١

SELENA – An open-source tool for seismic risk and loss ٢

assessment using a logic tree computation procedure ٣

٤

S. Molina 1,3, D.H. Lang 2,3, and C.D. Lindholm 2,3 ٥

٦

(1) Dpto. de Ciencias del Mar y Biología Aplicada (Área de Física de la Tierra), Facultad de Ciencias, Fase II, ٧

Universidad de Alicante, Campus de S. Vicente del Raspeig s/n, 03690 Alicante, Spain ٨

(2) NORSAR, P.O. Box 53, 2027 Kjeller, Norway ٩

(3) International Centre for Geohazards (ICG), 0806 Oslo, Norway ١٠

١١

١٢

Abstract ١٣

١٤

The era of earthquake risk and loss estimation basically began with the seminal paper on ١٥

hazard by Allin Cornell in the year 1968 (Cornell, 1968). Following the 1971 San Fernando ١٦

earthquake, the first studies placed strong emphasis on the prediction of human losses ١٧

(number of casualties and injured) and to foresee the needs in terms of health care and ١٨

shelters in the immediate aftermath of a strong event. In contrast to the early risk modeling ١٩

efforts, the later studies have focused on the disruption of the serviceability of roads, ٢٠

telecommunications and other important lifeline systems. In the 1990’s the National ٢١

Institute of Building Sciences (NIBS) developed a tool (HAZUS®99) for the Federal ٢٢

Emergency Management Agency (FEMA), where the goal was to incorporate the best ٢٣

quantitative methodology in the earthquake loss estimates. ٢٤

Herein, the current version of the open-source risk and loss estimation software SELENA ٢٥

v4.1 is presented. While using the spectral displacement-based approach (capacity spectrum ٢٦

method), this fully self-contained tool analytically computes the degree of damage on specific ٢٧

building typologies as well as the connected economic losses and number of casualties. The ٢٨

earthquake ground shaking estimates for SELENA v4.1 can be calculated or provided on ٢٩

three different ways: deterministic, probabilistic or based on near-real-time data. The main ١

distinguishing feature of SELENA compared to other risk estimation software tools is that it ٢

is implemented in a ‘Logic Tree’ computation scheme which accounts for uncertainties of ٣

any input (e.g., scenario earthquake parameters, ground-motion prediction equations, soil ٤

models) or inventory data (e.g., building typology, capacity curves and fragility functions). ٥

The data used in the analysis is assigned with a decimal weighting factor defining the weight ٦

of the respective branch of the Logic Tree. The weighting of the input parameters accounts ٧

for the epistemic and aleatoric uncertainties that will always follow the necessary ٨

parameterization of the different types of input data. ٩

Like previous SELENA versions, SELENA v4.1 is coded in MATLAB which allows an easy ١٠

dissemination among the scientific-technical community. Furthermore, any user has access ١١

to the source code in order to adapt, improve or refine the tool according to his particular ١٢

needs. The handling of SELENA’s current version and the provision of input data is ١٣

customized to an academic environment which might support decision-makers of local, state ١٤

and regional governmental agencies in estimating possible losses from future earthquakes. ١٥

١٦

1. Introduction ١٧

١٨

Seismic risk estimation is based on the need to quantify the expectations of ground shaking ١٩

and the corresponding performance of structures. Based on such investigations and efforts ٢٠

construction techniques have improved over time, and appropriate countermeasures can be ٢١

taken. The scientific field of seismic risk and loss assessment is a growing research area ٢٢

which traditionally has been either based on macroseismic intensity or peak ground ٢٣

acceleration (PGA). In recent years different risk assessment methodologies have been ٢٤

developed which are incorporated in a considerable number of different software (Crowley ٢٥

et al., 2004; McGuire, 2004; Oliveira et al., 2006). ٢٦

Unfortunately, it is a large damaging earthquake of all which is able to verify or refute the ٢٧

estimated seismic scenario, the chosen methodology and the defined assumptions. But still ٢٨

this creates a fruitful situation when we are able to calibrate our models and input ٢٩

parameters on this experience and results of different risk estimation methods can be ١

compared with each other. In general nearly all available risk and loss assessment studies ٢

defined individual scenario earthquakes as a main basis for planning (ATC, 1996a; ٣

Algermissen et al., 1988; CDMG, 1982a, 1982b, 1987, 1988, 1990, 1995; CUSEPP, 1985; ٤

Dames and Moore, 1996; EQE, 1997; Harlan and Lindbergh, 1988; NOAA, 1973). All ٥

these studies use already existing knowledge of regional geology and seismic activity to ٦

generate maps with estimated intensities I or ground motion accelerations a. This, in ٧

combination with other types of input data (e.g., building stock, population density), is used ٨

to calculate the extent of damages to structures and life-lines as well as the impacts on ٩

population. Some of these studies additionally address potential secondary hazards such as ١٠

fire, flood, and hazardous materials release. Earthquake scenarios of this type have been ١١

employed by governmental institutions and public utilities to prepare for and to mitigate ١٢

the degree of damage from future events. Thus it appears that the typical loss study has ١٣

been focused on a single event, applied in the long-term pre-event period, and utilized ١٤

primarily by those concerned with seismic safety planning and disaster management. ١٥

In this respect, the Federal Emergency Management Agency (FEMA 366, 2001, 2008) ١٦

initiated a study on seismic risk estimation for all regions of the United States using the ١٧

national loss estimation tool HAZUS®99 and HAZUS®MH, respectively. The study’s main ١٨

task was to analyze and compare the seismic risk across regions in the U.S. which have ١٩

different hazard levels, characterized by different population density or physical building ٢٠

vulnerability. ٢١

The advent of high-speed computing, satellite telemetry and Geographic Information ٢٢

Systems (GIS) made it possible to electronically generate loss estimates for multiple ٢٣

earthquake scenarios, to provide a nearly unlimited mapping capability, and perhaps most ٢٤

important, to develop estimates for a current earthquake event in near real-time given its ٢٥

source parameters, i.e. magnitude and location, are provided. ٢٦

Currently, a number of different computer tools able to estimate the seismic risk using ١

different methodologies are available. Table 1 lists some of them and briefly describes their ٢

main principles and outputs. ٣

A very powerful approach being attractive from a scientific-technical perspective is the ٤

HAZUS software (HAZUS®97, HAZUS®99, HAZUS®99-SR1, HAZUS®99-SR2, ٥

HAZUS®MH, HAZUS®MH MR1, HAZUS®MH MR2, HAZUS®MH MR3) which was ٦

developed by the National Institute of Building Sciences (NIBS) for the Federal ٧

Emergency Management Agency (FEMA, 1997, 1999, 2001, 2002, 2004, 2005a, 2006, ٨

2007). The HAZUS tool is built upon the integrated geographic information system ٩

platform ArcGIS (ESRI, 2004) and can be considered as a software extension to ArcGIS. ١٠

HAZUS is directly integrated with the national and regional databases on building stock ١١

and demography data of the United States (FEMA 366, 2008). This enables any larger ١٢

community in the United States to simulate earthquake risk scenarios with a minimum ١٣

effort since most of the necessary data is already prepared. The basic methodology behind ١٤

HAZUS represented the starting point for the development of alternative tools (see Table ١٥

1) in order to compute seismic risk and loss estimates as well as initiated numerous ١٦

application studies (i.e. Kircher, 2003; Wang et al., 2005; Kircher et al., 2006; Nielson and ١٧

DesRoches, 2007; Rojas et al., 2007). The fact that HAZUS is tailored so intimately to U.S. ١٨

situations which makes it very difficult to be applied to other environments or geographical ١٩

regions, has recently activated the development of GIS-based methodologies which ٢٠

facilitate the application of HAZUS to other parts of the world (Hansen and Bausch, ٢١

2006). ٢٢

Aware of the importance of a proper seismic risk estimation, the International Centre for ٢٣

Geohazards ICG, through NORSAR (Norway) and the University of Alicante (Spain), has ٢٤

developed a software tool running under MATLAB (The MathWorks, Inc.) in order to ٢٥

compute seismic risk of urban areas using the capacity-spectrum method. The tool named ٢٦

SELENA (SEimic Loss EstimatioN using a logic tree Approach) is open for any user-١

defined data and thus can be applied to any part of the world. The user has to supply a ٢

number of input files which contain the necessary input data (e.g., building inventory data, ٣

demographical data, definition of seismic scenario etc.) in a simple ASCII format. A more ٤

detailed description of the type of input data will be given below. ٥

It should be explicitly stated that the core of the HAZUS methodology (FEMA, 1999, ٦

2003) was adopted for SELENA. However, one of the main differences between both ٧

tools is that SELENA works independent of any Geographic Information System, while ٨

HAZUS is connected to the ArcGIS software (ESRI, Inc.). In addition, a Logic Tree-٩

computation scheme has been implemented in SELENA which allows the user to define ١٠

weighted input parameters and thus being able to properly account for epistemic ١١

uncertainties. Consequently, any type of output is provided with corresponding confidence ١٢

levels. ١٣

١٤

2. Technical components of the SELENA-tool ١٥

١٦

2.1 Basic procedure ١٧

The current version of SELENA (Molina et al., 2009) allows for three analysis types which ١٨

are differing in the way the seismic impact is described. In general, spectral ordinates of ١٩

seismic ground motion at different reference periods have to be provided for each ٢٠

geographical unit (i.e. census tract), in order to allow the construction of a design spectra ٢١

following a selectable seismic code provision, i.e. spectral ordinates (PGA, Sa) at reference ٢٢

periods T = 0.01, 0.3 and 1.0 s for IBC-2006 (International Code Council, 2006), and T = ٢٣

0.01 s (PGA) for Eurocode 8 (CEN, 2002) as well as for Indian seismic building code IS ٢٤

1893 (Part 1) : 2002 (BIS, 2002). ٢٥

Once the seismic ground motion in each geographical unit is defined, the computation of ١

physical damage to the building stock, the total economic loss related to these damages, ٢

and the number of casualties, i.e. the number of injured people and fatalities, is conducted. ٣

Damage results are given in terms of cumulative probabilities P of being in, or exceeding ٤

one particular damage state ds following the classification scheme given by HAZUS ٥

(FEMA, 1999) into none, slight, moderate, extensive and complete damage. Absolute numbers of ٦

damaged buildings or the damaged building (floor) area in [m2] are calculated by combining ٧

damage probabilities for each building type with the inventory data of the building stock. ٨

Figure 1 exemplarily illustrates the sequence of SELENA for a deterministic analysis and ٩

shows which input layers are required for the different program outputs. ١٠

١١

Since the described methodology of risk assessment is based on statistical data, the level of ١٢

resolution is of utmost importance. Since a resolution of the damage outputs on the level ١٣

of individual buildings would require huge computation efforts, SELENA as most other ١٤

risk estimation software tools considers the geographical unit (GEOUNIT) as the smallest ١٥

area unit. In practice, this unit is related to building blocks or city districts. The decision on ١٦

the extent of each geographical unit has to be made considering different aspects such as ١٧

having equal soil conditions, constant surface topography or a homogeneous level of ١٨

building quality within the demarcated area. ١٩

٢٠

٢١

2.2 Specification of the seismic input (demand) ٢٢

٢٣

A key point in any seismic risk assessment is the provision of seismic ground motion. In ٢٤

SELENA v4.1 this can be done by: ٢٥

- the provision of spectral ordinates (e.g., PGA, Sa at 0.3 and 1.0 seconds) taken out from ١

probabilistic shaking maps) and assigned to the geographical units (probabilistic analysis), ٢

- the definition of deterministic scenario earthquakes (e.g. historical or user-defined ٣

events) and adequate/suitable ground-motion prediction equations in order to compute ٤

the spectral ordinates in each geographical unit (deterministic analysis), ٥

- provision of spectral amplitudes of recorded ground motion at the locations of seismic ٦

(strong-motion) stations (analysis with near-real-time data). ٧

The first results of each analysis type consist in the provision of seismic ground motion ٨

amplitudes at the center of each geographical unit. Thereby it has to be considered that ٩

these amplitudes can either represent the motion on rock conditions (deterministic and ١٠

probabilistic analysis) or already include amplification effects of local (near-surface) subsoil ١١

conditions (analysis with near-real-time data). Based on these acceleration values SELENA ١٢

generates an elastic response spectrum (damping factor = 5 %) following a selectable ١٣

seismic code provision. Currently, the provisions of the U.S. seismic code IBC-2006 ١٤

(International Code Council, 2006), Eurocode 8 – Type 1 and Type 2 (CEN, 2002) and ١٥

Indian seismic building code IS 1893 (Part 1) : 2002 (BIS, 2002) with their soil ١٦

amplification factors are incorporated. Figure 2 schematically depicts the shape of a design ١٧

response spectra in the T-Sa-domain. ١٨

١٩

2.3 Structural performance under seismic action ٢٠

٢١

The core methodology of SELENA was adopted from the HAZUS-MH software whose ٢٢

basic approach in order to identify the level of structural damage under a given seismic ٢٣

impact is the capacity-spectrum method (FEMA, 1996). During the last years, several ٢٤

modifications and extensions of this powerful and efficient technique have been developed. ٢٥

Two of these procedures are incorporated in SELENA v4.1: ٢٦

- Procedure 1: The conventional Capacity Spectrum Method as proposed in ATC-40 (Applied ١

Technology Council ATC, 1996b) ٢

- Procedure 2: The Modified Capacity Spectrum Method (MADRS) as proposed in FEMA 440 ٣

(FEMA, 2005b) ٤

٥

To apply any of the available capacity spectrum methods, both seismic demand and the ٦

capacity curve have to be transformed into the spectral acceleration-spectral displacement ٧

(Sa-Sd) domain (Figure 3). Thereby, seismic demand is represented by the elastic response ٨

spectrum while the capacity curve reflects the building’s lateral displacement as a function ٩

of a horizontal force V applied to the structure. Beside a number of factors, building ١٠

capacity curves mainly depend on the building type (working materials and construction), ١١

number of stories (height), and also from its region reflecting local building regulations as ١٢

well as local construction practice and quality. ١٣

The main task of the capacity-spectrum method is to find that point on the capacity curve ١٤

consistent with the seismic demand being reduced for nonlinear effects. Since each point ١٥

on the capacity curve represents a certain state of structural damage and thus reflects an ١٦

increase in structural damping as the damages accumulate, the performance point can only ١٧

be found iteratively. As Figure 3 illustrates, the performance point finally is characterized ١٨

by a spectral acceleration Sa and spectral displacement Sd (and establishing the basis for ١٩

assigning discrete damage probabilities P.) ٢٠

Once the performance point and its corresponding spectral displacement Sd are found, ٢١

structural vulnerability (fragility) functions for each damage state ds are required to assign ٢٢

damage probabilities P (Figure 4). These represent cumulative probabilities of a certain ٢٣

building type of being in or exceeding one of the different damage states ds dependent on ٢٤

spectral displacement Sd . Figure 4 shows a set of fragility functions for a random model ٢٥

building type as described by HAZUS (FEMA, 1999). The differences between the ٢٦

intersection points of two neighboring fragility functions for a given spectral displacement ١

are discrete damage probabilities (Figure 5) which establish the basis for the calculation of ٢

absolute damage values. Generally, fragility functions are provided in dependence on ٣

building typology and level of seismic code design reflecting the general quality of ٤

construction practice. The fragility functions given by HAZUS (FEMA, 1999) are ٥

described by the following form: ٦

dsd

d

dsd S

SSdsP

,

ln1

(1) ٧

in which: ٨

٩

dsdS , - median value of spectral displacement at which the building reaches the ١٠

threshold of damage state ds, ١١

ds - standard deviation of the natural logarithm of spectral displacement for damage ١٢

state ds, ١٣

- standard normal cumulative distribution function. ١٤

١٥

2.4 Risk and loss assessment based on statistical inventory data ١٦

١٧

Seismic risk results are essentially represented by the total extent of physical damage an ١٨

earthquake scenario likely to occur may produce to the building stock. Within SELENA ١٩

v4.1 the extent of structural damage can either be quantified in the number of buildings or ٢٠

building floor area affected by a certain damage state ds. ٢١

Based on the damage results, both economic losses and number of casualties are calculated. ٢٢

The first requires a suitable economic model, which provides realistic costs for the repair or ٢٣

replacement of damaged or collapsed buildings and which then allows the appraisal of the ٢٤

total amount of loss in each geographical unit. ٢٥

The computation of economic losses caused by direct structural damage and which are ١

required for building repair and in case of complete damage or collapse for replacement is ٢

done in the following way adopting the methodology described by FEMA (2003): ٣

oct

i

mbt

j

ds

kkjikjji CPACILoss

1 1 1,,,, (2) ٤

in which: ٥

CI - regional cost multiplier accounting for the geographic cost variations ٦

between the different geographical units, ٧

Ai, j - built area of occupancy type (oct) i and model building type (mbt) j (in [m2]) ٨

Pj, k - damage probability for model building type j to experience structural ٩

damage of damage state ds k (slight, moderate, extensive or complete), ١٠

Ci, j, k - cost of repair or replacement for each occupancy type i and model building ١١

type j suffering structural damage state k in input currency per floor area, ١٢

e.g. [$/m2]. ١٣

١٤

To determine the estimated number of casualties (i.e. injured people and fatalities) which ١٥

are mainly caused by the total or partial collapse of buildings, reliable statistical data on the ١٦

study area’s demography is required. This does not only consist of population statistics, like ١٧

e.g. total number of inhabitants per district, but also in average numbers of people staying ١٨

in buildings of different type or occupancy and percentages of people staying outdoors or ١٩

indoors at different times of the day. ٢٠

SELENA’s current version 4.1 facilitates the computation of casualty numbers using two ٢١

different methodologies. During the preparation of input files, the user can explicitly ٢٢

choose whether the HAZUS approach (FEMA, 2003) or the basic approach following ٢٣

Coburn and Spence (2002) will be applied. ٢٤

The computation of the estimated number of casualties Ks, i of severity level i based on the ١

basic approach is based on the following equation: ٢

mbt

j

ds

kjkjkjis popPCRK

1 1

.*,,, (3) ٣

in which: ٤

i - injury severity level ranging from light injuries (i = 1), hospitalized injuries (i ٥

= 2), life threatening injuries (i = 3) to deaths (i = 4), see FEMA (2003), ٦

CRj, k - casualty rate of each severity level i for model building type j and damage ٧

state k, ٨

P*j, k - damage probability for model building type j to experience structural ٩

damage of damage state ds k (slight, moderate, extensive, complete or ١٠

complete with collapse), ١١

pop.j - average number of people occupying model building type j. ١٢

١٣

Irrespective of the methodology chosen, casualty numbers are computed for three different ١٤

day time scenarios (night time, day time, commuting time). Thus, extreme cases of ١٥

occupancy are covered which are strongly dependent on the time of the day. These ١٦

scenarios are expected to generate the highest casualty numbers for the population at home ١٧

(night time), the population at work or educational facilities (day time), and the population ١٨

during rush hour (commuting time). ١٩

٢٠

2.5 Accounting for uncertainties by a ‘Logic Tree’ approach ٢١

٢٢

The main difference between SELENA and other risk estimation software tools, as e.g. ٢٣

HAZUS (FEMA, 1999) lies in the fact that uncertainties of any input type can be ٢٤

accounted for by a ‘Logic Tree’ approach. The diversification of each input type thereby ٢٥

increases the number of branches of the logic tree and thus computation runs as well as ١

computation time. Such logic tree methodologies have been widely used in seismic hazard ٢

computation in order to capture epistemic uncertainties which are connected to knowledge ٣

deficits of the input data (Scherbaum et al., 2005; Bommer et al., 2005) and the logic tree ٤

implementation in SELENA is to the best of our knowledge the first realization of that ٥

kind in risk analysis. Based on the often uncritical use of logic tree procedures in ٦

earthquake hazard, Bommer et al. (2005) bring up some important mementos to the naive ٧

use of logic tree computations: The main question raised is if the user is sufficiently sure ٨

that the logic tree branches are mutually exclusive (no implicit overlap), and possibly most ٩

important, are the branches exhaustive so that all possibilities are covered. We do not ١٠

pretend that the SELENA logic tree scheme is foolproof, since with any such specialized ١١

software, the results always critically depend on the expertise of the operating scientist or ١٢

engineer. The logic tree of SELENA covers mainly inventory parameters (such as subsoil ١٣

classes, building vulnerabilities or socio-economic values within the same geographical unit) ١٤

that are regarded as uncertain model parameters, and the respective weighting factors can ١٥

be easily defined in accordance with the given situation. The current setup may be changed ١٦

in the future, depending on the improved understanding. ١٧

The logic tree approach leads to the computation of median values as well as 16%- ١٨

respectively 84%-fractile values for each output type. An exemplary illustration of the logic ١٩

tree computation scheme for the deterministic analysis is given in Figure 6. Uncertainties of ٢٠

the earthquake source, the applied ground-motion prediction relations (attenuation laws), ٢١

the subsoil model, the vulnerability of the building stock, and economic loss model are ٢٢

considered. ٢٣

For the first time, Molina and Lindholm (2005) applied this approach for Oslo (Norway) in ٢٤

order to compute estimated damages and losses. The epistemic uncertainties considered in ٢٥

this work concentrated on ground motion prediction equations and fragility information ٢٦

(capacity curves and vulnerability functions) assigned to the different model building types. ١

This because neither specific ground motion prediction equations nor fragility functions ٢

for Norway and Norwegian building typologies are yet developed or available. ٣

٤

3. Implementation of the risk estimation method in MATLAB ٥

٦

3.1 Why MATLAB? ٧

٨

Like previous versions, SELENA v4.1 is coded in MATLAB since it provides a suitable ٩

scientific and technical computing environment for the purpose and required computation ١٠

needs of SELENA. Further reasons for choosing the MATLAB platform are its ١١

dissemination and prevalence among the scientific community and its plain computing ١٢

language making it more attractive to a large number of potential users. Thereby it allows ١٣

the user to access SELENA’s source code and to adapt, improve or refine the tool to his ١٤

own needs. ١٥

١٦

3.2 Program structure ١٧

١٨

Following the previously described methodology to compute damages and losses, the ١٩

flowchart in Figure 7 demonstrates the main sequence of SELENA v4.1. ٢٠

The whole computation process of SELENA v4.1 is subdivided into 33 separate m-script ٢١

files (format *.m) which are consecutively accessed during the program sequence. Their ٢٢

respective functions and tasks are briefly described in Table 2. The main reason for ٢٣

subdividing the code into separate script files was to allow the user an easier inspection of ٢٤

the code and thus to implement own changes or amendments. ٢٥

٢٦

In general, the execution of SELENA requires less interactive action by the user. Most of ١

the required information has to be prepared in the forefront and provided in the form of ٢

input files (paragraph 3.3). After starting the program, the user has to make decisions ٣

interactively on: ٤

- the type of analysis (deterministic, probabilistic, or (near-)real-time data) ٥

- the source directory of the different input files. ٦

٧

An exemplary illustration of the prompting windows consecutively popping up after ٨

starting the program sequence in MATLAB is given in Figure 8. ٩

١٠

Besides a number of programming tools implemented in MATLAB, SELENA v4.1 reverts ١١

to the wavelet, mapping and statistics toolboxes which have to be installed by the user. It ١٢

will be a matter of future development to reprogram the SELENA code such that the ١٣

toolboxes will not be required. ١٤

١٥

3.3 Preparation of input files ١٦

١٧

Most of the input files required by SELENA have to be prepared in ASCII-format and ١٨

provided as plain text files (*.txt) in the same (source) directory as the different m-scripts. ١٩

A variety of input files containing the data of an exemplary case study are included in the ٢٠

SELENA-package which can be received through the NORSAR webpage free of charge ٢١

(www.norsar.no/seismology/selena.htm). It is supposed that any user simply modifies ٢٢

these files according to his own data. Figure 9 exemplarily illustrates the format of some ٢٣

input files which are required as txt-files. ٢٤

As it was already addressed, the main difference between the three analysis types consists in ٢٥

the way the seismic ground motion parameters are provided or calculated. Table 3 lists the ٢٦

different input files necessary for the provision of seismic ground motion parameters. The ١

remaining input files are common for all analysis types (Table 4). While most of them have ٢

to be provided as text files (*.txt), some need to be provided in MATLAB-format ٣

(*.mat). The latter especially applies for files containing fixed parameter values (variables), ٤

e.g. ubcampfact.mat, which will normally not be modified by the user, and files which ٥

include the spectral acceleration and displacement values of the single capacity curves to be ٦

provided by the user himself. ٧

٨

3.4 Computation duration ٩

١٠

As already stated, a major computation component of SELENA is working iteratively in ١١

order to determine the performance point of a building type under a given seismic demand ١٢

(capacity spectrum method). Consequently, this has serious effects on the computation ١٣

time especially when the amount of input data (diversification of building types, size of the ١٤

studied region i.e. number of geographical units) is increasing. Additionally, the number of ١٥

branches defined for the Logic Tree methodology should also kept in certain limits since ١٦

this also affects the computation time (Lang et al., 2008a). ١٧

In order to get an idea of SELENA’s computation duration and velocity, the risk and loss ١٨

assessment calculation for Oslo (Norway) took 190 minutes on a standard PC. The study ١٩

area covering the whole Oslo municipality with an area of 226 km2 and around 550 000 ٢٠

inhabitants was divided in 84 different geographical units. The building stock consisted of ٢١

nearly 71 500 single buildings which were classified into 15 different model building types ٢٢

and 23 different occupancy types. The Logic Tree comprised 108 branches formed by 3 ٢٣

different scenario earthquake sources, 2 regarded ground-motion prediction equations ٢٤

(attenuation relations), 3 different soil models, 3 different levels of building quality, and 2 ٢٥

different economic models. Consequently the computation of one branch lasted 102 ١

seconds (Molina and Lindholm, 2005). ٢

٣

3.5 Managing the results, program outputs ٤

٥

After running one of the three analysis types which are incorporated in SELENA v4.1, a ٦

considerable number of output files is generated and written in a sub-directory. How many ٧

of these output files are written mainly depends on the defined number of Logic Tree ٨

branches as well as the analysis type. An overview and brief description of the output files ٩

is given in Table 5. It should be noted that in some cases differences both in the name of ١٠

the output files and in the units of the results exist whether the user requests the output ١١

dependent on the number of damaged buildings or damaged area. ١٢

١٣

All output files are given in plain ASCII text format (.txt). The data contained in the ١٤

output files is in delimited format separated by tabulators, such that it can be imported into ١٥

any program or converted into any user-defined format. Since most of the results are ١٦

referring to the geographic location of the minimum geographical units, the main interest ١٧

of the user presumably lies in plotting the results with Geographical Information Systems, ١٨

as e.g., ArcView, ArcGIS, MapInfo. Figure 10 exemplary illustrates selected results taken ١٩

from a comprehensive risk assessment study for Oslo, Norway (Molina and Lindholm, ٢٠

2005). Predicted physical damages (in the five different damage states) for one particular ٢١

building type (URML – unreinforced masonry, low-rise) are plotted as column bars on the ٢٢

level of geographical units. The damage results are underlain with predicted seismic ٢٣

ground-motion values (PGA) considering soil amplification effects for the assumed ٢٤

deterministic earthquake scenario (M 6.0). The illustration of results is done with the ٢٥

Google Earth-based software tool RISe (Risk Illustrator for SELENA; Lang et al., 2008b; ١

Lang and Gutiérrez, 2009). ٢

٣

4. Conclusions and further development ٤

٥

It should be noticed that the presented tool for seismic risk and loss assessment SELENA ٦

v4.1 is an ongoing development and thus will undergo a number of changes and ٧

extensions. In order to be able to do this, the authors depend on the users’ feedback and ٨

suggestions. ٩

Since the applied methodologies for assigning damage probabilities, damage extent and loss ١٠

are scientifically substantiated and accepted by the scientific community further ١١

developments will mainly concentrate on the following aspects: ١٢

1. Provision of alternatives for the user regarding the incorporation of further ١٣

scientific developments and innovations, e.g., capacity spectrum methods, type of ١٤

design spectra, etc. ١٥

2. Improvement of the graphic user interface. ١٦

3. Modification of output files following the structure and requested data format of ١٧

commercial (e.g., ArcView, MapInfo) as well as open-source Geographic ١٨

Information Systems (e.g., Google Earth, GRASS GIS) in order to simplify the ١٩

graphical illustration of results. ٢٠

4. Initiation of a database and web-forum for possible users of SELENA ٢١

accommodating all types of input data for different parts of the world such as ٢٢

capacity curves and vulnerability functions for different model building types, ٢٣

national economic loss models for building stock or statistical data on the ٢٤

demography. ٢٥

٢٦

Acknowledgments ١

٢

This work has been developed thanks to the agreement between NORSAR and the ٣

University of Alicante (Sergio Molina) under the umbrella of the International Centre for ٤

Geohazards ICG. The funding through the SAFER project, ICG, NORSAR1-03I and ٥

GV07/045 has facilitated major developments of SELENA from its first version in 2004. ٦

Two anonymous reviewers are thanked for their thoughtful comments which improved the ٧

manuscript. ٨

٩

١٠

References ١١

١٢

Applied Technology Council, 1996a. Earthquake Loss Evaluation Methodology and Databases for ١٣

Utah, Prepared for the Federal Emergency Management Agency. ١٤

١٥

Applied Technology Council ATC, 1996b. Seismic Evaluation and Retrofit of Concrete Buildings, ١٦

Report ATC-40, Redwood City, California. ١٧

١٨

Algermissen, S.T., Arnold, E.P., Steinbrugger, K.V., Hopper, M.G., Powers, P.S., 1988. Earthquake ١٩

Losses in Central Utah, U.S. Geological Survey Open-File Report 88-680. ٢٠

٢١

Bommer, J.J., Scherbaum, F., Bungum, H., Cotton, F., Sabetta, F., and Abrahamson, N.A. (2005). ٢٢

On the Use of Logic Trees for Ground-Motion Prediction Equations in Seismic-Hazard Analysis. ٢٣

Bulletin of the Seismological Society of America 95(2): 377–389. ٢٤

٢٥

Bureau of Indian Standards BIS (2002). IS 1893 (Part 1) : 2002, Indian Standard - Criteria for ١

Earthquake Resistant Design of Structures, Part 1 - General Provisions and Buildings. ICS ٢

91.120.25, New Delhi, 2002. ٣

٤

California Division of Mines and Geology, 1982a. Earthquake Planning Scenario for a Magnitude ٥

8.3 Earthquake on the San Andreas Fault in Southern California, California Division of Mines and ٦

Geology, Special Publication 60. ٧

٨

California Division of Mines and Geology, 1982b. Earthquake Planning Scenario for a Magnitude ٩

8.3 Earthquake on the San Andreas Fault in the San Francisco Bay Area, California Division of ١٠

Mines and Geology, Special Publication 61. ١١

١٢

California Division of Mines and Geology, 1987. Earthquake Planning Scenario for a Magnitude 7.5 ١٣

Earthquake on the Hayward Fault in the San Francisco Bay Area, California Division of Mines and ١٤

Geology, Special Publication 78. ١٥

١٦

California Division of Mines and Geology, 1988. Earthquake Planning Scenario for a Major ١٧

Earthquake on the Newport-Inglewood Fault Zone, California, California Division of Mines and ١٨

Geology, Special Publication 99 (Preprint). ١٩

٢٠

California Division of Mines and Geology, 1990. Earthquake Planning Scenario for a Major ٢١

Earthquake, San Diego-Tijuana Metropolitan Area, California Division of Mines and Geology, ٢٢

Special Publication 100. ٢٣

٢٤

California Division of Mines and Geology, 1995. Earthquake Planning Scenario in Humboldt and ٢٥

Del Norte Counties, California for a Great Earthquake on the Cascadia Subduction Zone, ٢٦

California Division of Mines and Geology, Special Publication 115. ٢٧

٢٨

Campos Costa, A., Sousa, M.L., Carvalho, A., Coelho, E., 2006. Seismic loss scenarios based on ١

hazard disaggregation. Application to the metropolitan region of Lisbon, Portugal, In: Oliviera, ٢

C.S., Goula, X. and Roca, A. (Eds.) Assessing and Managing Earthquake Risk. Geo-scientific and ٣

Engineering Knowledge for Earthquake Risk Mitigation: Developments, Tools, Techniques, ٤

Springer Netherlands, pp. 449–462,. ٥

٦

Central United States Earthquake Preparedness Project CUSEPP, 1985. An Assessment of Damage ٧

and Casualties for Six Cities in the Central United States Resulting from Earthquakes in the New ٨

Madrid Seismic Zone, Prepared for the Federal Emergency Management Agency (FEMA). ٩

١٠

Coburn, A., Spence, R., 2002. Earthquake protection. Second edition. John Wiley & Sons Ltd., ١١

2002. ١٢

١٣

Cornell, C.A., 1968. Engineering seismic risk analysis. Bulletin of the Seismological Society of ١٤

America 58 (5), 1583–1606. ١٥

١٦

Crowley, H., Pinho, R., Bommer, J., 2004. A probabilistic displacement-based vulnerability ١٧

assessment procedure for earthquake loss estimation. Bulletin of Earthquake Engineering 2, 173–١٨

219. ١٩

٢٠

Dames and Moore, Inc., 1996. Earthquake loss estimation pilot study for the Portland Metropolitan ٢١

Region. Prepared for the National Institute of Building Sciences. ٢٢

٢٣

Eguchi, R.T., Goltz, J.D., Seligson, H.A., Flores, P.J., Heaton, T.H., Bortugno, E., 1997. Real-time ٢٤

loss estimation as an emergency response decision support system: The Early Post-Earthquake ٢٥

Damage Assessment Tool (EPEDAT). Earthquake Spectra 13, 815–832. ٢٦

٢٧

EQE International, Inc., 1997. Second pilot test study of the standardized nationally applicable loss ١

estimation methodology, Boston, Massachusetts. Prepared for the National Institute of Building ٢

Sciences. ٣

٤

Erdik, M., Aydinoğlu, N., 2002. Earthquake risk to buildings in Istanbul and a proposal towards its ٥

mitigation. In: Proceedings of the Second Annual IIASA-DPRI Meeting on Integrated Disaster ٦

Risk Management: Megacity Vulnerability and Resilience, Laxenburg, Austria 2002, pp. 1–10. ٧

٨

ESRI (2004). ArcGIS 9 – What is ArcGIS? White paper. ESRI, Redlands, United States. ٩

١٠

European Committee for Standardization CEN, 2002. prEN 1998-1:200X, Eurocode 8: Design for ١١

structures for earthquake resistance, Part 1: General rules, seismic actions and rules for buildings, ١٢

May 2002. ١٣

١٤

FEMA, 1996. NEHRP Guidelines for the seismic rehabilitation of buildings, FEMA 273, Federal ١٥

Emergency Management Agency, Washington, D.C. ١٦

١٧

FEMA, 1997. HAZUS®97: Earthquake Loss Estimation Methodology: User’s Manual, Federal ١٨

Emergency Management Agency, Washington, DC, Unites States. ١٩

٢٠

FEMA, 1999. HAZUS®99: Earthquake Loss Estimation Methodology: User’s Manual, Federal ٢١

Emergency Management Agency, Washington, DC, Unites States. ٢٢

٢٣

FEMA, 2001. HAZUS®99-SR1 Estimated Annualized Earthquake Losses for the United States ٢٤

(FEMA 366), Federal Emergency Management Agency, Washington, DC, Unites States, February ٢٥

2001, 33 pp. ٢٦

٢٧

FEMA, 2002. HAZUS®99-SR2 (Service Release 2), HAZUS Technical Manual, Federal Emergency ١

Management Agency, FEMA and National Institute of Building Sciences, NIBS, Washington D.C., ٢

Vol. 1–3. ٣

٤

FEMA, 2004. Using HAZUS-MH for Risk Assessment (FEMA 433), Federal Emergency ٥

Management Agency, Washington, DC, United States, August 2004. ٦

٧

FEMA, 2005a. HAZUS®MH MR1: Multi-hazard Loss Estimation Methodology Maintenance ٨

Release 1 (MR1), Federal Emergency Management Agency. ٩

١٠

FEMA, 2005b. Improvement of nonlinear static seismic analysis procedures (FEMA 440). Prepared ١١

by Applied Technology Council (ATC-55 Project) for Federal Emergency Management Agency, ١٢

Washington, DC, United States, June 2005. ١٣

١٤

FEMA, 2006. HAZUS®MH MR2: Multi-hazard Loss Estimation Methodology Maintenance ١٥

Release 2 (MR2), Federal Emergency Management Agency. ١٦

١٧

FEMA, 2007. HAZUS®MH MR3: Multi-hazard Loss Estimation Methodology Maintenance ١٨

Release 3 (MR3), Federal Emergency Management Agency. ١٩

٢٠

FEMA, 2008. HAZUS-MH Estimated Annualized Earthquake Losses for the United States ٢١

(FEMA 366). Federal Emergency Management Agency, Washington, DC, United States, April ٢٢

2008, 66 pp. ٢٣

٢٤

Field, E.H., Seligson, H.A., Gupta, N., Gupta, V., Jordan, T.H., Campbell, K.W., 2005. Loss ٢٥

estimates for a Puente Hills blind-thrust earthquake in Los Angeles, California. Earthquake Spectra, ٢٦

21, 2, 329–338. ٢٧

٢٨

Hansen, R., Bausch, D., 2006. A GIS-Based Methodology for Exporting the Hazards U.S. ١

(HAZUS) Earthquake Model for Global Applications. Earthquake Workshop – Malta, ٢

http://www.usehazus.com/docs/gis_global_hazus_paper.pdf , 64 pp. ٣

٤

Harlan, M.R., Lindbergh, C. 1988. An Earthquake vulnerability analysis of the Charleston, South ٥

Carolina Area, Department of Civil Engineering, The Citadel, Charleston, South Carolina. ٦

٧

International Code Council ICC, 2006. 2006 International Building Code (IBC-2006), United ٨

States, January 2006, 664 pp. ٩

١٠

Kircher, C.A. (2003). Earthquake loss estimation methods for welded steel moment-frame ١١

buildings. Earthquake Spectra, 19, 2, 365–384 ١٢

١٣

Kircher, C.A., Seligson, H.A., Bouabid, J., Morrow, G.C., 2006. When the Big One strikes again – ١٤

Estimated Losses due to a Repeat of the 1906 San Francisco Earthquake. Earthquake Spectra, 22, ١٥

S2, S297–S339. ١٦

١٧

Lang, D.H., Molina, S., Lindholm, C.D., 2008a. Towards near-real-time damage estimation using a ١٨

CSM-based tool for seismic risk assessment. Journal of Earthquake Engineering, 12: 1, 199–210. ١٩

٢٠

Lang, D.H., Gutiérrez, V., Lindholm, C.D., 2008b. RISe – Risk Illustrator for SELENA, User ٢١

Manual v1.0, NORSAR, Kjeller (Norway), December 2008. ٢٢

٢٣

Lang, D.H., Gutiérrez C., F.V. (2009). RISe: Illustrating geo-referenced data of seismic risk and loss ٢٤

assessment studies using Google Earth, Earthquake Spectra, Technical Note (accepted). ٢٥

٢٦

Luco, N., 2007. Modeling Vulnerability, e.g. for Residential Risk – “ResRisk”. First International ٢٧

Workshop on Open-Source Risk Software, Pasadena, CA, February 2007. ٢٨

١

McGuire, R.K., 2004. Seismic Hazard and Risk Analysis. EERI Publication No. MNO-10. 221 pp. ٢

Earthquake Engineering Research Institute, Oakland, CA. ٣

٤

Mitrani-Reiser, J., 2007. An ounce of prevention: Probabilistic loss estimation for performance-٥

based earthquake engineering. Tech. Rep. EERL 2007-01, Earthquake Engineering Research ٦

Laboratory, California Institute of Technology, Pasadena, California, USA. ٧

٨

Molina, S., Lindholm, C.D., 2005. A logic tree extension of the capacity spectrum method ٩

developed to estimate seismic risk in Oslo, Norway. Journal of Earthquake Engineering 9(6), 877–١٠

897. ١١

١٢

Molina, S., Lang, D.H., Lindholm, C.D., 2009. SELENA v4.1 – User and Technical Manual v4.1, ١٣

NORSAR, Kjeller (Norway), April 2009. ١٤

١٥

Muto, M., Krishnan, S., Beck, J. L., Mitrani-Reiser, J., 2008. Seismic Loss Estimation Based on ١٦

End-to-End Simulation. Proceedings of the first international symposium on life-cycle civil ١٧

engineering, Varenna, Lake Como, Italy. ١٨

١٩

Naeim, F., Alimoradi, A., Hagie, S., Comartin, C., 2007. PACT – Performance Assessment ٢٠

Calculation Tool - User Guide for Beta Version 1.0, May 27, 2007. ATC-58 Guidelines for Seismic ٢١

Performance Assessment of Buildings. Prepared for Department of Homeland Security, Federal ٢٢

Emergency Management Agency. ٢٣

٢٤

National Oceanic and Atmospheric Administration NOAA, 1973. A study of earthquake losses in ٢٥

the Los Angeles, California Area. Prepared for the Federal Disaster Assistance Administration, ٢٦

Department of Housing and Urban Development. ٢٧

٢٨

Nielson, B.G., DesRoches, R., 2007. Analytical seismic fragility curves for typical bridges in the ١

central and southeastern United States. Earthquake Spectra, 23, 3, 615–633. ٢

٣

Oliveira, C.S., Roca, A., Goula, X., 2006. Assessing and managing earthquake risk: geo-scientific ٤

and engineering knowledge for earthquake risk mitigation: developments, tools, techniques. Ed. ٥

Springer, 543 pp. ٦

٧

Porter, K.A., Kiremidjian, A.S., 2001. Assembly-Based Vulnerability of Buildings and its Uses in ٨

Seismic Performance Evaluation and Risk-Management Decision-Making, Report 139, John A. ٩

Blume Earthquake Engineering Research Center, Stanford, CA. ١٠

١١

Scherbaum, F, Bommer, J.J., Bungum, H., Cotton, F., Abrahamson, N.A., 2005. Composite ١٢

ground-motion models and logic trees: methodology, sensitivities and uncertainties. Bulletin of the ١٣

Seismological Society of America, 95, 5, 1575–1593. ١٤

١٥

Rojas, H.A., Pezeshk, S., Foley, C.M., 2007. Performance-based optimization considered both ١٦

structural and non-structural components. Earthquake Spectra, 23, 3, 685–709. ١٧

١٨

Spencer, B., Elnashai, A., Myers, J., Hampton, S., 2008. MAEviz Overview. ١٩

http://mae.ce.uiuc.edu/software_and_tools/maeviz.html, Mid-America Earthquake Center, ٢٠

2008. ٢١

٢٢

Webb, T.M., 1999. NHEMATIS: Project overview and future framework. A report prepared by ٢٣

Nobility Environmental Software Systems Inc., Vancouver, B.C. for Emergency Preparedness ٢٤

Canada, 23 pp. ٢٥

٢٦

Wong, I., Bouabid, J., Graf, W., Huyck, C., Porush, A., Silva, W., Siegel, T., Bureau, G., Eguchi, R., ١

Knight, J., 2005. Potential losses in a repeat of the 1886 Charleston, South Carolina Earthquake. ٢

Earthquake Spectra, 21, 4, 1157–1184. ٣

٤

Wyss, M., 2005. Earthquake loss estimates applied in real time and to megacity risk assessment, In: ٥

Carle, B. and Van de Walle, B. (Eds.) Proceedings of the Second International ISCRAM ٦

Conference, Brussels, Belgium, April 2005, pp. 297–299. ٧