ARTICLE IN PRESS

Soil Dynamics and Earthquake Engineering 28 (2008) 851–865

www.elsevier.com/locate/soildyn

Seismic damage evaluation in urban areas using the capacity spectrum

method: Application to Barcelona
Alex H. Barbata,, Luis G. Pujadesb, Nieves Lantadab
a
Department of Structural Mechanics, Civil Engineering School, Technical University of Catalonia, Edificio C1, Campus Norte UPC,
Jordi Girona 1-3, 08034 Barcelona, Spain
b
Department of Geotechnical Engineering and Geosciences, Civil Engineering School, Technical University of Catalonia, Edificio D2,
Campus Norte UPC, Jordi Girona 1-3, 08034 Barcelona, Spain

Received 6 May 2006; received in revised form 5 July 2007; accepted 11 October 2007

Abstract

Conceptual aspects related to seismic vulnerability, damage and risk evaluation are discussed first, together with a short review of the
most widely used possibilities for seismic evaluation of structures. The capacity spectrum method and the way of obtaining seismic
damage scenarios for urban areas starting from capacity and fragility curves are then discussed. The determination of capacity curves for
buildings using non-linear structural analysis tools is then explained, together with a simplified expeditious procedure allowing the
development of fragility curves. The seismic risk of the buildings of Barcelona, Spain, is analyzed in the paper, based on the application
of the capacity spectrum method. The seismic hazard in the area of the city is described by means of the reduced 5% damped elastic
response spectrum. The information on the buildings was obtained by collecting, arranging, improving and completing a broad database
of the dwellings and current buildings. The buildings existing in Barcelona are mainly of two types: unreinforced masonry structures and
reinforced concrete buildings with waffled-slab floors. The ArcView software was used to create a GIS tool for managing the collected
information in order to develop seismic risk scenarios. This study shows that the vulnerability of the buildings is significant in Barcelona
and, therefore, in spite of the low-to-moderate seismic hazard in the region, the expected seismic risk is considerable.
r 2007 Elsevier Ltd. All rights reserved.

Keywords: Seismic hazard; Seismic vulnerability; Seismic risk; Damage evaluation; Capacity spectrum; Fragility curve; Damage probability matrix; Risk
scenario

1. Introduction scale, for instance, an urban area. There are a number of

methodologies to estimate the vulnerability, damage and
The major part of losses due to earthquakes has its risk in seismic areas (e.g. Ref. [1–3]) and all of them have
origin in the deficient seismic behavior of structures. In difficulties arising basically from the lack or low quality of
spite of the advances of research in earthquake engineering available data when they are applied in low-to-moderate
in general and on seismic design codes in particular, seismic areas.
catastrophic losses have occurred recently in many Damage probability matrices, vulnerability functions
countries in the world even in those where earthquake and fragility curves, obtained from structural damages
engineering studies are priority tasks. The aim of risk observed during earthquakes, were the preferred tools in
studies is to estimate and map the expected damage in the seismic risk studies in the past [1,4–6]. A complete
structures due to a specified earthquake at a territorial observed damage database would be necessary for applying
such approaches; however, this is only possible in high-
seismicity areas where properly performed post-earthquake
Corresponding author. Tel.: +34 934016496; fax: +34 934011048. surveys are available. In areas where the data for the
E-mail addresses: [email protected] (A.H. Barbat), existing building typologies and for the seismic intensities
[email protected] (L.G. Pujades). are limited or incomplete, local expert opinion has been

0267-7261/$ - see front matter r 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.soildyn.2007.10.006
 ARTICLE IN PRESS
852 A.H. Barbat et al. / Soil Dynamics and Earthquake Engineering 28 (2008) 851–865

used to support or completely replace probabilistic 2. Capacity spectrum-based method

processing of the observed data [2,7]. Building modeling
and non-linear structural analysis are other methods to The capacity spectrum method has been used in this article
stand in for the shortage of data [8–10]. In areas without to estimate the expected performance of the buildings of
any available damage database, the information obtained Barcelona when subjected to selected earthquakes scenarios.
in other similar areas was applied, but at the same time A summary of displacement-based methods of earthquake
using an expert judgment [11]. To complete the earthquake damage estimation can be found in Calvi et al. [20].
damage information in areas with lack of data, Monte In this article, the earthquake ground motion is modeled
Carlo simulation procedures have been used in the past by means of 5% damped elastic response spectra in
[12,4–6]. Accordingly, the probabilistic analysis of compu- ADRS format according to the procedure given by Faccioli
ter-generated structural responses, obtained by using non- [21]. Deterministic and probabilistic hazard scenarios are
linear analysis procedures of representative buildings, has considered.
provided fragility curves, damage probability matrices and Capacity curves mainly depend on the structural design
vulnerability functions. and on the construction practice. Reliable structural models
There are certain aspects of the seismic damage have been used, allowing calculation of capacity curves
evaluation of an urban area which have to be pointed which consider the constructive peculiarities of the existing
out: (1) uncertainties are high in each step of seismic risk buildings. A non-linear macro-element model, representative
evaluation, particularly in the evaluation of the seismic of a whole masonry panel, proposed by Gambarotta and
hazard in low-to-moderate seismic areas and of the Lagomarsino [22], has been used to model the unreinforced
vulnerability of existing buildings. It is not the purpose of masonry buildings of Barcelona. This model permits
this paper to perform a probabilistic study in the strict representing, with a limited number of degrees of freedom,
sense, but to perform analyses based on average or the two main masonry failure modes, which are based on
most likely values. (2) For management purposes, risk bending-rocking and shear-sliding mechanisms. It is one of
requires a multidisciplinary evaluation that takes into the few existing models describing the seismic behavior of
account not only the expected physical damage, the masonry buildings which considers, by means of internal
number and type of casualties or the economic losses, but variables, the shear-sliding damage evolution, the strength
also the conditions related to social fragility and lack of deterioration and the stiffness degradation. Equivalent
resilience, which favor the indirect effects when a hazard column-and-beam models have been used to assess the
event strikes a urban center [13,14]. In this paper, we study expected damage of reinforced concrete buildings with
only the physical risk of urban areas. (3) The most recent waffled-slab floors. Capacity curves were obtained in this
trends in the field of vulnerability evaluation for risk case by using the computer code RUAUMOKO [23].
analysis operate with simplified mechanical models essen- Structures were modeled by means of several plane frames
tially based on the capacity spectrum method [3,15,16] connected to one another. The rigid diaphragm effect was
and this will be the method used in this paper. The method considered by constraining the nodes belonging to the same
permits evaluating the expected seismic performance of storey. The obtained capacity curves have been represented
structures by comparing, in spectral coordinates, their in the same spectral acceleration (Sa)–spectral displacement
seismic capacity with the seismic demand, described by (Sd) domain as the demand spectrum. Finally, these curves
acceleration–displacement response spectra (ADRS) ade- have been described in their bilinear form defined by yielding
quately reduced in order to take into account the inelastic (Dy, Ay) and ultimate (Du, Au) points. Crossing the demand
behavior [17,18]. and capacity spectra, the performance point is determined
The seismic risk of the city of Barcelona (Spain) is and thus the expected spectral displacement. The perfor-
analyzed in this paper. This city is located in an area of mance point is calculated by using the iterative method
low-to-moderate seismic hazard [19], but its buildings have (procedure A) of ATC-40 [24].
a high vulnerability. Most of the residential buildings have In order to analyze the seismic damage, we considered
been designed and built without the consideration of any for the buildings in Barcelona five damage states ds: none,
earthquake resistant criterion and many of their particular slight, moderate, severe and complete. For unreinforced
features, typical for the constructive techniques of the city, masonry buildings and reinforced concrete buildings with
have been identified as potential damage sources. Detailed waffled-slab floors, these damage states have the same
information on their design has been obtained through the meaning as in HAZUS (2002) but the complete damage
years by collecting, arranging, improving and completing grade includes the very heavy damage and the destruction
the database of the housings and buildings of the city. It is states (grades 4 and 5) corresponding to the 1998 European
important to note that 97% of the housings in Barcelona Macroseismic Scale, EMS’98 [25]. A weighted average
are unreinforced masonry and reinforced concrete with damage index, DSm, can be calculated as
waffled-slab floors buildings. In order to develop damage
and risk scenarios, capacity spectra and fragility curves X
4
have been developed and are used to simulate earthquake DSm ¼ dsi P½dsi , (1)
risk scenarios. i¼0
 ARTICLE IN PRESS
A.H. Barbat et al. / Soil Dynamics and Earthquake Engineering 28 (2008) 851–865 853

where dsi takes the values 0, 1, 2, 3 and 4 for the damage

states i considered in the analysis and P[dsi] are the
corresponding occurrence probabilities. DSm is close to
the most likely damage state of the structure. Table 1
shows the most probable damage grade as a function of the
average damage index.
This damage index is useful for mapping and analyzing
damage distributions by using a single parameter. Of
course, alternative maps may plot the spatial distribution
of the probability of occurrence of a specified damage state
dsi, that is P[dsi].
Fragility curves define the probability that the expected
global damage d of a structure exceeds a given damage
state dsi, as a function of a parameter quantifying the Fig. 1. Damage state thresholds from capacity spectrum.
severity of the seismic action. Here, this parameter is
the spectral displacement Sd. Thus, for each damage state, Table 2
the corresponding fragility curve is completely defined by Damage state thresholds (see Fig. 1)
plotting P[dXdsi] as ordinate and the spectral displacement
Sd as abscissa. If we assume that fragility curves follow a Sd1 ¼ 0:7Dy Slight
lognormal probability distribution, they are completely Sd2 ¼ Dy Moderate
Sd3 ¼ Dy þ 0:25ðDu  Dy Þ Severe
defined by only two parameters, which are the mean
Sd4 ¼ Du Complete
spectral displacement Sddsi and the corresponding standard
deviation bdsi . Thus, for a given damage state dsi, a fragility
curve is well described by the following lognormal
probability density function: Table 3
Probabilities of the expected damage states when fixing a 50% probability
  " !#
for each damage state: (1) slight, (2) moderate, (3) severe and (4) complete
dsi 1 Sd
P ¼F ln , (2)
Sd bdsi Sddsi Condition mD Pb (1) Pb (2) Pb (3) Pb (4)

Pb (1) ¼ 0.5 0.911 0.500 0.119 0.012 0.00

where Sddsi is the threshold spectral displacement at which Pb (2) ¼ 0.5 1.919 0.896 0.500 0.135 0.008
the probability of the damage state dsi is 50%, bdsi is the Pb (3) ¼ 0.5 3.081 0.992 0.866 0.500 0.104
standard deviation of the natural logarithm of this spectral Pb (4) ¼ 0.5 4.089 1.000 0.988 0.881 0.500
displacement, F is the standard normal cumulative
distribution function and Sd is the spectral displacement.
Fragility curves can be obtained in a simplified way 1
starting from the bilinear representation of the capacity
curves. Fig. 1 and Table 2 show how the thresholds Sddsi
Probability of exceedance

0.75
are obtained in this case in function of the yielding
displacement Dy and the ultimate displacement Du of the No damage
structure. 0.5
Concerning bds, it is well known that the expected Slight

seismic damage in buildings follows a binomial probability Moderate

distribution [25]. Therefore, it is assumed that at the Sdds 0.25 Severe
threshold, the probability of this damage state is 50%.
Complete
Table 3 shows the cumulative expected probabilities of all
the damage states when a particular damage state
probability is fixed to 50% and a binomial probability Sd1 Sd2 Sd3 Sd4
Spectral displacement
Table 1
Damage states and mean damage index values Fig. 2. Example of fragility curves construction.

Mean damage index intervals More probable damage state

distribution is assumed. The parameter mD controls the
0–0.5 No damage assumed binomial probability distribution.
0.5–1.5 Slight damage Finally, the function expressed by Eq. (2) is fitted to the
1.5–2.5 Moderate damage obtained points by means of a least-square criterion. Fig. 2
2.5–3.5 Severe damage
shows an example of fragility curves obtained by using this
3.5–4.0 Complete damage
method.
 ARTICLE IN PRESS
854 A.H. Barbat et al. / Soil Dynamics and Earthquake Engineering 28 (2008) 851–865

Finally, for each hazard scenario and for each class buildings of Barcelona are not moment resisting frames,
of building, damage probability matrices can be obtained but they consist of columns and slabs in their waffled-slab
by entering with the spectral displacement corresponding floors version, which is a structural type not adequate for
to the performance point into the corresponding fragility seismic areas due to their low ductility. Most of them also
curves. have a soft first storey. The Spanish code limits their
ductility factor to 2, while earthquakes like that of Kokaeli,
3. Description of the urban area of Barcelona Turkey, 1999, have dramatically shown the high seismic
vulnerability of this kind of buildings. In the seismic areas
3.1. The building types of Europe, the seismic design of reinforced concrete
buildings varies extremely and structures show a large
The most representative type of buildings of the central variation of earthquake resistance. Accordingly, the EMS-
part of Barcelona is the unreinforced masonry one, whose 98 scale [25] assigns a very wide range of vulnerability to
large number greatly influences the overall seismic vulner- the framed reinforced concrete buildings used in Europe,
ability of the city. Only a part of the structures of this area which covers the whole vulnerability range from buildings
are made of reinforced concrete, substituting demolished without earthquake resistant design to buildings design
unreinforced masonry buildings, but they are found in a with high-level seismic codes. In an extreme case, their
significant number in other more recently built districts of vulnerability can be comparable with that of low-quality
the city. The most emblematic and representative district of unreinforced brick masonry buildings. The reinforced
Barcelona is the ‘‘Eixample’’, in the central part of the city, concrete buildings of Barcelona fall within the high
designed in the middle of the nineteenth century and vulnerability part of the EMS-98 scale, for which
special attention has been paid in this study to the seismic significant damage for relatively low seismic intensities is
risk evaluation of this area. Covering about 750 ha, it expected.
consists of square blocks sizing about 113 m  113 m. These In general, the buildings of Barcelona are part of
blocks are perfectly aligned and are beveled in their vertices aggregates, forming building blocks. In the past, designers,
by edges of about 20 m. The construction of this district architects and builders have not been careful at all when
took place between 1860 and 1950, with an average of 25 joining new buildings to older ones. Important differences
buildings for each block, which were designed only for in the number of storeys and in the level of the floors are
vertical static loads, without any consideration of seismic frequent within a block. Although there is no possibility of
design criteria. pounding because adjacent buildings have a common wall,
The unreinforced masonry buildings of Barcelona are this characteristic increases their seismic vulnerability. The
tall and have openings of considerable size in their walls, seismic analysis of a whole block is a complex problem
which affect their vulnerability increased even more by which is not an objective of this paper.
long walls without perpendicular stiffening. Their particu- Detailed information on the design and construction of
lar features, typical for the constructive techniques of the the buildings of Barcelona has been obtained through the
city at that time, have been identified as potential damage years by collecting, arranging, improving and completing
sources. As an example, the floors of these unreinforced the database of the housings and buildings of the city.
masonry buildings are made of wooden, steel or precast Three main information sources were used in obtaining
concrete beams with small ceramics vaults in between, data for the risk assessment at urban scale: (1) a territorial
according to the building period, showing a poor stiffness information system [26], (2) a file with the construction
both to bending moment and to axial forces. Another year of each building, and (3) a file containing information
example is that almost all of these buildings have two soft about the building types. The territorial information
storeys, due to the greater height of their first two floors. system contains the most important data, namely the
Furthermore, cast iron columns were used in many cases cadastre information. The total number of cadastre units
instead of masonry walls at the base and ground floors, or lots in the city is 80,715 and they may contain a building
reducing even more their lateral stiffness, because their or may be empty. This information was used to obtain
upper and lower edges are not perfectly clamped. Similar the geometry and the core features of the buildings of the
masonry buildings can also be found in many other studied area, like perimeters and numbers of storeys of
European and Mediterranean cities. In some of them, each built lot. A set of blocks composes the so-called small
certain measures of seismic protection have been applied in statistical zones, which are used for administrative purposes
the past, but there are also many buildings in such poor and are the basis for the census. These census zones,
conditions that they have to be classified in the highest numbering 248, have been used to map risk scenarios; a
vulnerability class of the European Macroseismic Intensity relatively small number of them make up a neighborhood,
Scale, EMS-98 [25]. whose total number is 38, and a small number of
Since the middle of the twentieth century, the number neighborhoods make up a district whose total number
of reinforced concrete buildings increased significantly in is 10.
Barcelona, becoming nowadays the most frequent typology According to the official statistics of Barcelona corre-
for new buildings. Most of the reinforced concrete sponding to the year 2001 [27], Barcelona has about 1.566
 ARTICLE IN PRESS
A.H. Barbat et al. / Soil Dynamics and Earthquake Engineering 28 (2008) 851–865 855

millions inhabitants, 700,000 housings and 69,000 build- Table 4

ings, with an average of about 2.24 inhabitants in each. The Building types in Barcelona
Municipality of Barcelona provided the cadastre database, Code Building description
which is well described in the report Infocca [26], and
detailed information concerning the age and typology of M31 Unreinforced masonry buildings (UMB) with wooden floors.
the buildings. The collected data allowed complete M32 UMB with masonry vaults.
M33 UMB with steel beams and masonry vaults.
characterization of the geometrical features and geogra-
M34 UMB with reinforced concrete (RC) beams and masonry vaults.
phical location together with the type and year of RC32 RC buildings with irregular structural system, irregular infill and
construction of about 63,000 buildings (more than 91% soft/weak storey.
of the total number of buildings), which mainly correspond S1 Steel buildings with moment resisting frames.
to residential buildings. For the others, that is, about 6000 S2 Steel buildings with braced frames.
S3 Steel buildings with frames with unreinforced masonry infill walls.
buildings, there is a lack of information about one or more S5 Steel and RC composite buildings.
of the mentioned characteristics. There are 60,653 un-
reinforced masonry and reinforced concrete buildings,
which represent the 96.3% of the total number of 3.2. Seismic hazard
residential buildings and only these have been included in
the simulations performed in this study. Barcelona, city located in the Northeastern part of
Fig. 3 shows the age distribution of the buildings as a Spain, has a low-to-moderate seismicity and weak tectonic
function of the constructive periods in Spain that have been motions. Deterministic and probabilistic hazard scenarios
defined based on earthquake-resistant considerations. It have been developed by the geological survey of Catalonia
can be seen that almost the 80% of the building stock of [30–33] following the guidelines of the RISK-UE project
Barcelona was constructed prior to the first Spanish ([21]; see also Refs. [34,35]). The seismic action is defined in
Seismic Code [28]. Fig. 4 shows the building typology terms of elastic response acceleration spectra. The deter-
distribution. Table 4 describes these building types. These ministic magnitude–distance scenario corresponds to a
classification codes have been developed within the RISK- historical earthquake occurred quite near to the city on 25
UE European project to describe, in a detailed way, all the May 1448 with MS=5.1 and the probabilistic one has a 475
building types identified in Europe [29]. As it can be seen in years return period and corresponds to the 1428 Pyrenees
Fig. 4, 97% of buildings in Barcelona are unreinforced earthquake with MS=6.5 [36]. The seismic zoning [37,38]
masonry buildings and reinforced concrete buildings. has been taken into account to obtain specific response
spectra for areas with different soil characteristics of
Barcelona. Fig. 5 shows the four main seismic zones of
1950-1962 the city: zone R corresponds to rock outcrops; soils in zone
17.3%
I are soft, while zones II and III are made of intermediate
soils. Typical shear-wave velocities for these zones are 800,
<1950 225, 394 and 405 m/s, respectively.
50.7% 1963-1968 A simplified analytical equation based on the Eurocode
10.9%
EC8 has been used to fit average response spectra:
8  
1969-1974 > T
1995-1996 >
> ag 1 þ ðBC  1Þ 0pTpT B ;
1975-1994 9.8% >
> TB
0.2% >
>
11.1% >
> T B pTpT C ;
< ag B
> C
 
Fig. 3. Distribution of buildings by age. Sa ðTÞ ¼ T C (3)
>
> ag BC T C pTpT D ;
>
> T
>
>  
>
> T CT D
M32 >
> ag BC T D pTp4 s;
:
1% T2
M33
M31 28%
32%
where Sa(T) is the spectral acceleration; T is the period; ag
is the peak ground acceleration; BC is an amplification
factor defined as S amax =ag ; TB and TC are the limits of the
constant acceleration interval; and TD defines the begin-
ning of the constant displacement section.
S5 Fig. 6 shows the smoothed response spectra for the four
2% soil types, both for the deterministic and probabilistic
M34
S1/S2/S3 cases. The differences between the probabilistic and
RC32 18%
1% deterministic response spectra are due the fact that
18%
PSHA-derived spectrum includes the strong influence of
Fig. 4. Distribution of the buildings by typology. the variability in the ground motion prediction equation
 ARTICLE IN PRESS
856 A.H. Barbat et al. / Soil Dynamics and Earthquake Engineering 28 (2008) 851–865

whereas the deterministic case corresponds to the median building typology matrix with only six models, which
ground motion for the given seismic hazard scenario reasonably represents the wide number of buildings of the
[39,40]. Table 5 contains the parameters of the smoothed city. Selected, representative, existing buildings have been
spectra according to Eq. (3) [30,32]. then analyzed by using simplified structural analysis
models but adequately describing their seismic behavior.
4. Structural capacity of the buildings of Barcelona The seismic performance of a building can be characterized
by its capacity spectrum obtained by means of a pushover
An important objective of this paper was to produce analysis [24], modeled in its bilinear form.
building-by-building risk scenarios for Barcelona, which Detailed structural plans have been used to model
required the seismic characterization of about 60,653 representative buildings for low-rise (two storeys, 5.2 m
residential buildings. It was decided to adopt a simplified tall) mid-rise (five storeys, 15.8 m tall) and high-rise (eight
storeys, 24.0 m tall) reinforced concrete buildings. Capacity
curves were obtained by performing non-linear static
analyses using the 2D version of the computer code
RUAUMOKO [23]. Structures were modeled by means
of several plane frames connected one to other. The rigid
diaphragm effect was considered by constraining the nodes
belonging to the same storey. High-rise and mid-rise
buildings have a rectangular floor size of 25.65 m  21.90 m
while the low-rise one has a 13.6 m  13.6 m floor area. The
following mean mechanical properties have been assumed:
concrete compression strength fck ¼ 20 MPa, steel yield
stress fy ¼ 510 MPa, elastic modulus Ec ¼ 30 GPa, and shear
modulus G ¼ 12.5 GPa.
In a similar way, based on detailed structural plans, three
unreinforced masonry buildings of the Eixample district of
Barcelona have been modeled. An old building, but still in
use today, located in the ‘‘Eixample’’ district, was used as a
sample to develop masonry building models. The analyzed
unreinforced masonry models correspond to two storeys
(low-rise), four storeys (mid-rise) and six storeys (high-rise)
buildings of Barcelona. Both the mid- and high-rise
buildings have the same floor size (18.9  24.5 m) but
different height (17 m and 24 m). The low-rise building has
a 7.3 m  9.3 m in floor and is 6.65 m tall. This last case
corresponds to single-family houses of the residential areas
of Barcelona. All the analyses have been performed with
Fig. 5. Seismic zonation of Barcelona [37,38]. TreMuri program [41], a useful tool to study the non-linear

0.6 0.6

0.5 0.5

0.4 Deterministic scenario 0.4 Probabilistic scenario

Sa (g)

Sa (g)

0.3 0.3
Zone R Zone R
Zone I Zone I
0.2 Zone II 0.2 Zone II
Zone III Zone III

0.1 0.1

0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2

Period (s) Period (s)

Fig. 6. Smoothed 5% damped response spectra for the deterministic (left) and probabilistic (right) scenarios [31].
 ARTICLE IN PRESS
A.H. Barbat et al. / Soil Dynamics and Earthquake Engineering 28 (2008) 851–865 857

Table 5
Parameters for the 5% damped elastic response spectra for the deterministic and probabilistic scenarios (see Eq. (3)) [30]

Deterministic scenario Probabilistic scenario

Zone R Zone I Zone II Zone III Zone R Zone I Zone II Zone III
2
ag (cm/s ) 71 133 138 120 98 184 190 166
TB (seg) 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1
TC (seg) 0.23 0.39 0.22 0.22 0.25 0.40 0.23 0.19
TD (seg) 1.75 2.30 2.20 2.00 1.75 2.85 2.21 1.77
BC 2.26 1.91 2.45 2.29 2.29 2.0 2.50 2.57
RMS Sa (g) 0.01 0.03 0.03 0.05 0.01 0.03 0.04 0.06
RMS Sd (cm) 0.08 0.12 0.11 0.14 0.07 0.19 0.19 0.18

Table 6
Yield and ultimate capacity for reinforced concrete (RC) and unreinforced masonry (M) buildings

Building class No. of storeys range Period (s) Yield capacity Ultimate capacity

Dy (cm) Ay (g) Du (cm) Au (g)

Low-rise, RC 1–3 0.47 0.70 0.13 5.24 0.14

Mid-rise, RC 4–6 0.83 1.42 0.08 5.11 0.12
High-rise, RC 7+ 1.14 1.89 0.06 4.67 0.08
Low-rise, M 1–2 0.13 0.27 0.65 1.36 0.56
Mid-rise, M 3–5 0.44 0.63 0.13 2.91 0.12
High-rise, M 5+ 0.51 0.68 0.11 2.61 0.08

in-plane mechanical behavior of masonry panels and to the modeled reinforced concrete and masonry buildings.
assess the expected damage for masonry buildings due to The number of storeys ranges for the corresponding
earthquakes by means of 2D and 3D models. In spite of the building classes has been also included in this table.
local flexural behavior of floors and walls, the out-of-plane Fig. 7 shows the bilinear capacity spectra for reinforced
response was not included in the analysis because its effect concrete buildings and unreinforced masonry buildings
on the global building response was not considered to be together with the 5% damped elastic response spectra in
significant in this case, in which vertical and horizontal ADRS format for the deterministic and probabilistic
structural elements are not properly connected. The use of scenarios. Observe that the crossing points are perfor-
these tools guaranteed the computation of fragility curves mance points only when they belong to the linear branch of
and damage probability matrices for more than 95% of the the capacity curves. But even when they are on the non-
residential building stock of the city, allowing developing linear branch, a graphic estimate of the performance point
representative risk scenarios which was the main goal of can be visualized by taking into account the equivalent
the study. linear displacement method. This fact becomes important
The highest expected period of the residential buildings when evaluating damage by using fragility curves because it
existing in Barcelona is about one second and it greatly influences on the damage probability matrices.
corresponds to high-rise reinforced concrete buildings. As A significant ductility can be observed for mid-rise and
capacity curves are based on the assumption that the high-rise masonry buildings. This fact should be related to
response of the structure is well represented by the an excessive slenderness of these buildings and also to the
fundamental mode of vibration, they describe adequately failure criterion used in the pushover analysis performed
the seismic behavior of buildings with a fundamental by means of the model proposed by Gambarotta and
period lower than 1 s [24]. Consequently, for the purpose of Lagomarsino [22], adequate to the masonry buildings
this study, which is the calculation of seismic risk scenarios of the type existing in Barcelona. It can be also seen in
for the whole city, capacity curves, in spite of their Fig. 7 how the capacity decreases with the height of
limitations, provide a reasonably good structural damage the building both for masonry and for RC buildings.
description for the residential buildings of Barcelona. The capacity spectrum for low-rise masonry buildings
Pushover analyses allowed the capacity curves for each shows high stiffness and strength. In fact, this type of
building class to be determined and, starting from these building, representative for a number of one-family houses,
curves, capacity spectra have been obtained [24]. Table 6 mainly located in the residential districts of the city, is
shows the fundamental period and the yield and ultimate completely different from the mid- and high-rise masonry
capacity points defining the bilinear capacity spectra for buildings.
 ARTICLE IN PRESS
858 A.H. Barbat et al. / Soil Dynamics and Earthquake Engineering 28 (2008) 851–865

0.6 0.6
Low-rise (M) Deterministic scenario

0.5 0.5
Deterministic scenario Zone R
Zone I
0.4 0.4 Zone II
Zone R Zone III
Zone I Low-rise (RC)
Sa (g)

Sa (g)
0.3 Zone II 0.3
Zone III Mid-rise (RC)

0.2 0.2 High-rise (RC)

Mid-rise (M)

0.1 0.1
High-rise (M)
0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Sd (cm) Sd (cm)

0.6 0.6
Low-rise (M) Probabilistic scenario Zone R
0.5 0.5 Zone I
Zone II
Probabilistic scenario Zone III
0.4 0.4
Low-rise (RC)
Sa (g)
Sa (g)

Zone R
0.3 0.3
Zone I Mid-rise (RC)
Zone II
Zone III 0.2
0.2 High-rise (RC)
Mid-rise (M)

0.1 0.1
High-rise (M)
0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2 2.5 3
Sd (cm) Sd (cm)

Fig. 7. Bilinear capacity spectra for reinforced concrete buildings (RC) and unreinforced masonry buildings (M). The 5% damped elastic response spectra
in ADRS format are also shown for the deterministic and probabilistic cases.

5. Fragility curves and damage probability matrices for the buildings with waffled-slab floors are really weak. This is a
buildings of Barcelona reasonable result when taking into account the particular
structural type described above. In any case, for each
Specific fragility curves have been developed for both hazard scenario, damage probability matrices strongly
masonry and reinforced concrete buildings of Barcelona. depend on the spectral displacement of the performance
Fig. 8 shows the obtained curves. Table 7 shows the values point. Table 8 shows these matrices for the masonry
of the parameters Sdi and bi, where i ¼ 1, y, 4, which buildings of Barcelona. Both scenarios, deterministic and
define the corresponding cumulative lognormal distribu- probabilistic, are considered for the four seismic zones of
tion (see Eq. (2)). It can be observed that reinforced the city and for the three building classes corresponding to
concrete buildings are more ductile than the masonry ones, low-rise, mid-rise and high-rise buildings. Similar matrices
showing a better seismic performance. For example, for have been obtained for reinforced concrete buildings (see
mid-rise reinforced concrete buildings in Fig. 8, in case of a Table 9).
4 cm spectral displacement, the expected probability for the In Tables 8 and 9, DSm is the weighted average damage
complete damage state is about 30%, but it is more than state calculated by using Eq. (1), which can be considered
60% for unreinforced masonry buildings. Fortunately, close to the most likely damage state of the structure. These
Barcelona is located in an area of low seismic hazard, but tables show how the expected damage for a relatively small
the analyses clearly point out the very high vulnerability of earthquake can be relatively high. In the deterministic case
the buildings and, consequently, a significant probability (Table 8), for example, there is a probability of 0.281+
of damage even in the case of a not too severe earthquake. 0.178 ¼ 0.459 that high-rise unreinforced masonry build-
It is somewhat surprising that the obtained results show ings located in zone II have a damage state between
high expected seismic damage for relatively low spectral moderate and severe, due to the high vulnerability of this
displacements. They indicate that the reinforced concrete type of buildings.
 ARTICLE IN PRESS
A.H. Barbat et al. / Soil Dynamics and Earthquake Engineering 28 (2008) 851–865 859

1.00 1.00
Probabylity of exceedance

Probabylity of exceedance
0.75 Low-rise (M) 0.75 Low-rise (RC)

0.50 0.50
Sligth
Moderate
Sligth
Severe
0.25 Moderate 0.25 Complete
Severe
Complete

0.00 0.00
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.0 0.5 1.0 1.5 2.0 2.5 3.0
Sd (cm) Sd (cm)

1.00 1.00
Probabylity of exceedance

Probabylity of exceedance
0.75 Mid-rise (M) 0.75 Mid-rise (RC)

0.50 0.50
Sligth
Moderate Sligth
Severe Moderate
0.25 Complete 0.25 Severe
Complete

0.00 0.00
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.0 0.5 1.0 1.5 2.0 2.5 3.0
Sd (cm) Sd (cm)

1.00 1.00
Probabylity of exceedance

Probabylity of exceedance

0.75 High-rise (M) 0.75 High-rise (RC)

Sligth
Moderate
Severe
0.50 Sligth 0.50 Complete
Moderate
Severe
Complete
0.25 0.25

0.00 0.00
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.0 0.5 1.0 1.5 2.0 2.5 3.0
Sd (cm) Sd (cm)

Fig. 8. Fragility curves for masonry (M) and reinforced concrete (RC) buildings.

Table 7 Barcelona may be included in the vulnerability classes B

Parameters characterizing the fragility curves, for reinforced concrete and C in a scale ranging from A (very high) to F (very low).
buildings (RC) and unreinforced masonry buildings (M)
EMS-98 predicts significant damage for these buildings
Building class Damage states thresholds even for low macroseismic intensities like VI and V. Note
that macroseismic scales are based on real damage data
Sd1 (cm) b1 Sd2 (cm) b2 Sd3 (cm) b3 Sd4 (cm) b4 observed during past earthquakes, covering a wide period
Low-rise, RC 0.49 0.28 0.70 0.37 1.84 0.82 5.24 0.83
of time and a wide range of earthquake sizes and building
Mid-rise, RC 0.99 0.28 1.42 0.36 2.34 0.50 5.11 0.61 classes. Mid-rise and high-rise RC buildings show a better
High-rise, RC 1.33 0.28 1.89 0.29 2.59 0.34 4.68 0.45 seismic performance for the selected ground motion
Low-rise, M 0.19 0.28 0.27 0.37 0.54 0.54 1.36 0.72
scenarios but the expected damage is greater for low-rise
Mid-rise, M 0.44 0.40 0.63 0.50 1.20 0.75 2.91 0.70 RC buildings.
High-rise, M 0.46 0.30 0.68 0.65 1.68 0.65 2.61 0.65
6. Risk scenarios for Barcelona

According to EMS-98 [25] and to the calculated damage Results for seismic risk scenarios simulated for Barcelo-
probability matrices for masonry buildings of Table 8, na are shown in this section. Spatial convolution between
the vulnerability of most of the residential buildings of seismic hazard and fragility, through damage probability
 ARTICLE IN PRESS
860 A.H. Barbat et al. / Soil Dynamics and Earthquake Engineering 28 (2008) 851–865

Table 8
Damage probability matrices for masonry buildings for deterministic and probabilistic hazard scenarios

Zone Deterministic hazard scenario Probabilistic hazard scenario

Damage state probabilities DSm Damage state probabilities DSm

0 1 2 3 4 0 1 2 3 4

Low-rise
I 0.950 0.037 0.011 0.002 0.000 0.066 0.632 0.257 0.095 0.014 0.002 0.498
II 0.737 0.189 0.063 0.009 0.001 0.349 0.287 0.416 0.249 0.042 0.006 1.065
III 0.917 0.061 0.018 0.003 0.001 0.109 0.431 0.365 0.173 0.027 0.004 0.807
R 1.000 0.000 0.000 0.000 0.000 0.001 0.981 0.013 0.004 0.002 0.000 0.025
Mid-rise
I 0.003 0.166 0.399 0.353 0.079 2.339 0.000 0.032 0.322 0.453 0.193 2.807
II 0.121 0.384 0.289 0.189 0.017 1.598 0.007 0.160 0.419 0.347 0.067 2.306
III 0.273 0.364 0.215 0.139 0.009 1.247 0.029 0.270 0.395 0.271 0.035 2.012
R 0.623 0.193 0.105 0.076 0.003 0.642 0.109 0.391 0.305 0.181 0.014 1.600
High-rise
I 0.003 0.145 0.389 0.371 0.092 2.404 0.000 0.019 0.248 0.464 0.269 2.983
II 0.135 0.388 0.281 0.178 0.018 1.556 0.002 0.133 0.385 0.381 0.099 2.441
III 0.307 0.369 0.195 0.120 0.009 1.155 0.014 0.23 0.386 0.307 0.056 2.154
R 0.647 0.205 0.086 0.059 0.003 0.566 0.632 0.257 0.095 0.014 0.002 0.498

Zones corresponding to the seismic microzonation of the city: R, rocky outcrops and I, soft soils. Damage states: 0, none; 1, slight; 2, moderate; 3, severe;
and 4, complete. DSm: weighted mean damage state.

Table 9
Damage probability matrices for reinforced concrete buildings for deterministic and probabilistic hazard scenarios

Zone Deterministic hazard scenario Probabilistic hazard scenario

Damage state probabilities DSm Damage state probabilities DSm

0 1 2 3 4 0 1 2 3 4

Low-rise
I 0.020 0.280 0.450 0.220 0.030 1.960 0.020 0.240 0.450 0.250 0.040 2.050
II 0.180 0.420 0.300 0.090 0.010 1.330 0.080 0.380 0.400 0.130 0.010 1.610
III 0.420 0.340 0.170 0.070 0.000 0.890 0.200 0.400 0.310 0.080 0.010 1.300
R 0.760 0.180 0.020 0.040 0.000 0.340 0.470 0.340 0.130 0.060 0.000 0.780

Mid-rise
I 0.220 0.440 0.240 0.090 0.010 1.230 0.120 0.410 0.330 0.120 0.020 1.510
II 0.680 0.230 0.060 0.030 0.000 0.440 0.430 0.370 0.150 0.040 0.010 0.830
III 0.840 0.120 0.030 0.010 0.000 0.210 0.610 0.280 0.080 0.030 0.000 0.530
R 0.920 0.060 0.010 0.010 0.000 0.110 0.670 0.240 0.070 0.020 0.000 0.440

High-rise
I 0.480 0.340 0.080 0.080 0.020 0.820 0.290 0.420 0.140 0.110 0.040 1.190
II 0.800 0.150 0.010 0.030 0.010 0.300 0.560 0.300 0.060 0.060 0.020 0.680
III 0.920 0.060 0.020 0.000 0.000 0.100 0.700 0.220 0.030 0.040 0.010 0.440
R 0.930 0.050 0.000 0.020 0.000 0.110 0.660 0.240 0.040 0.040 0.020 0.520

Zones corresponding to the seismic microzonation of the city: R, rocky outcrops and I, soft soils. Damage states: 0, none; 1, slight; 2, moderate; 3, severe;
4, complete. DSm: weighted mean damage state.

matrices, allows estimating the spatial distribution of entire city is 0.86 for the deterministic scenario, with a
expected damage. Although damage is assessed building standard deviation of 0.71, while for the probabilistic case
by building, we show damage results for the main we obtained a mean damage grade of 1.5170.87 which
administrative divisions of the city: districts, neighbor- respectively correspond to slight and moderate damage
hoods and census zones. The mean damage grade for the states according to Table 1.
 ARTICLE IN PRESS
A.H. Barbat et al. / Soil Dynamics and Earthquake Engineering 28 (2008) 851–865 861

Fig. 9 depicts the physical seismic risk scenario for the earthquake. Eixample means widening and this district
deterministic hazard case. Fig. 10 shows the scenarios for includes the urban area designed and projected in the
the probabilistic case. These scenarios are mapped for second half of the XIX century in order to plan and
different administrative units of the city, namely: (a) organize the growth of the city between Ciutat Vella and
districts, (b) neighborhoods and (c) census zones, thus the near small towns or villages which give the names to
allowing different resolution maps. Maps at district and another eight districts of the city (e.g. Gracia and Sarriá-
neighborhood scales smooth the maps of census zones and Sant Gervasi). The Eixample district was built between the
they provide average information that is easier to analyze end of the XIXth century and the beginning of the XXth
and interpret. Maps at census zones level are more detailed. century. In the deterministic case, the expected mean
They show greater resolution and are more useful, for damage state for Ciutat Vella and Eixample districts are
instance, for preparedness, risk management and emer- 1.77 and 1.16, respectively. In the probabilistic case, these
gency planning. values are 2.49 and 1.9, respectively. Fig. 11 shows detailed,
The districts of Ciutat Vella (01) and Eixample (02) are building-by-building, scenarios. In this figure, special and
the oldest in the city and they show the greatest expected other types of buildings have been excluded from the
damage. Ciutat Vella means Old City and it is the damage analysis due to their negligible effect on the final
downtown of Barcelona, with the oldest buildings of the damage scenarios.
city, mainly masonry buildings. It is expected that Ciutat Figs. 12 and 13 are supplementary examples of risk
Vella would be the most damaged district in case of scenarios showing the probability of a given damage state;

Fig. 9. Damage scenario corresponding to the deterministic hazard scenario: (a) districts, (b) neighborhoods and (c) census zones.

Fig. 10. Damage scenario corresponding to the probabilistic hazard scenario: (a) districts, (b) neighborhoods and (c) census zones.
 ARTICLE IN PRESS
862 A.H. Barbat et al. / Soil Dynamics and Earthquake Engineering 28 (2008) 851–865

these figures display the probability distribution in the them are located in the soil Zone II. Thus, according to the
Ciutat Vella and the Eixample districts, respectively, for values in Table 8, for the probabilistic case, the occurrence
the moderate and severe damage states. Fig. 12 shows how probability of the severe damage state is 0.381, which is
the probabilistic scenario would produce a significant consistent with the probability distribution mapped in Fig. 13.
number of severe damaged buildings in the Ciutat Vella From Tables 8 and 9 and Figs. 9–13, it is followed that
district, since about a half of its buildings show severe the probabilistic hazard scenario is more damaging than
damage state probabilities in the range between 40% and the deterministic one. This fact is reasonable according to
50%. Fig. 13 shows how for the probabilistic earthquake in the differences between the probabilistic and deterministic
the Eixample district, most of the buildings attain response spectra explained in Section 3.2. Sets of maps, like
significant probability for the severe damage state, namely those in Figs. 9–13, have been developed for the entire city,
in the probability range 30–40%. It has to be noticed that districts, neighborhoods and census zones. In addition, it is
more than 70% of the buildings in this district are old, possible to develop any type of detailed seismic risk
unreinforced, high-rise masonry buildings and that all of scenario for any seismic hazard case.

Fig. 11. Detailed damage scenarios, building by building, for Ciutat Vella (above) and Eixample (below) districts. Both deterministic (left) and
probabilistic (right) seismic scenarios are shown. These two districts are where the expected damage is more important (see also Figs. 9 and 10).
 ARTICLE IN PRESS
A.H. Barbat et al. / Soil Dynamics and Earthquake Engineering 28 (2008) 851–865 863

Fig. 12. Probability maps for the damage states moderate (above) and severe (below) for the Ciutat Vella (district 01). Both deterministic (left) and
probabilistic (right) seismic scenarios are shown.

7. Conclusions capacity spectra, fragility curves are also estimated in

a simplified way for each considered building type.
The seismic risk evaluation method used in this paper Fragility curves are used to characterize the expected
incorporates last generation methodologies for hazard, structural damage in a probabilistic way. Together with the
damage and risk estimation. It is efficient in coping with performance of the building when submitted to a specific
the enormous number of buildings existing in modern seismic action, they lead to damage probability matrices
urban areas. We solved this problem by classifying the for each seismic zone which are the key result for
buildings in typological groups. The vulnerability of the calculating seismic risk scenarios.
different building classes is characterized by bilinear The adopted method has been applied to Barcelona,
capacity spectra obtained by using CMS methods. The which is a typical Mediterranean city, located in a low-
basic seismic hazard in the studied area is defined by 5% to-moderate seismic hazard area. One of the most
elastic response spectra starting from which demand important results, which have been obtained, is the seismic
spectra are obtained. The seismic microzonation allows vulnerability characterizations of the buildings. Reliable
obtaining specific elastic response and demand spectra for capacity curves have been thus obtained, which show a
the different soil types of the urban area. Starting from wide vulnerability range for both the masonry and the
 ARTICLE IN PRESS
864 A.H. Barbat et al. / Soil Dynamics and Earthquake Engineering 28 (2008) 851–865

Fig. 13. Probability maps for the damage states moderate (above) and severe (below) for the Eixample district. Both deterministic (left) and probabilistic
(right) seismic scenarios are shown.

reinforced concrete buildings. Capacity and fragility curves for the buildings of the city. Seismic risk scenarios have
have been developed for about 97% of the residential been developed based on a building-by-building analysis.
building stock of the city, which is well represented by six These physical damage scenarios have been mapped
building classes. Credible hazard scenarios in ADRS according to different territorial or political areas of the
format have been used for the studied urban area. city like districts, neighborhoods and census zones. They
Significant damage is obtained for mid-rise and high-rise constitute excellent information sources and tools for risk
masonry buildings, due to the slenderness and low strength management, emergency planning and also useful for civil
of these buildings. Reinforced concrete buildings with protection, prevention and preparedness.
waffle slabs also show low seismic capacity leading to
significant expected damage. Damage probability matrices Acknowledgments
have been obtained for the four seismic areas of the city,
allowing development of representative risk scenarios, This work has been partially sponsored by the Spanish
which are based on a complete and highly reliable database Ministry of Education and Science and with FEDER funds
 ARTICLE IN PRESS
A.H. Barbat et al. / Soil Dynamics and Earthquake Engineering 28 (2008) 851–865 865

(Projects: REN2002-03365/RIES, REN2003-07170/RIES [21] Faccioli E. Seismic hazard assessment for derivation of earthquake
and CGL2004-22325-E, CGL-2005-04541-C03-02/BTE) scenarios in Risk-UE. Bull Earthquake Eng 2006;4(4):341–64.
[22] Gambarotta L, Lagomarsino S. A microcrack damage model for
and by the European Commission (RISK-UE Project,
brittle materials. Int J Solids Struct 1993;30:177–98.
contract EVK4-CT-2000-00014). [23] Carr A. Inelastic dynamic analysis program: RUAUMOKO and
post-processor for RUAUMOKO. New Zealand: Department of
Civil Engineering, University of Canterbury; 2000.
References [24] ATC-40, Seismic evaluation and retrofit of concrete buildings.
Applied Technology Council, Report: SSC 96-01, vol. 1. Redwood
[1] Benedetti D, Petrini V. Sulla vulnerabilit. L’industria delle Con- City, CA: Seismic Safety Commission, 1996.
struzioni 1984;149:66–74. [25] Grünthal G, editor. European Macroseismic Scale 1998, vol. 15.
[2] ATC-13. Earthquake damage evaluation data for California. Red- Luxemburg: Centre Européen de Géodynamique et Séismologie,
wood City, CA: ATC Applied Technology Council; 1985. Cahiers du Centre Européen de Géodynamique et de Séismologie;
[3] HAZUS 99-SR2. HAZUS technical manual, vols. 1–3. Washington, 1998.
DC: Federal Emergency Management Agency, FEMA & National [26] Infocca. Funcions d’informació de Barcelona i Cartografia. IMI,
Institute of Building Sciences, NIBS; 2002. Ajuntament de Barcelona, 1999.
[4] Kappos A, Pitilakis K, Stylianidis K, Morfidis K. Cost-benefit [27] Departament d’Estadı́stica. Anuari estadı́stic de la ciutat de
analysis for the seismic rehabilitation of buildings in Thessaloniki, Barcelona 2002. Departament d’Estadı́stica. Ajuntament de Barcelo-
based on a hybrid method of vulnerability assessment. In: Proceed- na. Barcelona, 2002.
ings of the 5th international conference on seismic zonation, Nice, [28] PGS-1. Norma Sismorresistente P.G.S.-1, Comisión Interministerial
1995, vol. 1. p. 406–13. de la Presidencia del Gobierno, Decreto 106/1969 de 16 de enero de
[5] Singhal A, Kiremidjian AS. Method for probabilistic evaluation 1969, 1968.
of seismic structural damage. J Struct Eng ASCE 1996;122(12): [29] Lungu D, Aldea A, Arion A, Vacareanu R, Petrescu F, Cornea T.
1459–67. European distinctive features, inventory database and typology,
[6] Barbat AH, Yépez Moya F, Canas JA. Damage scenarios simulation Work Package 1 of RISK_UE Project. European Commission,
for seismic risk assessment in urban zones. Earthquake Spectra EVK4-CT-2000-00014, 2001.
1996;12(3):371–94. [30] Irizarry J, Goula X, Susagna T. Analytical formulation for the elastic
[7] Anagnos T, Rojahn C, Kiremidjian A. NCEER-ATC joint study on acceleration–displacement response spectra adapted to Barcelona soil
fragility of buildings. Report NCEER-95-0003. Buffalo, New York: conditions. Technical Report. Barcelona: Institut Cartogràfic de
National Center for Earthquake Engineering Research, 1995. Catalunya, 2003.
[8] Milutinovic ZV, Trendafiloski GS. Vulnerability of current buildings. [31] Irizarry J. An advanced approach to seismic risk assessment.
Work Package 4 of RISK-UE Project. European Commission, Application to the cultural heritage and the urban system of
EVK4-CT-2000-00014, 2003. Barcelona. Ph.D. thesis. Barcelona: Universitat Politècnica de
[9] Barbat AH, Pujades LG, Lantada N. Performance of buildings under Catalunya, 2004.
earthquake in Barcelona, Spain. Comput-Aided Civil Infrastruct Eng [32] Goula X, Susagna T. Observation, characterization and prediction of
2006;21:573–93. strong ground motion. In: Oliveira CS, Roca A, Goula X, editors.
[10] Barbat AH, Lagomarsino S, Pujades LG. Vulnerability assessment of Assessing a managing earthquake risk. Dordrecht, The Netherlands:
dwelling buildings. In: Oliveira CS, Roca A, Goula X, editors. Springer; 2006. p. 47–65 [Chapter 3].
Assessing a managing earthquake risk. Dordrecht, The Netherlands: [33] Roca A, Oliveira CS, Ansal A, Figueras S. Local site effects and
Springer; 2006. p. 115–34 [Chapter 6]. microzonation. In: Oliveira CS, Roca A, Goula X, editors. Assessing
[11] Moreno R, Bairán JM, Pujades LG, Aparicio AC, Barbat AH. a managing earthquake risk. Dordrecht, The Netherlands: Springer;
Evaluación probabilista del comportamiento de edificios porticados 2006. p. 67–89 [Chapter 4].
de hormigón armado. Hormigon y acero 2004;232:125–36. [34] Faccioli E, Pessina V. Use of engineering seismology tools in ground
[12] Barbat AH, Mena U, Yépez F. Evaluación probabilista del riesgo shacking scenarios. In: Lee WHK, Kanamori H, Jennings P,
sı́smico en zonas urbanas. Revista internacional de métodos Kisslinger C, editors. International handbook of earthquake and
numéricos para cálculo y diseño en ingenierı́a 1998;14:247–68. engineering seismology, Part B (International Geophysics). Academic
[13] Carreño ML, Cardona OD, Barbat AH. Urban seismic risk Press; 2002. p. 1031–49.
evaluation: a holistic approach. Nat Hazards 2007;40:137–72. [35] Rey J, Faccioli E, Bommer J. Derivation of design soil coefficient (S)
[14] Carreño ML, Cardona OD, Barbat AH. Disaster risk management and response spectral shapes for Eurocode 8 using the European
performance index. Nat Hazards 2007;41:1–20. Strong-Motion Database. J Seismol 2002;6(4):547–55.
[15] Freeman SA. Prediction of response of concrete buildings to severe [36] Olivera C, Redondo E, Lambert J, Riera Melis A, Roca A. Els
earthquake motion. In: Proceedings of Douglas McHenry interna- terratrèmols del segles XIV i XV a Catalunya. Barcelona: Institut
tional symposium on concrete and concrete structures, Publication Cartogràfic de Catalunya, Generalitat de Catalunya; 2006.
SP-55. Detroit, MI, USA: American Concrete Institute, 1978. [37] Cid J. Zonificación sı́smica de la ciudad de Barcelona basada en
[16] Freeman SA. The capacity spectrum method. In: Proceedings of the métodos de simulación numérica de efectos locales. Ph.D. thesis.
11th European conference on earthquake engineering, Paris, 1998. Barcelona, Spain: Technical University of Catalonia, 1998.
[17] Fajfar P. A non linear analysis method for performance-based seismic [38] Cid J, Susagna T, Goula X, Chavarria L, Figueras S, Fleta J, Casas
design. Earthquake Spectra 2000;16(3):573–5924. A, Roca A. Seismic zonation of Barcelona based on numerical
[18] Fajfar P. Structural analysis in earthquake engineering—a break- simulation of site effects. Pure Appl Geophys 2001;158:2559–77.
through of simplified non-linear methods. In: Proceedings of the 12th [39] Bommer JJ, Abrahamson NA. Why do modern probabilistic seismic
European conference on earthquake engineering, London, 2002. hazard analyses often lead to increased hazard estimates? Bull
[19] Egozcue JJ, Barbat A, Canas JA, Miquel J, Banda E. A method to Seismol Soc Am 2006;96(6):1967–77.
estimate intensity occurrence probabilities in low seismic activity [40] Oliveira CS, Campos-Costa A. Overview on earthquake hazard
regions. Earthquake Eng Struct Dyn 1991;20:43–60. assessment—methods and new trends. In: Oliveira CS, Roca A,
[20] Calvi GM, Magenes G, Bommer JJ, Pinho R, Crowley H, Restrepo- Goula X, editors. Assessing a managing earthquake risk. Dordrecht,
Vélez LF. Displacement-based methods for seismic vulnerability The Netherlands: Springer; 2006. p. 15–46 [Chapter 2].
assessment at variable geographical scales. ISET J Earthquake [41] Galaso A, Lagomarsino S, Penna A. TREMURI Program: seismic
Technol 2006;43(3):75–104. analysis of 3D masonry buildings. Italy: University of Genoa; 2002.