Comparison of simplified and detailed analyses for assessment of

dynamic displacement of a centerline tailings dam

1
A.R. Pretell Ductram
2
J.N. Dismuke
1
Golder Associates, Denver, USA.
2
Golder Associates, Christchurch, New Zealand.

ABSTRACT: Tailings dam failure is one of the most catastrophic events in geotechnical engi-
neering due to the environmental and economic impact a failure could cause, as well as the po-
tential for loss of human life. Centerline tailings dams are increasing in popularity because they
are efficient and allow maximum storage capacity with a minimum volume of dam material.
Despite these advantages, centerline tailings dams require sophisticated design analyses in order
to minimize the probability of dam failure occurrence. This article presents static and dynamic
analyses of a centerline tailings dam located in the Peruvian Andes, in an area with challenging
topographic and seismic conditions that are typical of many of the existing and proposed dams
in South America. Static deformation analyses are presented and discussed in this paper, and
permanent displacements obtained from dynamic analysis are compared with displacements de-
termined using several different simplified methods.

1 INTRODUCTION

The centerline construction method has become an attractive alternative for dam designers in
the mining industry since it allows optimization of available storage while minimizing the vol-
ume of dam material. In seismically active areas of the world the dam design is often governed
by earthquake loading. Detailed dynamic numerical analyses to assess the effects of earthquake
loading on dams are difficult and time consuming, so designers typically rely on simplified de-
sign methods because simplified analysis methods are comparatively straightforward and cost
effective to implement. These methods are acceptable for a wide range sites, but as design dam
heights and slopes become more aggressive there is need for reliable detailed dynamic analyses
to justify taller and steeper dams. Dynamic numerical modelling techniques have become more
accessible to practicing geotechnical engineers as computing power increases, as has access to
current seismologic data and modern ground motion prediction tools to develop input ground
motions for dynamic analyses.

The main purpose of this article is to demonstrate the difference between dam displacements
determined using simplified and detailed dynamic analyses. Permanent seismic displacements
obtained from simplified methods that are common in engineering practice are estimated for a
230 m high centerline tailings dam located in Peru. These displacements are compared with
seismic displacements directly computed using dynamic analyses. The dynamic analyses were
conducted using FLAC (Itasca, 2010), and comprise material characterization for static and dy-
namic conditions, development of design acceleration response spectra, ground motion selection
criteria, spectral matching of input time histories, and both static and dynamic analyses. Com-
parison of simplified and numerical results provides valuable insight to the limitations of simpli-
fied analyses and benefits of detailed analyses for seismic design of centerline tailings dams.
2 LOCATION AND CENTERLINE TAILINGS DAM DESCRIPTION

The centerline tailings dam used for the analyses (the dam) discussed in this paper is located in
the middle of the Peruvian highlands in a deep valley surrounded by the Andes, which is the
typical location of most mining projects not only in Peru, but also in South America.
The dam is 230 m in height, and thought to be built in 10 stages. Both upstream and down-
stream slopes are equal to 1.6H:1.0V. The dam is mainly composed of rockfill, but core, filter
and transition materials are also included in a minor extension. The core consists of till material
(gravelly clay, GC), with a minimum plasticity index (PI) of 10. The filter and transition are
composed of coarse sand and sandy gravel materials respectively. The dam also includes an up-
stream geomembrane to prevent contact water from getting into the dam body.
Stages 1 and 2 of dam construction were built as a downstream dam, with average height of
130 m; here core, filter and transition are placed at the upstream slope. Stages 3 to 11 were built
following centerline construction method; in this case the crest of every stage is composed by a
sequence of rockfill, upstream filter, core, downstream filter, transition a rockfill from upstream
to downstream.

3 NUMERICAL MODEL

The dam was modeled using FLAC (Fast Lagrangian Analysis of Continua) code, which al-
lows users to solve stress-strain geotechnical issues considering a finite differences method. The
Model is composed of 157,600 zones strategically spread to properly model areas of interest, i.e.
the centerline core. The dimensions of the model and zones satisfy seismic wave transmission
requirements according to Kulhemeyer & Lysmer (1973) recommendations, which state that
maximum zone dimension should be less than one tenth of the maximum shear wave length for
a given material. Figure 1 presents FLAC dam model.

Figure 1. Centerline tailings dam model.

4 STATIC ANALYSIS

A static analysis was performed in order to verify dam deformations observed during con-
struction. Dam construction was simulated for every 10 m of the massive rockfill zone and eve-
ry 1 m for the last 15 m before the crest is reached for all stages. The tailings elevation was kept
5 m under the crest. Horizontal and vertical initial stresses were taken into account to simulate
stresses transmitted by compaction equipment.
The upstream geomembrane is assumed to behave correctly during construction, so no seep-
age into the dam body is taken into account for the analysis and hydrostatic pore pressure distri-
bution is assumed for the tailings deposit.
4.1 Material properties
Foundation, dam materials and tailings properties are based on Golder field and lab geotechnical
investigations as well as published technical literature. The Mohr-Coulomb plastic model was
considered for tailings and dam materials, but an elastic model was used for bedrock.
Bedrock elastic parameters were obtained from geophysics tests comprising multichannel
analysis of surface waves (MASW) and cross-hole wave velocity measurements. Dam materials
were characterized based on large scale triaxial tests performed by Golder. Thus, elastic parame-
ters take into account particle crushing and interlocking, which has great impact on deformation
results during construction.
Tailings parameters were estimated considering three different conditions variable with
depth: tailings recently disposed, during consolidation and normally consolidated. They were
characterized based on cone penetration tests (CPTs) carried out by Golder. Table 1 presents
properties considered for the analysis at a reference confining pressure of 2000 kPa.

Table 1. Material properties for static analysis.

Density Bulk Modulus Shear Modulus Cohesion Friction angle
Material
(kg/m3) (MPa) (MPa) (kPa) (°)
Rockfill 2100 18 12 0 37.5
Core 1950 150 70 10 37.5
Filter 1900 180 70 0 37.5
Transition 1900 180 85 0 37.5
Tailings 1900 60 21 0 34.0

4.2 Construction displacements

Figures 2 and 3 respectively show horizontal and vertical displacement distributions due to
the construction process. Horizontal and vertical displacements reach 4.5 m and 7.0 m in the
middle of the dam, where centerline construction takes place, and both tailings and rockfill are
in contact. Horizontal and vertical displacements in the centerline dam zone show considerably
uniform distributions, which suggest low levels of differential displacements. Furthermore, in
both horizontal and vertical cases, bedrock presence has a relevant and notorious influence on
displacements.
Rockfill crushing and interlocking is well reflected in these results, which are important be-
cause they affect volume estimation during planning. Construction schedule and associated costs
should include crushing and interlocking when dealing with rockfill dams.

Figure 2. Horizontal displacement distribution.

Figure 3. Vertical displacement distribution.

5 DYNAMIC ANALYSIS

The main purpose of this study is to evaluate permanent horizontal and vertical displacements
after earthquake shaking, especially at the dam crest. This was done using a dynamic analysis by
inputting an earthquake acceleration time history at the base of the FLAC model described
above.

5.1 Site seismic conditions

Peru is one of several countries highly affected by earthquakes since it is located in the Pacif-
ic Ring of Fire. Peruvian seismic activity is dominated by the convergent plate boundary be-
tween the South America Plate at the east and Nazca Plate at the west of the coastline and be-
neath Peru, where more than 18 earthquakes higher than 7.5 M have been recorded since 1900
(Bilek, 2009). Intraslab events are considered to dominate the seismic activity of the dam loca-
tion as they do in most of the Peruvian Andes.

5.1.1 Design acceleration response spectrum definition

The National Service of Training for Industry and Construction (SENCICO from its Spanish
initials) implemented a seismic hazard service for Peruvian territory, which provides geotech-
nical engineers with uniform hazard spectra (UHS) by entering coordinates of the site of interest
into their web based application. UHS obtained from this service are based on a seismic cata-
logue updated to October 2013 for bedrock, or a soil profile Class B in the International Build-
ing Code (ASCE, 2005). Seismic hazard information is expressed in terms of annual exceedance
probability, and spectral acceleration is given for a range of periods from peak ground accelera-
tion (PGA) at 0s to 3.0 s with critical damping ratio between 2% and 10%. Figure 4 shows the
ground motion hazard curves (probability of exceedance curves) obtained for the dam deter-
mined using SENCICO’s tool, and Figure 5 shows the UHS for the 2475 year return period (2%
probability of exceedance in 50 years).
Figure 4. Ground motion hazard curves for various periods.

5.1.2 Seed ground motion record selection

Ground motions for input into the dynamic model are developed by matching observed ground
motion to the design UHS. The first step in this process is to select seed ground motion records.
Ground motion record selection is a task that should be done carefully because it has an im-
portant effect on the dynamic analysis results.

There are several parameters to be considered by engineers when records are selected, i.e. earth-
quake source, where distinction should be made not only between crustal and subduction
sources, but also intraslab and interface (Bommer & Acevedo, 2004), distance and magnitude
(Steward et al., 2001), duration (Bommer & Acevedo, 2004), Arias intensity (Hancock et al.,
2006), spectral shape, site classification, etc. Typically, the ground motion hazard is assessed us-
ing a process called deaggregation to determine earthquake rupture scenarios that control the
hazard for a specific site, that is, to determine target parameters to aid in ground motion record
selection. Using deaggregation many of the parameters discussed in this section may be ob-
tained, but for brevity, discussion of the deaggregation for the dam site is not included in this
paper.
Ground motion records to perform the spectral matching (shown in Table 2) have been col-
lected from various institutions including PEER, COSMOS, REDACIS and RENADIC.

Table 2. Candidate events for dynamic analysis.

__________________________________________________________________________________________________________
Event Name Date Magnitude Failure Mechanism Country Station
__________________________________________________________________________________________________________
Peru Coast 31 May 1970 8.0 Subduction Intraslab Peru Parque de la Reserva
Lima 3 October 1974 8.1 Subduction Interface Peru Parque de la Reserva
Valparaiso 3 May 1985 7.9 Subduction Interface Chile Llolleo
Punitaqui 15 October 1997 7.1 Subduction Intraslab Chile Illapel
Iquique 13 June 2005 7.9 Subduction Intraslab Chile Idiem
__________________________________________________________________________________________________________

5.1.3 Spectral matching

Once a seed ground motion is selected, it may need to be modified to fit the site specific seismic
hazard. There are several methodologies to generate time series for dynamic analysis of struc-
tures, one of the most used and recommended is time domain spectral matching where a seed
time history is modified to have a response spectra that matches closely with the design UHS.
The main attribute of this method is the preservation of the non-stationary behavior of the origi-
nal ground motion (Hancock et al., 2006; Takhirov et al., 2004).
 The spectral matching process for this study was performed using the software SeismoMatch
(SeismoSoft, 2013), which uses a wavelet algorithm proposed by Abrahamson (1992) and Han-
cock et al. (2006). SeismoMatch performs time-domain spectral matching by adding wavelets to
reduce the difference between input and target spectra.
It is typically recommended to select a suite of three records for spectrally matching for dy-
namic analysis of dams as a minimum to assess the variability in time histories and compensate
for peaks and valleys present in different spectra relative to the target (NIST 2011). For practical
purposes, this article discusses dynamic analysis using a single matched record from the Iqui-
que Earthquake in 2005. Figure 5 shows the UHS target spectrum, and both the original and
matched spectra, while the original and matched time-histories of acceleration are presented in
Figure 6.

5.2 Material properties for dynamic analysis

Over the last four decades, several authors have demonstrated that small strain (dynamic) soil
stiffness is sensitive to several parameters, i.e. shear strain, void ratio, confining pressure, PI,
overconsolidation ratio (OCR) and number of loading cycles (Kramer, 1996). In dynamic anal-
yses the dynamic soil stiffness is modelled as nonlinear shear modulus reduction and damping
ratio curves, which were introduced to geotechnical engineering when dealing with equivalent
linear soil response analysis (Seed & Idriss, 1970); however, since that time their use has spread
extensively until today, and they are used even in nonlinear analysis.

Figure 5. Target, original and matched acceleration response spectra used in dynamic analysis.
Figure 6. Matched time-history of acceleration used in dynamic analysis.

Dynamic properties comprising small strain shear stiffness, Gmax, and shear modulus reduc-
tion and damping ratio curves were selected using several references: Rollins et al. (1998) was
considered for rockfill and transition materials, Vucetic & Dobry (1991) for the core with PI of
10, and Seed & Idriss (1984) was chosen for the filter material. Figure 7 shows the dynamic
properties curves for dam materials. Dynamic properties selection is a task that should be done
judiciously because the difference between design references could have a great impact on the
results (Pretell et al., 2012).
Maximum elastic properties of rockfill were estimated based on shear wave velocity meas-
urements, and Gmax and bulk modulus values vary with depth. Likewise tailings properties were
obtained from CPT tests results and advanced laboratory tests.
On the other hand, core, filter and transition elastic properties were considered to vary with
mean confining pressure according to Equation 1, proposed by Seed et al. (1970, 1984).
0,5
𝜎′
𝐺𝑚𝑎𝑥 = 21.7 𝐾2 𝑃𝑎 ( 𝑜 ) (1)
𝑃𝑎

where Gmax = maximum shear modulus (kPa); K2 = modulus coefficient; Pa = atmospheric pres-
sure (kPa); and ’o = mean confining pressure (kPa). The K2 parameter was taken from technical
literature (Vrymoed, 1981).

Figure 7. Comparison of dynamic properties for dam materials.

 Table 3 presents Poisson ratio and Gmax for each material at 2000 kPa confining pressure.

Table 3. Material properties for dynamic analysis.

__________________________________________________________________________________________________________
Material K2 Poisson Ratio () Gmax (MPa)
__________________________________________________________________________________________________________
Rockfill - 0.25 1750
Core 80 0.30 650
Filter 70 0.33 625
Transition 90 0.30 675
Tailings - 0.13 160
__________________________________________________________________________________________________________

5.3 Seismic response

The dynamic model was run for the selected input time history. The computed response spec-
tra of the ground at a free-field location, i.e., the top of the bedrock downstream of the dam in
the model is compared with the response spectra of the input ground motion in Figure 8. As ex-
pected, the relatively stiff bedrock amplified the ground motion at short periods. The compari-
son between these spectra verifies that input ground motion into the dam is represented correct-
ly, as well as verifies the absence of model boundaries influence on the analysis. In addition,
Figure 8 presents the acceleration response spectrum at the crest (core). Spectral accelerations at
the crest for long periods have been amplified compared to the input motion, while shorter peri-
ods have been deamplified.

Figure 8. Comparison of input, free-field and crest acceleration response spectra.

5.4 Dynamic displacements

The model computed maximum horizontal displacement at the dam crest of 3.7 m and maxi-
mum vertical deformation of 2.0 m as shown in Figures 9 and 10. Maximum displacements are
located at the downstream slope near the dam crest. These predicted displacements are consid-
ered acceptable since the dam configuration after shaking is stable and freeboard has not suf-
fered a loss that compromises dam function.
The upstream filter, core, downstream filter and transition sequence experienced a range of
horizontal displacements between 1.2 to 2.0 m in the crest and up to 1.0 m of vertical displace-
ment. These displacements are considered to be acceptable as it is not expected that they could
affect continuity of core materials, which have an original width of 4 m.

Figure 9. Horizontal crest permanent displacements.

Figure 10. Vertical crest permanent displacements.

5.5 Permanent deformation from simplified procedures

Four simplified procedures to estimate permanent displacements are usually considered when
dealing with dam crest or slope deformations: Makdisi & Seed (1978), Jibson (2007), Bray &
Travasarou (2007) and Swaisgood (2013). These approaches are briefly described and used to
estimate displacement for the centerline tailings dam previously discussed in order to validate
their application to this specific dam.

5.5.1 Makdisi & Seed (1978)

This simplified methodology can be considered as the most widely used decoupled analysis, i.e.
the seismic response of the dam is determined separately from the deformation of the dam. De-
spite its rational approach and longevity in practice, this method and its suggested procedure of
estimating dam crest acceleration is outdated, yet it provides a good comparison for this study.
Makdisi & Seed’s procedure to estimate permanent displacements was developed using 3 rec-
ords from 2 different events and one synthetic accelerogram (M = 8.25). Although magnitude
measure was not specified by the authors, equivalent moment magnitudes correspond to 6.6 and
7.7 (PEER, 2005).A spreadsheet to estimate permanent deformation using Makdisi and Seed
formulation was developed by the authors and used in this estimation.

5.5.2 Jibson (2007)

Jibson, in 1993, proposed a formulation to estimate seismic displacements based on Newmark’s
sliding block. This proposal was revised in 1998, 2000 and updated in 2007 as a larger ground
motion database was available. A total of 2270 records from 30 different events with magni-
tudes between 5.3 and 7.6 were used to develop three equations to determine displacements as a
function of three parameters: (1) critical acceleration (a c), (2) peak ground acceleration (amax)
and (3) moment magnitude (M). Equation 2 shows the equation proposed by Jibson that relates
these parameters.
𝑎𝑐 2.335 𝑎 −1.478
𝑙𝑜𝑔(𝐷 ) = −2.71 log ((1 − ⁄𝑎𝑚𝑎𝑥 ) ( 𝑐⁄𝑎𝑚𝑎𝑥 ) ) + 0.424 𝑀 ± 0.454 (2)

where D = permanent displacement (cm)

5.5.3 Bray & Travasarou (2007)

A semi-empirical relationship was proposed by Bray and Travasarou based on a fully coupled
stick-slip sliding block model to simulate dynamic behavior of dams. This method predicts de-
viatoric slope displacements (i.e. shear). This method was developed considering a suite of in-
put ground motions from 41 shallow crustal earthquakes with magnitudes between 5.5 and 7.6
and comprising 688 time histories recorded at distances of less than 100 km.
This proposal considers two parameters: (1) Slope’s yield coefficient (k y), and initial funda-
mental period (TS) as representation of the dynamic strength and stiffness respectively. Equation
3 shows the predictive equation for estimating potential seismic displacements for earth and
waste structures. Additionally, Equation 4 presents a simplified formulation to calculate funda-
mental period for triangular-shaped sliding mass.
𝑙𝑛(𝐷) =
2
−1.10 − 2.83 𝑙𝑛(𝑘𝑦 ) − 0.333 (𝑙𝑛(𝑘𝑦 )) + 0.566 𝑙𝑛(𝑘𝑦 ) 𝑙𝑛(𝑆𝑎 (1.5𝑇𝑠 )) + 3.04 𝑙𝑛(𝑆𝑎 (1.5𝑇𝑠 )) −
2
0.244 (𝑙𝑛(𝑆𝑎 (1.5𝑇𝑠 ))) + 1.5𝑇𝑠 + 0.278(𝑀 − 7) (3)

2.6H
𝑇𝑠 = (4)
Vs
where D = permanent displacement (cm); ky = yield acceleration coefficient; TS = fundamental
period (s), H = dam height (m), VS = shear wave velocity through the dam (m/s); Sa(1.5 TS) =
spectral acceleration of the ground at a period equal to 1.5 times fundamental period (g).

5.5.4 Swaisgood (2013)

Swaisgood statistically analyzed historical cases of dam deformation between 1990 and 2003 to
relate crest settlements with peak ground acceleration (PGA) and earthquake magnitude (M).
This formulation was updated in 2003 based on a database of eighty two (82) earthquakes with
magnitudes between 5.3 and 9.0. These relations developed by regression analyses are shown in
Equations 5 and 6:
% Settlement = e(5.70 PGA+0.47 M−7.22) (5)
Settlement
% Settlement = x100 (6)
Dam height+Thickness of the foundation

where PGA = Peak ground acceleration (g); and M = Seismic moment magnitude.
5.6 Comparison of Simplified Deformation Analyses
Table 4 shows the four main parameters involved in simplified displacement estimation. Dam
height corresponds to the dam design considered by the authors for the present article; earth-
quake magnitude was obtained from deaggregation information for the area of study. Yield ac-
celeration coefficient and sliding mass height were obtained from limit equilibrium pseudostatic
analysis performed using Slide version 6.036 (Rocscience Inc., 2015) considering the method
suggested by Spencer for force and moment equilibrium. It is important to mention that failure
surface is along the downstream rockfill material. Finally, Arias Intensity of the input motion
was calculated using SeismoSignal (SeismoSoft, 2004). Table 5 presents the estimated perma-
nent displacements.

Table 4. Parameters used in simplified permanent displacement estimation.

__________________________________________________________________________________________________________
Parameter Value Methodology
__________________________________________________________________________________________________________
Dam height 230 m (a) and (d)
Earthquake magnitude 8.25 (a), (c) and (d)
Yield acceleration coefficient 0.17 g (a), (b), (c) and (d)
Peak ground acceleration 0.66 g (b) and (d)
__________________________________________________________________________________________________________
(a) Makdisi & Seed (1978)
(b) Jibson (2007)
(c) Bray & Travasarou (2007)
(d) Swaisgood (2013)

Table 5. Summary of simplified permanent displacement.

__________________________________________________________________________________________________________
Methodology Permanent Displacement (cm) Note
__________________________________________________________________________________________________________
Makdisi & Seed (1978) 20 (deviatoric) Average y/h vs. kmax/amax bound
Jibson (2007) 25 (deviatoric) Mean value
Bray & Travasarou (2007) 65 (deviatoric) Mean value
Swaisgood (2013) 350 (settlement) Mean value
__________________________________________________________________________________________________________

Figure 11 plots the variation of the horizontal and vertical displacements obtained from the
dynamic analysis described in the article for the crest and downstream slope, against the perma-
nent displacements estimated from simplified methods. Horizontal and vertical displacements
considered for the crest correspond to the downstream extreme area of crest, while slope dis-
placements represent the upper third area of the downstream slope.
Figure 11. Comparison between permanent displacements obtained from simplified procedures and the
dynamic analysis discussed herein.

6 CONCLUSIONS

This paper presents static and dynamic analyses to estimate permanent slope deformation of a
centerline tailings dam. For a large dam, such as the 230 m tall centerline dam considered in
this study, numerical analyses indicated significantly different permanent displacement com-
pared to assessments made with simplified analysis methods typically adopted in engineering
practice.
The three simplified procedures proposed by Makdisi and Seed, Jibson and Bray and Trava-
sarou, respectively, were developed to estimate shear induced displacements, but no volumetric
displacement is included in these methods. These predicted deviatoric displacements are gener-
ally consistent between simplified methods, but they are an order of magnitude less than the to-
tal crest displacement computed in the dynamic model.
Swaisgood’s empirical equation estimates the amount of crest settlement. The crest settlement
prediction, which inherently includes volumetric displacements, is reasonably consistent with
the total crest displacement computed in the dynamic model.
A key point highlighted in this study is that volumetric displacements may be significant and
should be considered in design. That is, screening level analysis aimed at predicting deviatoric
displacements (e.g. Makdisi and Seed, Jibson, and Bray and Travasarou methods) should be
supplemented with assessment of volumetric displacements, such us as using Swaisgood (2013)
or estimated with Tokimatsu and Seed (1987).
The various simplified procedures were developed from data that does not match well with
the conditions considered for the dam, i.e. centerline dam 230 m height, steep slopes (1.6 H : 1.0
V), subduction failure mechanism as earthquake source, 8.25 earthquake magnitude, etc. How-
ever, these methods are regularly adopted in engineering practice as screening-level and some-
times design-level calculations.
Dynamic analyses require a much greater level of care to conduct compared with simplified
analyses, including a detailed understanding of seismic hazard for a site, but offer a more accu-
rate understanding of seismic behavior. Simplified analyses are appropriate for screening-level
assessments for assessing dam displacements, but detailed design decisions should be based on
more advanced analyses, especially when the dam and seismic setting push on boundaries of
applicability for simplified methods.
REFERENCES

Abrahamson, N. & Silva, W. 2008. Summary of the Abrahamson & Silva NGA ground-motion relations.
Earthquake Spectra, 24(1): 67-97.
ASCE. 2005. Minimum design loads for building and other structures. 7-05, American Society of Civil
Engineers.
Bilek, S. 2009. Seismicity along the South American subduction zone: Large earthquakes, tsunamis, and
subduction complexity. Technophysics, doi: 10.1016/jtecto.2 009.02.037.
Boomer, J., & Acevedo, A., 2004. The use of real earthquake accelerograms as input to dynamic analysis.
Journal of Earthquake Engineering, 8(1).
Boore, D., & Atkinson, G. 2008. Ground-motion prediction equations for the average horizontal compo-
nent of PGA, PGV, and 5%-damped PSA at spectral periods between 0.01 s and 10.0 s. Earthquake
Spectra, 24(1): 99–138.
Bray, J., & Travasarou, T. 2007. Simplified procedure for estimating earthquake-induced deviatoric slope
displacements. Journal of Geotechnical and Geoenvironmental Engineering, ASCE, 135: 1336-1340.
Campbell, K., & Bozorgnia, Y. 2008. NGA ground motion model for the geometric mean horizontal
component of PGA, PGV, PGD and 5% damped linear elastic response spectra for periods ranging
from 0.01 to 10 s. Earthquake Spectra, 24(1): 139-171.
Hancock, J., Watson-Lamprey, J., Abrahamson, N., Bommer, J., Markatis, A., McCoy, E. & Mendis, R.
2006. An improved method of matching response spectra of recorded earthquake ground motion using
wavelets. Journal of Earthquake Engineering, 10(1). Imperial College Press, London.
Itasca Consulting Group, Inc. 2010. FLAC – Fast Lagrangian Analysis of Continua, Ver. 7. Minneapolis:
Itasca.
Jibson, R., 2007. Regression models for estimating coseismic landslide displacement. Engineering Geo-
logy, 91 (2-4), 209-218.
Kramer, S., 1996. Geotechnical earthquake engineering. Prentice-Hall, Inc., Upper Saddle River, New
Jersey, 1996.
Kihlemeyer, R.L. & Lysmer J., 1993. Finite element method accuracy for wave propagation problems. J.
of Soil Mechanics and Foundations. Div. ASCE, 99(SM5): 421-427.
Makdisi, F. & Seed, H. 1978. Simplified procedure for estimating earthquake-induced deformations.
Journal of Geotechnical Engineering Division, ASCE, 104 (GT7): 849-867.
McVerry G., Zhao J., Abrahamson N. & Somerville, P. 2006. New Zealand acceleration response spec-
trum attenuation relations for crustal and subduction zone earthquakes. Bulletin of the New Zealand
Society for Earthquake Engineering, 39(1).
MEM. 1998. Guía ambiental para la estabilidad de taludes de depósitos de desechos sólidos de mina.
Monroy, M., Hull, A., Martinez, M. & Bolaños, A. 2011. Estimation of parameters for seismic design in
Peru. In: 5th International Conference on Earthquake Geotechnical Engineering. Santiago, Chile. Janu-
ary, 2011.
NIST (National Institute of Standards and Technology). 2011. Selecting and scaling earthquake ground
motions for performing response-history analyses.
PEER, 2005. PEER Strong Motion Database.
Pretell, A.R., Borja, R.N. & Parra, D. 2012. Comparación de curvas del módulo de corte con resultados
triaxiales cíclicos y análisis de respuesta sísmica. XX National Congress of Civil Engineering and VIII
International Congress of Civil Engineering. Huaraz, Peru. August, 2012.
Rocscience Inc., 2015. Software for Slope Stability Modeling.
Rollins, K., Evans, M., Diehl, N. & Daily, III W. 1998. Shear modulus and damping relationships for
gravels. Journal of Geotechnical and Geoenvironmental Engineering, 124(5), Paper No. 11141.
SENCICO. Norma Sísmica Peruana E.030-2003. Diseño Sismorresistente. Norma Técnica de Edificación.
Seed, B. & Idriss, I. 1970. Soil moduli and damping factors for dynamic response analysis. Report No.
EERC 70-10, Earthquake Engineering Research Center, Univ. of California at Berkeley, Berkeley, Cal-
ifornia.
Seed, B., Wong, R., Idriss, I. & Tokimatsu, K. 1984. Moduli and damping factors for dynamic analyses of
cohesionless soils. Report No. UCB/EERC-84/14. Earthquake Engineering Research Center. Universi-
ty of California, Berkeley, CA.
SeismoSoft, 2004. SeismoSignal – A computer program for signal processing of strong-motion data.
SeismoSoft, 2013. SeismoMatch – Acomputer program for spectrum matching of earthquake records.
Spencer, E., 1967. A Method of Analysis of the Stability of Embankments Assuming Parallel Inter-slice
Forces, Geotechnique, 17: 11-26.
Stewart, J., Chiou, S., Bray, J., Graves, R., Somerville, P. & Abrahamson, N. 2001. Ground motion evalu-
ation procedures for performance-based design. Pacific Earthquake Engineering Research Center, No.
PEER-2001/09. University of California.
Swaisgood, J.R. 2013. Predicting dam deformation caused by earthquakes – An update. 2003 Pacific
Earthquake Engineering.
Tokimatsu, K. and Seed, H. 1987. Evaluation of Settlements in Sands due to Earthquake Shaking. Journal
of Geotechnical Engineering, 113(8), 861-878.
Vrymoed, J. 1981. Dynamic FEM model of Oroville Dam. ASCE Journal of the Geotechnical Engineer-
ing, 107(GT-8).
Vucetic, M.3 & Dobry, R., 1991. Effect of soil plasticity in cyclic response. ASCE, Journal of Geotech-
nical Engineering, 117(1): 89-107.
Youngs, R., Chiou, S., Silva, W. & Humphrey, J. 1997. Strong ground motion attenuation relationship for
subduction zone earthquakes. Seismological Research Letters, 68(1).
Zhao, J., Zhang, J., Asano, A., Ohno, Y., Oouchi, T., Takahashi, T., Ogawa, H., Irikura, K., Hong, K.,
Thio, Somerville, P. & Fukushima, Y. 2006. Attenuation relations of strong ground motion in Japan
using site classification based on predominant period. Bulletin of the Seismological Society of America,
96:898-913.