7995 JIGE 42-1SEGURIDAD.indb 55 19/05/2016 14:21:22 56 Sanz-Pérez et al. / Journal of Iberian Geology 42 (1) 2016: 55-68 1. Introduction Rotational landslides are very frequent movements of slope worldwide, which largely affect cases of clayey materials with some degree of consolidation. The small landslides usually have a clearly defined circular morphology, especially if the materials that make up the slope are homogeneous. A single rotational slide or slump is a “more or less rotational movement, about an axis that is parallel to the slope contours, involving shear displacement (sliding) along a concavely upward-curving failure surface, which is visible or may reasonably be inferred” (Varnes, 1978). The sliding surface need not be perfectly circular. An additional feature of a genuine rotational slide is the small degree of internal deformation of the displaced material, which distinguishes it from flow-like mass movement types, although sometimes soil slump material liquefies and transforms into a flow at its downslope end. Rotational mass movements can vary from terracettes with an area of only a few square meters to large complexes of several hectares. Great slopes (tens of meters) often produce complex movements. Sharpe (1938) called this form a slump-earth flow, and Varnes (1978) provided an idealized diagram showing the features of this complex slide, that usually affects engineering soils (debris, predominantly coarse; or earth, predominantly fine). For instance, La Conchita Landslide, Southern California occurred in 1995 is a good example of these complex (rotationaland flow-type) slides (Turner and Schuster, 1996; USGS, 2004). Large rotational landslides are frequent on UK coastal cliffs and occasionally they affect layers of ‘weak rocks or strong soils’ under saturated conditions. Stratigraphic series also includes rocky strata (gypsum, limestone, marl, sandstone). For instance, St Catherine’s Point landslide, Isle of Wight, is a good example of a major bedding controlled slide (Hutchinson et al., 1991). The sliding surface is located in a thin clay layer, which is below sea level and the geological sequence is primarily mid to Lower Cretaceous (Bromhead and Ibsen, 2007). The Pico de Navas slump (Burgos, Spain) in the Iberian Cordillera can be compared in some way to the two models of landslides referred to above (earth slide-earth complex flow model of Varnes -1978-) and St. Catherine’s Point landslide model (Hutchinson et al., 1991) as some of its features are similar. Typical and very distinctive Utrillas facies materials are notably known for their extensive presence in the Iberian Peninsula. Their characteristics have been the topic of copious research throughout the twentieth century. So we can quote references from its very first name by Tricalinos (1928) to Hanne (1930), Brinkmann (1931), Richter and Teichmuller (1933), Aguilar et al. (1971), Fallot and Bataller (1927), Canerot et al. (1982), Vilas et al. (1982), Floquet et al. (1982), Alonso (1981), among others, indicating the importance of studying of this type of formations. The Utrillas facies which dates from the Albian (last unit of the Lower Cretaceous) is very common in the Cretaceous geological formations of mainland Spain. It mainly consists of sand, silt and clay minerals from the kaolin group and it shows a very characteristic variety of colours. Its extensive presence in the Iberian Peninsula, the constancy of its lithological characteristics, and the geotechnical behaviour of its deposits (which are clearly unfavourable to the stability of natural and artificial slopes), have made it infamous as a natural hazard and very problematical in terms of civil works constructions. Menéndez-Pidal (2006) lists more than thirty locations in Spain where difficulties arise in the execution and exploitation of large engineering works on this facies over the twentieth century. It is highly significant that, in nearby terrains of this facies, serious problems have been caused due to instability, such as the case of the roadside cuttings of the Soria to Burgos highway, the Puerto del Mojón Pardo (Navaleno, Soria), and the Santander Mediterranean railway in the town of Navaleno (Soria) (some 15 km from Pico de Navas) when it was constructed in the nineteenth century. The grain size and mineralogy of the Utrillas facies in the Pico de Navas is almost the same. Deposits around Navas del Pinar were tested to estimate the fluidity capacity of the sands, by observing their behaviour in moulds loaded with water (bucket tests). These tests showed that, with little water, the material was extremely fluid and completely liquefied. The probable causes of the Pico de Navas landslide should be related to the geomechanical features of the Utrillas facies (Albian, Lower Cretaceous) and specific hydrological conditions in the area that ought to facilitate the movement. In fact, numerous civil engineering constructions all along the Iberian Range have been affected by the potentially plastic behavior of these facies causing the failure of many slopes, mainly after rainfall periods (Menéndez-Pidal, 2006). The mechanisms of rainfall induced landslides have been extensively studied and some of the conclusions assert that El modelo geomecánico ha permitido simular las condiciones de estabilidad y rotura del deslizamiento, cuyos parámetros y condiciones coinciden con las anteriormente indicadas. El deslizamiento se produjo en el Holoceno Reciente o Medio, con certeza antes del S.V AC y posiblemente durante un periodo climático húmedo. Ahora el deslizamiento es inactivo. Este trabajo ayuda a entender las frecuentes inestabilidades de laderas y taludes asociados a las facies Utrillas, a lo largo de la Cordillera Ibérica. Palabras clave: deslizamiento rotacional, Cordillera Ibérica, facies Utrillas, Cretácico, simulación numérica JIGE 42-1SEGURIDAD.indb 56 19/05/2016 14:21:22 57 Sanz-Pérez et al. / Journal of Iberian Geology 42 (1) 2016: 55-68 the amount of rain and weathering are the major factors predisposing a slope to failure (Iverson, 2000; Corominas, 2006; Msilimba and Holmes, 2010; Wang et al., 2002; Sassa et al., 2004; Guzzetti et al., 2008). Water infiltration is a significant triggering factor for slope failures around the world. This paper is focus on four main objectives on the landslide: (1) geological and geomorphological features; (2) geomechanical characterization of the Utrillas facies (Albian, Lower Cretaceous) and interpretation of the hydrogeological conditions in the area around the slide; (3) study of the age and state of activity; (4) reconstruction of the original slope, analysis of its failure by mean of adequate numerical software (back analysis) in order to explain the possible causes and behavior of the movement. 2. Methods A variety of methods and techniques have been used to study this landslide, which can be gathered as: desk work; field work and laboratory tests (Table 1). Firstly, desk work included: a previous review of scientific literature and maps related to the study area; geo-interpretation of aerial photographs; drawing of a geological map on scale 1:5000; stability analysis of the slope using FLAC program and interpretation of the whole data we have got. Secondly, field work consisted of: geological surveys for the recognition of the landslide area and the surroundings, including a description of geomorphologic features and the characterization of geological formations; an electrical geophysical survey (Shlumberger method) with 3 lines of about 150 m long and a boring survey with three boreholes (30-50 m long) with adequate sampling, SPT (Standard Penetration Test) and logging. In addition, 10 samples of silty and clayey sands of the Utrillas facies were specifically taken from the stable bedrock out the landslide (former kaolin quarry at Navas del Pinar). Besides, an archaeological survey was done in the area because the existence of an old Celtic-Iberian fortress on the top of the upper scarp. We recognized the archaeological remains of this ancient building and its characteristics to achieve its age, which have to be previous to the landslide age (Table 1) And thirdly, laboratory tests were mainly focused on the geomechanical and mineralogical characterization of silty and clayey sand of the Utrillas facies, which corresponds to the main layer affected by the sliding surface (Table 1). They were: 10 Particle size tests of soils by sieve analysis (ASTM D-422-63, 2007) and 3 by ultrasound spectroscopy; 9 Standard tests for Liquid Limit, Plastic Limit and Plasticity Index of soils (ASTM D4318) and 4 Direct shear tests (Consolidated and drained) (BS 1377: Part 7:1990) which were performed without drainage to determine the internal angle of friction (φ) and cohesion (c) of the materials involved in the movement. In addition, the mineralogy of the samples was studied using SEM (Scanning Electron Microscope) technique (Table1). 3. Geological and Geomorphological Setting The Sierra de Navas (Burgos, Spain) forms part of the western end of the massif of the River Lobos Canyon, which comprises Cretaceous materials belonging to the northern zone of the Iberian Cordillera of northern and central Spain. The form of an inverted ship’s prow is a characteristic shape of mountain promontories in the Cretaceous calcareous region of this sector (Pico Frentes, Peña Type of work Description Desk Work Aerial photographic image interpretation. Drawing of geological map (scale 1:5,000). Stability analysis with FLAC software. Laboratory techniques and tests 10 Particle-size analyses of soils. 3 Particle-size (ultrasound method). 9 Standard tests for Limit Liquid, Plastic Limit and Plasticity Index of soils. 4 Determination of shear strength by direct shear box, CD test type (consolidated and drained). Mineralogical characterization by SEM method (Scanning Electron Microscope). Field Work Geological and geomorphological recognition and mapping. Electrical Geophysics (Shlumberger method) 3 lines, 150 m long. Boreholes (3), 30-50 m deep, SPT (Standard Penetration Test), logging and sampling; 10 specific samples from Utrillas sand. Archaeological survey. Table 1.- Summary of methods and techniques used to study the Pico de Navas landslide. Fig. 1.- Location of the Pico de Navas Landslide (Burgos, Spain). JIGE 42-1SEGURIDAD.indb 57 19/05/2016 14:21:22 58 Sanz-Pérez et al. / Journal of Iberian Geology 42 (1) 2016: 55-68 normal in type and give rise to small fault throws (Figs. 2, 3). The geometry of these structures determines that on the slope were predominantly present soft marls and clays (base of the Cretaceous) rather than the upper calcareous series, which are stronger and outcrops in the elevated NW block of the Pico de Navas fault. 4. The Pico de Navas landslide 4.1. Landslide features and dimensions The features of this landslide are very conditioned by geological structures and by hydrological aspects. On one hand, it was developed on the axis zone of a syncline which also has a number of SW-NE parallel faults. And on the other hand, the hydrogeology had a direct influence on the movements of the slide, not only in the rupture surface but also in the erosion of the foot. The area of the landslide is about 1.25 km wide plus 0.8 km long on average, i.e. 1 km 2 in extension and it can be divided in three different zones (Figs. 3, 5): upper, middle and lower zones, which are about 0.15; 0.65 and 0.2 km 2 in extension, respectively. (a) The upper zone has a main escarpment, 60-70 m high and about 1,600 m long, coincides superficially with a SWNE fault, corresponding to the main and upper scar of the landslide (Fig. 3). Nevertheless, the fault and the sliding surface have different traces once their surfaces are followed inside the rocky massif. The crown is very near to Pico de Navas peak (height = 1,351 m) At the base of the escarpment there is a tilted rock mass composed of Turonian beds, broken into large and fragmented blocks due to the slide movement; (b) the middle zone comprises middle angle slopes (10º-15º) with secondary scars and many slide lobes; and (c) the lower zone, though being the foot of the landslide deposit did not have an unique toe but a wide and semicircular apron, forming gentle and low angle slopes (<5º). Their semicircular contour is surrounded by a little stream, whose alluviums were invaded northward by the slide mass. The main dimensions of the landslide are summarized in Table 2. The width and length of the displaced mass are 1,280 and 1,000 m respectively, the depth of the rupture surface is 153 m and the landslide achieve a total length of 1,030 m. Carazo, Sahuquillo de Alcázar, etc.). On these topographical high points the geological formations are well preserved and their stratigraphy well exposed. The Pico de Navas (1,351 m high) also shows this ship’s prow form; being the last and highest peak of the Sierra de Navas, which has been affected by a massive gravitational landslide that displaced the northwestern end of the mountain (Figs. 1, 2). The Cretaceous stratigraphic series that outcrops on the northern slopes of Sierra de Navas (Figs. 3 and 4) (bottom to top) consist of: (1) a layer of the Utrillas facies, comprising some 150 m of sand and gravel, with silt and clay minerals. Quartzarenites are predominant and kaolinite is abundant in their matrix. In fact, a kaolin mine at the base of the southern slope of Pico de Navas (Figs. 2, 3) was intensively exploited in the 1980s; (2) layer 15-30 m thick of very soft, yellowish-grey marls and clays of the Cenomanian; (3) layer 50 m thick of a hard marl limestone of the Turonian; and (4) massive grey limestone dated lower Coniacian-Santonian, 200 m thick, though in the Pico de Navas area only the lower 50 to 60 m are preserved. The geological structure of Pico de Navas corresponds to a syncline with W.NW-E.SE strike and 10º E.SE. axis plunge. The syncline dips 10º-15º on its southern flank, giving rise to a scarp face, where the stratigraphic section described above comes from. All the resting outcrops of the fold correspond to its axial zone, which has a subhorizontal dip. Besides, the fold is crossed by a set of faults orientated N.NE-S.SW and separated about 150 m each fault to the other, and dipping 80-90º towards the E. They are Dimensions (m) Measurement type Wd Ld Dd Hd Width of the displaced mass Length of the displaced mass Depth of the displaced mass Height of the displaced mass 1,280 1,000 187 300 direct (field) direct (field) deduced (profile and modeling) deduced (profile and modeling) Wr Lr Dr Hr Width of the rupture surface Length of the rupture surface Depth of the rupture surface Height of the rupture surface 1,280 907 153 320 direct (field, bore hole) deduced (profile) deduced (profile) deduced (profile and modeling) L Total length of the landslide 1,030 direct (field) Table 2.- The Pico de Navas landslide main dimensions. Terminology from Dikau et al. (1996). Fig. 2.- Geological features interpretation on an aerial photo of the Pico de Navas landslide and its surroundings. JIGE 42-1SEGURIDAD.indb 58 19/05/2016 14:21:23 59 Sanz-Pérez et al. / Journal of Iberian Geology 42 (1) 2016: 55-68 natural Kaolin content varies from 1.25 to 10.46% (GalánHuertos and Martin-Vivaldi, 1974) meanwhile for mining exploitation of kaolin are necessary much bigger percentages. In fact, at the SW of the landslide there is an abandoned mine achieving 43% of Kaolinite (white zone in Figs. 2, 6). It is very close to the slip surface and is composed of many white beds very rich in Kaolinite. Ten samples were taken in the area of the landslide representing the lithological variety within the layer with Utrillas Sand Fines Coarse sand Medium sand Fine sand Silt Clay 2.32 (0-22) 8.44 (0-59) 12.16 (0-34) 71.06 (5-89) 6.02 (2-43) 22.92 77.08 Soil description: silty sand with clay, 77% of fines content. 4.2. Characterization of the Utrillas facies layer 4.2.1. Basic properties The lower layer of the stratigraphic section of Pico de Navas is 100-150 m thick and is composed of silty and clayey sand of Albian age (Utrillas facies); it is the upper unit of the Lower Cretaceous (Fig. 4). Above, there are mainly marl and limestone layers of the Upper Cretaceous, which are harder and with more strength than the lower strata. In the context of the Iberian Range, the Utrillas facies consist of various sedimentary materials: loose white sand, quartz gravel, purple, white or red clays, and locally clayey sands, rich in kaolin. The stratigraphic series of Pico de Navas consists of a succession of thin (decimetric) beds of these various lithologies, typical of Utrillas facies, although dominated by silt-clayey sands. The presence of kaoliniferous beds has to be pointed out because its strong influence in the geotechnical behaviour of this stratigraphic unit. The Table 3.- Particle size percentages (%) on average of the Utrillas facies unit at Pico de Navas (interval of values in brackets), according to BS 5930:1999 classification. Fig. 3.- Geological Map of the Pico de Navas Landslide. Geological units in color. Note that even being general dip to NE, there is local and slightly dip variations collected into this figure. JIGE 42-1SEGURIDAD.indb 59 19/05/2016 14:21:23 60 Sanz-Pérez et al. / Journal of Iberian Geology 42 (1) 2016: 55-68 facies. They are grouped into the following types: 5 sandy clay silt, gray to brown in colour; 3 white silty sand; 1 red sandy clay silt and other slightly rounded cemented gravel. These samples do not react to attack of hydrochloric acid, indicating low carbonate content, and do not contain organic matter (Table 1). 4.2.2. Geotechnical index parameters: grain size and plasticity Using the soil classification methods for engineering purposes (e.g. USCS: Unified Soil Classification System) formations with Utrillas facies basically consist of a silty sand with some clay (Table 3). Fine particles content is high (77.08% on average) but clay minerals content is generally low (6.02% on average). Nevertheless in some cases it reaches 11%, or even 43%, when kaolinite content is high. In any case, there should be a considerable influence of fine particles (>35%) on the geomechanical behaviour of this kind of soils. Attending to standard plasticity parameters for fine soil classification (Table 4), the samples corresponds to CL-ML group (low plasticity clay and silt). Moreover, Liquidity index (LI) indicates solid to plastic consistency, depending on water content. Finally, activity values range within a wide data interval (0.2-2.21), showing low (A<0.5) to high (A>1) potential swelling, which primarily depends on the content of clay in the soil. 4.2.3. Geotechnical parameters Since in detail Utrillas facies unit consists of a succession of beds, decimetric to metric in thickness and of various lithologies, is difficult to obtain representative geotechnical parameters of the unit as a whole. Therefore, samples with four common lithologies were chosen (Table 5) and their geotechnical parameters were obtained by laboratory tests. The unit weight ranges between 19 and 23 kN/m 3 ; friction angles are 13º to 38º and the cohesion ranges between 10 and 48 kN/m 2 . For so wide range of values in the parameters is very difficult to establish a representative mean value. Nevertheless, it is likely that in wet condition and considering the pore pressure these parameters considerably decrease, resulting quite lower values. Such values have to be considered the effective ones (Table 5). Furthermore, the layer sequence of limestones and marls outcropping above Utrillas unit has the following strength parameters: cohesion, c = 10 kN/m 2 and friction angle φ = 30°. 4.2.4. Mineralogy and internal fabric (micro-structure) The Utrillas facies unit of Pico de Navas contains quartz minerals grains (≈80%), feldspar (≈10%) and clay minerals (≤10%) according to Galán-Huertos and Martin-Vivaldi (1974) and corresponds to an arkosic sand with a variable content of silt and clay of kaolinite. The observation of this Symbol Parameter Values and classification LL PL PI Liquid limit, % Plastic limit, % Plasticity index Fine soil classification 16.50 – 32.50 15.26 – 25.25 1.24 – 13.34 CL-ML W LI Water content, % Liquidity index Consistency 2.3 – 17 -0.61 – 1.4 Solid to plastic A Activity Potential swelling 0.2 – 2.21 Low to very high Table 4.- Plasticity parameters and fine soil classification, according to BS 1377:1999 standard. Fig. 4.- Stratigraphic sequence of Cretaceous units at Pico de Navas and stratigraphic section. Pictures on the left show outcrops of the main layers and their appearance. JIGE 42-1SEGURIDAD.indb 60 19/05/2016 14:21:24 61 Sanz-Pérez et al. / Journal of Iberian Geology 42 (1) 2016: 55-68 material under the electron microscope shows that the clay fraction (<2 μm) appears as a well crystallized kaolinite. About the 90% are heterometrical crystals (between 0.1 and 1μm), generally with hexagonal shapes, although some crystals show rounded edges. The resting minerals are feldspars and micas somehow altered. This silty clayey sand have a specific internal constitution, because the fine grains stick together in clayey clumps over the terrigenous grains. The presence of these aggregates reaches 85-90% of the whole soil as SEM (Scanning Electron Microscopy) observation has shown (Fig. 7). There are some studies on the behaviour of soils prone to form clumps or clods (Lambe, 1958; Daniel, 1984; Hermann and Elsbury, 1987; Benson and Daniel, 1990; Santiago, 2000) that shows most of the fluid flow through the clay soil occurs through the large pores between and around the clumps (macroporosity or secondary porosity) and not between the isolated clay particles (primary porosity or microporosity). Therefore, the relationship between the aggregates (abundance, size and density) and, between these and the voids of such aggregates, is crucial for the hydraulic conductivity of such materials. Then, these pseudo-spherical clumps, even visible to the naked eye, form a specific structure in the Utrillas facies sand, leaving the rest of the constituents (grains of quartz, feldspar and clay isolated particles) embedded within them. The size of these aggregates is between 10 and 200 μm, being this value range within the limits of the silt fraction, fine sand and even fine and medium sand (Menendez-Pidal, 2006). This feature of some clay particles, which generally tend to clump forming large aggregates, may be the cause of certain clayey beds behavior, despite having a general appearance of a sandy material. They can produce unexpected changes of plasticity and thus it produces deformability, even with a small amount of water within its structure. 4.3. Age of the Movement and state of activity The morphology of the slide deposit seems to be well preserved (Figs. 2, 3, 5) but, as far as we know, there was not any significant activity over the last few hundred years. The patchwork of fields over the lower part of the landslide date at least as far back as the tenth century, and no historic tradition amongst the local inhabitants refers to instabilities of the ground. The water supply for Navas del Pinar, for example, which crosses the middle and lower portions of the landslide, has never been cut as a result of ground deformation. The rocky outcrop of the head scarp has been altered, but only the chaos of blocks in the central part of the crown (Peña del Fraile) seems to be fresher and less colonized by lichen, suggesting that there might still be a very slow background movement. On the other hand, the oldest pottery items found on the surface date from the Iron Age. Nevertheless, there are three facts that can be taken as more certain: Sample Description % γ (kN/m3)φ (º) c (kPa) 4 Silty sand 20 20 27.5 <10 5 Sandy gravel poorly cemented 10 19 38 15 6 Sandy silt slightly clayey 50 23 22 35 7 Clayey and sandy silt 20 20 13 48 Table 5.- Basic geotechnical parameters of the Utrillas unit at Pico de Navas (%: approximate percentage of the presence of each soil type in the local stratigraphic section). Fig. 5.- Geological main section of the landslide: (a) before the slide the Marl and Clay layer retained the water and facilitated its infilling into the fault; (b) after the rotational landslide water was concentrated in a specific zone (below the broken limestone unit and above the Marl and Clay unit) coinciding with the main spring in the area; (c) frontal general view from the North of Pico de Navas, showing the main scarp and the slide mass (50x106 m3 in volume). Note that even being general dip to NE, there is local and slightly dip variations collected into this figure. c JIGE 42-1SEGURIDAD.indb 61 19/05/2016 14:21:25 62 Sanz-Pérez et al. / Journal of Iberian Geology 42 (1) 2016: 55-68 (1) The movement displaced the Navas stream that flows around the base of the mountain, narrowing its alluvial plain. Like all river deposits in this area, the alluvium below the lowest terrace can be dated to the Holocene. However, as commented above, at the foot of the landslide, subsequent erosion by the stream has occurred, though it is very blurred due to ploughing of the fields. The difficulty in distinguishing some sediment from others dissuaded us from opening trial pits to take samples of organic material for dating, since it could confuse more than clarify. Thus, it can be taken that the movement occurred during the Holocene. (2) It is not pointing any deposit of the Chalcolithic (Bronze Age) on landslide formation, although this is very common in the surrounding area as there is lithic industry of this period (Abásolo and García Rozas, 1980). (3) At the summit of the Pico de Navas there is an Iron Age hillfort, situated precisely to take advantage of the natural barrier to the NW formed by the scarp of the landslide. The more accessible SW edge still conserves two walled alignments, one of which is linked to a spectacular limestone karren, which served as a defensive device (as in Celtic hillforts in Iberia). Since the archaeological site dates from the fourth or fifth century BC, we can be sure that the movement had already taken place by this time. Landslide treads the Holocene alluvial deposit of the stream; therefore it is dated later than 10,000 BC. Thus, in principle, the landslide occurred between the Neolithic (5,000 BC to the Iron Age). It must not be a very ancient movement as it also retains its fresh geomorphology. According to the classification of Dikau et al. (1996) on the states of activity of landslides, Pico de Navas slide can be considered inactive, being an abandoned landslide (“inactive landslide which is no longer affected by its original causes”) or even a relict landslide (“inactive landslide which developed under climatic or geomorphological conditions considerably different from those at present”). 4.4. Back analysis and evolution Based on the present day topography of the landslip and the surrounding landforms we have done a hypothetical reconstruction of the topography previous to the landslide. This antecedent geometry is shown in Figure 5a (main cross section). If a stability analysis of the ancient geometric dimension is performed, the conditions at the onset of the landslide can be inferred. Since the current topography is known (Fig. 5a, b), it is also possible to recreate the original topography before landslide and obtain an accurate prediction in the initial situation of instability to reality by a numerical modeling using the geotechnical calculation software FLAC. Figure 8 shows a model with the main cross section of the topography of the situation before the landslide which has been introduced in the calculation software. The main material is a sequence of limestones and marls above Utrillas facies layer as obtained from the geological section described above (Figs. 4 and 5a). All the involved materials are modelled as Mohr-Coulomb type; in the case of Utrillas facies unit, the parameters are obtained from available tests presented in Table 5. It is intended to perform a geomechanical modeling only at the initial moment of sliding; therefore should be chosen parameters of Mohr-Coulomb failure criterion corresponding to that instant. An average value of the density equal to 21 kN / m 3 is adopted. It is considered a null value of cohesion to represent the initial state of the creep process in the situation of initial instability (since the cohesive behavior is limited by the coarser fraction); meanwhile, the frictional behavior is mainly marked by fines fraction which is predominant in Utrillas facies unit, so that their average value of friction angle equal to 18° is chosen. An elastic modulus of 10 MPa for both materials and Poisson’s ratio of 0.3 is considered; however, as is well known generally for such problems, the deformation moduli do not affect the value of the safety factor. These parameters are given in Table 6. For the boundary conditions, horizontal constraints for laterals contours and both horizontal and vertical constraints for the lower boundary of the model are considered. The position of the fault has been incorporated into the model as shown in figure 8. The model includes an interface with residual friction angle equal to one third of the friction angle of upper material of limestones and marls (10º). For calculation purposes, water table has been calculated to level of 1,285 m for strict equilibrium. First of all, initial stresses under static conditions (only under vertical gravity loading) are calculated, considering the water pressure caused by the water table (Fig. 9), as previously described. The estimation of horizontal effective stresses has been performed with Jaky’s formula, since, even if this is clearly not a horizontal layered soil configuration, it can be considered accurate enough for this estimation. As a conclusion of this Fig. 6.- Albian layer (Utrillas facies) densely fractured at the SW lateral zone of the slide mass. White beds reach more than 40% of Kaolinite in content. JIGE 42-1SEGURIDAD.indb 62 19/05/2016 14:21:25 63 Sanz-Pérez et al. / Journal of Iberian Geology 42 (1) 2016: 55-68 beds, causing them to lose their resistance and provoking an extrusion or radial plastic flow over the open boundary and into the valley at the base of the mountain. Even so, the movement was also conditioned by the dip of the strata towards the NW (Note that even being general dip to NE, there is local and slightly dip variations collected into figures 3 and 5). The extrusion of the marls must have been enhanced by calculation, stability of this slope under static conditions have been found. For proper back analysis, it has performed a sensitivity analysis by varying the strength parameters of the two geotechnical units and the position of the water table (Table 7). From this analysis, the considered situation of strict equilibrium can be identified. The obtained 2D failure surface is the one depicted in figure 10, which fits the field observations. In this way, the current geometric situation is analysed. The stability is calculated before slippage occurs using the geomechanical parameters tested. This yields a surface with a safety factor close to 1,000, which indicates the strict equilibrium of the mass considered in this calculation. In addition, it is intended to represent the correct direction of movement of the landslide for the final equilibrium situation. To do this, 3D model using geotechnical calculation software FLAC 3D, has been implemented. Geotechnical properties of both materials shown in Table 6 were considered. It has simplified model considering the fault as a boundary where the failure is initiated and water table with the maximum level of 1285 m indicated above is adjusted (Fig. 11). Figure 12 shows the direction of movement of the failure surface in the 3D model. The Pico de Navas landslide has all the characteristics to classify it as a large magnitude rotational movement (Varnes, 1978; Corominas, 1989; Cruden and Varnes, 1996), affecting more than 50·10 6 m 3 earth. The main rupture of the rotation affected the Utrillas sands. However, the rotational movement also involved the rest of the overlying stratigraphic sequence of limestone and marls, which is 150 m thick, and clays and marls, about 30 m thick. They underwent a series of expansion changes in the existing stratigraphic context, meant that the weight of the limestones would deform the underlying Material Apparent specific weight (kN/m3) Cohesion (kN/m2) Friction angle (º) Elastic modulus (MPa) Poisson´s ratio Sequence of limestones and marls 25 10 30 10 0.3 Utrillas facies 21 0 18 10 0.3 Maximum level of water table (m) Utrillas facies: Friction angle (º) Limestones and marls: Cohesion (kN/m2) 1290 1285 1275 16 7.5 0.87 0.91 0.96 10 0.90 0.92 0.98 12.5 0.91 0.93 0.99 18 7.5 0.93 0.97 1.04 10 0.96 1.00 1.07 12.5 0.99 1.02 1.08 20 7.5 1.01 1.05 1.12 10 1.03 1.07 1.15 12.5 1.06 1.10 1.18 Table 7.- Sensitivity analysis. Safety factors in the 2D model. Table 6.- Geotechnical parameters of the sequence of limestones and marls and Utrillas facies. Fig. 7.- SEM (Scanning Electron Microscopy) images of a clayey and silty sand from the Utrillas facies unit. (A) General view (x50) showing little matrix among sand grains but very abundant on their surfaces. (B) Detailed view (x750) of a quartz grain almost embedded in lamellar kaolinite particles with a large lateral development and medium thick size. Note the scarce presence of voids between the grains and particles constituting the matrix. (C) Above image in more detail (x2.000): well developed lamellar particles of kaolinite. Note the absence of impurities and their pseudo-hexagonal shaped crystals. (D) Piles of kaolinite (x3.500), showing voids among kaolinite aggregates. JIGE 42-1SEGURIDAD.indb 63 19/05/2016 14:21:25