Geotechnical Assessment of Volcanic Edifice Collapse Events, Pico De Orizaba Volcano, Mexico

Concha-Dimas, Aline (aline@unr.nevada.edu)

Watters, Robert J. (watters@mines.unr.edu)

Introduction

Pico de Orizaba volcano (see Figure 1, for location) has collapsed twice (Carrasco-Nuñez, 2000). The initial stage of evolution for this volcano is known as the Torrecillas cone, which ended with collapse at 0.21 Ma b.p., and the related deposits formed the Jamapa avalanche which traveled eastward 75 km. A second, superimposed constructional stage is the Espolón de Oro cone that also ended with a collapse 20,000 years b.p., forming the Tetelzingo avalanche-lahar that traveled 85 km (see Figure 2 and 3). The present volcanic cone is the last constructive stage, known as the Citlatépetl stage.
 

Figure 1. Location and aerial view looking from southeastern portion of Pico de Orizaba

As a first step for geomechanical approaches in evaluations of edifice slope stability (Voight, 2000), we examine prior failures of Pico de Orizaba and obtain mechanical properties of the materials involved in those failures in order to:

1. characterize intact rock strength;
2. compare mechanical properties of the involved materials with intact edifice rock; and
3. get input values for future numerical models to evaluate volcanic flank stability of old failures and possible failures of the present cone.

Figure 2. Schematic reconstructions (section A-A') of the three stages at Pico de orizaba (Modified from Carrasco-Nuñez, 2000; Hubbard, 2001)

Estimations of rock mass properties

Geomechanical analyses for volcano stability require estimations of rock mass strength (commonly used for design of slopes, foundations and underground excavations). Many numerical models such as Finite Element (FE) and Finite Differences (FD) are expressed in Mohr-Coulomb linear failure criterion, see Eq. (1). Non linear failure criterions are more adequate and can be conveniently applied as proposed by Hoek and Brown (Hoek, 2000), see Eq. 2. This criterion is based on empirical relationships between numerical constants and physical characteristics of the rock: joints, interlocking blocks and condition of surfaces.

t = c'A+ s _n' tan f ' (Eq. 1)

t : shear stress acting on sliding plane

c': cohesion at effective stress

A: area of sliding plane

s _n': effective normal stress on sliding plane

f ': friction angle at effective stress

s '₁= s '₃ + s _ci [(m_bs '₃/s _ci)+s]^a (Eq. 2)

s '₁ , s '₃: maximum and minimum effective stresses at failure

m_b= f(m_i , GSI ): Hoek and Brown constant

or the rock mass characteristics (i.e. Geological Strength Index, GSI)

and the intact rock properties (m_i)

s, a= f(GSI); constants that depend of the rock mass characteristics

s _ci: Uniaxial Compressive Strength (UCS)

Figure 3. Avalanche deposits distribution, proposed avalanche calderas, and location of section A-A' (After hubbard, 2001, and Carrasco-Nuñez, 2000)
 
 

UCS was indirectly assessed by Point Load Index on boulders of the proximal avalanche deposits. Boulders of the proximal facies ensure representative values of the intact rock, since this facies mainly contains big blocks of the different units from the former edifice (Glicken, 1991) and also the weathering degree is low, in the order of II-III according to Krank and Watters (1983). Cohesion and friction were obtained from shear test on samples from core remnants. Figure 3 shows the location of the collected samples. We classify the samples from the deposits in three groups: fresh lavas, hydrothermal-alterated (argillic or silicic) according to Hubbard (2000), and pyroclasts.
 
 

Obtained Data

Table Figure 4 shows the values obtained for the mechanical properties of avalanche deposits used in linear and non-linear failure criteria. In order to compare, we include values obtained for the new cone samples.

Figure 4. Values of UCS, c, f from point load and shear tests. Torrecillas (Torr), Espolón de Oro (EO), and Citlaltépetl (present) cone (C)
 
 

Discussion

Strength values obtained from proximal deposits give good approximations of the intact rock strength for the old edifices, since they are representative blocks of the different units that constituted the old volcanic edifice and samples of adequate sizes for point load test had low degrees of weathering.

UCS values for the different type of rocks are similar. Even though fresh lavas are compositionally different, they show a similar range in mechanical strengths: Torrecillas: 50-216 [MPa]; Espolón de Oro: 56-236[MPa]; Citlaltépetl: 60-186 [MPa]. Also, argillic alterated material shows the same range of strengths: 10- 60 and 36-90[MPa] for the two known stages. Therefore, properties for materials that we were not able to get, because they are not preserved or because appropriate type of samples for specific type of tests are not available, can be extrapolated from existing data of similar rock types.

The obtained UCS values represent the highest input values for the rock mass in numerical models of previous failures since they represent the strength of the intact rock (see Figure 5). Future work on uniaxial and triaxial tests will restrain m_i values. Since rock mass properties by means of the GSI parameter in the Hoek and Brown criterion, are not known, it is necessary to perform sensitivity analyses for this. For DE analysis and flow codes the maximum displacement of the avalanche deposits can be considered restraining (final) conditions for numerical models. Distribution and types of materials at the actual cone are used to refine re-constructions of old cones in order to create a simplified geological model for input in numerical analyses (See Figure 6).

Figure 5. Scale effects on the rock mass strength (After Voight, 2000)
 
 

Figure 6. Schematic Section (A-A') showing the distribution of different strenght zones for Citlaltepetl stage.
 
 

Conclusions

The first step for geomechanical evaluations of volcanic flank stability at Pico de Orizaba volcano was to analyze mechanical properties of geological materials involved in previous failures. We classified them as fresh lavas, hydrothermally alterated and pyroclasts that represent different zones of strength inside the volcanic edifice. These strength-zones represent a simplified geological model for numerical models. Some of their mechanical properties, for linear and non linear failure criterions, were measured directly or can be inferred from the same type of rock from other stages when missing.

Numerical models require intact rock parameters (UCS and m_i) and rock mass parameters (GSI). Sensitivity analyses will be performed to generate realistic models for old avalanches. Final displacement from the source will represent restraining conditions.

References