Eqrmtech Manual

EQRM: Geoscience Australias Earthquake RiskModel

Technical Manual

Version 4.0 - DRAFT

This document is in draft form and has not been internally

reviewed and officially released by Geoscience Australia

David Robinson, Nick Horspool, Hadi Ghasemi and Duncan Gray

Geoscience Australia

August 22, 2012

This page is intentionally left blank.

i

This page is intentionally left blank.

ii

AcknowledgementsThe authors wish to thank John Schneider, leader of the Risk and Impact Ananl-ysis Group at Geoscience Australia. Johns expertise in the areas of earthquakehazard and risk analysis have proven invaluable throughout the project. Withouthis ongoing support and advice the EQRM would not be half the product that itis today.

We wish to thank Glenn Fulford and Trevor Dhu who drove and supported EQRMduring its initial developement. They were part of the original team that devel-oped the scientific concepts of EQRM. Peter Row and Ken Dale are also acknowl-edged for their coding efforts in particular modules of the EQRM software.

Andres Mendez from Aon Re in Chicago is acknowledged for providing softwarethat formed the backbone of the earthquake catalogue generation. This generouscontribution, as well as his high level of collaborative support, was fundamentalto the projects success.

The authors have sought advice on specific topics from a range of people involvedin the earthquake hazard and risk fields. The following are acknowledged for theirexpertise and assistance; George Walker of Aon Re, Professor John McAneney ofRisk Frontiers, Walt Silva of Pacific Engineering and Analysis, Don Windeler andhis team at Risk Management Solutions (RMS), Mike Griffith at The Universityof Adelaide, Gary Gibson of the Seismology Research Centre and Brian Gaullfrom Guria Consulting.

A number of people at Geoscience Australia have contributed to the developmentof the EQRM. Members of the Newcastle and Lake Macquarie project team leadby Trevor Jones are acknowledged for their assistance in shaping the early de-sign of the EQRM. Mark Edwards and Ken Dale are thanked for the countlessengineering questions that they have fielded over the four year project. JaneSexton, project leader of the Natural Hazard Impacts Project is thanked for cre-ating the productive and supportive environment that allowed this project toflourish. EQRM users, Cvetan Sinadinovski, Annette Patchett, Ken Dale andAugusto Sanabria are thanked for their feedback and for continually inspiringthe incorporation of new functionality. Ole Nielsen is are acknowledged for theprovision of software engineering advice and for generally improving the codingabilities of the authors. Angie Jaensch, Greg Michalowski and Fiona Watford areacknowledged for drafting some of the figures. Finally, the authors wish to thankDavid Burbidge, Hyeuk Ryu and Jonathan Griffin for reviewing the drafts of thisreport. Their reviews lead to substantial improvements to this manual.

iii

Contents

1 Introduction 1

1.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.2 Using this manual . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.3 About this manual . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2 The EQRM application 5

2.1 The EQRM Control File . . . . . . . . . . . . . . . . . . . . . . . 6

2.2 The Source Files . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

2.2.1 Source Zone File . . . . . . . . . . . . . . . . . . . . . . . 24

2.2.2 Source Fault File . . . . . . . . . . . . . . . . . . . . . . . 29

2.3 Event Type Control File . . . . . . . . . . . . . . . . . . . . . . . 32

2.4 Site Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

2.4.1 Hazard Site File . . . . . . . . . . . . . . . . . . . . . . . . 36

2.4.2 Risk Site File (Building Database) . . . . . . . . . . . . . 36

Building construction types . . . . . . . . . . . . . . . . . 38

Building usage types . . . . . . . . . . . . . . . . . . . . . 39

Replacement costs . . . . . . . . . . . . . . . . . . . . . . 41

2.4.3 Bridge Site File . . . . . . . . . . . . . . . . . . . . . . . . 44

iv

2.4.4 Fatality Site File . . . . . . . . . . . . . . . . . . . . . . . 47

2.5 Amplification File . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

2.6 Vulnerability File . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

3 Earthquake source generation 55

3.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

3.2 Creating an earthquake catalogue for probabilistic seismic hazardanalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

3.3 Simulated events and virtual faults . . . . . . . . . . . . . . . . . 57

3.4 Magnitude selection and event activity . . . . . . . . . . . . . . . 59

3.4.1 Activity rate for areal sources . . . . . . . . . . . . . . . . 62

3.4.2 Activity rate for faults with bounded Gutenberg Richtermodel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

3.4.3 Activity rate for faults with bounded Characteristic Earth-quake model . . . . . . . . . . . . . . . . . . . . . . . . . . 63

3.4.4 Event Magnitude Examples . . . . . . . . . . . . . . . . . 65

3.5 Generating synthetic earthquakes in areal source zones . . . . . . 70

3.5.1 Dimensions and position of the rupture plane . . . . . . . 70

(Wells and Coppersmith, 1994) Scaling Laws . . . . . . . . 70

Modified (Wells and Coppersmith, 1994) Scaling Laws . . 71

Leonard (2010) Stable Continental Regions Scaling Law . . 72

3.5.2 Azimuth and dip of rupture . . . . . . . . . . . . . . . . . 72

3.5.3 Location of synthetic events in areal sources . . . . . . . . 73

3.5.4 Overlapping source zones . . . . . . . . . . . . . . . . . . . 74

3.6 Generating synthetic earthquakes on pre-defined fault sources . . 74

3.6.1 Dimensions of the rupture plane . . . . . . . . . . . . . . . 74

v

3.6.2 Location of synthetic events on defined fault sources . . . . 76

3.6.3 Crustal Faults and Subduction Interface Faults . . . . . . 76

3.6.4 Intraslab Faults . . . . . . . . . . . . . . . . . . . . . . . . 79

Case if 0 r 90 . . . . . . . . . . . . . . . . . . . . . . 80

Case if 90 < r 175 . . . . . . . . . . . . . . . . . . . . . 84

Case if 185 < r f + . . . . . . . . . . . . . . . . . . 84

3.7 Spawning events . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

3.8 Analysing a scenario event . . . . . . . . . . . . . . . . . . . . . . 88

4 Ground-Motion Prediction Equations 89

4.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

4.2 Background theory . . . . . . . . . . . . . . . . . . . . . . . . . . 89

4.3 Implemented GMPEs in EQRM . . . . . . . . . . . . . . . . . . . 93

4.4 Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

4.4.1 Further Comments on Specific GMPEs . . . . . . . . . . . 99

Gaull et al. (1990a) . . . . . . . . . . . . . . . . . . . . . . 99

Somerville (2009) . . . . . . . . . . . . . . . . . . . . . . . 101

4.5 Accounting for uncertainties . . . . . . . . . . . . . . . . . . . . . 102

4.6 Incorporating aleatory variability . . . . . . . . . . . . . . . . . . 103

4.6.1 Random sampling of a response spectral acceleration . . . 103

4.6.2 Sampling the probability density function of the responsespectral acceleration (spawning) . . . . . . . . . . . . . . . 104

4.6.3 Recomendation for sampling GMPE aleatory Variability . 106

4.7 Using multiple GMPEs - Incorporating epistemic uncertainty . . . 106

4.8 Collapse versus no-collapse . . . . . . . . . . . . . . . . . . . . . . 107

vi

5 Regolith amplification 109

5.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

5.1.1 Background theory . . . . . . . . . . . . . . . . . . . . . . 109

5.2 Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

5.3 Incorporating aleatory uncertainty . . . . . . . . . . . . . . . . . . 113

6 Building damage 114

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114

6.2 The capacity spectrum method . . . . . . . . . . . . . . . . . . . 114

6.2.1 The building capacity curve . . . . . . . . . . . . . . . . . 116

Fitting the building capacity curve . . . . . . . . . . . . . 117

Variability of the capacity curves . . . . . . . . . . . . . . 118

6.2.2 Damping the demand curve . . . . . . . . . . . . . . . . . 118

Modification of elastic damping . . . . . . . . . . . . . . . 119

Hysteretic damping . . . . . . . . . . . . . . . . . . . . . . 121

6.2.3 Finding the intersection point . . . . . . . . . . . . . . . . 123

6.3 Fragility curves . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124

6.3.1 Form of fragility curves . . . . . . . . . . . . . . . . . . . . 125

6.3.2 Damage state thresholds . . . . . . . . . . . . . . . . . . . 126

6.3.3 Variability of the damage states . . . . . . . . . . . . . . . 127

6.3.4 Incremental probabilities . . . . . . . . . . . . . . . . . . . 128

6.4 Differences from HAZUS methodology . . . . . . . . . . . . . . . 128

6.4.1 Extra features . . . . . . . . . . . . . . . . . . . . . . . . . 129

vii

7 Losses 130

7.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130

7.2 Direct financial loss . . . . . . . . . . . . . . . . . . . . . . . . . . 130

7.2.1 General financial loss equations: loss for a single building . 130

7.2.2 Aggregated loss and survey factors . . . . . . . . . . . . . 134

7.2.3 Cutoff values . . . . . . . . . . . . . . . . . . . . . . . . . 134

7.3 Social losses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134

8 Vulnerability 137

8.1 MMI conversion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137

8.2 Determining mean loss and sigma . . . . . . . . . . . . . . . . . . 138

8.3 Probabilistic sampling . . . . . . . . . . . . . . . . . . . . . . . . 139

8.4 Calculating loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

9 Hazard and risk results 141

9.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141

9.2 Calculating hazard and risk . . . . . . . . . . . . . . . . . . . . . 141

9.2.1 Computing the annual exceedance rate . . . . . . . . . . . 142

9.3 Earthquake hazard results . . . . . . . . . . . . . . . . . . . . . . 143

9.3.1 Hazard maps . . . . . . . . . . . . . . . . . . . . . . . . . 143

9.3.2 Hazard exceedance curves . . . . . . . . . . . . . . . . . . 143

9.4 Earthquake risk results . . . . . . . . . . . . . . . . . . . . . . . . 145

9.4.1 Risk exceedance curve . . . . . . . . . . . . . . . . . . . . 146

9.4.2 Annualised loss . . . . . . . . . . . . . . . . . . . . . . . . 147

9.4.3 Disaggregated annualised loss . . . . . . . . . . . . . . . . 148

9.5 Earthquake scenario results . . . . . . . . . . . . . . . . . . . . . 149

viii

A Appendicies 151

A.1 A selection of GMPE formulae . . . . . . . . . . . . . . . . . . . . 151

A.1.1 Toro ground motion formula . . . . . . . . . . . . . . . . . 151

A.1.2 Gaull ground motion formula . . . . . . . . . . . . . . . . 152

A.1.3 Atkinson and Boore ground motion formula . . . . . . . . 153

A.1.4 Sadigh ground motion formula . . . . . . . . . . . . . . . . 153

A.1.5 Somerville ground motion formula . . . . . . . . . . . . . . 154

ix

Chapter 1

Introduction

1.1 Overview

The EQRM application is a computer model for estimating earthquake hazardand earthquake risk. Modelling earthquake hazard involves assessing the proba-bility that certain levels of ground motion will be exceeded. Modelling of earth-quake risk involves estimating the probability of a building portfolio experiencinga range of earthquake induced losses. For any number of synthetic earthquakes,the EQRM application can be used to estimate:

1. the ground motion and its likelihood of occurrence (earthquake hazard),

2. the direct financial loss and its likelihood of occurrence (earthquake risk),and

3. less reliably the number of fatalities and their likelihood of occurrence(earthquake risk).

The EQRM application is Geoscience Australias centerpiece for modelling earth-quake hazard and risk. Its use formed the basis for Geoscience Australias re-cent reports on Earthquake risk in the Newcastle and Lake Macquarie (Dhu andJones, 2002) and Perth (Sinadinovski, Edwards, Corby, Milne, Dale, Dhu, Jones,McPherson, Jones, Gray, Robinson and White, 2005) regions.

EQRM is a product of Geoscience Australia, an Australian Government Agencyand is open-source. It can be downloaded from https://sourceforge.net/projects/eqrm/.

The process for computing earthquake hazard can be described by the followingsteps:

1

1. the generation of a catalogue of synthetic earthquakes (or events) (see Chap-ter 3),

2. the propagation or attenuation (see Chapter 4) of the seismic wave fromeach of the events in 1 to locations of interest (see Chapter 2),

3. accounting for the interactions between the propagating seismic wave andthe local geology or regolith (see Chapter 5), and

4. accounting for the probability of each event and the estimation of hazard(see Chapter 9).

The process for computing earthquake risk shares the same first three steps asthe earthquake hazard. The fourth step and onwards can be described as follows:

4. estimating the probability that the portfolio buildings (see Chapter 2) willexperience different levels of damage (see Chapter 6),

5. the computation of direct financial loss as a result of the probabilities com-puted in 4 (see Chapter 7), and

6. assembling the results to compute the risk (see Chapter 9).

1.2 Using this manual

This manual describes the EQRM application; it has been designed to serve thefollowing three purposes:

1. describe the theory and methodology behind the EQRM application;

2. explain how to use the EQRM application to model earthquake hazard andrisk; and

3. provide enough information to assist those who may wish to access individ-ual modules and modify them.

A number of features have been included to assist readers of the manual. Thesefeatures include:

Text highlighting - Important parameters, file names and code are iden-tified by a typeset font.

Index - An index has been introduced to allow speedy of navigation of themanual.

2

1.3 About this manual

Chapter 2: The EQRM application introduces the EQRM software. The chapterdescribes the directory structure and introduces important input parameter files.Setup and operational processes are discussed and the conversion of the EQRMto a stand-alone executable described. Readers who do not wish to familiarisethemselves with details of the EQRM application may wish to skip this chapter.

Chapter 3: Earthquake source generation discusses the creation of an earthquakecatalogue. At the core of any hazard or risk assessment is the simulation ofsynthetic earthquakes and the creation of a catalogue of synthetic events. Thechapter describes how the simulation of earthquakes can focus around a specificearthquake of interest (a scenario based simulation) or how it can model the effectof all foreseeable events (a probabilistic simulation).

Chapter 4: Ground-Motion Prediction Equations discusses the propagation (orattenuation) of the motion from each of the synthetic earthquakes in the eventcatalogue to locations of interest. Different measures of distance between an eventand a site of interest are introduced. The chapter also describes the attenuationmodels that can be used with the EQRM application.

Chapter 5: Regolith amplification discusses how local geology, or regolith, can beincorporated in an earthquake hazard or risk assessment. The chapter illustrateshow the presence of regolith can lead to amplification (or de-amplification) ofboth ground and building motion, and consequently result in higher (or lower)levels of hazard and risk.

Chapter 6: Building damage describes how the EQRM application estimatesthe probability that each building will experience different levels of damage. Thechapter gives a brief discussion of the theory behind making such estimates as wellas outlining how to extrapolate the estimates to an entire portfolio of buildingsin a region of interest.

Chapter 7: Losses illustrates how to model the direct financial loss as a result ofthe damage estimates described in Chapter 6. The chapter also provides a briefdescription of how the EQRM application can estimate fatalities.

Chapter 9: Hazard and Risk Results describes how the results from the EQRMapplication can be summarised and displayed. There are many ways to displayestimates of earthquake hazard and risk. Earthquake hazard is commonly mod-elled in terms of a probability (often 10%) of a particular ground motion (usuallyacceleration) being exceeded in some time frame (often 50 years). In the case ofa scenario based simulation, earthquake risk can be modelled in terms of dollar

3

estimates of building loss. In the case of a probabilistic simulation, earthquakerisk is often modelled in terms of one or both of the following:

1. the loss per year averaged over a long period of time, typically 10000 years.This is known as the annualised loss.

2. the probability that different levels of loss will be exceeded. The plot of lossagainst such exceedance probabilities is referred to as a probable maximumloss curve.

4

Chapter 2

The EQRM application

The Earthquake Risk Model (EQRM) is capable of:

1. earthquake scenario ground motion modeling;

2. scenario loss forecasts;

3. probabilistic seismic hazard analysis (PSHA); and

4. probabilistic seismic risk analysis (PSRA).

This chapter describes the EQRM application. Input files and parameters arediscussed and directions on how to run the EQRM provided. Readers who areinterested in only the EQRM methodology and not the EQRM software packagemay wish to skip this chapter.

The input files required by the EQRM depend on the nature of the simulationconducted. For example, the inputs for a scenario loss simulation are differ-ent to those required for a probabilistic seismic hazard analysis. Table 2.1 pro-vides a summary of the inputs required by the EQRM. The EQRM Demos in*/eqrm core/demo provide examples of each input file and demonstrate how torun the EQRM for each of the four main simulation types. The following sectionprovide an overview of each of the main input files.

Except for the control file all input files are assumed to be in the input directory,input dir, specified in the EQRM control file. If a file is not there it is lookedfor in */eqrm core/resources/data.

5

Table 2.1: Input files required for different types of simulation with the EQRM.The asterisks indicate optional input files, the requirement for which depends onsettings in the EQRM control file

hazard riskscenario EQRM control file EQRM control file

amplification factors* amplification factors*hazard grid building database

probabilistic EQRM control file EQRM controlfilesource file(s) source file(s)event type control file event type control fileamplification factors* amplification factors*hazard grid building database

2.1 The EQRM Control File

The EQRM control file is the primary input file for an EQRM simulation. It:

1. contains a series of input variables (or parameters) that define the mannerin which the EQRM is operated; and

2. initialises the simulation.

For example, there is a parameter to control whether the EQRM models hazardor risk. Other parameters can be used to identify return periods or indicatewhether site amplification is considered. A description of all of the parameters isgiven below.

Note that it not essential to supply all parameters for each simulation. Forexample, if amplification is not being used (i.e. use amplification = False) itis not necessary to supply the remaining amplification parameters. Furthermore,default values are set by the EQRM for several parameters. These are indicatedbelow when applicable. Omission of these input parameters in the EQRM controlfile will lead to use of the default values. For example, the default value foratten threshold distance is 400 km.

The following also provides suggested values for several parameters. Users arefree to change these values as desired. The developers are merely suggesting thevalue they would use in most circumstances. For example, the suggested valuefor loss min pga is 0.05.

6

The term preferred is used to indicate those parameters that the developersbelieve to be most appropriate. For example, the preferred value for csm -hysteretic damping is curve. In this case the alternative choices of None andtrapezoidal would typically only be used for experimental purposes.

A simulation can be started by executing an EQRM control file or by executinganalysis.py with a control file as the first parameter.

Two ways of running EQRM:>python EQRM control file.py

OR>python analysis.py EQRM control file.py

The control file is a Python script, so Python code can be used to manipulateparameter values. Note though that all variables other than the parameter valuesmust be deleted to avoid passing unknown variables into EQRM.

Acronyms:PSHA is probabilistic seismic hazard analysisPSRA is probabilistic seismic risk analysisGMPE is ground motion prediction equationPGA is peak ground acceleration (usually in units of g)RSA is response spectral acceleration (usually in units of g)CSM is capacity spectrum method

7

General Input:

run type: (Mandatory)Defines the operation mode of the EQRM:hazard Scenario RSA and PSHA (probabilistic hazard);risk csm Scenario Loss and PSRA (probabilistic risk) based on fragility

curves;risk mmi Scenario Loss and PSRA (probabilistic risk) based on user-defined

vulnerability curves;fatality Scenario Fatality forecast (based on MMI and population from

USGS Open-File Report 2009-1136);bridge Bridge damage (based on fragility curves from Dale 2004).

is scenario: (Mandatory)Event simulation type:True a specific scenario event (Use Scenario input);False probabilistic simulation, PSHA or PSRA (Use Probabilistic input)

site tag: (Mandatory)String used in input and output file names. Typically used to define the city orregion of interest (e.g. newc is used in the demos).

site db tag:DEFAULT = String used to specify the exposure or building data base. The building data basefile name is sitedb .csv The exposure data base filename is par site.csv

return periods:List whose elements represent the return periods to be considered for PSHA.

input dir:Directory containing any local input files.

output dir:Directory for output files. This directory will be created if not present

use site indexes:DEFAULT = FalseTrue sample sites with indices in site indexes (for testing simulations);False No sub-sampling.

8

site indexes:List whose elements represent the site indices to be used (if use site indexes =True). The index of the first row of data (i.e. first data row in site file) is 1.

fault source tag:Tag for specifying a source fault file.If fault source tag is defined the filename for the fault source file is fault source .xml. Otherwise it is not used.Note that one of fault source tag and zone source tag must be set.

zone source tag:Tag for specifying a source zone file.If zone source tag is defined the filename for the zone source file is zone source .xml. Otherwise it is not used.Note that one of zone source tag and fault source tag must be set.

event control tag:Tag for specifying a event control file.If event control tag is defined the filename for the event control is event control .xml Otherwise it is not used.

9

Scenario Input:

scenario azimuth:Azimuth of the scenario event (degrees from true North).

scenario depth:Depth to event centroid (km).

scenario latitude:Latitude of rupture centroid.

scenario longitude:Longitude of rupture centroid.

scenario magnitude:Moment magnitude of event.

scenario dip:Dip of rupture plane (degrees from horizontal).

scenario max width: (Optional)Maximum width along virtual faults i.e. rupture width can not exceedscenario max width (km).

scenario width:DEFAULT = NoneWidth of rupture centroid (km).If None, dip, magnitude, area and scenario max width used to calculate using aWells and Coppersmith (1994) conversion.

scenario length:DEFAULT = NoneLength of rupture centroid (km).If None, calculated as area/width.

scenario number of events:The desired number of copies of the event to be generated. Typically, copies aretaken if random sampling is used to incorporate aleatory uncertainty in GMPE(i.e. atten variability method= 2), amplification (i.e. amp variability -method= 2) or the CSM (csm variability method= 3).

10

Probabalistic Input:

All of the parameters in this section can be specified in the fault source or zonesource xml files. Specifying them here will override the values in the xml files.

prob number of events in zones:A list, where each element is the desired number of events for each zone source.

prob number of events in faults:A list, where each element is the desired number of events for each fault source.

11

Ground Motion Input:

atten models:The list of GMPEs, for the logic tree. Specifying them here will override thevalues in the xml files. This should only be used for scenario simulations. SeeChapter 2.3 for GMPE name values.

atten model weights:The list of GMPE weights, for the logic tree. The weights must sum to one.Specifying them here will override the values in the xml files.This should only beused for scenario simulations.

atten collapse Sa of atten models:DEFAULT = FalseSet to True to collapse the surface accelerations when multiple GMPEs are used.

atten variability method:DEFAULT = 2Technique used to incorporate GMPE aleatory uncertainty:None No sampling;1 spawning;2 random sampling;3 +2;4 +;5 ;6 2.

atten periods:Periods for RSA. Values must ascend. The first value must be 0.0.

atten threshold distance:DEFAULT = 400Threshold distance (km) beyond which motion is assigned to zero.

atten spawn bins:DEFAULT = 1Number of bins created when spawning.

12

atten override RSA shape:DEFAULT = NoneUse GMPE for PGA only and change shape of RSA. If None use RSA as definedby GMPE, otherwise ifAust standard Sa use RSA shape from Australian earthquake loading

standard;HAZUS Sa use RSA shape defined by HAZUS;

Supported run type = risk csm

atten cutoff max spectral displacement:DEFAULT = FalseTrue cutoff maximum spectral displacement.False no cutoff applied to spectral displacement.


atten pga scaling cutoff:DEFAULT = 2The maximum acceptable PGA in units g. RSA at all periods re-scaled accord-ingly.

atten smooth spectral acceleration:DEFAULT = FalseTrue Smooth RSA;False No smoothing applied to RSA.

13

Amplification Input:

use amplification:If set to True use amplification associated with the local regolith. Nature ofamplification varies depending on the GMPE. If GMPE has a VS30 term thenthis will be used to compute RSA on regolith. Otherwise, RSA is computed onbedrock and amplification factors used to transfer this to regolith surface.

amp variability method:DEFAULT = 2Technique used to incorporate amplification aleatory uncertainty:None No sampling;2 random sampling;3 +2;4 +;5 ;6 2.7 2.

amp min factor:SUGGESTED = 0.6Minimum accepted value for amplification factor. This minimum is not used forVs30 models.

amp max factor:SUGGESTED = 10000Maximum accepted value for amplification factor. This maximum is not used forVs30 models.

14

Building Classes Input:

buildings usage classification:Building usage classification system - HAZUS or FCB

Supported run type = risk csm, risk mmi

buildings set damping Be to 5 percent:SUGGESTED = FalseIf True a damping Be of 5% will be used for all building structures.


bridges functional percentages:Functional percentages used to estimate the number of days to complete repairs.Normal curves are used using mean and sigma for each damage state as specifiedin table 8, Dale 2004.Setting this parameter will produce one file per functional percentage, in theformat bridge days to complete fp[]>.csv.

Supported run type = bridge

15

Capacity Spectrum Method Input:

csm use variability:SUGGESTED = TrueTrue use the variability method described by csm variability method;False no aleatory variability applied.


csm variability method:SUGGESTED = 3Method used to incorporate variability in capacity curve:None No sampling;3 Random sampling applied to ultimate point only and yield point

re-calculated to satisfy capacity curve shape constraint.


csm standard deviation:SUGGESTED = 0.3Standard deviation for capacity curve lognormal PDF.


csm damping regimes:PREFERRED = 0Damping multiplicative formula to be used:0 PREFERRED: use Ra, Rv, and Rd;1 use Ra, Rv and assign Rd = Rv;2 use Rv only and assign Ra = Rd = Rv.


csm damping modify Tav:PREFERRED = TrueModify transition building period i.e. corner period Tav:True PREFERRED: modify as in HAZUS;False do NOT modify.


16

csm damping use smoothing:PREFERRED = TrueSmoothing of damped curve:True PREFERRED: apply smoothing;False NO smoothing.


csm hysteretic damping:PREFERRED = curveTechnique for Hysteretic damping:None no hysteretic dampingtrapezoidal Hysteretic damping via trapezoidal approximation;curve PREFERRED: Hysteretic damping via curve fitting.


csm SDcr tolerance percentage:SUGGESTED = 1.0Convergence tolerance as a percentage for critical spectral displacement in non-linear damping calculations.


csm damping max iterations:SUGGESTED = 7Maximum iterations for nonlinear damping calculations.


building classification tag :DEFAULT = Reference for structural damage file. The base file name isbuilding parameters.csv.


damage extent tag:DEFAULT = Reference for damage extent file. The base file name isdamage extent.csv.


17

Loss Input:

loss min pga:SUGGESTED = 0.05Minimum PGA(g) below which financial loss is assigned to zero.


loss regional cost index multiplier:SUGGESTED = 1Regional cost index multiplier to convert dollar values in building database todesired regional and temporal (i.e. inflation) values.


loss aus contents:SUGGESTED = 0Contents value for residential buildings and salvageability after complete buildingdamage:0 contents value as defined in building database and salvageability of

50%;1 60% of contents value as defined in building database and salvageability

of zero.


Vulnerability Input:

vulnerability variability method:DEFAULT = 2Technique used to sample mean loss derived from vulnerability curves:None No sampling;2 random sampling;3 +2;4 +;5 ;6 2.7 2.

Supported run type = risk mmi

18

Save Input:

save hazard map:DEFAULT = FalseTrue Save data for hazard maps (Use for saving PSHA results). Specificallyspectral acceleration with respect to location, return period and period.

save total financial loss:DEFAULT = FalseTrue Save total financial loss.


save building loss:DEFAULT = FalseTrue Save building loss.

Supported run type = risk csm, risk mmi

save contents loss:DEFAULT = FalseTrue Save contents loss.


save motion:DEFAULT = FalseTrue Save RSA motion (use for saving scenario ground motion results). Specif-ically spectral acceleration with respect to spawning, ground motion model, re-curence model, location, event and period.

save prob strucutural damage:DEFAULT = FalseTrue Save structural non-cumulative probability of being in each damage state.Note this is only supported for a single event.

Supported run type = risk csm, bridge

save fatalities:DEFAULT = FalseTrue Save fatality forecast (based on MMI and population from USGS Open-File Report 2009-1136).Setting this parameter will produce the file fatalities.txt.

Supported run type = fatality

19

Data Input:

event set handler:DEFAULT = generateSets the mode that the event set generator uses to produce an event set to workon.generate Generate a new event set sample.save Generate a new event set sample, save and exit.load Load an event set (generated using save).

event set name:DEFAULT = current event setName used to identify the event set data.For event set handler options:save Save event set to data dir/event set name.load Load event set from data dir/event set name.

data dir:Directory used to save to and load from event set data files.For event set handler options:generate If not set default to output dir.save Mandatory and must exist.load Mandatory and must exist.

data array storage:DEFAULT = data dirDirectory used to store internal data files. Used to reduce the memory footprintof EQRM.Note: It is recommended that this be on a fast local filesystem.

file array:DEFAULT = TrueTurn on/off file based array support.

20

Log Input:

file log level:DEFAULT = debugLevel of verbosity in file log.Options (in decreasing verbosity):debug

info

warning

error

critical

console log level:DEFAULT = infoLevel of verbosity in console output.Same options as for log level. Must be set to less than or equal verbosity tolog level.

21

The following grey shaded box provides an example of an EQRM controlfile toundertake a PSHA.

EQRM parameter f i l e

Al l input f i l e s are f i r s t searched f o r in the input d i r , then in ther e s ou r c e s /data d i r e c to ry , which i s part o f EQRM.

Al l d i s t an c e s are in k i l omet e r s .Acc e l e r a t i on va lue s are in g .Angles , l a t i t u d e and long i tude are in decimal degree s .

I f a f i e l d i s not used , s e t the value to None .

from os . path import j o i nfrom eqrm code . pa r s e i n pa ramete r s import eqrm data home , g e t t ime u s e r

# Operation Moderun type = hazardi s s c e n a r i o = Fal semax width = 15s i t e t a g = newcs i t e d b t a g = r e t u rn p e r i od s = [10 , 50 , 100 , 10000 ]i npu t d i r = j o i n ( . , input )ou tput d i r = j o i n ( . , output , prob haz )u s e s i t e i n d e x e s = Trues i t e i n d e x e s = [2255 , 11511 , 10963 , 686 ]z one sou r c e t ag = ev en t c on t r o l t a g =

# Scenar io input

# P r o b a b i l i s t i c input

# Attenuationatten mode l s = [ Sadigh 97 ]a t ten mode l we ight s = [ 1 ]a t t e n c o l l a p s e S a o f a t t e n mod e l s = Truea t t en va r i ab i l i t y me thod = 2a t t en p e r i od s = [ 0 . 0 , 0 .29999999999999999 , 1 . 0 ]a t t e n t h r e s h o l d d i s t a n c e = 400atten overr ide RSA shape = Nonea t t en cu t o f f max sp e c t r a l d i s p l a c emen t = Falsea t t e n p g a s c a l i n g c u t o f f = 2a t t e n smoo th s p e c t r a l a c c e l e r a t i o n = Nonea t t en l og s i gma eq we i gh t = 0

# Ampl i f i c a t i onu s e amp l i f i c a t i o n = Trueamp var iab i l i ty method = 2amp min factor = 0 .6amp max factor = 10000

# Bui ld ings

# Capacity Spectrum Method

# Loss

22

# Savesave hazard map = Trues a v e t o t a l f i n a n c i a l l o s s = Fal ses a v e b u i l d i n g l o s s = Fal ses a v e c o n t e n t s l o s s = Fal sesave motion = Falsesave prob s t ruc tura l damage = None

f i l e a r r a y = False

# I f t h i s f i l e i s executed the s imu la t i on w i l l s t a r t .# Delete a l l v a r i a b l e s that are not EQRM at t r i b u t e s v a r i a b l e s .i f name == main :

from eqrm code . a n a l y s i s import mainmain ( l o c a l s ( ) )

23

2.2 The Source Files

The EQRM source files for probabilistic modeling (PSHA and PSRA) come intwo forms. These are:

source zones, and faults.

The EQRM can be run with either of these inputs separately or both together

2.2.1 Source Zone File

If is defined then the file name is: zone source .xml

Otherwise the file name is: zone source.xml

The source zone file is used to describe one or more areal source zones. Earth-quakes are assumed to be equally likely to occur anywhere within a sourcezone. The magnitude recurrence relationship for each source zone is defined by abounded Gutenberg-Richter relationship. The following grey shaded box providesan example of a source zone file. A description of the parameters follows.

32.4000 151.150032.7500 152.170033.4500 151.430032.4000 151.1500

recurrence min mag = 3.3recurrence max mag = 5.4A min = 0.568b = 1>

31.0000 149.500032.4000 149.500032.4000 151.150032.7500 152.170032.7500 152.760032.7000 152.800032.0000 153.110031.0000 153.290031.0000 149.5000

35.0000 149.500032.4000 149.500032.4000 151.150033.4500 151.430032.7500 152.170032.7500 152.760034.4000 151.350034.7400 151.1500

25

35.0000 151.100035.0000 149.5000

32.9250 151.400032.7500 151.750033.2500 152.250033.5000 151.900032.9250 151.4000

31.0000 149.500032.9250 149.500032.9250 151.400032.7500 151.750033.2500 152.250033.2500 152.330032.7000 152.8000

26

32.0000 153.110031.0000 153.290031.0000 149.5000

35.0000 149.500032.9250 149.500032.9250 151.400033.5000 151.900033.2500 152.250033.2500 152.330034.4000 151.350034.7400 151.150035.0000 151.100035.0000 149.5000

General inputs (source model zone)

magnitude type: Earthquake magnitude used to derive the recurrence pa-rameters. NOTE - the EQRM only supports moment magnitude Mw.

General zone inputs (zone)

27

area: Area of the source zone in km2. This is required but currently thevalue is not used.

name: Name for the source zone. event type: Pointer to the collection of inputs described in the event typecontrol file.

Geometry inputs (geometry)

azimuth: Center azimuth for randomly generated synthetic ruptures (de-grees).

delta azimuth: Range over which randomly generated azimuths will besampled. That is, the azimuth of all synthetic earthquake will be randomlydrawn from a uniform distribution between azimuthdelta azimuth (de-grees).

dip: Center dip for randomly generated synthetic ruptures (degrees). delta dip: Range over which randomly generated dips will be sampled.That is, the dip of all synthetic earthquake will be randomly drawn from auniform distribution between dipdelta dip (degrees).

depth top seismogenic: Depth (km) to the top of the seismogenic zone inkm. No component of a synthetic rupture will be located above this value.

depth bottom seismogenic: Depth (km) to the bottom of the seismogeniczone in km. No component of a synthetic rupture will be located below thisvalue.

boundary: Boundary of the areal source zone as defined on the surface ofthe Earth in latitude (column 1) and longitude (column 2). The first andlast points must be the same to close the polygon.

excludes: Boundary of any regions in the source in which events are notrequired. Boundary defined on the surface of the Earth in latitude (column1) and longitude (column 2). The first and last points must be the same toclose the polygon.This parameter is optional. the source zone file may haveno exclude zones, a single entry or multiple entries.

Recurrence inputs (recurrence model) Note, there is only one recurrencemodel per zone.

28

distribution: Distribution used to define the magnitude recurrence rela-tions. Note that the EQRM currently only supports a Bounded Gutenberg-Richter recurrence relationship for source zones (i.e. bounded gutenberg-richter).

recurrence min mag: Minimum magnitude used to define the A min. recurrence max mag: Maximum magnitude used to define the recurrencerelationship. Typically, this is the magnitude of the largest earthquakeexpected in the zone.

A min: Expected number of earthquakes with magnitude recurrence min magor higher in the source zone per year.

b: Gutenberg-Richter b value for bounded Gutenberg-Richter recurrencerelationship

generation min mag: Minimum magnitude for synthetic earthquake gen-eration. The EQRM will only generate synthetic earthquakes with magni-tudes equal to or greater than generation min mag.

number of mag sample bins: Number of magnitude bins used to discretisethe recurrence relationship in the magnitude range generation min mag torecurrence max mag

number of events: Number of syntectic ruptures to be generated in thesource zone.

2.2.2 Source Fault File

If is defined then the file name is: fault source .xml

Otherwise the file name is: fault source.xml

The source faults file is used to describe one or more faults (including crustalfaults and subduction interfaces) and/or one or more dipping slabs for intraslabearthquakes. Earthquakes are assumed to be equally likely to occur anywherewithin the fault (or slab). The magnitude recurrence for faults can be defined by a

29

bounded Gutenberg-Richter relationship or a combination of bounded Gutenberg-Richter and Characteristic. The magnitude recurrence for the intraslab earth-quakes must be defined by a bounded Gutenberg-Richter relationship. The fol-lowing grey box provides an example of a source fault file with the followingsource types:

1. crustal fault with recurrence defined by a bounded Gutenberg-Richter rela-tionship (fault 1),

2. crustal fault with recurrence defined by a combined bounded Gutenberg-Richter (for small earthquakes) and a characteristic recurrence for largerearthquakes (fault 2),

3. a subduction interface with recurrence defined by Gutenberg-Richter (forsmall earthquakes) and/or a characteristic recurrence for larger earthquakes(fault 3),

4. a 3D dipping volume to represent intraslab earthquakes in the subductingslab (intraslab 1).

Many of the parameters in the source fault file are identical to those describedin Section 2.2.1 and are not described separately here. A description of the newparameters is provided below.

< f a u l tname = f a u l t 1event type = c r u s t a l f a u l t >

30

< f a u l tname = f a u l t 2event type = c r u s t a l f a u l t >

< f a u l tname = i n t r a s l a b 1event type = i n t r a s l a b>

31

Parameters unique to the source fault file See Chapter 4 for further detail.

dip: Dip of fault, defined as angle in degrees from horizontal. out of dip theta: Out of plane dip, used for intraslab events. Angle be-tween fault plane and out of dip rupture plane (degrees).

delta theta: Bounds the range of dips for intraslab events. That is, allsynthetic ruptures will have uniformly random sampled dips in the rangedip + out of dip theta delta theta (degrees).

slab width: Width of slab (km) when using a fault source to representintraslab earthquakes in the subducting slab.

trace: Surface trace of the fault along the surface of the Earth. Note thatit is the projection of the fault along the direction of dip. It is defined bythe latitude (lat) and longitude (lon) of the start and end of the trace.

slip rate: Slip rate of fault in mm per year. distribution: Distribution used to define the magnitude recurrence re-lations. For faults the EQRM supports (i) a Bounded Gutenberg-Richterrecurrence relationship (bounded gutenberg richter) or (ii) a combinedBounded Gutenberg-Richter and Characteristic model (characteristic).For intraslab earthquakes the EQRM supports only bounded gutenberg richter.

recurrence max mag: Maximum magnitude used to define the recurrencerelationship. Typically, this is the magnitude of the largest earthquakeexpected on the fault (or in the subducting slab).

2.3 Event Type Control File

If is defined then the file name is: event control .xml

Otherwise the file name is: event control.xml

32

The event type control file is a second level control file facilitating the variationof selected EQRM parameters with event types. The mechanism for this is anevent type parameter which links the event type control file with individualsources (i.e. specific zones, faults or dipping slabs) in the fault source and/orzone source files.

Parameters in the event type control file are separated into event groups. Theseare blocks of input parameters defined by

The parameters enclosed within

Current options for the GMPE are:"Gaull 1990 WA" Gaull et al. (1990);"Toro 1997 midcontinent" Toro et al. (1997) model for

mid-continent USA;"Atkinson Boore 97" Atkinson et al. (1997);"Sadigh 97" Sadigh et al. (1997), using Campbell (2003) convention;"Sadigh Original 97" Sadigh et al. (1997);"Youngs 97 interface" Youngs et al. (1997) interface (ZT=0);"Youngs 97 intraslab" Youngs et al. (1997) intraslab (ZT=1);"Combo Sadigh Youngs M8" combined Youngs et al. (1997) and Sadigh

et al. (1997);"Boore 08" Boore et al. (2008);"Somerville09 Yilgarn" Somerville (2009) Yilgarn Craton;"Somerville09 Non Cratonic" Somerville (2009) Average Non

Cratonic model."Allen 2012" Allens yet to be published model;"Liang 2008" Liang et al. (2008)"Atkinson06 hard bedrock" Atkinson and Boore (2006) model for hard

bedrock (Vs30=760ms1)

"Atkinson06 bc boundary bedrock" Atkinson and Boore (2006) model forVs30 at the NEHRP BC boundary

"Campbell03" Campbell (2003) hybrid empirical model"Abrahamson08" Abrahamson et al. (2008) NGA model"Chiou08" Chiou and Youngs (2008) NGA model"Campbell08" Campbell and Borzorgnia (2008) NGA model"Akkar 2010 crustal" Akkar and Bommer (2010) model for Mediterranean

and Middle East"Zhao 2006 interface" Atkinson and Boore (2003) model for

earthquakes in the subducting slab"Atkinson 2003 intraslab" Zhao et al. (2006) model for

earthquakes in the subducting slab near Japan"Atkinson 2003 interface" Atkinson and Boore (2003) model for

earthquakes on the subduction interface"Zhao 2006 intraslab" Zhao et al. (2006) model for earthquakes

on the subduction interface near Japan

2.4 Site Files

The EQRM requires a site file at which either hazard or loss will be modeled.

35

2.4.1 Hazard Site File

Filename: par site.csv

The site file for hazard is a csv file containing a header and a list of points atwhich the hazard (PSHA simulation) or ground motion (scenario simulation) willbe computed. An example is given below in the grey shaded box:

LATITUDE,LONGITUDE, SITE CLASS , VS306.4125 ,110.879166 ,D,3466.4125 ,110.887497 ,D,3506.4125 ,110.895836 ,D,3566.4125 ,110.904167 ,C,4316.4125 ,110.912498 ,C,5326.4125 ,110.92083 ,C,5146.4125 ,110.929169 ,C,4836.4125 ,110.962502 ,D,2826.4125 ,110.970833 ,D,2166.4375 ,110.904167 ,B,7606.4375 ,110.912498 ,B,7606.4375 ,110.92083 ,B,7606.4375 ,110.929169 ,B,760

Parameters in the hazard site file:

Latitude: Latitude of the points of interest. Longitude: Longitude of the points of interest. SITE CLASS: Regolith site class. Typically, this is defined by a letter. Notethat the value of this parameter must match with an amplification factordefined in the amplification file (see Section 2.5)

VS30: Average velocity in the top 30m (i.e. Vs30). This is used to incorpo-rate regolith for GMPEs with a Vs30 term.

2.4.2 Risk Site File (Building Database)

Filename: sitedb .csv

The site file for risk is a csv file representing a building portfolio. It contains alist of points at which the risk (PSHA simulation) or loss (scenario simulation)will be computed. An example is given below in the grey shaded box:

36

BID ,LATITUDE,LONGITUDE,STRUCTURE CLASSIFICATION,STRUCTURECATEGORY, . . .. . . HAZUS USAGE,SUBURB,POSTCODE,PRE1989 ,HAZUS STRUCTURE CLASSIFICATION, . . .. . . CONTENTS COST DENSITY,BUILDING COST DENSITY,FLOORAREA,SURVEY FACTOR, . . .. . . FCB USAGE, SITE CLASS ,

1 ,32.945 ,151.7513 , W1BVTILE, BUILDING, RES1 , MEREWETHER,2291 , 0 , W1, . . .. . . 3 4 4 . 4 4 5 1 , 6 8 8 . 8 9 03 , 1 5 0 , 9 . 8 , 1 1 1 , C,

2 ,32.9442 ,151.7512 , S3 , BUILDING, RES3 , MEREWETHER,2291 , 0 , S3 , . . .. . . 4 3 0 . 5 564 , 8 61 . 1 128 , 4 80 , 1 , 1 31 , C,

3 ,32.9419 ,151.7495 , W1TIMBERMETAL, BUILDING, RES1 , MEREWETHER,2291 , 0 , W1, . . .. . . 2 9 2 . 7 7 8 4 , 5 8 5 . 5 5 67 , 1 2 0 , 9 . 8 , 1 1 1 , D,

4 ,32.9414 ,151.7492 , URMLTILE, BUILDING, RES1 , MEREWETHER,2291 , 0 , URML, . . .. . . 3 7 8 . 8 8 9 7 , 7 5 7 . 7 7 9 3 , 8 0 , 9 . 8 , 1 1 1 , D,

5 ,32.9412 ,151.7486 , W1TIMBERTILE, BUILDING, RES1 , MEREWETHER,2291 , 0 , W1, . . .. . . 2 9 2 . 7 7 8 4 , 5 8 5 . 5 5 67 , 1 2 0 , 9 . 8 , 1 1 1 , C,

6 ,32.9409 ,151.7498 , URMLMETAL, BUILDING, REL1, MEREWETHER,2291 , 0 , URML, . . .. . . 9 2 5 . 6 963 , 9 25 . 6 963 , 1 50 , 1 , 4 21 , D,

7 ,32.9431 ,151.7558 , S3 , BUILDING, RES3 , MEREWETHER,2291 , 0 , S3 , . . .. . . 4 3 0 . 5 564 , 8 61 . 1 128 , 2 88 , 1 , 1 31 , D,

8 ,32.9431 ,151.7549 , W1TIMBERMETAL, BUILDING, COM8, MEREWETHER,2291 , 0 , W1, . . .. . . 1 0 87 . 1 55 , 1 087 . 1 55 , 6 00 , 1 , 4 51 , D,

9 ,32.9416 ,151.7545 , C3L , BUILDING, RES3 , MEREWETHER,2291 , 0 , C3L , . . .. . . 4 3 0 . 5 564 , 8 61 . 1 128 , 7 20 , 1 , 1 31 , E,

10 ,32.9386 ,151.7609 , C1LMEAN, BUILDING, COM1, THE JUNCTION,2291 , 1 , C1L , . . .. . . 5 4 8 . 9 594 , 5 48 . 9 594 , 4 500 , 1 , 2 11 , G,

Parameters in the building database:

BID: Integer site identifier for EQRM LATITUDE: Latitude of building LONGITUDE: Longitude of building STRUCTURE CLASSIFICATION: Expanded HAZUS building type. See Sec-tion 2.4.2 Table 2.3, Table 2.2) for a full description of the types.

STRUCTURE CATEGORY: Type of structure. Currently BUILDING only HAZUS USAGE: HAZUS usage classification (Table ??) SUBURB: within which building is located POSTCODE: Postcode within which building is located PRE1989: Logical index stating whether the building is pre- (0) or post- (1)the 1989 Newcastle earthquake

HAZUS STRUCTURE CLASSIFICATION: HAZUS building type, not expanded.See Section 2.4.2 for a full description of the types.

CONTENTS COST DENSITY: Replacement cost of contents in dollars per squaremeter

37

BUILDING COST DENSITY: Replacement cost of building in dollars per squaremeter

FLOOR AREA: Total floor area in square meters (summed over all stories) SURVEY FACTOR: Survey factor indicating how many real buildings thedatabase entry represents. Multiple buildings are represented by singlebuildings to reduce computational time.

FCB USAGE: FCB usage type (Table 2.4) SITE CLASS:Regolith site class. Typically, this is defined by a letter. Notethat the value of this parameter must match with an amplification factordefined in the amplification file (see Section 2.5)

VS30: Average velocity in the top 30m (i.e. Vs30).

Typically the building database used with the EQRM represents a subset of thetrue portfolio of interest. When creating a database that sub-samples a largerportfolio, individual database entries are used to represent more than one realbuilding. Such sub-sampling is undertaken to reduce run times and memory re-quirements. Results from an EQRM loss simulation are scaled to the full portfoliousing the survey factor. The script aggregate building db.py in eqrm codecan be used to produce a sub-sampled database.

Building construction types

Buildings have been subdivided into a number of building types each with theirown set of engineering parameters uniquely defining the median capacity curveand the random variability around the median. The building construction typesare based upon the HAZUS definitions (FEMA, 1999),with some further subdivi-sions recommended by Australian engineers for Australian building constructiontypes (Stehle et al., 2001).

In essence, the seven basic HAZUS types are

Timber frame (W) Steel frame (S) Concrete frame (C) Pre-cast concrete (PC)

38

Reinforced masonry (R) Unreinforced masonry (URM) Mobile homes (MH)

There are further subdivisions of the HAZUS types into subtypes according tonumbers of stories in the building. The complete list of HAZUS types is inTable 2.2.

The new Australian sub-types, developed by Australian engineers, create furthersubdivisions of the HAZUS types (Stehle et al., 2001). In particular, the timberframe category (W1) is subdivided into wall types (timber or brick veneer walls)and roof types (metal or tiled); the unreinforced masonry types (URML andURMM) into roof type (metal, tile or otherwise), and the concrete frame types aresubdivided into soft-story or non-soft story types. Soft-story refers to buildingsthat may have a concrete basement or parking area but wood frame stories.

In total, we currently have 56 possible construction types although some arerarely used. For example; the original HAZUS W1 is still there, however thisis rarely used in favor of the more detailed classification into W1TIMBMETAL,W1BVTILE, etc. The expanded HAZUS types is given in Table 2.3.

Building usage types

The cost models used by the EQRM require knowledge of the buildings use insociety. For example the value of a factorys contents will vary from the value ofa residents house. Similarly, the cost associated with building a hospital and thecost of building a local shop may differ even if the same materials are used becausethe buildings may be built to different standards. To transfer this information tothe EQRM the building database stores information about each buildings usage.There are two different schemes that can be used; the functional classification ofbuilding (FCB) usage (ABS, 2001) and the HAZUS usage classification (FEMA,1999).

The FCB usage is summarised in Table 2.4 and the HAZUS usage classification issummarised in Table 2.5. The EQRM control file parameter buildings usage-classification can be used to switch between the two usage classifications.

39

Table 2.2: Definitions of the basic HAZUS building construction types.

.

code description StoriesW1 timber frame < 5 000 square feet (12)W2 timber frame > 5 000 square feet (All)S1L Low-Rise (13)S1M steel moment frame Mid-Rise (47)S1H High-Rise (8+)S2L Low-Rise (13)S2M steel light frame Mid-Rise (47)S2M High-Rise (8+)S3 steel frame + cast (All)

concrete shear wallsS4L steel frame + Low-Rise (13)S4M unreinforced masonry Mid-Rise (47)S4H in-fill walls High-Rise (8+)S5L steel frame + Low-Rise (13)S5M concrete shear Mid-Rise (47)S5H walls High-Rise (8+)C1L Low-Rise (13)C1M concrete moment frame Mid-Rise (47)C1H High-Rise (8+)C2L Low-Rise (13)C2M concrete shear walls Mid-Rise (47)C2H High-Rise (8+)C3L concrete frame + Low-Rise (13)C3M unreinforced masonry Mid-Rise (47)C3H in-fill walls High-Rise (8+)PC1 pre-cast concrete tilt-up walls (All)PC2L pre-cast concrete Low-Rise (13)PC2M frames with concrete Mid-Rise (47)PC2H shear walls High-Rise (8+)RM1L reinforced masonry walls + Low-Rise (13)RM1M wood or metal diaphragms Mid-Rise (4+)RM2L reinforced masonry Low-Rise (13)RM2M walls + pre-cast Mid-Rise (47)RM2H concrete diaphragms High-Rise (8+)URML unreinforced Low-Rise (12)URMM masonry Mid-Rise (3+)MH Mobile homes (All)

40

Table 2.3: Complete list of all building construction types (with those that arerarely used in italics).The integers corresponding to each building constructiontype represent the integer index used in the building database Column 4 forexpanded HAZUS types (column 12 for HAZUS only types).

.

1: W1 15: S5H 29: RM1L 43: C1LSOFT2: W2 16: C1L 30: RM1M 44: C1LNOSOFT3: S1L 17: C1M 31: RM2L 45: C1MMEAN4: S1M 18: C1H 32: RM2M 46: C1MSOFT5: S1H 19: C2L 33: RM2H 47: C1MNOSOFT6: S2L 20: C2M 34: URML 48: C1HMEAN7: S2M 21: C2H 35: URMM 49: C1HSOFT8: S2H 22: C3L 36: MH 50: C1HNOSOFT9: S3 23: C3M 37: W1MEAN 51: URMLMEAN10: S4L 24: C3H 38: W1BVTILE 52: URMLTILE11: S4M 25: PC1 39: W1BVMETAL 53: URMLMETAL12: S4H 26: PC2L 40: W1TIMBERTILE 54: URMMMEAN13: S5L 27: PC2M 41: W1TIMBERMETAL 55: URMMTILE14: S5M 28: PC2H 42: C1MMEAN 56: URMMMETAL

Replacement costs

The replacement cost in dollars per square meter for each building and the re-placement cost of the contents of each building are contained within the buildingdatabase. Typically these costs are a function of the usage classification of thebuilding and are hence also dependent on whether the HAZUS or FCB classifi-cation system is used. The EQRM does not cross check how the costings werecreated. In some instances it may be appropriate to use costings created fromone usage classification with the EQRM using the other usage mode (effects costsplits - see below) and in some instance it may not be appropriate to do so. Usersare encouraged to familiarise themselves with database metadata to ensure thatthey are using the EQRM appropriately for their own application.

41

Table 2.4: Functional classification of building (FCB) (Australian Bureau ofStatistics, 2001) .

.

Residential: Separate, kit and transportable homes111: Separate Houses 112: Kit Houses113: Transportable/relocatable homes

Residential: Semi-detached, row or terrace houses, townhouses121: One storey 122: Two or more storeysResidential: Flats, units or apartments

131: In a one or two storey block132: In a three storey block

133: In a four or more storey block134: Attached to a house

Residential: Other residential buildings191: Residential: not otherwise classified

Commercial: Retail and wholesale trade building211: Retail and wholesale trade buildingsCommercial: Transport buildings221: Passenger transport buildings

222: Non-passenger transport buildings223: Commercial carparks

224: Transport: not otherwise classifiedCommercial: Offices

231: OfficesCommercial: Other commercial buildings

291: Commercial: not otherwise classifiedIndustrial: Factories and other secondary production buildings

311: Factories and other secondary production buildingsIndustrial: Warehouses

321: Warehouses (excluding produce storage)Industrial: Agricultural and aquacultural buildings

331: Agricultural and aquacultural buildingsIndustrial: Other industrial buildings391: Industrial: not otherwise classified

Other Non-Residential: Education buildings411: Education buildings

Other Non-Residential: Religion buildings421: Religion buildings

Other Non-Residential: Aged care buildings431: Aged care buildings

Other Non-Residential: Health facilities (not in 431)441: Hospitals

442: Health: not otherwise classifiedOther Non-Residential: Entertainment and recreation buildings

451: Entertainment and recreation buildingsOther Non-Residential: Short term accommodation buildings

461: Self contained, short term apartments462: Hotels (predominately accommodation), motels, boarding houses, hostels or lodges

463: Short Term: not otherwise classifiedOther Non-Residential: Other non-residential buildings

491: Non-residential:not otherwise classified

42

Table 2.5: HAZUS building usage classification (FEMA, 1999).

ResidentialRES1: Single family dwelling (house)

RES2: Mobile homeRES3: Multi family dwelling (apartment/condominium)

RES4: Temporary lodging (hotel/motel)RES5: Institutional dormitory (jails, group housing - military, colleges)

RES6: Nursing homeCommercial

COM1: Retail trade (store)COM2: Wholesale trade (warehouse)

COM3: Personal and repair services (service station, shop)COM4: Professional and technical services (offices)

COM5: BanksCOM6: Hospital

COM7: Medical office and clinicCOM8: Entertainment and recreation (restaurants, bars)

COM9: TheatersCOM10: Parking (garages)

IndustrialIND1: Heavy (factory)IND2: Light (factory)

IND3: Food, drugs and chemicals (factory)IND4: Metals and mineral processing (factory)

IND5: High technology (factory)IND6: Construction (office)

AgricultureAGR1: Agriculture

Religion/Non/ProfitREL1: Church and non-profit

GovernmentGOV1: General services (office)

GOV2: Emergency response (police, fire station, EOC)Education

EDU1: Grade schoolsEDU2: Colleges and Universities (not group housing)

43

2.4.3 Bridge Site File

Filename: bridgedb .csv

The site file for bridges is a csv file containing a header and a list of points atwhich the ground motion will be computed, and bridge damage estimated. Anexample is given below in the grey shaded box:

BID ,LATITUDE,LONGITUDE,STRUCTURE CLASSIFICATION,STRUCTURECATEGORY, . . .. . . SKEW,SPAN, SITE CLASS

2 ,35.352085 ,149.236994 ,HWB17,BRIDGE, . . .. . . 0 , 2 ,E

3 ,35.348677 ,149.239383 ,HWB17,BRIDGE, . . .. . . 3 2 , 3 ,F

4 ,35.336884 ,149.241625 ,HWB17,BRIDGE, . . .. . . 2 0 , 6 ,G

5 ,35.345209 ,149.205986 ,HWB22,BRIDGE, . . .. . . 4 , 2 ,D

6 ,35.340859 ,149.163037 ,HWB3,BRIDGE, . . .. . . 0 , 1 ,E

7 ,35.301472 ,149.141364 ,HWB17,BRIDGE, . . .. . . 0 , 1 , F

8 ,35.293012 ,149.126767 ,HWB10,BRIDGE, . . .. . . 1 2 , 3 ,G

9 ,35.320122 ,149.063810 ,HWB28,BRIDGE, . . .. . . 0 , 3 ,C

10 ,32.822962 ,151.685346 ,HWB17,BRIDGE, . . .. . . 0 , 4 ,E

11 ,32.823370 ,151.685797 ,HWB22,BRIDGE, . . .. . . 0 , 6 ,C

12 ,32.872624 ,151.717496 ,HWB3,BRIDGE, . . .. . . 0 , 7 , F

13 ,32.878718 ,151.733289 ,HWB10,BRIDGE, . . .. . . 0 , 1 0 ,G

14 ,32.884673 ,151.786362 ,HWB28,BRIDGE, . . .. . . 0 , 2 ,D

15 ,32.848043 ,151.696107 ,HWB10,BRIDGE, . . .. . . 0 , 3 ,C

16 ,32.753763 ,151.744744 ,HWB3,BRIDGE, . . .. . . 0 , 3 ,D

17 ,32.751578 ,151.727342 ,HWB17,BRIDGE, . . .. . . 0 , 2 ,C


Latitude: Latitude of the bridge of interest. Longitude: Longitude of the bridge of interest. STRUCTURE CLASSIFICATION: HAZUS bridge type. See Table 2.6 for a de-scription of the types.

44

STRUCTURE CATEGORY: Type of structure. Currently BRIDGE only. SKEW: Bridge skew (degrees). SPAN: Number of spans in the bridge. SITE CLASS: Regolith site class. Typically, this is defined by a letter. Notethat the value of this parameter must match with an amplification factordefined in the amplification file (see Section 2.5).

45

Table 2.6: Definitions of the HAZUS Highway Bridge Classification.

. .

code descriptionHWB1 Major Bridge - Length > 150m (Conventional Design)HWB2 Major Bridge - Length > 150m (Seismic Design)HWB3 Single Span (Not HWB1 or HWB2) (Conventional Design)HWB4 Single Span (Not HWB1 or HWB2) (Seismic Design)HWB5 Concrete, Multi-Column Bent, Simple Support (Conventional Design), Non-

California (Non-CA)HWB6 Concrete, Multi-Column Bent, Simple Support (Conventional Design), Cal-

ifornia (CA)HWB7 Concrete, Multi-Column Bent, Simple Support (Seismic Design)HWB8 Continuous Concrete, Single Column, Box Girder (Conventional Design)HWB9 Continuous Concrete, Single Column, Box Girder (Seismic Design)HWB10 Continuous Concrete, (Not HWB8 or HWB9) (Conventional Design)HWB11 Continuous Concrete, (Not HWB8 or HWB9) (Seismic Design)HWB12 Steel, Multi-Column Bent, Simple Support (Conventional Design), Non-

California (Non-CA)HWB13 Steel, Multi-Column Bent, Simple Support (Conventional Design), Califor-

nia (CA)HWB14 Steel, Multi-Column Bent, Simple Support (Seismic Design)HWB15 Continuous Steel (Conventional Design)HWB16 Continuous Steel (Seismic Design)HWB17 PS Concrete Multi-Column Bent, Simple Support - (Conventional Design),

Non-CaliforniaHWB18 PS Concrete, Multi-Column Bent, Simple Support (Conventional Design),

California (CA)HWB19 PS Concrete, Multi-Column Bent, Simple Support (Seismic Design)HWB20 PS Concrete, Single Column, Box Girder (Conventional Design)HWB21 PS Concrete, Single Column, Box Girder (Seismic Design)HWB22 Continuous Concrete, (Not HWB20/HWB21) (Conventional Design)HWB23 Continuous Concrete, (Not HWB20/HWB21) (Seismic Design)HWB24 Same definition as HWB12 except that the bridge length is less than 20

metersHWB25 Same definition as HWB13 except that the bridge length is less than 20

metersHWB26 Same definition as HWB15 except that the bridge length is less than 20

meters and Non-CAHWB27 Same definition as HWB15 except that the bridge length is less than 20

meters and in CAHWB28 All other bridges that are not classified (including wooden bridges)

46

2.4.4 Fatality Site File

Filename: popexp.csv

The site file for fatalities is a csv file containing a header and a list of points atwhich the ground motion will be computed and converted to Modified MercaliIntensity (MMI), with which the population fatalities are estimated. An exampleis given below in the grey shaded box:

LATITUDE,LONGITUDE, SITE CLASS ,VS30 ,POPULATION6.4125 ,110.837502 ,D,301 ,611936.4125 ,110.845833 ,D,273 ,1060316.4125 ,110.854164 ,D,299 ,1201686.4125 ,110.862503 ,D,318 ,275946.4125 ,110.870834 ,D,338 ,836776.4125 ,110.879166 ,D,346 ,859916.4125 ,110.887497 ,D,350 ,578446.4125 ,110.895836 ,D,356 ,1464826.4125 ,110.904167 ,C,431 ,1050036.4125 ,110.912498 ,C,532 ,379836.4125 ,110.92083 ,C,514 ,632466.4125 ,110.929169 ,C,483 ,582696.4125 ,110.962502 ,D,282 ,149866.4125 ,110.970833 ,D,216 ,45301


Latitude: Latitude of the points of interest. Longitude: Longitude of the points of interest. SITE CLASS: Regolith site class. Typically, this is defined by a letter. Notethat the value of this parameter must match with an amplification factordefined in the amplification file (see Section 2.5)

VS30: Average velocity in the top 30m (i.e. Vs30). This is used to incorpo-rate regolith for GMPEs with a Vs30 term.

POPULATION: Population of the point of interest. An estimate of fatali-ties comes from the fatality rate as per the formula from USGS Open-FileReport 2009-1136 multiplied by POPULATION.

47

2.5 Amplification File

Filename: par ampfactors.xml

Local soil conditions (or regolith) are capable of amplifying bedrock (or hardrock) ground motion. Consequently, it can be important to incorporate regolith inhazard and/or risk studies. The choice to use regolith is controlled by the EQRMcontrol file parameter use amplification. The manner in which regolith (oramplification) is considered depends on the GMPEs used. If a GMPE explicitlyincorporates regolith with a VS30 term, then the EQRM will use this. Otherwisethe RSA is computed on bedrock and then amplified to the regolith surface using atransfer function (or amplification factor). An example of an input file containingamplification factors is provided in the grey box below.

4.50000000 5.50000000

0.00000000 0.15290000 0.2548000 0.35680000

< s i t e c l a s s e s >

CDE

0.00000000 0.40000000 0.50000000 2.00000000

< s i t e c l a s s c l a s s=C>

< l o g amp l i f i c a t i o n

s i t e c l a s s = Cmoment magnitude = 4.50000000pga bin = 0.00000000>

0.13976194 0.13976194 0.25464222 0.25464222

< l o g s t d


0.01000000 0.01000000 0.01000000 0.01000000



0.09531018 0.09531018 0.23111172 0.23111172

< l o g s t d

48


0.01000000 0.01000000 0.01000000 0.01000000



0.03922071 0.03922071 0.20701417 0.20701417

< l o g s t d s i t e c l a s s = C

moment magnitude = 4.50000000pga bin = 0.2548000>

0.01000000 0.01000000 0.01000000 0.01000000

< l o g amp l i f i c a t i o ns i t e c l a s s = Cmoment magnitude = 4.50000000pga bin = 0.35680000>

0.02020270 0.02020270 0.17395331 0.17395331

< l o g s t d s i t e c l a s s = C

moment magnitude = 4.50000000pga bin = 0.35680000>

0.01000000 0.01000000 0.01000000 0.01000000

. . .

< s i t e c l a s s c l a s s=D>

. . .

< s i t e c l a s s c l a s s=E>

. . .

. . .

49

The amplification of seismic ground motion depends on the composition of theregolith. The EQRM accounts for variation in regolith material by assigningamplification factors to different site classes. The EQRM also recognises thatamplification of seismic waves is a non-linear process. That is, the degree ofamplification is a function of the level of ground motion. To account for this non-linearity, the EQRM allows users to specify a number of amplification factorswhich are grouped according to the level of bedrock ground motion (as measuredby PGA) and the size of the event (as measure by Mw).

The amplification factor file must specify the following parameters at the begin-ning:

moment magnitude bins: centroids of the moment magnitude Mw bins forwhich amplification factors are defined.

pga bins: centroids of the PGA bins for which amplification factors aredefined.

site classes: List of site classes for which amplification factors are de-fined. The EQRM assumes that each site class is defined by a single letter(e.g. site class = B).

periods: RSA periods at which the amplification factors are defined. Notethat these periods need not be the same as those in atten periods fromthe EQRM control file. The EQRM will interpolate as required.

The xml amplification factor file is then composed of a sequence of blocks, eachof which defines:

the site class using the parameter class. the moment magnitude bin centroid using the parameter mag bin. the pga bin centroid using the parameter pga bin.

Finally the inside of each block specifies:

log amplification: the logarithm of the median amplification factor de-fined at each of the RSA periods in periods.

50

log std: the standard deviation of the amplification factor. The EQRMassumes that the amplification factor is log-normally distributed when usingthis standard deviation. The standard deviation can be set to an arbitrarilysmall number such as 0.01 (as shown in grey shaded box above) when notknown. Use of this standard deviation is controlled by the EQRM controlfile parameter amp variability method which can also be set to None.

51

2.6 Vulnerability File

Filename: vulnerability.xml

Vulnerability curves in EQRM use the Natural Hazards Risk Markup Language(NRML) specified by The GEM Foundation for use in OpenQuake. An examplevulnerability file is shown in the grey box below.

5 5 .5 6 6 .5 7 7 .5 8 8 .5 9 9 .5 10 10 .5 11

< l o s sRa t i o>2 .01E07 1 .94E05 0.000587335 0.007251983 0.0442416020.155404744 0.356384523 0.594100748 0.790940996 0.9115501870.968789668 0.990643449 0.997571724

0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .30 .3 0 .3

< l o s sRa t i o>1 .01E13 3 .73E10 2 .01E07 2 .37E05 0.0008406240.011332954 0.070471569 0.237465156 0 .5 0.750428967 0.9060817680.973002925 0.993935756

0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .30 .3 0 .3

< l o s sRa t i o>8 .80E17 3 .08E12 8 .87E09 3 .57E06 0.0003011380.007243107 0.063879826 0.255612903 0.563409229 0.8243336640.952020555 0.990906067 0.998762343

0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .30 .3 0 .3

< l o s sRa t i o>1 .60E25 1 .23E17 8 .52E12 1 .46E07 0.0001329540.01170814 0.164580979 0.593551025 0.916184727 0.9930763780.999755826 0.999995957 0.999999965

0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .30 .3 0 .3

p r o b a b i l i s t i cD i s t r i b u t i o n=N>< l o s sRa t i o>1 .48E08 1 .46E06 5 .10E05 0.00079375 0.0064869510.031635882 0.101942058 0.235937951 0.420815277 0.6147465640.775778887 0.885233282 0.947862317

0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .30 .3 0 .3

< l o s sRa t i o>0.000113882 0.00240425 0.021250659 0.0966434740.265261475 0 .5 0.721302041 0.872407761 0.951287624 0.984182360.995542434 0.998889016 0.999750924

0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .30 .3 0 .3

< l o s sRa t i o>4 .98E06 7 .76E05 0.000681337 0.003805858 0.0148236190.043189199 0.099472641 0.18932183 0.308769217 0.4447669910.580456473 0.701364547 0.799114523

0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .30 .3 0 .3

< l o s sRa t i o>3 .78E07 8 .13E06 9 .54E05 0.000693954 0.0034347960.012443628 0.034924284 0.079457911 0.15205226 0.2523148330.372489214 0 .5 0.621770257

0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .30 .3 0 .3

< l o s sRa t i o>1 .38E07 9 .65E06 0.000249076 0.002942808 0.0187454410.073088279 0.193120023 0.376258115 0.580939953 0.7567288690.877406954 0.945826618 0.978749341

0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .30 .3 0 .3

The top level attributes are defined in discreteVulnerabilitySet:

EQRM supports a single loss category option, economic loss, and the two otherattributes are not currently used.

53

The curves in the set are all based on the intensity measure levels defined inIML:

5 5 .5 6 6 .5 7 7 .5 8 8 .5 9 9 .5 10 10 .5 11

This is a list of intensity measure levels specified by the intensity measure type(IMT) attribute. EQRM supports conversion from spectral acceleration to Mod-ified Mercalli Intensity, MMI, as per the example above.

The mean loss ratio and sigma are defined for each site type in discreteVulnerability:

2 .01E07 1 .94E05 0.000587335 0.007251983 0.0442416020.155404744 0.356384523 0.594100748 0.790940996 0.9115501870.968789668 0.990643449 0.997571724

0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3 0 .3

Where:

vulnerabiltyFunctionID: This specifies the site type that for this curve.It must match the STRUCTURE CLASSIFICATION defined in the sitedb file.

probabilisticDistribution: Options are LN (lognormal) and N (nor-mal). The sampling technique will depend on the EQRM input parametervulnerability variability method.

lossRatio: Mean loss values that correspond to the configured IML. coefficientsVariation: Coefficient of variation that corresponds to theconfigured lossRatio, where

CoV =

(2.1)

54

Chapter 3

Earthquake source generation

3.1 Overview

The EQRM conducts probabilistic seismic hazard analysis (PSHA) and proba-bilistic seismic risk analysis (PSRA) using an event based approach. This meansthat the ground motions (hazard) and loss (risk) are computed for each event in-dividually and the results separately aggregated to form probabilistic estimates.The event based approach differs from the traditional approach to PSHA whichintegrates over all magnitude and distance combinations to attain probabilitylevels for exceeding a particular level of ground motion in a pre-defined period oftime. The traditional approach is introduced by Cornell (1968) and summarisedby McGuire and Arabasz (1990), while the event based approach is outlined byMusson (2000). A core component of any event based analysis with the EQRMis the generation of a simulated event catalogue. The generation of the event cat-alogue relies upon an existing model for the seismicity in the region. Typicallythe model of seismicity comes from an interpretation of historical earthquakes,geology and neotectonics. The current version of the EQRM application requiresa source model that consists of a set of areal source zones (defined by a poly-gon) and/or fault sources (defined by a plane). Users can use either a boundedGutenberg-Richter or characteristic earthquake model (Schwartz and Copper-smith, 1984) to describe the earthquake recurrence relationships.

Areal source zones capture the background seismicity in an area and are decribedby the acitivty rate within each source zone through its Gutenberg-Richter bvalue and Amin (the number of earthquakes per year greater than Mmin). Inevent-based PSHA calculations, such as the EQRM, synthetic ruptures will begenerated stochastically throughout an areal source zone. However, a user may

55

have sufficient information on the geometry and earthquake recurrence rates ofa known fault and may want to place synthetic ruptures on the fault by defininga fault source. In this instance, instead of placing the synthetic ruptures withinthe areal source zone, they will be restricted to a particular fault plane.

In addition, a novel technique has been developed to generate synthetic ruptureswithin a dipping 3D volume that may represent seismotectonic features such asthe intraslab of a subducting slab, therefore allowing the EQRM to realisticallysimulate intraslab ruptures.

This chapter describes the process of creating a synthetic earthquake catalogueand for defining an individual scenario event. The chapter is organised as fol-lows. First, the concept of a synthetic earhquake catalogue is discussed (Section3.2 and 3.3), then the approach for determining earthquake magnitudes and re-currence rates is presented in Section 3.4. This is followed by a description ofthe method for generating realistic synthetic ruptures in areal source zones (Sec-tion 3.5) or on a predefined fault (Section 3.6). Section 3.7 discusses the conceptof spawned events which is used to produce spatially correlate ground motions.Lastly, Section 3.8 describes the approach for scenario event generation.

3.2 Creating an earthquake catalogue for prob-

abilistic seismic hazard analysis

This section describes the method to generate a synthetic catalogue of plausibleevents. The event catalogue can be created for events that are defined withinan areal source zone or located on a known fault. The recurrence relationshipfor the events can be defined as a bounded Gutenberg-Richter (Kramer, 1996) orcharacteristic earthquake (Schwartz and Coppersmith, 1984) relationships.

A simulated event is represented by a 2D plane (or rupture) in 3D space thatsignifies the region where slip has occurred. The parameters descibing the geom-etry of a rupture plane are shown in Figure 3.1. The important parameters of asimulated event are its location, geometry, magnitude and activity (or likelihoodof occurrence). The rupture trace is the surface projection of the simulated eventalong the direction of dip. The position and geometry of the rupture trace isdefined by its start (rlats , r

lons ) and end (r

late , r

lone ) points, its azimuth razi and its

length rl. The position and geometry of the rupture plane are defined by itswidth (rw), dip (rdip) and the position of its centre (or rupture centroid). Theruputre plane dips to the right of the fault trace when looking along the trace.The rupture centroid is defined in cartesian coordinates (rx, ry, rz) in km using

56

a local coordinate system with origin at the rupture start point (see Figure 3.1).The orientation of the local coordinate system is such that the xaxis is orientedalong the rupture trace with positive direction pointing towards the rupture endpoint, the z axis is pointing vertically downwards and the y axis is orientedalong the surface of the Earth such that the axes form a right handed triad. Thevertical projection of the rupture centroid is also described by its latitude rlatcand longitude rlonc . The ruptures are not allowed to extend above the surface ofthe Earth. The event magnitude is represented as a moment magnitude and isgenerated by the process described in Section 3.4. The activity (or likelihood ofoccurrence) of the simulated event is described by the event activity (rv) whichrepresents the number of times a given simulated event (conditional on magnitudeand position) occurs in one year (see Section 3.4).

(a) (b)

x

z

y

Start of

trace

End of

trace

Rupture width

Rupture

length

04-236-2

Dip

Rupture

plane

(r , r

, r )

x y

z

x y

z

Groundsurface

Rupture

Trace

Rupture

centroid

TrueNorth

Azimuth

x

y

Surface projectionof rupture plane

04-236-1

Ruptu

re leng

th

Ruptu

re Trac

e

Surface projectionof rupture centroid

Figure 3.1: Orientation and dimension of the rupture plane in (a) 3D and (b)2D vertical projection to ground surface.

3.3 Simulated events and virtual faults

The terms simulated event, simulated earthquake and simulated rupture are con-gruent and will be used interchangeably throughout this document. Sometimes

57

Table 3.1: Definition of parameters used in generation of the synthetic earth-quake catalogue.Parameter Definition

fA Fault areafw Fault widthfl Fault lengthf Fault azimuthf Fault dipf lons Longitude of start of fault tracef lats Latitude of start of fault tracef lone Longitude of end of fault tracef late Latitude of end of fault tracef topz The top of the seismogenic zonef botz The bottom of the seismogenic zonerA Rupture arearw Rupture widthrl Rupture lengthrlons Longitude of start of rupture tracerlats Latitude of start of rupture tracerlone Longitude of end of rupture tracerlate Latitude of end of rupture tracerx Rupture centroid x in local coordinate system

(distance in km along rupture trace from start of fault trace)ry Rupture centroid y in local coordinate system

(distance in km perpendicular from rupture trace in direction of dip)rz Rupture centroid depth (km)rlonc Longitude of rupture centroidrlatc Latitude of rupture centroidr Rupture azimuth (degrees from true North)r Rupture dip (degrees from horizontal)f topz The upper limit of the seismogenic zonef botz The lower limit of the seismogenic zone Out of dip theta of rupture plane Range of dips to be sampled either side of rupture planesw Width of slab for intraslab events. This is used to constrain the

out-of-plane rupturesAmin number of earthquake of magnitude Mmin or greater per year.rm The event magnituderv The event activity, the number of magnitude rm earthquakes

expected per year scaled to the number being simulated.r Source epsilon from spawning.we Weight derived from event spawning.

58

the adjective simulated will be omitted for brevity. An adjective actual willbe used in place of simulated to refer to a historic earthquake (i.e. one that hasactually occurred rather than one that is simulated).

A virtual fault refers to a plane in 3D space upon which an event can occur. Thevirtual fault can be located within an areal source zone or defined by a knownfault. The EQRM works by first creating a virtual fault, either within an arealsource zone, or by a user defined fault. The rupture is then placed on the virtualfault (the rupture is not allowed to exceed the bounds of the virtual fault). Theintroduction of a virtual fault is the mechanism by which the EQRM applicationconstrains the location and extent of each rupture. The depth to the top of avirtual fault is defined as the depth to the seismogenic region fz (see Table 3.1).Other geometrical parameters of virtual faults include the width fw and lengthfl.

3.4 Magnitude selection and event activity

A stratified Monte-Carlo technique is used to assign the event magnitudes. Thestratified nature of the technique ensures that the full range of magnitudes isadequately sampled. The stratified Monte-Carlo technique is illustrated in Fig-ure 3.2 for the general case. This approach is distinctly different to a bruteforce Monte-Carlo technique that would preferentially sample the more probablelower magnitude events. Such an approach would require the sampling of a largenumber of small events to ensure that a handful of large events are sampled.

The Probability Density Function (PDF) for the event magnitudes is based oneither the bounded Gutenberg-Richter model (Kramer, 1996) or characteristicearthquake model (Schwartz and Coppersmith, 1984). The activity rates of anareal source are computed using the bounded Gutenberg-Richter earthquake fre-quency model which is represented by its Amin and b values (see table 3.1). Forfault sources, the magnitude-frequency model is based on either the boundedGutenberg-Richter or characteristic Earthquake model and requires input of theslip rate of the fault by the user for the EQRM to calculate the activity rate.

The algorithm described in this section is the algorithm applied to both arealsources and fault sources. While the sampling of the PDF is identical for allmodels, there are some minor modifications where the activity rate is calculatedand these differences are discussed below.

The EQRM application simulates a number Ns of earthquake events which isunique to each source. Considering the ith source zone (areal source or fault

59

x min x max

fX(x)

(a)

x

PDF of the random variable X.

x min x max

fX(x)

(b)

Division of the PDF range

into Nb = 7 bins.

x

1 2 3 4 5 6 7

x4L

x4U

1/7

(c)

Ns samples are drawn from

the discrete uniform PDF where

each value of y corresponds to a

different bin. Note that the solid

lines indicate the bin centroids

and the dashed lines outline the

original bins.y1 2 3 4 5 6 7

fY(y)

fX(x)

(d)

A sample of X is created for each of the Y values generated in (c) by

selecting a random sample from the continuous uniform PDF spanning the

range of the bin. This is illustrated here for two separate samples selected

randomly from the 5th bin.

1

x1U

- x1L

(ii)

x5L

x5U

fX(x)(i)

5 5

x5L

x5U

x x

1

x1U

- x1L

Figure 3.2: Stratified Monte Carlo technique for sampling a PDF. In this casea normal distribution is shown but any PDF can be used.

60

source), the algorithm for choosing the magnitude of each event can be sum-marised as follows:

1. Bound the domain of the PDF with mmin and mmax (called xmin and xmaxin Figure 3.2a).

2. Separate the interval [mmin,mmax] intoNb bins, and return the bin centroids(Figure 3.2b).

3. For each of the Ns,i events randomly select a bin from a discrete uniformdistribution, where Ns,i refers to the number of events in the i

th source zone.This effectively leads to Ns,i/Nb earthquakes in each magnitude bin and en-sures that the full range of magnitudes is adequately sampled (Figure 3.2c).

4. For each of the Ns,i events randomly select a magnitude denoted rm, froma continuous uniform distribution that spans the complete range of magni-tudes in the bin. Note that Step 3 ensures that the entire range of magni-tudes is adequately sampled whereas this step ensures that all magnitudescan be attained (Figure 3.2d).

5. F

Documents

Eqrmtech Manual