nueva página del texto (beta)
Inglés (pdf)
Artículo en XML
Referencias del artículo
Traducción automática
Enviar artículo por email
Citado por SciELO 
Similares en
SciELO 
Permalink
Introduction
Landslides represent a major socio-economic impact in most world countries, they are responsible for direct and indirect costs as well as for the loss of human lives and ecosystems damage in most mountainous and hilly regions of the world (Nadim et al., 2006; Dikau et al., 1996). The causes of landslides are differentiated into two categories: the triggering factors and the conditioning (susceptibility) factors.
The most common natural landslide triggers include intense rainfall, water-level change, volcanic eruption, and earthquakes. These factors cause an immediate reaction by increasing the stress or strains and reducing the strength of the slope materials. Rainfall is one of the most important triggering factors of landslides (Wieczorek, 1987; Saucedo et al., 2008; Persichillo et al., 2016; Mikoš et al., 2017).
The landslide occurrence varies according to different conditioning factors such as geology, tectonic history, topographic profile, weathering, and erosional history (Varnes, 1978). The conditioning factors are the internal terrain elements susceptible to generating a landslide due to a stimulating factor that modifies the slope condition. The hydrothermal alteration of rocks is an important precursor to landslides and is considered a conditioning factor. However, also hydrothermally altered areas can become a triggering factor (Reid et al., 2001, Maeda et al., 2012).
Rock properties are essential for slope stability, and they are fundamental aspects of geological hazards. The progressive alteration of rocks and the changing properties promote a range of geological phenomena as slope instability and rock failure with the triggering of different landslide types (Aladejare and Wang, 2016; Frolova et al., 2019). In the following, we indicate some examples that show the importance of carrying out a characterization of the geomechanical properties of both fresh and totally altered rock to study the generation of landslides. Frolova et al. (2019) studied the alteration of volcanic rocks and changes in the physical-mechanical properties of the South-Kambalny thermal field (Kamchatka). They concluded that, under the influence of thermal water, the changes in the mineral composition of the original rocks gradually modify to clays and secondary minerals; these transformations promote landslides. The geotechnical characterization of crater wall sequences at White Island Volcano, New Zealand (Moon et al., 2005) showed that hydrothermally altered rocks (andesite) exhibit considerable loss of strength, essentially frictional resistance, compared with the fresh lava units. An important study carried out by Reid et al. (2001) deals with large collapses of hydrothermally altered parts of Mount Rainier (Cascade Range) that generated far traveling debris flows. As part of their methodology, they used a three-dimensional gravitational slope stability tool and mapped the distribution of fresh and altered rocks to analyze the potential volcano instability. They conclude that weakening of edifice rocks by hydrothermal alteration could be important for promoting and localizing the flank collapse.
The above examples show that rock alteration processes are an important factor for slope instability. However, other phenomena such as rainfall that could trigger landslides should also be considered. An example for these conditions is the study carried out by Kim and Song (2015) in South Korea. They analyzed three types of rocks (sedimentary, metamorphic, and igneous) to determine the geotechnical characteristics of landslides under various geological conditions using laboratory test, field surveys and precipitation records. They highlight that rock alteration, rainfall, and the slope of the terrain between 26 and 30 degrees are important factors for slope failure.
In the year 2010, in Mexico occurred extraordinary rainfall in the region of the Sierra Angangueo, Michoacán which is the site of the Monarch Butterfly Sanctuaries. This atypical rainfall produced mass movements in the Sierra de Angangueo. These failures generated several debris flows that impacted the municipality of Ocampo. They caused at least 31 casualties and infrastructure damages throughout the region (Alcántara-Ayala et al., 2012).
In this study, we analyzed the physical-mechanical properties of andesitic rocks that sourced debris flow in the Monarch Butterfly Sanctuary, which belongs to the municipality of Ocampo. We collected samples of fresh and altered andesites and analyzed them in the laboratory to define their mechanical properties. The objective of this study is to know how the hydrothermal alteration diminished the strength of rocks, facilitating the generation of debris flows.
Study area and geological setting
The study area is located in the south-central part of the Trans-Mexican Volcanic Belt (TMVB) (Figure 1a) in the Mil Cumbres physiographic subprovince (INEGI, 2009), within the municipality of Ocampo, located in the northeast of the state of Michoacán between parallels 19°35' and 19°41’ N, 100°14’ and 100°22’ W. This municipality limits to the north with the municipalities of Aporo and Angangueo, to the south with the municipality of Zitácuaro, to the west with the municipality of Tuxpan, and to the east with San José del Rincón, belonging to the state of Mexico.
Figure 1 a) Location of the Trans-Mexican Volcanic Belt (TMVB), the largest physiographic volcanic province in Mexico, and the study area in the State of Michoacán. b) Location of El Rosario basin (red line), Sierra Angangueo (dashed blue line), and the Monarch Butterfly Biosphere Reserve (green line). c) Detail of El Campanario hill and El Rosario basin area.
More locally, the study area lies within the Sierra Angangueo, a NNW-SSE elongated landform that represents the most important morphotectonic feature of the region (Figure 1b). El Campanario hill, with an elevation 3640 m a.s.l., is the highest peak of the Sierra de Angangueo and is the main sanctuary of the Monarch Butterfly Biosphere Reserve (MBBR), located nearby the El Rosario locality (Figure 1c, 2). The hydrothermal alteration zone in El Campanario hill is associated with the mineral deposits of the Angangueo Mining District (Ostrooumov, 2014; Hernández-Bernal et al., 2016). This alteration process was correlated with the 2010 landslides that partly swept the town of Angangueo, Michoacán (López-García and Alcántara-Ayala, 2012).
Figure 2 a) Simplified geological map of the El Campanario hill area, located to the northeast of the Monarch Butterfly Biosphere Reserve (modified from Hernández-Bernal et al., 2016). b) ArcGis profile (A-A´) showing the andesitic lavas at the El Campanario hill covering the metamorphic basement in the Sierra de Angangueo (modified from SGM, 2000).
The basement in the region is represented by Upper Jurassic to Lower Cretaceous rocks of the Guerrero Terrain (151-140 Ma) (Campa and Coney, 1983). The oldest exposed rocks in the area are of Lower Cretaceous age and correspond to a metamorphosed sequence of andesite and volcano-sedimentary sediments interstratified with metalimestone, metalutites and metasandstone (Campa et al., 1981; Campa and Coney, 1983). On top of these old lithological successions, the volcanic rocks of Sierra de Angangueo are exposed (Pasquaré et al., 1991; SGM, 2000; Ferrari et al., 2012; Hernández-Bernal, 2016). The Sierra de Angangueo is made of three volcanic successions emplaced during the early-middle Miocene. They consist of massive andesite lava flows and several andesitic cones aligned in a NNW-SSE direction.
The oldest volcanic unit in Sierra de Angangueo (Lamt) is the most widespread rock in the study area and consists of several volcanic cones of andesitic composition, dated between 23.9 and 13 Ma (Hernández-Bernal et al., 2016). The Lamt is covered by Quaternary basaltic lavas (Abpl), which in turn is covered by alluvium and piedmont sediments that represent the youngest deposits (Figure 2). From a structural point of view, in the Sierra Angangueo predominates a NNW-SSE fault system formed during the Laramide Orogeny in Late Cretaceous to early Paleocene times (Demant, 1978; SGM, 2000). According to SGM (2000), H. Ayuntamiento Constitucional de Angangueo, Michoacán (2014), and Hernández-Bernal et al. (2016) the study area is affected by three successive structural systems that are: a NW-SE system (from 300° to 340°), a NE-SW system (e.g. El Rosario fault of ca. 40 km length), and an E-W system (Morelia-Acambay) that cuts the El Rosario fault (Figure 2b). Faults and fractures of these structural systems affected the stratigraphic succession in the Sierra de Angangueo, generating weak areas in the lithological units.
Methodology
To produce the geological map and the Digital Terrain Model (DTM) of the MBBR and Sierra de Angangueo areas, we combined the digital topography of 5 m contour lines of the Instituto Nacional de Estadística y Geografía (INEGI), and the 12.5 m contour lines of the Alaska Satellite Facility (ASF) (UAF-NASA). The DTM and map are georeferenced with respect to the WGS-1984-UTM-Zone-14N coordinate system. The resulting DTM was used to prepare shaded relief models and an anaglyph. We used Landsat TM8 images, panchromatic (15 m resolution), multispectral (30 m resolution), and Spot-6 satellite images (1.5-m panchromatic and 6-m multispectral resolutions). For the construction of thematic maps (slopes, dissection, and altitude) we obtained a DTM with 15 m resolution (INEGI, 2010). This information was edited in ArcGis 10.2 Base Map, the Google Earth imagery server, and using vectorial cartography (e.g. geographic limits, urban infrastructure, roads, and towns).
Sampling procedure
All physical and mechanical laboratory tests were performed according to international standards to place our results in an international context as follows: We collected two samples according to the distribution of rock units and slope stability phenomena. As shown in Figure 3, the fresh sample (MM193) was collected atop the El Campanario hill. The sample was in situ, away from the collapse zone, without fractures, with a minimum of discontinuities and avoiding metamorphic or weathered areas. The minimum dimensions of the sample were 30 × 30 cm. The completely altered rock sample (MM192) was collected in situ, on the southern flank of El Campanario hill, on the slip plane, near the fault scarp and the collapse zone. The minimum sample dimensions were 30 × 30 cm (Table 1). The ID (identification tag) of samples corresponds to the outcrop number, composed by the first letters of the name of the project, followed by a sequential numbering. To perform a detailed geomechanical characterization, we collected two blocks of fresh and two blocks of altered rocks of different dimensions (~30 × 30 × 30 cm). We segmented each block into twenty-five specimens (cylinders and discs of 53, 32, and 25 mm in diameter) (ASTM D4543-01; ASTM, 2001a) and verified with normalized tools, including the Specimen Flatness Gauge. Cylindrical specimens of altered rock (MM192) could not be extracted and prepared by Core Drilling because the material was very soft. Instead, we, prepared this sample with a lathe, trimmer, and extruder apparatus used to prepare cylindrical soil samples, by taking a mold of a cylindrical specimen of 53 mm in diameter by 90 mm in height.
Figure 3 a) and b) Panoramic view of the outcrops at Monarch Butterfly Biosphere Reserve. c) MM193 outcrop where fresh sample was collected. d) MM192 outcrop where completely altered sample was collected.
Table 1 Sample characteristics of rock analyzed from El Campanario hill at Monarch Butterfly Biosphere Reserve, Sierra Angangueo, Mexico.
| Sample ID | Coordinates (UTM) | Description | ||
|---|---|---|---|---|
| X | Y | Z (m a.s.l.) | ||
| MM192 | 368261 | 2167090 | 3433 | Altered andesite |
| MM193 | 368211 | 2167146 | 3496 | Fresh andesite |
General characteristics of the outcrops and hand specimen are described as follows: The MM192 ochre to yellow outcrop is located in the southern flank of the "El Campanario" hill. It is about 10-m thick. Samples were collected at the slope instability surface that generated a debris flow (Figure 3a and 3b). The entire rock mass was classified as Disintegrated being completely transformed, broken with a mixture of fine grains and angular to rounded rock fragments (Figure 3d). Some silica refilled or recrystallized fractures are observed. In hand specimen, the rock is massive, strongly hydrothermally altered, with plagioclase phenocrysts and pyroxene in minor proportion. At outcrop scale, the entire rock was water-saturated with runoff features.
The MM193 outcrop consists of a sequence of lava flows (ca. 150 m thick) located at the top of the El Campanario hill (Figure 3a and 3c). The color of the rock changes along the outcrop, from gray and brown to ochre yellow in the most altered part. The entire rock mass could be classified as Very blocky: is partially disturbed with multifaceted angular blocks (e.g., wedge-shaped) formed by at least three sets of discontinuities, some of them produced by the cooling process. The hand specimen could be described as massive with porphyritic texture containing plagioclase, pyroxene, and amphibole in minor proportion (Figure 3c).
Physical and petrological characterization
We determined the particle density or real density (ρ) (i.e., the specific gravity of a material) and the bulk density (γ) (i.e., the mass of a material divided by the total volume). The ρ of each sample was defined with a Quantachrome 1200e pycnometer designed to measure the true volume of solid materials by fluid displacement and gas expansion (Boyle's Law). Ultra-high purity helium (99.999 %) is used as the displacing fluid since it penetrates the finest pores assuring maximum accuracy (ASTM D5550-14; ASTM, 2014). The γ was obtained with basic methods (e.g., water saturation) (ASTM D7263-21; ASTM, 2021). A major issue was to obtain an accurate estimate of the total volume of the rock sample. Different methods for calculating the γ of a material are used based on how the void space in the samples is handled (see ASTM D7263-21) for a good review).
The Porosity (η) of rock includes pores, cracks, and inter- and intra-crystalline spaces, and varies between 0 and 1. Porosity keeps a close relation with strength, strain and hydric properties depending on the rock type and grain arrangements (grain size and distribution, grain packing, particle shape). Total porosity (ηt) considers the total porous volume regardless of whether they are interconnected or not, and effective porosity (ηe), is the interconnected portion of the total void space allowing the circulation of fluids. General characteristics of the pore structure (e.g., pore size distribution and porosity) are related to mechanical properties (e.g., compressive strength, modulus of elasticity and Poisson ratio), as well as hydric characteristics (e.g., capillarity and permeability).
Effective porosity (ηe) was obtained following the procedure recommended by ISRM (2007) employing a vacuum water saturation test. Samples were oven dried at a temperature of 70º C for 48 hours until a constant mass was achieved. The dried samples were placed in a vacuum at 20 ± 7-mbar pressure in three 5-hour cycles. As described in the European standard EN 1936:2006 (EN, 2006), the trapped gases in the porous system were eliminated during the first cycle. During the second cycle, samples were slowly introduced into distilled water over 15 minutes until they were submerged 5 cm in water. The samples were submerged for 5 hours. Atmospheric pressure was re-established and maintained throughout the last cycle. Thus, connected porosity is defined as the measure of the void spaces in a material (ratio of absorbed water to water density), and is a fraction of the volume of voids over the total volume, expressed as a percentage.
Hydric characterization
Permeability (k) is a measure of how readily a fluid can flow through a porous material and it is a property inherent to the material which does not depend on the type of fluid used to measure it. Permeability depends on intrinsic properties including porosity, pore size, pore-size distribution, pore connectivity, and tortuosity. Permeability (k) test was carried out following standard procedures (ASTM D4525-13e2; ASTM, 2013) and those described in the literature (Sander et al., 2017). Permeability values were measured by using a GasPerm AP-123-002-0 gas permeameter, designed for the determination of permeability to gas at steady state condition (constant pressure and flow through the sample). In general terms, k is obtained by the injection of nitrogen (99.999 % high purity) at flowrate Q and pressure P through a confined cylindrical rock specimen. The confining pressures range from 300 to 400 psi, while flow rate range from 0 to 600 cm3/min. Initial flow rate and differential pressure during the test are subjected to the characteristics of each rock type and the length of the specimens. Results were collected by applying the law of Darcy for ideal horizontal laminar flow of gas under steady state isothermal conditions. To account for the Klinkenberg effect (gas permeability of a core is always higher than its permeability to a single saturating inert liquid), we constructed a graph. Here, if the gas permeability values obtained at different mean core pressures are plotted against reciprocal mean pressure (1/P m) a straight line should fit the points (frequently five). Finally, the liquid permeability is obtained by the extrapolation of this line to infinite mean pressure where it intersects the gas permeability axis. The spatial location of these five points and the slope of the trend line are directly related to the distribution of pore sizes and the morphology of the entire system.
Mechanical characterization
The mechanical behavior of a rock material directly depends on its physical properties (e.g.,Heap and Kennedy, 2015; Bubeck et al., 2017) and it shows a strong link with porosity (an increase in porosity or pore size is typically associated with a decrease in strength), the distribution of the minerals and pores, and the fracture content and arrangement (the relative orientation of the applied load and fracture plane).
Uniaxial compressive tests and splitting tensile tests were performed following standard procedures ASTM D2938-95(2002) (ASTM, 2002a), D3148-02 (ASTM, 2002b), ASTM D3967-95a(2001) (ASTM, 2001b), using a Geotechnical Digital Virtual Infinite Stiffness Loading System (250 kN GDS VIS) at a constant displacement rate of 4 mm/h. Stress and strains were recorded continuously during each test by two load cells (internal [inside the load cell] and external [integrated into the compression ring]). The elastic properties (Young’s modulus and Poisson's ratio) were obtained by recording local deformations through the use of three Linear Variation Differential Transformers (LVDT), two axial and one radial.
Results
Through physical and mechanical laboratory tests, we determined that the instability factors are directly related to the alteration of the rocks and the changes in their physical-mechanical properties. Other conditioning factors in the study area are steep slopes, altimetry, and geometry of the terrain that were determined with photo interpretation techniques, digital images, and relief analysis.
Figure 4a shows the regional landform conditions in a hypsometric map of the El Rosario basin. We observe that El Campanario hill is the most prominent elevation of the basin (yellow circle). Figure 4b shows the slope map in which the steeper slopes in the area occur near the top of the hill (black circle). Figure 4c (2005 satellite image) shows a close-up on El Campanario hill before the slope failure. We observe the hillside with undisturbed vegetation. The white dotted line shows the collapse area and the direction followed by the displaced material. Figure 4d shows a satellite image (2015) after the slope failure. The extension of the debris flow deposit to the SW of El Campanario hill has a straight to convex geometry from and a steep slope near the scarp. Additionally, we observe the debris flow source area, the scarp, and the rock alteration zone located at 3475 m a.s.l. with a slope gradient > 35°. The toe of the debris flow deposit occurs at 2988 m a.s.l. with a slope gradient of < 10°.
Figure 4 a) Hypsometric map of El Rosario basin showing the difference of altitudes in the study area. b) Slope map showing the grade of the slope at El Campanario hill and its influence on hillside instability. c) Photogrammetric image from 2005 showing the study area before the slope failure. The route followed by the debris flow is shown with a white dotted line. d) Satellite image Spot 6 (February 2015) showing the debris flow in the Monarch Butterfly sanctuary developed in 2010. We can observe its magnitude, length and area of distribution.
About the rainfall in the study area, the rain records from five meteorological stations near the Monarch Butterfly Biosphere Reserve indicate that data from 2010 were atypical (Figure 5). As a result, the National Water Commission (Comisión Nacional del Agua) declared several municipalities as disaster area in Michoacán (DOF, 2010). Figure 5a shows the rainfall in February from four hydrometeorological stations in the Sierra de Angangueo, highlighting the Chilesdo station with 312.4 mm of precipitation in 2 days. In turn, Figure 5b shows the historical record of the Chilesdo station, clearly showing an anomalous value in 2010. Alcántara-Ayala et al. (2012) pointed out this atypical precipitation with cumulative rainfall values (more than 144 mm in one day).
Figure 5 a) Histogram of rainfall in February 2010 recorded in five hydrometeorological stations in the Sierra de Angangueo. Values> 190 mm considered as atypical are observed for the month of February. b) Pluvial precipitation values recorded at the Chilesdo station, located just 14 km from the Monarch Butterfly Sanctuary.
Based on our field survey we identified the area affected by landslides during the February 2010 rainfall. The area exposes fresh andesite sampled at 3475 m a.s.l. (MM193), highly altered andesites at the debris flow source zone (MM192). At outcrop scale, the collected samples show substantial differences in effective porosity (ηe), permeability (k), and density. Table 2 summarizes the properties of fresh and altered samples, all results being given as an average value, with their corresponding standard deviation.
Table 2 Summary of physical and mechanical properties of fresh and completely altered samples (all values are given as an average, with their corresponding standard deviation).
| ID | MM193 | MM192 |
|---|---|---|
| γHe (g/m3) | 2.61±0.002 | 2.52±0.0007 |
| ρHe (g/m3) | 2.807±0.001 | 2.54±0.02 |
| k (mD) | 0.262±0.319 | 393.71 |
| k (m2) | 2.58E-16±3.15E-16 | 3.88E-13 |
| ηeHe (%) | 12.06±2.46 | 59.88 |
| ηe (%) | 10.04±2.16 | |
| ηT (%) | 17.14±2.54 | 60.93 |
| σs (MPa) | 63.51±19.67 | 0.26±0.03 |
| E (GPa) | 26.7±6.40 | - |
| ν (-) | 0.38 | - |
| σt (MPa) | 6.3±1.24 | - |
Abbreviations are: γHe = bulk density by helium pycnometer; ρHe = real density by helium pycnometer; k = permeability; ηe He = effective porosity by helium pycnometer; ηe = effective porosity; ηT = total porosity; σs = uniaxial compressive strength; E = Young modulus; v = Poisson ratio; σt = tensile strength.
The strength of a rock is closely related to its texture and structure, and fracture mode and fracture evolution depends on the distribution and size of grains, minerals and pores. Figure 6 shows the stress-strain behavior of both MM192 and MM193 samples: the strength of the fresh sample (∼60 MPa, MM193 from 1 to 10), in comparison to fresh sample with preexisting damage (microcrack) (∼116 MPa, MM193-11) is abnormally lower, probably because both the orientation and the characteristics of microcracks (e.g. roughness, cementation grade) (see Figure 6c) contributed to the increase in the final strength. On the contrary, as shown in Figure 4a, the slope of the stress-strain curves (E values) is similar. Figure 6c shows the failure surface, developed in sample MM192-02, MM193-11, and MM-193-02 specimens. As expected, in the most homogeneous specimens, cracks developed in the direction of the maximum applied shear stress, while in the specimens with preexisting damage the cracks developed exactly on the preexisting microcracks. Moreover, cracks of the completely altered samples are characterized by conical and vertical shearing, mode of fracture associated or described for porous or granular materials.
Discussion
According to the geomechanical results of the fresh and altered rock samples (stress-strain, porosity, permeability, compressive strength) (see Table 2), the debris flow deposit contains particles ranging from 2 mm to metric-sized blocks of fresh rock. The fraction of fine particles <2 mm was provided by the altered rocks. Therefore, at the time of the debris flow emplacement it was channeled and had a medium mobility typical for cohesive flows with a matrix (Sánchez-Núñez et al. 2015). The maximum distance traveled by the debris flow was about 1.5 km prior to stop.
Frolova et al. (2019) found similar geological settings at the largest geothermal fields in Kamchatka. They argued that thermal water changes the mineral composition of the original rock (andesite) promoting landslides. Moon et al. (2005) presented similar conditions at White Island (stratovolcano of New Zealand). They showed the susceptibility to crater wall collapse due to a very steep slope and the significantly hydrothermally altered rock. Other authors, such as Jazouli et al. (2020), Upton et al. (2018) and Reid et al. (2001), presented considerations relevant to slope stability assessment related to rock properties and slope gradient. They performed a geomechanical study and characterized the strong (fresh) and weak (altered) rocks in the basin of the Oum Er Rbia river (Morocco), the Alpine Fault (New Zealand) and from Mount Rainier (Washington).
Rainfall is one of the most important triggering factors for landslides worldwide; this is shown by many studies such as those carried out by Rong et al. (2020) in Shuicheng Country, China; Zhang et al. (2019) in Shaanxi Province, China, Saucedo et al. (2008) in Nevado de Colima volcano, Caballero et al. (2006) in Chiapas, Mexico, Capra et al. (2003) in Puebla, Mexico. Polemio and Petrucci (2000) presented an overview of international research on rainfall as a landslide triggering factor.
According to Table 2, we observe a clear decay of the properties between fresh and completely altered lava is visible. Effective porosity (ηe), permeability (k), and uniaxial compressive strength (UCS) changes from approximately 12 to 60 %, 0.26 to 393.71 mD, and 63 to 0.28 MPa, respectively. Values of the fresh samples are in the range of similar volcanic rocks reported elsewhere (e.g., Pola et al., 2014; Heap et al., 2014; Farquharson et al., 2015), while the properties of the completely altered samples have not been placed in a global context because the degree of alteration has not been analyzed in detail in this work. In any case, a literature review suggests that strength values are within the range of those for hard soils (e.g., Avşar et al., 2015) and weak rocks.
As described in literature (e.g., Palchik and Hatzor, 2002) porosity (effective and total) is an important parameter of rock materials which directly determines their mechanical strength. As graphically described in Figure 4c, the intense hydrothermal alteration could completely modify all physical properties, including rock fabric, porosity, density, and consequently mechanical behavior.
According to Benavente et al. (2015), the fluid permeability could be classified into four different groups: very low (k value under 1 mD), low (1 < k < 100 mD), permeable (100 < k < 10000 mD), and high (k values higher than 10000 mD). The altered and fresh samples could be classified as permeable and of very low permeability, respectively (see Table 2 and Figure 7). The five- point reciprocal mean pressure test (1/Pmean [Abs Atm-1]) is efficient not only for characterizing the permeability, but also for understanding the very complex interconnected pore structure and the effect of some interconnected pores that contribute very little to the flow (e.g., dead-end pores, see Sander et al., 2017 for a good review). In this way, if we want to respect always the Darcy’s law for ideal horizontal laminar flow of gas under steady state isothermal conditions, in certain mechanisms of flow such as diffusion and dispersion it is important to pay attention to the effects of this particular type of pores. Particularly, the dispersion mechanism is due to the differences in the fluid flow velocity fields in the fracture networks, caused mainly by the variations in trace lengths, orientations and apertures of the interconnected pores and the fractures.
Figure 7 Empirical permeability of fluid (mD) versus reciprocal mean pressure (1/Pmean [Abs Atm-1]). Solid black lines indicate the permeability classification according to that discussed in Benavente et al. (2015), while red and orange solid circles indicate the position of values of the fresh (MM193) and completed altered samples (MM192), respectively.
Values of k of the specimens range from 0.012 to 1.03 mD for fresh specimens to 396 to 409 mD for completely altered specimens (Figure 7). If we compare these results with those obtained by the implementation of the same methodology on different rocks, we observe that permeability values of the completely altered samples are similar to the values obtained for very porous materials such as pumice rocks. On the contrary, large differences are observed in the distribution of the five-point reciprocal mean pressures. It should be remembered that these distributions could be associated to the geometric and physical characteristics (e.g., tortuosity, roughness) of the microstructural elements of the specimens. In turn, permeability values of fresh samples are similar to the values obtained for dense but microfractured materials, as well as for hydrothermally altered rocks (Figure 7). However, a considerable increase in permeability values and therefore in effective porosity between MM193 and MM192 samples could be easily identified.
Concluding remarks
We investigated the differences in physical and mechanical properties between fresh (MM193) and completely altered (MM192) lava samples. The main findings are as follows:
The intense hydrothermal alteration has completely modified all physical properties, including porosity, density, and consequently mechanical behavior of the fresh lava MM193.
The strength of fresh samples is in the range of strength of similar andesite rocks, while the strength of fresh sample with preexisting damage (microcrack) is abnormally higher, probably because both the orientation and the characteristics of microcracks (e.g., roughness, cementation grade) are contributing to the increase in the final strength.
The strength of altered samples is very low; it is within the range of strength values of a weak rock. Values obtained in dry conditions indicate that strength decreases about 99 % from fresh to completely altered sample (from 70 to 0.26 MPa).
Permeability values change drastically between fresh and altered samples. Mechanism of flow such as diffusion and dispersion are graphically described by the distribution of the reciprocally mean pressure.
The conditioning factors involved in the generation of the debris flow in El Campanario were: a substrate altered by hydrothermalism, the slope of the land (> 35°), and the morphology of the slope. The triggering factor was the unusual rainfall (> 144 mm on a single day) in February 2010.
