CRISIS Manual PDF

Chapter 1
CRISIS USER'S MANUAL
-1-
Chapter 1
1. Introduction
1.2 About CRISIS
CRISIS has been mainly developed at Instituto de Ingeniera, UNAM, Mexico. It has been developed by:
M. Ordaz (II-UNAM, Mexico)
E. Faccioli (Politecnico di Milano, Italia)
F. Martinelli (INGV, Italia)
A. Aguilar (II-UNAM, Mexico)
J. Arboleda (II-UNAM, Mexico)
C. Meletti (INGV, Italia)
V. D'Amico (INGV, Italia)
Development of CRISIS (to obtain CRISIS 2008) was funded, between 2008 and 2010, by the Department
of Civil Protection, Government of Italy.
All rights reserved, 1987-2014
1.1 Brief description

CRISIS gives a friendly environment to calculate seismic hazard. The program computes seismic hazard
using a probabilistic model that considers the earthquake occurrence probabilities, attenuation characteristics
and geographical distribution of earthquakes. A friendly graphical interface facilitates data input.
Hazard results are mainly given, for each computation site, in terms of probabilities of exceeding a given
intensity value in different time frames. It is also possible to obtain hazard results in terms of non-
exceedance probability and equivalent exceedance rate. Details on the hazard computation algorithm can
be found following the link.
Some of the main features of CRISIS are:
1.1.1. Earthquake occurrence

Earthquake occurrence can be modeled either as a Poissonian process or as a non-Poisson process. For the
Poissonian case, CRISIS admits several types of magnitude- frequency relations: modified Gutenberg-
Richter law , Characteristic Earthquake ,and Generalized Poissonian model. For Non- Poissonian
occurrences, CRISIS can work with a generalized Non- Poissonian model with which earthquake
occurrence probabilities are explicitly given for various time frames.
In the seismicity screen there are two options to apply the modified Gutenberg-Richter law: the G-R option
and the Gridded option.
-2-
Chapter 1
1.1.2. Source geometry

Regarding their spatial distribution, sources can be modeled as areas, lines, SSG,Grid, Area Planes .
1.1.3. Attenuation models

Attenuation models (also called Ground Motion Prediction Models, or GMPM) relate, in probabilistic terms,
the earthquake characteristics (e.g., magnitude, hypocentral location) and the site location relative to the
source, with the intensities produced at the site by the earthquake.
CRISIS admits three families of GMPM: Attenuation tables furnished by the user, built- in parametric
models, and generalized attenuation models. These possibilities give CRISIS great flexibility to perform the
hazard computations.
1.1.4. Spatial integration procedure

CRISIS operates with a dynamic integration procedure which allows fast computation of hazard in
extended areas.
1.2 What is new in CRISIS (since CRISIS 2013 up to CRISIS

2015 v1.1)
l Reading source parameters from shape files
l GMPM viewer
l Sigma grids in site-effects
l New tools
l Generalized Poissonian (new seismicity model)
l Gridded seismicity (new seismicity model)
l Grid source (new geometry model)
l Area Planes (new geometry model)
l Various forms of writing results
l Event set generator
l Map file is back
l Disaggregation file (*.des) is back
l Some corrections were made to GMPM Atkinson and Boore, 2006
l New collection of GMPM
l A small correction in computation of areas in Area and AreaPlanes
l New options in Area geometries
l A new xml file that stores the same information than the *.dat file, but in a more convenient format. This
file will progressively substitute the old *.dat file

l Computations in CRISIS are now partially parallelized, so a substantial reduction in computing time
might be expected in computers that have more than one processor. In limited tests carried out in one of the
authors' machine, with 8 processors, computation times were reduced by a factor of 3.
l DEM File
l CAV filters
-3-
Chapter 1
1.2 What is new (from CRISIS2015 v1.2 on)

l Some back compatibility errors while reading source parameters from shape files have been corrected
l As suggested by Dr. Thomas Chartier, a batch computation utility has been created, which allows
running CRISIS from a command line at the OS prompt.
l Upon a suggestion of Dr. Thomas Chartier and Dr. Marco Santulin, CRISIS has now the ability to
compute *.map files that result from a logic-tree computation. Also, the computation of a logic tree gives, as
a new result, the values of the percentiles of the hazard measure chosen (probability of exceedance, non-
exceedance probability or equivalent exceedance rate) for the intensity values selected by the user. See here
some more details.
l A more precise algorithm was introduced to compute distances R and RJB for elliptical ruptures with
RUP
very large or very small aspect ratios. This change does not affect circular ruptures, for which RRUP and
RJB was already computed exactly.
l Previously, elliptical ruptures were constructed with the aspect ratio indicated by the user until the
rupture did not fit in the area source with that aspect ratio. After that, aspect ratio was changed to that of the
area source, in order to accommodate the largest possible rupture area. This has been changed for a
smoother aspect ratio assignment. Now, when the elliptical rupture does not fit with the given aspect ratio,
CRISIS has a smooth transition between the aspect ratio given by the user and the area source aspect ratio.
This slightly affects computation of RRUP and RJB for relatively large earthquakes.
l Some corrections were made to Campbell and Bozorgnia's 2014 GMPE.
l PEER benchmarks tests (Set 1 and Set 2) have been recomputed with the newer versions, with excellent
results, which are presented here.
-4-
Chapter 2
2. Building a data file

CRISIS data files are constructed via a user graphic interface that is comprised of several screens and menu
items. See below what pieces of information are given in each screen. Click on the icon for more
information.
Open a *.dat or *.xml data file
Save a data file
Give the names of the map and cities reference files
Define the computation-sites grid
Give the geometry of the sources
Give data on spectral ordinates
Give seismicity of the sources
Give data about GMPM
View GMPM
Give site-effects information
Give in this screen information to read a digital elevation model, to be used in hazard computation
Give global general parameters
Do logic-tree calculations
Validate, save, and start execution
See hazard maps
Show disaggregation chart
Give batch disaggregation parameters
CAPRA seismic scenario generation
Compute event-set for a selected computation site
Show a screen with various useful tools
-5-
Chapter 2
Show a screen to compute optimum spectra
2.2 Menus
CRISIS has the following menus:
- File
- Input
- GMPE Analyzer
- Run
- Hazard
- Tools
- Help
2.2.1. File Menu
l New. This option allows building a new data file.

l Open. This option allows building a data file (*.dat or *.xml).
l Open logic tree. This option allows opening a file (*.ltc) that contains the definition of a logic tree.
l Save As. This option allows save the data file with other name or in other path.
l Add source data from shape. This option allows reading a shape file that contains data about the
seismic sources that can be used to compute seismic hazard.
l Export source data to shape. This option allows generating a shape file that contains the existing data
in CRISIS about the geometry and seismicity of the seismic sources. It is important to mentioned that this
command generated three files, the shape file (*.shp) and the files *.dbf and *.shx.
l Exit. This option allows closing CRISIS.
2.2.1. Input Menu
l Maps(optional). This option activate a screen where it is possible choose the paths where the cities and
map files are located. The cities file contains geographic data to locate cities and data the map file contains
geographic data to draw a map. Both files are optional but both are an important help in order to introduce
data and represent results.
l Grid of sites. This option open a screen that allows you to input the grid or list of sites for which
seismic hazard will be computed. There are two options: Grid of sites or list of sites. The grid of sites can be
generated in this screen or a list of sites can be loaded from a text file.
l Source Geometry. This screen allows entering the geometry of each seismic source. Sources can be
areas, lines, points or grids.
-6-
Chapter 2
l Source Seismicity. This screen allows you to enter the information about the seismicity of each source.
The seismicity can be assigned after the geometry of the source has been defined.
l Attenuation data. This screen allows entering information about the attenuation relations to be used in
the hazard analysis.

l Spectral ordinates . This screen allows entering the parameters for each spectral ordinate (or, in
general, intensity measure) for which seismic hazard will be computed.
l Global parameters. This screen allows you to enter information concerning to: the spatial integration
procedure, the value of the time frames for which seismic hazard will be computed and the distance to be
used for M-R disaggregation.
l Set output files (optional) . This screen allows you to choose between different types of files that
CRISIS can generate with seismic hazard results.
l Site effects (optional) . This screen allows you to enter a grid of amplification factors in order to
considered site effects.
2.2.1. GMPM Viewer

This screen is designed to view, simultaneously, the predictions of one or more of the GMPM that are part
of a CRISIS projects. That is, all the GMPM that have been added with the attenuation models screen.
In the following lines we describe the different input and output values and graphs offered by this screen.
Models
In this portion of the screen a list appears of all the GMPM that are part of the current CRISIS project. The
list includes the model's ID (the number with which each GMPM is identified for the purpose of assignment
to the sources), its name, the type of distance that it uses, the truncation parameter employed, a checkbox
that indicates whether the model graphs will be drawn or not, and an indication of the color with which the
correponding grpahs will be drawn.
Data
General
Give in this tab page:

l The magnitude of the earthquake whose spectra or attenuation curve are to be drawn
l The number of spectral ordinate for which attenuation curve or magnitude scaling curve are required
l The percentile for which intensities are to be drawn
Hypocenter
Give in this tab page:
-7-
Chapter 2
l The hypocentral coordinates X (along the longitude axis, in Km) and Z (depth, in Km). The hypocenter
is located always at Y=0, while the computation site is always located at (X,Y,Z)= (0,0,0)
l Parameters K1 and K2 that allow computation of a circular rupture area as a function of magnitude
l Ux, Uy and Uz, which are the components of a unit vector that defines the orientation of the rupture
plane. In the following figure examples are presented of some plane orientations with their corresponding
normal vector, as understood by CRISIS:
l
Distances
Find in this tab page, for given hypocentral location, magnitude (and hence rupture area) and rupture plane
orientation, the different measures of distance to the computation site obtained by CRISIS. Please recall that
the computation site is always located at coordinates (X,Y,Z)=(0,0,0)
Type of graph
Choose which type of graph will be drawn:
l A spectrum, for given magnitude, hypocentral location and percentile
l An attenuation curve, for the distance metric chosen, and given magnitude, depth, rupture plane,
percentile and spectral ordinate. In order to obtain various distances, CRISIS will change the X coordinate
of the hypocentral location
l A magnitude scaling curve, for given spectral ordinate, hypocentral location and percentile
Copy button
Copy to the clipboard all the curves currently drawn. These curves can later be pasted into other
applications, like Word and Excel.
Graph
Use the checkboxes in the graph area to switch between arithmetic scale (not checked) and log scale
(checked)
-8-
Chapter 2
2.2.1. Run Menu
l Validate and Run. This command allows you to execute a run after you have finished with the input.
The execution stars if CRISIS considers that there are enough information in order to compute seismic
hazard. If CRISIS identify possible inconsistencies then a messages is showed.
2.2.1. Hazard Menu

See Hazard Maps
This command is enabled after a successful CRISIS run and allows seeing hazard maps , exceedance
probability curves and uniform hazard spectra, with several options.
Disaggregation charts
This screen allows you see disaggregation results computed on the fly
Batch disaggregation
Give in this screen parameters to compute and save, in batch mode, disaggregation results.
CAPRA scenario generation

This option allows to compute scenarios that can be used in the CAPRA system in order to compute seismic
risk.
Event-set generation
Use this screen to find the event set (that is, the set of earthquakes that will be used to compute hazard) for a
given hazard computation site.
2.2.1. Tools Menu
l GMPE branch constructor. This tool allows construction of a set of CRISIS hazard models that
collectively constitute a logic tree, in which each hazard model, represented by a *.dat file, is one of the
branches of the tree.
l Map comparer. This tool allows compare two maps of seismic hazard that were calculated with
different considerations.
l Site-effects files conversion. This tool allows converting a text file with a grid of amplification factors
into a binary file.

l Non-Poisson files conversion. Use this tool to convert generalized ASCII non-Poisson seismicity files
(nps extension) into the binary format required by CRISIS or vice-versa (from binary to ASCII).
2.2.1. Help Menu
-9-
Chapter 2
l Index. This option show you the help file.

l About CRISIS. This option show you the screen with the about data.
l Supported GMPE. This option shows you all the Ground Motion Prediction Models available in order
to choose some of these models in order to compute seismic hazard. It is also important to mention that
CRISIS allows you to include new GMPM in order to compute seismic hazard.
l Supported dimensions. This screen shows you the dimensions that are supported by CRISIS.
2.1 Open an existing input file

Button:
Menu: File - Open
This command triggers an open file dialog to open an existing data file with extension *.dat or with the new
extension *.xml
2.2 Save data to a file

Button:
Menu: File - Save As...
This command allows saving data into a data file in two possible formats:
l The usual plain-text file with extension *.dat. Care must be taken when editing this file by hand
l A new xml file that stores the same information than the *.dat file,but in a more convenient format. This
file will progressively substitute the old *.dat file
2.3 Map and cities file selection
Button:
Menu: Input - Maps
Give the name and path of the map file and the cities file. Both files are optional. By double clicking in the
text box you can choose an existing file.
The map and cities information is a helpful visual reference but has not any influence on the computations.
2.4 Grid or list of sites
Button:
Menu: Input - Grid of sites
- 10 -
Chapter 2
This screen allows you to input the grid or list of sites for which seismic hazard will be computed. There are
two options:
Grid of sites
Compute for a grid, defined by its origin, longitude and latitude increments, and number of lines in both
directions. Hazard will be computed at the nodes of this grid.
List of sites
Select this option if you want to compute hazard for a list of sites with given coordinates. Double-click in
the box to read the name of the file that contains the list of sites. The format of this text file is the following:
Number of cities
State_1, City_1, Longitude_1, Latitude_1
State_2, City_2, Longitude_2, Latitude_2
Grid reduction
It is possible to modify the basic rectangular grid by using optional polygons. Introducing one or more
boundary polygons can reduce the initial rectangular grid of points. If polygons are given, the computation
of hazard will be performed only for those points of the grid which are inside at least one of the polygons.
The polygon must be described in counter-clockwise order.
Select the Start polygon command to start drawing the polygon. Each click of the mouse defines a point of
the polygon.
Choose End polygon command to close the polygon.
Command Delete selected polygon allows you to remove the selected polygon. To see the selected
polygon, choose a polygon number and press the command draw. The polygon with the widest line is the
selected polygon.
2.5 Geometry of the sources

Button:
Menu: Input - Source Geometry
This screen allows entering the geometry of each seismic source. Sources can be areas, lines, points or grids
Name: Select the active source using this combo box

Source is alive: Select if the source is alive or not. A source that is not alive is simply ignored in the hazard
computations.
- 11 -
Chapter 2
The total number of sources and the number of the active source, as well as the number of vertex of the
active source will be shown in the corresponding labels.
2.5.1. Source operations

Add Area: Use this button to add an area source to the hazard model.
Add Line: Use this button to add a line source to the hazard model.
Add SSG: Use this button to add point sources (smoothed seismicity geometry) to the hazard model.
Add Grid: Use this button to add a grid source to the hazard model.
Delete: Delete the active source
Rename: Rename the active source
Set alive: Use this button to rapidly set alive a group of sources
2.5.2. Draw options

Choose whether the map, cities and computation sites will appear, for your reference, in the graphs. Choose
also if the triangularization of the area sources will appear in the graphs.
2.5.3.
2.5.4. Sources to draw
Active: Graphs will show only the active source
Selection: Graphs will show only the sources selected
Range: Graphs will show all sources with numbers in the range Start to End
Long/Lat plane: See the sources selected to draw in the horizontal (longitude-latitude) plane
Several planes: See the sources selected to draw in the three different planes
2.5.5.
2.5.6. Shape import/export
Export SHP: Exports fault geometry and seismicity information to a shape file
Import SHP: Imports fault geometry and seismicity information from a shape file
In addition to this general information, there are other parameters required depending on the geometry class
assigned to the source:
Parameters for area source
- 12 -
Chapter 2
Parameters for line source

Parameters for point sources (smooth seismicity geometry)
Parameters for grid sources
Parameters for AreaPlanes sources
2.6 Rupture Area/Rupture Length

CRISIS allows choosing, for each source, the parameters that relate rupture area or rupture length with
magnitude. These parameters can be either given by the user or chosen from a set of constants.
2.6.1. Area sources or Smoothed-seismicity geometries

The general relation is the following:
where A is the source area (in km2), M stands for magnitude and K1 and K 2 are constants given by the user
or chosen from a set of constants. CRISIS has the following built-in sets of constants:
Model K1 K2
Brune (1970) 0.00381 1.15130
Singh et al. (1980) 0.00564 1.15130
Wells and Coppersmith (1994) Strike-slip 0.01100 1.03616
Wells and Coppersmith (1994) Reverse 0.00571 1.12827
Wells and Coppersmith (1994) Normal 0.02072 0.94406
Wells and Coppersmith (1994) All 0.01015 1.04768
2.6.2. Line sources

In this case, the fault length L (in km) is related to magnitude through:
The corresponding built-in set of constants, taken from Wells and Coppersmith (1994) are
Model K3 K4
Surface Rupture Length (SRL), Strike-Slip 0.00028 1.70391
Surface Rupture Length (SRL), Reverse 0.00138 1.45063
Surface Rupture Length (SRL), Normal 0.00977 1.15129
Surface Rupture Length (SRL), All 0.00060 1.58878
Subsurface Rupture Length (RLD), Strike-Slip 0.00269 1.42760
Subsurface Rupture Length (RLD), Reverse 0.00380 1.33550
Subsurface Rupture Length (RLD), Normal 0.01318 1.15129
- 13 -
Chapter 2
The built-in sets of constants presented in the previous tables are given in file CrisisRPR.ini, located in the
installation directory of the CRISIS executable file. These constants can be edited, or new constants can be
manually added to this file. The general format for a new constant would be, in a single row, separated by
commas, the following:
Author , K1, K2, K3, K4
Where Author is a string used for identification purposes (normally indicating the author of the
corresponding relation) and K1, K2, K3, K4 are the given constants. For area sources, it is required that K3=
K4=0, while for line sources it is required that K1= K2=0.
2.6.3. References:
Donald L. Wells and Kevin J. Coppersmith (1994). New Empirical Relationships among Magnitude,
Rupture Length, Rupture Width, Rupture Area, and Surface Displacement. Bulletin of the Seismological
Society of America, Vol. 84, No. 4, pp. 974-1002, August 1994.
S. K. Singh, E. Bazan, and L. Esteva (1980). Expected Earthquake Magnitude from a Fault. Bulletin of the
Seismological Society of America, Vol. 70, No. 3, pp. 903-914, June 1980.
2.7 Data on spectral ordinates
Button:
Menu: Input - Spectral ordinates
This screen allows entering the parameters for each spectral ordinate (or, in general, intensity measure) for
which seismic hazard will be computed.
2.7.1. Spectral Ordinates

Total number of spectral ordinates
Is the total number of different intensity measures for which hazard is to be computed. Frequently, the
different intensity measures refer to spectral ordinates for different structural periods. In this case, spectral
attenuation relations are needed.
Actual spectral ordinate

Use this control to move from one intensity measure to the other.
Structural period of spectral ordinate

This is the value of the structural period associated to this measure of intensity. The values given in the
attenuation tables must be coherent with the period values given here.
Lower limit of intensity level

See Points defining exceedance rate curves
- 14 -
Chapter 2
Upper limit for intensity level

2.7.2. Spacing
Log: Logarithmic spacing between acceleration points. See Points defining exceedance rate curves
Linear: Constant (arithmetic)spacing between acceleration points. See Points defining exceedance rate
curves
PEER and Large PEER: Used for PEER validation tests
2.7.3. General values

Units
The units of the intensity measures (for reference only).
Number of levels of intensity for which seismic hazard will be computed.

2.8 Points defining exceedance probability curves

Lower limit, upper limit and number of levels for hazard computation
Exceedance probabilities will be computed for the number of levels selected and between the lower and
upper limit given by the user, with either logarithmic or linear (arithmetic) spacing. For instance, if 10 levels
of intensity (PGA in this case) are chosen between, say, 1 gal and 300 gal, with logarithmic spacing, the
exceedance probabilities will be given for the accelerations shown in the following figure:
If linear (arithmetic) spacing is selected, then the distance between acceleration values will be constant.
- 15 -
Chapter 2
There is always a compromise between speed and precision: the larger the number of points to define the
curve or the larger the intensity range, the slower the computation time. Usually, not more than 20 points are
required to accurately define the exceedance probability curves.
2.9 Seismicity
Button:
Menu: Input - Source seismicity
This screen allows you to enter the information about the seismicity of each source. See Seismicity models
used for more details.
2.9.1. For the Gutenberg-Richter model

GR Model
Threshold magnitude(M0 ). The catalog of earthquakes is assumed to be complete for M>M 0. Earthquakes
with M<M0 are absolutely ignored.
Lambda(M0): Exceedance rate of magnitude M0. The units are earthquakes/year.
Expected value of beta. Expectation of the "b-value" for the source, given in terms of the natural logarithm.
Coefficient of variation of beta. Coefficient of variation of the "b-value" for the source, given in terms of the
natural logarithm.
Parameters defining Mu
Expected value. Expected value of the maximum magnitude for the source. See details
Uncertainty range (+/-). A number indicating that the maximum magnitude will have a uniform probability
density function, centered at its expected value, plus minus this number. See details
2.9.2. For the characteristic model

Median value of the times between characteristic earthquakes with M>M 0 . This is the inverse of the
exceedance rate for M>M0.
Standard deviation of the magnitude of the characteristic earthquake. It is assumed independent of time.
Minimum possible magnitude of a characteristic earthquake. Earthquakes with M<M 0 are absolutely
ignored
Maximum magnitude of the characteristic earthquake to be used in the integration process.
- 16 -
Chapter 2
Parameters D and F define the expected magnitude as a function of time, as in the slip-predictable model. It
is assumed that
E(M|T00)=max(M0,D+F*LN(T00))
where T00 is the time elapsed since the last characteristic event. Of course, if F is set to zero, then D
becomes the expected time-independent magnitude of the characteristic earthquake.
2.9.3. For the Non-Poisson model

Give, by double-clicking, the name of the nps file that stores the non-Poissonian earthquake occurrence
probabilities associated to this source.
2.9.4. For the Gridded Seismicity model

Give the threshold magnitude, M0, that will be the same for all grid points.
Give, by clicking in the appropriate button (...):
l The name of the file that stores the grid of values values of l0 (M0 ), that is, the exceedance rate of the
threshold magnitude. In the case of grids, the value given in each node is the number of events/year with
M>M 0 that take place in the small square represented by the node, that is, in a square with width Dx and
height Dy, centered at the node. This means that, in grids, the total number of events/year with M>M0 is the
sum of the values given at each node. This grid must be given in Surfer 6 ASCII or Binary formats.
l The name of the file that stores the grid of values of E( b), that is, the expected value of coefficient b in
the Modified Gutenberg-Richter model. This grid must be given in Surfer 6 ASCII or Binary formats.
l The name of the file that stores the grid of values of E(M ), that is, the expected value of the maximum
u
magnitude in the Modified Gutenberg-Richter model. This grid must be given in Surfer 6 ASCII or Binary
formats.
2.9.5.
2.9.6. Moment rate and slip rate
This screen shows the values of moment rate and - in some cases- slip rate associated to the
geometry/seismicity combination of parameters given. See details of the moment and slip rate computations.
2.10 Attenuation data
Button:
Menu: Input - Attenuation data
- 17 -
Chapter 2
This screen allows entering information about the attenuation relations to be used in the hazard analysis. In
general, an attenuation relation describes the probabilistic link between earthquake magnitudes, source to
site distance, and intensity (see Probabilistic interpretation of attenuation relations).
In general, CRISIS must know what relation to use to attenuate earthquakes generated in each source. In
principle, each source could have its associated attenuation relation. In practice, only a few different
attenuation relations are used in a particular analysis (e.g., one for subduction events and another for
shallow crustal earthquakes).
CRISIS can perform a simultaneous hazard analysis for several intensity measures (e.g., PGA and spectral
accelerations for different periods). Therefore, CRISIS must also know for how many different intensity
measures the analysis will be carried out, and the associated attenuation relations. Frequently, the different
intensity measures are spectral response values for different periods, so a uniform hazard spectrum can be
constructed.
In view of this, the general operations that have to be performed in this screen are:
l Select and add the attenuation models to be used in the analysis.
l Assign one of these models to each source.
2.10.1. Selection of attenuation models

CRISIS admits three families of attenuation models:
Attenuation tables. In these tables, relations between earthquake characteristics and intensities at a site are
given in terms of the following parameters: magnitude, structural period, source-site distance and depth. For
the first moment (usually the median of a lognormal distribution), the attenuation relations are matrices in
which the rows run for the magnitude and the columns run for the distance. Note that when using
attenuation tables, the relations between magnitude, distance and intensity do not need to be of parametric
nature, since the intensity medians are given, point by point, for magnitude-distance combinations.
Built-In models. These are popular models, published in the literature, in which magnitude, distance and
intensity are probabilistically related by, usually, a set of formulas or parametric equations. There is a set of
built-in models ready to use in CRISIS and there is also the possibility of adding new models.
Generalized models. Generalized attenuation models are non-parametric probabilistic descriptions of the
ground motions produced by an earthquake. In the context of CRISIS, a generalized attenuation model is a
collection of probabilistic footprints, one for each of the events considered in the analysis. Each footprint
gives, in probabilistic terms, the geographical distribution of the intensities produced by this event.
Use the following buttons to create and edit the collection of attenuation models:
Add Model
Edit model
Delete model
- 18 -
Chapter 2
2.10.2. Assignment of attenuation models

Once one or several attenuation models have been added, the user must assign one attenuation model to
each source.
See special attenuation models for a more complex use of attenuation models.
2.11 Site effects

Button:
Menu: Input - Site effects
Add or delete site effects grids to be used in the hazard analysis. See the format of the site effects grids in
this link.
2.12 Global parameters

Button:
Menu: Input - Global parameters
This screen allows you to enter information concerning:
The spatial integration procedure
The value of the time frames for which seismic hazard will be computed
The distance to be used for M-R disaggregation
2.12.1. Integration parameters

Parameter controlling the integration process
All sources or sub-sources farther away than this number (in Km) will be ignored in the spatial integration
process.
Minimum triangle size

Sources will be subdivided into sub-sources whose characteristic size will not be less than this number. For
area sources, the characteristic size is the square root of its area. For a fault source, the characteristic size is
its length.
Minimum (Distance/Triangle size) ratio

Sources will be subdivided until the ratio between source-site distance and characteristic size of the sub-
source is larger than this number.
- 19 -
Chapter 2
2.12.2. Time frames for which hazard will be computed

Give in this table the values of the time frames for which hazard will be computed. These values must be
coherent with those given in the non-Poissonian seismicity files (nps) associated to the sources.
2.12.3. Return periods for map file generation

There is again the possibility of generating a *.map file. This file gives, for each computation site, the
intensities associated to 5 different fixed equivalent return periods. For a non- Poisson process, the
equivalent return period is the one that, in a truly Poisson process, would yield a given exceedance
probability in a given time frame. For instance, an equivalent return period of 100 years corresponds, in a
non-Poisson process, to an exceedance probability of 0.00995 in one year and 0.632 in 100 years. The
equivalent return periods to be used in the map file are given in this data grid. See here an example of the
map file.
2.12.4. CAV filters

Select in this Combo Box the type of CAV filter (if any) that CRISIS will use in the computation
2.13 Logic tree computation

Button:
This screen allows definition of a logic tree , a tool that is frequently used to account for epistemic
uncertainty in the computation of seismic hazard.
In the context of CRISIS, each branch of a logic tree is formed by one data file (usually with extension
*.dat) along with a measure of the degree of belief that the analyst has on each of the branches being the
"true" one. Therefore, this screen allows construction of the logic tree by way of informing CRISIS which
*.dat files form the branches of the tree and what weight is assigned to each of the branches. The functions
provided for this aim are the following:
2.13.1. Input
Use this option to create a new logic tree.
Use this option to open a previously created logic tree. Logic trees are defined in text files, usually with
extension *.ltc (logic.tree combination) that contains, for each branch, the name (path included) of the
CRISIS input data file associated to this branch, as well as the weight assigned to each branch, in the form
of a numerical integer. This weight, normalized by the sum of the weights of all branches, is interpreted as
the probability of being the "true" one.
The format of the text file is the following:
FileName1.dat, Weight 1
FileName2.dat, Weight 2
- 20 -
Chapter 2
...
FileNameN.dat, Weight N
Note that the file name and its associated probability must be separated by a comma.
Use this option to save a logic tree, usually in a text file with extension *.ltc
Add a new branch to the logic tree
Delete the selected branch of the logic tree
Change the weight of the selected branch
Perform the logic-tree combination. Before proceeding to do the logic-tree computations, CRISIS will
perform the following checks:
1) That all *.dat files exist and contain data of a valid hazard model.
2) That there is coherency among the various *.dat files.
Also, CRISIS will only recompute the branches whose associated *.dat files have changed since the last
execution. In other words, CRISIS will not recompute branches that have already been computed, unless
the "Force branch recomputation" option is selected.
2.13.2. Output
Results of the logic-tree combination will be given in the form of a new hazard model, with an associated
*.dat file that will have the base name of the *.ltc file that described the combination, but with the extension
*.dat. This new hazard model can be loaded into CRISIS and the corresponding hazard results can be
analyzed with CRISIS (hazard maps, exceedance probability curves, uniform hazard spectra) as if they
were the results of a regular *.dat file.
In terms of text files, the main results of the logic-tree combination are given in the *.gra file, which has the
structure of a regular *.gra file, but reports the average hazard, that is, the hazard obtained from the logic-
tree combination. Also, while CRISIS generates a *.gra file with the average (or mean) hazard, it also
generates a file with extension *.pct that contains some percentiles of the hazard measure associated to a
given intensity measure value. See here an example of *.pct file.
CRISIS computes hazard values for the following percentiles: 0, 5, 10, 25, 50, 75, 90, 95 and 100. By
definition, the hazard associated to percentile 0 is reported as 0, and the hazard reported for percentile 100 is
the maximum hazard value obtained, for a given intensity value, in all the logic-tree branches.
The computation of a logic tree will also produce a *.map file constructed with the hazard resulting from the
logic-tree combination.
Disaggregation results, however, can not be obtained for the hazard resulting from the logic-tree
combination.
- 21 -
Chapter 2
2.14 Validate data, save and start execution

Button:
Menu: Run - Validate and Run
This command allows you to execute a run after you have finished with the input.
If data are still required when this command is executed, CRISIS will issue a message showing the data
required. Enter the data needed, save the data file, and choose Run to initiate execution.
Also, in some cases CRISIS will issue a set of warnings, pointing to possible inconsistencies in the input
data. While these warning do not prevent the user from computing hazard, CRISIS issues them so the user
is aware of potential problems.
After a successful run, a success screen appears showing which output files were generated and where they
are located.
2.15 Hazard maps

This command is enabled after a successful CRISIS run and allows seeing hazard maps, exceedance
probability curves and uniform hazard spectra, with the following options:
2.15.1.
2.15.2. Types of hazard maps
Two types of hazard maps can be generated:
1. If switch is selected, the map will show intensities associated to a fixed exceedance probability in a
given time frame. Give the required exceedance probability and time frame the corresponding boxes.
2. If switch is selected, the map will show exceedance probabilities associated to fixed values of time
frame and intensity. Give the required values of time frame and intensity in the corresponding boxes.
In both cases, the map will be generated for the intensity measure chosen in the box Intensity.
2.15.3.
2.15.4. Intensity measure
Select in this combo box the intensity measure for which maps and hazard curves are to
be generated.
2.15.5.
2.15.6. Time frame
Select in this combo box the time frame for which maps and hazard curves are to
be generated
- 22 -
Chapter 2
2.15.7.
2.15.8. City Selection
Select a city of the list in order to see:

1. The hazard curve at the city for the selected intensity measure and time frame.
2. Depending on whether the fixed intensity or the fixed exceedance probability switch is selected:
l The uniform hazard spectrum
l A graph showing the exceedance probabilities associated to a fixed value of spectral intensity and time
frame
2.15.9.
2.15.10. Exceedance probability
Give the exceedance probability that will be used to draw a hazard map or a
uniform hazard spectrum.
2.15.11.
2.15.12. Intensity value
Give the fixed intensity value that will be used to draw a map with show
exceedance probabilities associated to fixed values of time frame and intensity or to draw a spectrum
showing the exceedance probabilities associated to a fixed value of spectral intensity and time frame, as a
function of period.
2.15.13.
2.15.14. Save hazard maps
Give the format for the maps and their characterisitics, as well as the name of the directory in which
hazard maps will be saved.
2.15.15.
2.15.16. Draw hazard map
Draws the hazard map with the selected options
2.15.17.
2.15.18. Site Selection
Click into a point of the hazard map in order to see:
1. The hazard curve for the selected intensity measure and time frame
2. Depending on whether the fixed intensity or the fixed probability switch is selected:
- 23 -
Chapter 2
l The uniform hazard spectrum

l A graph showing the exceedance probabilities associated to fixed values of time frame and spectral
intensity
2.15.19.
2.15.20. Zoom Tools
Use these tools to zoom-in and zoom-out from the map, or to define a rectangular area to
zoom-in.
2.15.21.
2.15.22. Draw options
Check in this frame the drawing items that will appear in the hazard map.
2.15.23.
2.15.24. Color scale
- 24 -
Chapter 2
Select in this section whether or not you want CRISIS to auto-scale the colors associated to the map. In case
that auto-scale is not chosen, the user must give upper and lower values for the scaling process. Also,
moving the mouse along the color scale will indicate the numerical values associated to a particular color.
2.16 Magnitude-Distance disaggregation results

CRISIS can generate exceedance rates disaggregated by magnitude, distance and "epsilon" value. The
program presents these values graphically, in the disaggregation screen.
To see disaggregation results in this screen the following operations are needed:
Select the site for which results are desired by pointing to it with the mouse in the right-hand side picture
box, which shows either the computation grid or the list of sites. The actual point used for the
disaggregation computations will be the one belonging to the grid or list of computation sites that is the
closest to the point clicked by the user. The actual point used for disaggregation will be indicated in the
"Site Location" frame:
Choose the desired intensity measure (usually, a spectral ordinate associated to a structural period), time
frame and epsilon value in the "Options" frame:
Choose, in the "Intensity/Return period" frame, the value of intensity for which disaggregation results will
be presented or choose the desired exceedance probability (CRISIS will compute exceedance probability if
intensity is given or intensity if exceedance probability is given):
Use the "Grid options" frame to define the size of the disaggregation chart, giving the limits for magnitude
and distance, as well as the number of cells in each direction:
- 25 -
Chapter 2
In general, disaggregation charts will be redrawn every time a parameter change is made.
Results will be shown in a disaggregation chart like the following:
The value in each cell is the probability that the selected intensity level is exceeded in a given time frame if
only earthquakes with magnitudes and distance within the cell's range are accounted for. The color scale
will adjust automatically if "Autoscale" is selected. The user, however, can change the upper (red color) and
lower (white color) limits of the scale, once the "Autoscale" option is disactivated.
On top of the disaggregation chart, CRISIS shows the following legend:
"Total probability in chart: 0.000E+00 (100.00% of total)"
This legend indicates that, with the current grid settings (magnitude and distance limits) and the selected
"epsilon" level, the total probability of exceedance is a certain percentage of the total exceedance probability
(for all magnitudes and distances, and epsilon equal to minus infinity).
However, the total probability is computed by interpolation of a previously computed hazard curve for the
site. If computation of this hazard curve was made for a small number of intensity levels, the interpolation
will not be exact, and percentages reported by the legend could be somehow inexact. To solve this problem,
simply compute the hazard curves with a larger number of intensity levels.
- 26 -
Chapter 2
When seeing CRISIS disaggregation results, the user must not expect that the exceedance probabilities
associated to each cell used for the disaggregation add up to the total exceedance probability computed for
the same site, intensity value and time frame. As shown in this link, arithmetic of exceedance probabilities is
more complex to that of intensity exceedance rates used in conventional hazard studies.
Disaggregation charts can be saved using button Save, which will save in a text file the currently
displayed chart settings, as well as the matrix of disaggregated hazard values.
2.17 Help
Opens the help file of CRISIS
2.17 Output Files

Upon the users selection, CRISIS can generate several output files. The possible output files are:
2.17.1. Results file (*.res)

This file contains a printout of the name of the run, the values assigned to the variables, characteristics of the
attenuation models, geometrical and seismicity description of the sources, the data defining the computation
grid, etc. It also gives a summary of the computations for each site, indicating which sources are of interest
the site and which sources were skipped. The computer times are also written.
If the users chooses to do so, this file also gives the final results, that is, exceedance probabilities for each
time frame, site and type of intensity. See an example of the *.res file
2.17.2. Graphics file (*.gra)

This file contains a brief identification header, and the exceedance rates for the types and levels of intensity
requested. This file can be used as input file to plot intensity versus exceedance rate curves. See an example
of the *.gra file.
2.17.3. Source by source results (*.fue)

This file contains exceedance probabilities by source, for each site and time frame. See an example of the
*.fue file.
Additionally, CRISIS will generate binary files (one for each intensity measure used in the analysis) to be
able to generate its own maps.
2.17.4. Map file (*.map)

This file contains intensity levels for the fixed return periodsentered in the data file, and for each type of
intensity, time frameand site. It also gives the coordinates of each site. This file can be used to generate
contour or 3D maps of intensity levels associated to constant exceedance rates. See an example of the *.map
file.
- 27 -
Chapter 2
2.17.5. M-R Disaggregation file (*.des)

This file contains results of seismic hazard disaggregation, as a function of magnitude and distance, for
given intensity levels, return periods, time frames and epsilon values. These disaggregated results indicate
which combinations of magnitude and distance contribute more to the seismic hazard at a site, for a given
intensity measure, time frame, and return period. See an example of the *.des file. The combinations of
parameters for whichdisaggregation is desired are given in the Batch Disaggregation screen.
2.17.6. Maximum earthquakes file (*.smx)

This file contains information about the maximum possible intensity values at each site. For a given site,
these values are computed using the worst combination of distance to a source and expected value of
maximum magnitude. The highest intensity computed for all sources is reported in this file, for different
values of epsilon. See details and an example of the *.smx file.
2.17 Event set generator

Use this screen to find the event set (that is, the set of earthquakes that will be used to compute hazard) for a
given hazard computation site. In order to calculate the event set, simply click at the desired computation
site and see how the epicenters that comprise the event set are drawn, with different colors for different
seismic regions.
The computed event set can be saved in a text file that contains the following information for each event:
Value Comments
Kx
Ky Hypocentral coordinates, in Km, measured with respect to the reference point (usually the first vertex) of
Kz
X
Epicentral location, in geographical coordinates (long, lat)
Y
Z Hypocentral depth
Distance from the hypocenter to the computation site. The type of distance is that used by the general a
R
corresponding seismic region
M Earthquake magnitude
MED(Sa) Median value of the acceleration, for the first structural period, caused by this earthquake at
SD(lnSa) Standard deviation of the logarithm of acceleration, for the first structural period, caused by this earth
NR Index of the seismic region to which this earthquake belongs
ZToR Depth (in km) to the shallowest point of the earthquake's rupture surface
Shortest horizontal distance from the site to a line defined by extending the fault trace (or the top edge o
Rx
directions. Values on the hanging-wall are positive and those on the foot-wall are
- 28 -
Chapter 2
2. General Overview
It is possible to obtain a general overview about CRISIS reading the Basic Theoretical Background
2.17 Basic theoretical background

The purpose of this section is not to describe in detail the techniques to compute seismic hazard; for that
purpose, an excellent and updated summary can be found in Manuela Villani's Ph.D. Thesis. However, we
will rapidly describe some of their main aspects.
Starting with CRISIS 2008, the code does not work anymore with intensity exceedance rates as measures of
seismic hazard. The more recent versions give seismic hazard in terms of probabilities of exceedance of
intensity values in given time frames. For instance, a valid measure of seismic hazard in the newer versions
is the probability of experiencing peak ground acceleration greater or equal than 0.2 g in the next 50 years at
a given location.
This change was made in order to allow users to introduce in the computations probabilities of earthquake
occurrences derived from non- Poissonian models. Poissonian computations, however, are still possible
since one can regard this case as a particular case of the non-Poisson computations. We will se later how to
compute the probabilities now required by CRISIS from conventional Poissonian models.
In order to compute seismic hazard, the territory under study is first divided into seismic sources according
to geotectonic considerations; in most cases, it is assumed that, within a seismic source, an independent
earthquake-occurrence process is taking place. For each seismic source, earthquake occurrence probabilities
are estimated by means of statistical analysis of earthquake catalogs.
In the more general case, earthquake occurrence probabilities must stipulate the probability of havingsevents
(s=0, 1, , Ns) of magnitude M i in the following T j years at a given source k. We will denote these
probabilities as Pk(s,Mi,Tj); they completely characterize the seismicity of source k.
Seismic hazard produced by an earthquake of magnitude Mi at a single point source, say the k-th, and for
the next Tj years, can be computed as:
(1)
where Pr(Aa | Mi, Rk) is the probability that intensity a is exceeded given that an earthquake of magnitude
Mi took place at source k, that is separated from the site of interest by a distance Rk . Please note that this
probability depends only on magnitude and source-to-site distance, and it is normally computed giving a
probabilistic interpretation to intensities predicted by ground motion prediction models or attenuation
relations. We also note that implicit in equation 1 is the assumption that exceedances of intensity values at
source k given that an earthquake of magnitude Mi occurred are independent from each other. This is the
reason why the non-exceedance probability of a given that s events of magnitude Mi took place at source k
can be computed as [1 - Pr(Aa | Mi, Rk)]s.
Seismic hazard contained in equation 1 is more easily expressed in terms of non-exceedance probabilities:
- 29 -
Chapter 2
(2)
Equation 2 gives the non- exceedance probability of intensity value a given that only earthquakes of
magnitudeM itook place. The non-exceedance probability of a associated to the occurrence of earthquakes
of all magnitudes at source k in the next Tj years can be computed as:
(3)
where Nm is the number of magnitude bins into which the earthquake occurrence process has been
discretized. Again, we have used the independence hypothesis among earthquakes of all magnitudes.
But seismic sources are usually points, lines, areas or volumes, so a spatial integration process must be
carried out to account for all possible focal locations. We will assume that the spatial integration process
leads to N sources. So finally, assuming that earthquake occurrences at different sources are independent
from each other, we obtain that the non-exceedance probability of intensity a in the next T j years due to
earthquakes of all magnitudes located at all sources, can be computed with
(4)
(5)
(6)
Finally,
(7)
Equation 7 is the one used by CRISIS to compute seismic hazard for situations in which the sources are
spatially distributed (k =1,,N), there are earthquakes of various magnitudes (M i , i=1,Nm ) and the
earthquake occurrence probabilities in known time frames Tj at source k are given by Pk(s,Mi,Tj), that is, the
probability of having s events of magnitude Mi in the next Tj years at source k.
- 30 -
Chapter 2
The equations presented here are, in general, applicable to non-Poisson occurrence process. But they are
also applicable to the Poisson process. Let us see what results we obtain if we assume that the occurrence
process is Poissonian.
Let us assume that at all sources a Poisson occurrence process is taking place for earthquakes of all
magnitudes. Under this assumption, Pk(s,Mi,Tj) takes the form of, precisely, a Poisson probability
distribution:
(8)
where Dlk(M i) is the number of earthquakes of magnitude Mi that, per unit time, take place at source k. In
other words, this quantity is the conventional exceedance rate of earthquakes in the range of magnitudes
represented by Mi, that is,
(9)
Replacing equation 8 in equation 2 we obtain:
(10)
Note that now the sum extends to infinity since, in the Poisson process, the possible range of values of s is 0
to infinity. The sum in equation 10 has analytical solution:
(11)
(12)
Hence, from equation 7 we get that
(13)
(14)
- 31 -
Chapter 2
But, under the Poisson assumption for the earthquake occurrences, the process of intensity exceedances is
also a Poisson process, for which the exceedance probability of intensity a during the next Tj years would
be given by:
(15)
where n(a) is the exceedance rate of intensity a. Comparing equations 14 and 15 we obtain that
(16)
Note that n(a), the well-known Poissonian intensity exceedance rate, does not depend anymore on Tj. In the
limit, the inner sum of equation 16 can readily be recognized as the integral with respect to magnitude that is
present in the conventional Esteva-Cornell approach to compute Poissonian seismic hazard. The outer sum
in equation 16 is simply the aggregation of intensity exceedance rates due to all sources. In other words,
(17)
(18)
Note that, due to the definition we used for Dlk(Mi ) in equation 9, its sign changed when we converted it to
its differential form. We have then shown that equation 7, derived for the general non-Poissonian case, is
also valid for the Poissonian case, leading to the well-known Esteva-Cornell expression to compute seismic
hazard.
2.17 Spatial integration procedure

CRISIS assumes that, within a source, seismicity is evenly distributed by unit area (area sources) or by unit
length (line sources). For point sources, of course, all seismicity is assumed to be concentrated at the points.
In order to correctly account for this modeling assumption, CRISIS performs a spatial integration by
subdividing the original sources. Once subdivided into sub-sources, CRISIS assigns to a single point all the
seismicity associated to a sub-source, and then the spatial integration adopts a summation form.
The subdivision procedure will be briefly described in the following paragraphs.
2.17.1. Area sources

These sources are originally given by the user as 3D polygons; the user gives the coordinates (longitude,
latitude and depth) of the N vertex defining the area source.
- 32 -
Chapter 2
First, the area source is subdivided into N-2 triangles. These triangles will be further subdivided until one of
the following two conditions are met:
1) The size of the triangle is smaller than the value Minimum triangle size given by the user. That is, the
triangle is subdivided if it is still big.
2) The ratio between the site-to-source distance and the triangle size is larger than the value Minimum
Distance/Triangle Size ratio given by the user. In other words, the triangle is subdivided if the site is still
not far enough.
The sub-sub divisions are performed by means of a recursive function.
The site-to-source distance is measured from the computation site to the centroid of the triangle whose
possible subdivision is being examined. The size of the triangle is simply the square root of its area. The
seismicity associated to each centroid is proportional to the triangles area.
If CRISIS decides that a triangle has to be subdivided, this is done dividing the triangle into four new ones,
whose vertexes are the mid-points of the three sides of the original triangle.
CRISIS uses the following as default parameters: Minimum triangle size = 11 Km; Minimum
Distance/Triangle Size ratio = 3. The following graph shows the resulting subdivision of a squared source
of size 1x1 when the computation site is located at the center of the source, using the default integration
parameters.
Figura 2.1. Source subdivision with Minimum triangle size = 11 Km, Minimum Distance/Triangle Size ratio = 3
Figure 2 shows the same subdivision but with Minimum triangle size = 5 Km, Minimum Distance/Triangle
Size ratio = 3. Note how this subdivision yields smaller triangles in the neighborhood of the computation
site.
- 33 -
Chapter 2
Figure 3 shows the same case but with Minimum triangle size = 5 Km, Minimum Distance/Triangle Size
ratio = 4. Note that the smaller triangles cover now a wider area around the computation site.
Finally, Figure 4 shows the resulting subdivision with Minimum triangle size =0.5 Km and Minimum
Distance/Triangle Size ratio = 4. Note how the density of triangles varies radially as we get away from the
computation site.
- 34 -
Chapter 2
Figura 2.4. Source subdivision with Minimum triangle size =0.5 Km, Minimum Distance/Triangle Size ratio = 4
2.17.2. Line sources

In this case, the subdivision is performed by bi-partition of a fault source segment, again until one of the
following criteria are met:
1) The size of the line is smaller than the value Minimum triangle size given by the user.
2) The ratio between the site- to- source distance and the line size is larger than the value Minimum
Distance/Triangle Size ratio given by the user.
The site-to-source distance is measured from the computation site to the midpoint of the line whose possible
subdivision is being examined. The size of the line is simply its length. The seismicity associated to each
centroid is proportional to the lines length.
- 35 -
Chapter 2
2. GeoSeismAtt Combinations
The different geometry / seismicity /attenuation models give raise to the combinations listed in the following
tables. The tables give a letter or a number that points to a note that gives details on the combinations. In the
tables, "Normal" attenuation refers to attenuation tables or user coded built-in models, while "General"
refers to generalized attenuation models.
The color codes indicate the following:
Green: Combination always valid, regardless of parameter values
Yellow: Combination valid or not depending on parameter values
Red: Combination never valid
Blue: Combination potentially valid, but not implemented yet
2.17.1. Normal attenuation

Modified Characteristic Generalized Generalized Gridded
Source/Seismicity
GR earthquake NP Poisson seismicity
Area A A B B E
Line A A B B E
SSG C C D D E
AreaPlanes A A B B E
Grid E E E E E
Note A
These are old CRISIS options, which are valid always.
Note B
In this option a source is geometrically modeled as a line or as an area, which means that every point that
belongs to the source has the same probability of being a hypocenter; this is the usual assumption when
using line or area sources in CRISIS. Attenuation, as in the older CRISIS option is modeled with a
parametric description (a "normal" GMPM). However, the new option permits stipulation of earthquake
occurrence probabilities with a generalized non-Poissonian model or a generalized Poisson model, and not
through a parametric frequency-magnitude relation (Gutenberg-Richter or Characteristic Earthquake).
The occurrence probabilities given in the non-Poissonian or Poissonian seismicity files correspond to the
whole source, that is, they are the probabilities of having and earthquake of given magnitude and in a given
time frame anywhere in the source. Using its standard spatial integration scheme, CRISIS will sample the
source in order to compute hazard, accounting for all possible locations of the earthquake within the source.
Note, however, that when probabilities are specified for the whole source, probabilities associated to
- 36 -
Chapter 2
segments of the source, or sub-sources, are not univocally defined. The following approach is adopted by
CRISIS in order to define the occurrence probabilities associated to sub-sources of known sizes.
Assume first that we have a conventional Poissonian source. The probability of having i events of
magnitude M in the next Tf years, due to the effect of the whole fault, P(i,M,Tf) would be given by:
P(i,M,Tf)=exp(-Dl(M)Tf) (1)
where Dl(M) is the Poissonian magnitude occurrence rate of earthquakes with magnitudes in the vicinity of
M, for the whole source. This occurrence rate can be written as:
Dl(M)=-ln[P(i,M,Tf)]/Tf (2)
In the case of Poissonian occurrences, occurrence rates are additive. Thus, the occurrence rate
corresponding to a sub-source of relative size wj, would simply be:
Dl (M)=Dl(M)wj (3)
j
Note that, for all sub-sources,Swj=1. Now we can go back to compute the occurrence probability associated
to sub-source j:
Pj(i,M,Tf)=exp(-Dl (M)Tf)=exp(-Dl(M)Tfwj)=exp(ln[P(i,M,Tf)]wj) (4)

j
from which we gather that:
Pj(i,M,Tf)=P(i,M,Tf)wj (5)
Note that, as we had mentioned, if only the occurrence probabilities for the whole source are specified, there
is not a unique way to specify occurrence probabilities associated to sub-sources. However, the path chosen
by CRISIS is, in our view, reasonable, and exact for the case of Poissonian sources.
The only compatibility restriction when using this option is that the file that contains the generalized non-
Poissonian occurrence probabilities must stipulate (in the *.nps file) that the number of sources is equal to 1,
that is, only a set of occurrence probabilities is given. See this link in order to see where this parameter is
stipulated.
Within the CRISIS development team, this combination is known as Peruzza- type, since Prof. Laura
Peruzza suggested its implementation an used it in her calculations in the context of project S2 (2008-2010)
funded by the Italian Civil Protection Authority.
Note C
In this option, source geometry is given in terms of points, there is a "normal" attenuation model and a
parametric seismicity description, either of Gutenberg-Richter or Characteristic Earthquake type. This is an
old CRISIS option, which has no compatibility restrictions.
Note D
In this option, source geometry is given in terms of points, a "normal" ground-motion prediction models is
used, and earthquake occurrence probabilities are given with a generalized non-Poissonian seismicity or a
- 37 -
Chapter 2
generalized Poisson model.

This option is mainly used to model the so called smoothed-seismicity, but now with probabilities obtained
with spatially arbitrarily complex Poissonian or non-Poissonian models. The only compatibility restriction in
this option is that the number of vertex given in the point-sources description must be equal to the number of
sources given in the non-Poissonian or Poissonian seismicity files.
Within the CRISIS development team, this combination is known as Warner- type, since Dr. Warner
Marzocchi suggested its implementation an used it in his calculations in the context of project S2 (2008-
2010) funded by the Italian Civil Protection Authority.
Note E
Currently, the gridded seismicity model works only with Grid sources
2.17.2.
2.17.3. Generalized attenuation models (GAF)
Source/Seismicity
Area AG AG BG BG EG
Line AG AG BG BG EG
SSG CG CG DG DG EG
AreaPlanes AG AG BG BG EG
Grid CG CG DG DG FG
Note AG
In this option, source geometry is a line or an area, but ground-motion characteristics are described with a
generalized attenuation model. This option is impossible, due to the fact that generalized attenuation models
are associated to known, fixed hypocentral locations, while line or area sources contain, implicitly,
uncertainty about future hypocenters. Thus, these source-attenuation choices are incompatible with each
other.
In addition, generalized attenuation models contain information about individual events with known
(although irrelevant) magnitudes. Since each event is associated to a fixed value of magnitude, occurrence
probabilities for each of the events contained in the attenuation model, cannot be computed for continuous,
arbitrary values of magnitude with the information provided by parametric seismicity descriptions, as
earthquake magnitude exceedance rates. It must be remembered that, starting with magnitude exceedance
rates, occurrence probabilities in given time frames can only be computed for magnitude intervals
(magnitude "bins") and not for point values.
Note BG
In this option, the source is a line or an area, seismicity is described with a generalized non-Poissonian
model an ground motion characteristics are given with a generalized attenuation model.
- 38 -
Chapter 2
This option is the only one in which generalized attenuation models can be used. Note that, when using this
type of ground motion model, locations of earthquake hypocenters are, in principle, unknown and
irrelevant. In consequence, specification of a source location is also, in principle, irrelevant. However, there
are two reasons why a source location must be specified: 1) when constructing a hazard model with CRISIS
interface, it is useful for the analyst to have a visual feedback of the source location; and 2) for hazard
disaggregation purposes, CRISIS must know the location to which the hazard coming from all events has to
be assigned. For the purpose of dissaggregation, earthquake location is conventionally considered to be the
geometrical center of the source area or line.
On the other hand, since also earthquake magnitudes are fixed (and irrelevant) in generalized attenuation
models, and each set of grids represents an individual event, it would be impossible to associate to this
events seismicity parameters using parametric descriptions. In view of this, the only possibility is that
earthquake occurrence probabilities are assigned using non-Poissonian generalized models.
Compatibility conditions in this option are the following:
1) The number of sources in the generalized attenuation model (*.gaf) must be the same that the number of
sites in the generalized non-Poissonian seismicity file (*.nps)
2) The number of magnitudes in the generalized attenuation model (*.gaf) must be the same that the number
of sites in the generalized non-Poissonian seismicity file (*.nps)
Within the CRISIS development team, this combination is known as of type Stupazzini-Villani -type, since
Marco Stupazzini and Manuela Villani are the two researchers in charge of developing this type of model in
the context of project S2 (2008-2010) funded by the Italian Civil Protection Authority
Note CG
In this option, source geometry is a collection of points, but ground-motion characteristics are described
with a generalized attenuation model. This option is considered impossible because generalized attenuation
models contain information about individual events with known (although irrelevant) magnitudes. Since
each event is associated to a fixed value of magnitude, occurrence probabilities for each of the events
contained in the attenuation model, cannot be computed for continuous, arbitrary values of magnitude with
the information provided by parametric seismicity descriptions, as earthquake magnitude exceedance rates.
It must be remembered that, starting with magnitude exceedance rates, occurrence probabilities in given
time frames can only be computed for magnitude intervals (magnitude "bins") and not for point values.
Note DG
Note that this option is similar to option BG except that source geometry in option DG is of the point-source
type. In principle, this option could have been regarded a valid, since, when using generalized attenuation
models, source geometry is irrelevant. However, we felt that option BG, in which the source is an area of a
line that is given only for the purpose of visual feedback and dissaggregation, was more useful, and we
inhibited this one to avoid confusion.
Note EG
This combination would be possible but is not implemented yet.
- 39 -
Chapter 2
2.17 GeoSeismAtt Combinations

The different geometry / seismicity /attenuation models give raise to the combinations listed in the following
tables. The tables give a letter or a number that points to a note that gives details on the combinations. In the
tables, "Normal" attenuation refers to attenuation tables or user coded built-in models, while "General"
refers to generalized attenuation models.
The color codes indicate the following:
Green: Combination always valid, regardless of parameter values
Yellow: Combination valid or not depending on parameter values
Red: Combination never valid
Blue: Combination potentially valid, but not implemented yet
Normal attenuation

Source/Seismicity
Area A A B B E
Line A A B B E
SSG C C D D E
AreaPlanes A A B B E
Grid E E E E E
Note A
These are old CRISIS options, which are valid always.

Note B
In this option a source is geometrically modeled as a line or as an area, which means that every point that
belongs to the source has the same probability of being a hypocenter; this is the usual assumption when
using line or area sources in CRISIS. Attenuation, as in the older CRISIS option is modeled with a
parametric description (a "normal" GMPM). However, the new option permits stipulation of earthquake
occurrence probabilities with a generalized non-Poissonian model or a generalized Poisson model, and not
through a parametric frequency-magnitude relation (Gutenberg-Richter or Characteristic Earthquake).
The occurrence probabilities given in the non-Poissonian or Poissonian seismicity files correspond to the
whole source, that is, they are the probabilities of having and earthquake of given magnitude and in a given
time frame anywhere in the source. Using its standard spatial integration scheme, CRISIS will sample the
source in order to compute hazard, accounting for all possible locations of the earthquake within the source.
Note, however, that when probabilities are specified for the whole source, probabilities associated to
segments of the source, or sub-sources, are not univocally defined. The following approach is adopted by
CRISIS in order to define the occurrence probabilities associated to sub-sources of known sizes.
- 40 -
Chapter 2
Assume first that we have a conventional Poissonian source. The probability of having i events of
magnitude M in the next Tf years, due to the effect of the whole fault, P(i,M,Tf) would be given by:
P(i,M,Tf)=exp(-Dl(M)Tf) (1)
where Dl(M) is the Poissonian magnitude occurrence rate of earthquakes with magnitudes in the vicinity of
M, for the whole source. This occurrence rate can be written as:
Dl(M)=-ln[P(i,M,Tf)]/Tf (2)
In the case of Poissonian occurrences, occurrence rates are additive. Thus, the occurrence rate
corresponding to a sub-source of relative size wj, would simply be:
Dl (M)=Dl(M)wj (3)
j
Note that, for all sub-sources,Swj=1. Now we can go back to compute the occurrence probability associated
to sub-source j:
Pj(i,M,Tf)=exp(-Dl (M)Tf)=exp(-Dl(M)Tfwj)=exp(ln[P(i,M,Tf)]wj) (4)

j
from which we gather that:
Pj(i,M,Tf)=P(i,M,Tf)wj (5)
Note that, as we had mentioned, if only the occurrence probabilities for the whole source are specified, there
is not a unique way to specify occurrence probabilities associated to sub-sources. However, the path chosen
by CRISIS is, in our view, reasonable, and exact for the case of Poissonian sources.
The only compatibility restriction when using this option is that the file that contains the generalized non-
Poissonian occurrence probabilities must stipulate (in the *.nps file) that the number of sources is equal to 1,
that is, only a set of occurrence probabilities is given. See this link in order to see where this parameter is
stipulated.
Within the CRISIS development team, this combination is known as Peruzza- type, since Prof. Laura
Peruzza suggested its implementation an used it in her calculations in the context of project S2 (2008-2010)
funded by the Italian Civil Protection Authority.
Note C
In this option, source geometry is given in terms of points, there is a "normal" attenuation model and a
parametric seismicity description, either of Gutenberg-Richter or Characteristic Earthquake type. This is an
old CRISIS option, which has no compatibility restrictions.
Note D
In this option, source geometry is given in terms of points, a "normal" ground-motion prediction models is
used, and earthquake occurrence probabilities are given with a generalized non-Poissonian seismicity or a
generalized Poisson model.
This option is mainly used to model the so called smoothed-seismicity, but now with probabilities obtained
with spatially arbitrarily complex Poissonian or non-Poissonian models. The only compatibility restriction in
- 41 -
Chapter 2
this option is that the number of vertex given in the point-sources description must be equal to the number of
sources given in the non-Poissonian or Poissonian seismicity files.
Within the CRISIS development team, this combination is known as Warner- type, since Dr. Warner
Marzocchi suggested its implementation an used it in his calculations in the context of project S2 (2008-
2010) funded by the Italian Civil Protection Authority.
Note E
Currently, the gridded seismicity model works only with Grid sources
Generalized attenuation models (GAF)

Source/Seismicity
Area AG AG BG BG EG
Line AG AG BG BG EG
SSG CG CG DG DG EG
AreaPlanes AG AG BG BG EG
Grid CG CG DG DG FG
Note AG
In this option, source geometry is a line or an area, but ground-motion characteristics are described with a
generalized attenuation model. This option is impossible, due to the fact that generalized attenuation models
are associated to known, fixed hypocentral locations, while line or area sources contain, implicitly,
uncertainty about future hypocenters. Thus, these source-attenuation choices are incompatible with each
other.
In addition, generalized attenuation models contain information about individual events with known
(although irrelevant) magnitudes. Since each event is associated to a fixed value of magnitude, occurrence
probabilities for each of the events contained in the attenuation model, cannot be computed for continuous,
arbitrary values of magnitude with the information provided by parametric seismicity descriptions, as
earthquake magnitude exceedance rates. It must be remembered that, starting with magnitude exceedance
rates, occurrence probabilities in given time frames can only be computed for magnitude intervals
(magnitude "bins") and not for point values.
Note BG
In this option, the source is a line or an area, seismicity is described with a generalized non-Poissonian
model an ground motion characteristics are given with a generalized attenuation model.
This option is the only one in which generalized attenuation models can be used. Note that, when using this
type of ground motion model, locations of earthquake hypocenters are, in principle, unknown and
irrelevant. In consequence, specification of a source location is also, in principle, irrelevant. However, there
are two reasons why a source location must be specified: 1) when constructing a hazard model with CRISIS
interface, it is useful for the analyst to have a visual feedback of the source location; and 2) for hazard
- 42 -
Chapter 2
disaggregation purposes, CRISIS must know the location to which the hazard coming from all events has to
be assigned. For the purpose of dissaggregation, earthquake location is conventionally considered to be the
geometrical center of the source area or line.
On the other hand, since also earthquake magnitudes are fixed (and irrelevant) in generalized attenuation
models, and each set of grids represents an individual event, it would be impossible to associate to this
events seismicity parameters using parametric descriptions. In view of this, the only possibility is that
earthquake occurrence probabilities are assigned using non-Poissonian generalized models.
Compatibility conditions in this option are the following:
1) The number of sources in the generalized attenuation model (*.gaf) must be the same that the number of
sites in the generalized non-Poissonian seismicity file (*.nps)
2) The number of magnitudes in the generalized attenuation model (*.gaf) must be the same that the number
of sites in the generalized non-Poissonian seismicity file (*.nps)
Within the CRISIS development team, this combination is known as of type Stupazzini-Villani -type, since
Marco Stupazzini and Manuela Villani are the two researchers in charge of developing this type of model in
the context of project S2 (2008-2010) funded by the Italian Civil Protection Authority
Note CG
In this option, source geometry is a collection of points, but ground-motion characteristics are described
with a generalized attenuation model. This option is considered impossible because generalized attenuation
models contain information about individual events with known (although irrelevant) magnitudes. Since
each event is associated to a fixed value of magnitude, occurrence probabilities for each of the events
contained in the attenuation model, cannot be computed for continuous, arbitrary values of magnitude with
the information provided by parametric seismicity descriptions, as earthquake magnitude exceedance rates.
It must be remembered that, starting with magnitude exceedance rates, occurrence probabilities in given
time frames can only be computed for magnitude intervals (magnitude "bins") and not for point values.
Note DG
Note that this option is similar to option BG except that source geometry in option DG is of the point-source
type. In principle, this option could have been regarded a valid, since, when using generalized attenuation
models, source geometry is irrelevant. However, we felt that option BG, in which the source is an area of a
line that is given only for the purpose of visual feedback and dissaggregation, was more useful, and we
inhibited this one to avoid confusion.
Note EG
This combination would be possible but is not implemented yet.
- 43 -
- 44 -
Chapter 3
3. Ground Motion Prediction Models (GMPM)

In general, ground motion prediction models, also called attenuation relations, establish probabilistic
relations between earthquake characteristics and intensities at a site of interest. These relations are
probabilistic since, for given earthquake characteristics, the intensities are regarded as random variables,
whose probability distribution is completely fixed by the GMPM. In most of the cases this means that at
least the first two statistical moments (e.g., the median and the standard deviation of the natural logarithm, in
the lognormal case) of the probability distribution must be furnished by the GMPM.
CRISIS recognizes three "families" of GMPM:
Attenuation tables. In these tables, relations between earthquake characteristics and intensities at a site are
given in terms of the following parameters: magnitude, structural period, source-site distance and depth. For
the first moment (usually the median of a lognormal distribution), the attenuation relations are matrices in
which the rows run for the magnitude and the columns run for the distance. Note that when using
attenuation tables, the relations between magnitude, distance and intensity do not need to be of parametric
nature, since the intensity medians are given, point by point, for magnitude-distance combinations.
Built-In models. These are popular models, published in the literature, in which magnitude, distance and
intensity are probabilistically related by, usually, a set of formulas or parametric equations. There is a set of
built-in models ready to use in CRISIS and there is also the possibility of adding new models.
Generalized models. Generalized attenuation models are non-parametric probabilistic descriptions of the
ground motions produced by an earthquake. In the context of CRISIS, a generalized attenuation model is a
collection of probabilistic footprints, one for each of the events considered in the analysis. Each footprint
gives, in probabilistic terms, the geographical distribution of the intensities produced by this event.
3.19 Adding new built-in GMPM

In addition to attenuation tables and generalized attenuation models, CRISIS admits built-in GMPM, which
are given to the code in the form of classes compiled in a Dynamic Link Library (dll). CRISIS includes a
number of built-in GMPM, which can be consulted in this link. However, this collection can be extended
by way of writing code for user-defined GMPM. Each new GMPM must be a new class that implements
some standard methods.
A detailed, fully documented example of the construction of a ground- motion prediction model
implementing the methods just described is presented in the form of VB.Net project GMPETutorial that is
distributed as part of CRISIS installation package, with the name GMPETutorial.zip. See here an example
of a built-in model.
3.1 Attenuation tables

These tables give CRISIS probabilistic relations between magnitude, source-site distance and intensities.
Each attenuation table must be in a different file and must contain the following information:
Attenuation table header
- 45 -
Chapter 3
This is a new part of the attenuation table, starting with CRISIS2008. All the lines of this portion are
optional, so as to keep back-compatibility with older attenuation tables. The reader, however, must be aware
of the default values that are used for the parameters that will be described in this numeral.
This header can contain up to 4 lines that give different characteristics of the attenuation table;lines can be
given in any order. Field names (including capital letters) are fixed. All header lines have the following
format:
# : Field name : Field value
The following table gives the four possible header fields recognized by CRISIS
Field name Field value Comments Default value

A string giving a brief description of
the attenuation table. Description
This information is for
usually includes: author, date of "Not
Description presentation purposes in the
publication, type of earthquakes for available"
Attenuation Data screen
which the model was derived, and so
on.
The original units will appear, for
information purposes, in the
Attenuation Data screen.
A string giving the units for which However, the original units might "Not
Units
the model was developed be relevant if a Units Coefficient available"
is given in order to convert from
these original units to user units,
given in the Intensities screen
Supported values are:
Normal=1
In integer number indicating the
Distribution probability distribution assigned to Lognormal=2 2 (Lognormal)
the residuals of the attenuation model
Beta=3
Gamma=4
A string giving the physical
See the meaning of physical
Dimension dimension of the intensities described "Acceleration"
dimension in this link
in the attenuation table
Parameters defining the magnitude limits (1 line)

MINF, MSUP, NMAG
Variable Meaning
MINF Lower limit of magnitude given in the table
MSUP Upper limit of magnitude given in the table
NMAG Number of magnitudes for which intensity is given
- 46 -
Chapter 3
CRISIS assumes than intensities are given for magnitudes M(K), where
M(K)=MINF+(K-1)*DMAG (1)
DMAG=(MSUP-MINF)/(NMAG-1) (2)
Parameters defining the distance limits and type (1 line)

RINF, RSUP, NRAD, TYPE
Variable Meaning
RINF Lower limit of distance given in the table
RSUP Upper limit of distance given in the table
NRAD Number of distances for which intensity is given
An integer indicating the type of distance used by the
TYPE
attenuation table
CRISIS assumes than intensities are given for distances R(K), where
log(R(K))=log(RINF)+(K-1)*DLRAD (3)
DLRAD=(log(RSUP)-log(RINF))/(NRAD-1) (4)
In other words, distances are supposed to be logarithmically spaced.

TYPE can have the following values, depending on the type of distance to be used:
Value Type of distance

1 (or blank) Focal
2 Epicentral
3 Joyner and Boore
4 Closest to rupture area (Rrup)
For each of the NT different intensity measures, the following blocks of lines:
T(J), SLA(J,0), AMAX(J), COEFH
Variable Meaning
Structural period of j-th spectral ordinate. It is used only for identification purposes and to
T(J) plot the uniform-hazard spectrum, so in the cases in which structural period has no
meaning, it can be just a sequential number
Standard deviation of the natural logarithm of the j-th measure of intensity. A value of
SLA(J,0)<=0 implies that the user will give standard deviations that vary with magnitude.
SLA(J,0)
In this case, the corresponding s values, one for each of the NMAG magnitudes has to be
given after the table of SA() values
AMAX(J) See Probabilistic interpretation of attenuation relations for a definition of this quantity
COEFH Depth coefficient. See explanation below
- 47 -
Chapter 3
Some modern attenuation relations have a coefficient to make the intensity explicitly dependent on focal
depth. This information is given with coefficient COEFH, so that:
MED(A|M,R)=SA(M,R)*exp(COEFH*H) (5)
where MED(A|M,R) is the (depth-dependent) median value of intensity for given values of magnitude M
and distance R; SA(M,R) is the median intensity given in the table for the same values of magnitude and
distance, and H is focal depth.
Matrix of median intensities, associated to a magnitude (row) and a distance (column)
SA(1,1,1), SA(1,1,2),,SA(J,K,L),.,SA(NT,NMAG,NRAD)
SA(J,K,M): Median value of the intensity, for the J-th spectral ordinate, the K-th magnitude and the L-th
distance.
Only if SLA(J) <= 0:
SLA(J,1)
SLA(J,2)
...
SLA(J,NMAG)
Example
In this example, an attenuation model for NT=2 periods (or measures of intensity) is given:
# : Description
: Sample attenuation file constructed for illustration purposes (2008)
# : Units : cm/sec/sec
# : Distribution
:2
# : Dimension : Acceleration
4.5 8.5 5 5 magnitudes between 4.5 and 8.5
5.0 500.0 10 1 10 distances, log-spaced between 5 and 500 Km; focal distance
0.0 0.7 0.0
0.0 Period 0. Sigma=0.7. Amax=0 (no truncation), CoefH=0
119.3 97.5 70.3
45.3 26.8 14.7 7.3 3.2 1.1 0.3
202.5 165.6 119.4 76.9 45.5 24.9 12.4 5.4 1.9 0.5
344.0 281.2 202.7 130.6 77.3 42.3 21.1 9.1 3.2 0.8
584.1 477.6 344.3 221.8 131.2 71.9 35.9 15.5 5.4 1.3
992.0 811.1 584.7 376.7 222.9 122.1 60.9 26.4 9.1 2.2
0.2 -0.7 0.0035 Period 0.2. Sigma variable with M. Amax=0 (no truncation), CoefH=
0.0
0.0035
250.4 203.2 145.2 92.7 54.2 29.4 14.5 6.2 2.1 0.5
420.4 341.3 244.0 155.7 91.2 49.4 24.3 10.4 3.6 0.8
708.3 575.2 411.3 262.6 153.8 83.4 41.1 17.6 6.0 1.4
1193.5 969.6 693.8 443.3 259.9 141.0 69.5 29.7 10.2 2.4
2014.4 1637.1 1172.1 749.4 439.6 238.6 117.7 50.4 17.3 4.1
0.830 5 values of magnitude-dependent Sigma, one for each magnitude
- 48 -
Chapter 3
0.784
0.615
0.623
0.514
3.2 Built-in attenuation models

These are popular attenuation equations, published in the literature, that the user can choose as GMPM for
CRISIS. These models relate, in probabilistic terms, earthquake magnitude and a certain distance metric
with the intensity at a site. (CRISIS can handle 4 types of distance metrics). Also, many of these attenuation
equations require specification of additional parameters that the user must select, such as style of faulting
and soil type.
See the GMPM that are currently built-in into CRISIS.
Note that, besides the parameters that each GMPM uses -such as soil type or style of faulting- all built-in
GMPM contain two extra parameters, called "Units coefficient" and "Sigma truncation". The first one is
used to change the original units of the model (see details) while the second one is used to truncate the
probability distribution of the residuals (see details).
There is also the possibility of adding new built-in models to CRISIS, that, in theory, can be of an arbitrary
level of complexity.
3.3 Generalized attenuation models

Generalized attenuation models are non- parametric probabilistic descriptions of the ground motions
produced by an earthquake.
Ground motions descriptions obtained when using traditional attenuation models -also called ground motion
prediction equations (GMPE)- are generally functions of earthquake magnitude and source-to-site distance.
But generalized attenuation models are not explicit functions of magnitude and distance. They are simply
probabilistic footprints of the ground motions produced by an individual event.
In the context of CRISIS, a generalized attenuation model is a collection of probabilistic footprints, one for
each of the events considered in the analysis. Each footprint gives, in probabilistic terms, the geographical
distribution of the intensities produced by this event.
For a given event, the footprint consists of several pairs of grids of values. Each pair of grids is associated to
one of the intensity measures for which hazard is to be computed. CRISIS needs two grids for each
intensity measure because, as with other ground motion prediction models, the intensity caused by the
earthquake is considered probabilistic, so CRISIS requires two statistical moments in order to fix a
probability density function of the intensity caused by an earthquake at a particular location.
For instance, assume that one generalized attenuation model will be used to describe the intensities caused
by 10 different earthquakes. Also, assume that the hazard analysis is being made for seven intensity
measures (for instance, the response spectral ordinates for seven different structural periods). For this
- 49 -
Chapter 3
example, each event will be described by 14 different grids, two for each intensity measure, the first one
giving the geographical distribution of the median intensity and the second one given the geographical
distribution of the standard deviation of the natural logarithm of the intensity. Hence, a total of 140 grids
will form the generalized attenuation model of this example. It would be natural that all the 140 grids
covered exactly the same region; however, there are no restrictions at this respect.
From this description, it is clear that it would be extremely difficult to perform a hazard study of regional (or
higher) size using generalized attenuation models. Usually, a hazard model of regional size contains
thousands of events, and the task of geographically describing the intensities caused by all of them in non-
parametric form would be titanic.
Rather, generalized attenuation models will very likely be used for local studies, for which the relevant
earthquakes are few and can be clearly identified. In this case, the grids of required values (geographical
distribution of statistical moments of one or more intensity measures for each event) can be constructed
using, for instance, advanced ground-motion simulation techniques.
Generalized attenuation models are given to CRISIS in the form of binary generalized attenuation files
(GAF). The reason for requiring the GAFs to be in binary format is the computational need of having
random access to individual intensity values. This need is basically dictated by computational speed.
The following tables illustrate the detailed format of GAFs.
Description Type Length Comments

Custom file Give a synthetic description of the main
String Variable
description features of the GAF
Original Units String Variable
Intensity
physical String Variable
dimension
Data type
(short, integer,
Integer 4
single, double,
long)
Probability
distribution
assigned to
intensity Integer 4
(normal,
lognormal,
beta, gamma)
Number of
intensity
measures Integer 4
(number of
periods)
- 50 -
Chapter 3

Number of
sources Integer 4
(locations)
Number of
magnitudes per Integer 4
location
Number of
statistical
Integer 4
moments of
intensity stored
Period values are required because the user
Period 1 Double 8
might want to compute for arbitrary periods
Period 2 Double 8

Period Number
of Intensity Double 8
measures
Magnitude values are required to compute
Magnitude occurrence rates when GR or Characterisitic
representative Double 8 models are used. When a non-Poissonian
of bin 1 seismicity file is given these magnitudes are
irrelevant
Magnitude
representative Double 8
of bin 2

Magnitude
representative Double 8
of last bin
occurrence rates when GR or Characterisitic
Scenario name Char 40 models are used. When a non-Poissonian
seismicity file is given these magnitudes are
irrelevant
Grid for
intensity
ModGRN 56+Nbytes*Nx1*Ny1
measure 1,
moment 1
Grid for
intensity
measure 1,
moment 2
- 51 -
Chapter 3

Grid for
intensity
measure 1, ModGRN 56+Nbytes*Nx1*Ny1
moment
NumMoments
Grid for
intensity
measure 2,
moment 1
Grid for
intensity
measure 2,
moment 2

Grid for
intensity
moment
NumMoments

Grid for
intensity
Then, the actual georeferenced probabilistic
measure ModGRN 56+Nbytes*Nx1*Ny1
intensity values follow
NumInt,
moment 1
Grid for
intensity
NumInt,
moment 2

Grid for
intensity
measure
NumInt,
moment
NumMoments
occurrence rates when GR or Characterisitic
Scenario name Char 40 models are used. When a non-Poissonian
seismicity file is given these magnitudes are
irrelevant
- 52 -
Chapter 3

Grid for
intensity
measure 1,
moment 1
Grid for
intensity
measure 1,
moment 2

Grid for
intensity
moment
NumMoments
Grid for
intensity
measure 2,
moment 1
Grid for
intensity
measure 2,
moment 2

Grid for
intensity
moment
NumMoments

Grid for
intensity
Then, the actual georeferenced probabilistic
intensity values follow
NumInt,
moment 1
Grid for
intensity
NumInt,
moment 2

Grid for
intensity ModGRN 56+Nbytes*Nx2*Ny2
measure
- 53 -
Chapter 3

NumInt,
moment
NumMoments
Similar blocks
continue for
all scenarios
3.4 Measuring distances (suggested by Dr. R. Secanell)

In CRISIS, there are four ways of measuring site-to-source distances:
1) Focal (RF)
2) Epicentral (REPI)
3) Joyner and Boore (closest distance to the projection of the fault plane on the Earths surface; RJB)
4) Closest distance to rupture area (RRUP)
The following figure illustrates the different distances:
- 54 -
Chapter 3
H is the focal depth. Computation of RF and REPI deserves no further comments. Computation of RRUP and
RJB, however, require the specification of a rupture area. In CRISIS, the area is specified with the following
criteria:
The rupture area is assumed to be circular with radius r, which depends on magnitude in a way specified by
the user (see Relation between magnitude and fault radius).
The circular fault is contained in the plane defined by the triangle resulting from source subdivision, whose
centroid is assumed to be the hypocentral location.
Note that, if the site is within the projection of the fault in the Earths surface, RJB=0 and RRUP=H.
The user must indicate to CRISIS what type of distance he wishes to use, depending on the characteristics
of the attenuation relation being used. This is done directly in the attenuation relations or ground motion
prediction models (GMPM)
Computation of the exact values of distances RJB and RRUP is cumbersome. To save computation time, the
exact values are approximated with simpler formulas that produce small errors.
3.5 Physical dimension

In order to have tighter checks of the compatibility among different ground-motion prediction models
(GMPM) when performing logic-tree computations, each GMPM must be assigned aphysical dimension of
the measures of intensity the model is describing. The physical dimension of most GMPM is acceleration
(because they are usually constructed for PGA and the response spectral ordinates at selected periods), but
other physical dimensions are also accepted. CRISIS, so far, accepts the following physical dimensions,
which correspond to classes defined for this purpose:
Physical dimension Assembly name

Acceleration Crisis2008.NewAttenuation.dll
Velocity Crisis2008.NewAttenuation.dll
Displacement Crisis2008.NewAttenuation.dll
MMI Crisis2008.NewAttenuation.dll
MCSI Crisis2008.NewAttenuation.dll
DuctilityDemand ExtraDimensions.dll
ISDrift ExtraDimensions.dll
Although only these physical dimensions are recognized by CRISIS, it is relatively simply to construct
additional classes associated to other intensity measures. To do so, the constructed class must implement the
following methods:
Method Purpose
Returns an integer indicating
Public ReadOnly Property distancePow() As Integer the distance power of this
dimension
Public ReadOnly Property forcePow() As Integer the force power of this
dimension
- 55 -
Chapter 3
Method Purpose
Public ReadOnly Property timePow() As Integer the time power of this
dimension
Public ReadOnly Property chargePow() As Integer the charge power of this
dimension
Provides a number specific to
Public MustOverride ReadOnly Property name() As String
the class
Checks if the types have same
Public Overrides Function Equals(ByVal obj As Object) As
power for MKSA elements
Boolean
describing dimensions
Classes constructed that implement these methods must be compiled to the form of a dll, which must be
reside in CRISIS application directory. In addition, file CRISISDimensions.ini must be edited to add the
new classes. The general format of the lines of this file is the following:
Full class name, Assembly name
3.6 Probabilistic interpretation of attenuation relations

In general, given a magnitude and a distance, intensity A is assumed to be a random variable with a given
probability distribution (usually lognormal). Attenuation relations (also called ground motion prediction
models, or GMPM) give the two first statistical moments of A given a magnitude and a distance, that is,
A|M,R. These two moments usually describe the mean or median value of A|M,R and a measure of its
uncertainty.
Up to now, CRISIS supports three probability distributions that can be used to describe intensities. These
distributions are presented in the following table, along with the two statistical moments that have to be
given in order to correctly describe A|M,R as a random variable.
Distribution 1st moment (m ) 2nd moment (m ) Lower limit Amax

1 2
Standard deviation
Lognormal Median of the natural 0 m1exp(Km2)
logarithm
Gamma Mean Standard deviation 0 m1+Km2
Normal Mean Standard deviation -infinity m1+Km2
As part of the hazard computations, CRISIS requires to compute the probability that intensity A at a given
site exceeds a known value, a, given that at some hypocentral location, H, an earthquake of magnitude M
took place, that is, Pr(A>a|M,H).
If no truncation is applied to intensity values, this probability is computed with the following expression:
Pr(A>a|M , H)=1 - FA[a; m1(M , H),m2(M , H)] (1)
- 56 -
Chapter 3
where m1(M , H) and m2 (M , H) are the first and second moments, respectively, of intensity A, given that at
hypocentral location H an earthquake of magnitude M took place. Depending on the probability distribution
assigned to A, the first and second moments have the interpretation presented in the previous table. FA [a;
m1 (M , H), m 2(M , H)] is the probability distribution of A (also called the cumulative probability function)
whose form depends on the type of distribution chosen.
The moments of A|M,R, that is, m 1(M , H) and m 2(M , H) are given by the user by means of attenuation
relations or GMPM.
In many cases, truncation is specified in the GMPM trough a parameter called "Sigma truncation", Tc.
This means that the integration across the attenuation relation uncertainty implied in the previous equations
is not carried out up to infinity, but up to a certain value, Tc.
Depending on the value of the truncation coefficient given in the GMPM, the following considerations are
made:
Tc=0
In this case, no truncation is applied, so Equation 1 is used.
Tc> 0
In this case, a truncated distribution between the lower limit of A and T c is assumed, regardless of
magnitude and distance. Hence,
(2)
Tc<0
In this case, ABS(Tc) = K, is interpreted as the number of standard deviations, for which integration will be
performed. Hence, the integration will be performed between the lower limit and Amax , both given in the
previous table. Therefore,
(3)
Depending on the distribution chosen, Amax takes the values indicated in the previous table. Note that in this
case, the actual truncation value for A depends on magnitude and distance.
In the following graph, the effect of the different truncation schemes can be observed:
- 57 -
Chapter 3
3.7 Relation between magnitude and fault size

In CRISIS, attenuation relations can be specified in terms of 4 different measures of distance. If distances
R RUP or R JB are used, CRISIS must have means to know the rupture area or the rupture length, as a
function of magnitude, in order to compute the required distances.
In general, CRISIS assumes that the relation between area/rupture length and magnitude is
A=K1exp(K2M) (for area sources, A in Km2) (1)
L=K3exp(K4M) (for line sources, L in Km) (2)
where A is the source area (in km2), L is the rupture length (in km), M stands for magnitude and K1 , K2, K3
and K4 are constants given by the user or chosen from a built-in set of constants.
In the case of area sources and point sources, CRISIS will assume that the earthquake takes place in a plane
defined by the source geometry, and that the rupture area will be a circle, within this plane, with area A and,
in consequence, with radius
r2=A/p (3)
In the case of line sources, CRISIS will assume that the earthquake takes place along a line defined by the
source geometry, and that the rupture length will be centered at the hypocenter.
- 58 -
Chapter 3
3.8 Special Attenuation Models

In the most frequent case, only one attenuation model will be assigned to a source. However, there is the
possibility to assign one ore more special attenuation models to a source, which will be effective only for
sites located inside corresponding polygons, called special attenuation regions given by the user.
If special attenuation models are given, then CRISIS will proceed in the following way:
When computing hazard from a source, CRISIS will check if this source has special attenuation models. If it
does not, then it will use the general attenuation model for the source. If the source was assigned special
models, then CRISIS will check if the site of computation is inside one of the user-given polygons. If
affirmative, CRISIS will use the model assigned to this source-site combination. If the site is not inside any
of the special polygons, then CRISIS will use the general attenuation model of the source.
It must be noted that if site-effects grids are given, the amplification factors will be applied on top of the
intensities computed either with the general attenuation model assigned to the source or with attenuation
models assigned to special attenuation regions.
3.9 Units coefficient

All GMPM used by CRISIS (attenuation tables, built-in models and generalized models) give probabilistic
relations between earthquake properties, site characteristics and intensities at a site. These relations are given
in terms of the first and second statistical moments of the intensities, given some earthquake and site
parameters. For instance, for the GMPM in which the residuals are lognormally distributed, the GMPM
gives the median and the standard deviation of the natural logarithm of the intensity, given a set of
earthquake and site parameters.
The GMPM are constructed to express the intensities in certain units. These units are called the "original
units". However, the user of CRISIS might want to perform calculations for intensities expressed in other
units. For instance, a user could be using a model whose original units are cm/sec 2, but he wants to make
hazard calculations for intensities expressed as a fraction of the acceleration of gravity, g. In these case, the
g units will be called the "user units", which are given by the user in the Intensities Screen.
The Units Coefficient (UC) is a positive number used to change from the model's original units to the user
units, using the following relation:
1 user unit = UC original units (1)
For instance, in the example given above, since the original units are cm/sec2 and the user units are fractions
of g, we would have that 1 g = UC cm/sec2, and hence UC=981, because 1 g = 981 cm/sec2.
- 59 -
Chapter 3
3. Hazard disaggregation
Magnitude-distance disaggregation
Consider the basic hazard computation equation (see the basic theoretical background):
(1)
where Pr(A<a|T j ) is the probability of not exceeding intensity a at a site in the next T j years, when
subjected to a seismic regime composed by N point sources, each of which produces earthquakes of
magnitudes M1 , M 2 ,...,MNM .. It can be noted that the product in equation 1 is composed by many terms,
each of which corresponds to a particular magnitude value, M i , and to a specific source-to-site distance,
which is the one from source k to the site for which hazard is being computed.
In view of this, the contributions to Pr (A<a|T j ) or to Pr (A>a|T j ) could be grouped for a range of
magnitudes (say, from 1 to M2 ) and a range of distances. This is the magnitude-distance disaggregation.
These results indicate which combinations of magnitude and distance contribute more to the seismic hazard
at a site, for a given intensity measure, for a given time frame, and at certain level of intensity, a in this case.
Let's say that hazard has been disaggregated, leading to a matrix of Ng rows (one for each magnitude range)
and Nr columns (one for each distance range). The contents of each cell must be such that the following
relation is satisfied:
(2)
In other words, the original non-exceedance probability must be equal to the product of the non-exceedance
probabilities disaggregated for each magnitude-distance bin. This means that, oppositely to what happens
with intensity exceedance rates, which are additive, non- exceedance probabilities (or exceedance
probabilities) are not additive, but multiplicative in the sense expressed by equation 2 above. In view of this,
when seeing CRISIS disaggregation results, the user must not expect that the exceedance probabilities
the same site, intensity value and time frame. As shown by the previous paragraphs, arithmetic of
exceedance probabilities is more complex to that of intensity exceedance rates used in conventional hazard
studies.
Epsilon disaggregation
In occasions, it is interesting to know which portions of the intensity probability density function contribute
most to the seismic hazard at a given site. Consider the following equation, which is equation 1 but written
in terms of exceedance probabilities:
- 60 -
Chapter 3
(3)
For given magnitude, time frame and source location, the term Pr(A>a|Mi, Tj, k) will be computed by
calculating the area shown in green in the following figure:
The shape of the probability density function of Sa (shown in black in the previous figure) depends on
magnitude, distance, and ground-motion prediction model employed, while a is an arbitrarily fixed value:
the one for which seismic hazard is being computed.
However, it is sometimes of interest to know how much of the probability marked in green comes from the
high percentiles of the distribution. For instance, how much of the probability comes from the area to the
right of value L, shown in blue in the previous figure. Normally, L is indexed to an "epsilon" (e) value, such
that:
(4)
where MED(A|Mi, Tj, k) and s LN (A|Mi, Tj, k) are, respectively, the median and the logarithmic standard
deviation of A given magnitude Mi at source k; the value of e is kept fixed for the whole analysis. In view of
- 61 -
Chapter 3
this, when an epsilon disaggregation is required, exceedance probabilities required to evaluate equation 4
are computed with:
(5)
where pA|Mi, Tj, k() is the probability density function of A given magnitude Mi at source k, and:
(6)
Interpretation of e for other probability distributions

Usually, intensity A is assigned a lognormal probability distribution, so Equation 4 can be used to compute
the lower integration limit, L. However, admits the possibility of using four different types of probability
distributions: Lognormal, Gamma, Normal and Beta. In the three last cases, the meaning of e is not
unambiguously defined. In CRISIS, the following interpretations of e are adopted:
For the Gamma distribution
(7)
For the Normal distribution
(8)
For the Beta distribution
(9)
In the three cases above, E(A|Mi, Tj, k) and s (A|Mi, Tj, k) are, respectively, the expected value and the
standard deviation of A given magnitude Mi at source k.
3.9 Hazard disaggregation

Magnitude-distance disaggregation
Consider the basic hazard computation equation (see the basic theoretical background):
- 62 -
Chapter 3
(1)
where Pr(A<a|T j ) is the probability of not exceeding intensity a at a site in the next T j years, when
subjected to a seismic regime composed by N point sources, each of which produces earthquakes of
magnitudes M1 , M 2 ,...,MNM .. It can be noted that the product in equation 1 is composed by many terms,
each of which corresponds to a particular magnitude value, M i , and to a specific source-to-site distance,
which is the one from source k to the site for which hazard is being computed.
In view of this, the contributions to Pr (A<a|T j ) or to Pr (A>a|T j ) could be grouped for a range of
magnitudes (say, from 1 to M2 ) and a range of distances. This is the magnitude-distance disaggregation.
These results indicate which combinations of magnitude and distance contribute more to the seismic hazard
at a site, for a given intensity measure, for a given time frame, and at certain level of intensity, a in this case.
Let's say that hazard has been disaggregated, leading to a matrix of Ng rows (one for each magnitude range)
and Nr columns (one for each distance range). The contents of each cell must be such that the following
relation is satisfied:
(2)
In other words, the original non-exceedance probability must be equal to the product of the non-exceedance
probabilities disaggregated for each magnitude-distance bin. This means that, oppositely to what happens
with intensity exceedance rates, which are additive, non- exceedance probabilities (or exceedance
probabilities) are not additive, but multiplicative in the sense expressed by equation 2 above. In view of this,
when seeing CRISIS disaggregation results, the user must not expect that the exceedance probabilities
the same site, intensity value and time frame. As shown by the previous paragraphs, arithmetic of
exceedance probabilities is more complex to that of intensity exceedance rates used in conventional hazard
studies.
Epsilon disaggregation
In occasions, it is interesting to know which portions of the intensity probability density function contribute
most to the seismic hazard at a given site. Consider the following equation, which is equation 1 but written
in terms of exceedance probabilities:
(3)
For given magnitude, time frame and source location, the term Pr(A>a|Mi, Tj, k) will be computed by
calculating the area shown in green in the following figure:
- 63 -
Chapter 3
The shape of the probability density function of Sa (shown in black in the previous figure) depends on
magnitude, distance, and ground-motion prediction model employed, while a is an arbitrarily fixed value:
the one for which seismic hazard is being computed.
However, it is sometimes of interest to know how much of the probability marked in green comes from the
high percentiles of the distribution. For instance, how much of the probability comes from the area to the
right of value L, shown in blue in the previous figure. Normally, L is indexed to an "epsilon" (e) value, such
that:
(4)
where MED(A|Mi, Tj, k) and s LN (A|Mi, Tj, k) are, respectively, the median and the logarithmic standard
deviation of A given magnitude Mi at source k; the value of e is kept fixed for the whole analysis. In view of
this, when an epsilon disaggregation is required, exceedance probabilities required to evaluate equation 4
are computed with:
(5)
- 64 -
Chapter 3
where pA|Mi, Tj, k() is the probability density function of A given magnitude Mi at source k, and:
(6)
Interpretation of e for other probability distributions

Usually, intensity A is assigned a lognormal probability distribution, so Equation 4 can be used to compute
the lower integration limit, L. However, admits the possibility of using four different types of probability
distributions: Lognormal, Gamma, Normal and Beta. In the three last cases, the meaning of e is not
unambiguously defined. In CRISIS, the following interpretations of e are adopted:
For the Gamma distribution
(7)
For the Normal distribution
(8)
For the Beta distribution
(9)
In the three cases above, E(A|Mi, Tj, k) and s (A|Mi, Tj, k) are, respectively, the expected value and the
standard deviation of A given magnitude Mi at source k.
- 65 -
Chapter 3
3. Logic trees
The following paragraphs, giving a brief introduction to logic trees in the context of seismic hazard analysis,
have been taken from "On the Use of Logic Trees for Ground-Motion Prediction Equations in Seismic-
Hazard Analysis" by Julian J. Bommer, Frank Scherbaum, Hilmar Bungum, Fabrice Cotton, Fabio Sabetta,
and Norman A. Abrahamson, Bulletin of the Seismological Society of America, Vol. 95, No. 2, pp. 377
389, April 2005, doi: 10.1785/0120040073:
"Logic trees are widely used in probabilistic seismic hazard analysis as a tool to capture the epistemic
uncertainty associated with the seismogenic sources and the ground-motion prediction models used in
estimating the hazard..."
"Logic trees were first introduced into probabilistic seismic hazard analysis (PSHA) by Kulkarni et al.
(1984) as a tool to capture and quantify the uncertainties associated with the inputs required to perform
such an analysis, and they have since become a standard feature of PSHA (Coppersmith and Youngs,
1986; Reiter, 1990)..."
"Handling uncertainties is a key element of SHA [Seismic Hazard Analysis]. Distinction is made between
two types of uncertainty in seismic hazard assessment, and these are given the adjectives aleatory and
epistemic (e.g., Budnitz et al., 1997), terms used to replace and distinguish between the terms randomness
and uncertainty, whose use has become ambiguous (Bommer, 2003). Uncertainties that are related to an
apparent randomness in nature, such as the scatter associated with empirical relationships, are referred to
as aleatory variability. If the aleatory variability can be measured, usually by fitting observations to an
assumed probability distribution, it is then straightforward to incorporate this variability directly into the
hazard calculations. The most important aleatory variability in SHA is that associated with ground-motion
prediction equations, which is generally represented by the standard deviation of the logarithmic residuals
of the predicted parameter. Standard practice in PSHA is now to integrate across this aleatory variability
within the hazard calculations..."
"Uncertainties reflecting the incomplete knowledge of, say, seismicity, rupture characteristics, and seismic
energy excitation, are referred to as epistemic. There are many epistemic uncertainties in any seismic
hazard assessment, including the characteristics of the seismic source zones (be these area zones or specific
faults), the model for the recurrence relationship, and the maximum earthquake magnitude. In PSHA, the
established procedure is to incorporate the epistemic uncertainty into the calculations through the use of
logic trees. The logic tree is set up so that for each of the steps in which there is epistemic uncertainty,
separate branches are added for each of the choices that the analyst considers feasible. To each of these a
normalized weight is assigned that reflects that analysts confidence that this is the most correct model, and
the weights are generally, but not necessarily, centered on a best estimate. The hazard calculations are then
performed following all the possible branches through the logic tree, each analysis producing a single
hazard curve showing ground motion against annual frequency of exceedance. The weighting of each
hazard curve is determined by multiplying the weights along all the component branches..."
"For every branch added to a logic tree, a penalty is paid in terms of additional calculations; if there are
multiple branches for each component of the hazard analysis, the total number of hazard calculations can
rapidly become very large. For this reason it is advisable to avoid using branches with very small
differences between the options that they carry, in cases when these options result in very similar nodes..."
- 66 -
Chapter 3
"An important principle to follow in setting up a logic tree, but not always taken into account, is that the
options represented by the branches extending from a single node should encompass the complete range of
physical possibilities that particular parameter could be expected to take. This is consistent with the
objective of the logic tree in capturing epistemic uncertainty, which arises from lack of knowledge. The
branches should be set up so that, as knowledge improves, mainly through the gathering of more and better
data, revised estimates for the parameters should fall within the bounds expressed by the logic-tree
branches..."
"true" one. Results from the different branches, along with the weights assigned to each branch, are
computed using the combination rule that will be described in the following paragraphs.
Assume that the probability of exceeding level a of a intensity measure A at a site, in the i-th time frame,
accoring to the j-th branch of a logic tree is P ij(A>a). Assume also that the probability of being the true one
assigned to the j-th branch is wj, j=1,...,N. It is required that the N weights add up to unity.
Then, the expected value of Pij(A>a) once all branches have been accounted for, Pi(A>a), is given by:
(1)
*.dat file that will have the base name of the file that described the combination but with the extension *.dat.
This new hazard model can be loaded into CRISIS and the corresponding hazard results can be analyzed
with CRISIS (hazard maps, exceedance probability curves, uniform hazard spectra) as if they were the
results of a regular *.dat file. Disaggregation results, however, can not be obtained for the hazard resulting
from the logic-tree combination.
3.9 Logic trees

The following paragraphs, giving a brief introduction to logic trees in the context of seismic hazard analysis,
have been taken from "On the Use of Logic Trees for Ground-Motion Prediction Equations in Seismic-
Hazard Analysis" by Julian J. Bommer, Frank Scherbaum, Hilmar Bungum, Fabrice Cotton, Fabio Sabetta,
and Norman A. Abrahamson, Bulletin of the Seismological Society of America, Vol. 95, No. 2, pp. 377
389, April 2005, doi: 10.1785/0120040073:
"Logic trees are widely used in probabilistic seismic hazard analysis as a tool to capture the epistemic
uncertainty associated with the seismogenic sources and the ground-motion prediction models used in
estimating the hazard..."
"Logic trees were first introduced into probabilistic seismic hazard analysis (PSHA) by Kulkarni et al.
(1984) as a tool to capture and quantify the uncertainties associated with the inputs required to perform
- 67 -
Chapter 3
such an analysis, and they have since become a standard feature of PSHA (Coppersmith and Youngs,
1986; Reiter, 1990)..."
"Handling uncertainties is a key element of SHA [Seismic Hazard Analysis]. Distinction is made between
two types of uncertainty in seismic hazard assessment, and these are given the adjectives aleatory and
epistemic (e.g., Budnitz et al., 1997), terms used to replace and distinguish between the terms randomness
and uncertainty, whose use has become ambiguous (Bommer, 2003). Uncertainties that are related to an
apparent randomness in nature, such as the scatter associated with empirical relationships, are referred to
as aleatory variability. If the aleatory variability can be measured, usually by fitting observations to an
assumed probability distribution, it is then straightforward to incorporate this variability directly into the
hazard calculations. The most important aleatory variability in SHA is that associated with ground-motion
prediction equations, which is generally represented by the standard deviation of the logarithmic residuals
of the predicted parameter. Standard practice in PSHA is now to integrate across this aleatory variability
within the hazard calculations..."
"Uncertainties reflecting the incomplete knowledge of, say, seismicity, rupture characteristics, and seismic
energy excitation, are referred to as epistemic. There are many epistemic uncertainties in any seismic
hazard assessment, including the characteristics of the seismic source zones (be these area zones or specific
faults), the model for the recurrence relationship, and the maximum earthquake magnitude. In PSHA, the
established procedure is to incorporate the epistemic uncertainty into the calculations through the use of
logic trees. The logic tree is set up so that for each of the steps in which there is epistemic uncertainty,
separate branches are added for each of the choices that the analyst considers feasible. To each of these a
normalized weight is assigned that reflects that analysts confidence that this is the most correct model, and
the weights are generally, but not necessarily, centered on a best estimate. The hazard calculations are then
performed following all the possible branches through the logic tree, each analysis producing a single
hazard curve showing ground motion against annual frequency of exceedance. The weighting of each
hazard curve is determined by multiplying the weights along all the component branches..."
"For every branch added to a logic tree, a penalty is paid in terms of additional calculations; if there are
multiple branches for each component of the hazard analysis, the total number of hazard calculations can
rapidly become very large. For this reason it is advisable to avoid using branches with very small
differences between the options that they carry, in cases when these options result in very similar nodes..."
"An important principle to follow in setting up a logic tree, but not always taken into account, is that the
options represented by the branches extending from a single node should encompass the complete range of
physical possibilities that particular parameter could be expected to take. This is consistent with the
objective of the logic tree in capturing epistemic uncertainty, which arises from lack of knowledge. The
branches should be set up so that, as knowledge improves, mainly through the gathering of more and better
data, revised estimates for the parameters should fall within the bounds expressed by the logic-tree
branches..."
"true" one. Results from the different branches, along with the weights assigned to each branch, are
computed using the combination rule that will be described in the following paragraphs.
- 68 -
Chapter 3
Assume that the probability of exceeding level a of a intensity measure A at a site, in the i-th time frame,
accoring to the j-th branch of a logic tree is P ij(A>a). Assume also that the probability of being the true one
assigned to the j-th branch is wj, j=1,...,N. It is required that the N weights add up to unity.
Then, the expected value of Pij(A>a) once all branches have been accounted for, Pi(A>a), is given by:
(1)
*.dat file that will have the base name of the file that described the combination but with the extension *.dat.
This new hazard model can be loaded into CRISIS and the corresponding hazard results can be analyzed
with CRISIS (hazard maps, exceedance probability curves, uniform hazard spectra) as if they were the
results of a regular *.dat file. Disaggregation results, however, can not be obtained for the hazard resulting
from the logic-tree combination.
- 69 -
- 70 -
Chapter 4
4. Output Files
Upon the users selection, CRISIS can generate several output files. The possible output files are:
4.10 Results file (*.res)

This file contains a printout of the name of the run, the values assigned to the variables, characteristics of the
attenuation models, geometrical and seismicity description of the sources, the data defining the computation
grid, etc. It also gives a summary of the computations for each site, indicating which sources are of interest
the site and which sources were skipped. The computer times are also written.
If the users chooses to do so, this file also gives the final results, that is, exceedance probabilities for each
time frame, site and type of intensity. See an example of the *.res file
4.10 Graphics file (*.gra)

This file contains a brief identification header, and the exceedance rates for the types and levels of intensity
requested. This file can be used as input file to plot intensity versus exceedance rate curves. See an example
of the *.gra file.
4.10 Source by source results (*.fue)

This file contains exceedance probabilities by source, for each site and time frame. See an example of the
*.fue file.
Additionally, CRISIS will generate binary files (one for each intensity measure used in the analysis) to be
able to generate its own maps.
4.10 Map file (*.map)

This file contains intensity levels for the fixed return periodsentered in the data file, and for each type of
intensity, time frameand site. It also gives the coordinates of each site. This file can be used to generate
contour or 3D maps of intensity levels associated to constant exceedance rates. See an example of the *.map
file.
4.10 M-R Disaggregation file (*.des)

This file contains results of seismic hazard disaggregation, as a function of magnitude and distance, for
given intensity levels, return periods, time frames and epsilon values. These disaggregated results indicate
which combinations of magnitude and distance contribute more to the seismic hazard at a site, for a given
intensity measure, time frame, and return period. See an example of the *.des file. The combinations of
parameters for whichdisaggregation is desired are given in the Batch Disaggregation screen.
4.10 Maximum earthquakes file (*.smx)

This file contains information about the maximum possible intensity values at each site. For a given site,
these values are computed using the worst combination of distance to a source and expected value of
- 71 -
Chapter 4
maximum magnitude. The highest intensity computed for all sources is reported in this file, for different
values of epsilon. See details and an example of the *.smx file.
4.10 Example DES file

This file contains results of seismic hazard disaggregation, as a function of magnitude and distance. These
disaggregated results indicate which combinations of magnitude and distance contribute more to the seismic
hazard at a site, for a given intensity measure, time frame, epsilon value and intensity level, measured with a
given equivalent return period.
The following file contains disaggregation results for one site, one intensity measure (Intensity 1), one time
frame (Time frame 1), an equivalent return period of 475 years -associated, in this case to an intenisty level
of 66.3, and epsilon=1. The chart was constructed for 10 magnitudes and 10 distances.
Site: -96.4024 19.7250

Intensity 1, Time frame 1, Return period=475, Intensity level=6.63E+01, Nbin=0, Epsilon=1,Probability
theory=9.991E-02,Q=4.98E-01
Distance class marks
0.000 22.222 44.444 66.667 88.889 111.111 133.333 1
4.000 2.7327E-05 1.5740E-05 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00 0.0
4.444 3.9338E-05 1.4185E-04 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00 0.0
4.889 9.4605E-06 1.1587E-04 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00 0.0
5.333 8.3647E-06 2.9080E-04 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00 0.0
Magnitude
class marks 5.778 1.6597E-06 1.4571E-04 1.9845E-06 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00 0.0
6.222 1.4378E-06 1.8297E-04 3.3678E-05 9.4434E-07 0.0000E+00 0.0000E+00 0.0000E+00 0.0
6.667 0.0000E+00 8.1256E-07 1.1676E-06 5.4940E-07 1.6995E-09 0.0000E+00 0.0000E+00 0.0
7.111 0.0000E+00 3.3237E-07 6.3414E-07 5.9588E-07 1.9163E-07 2.4454E-10 0.0000E+00 0.0
7.556 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00 0.0
8.000 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00 0.0
Using results of this file, disaggregation graphs, like the one shown in the following figure, can be
constructed:
- 72 -
Chapter 4
From this graph, for instance, it is clear that most of the hazard comes from earthquakes of M between 5.33
and 5.78 located at between 22 and 44 Km of focal distance from the site.
4.11 Example FUE file

This file gives of the contribution of each source to the exceedance probabilities at a site. In the example
below, for a single site, values of probability of exceedance generated by each of the 2 sources present in
the model are shown, for 2 different intensity measures, ten different intensity values and three different
rime frames.
The user should remember that probabilities are not additive. This means that, for the same site, intensity
measure, intensity level and time frame, the addition of the exceedance probabilities associated to all the
individual sources does not give the total probability of exceedance. Please see the basic theoretical
background in order to know how the arithmetic of exceedance probabilities works.
Site: -0.2 -0.25

Intensity 1 T=0.050 Time frame 1 Tf=10.000
Level Region 01 Region 02
1.00E+02 5.64E-01 5.64E-01
1.54E+02 2.21E-01 2.21E-01
2.39E+02 4.98E-02 4.98E-02
3.68E+02 6.34E-03 6.34E-03
5.69E+02 4.39E-04 4.39E-04
8.79E+02 1.64E-05 1.64E-05
1.36E+03 3.48E-07 3.48E-07
- 73 -
Chapter 4
2.10E+03 4.83E-09 4.83E-09

3.24E+03 4.99E-11 4.99E-11
5.00E+03 4.05E-13 4.01E-13
1.00E+02 8.74E-01 8.74E-01
1.54E+02 4.64E-01 4.64E-01
2.39E+02 1.20E-01 1.20E-01
3.68E+02 1.58E-02 1.58E-02
5.69E+02 1.10E-03 1.10E-03
8.79E+02 4.10E-05 4.10E-05
1.36E+03 8.71E-07 8.71E-07
2.10E+03 1.21E-08 1.21E-08
3.24E+03 1.25E-10 1.25E-10
5.00E+03 1.01E-12 1.00E-12
1.00E+02 9.84E-01 9.84E-01
1.54E+02 7.12E-01 7.13E-01
2.39E+02 2.25E-01 2.26E-01
3.68E+02 3.13E-02 3.13E-02
5.69E+02 2.19E-03 2.19E-03
8.79E+02 8.18E-05 8.19E-05
1.36E+03 1.74E-06 1.74E-06
2.10E+03 2.41E-08 2.42E-08
3.24E+03 2.49E-10 2.50E-10
5.00E+03 1.99E-12 2.00E-12
1.00E+02 8.60E-01 8.60E-01
1.62E+02 5.27E-01 5.27E-01
2.61E+02 1.92E-01 1.92E-01
4.22E+02 3.95E-02 3.95E-02
6.81E+02 4.46E-03 4.46E-03
1.10E+03 2.62E-04 2.62E-04
1.78E+03 7.68E-06 7.68E-06
2.87E+03 1.13E-07 1.13E-07
4.64E+03 9.20E-10 9.20E-10
7.50E+03 5.02E-12 5.02E-12
- 74 -
Chapter 4

1.00E+02 9.93E-01 9.93E-01
1.62E+02 8.46E-01 8.46E-01
2.61E+02 4.13E-01 4.13E-01
4.22E+02 9.58E-02 9.58E-02
6.81E+02 1.11E-02 1.11E-02
1.10E+03 6.55E-04 6.55E-04
1.78E+03 1.92E-05 1.92E-05
2.87E+03 2.82E-07 2.82E-07
4.64E+03 2.30E-09 2.30E-09
7.50E+03 1.25E-11 1.25E-11
1.00E+02 1.00E+00 1.00E+00
1.62E+02 9.76E-01 9.76E-01
2.61E+02 6.55E-01 6.55E-01
4.22E+02 1.82E-01 1.82E-01
6.81E+02 2.21E-02 2.21E-02
1.10E+03 1.31E-03 1.31E-03
1.78E+03 3.84E-05 3.84E-05
2.87E+03 5.64E-07 5.65E-07
4.64E+03 4.60E-09 4.60E-09
7.50E+03 2.50E-11 2.51E-11
4.12 Example GRA file

This files gives hazard results in terms of probabilities of exceeding intensity values in different time frames.
In the following example we show the portion of the file that corresponds to only one site. In the cases of
computations for several sites, blocks of data similar to the one presented will be written for each site.
The example presented was computed for 3 intensity measures, each with 10 intensity levels, and for 3
different time frames.
- 75 -
Chapter 4
Example gra file

Version
CRISIS 2009 Program name and version
2.8.0.0
File
17/12/2009 00:00 03:58:55 a.m. Date and time of the run
header
Prueba No Poissoniana Name of the run
Site: -0.2 -0.25 Site coordinates

This block gives hazard
Intensity 1 T=0.050
values for the same site
Level Time Frame 1 Time Frame 2 Time Frame 3
Probability of exceeding the
level of intensity number 1
1.00E+02 8.10E-01 9.84E-01 1.00E+00
indicated in the first column,
in three different time frames
1.54E+02 3.93E-01 7.13E-01 9.17E-01
2.39E+02 9.72E-02 2.26E-01 4.00E-01
3.68E+02 1.26E-02 3.13E-02 6.16E-02
5.69E+02 8.78E-04 2.19E-03 4.38E-03
8.79E+02 3.28E-05 8.19E-05 1.64E-04
1.36E+03 6.97E-07 1.74E-06 3.48E-06
2.10E+03 9.66E-09 2.42E-08 4.82E-08
3.24E+03 9.99E-11 2.50E-10 4.98E-10
5.00E+03 8.06E-13 2.01E-12 3.99E-12
Intensity 2 T=0.150
1.00E+02 9.80E-01 1.00E+00 1.00E+00
1.62E+02 7.77E-01 9.76E-01 9.99E-01
2.61E+02 3.47E-01 6.55E-01 8.81E-01
4.22E+02 7.74E-02 1.82E-01 3.32E-01
6.81E+02 8.90E-03 2.21E-02 4.37E-02
1.10E+03 5.24E-04 1.31E-03 2.62E-03
- 76 -
Chapter 4
Example gra file

1.78E+03 1.54E-05 3.84E-05 7.68E-05
2.87E+03 2.26E-07 5.65E-07 1.13E-06
4.64E+03 1.84E-09 4.60E-09 9.20E-09
7.50E+03 1.00E-11 2.51E-11 5.01E-11
Intensity 3 T=0.300
1.00E+02 9.62E-01 1.00E+00 1.00E+00
1.62E+02 7.33E-01 9.63E-01 9.99E-01
2.61E+02 3.38E-01 6.43E-01 8.72E-01
4.22E+02 8.60E-02 2.01E-01 3.62E-01
6.81E+02 1.23E-02 3.04E-02 5.99E-02
1.10E+03 9.69E-04 2.42E-03 4.84E-03
1.78E+03 4.03E-05 1.01E-04 2.02E-04
2.87E+03 8.43E-07 2.11E-06 4.22E-06
4.64E+03 8.60E-09 2.15E-08 4.30E-08
7.50E+03 4.31E-11 1.08E-10 2.15E-10
4.12 Example MAP file

There is again the possibility of generating a *.map file. This file gives, for each computation site, the
intensities associated to 5 different fixed equivalent return periods. For a non- Poisson process, the
equivalent return period is the one that, in a truly Poisson process, would yield a given exceedance
probability in a given time frame. For instance, an equivalent return period of 100 years corresponds, in a
non-Poisson process, to an exceedance probability of 0.00995 in one year and 0.632 in 100 years. The
equivalent return periods are given in the Global Parameters screen.
An example follows:
Time frame=50 years Time frame=100 year

Return periods
RP 1.00E+02 2.50E+02 5.00E+02 1.00E+03 2.50E+03 1.00E+02 2.50E+02 5.00E+02 1.00E+0
PE 3.93E-01 1.81E-01 9.52E-02 4.88E-02 1.98E-02 6.32E-01 3.30E-01 1.81E-01 9.52E-0
Probabilities of exceedance
NBIN
Intensity
1 6.95E+01 9.80E+01 1.23E+02 1.51E+02 1.92E+02 00000 6.95E+01 9.80E+01 1.23E+02 1.51E+0
measure 1
Intensity
2 1.39E+02 2.07E+02 2.68E+02 3.37E+02 4.42E+02 00000 1.39E+02 2.07E+02 2.68E+02 3.37E+0
measure 2
Intensity 3 4.87E+01 8.03E+01 1.11E+02 1.48E+02 2.05E+02 00000 4.87E+01 8.03E+01 1.11E+02 1.48E+0
- 77 -
Chapter 4
measure 3
1 1.36E+02 1.94E+02 2.46E+02 3.04E+02 3.90E+02 00000 1.36E+02 1.94E+02 2.45E+02 3.04E+0
2 2.41E+02 3.57E+02 4.63E+02 5.84E+02 7.68E+02 00000 2.41E+02 3.57E+02 4.63E+02 5.84E+0
3 7.17E+01 1.16E+02 1.58E+02 2.07E+02 2.84E+02 00000 7.16E+01 1.16E+02 1.58E+02 2.07E+0
1 6.88E+01 9.70E+01 1.22E+02 1.49E+02 1.90E+02 00000 6.88E+01 9.70E+01 1.22E+02 1.49E+0
2 1.38E+02 2.05E+02 2.66E+02 3.35E+02 4.39E+02 00000 1.38E+02 2.05E+02 2.66E+02 3.35E+0
3 4.84E+01 7.99E+01 1.11E+02 1.47E+02 2.04E+02 00000 4.84E+01 7.99E+01 1.11E+02 1.47E+0
1 1.35E+02 1.93E+02 2.45E+02 3.02E+02 3.88E+02 00000 1.35E+02 1.93E+02 2.45E+02 3.02E+0
2 2.41E+02 3.57E+02 4.62E+02 5.83E+02 7.66E+02 00000 2.41E+02 3.57E+02 4.62E+02 5.83E+0
3 7.17E+01 1.16E+02 1.58E+02 2.07E+02 2.84E+02 00000 7.17E+01 1.16E+02 1.58E+02 2.07E+0
1 2.28E+02 3.12E+02 3.85E+02 4.66E+02 5.88E+02 00000 2.28E+02 3.12E+02 3.85E+02 4.66E+0
2 3.99E+02 5.67E+02 7.15E+02 8.81E+02 1.13E+03 00000 3.99E+02 5.67E+02 7.15E+02 8.81E+0
3 1.12E+02 1.72E+02 2.28E+02 2.93E+02 3.91E+02 00000 1.12E+02 1.72E+02 2.28E+02 2.93E+0
1 1.36E+02 1.94E+02 2.45E+02 3.03E+02 3.89E+02 00000 1.36E+02 1.94E+02 2.45E+02 3.03E+0
2 2.41E+02 3.57E+02 4.63E+02 5.84E+02 7.67E+02 00000 2.41E+02 3.57E+02 4.63E+02 5.84E+0
3 7.18E+01 1.16E+02 1.58E+02 2.07E+02 2.84E+02 00000 7.18E+01 1.16E+02 1.58E+02 2.07E+0
1 6.91E+01 9.74E+01 1.22E+02 1.50E+02 1.91E+02 00000 6.91E+01 9.74E+01 1.22E+02 1.50E+0
2 1.39E+02 2.06E+02 2.66E+02 3.36E+02 4.40E+02 00000 1.39E+02 2.06E+02 2.66E+02 3.36E+0
3 4.85E+01 8.01E+01 1.11E+02 1.47E+02 2.04E+02 00000 4.85E+01 8.01E+01 1.11E+02 1.47E+0
1 1.35E+02 1.93E+02 2.45E+02 3.02E+02 3.88E+02 00000 1.35E+02 1.93E+02 2.44E+02 3.02E+0
2 2.41E+02 3.56E+02 4.62E+02 5.82E+02 7.65E+02 00000 2.41E+02 3.56E+02 4.62E+02 5.82E+0
3 7.17E+01 1.16E+02 1.58E+02 2.07E+02 2.84E+02 00000 7.17E+01 1.16E+02 1.58E+02 2.07E+0
1 6.94E+01 9.78E+01 1.23E+02 1.51E+02 1.91E+02 00000 6.94E+01 9.78E+01 1.23E+02 1.51E+0
2 1.39E+02 2.06E+02 2.67E+02 3.37E+02 4.41E+02 00000 1.39E+02 2.06E+02 2.67E+02 3.37E+0
3 4.86E+01 8.02E+01 1.11E+02 1.48E+02 2.05E+02 00000 4.86E+01 8.02E+01 1.11E+02 1.48E+0
In the example above, results are given for nine sites, three different intensity measures, five equivalent
return periods and two time frames. The relation between probability of exceedance, time frame and
equivalent return period can be found here.
The meaning of the NBIN flags is the following
If NNP=0 then interpolation was executed for the corresponding return period.
If NNP=1 then extrapolation was executed for the corresponding return period.
- 78 -
Chapter 4
If NNP=2 the value was impossible to calculate
4.1 Example RES file

The following is an example of the *.res file. Basically, it contains all the data given by the user to define
source geometry, source seismicity and attenuation characteristics, as well as other global parameters.
*********************************************
CRISIS 2009 Version 3.4.2.0
15/04/2010 07:33:20 p.m.
La Arbolada-Jalisco
VALUES OF PARAMETERS FOR THE PRESENT PROJECT

Number of regions or seismic sources: 45
Number of attenuation models: 3
Number of structural periods: 9
Number of intensity levels: 15
Number of magnitudes for integration: 9
Type of computation sites: Lista
Max. dist. of integration: 500.000
Min. distance/Triangle size ratio 5.000
Minimum triangle size (km) 7.000
INTENSITIES
I T(I) A0(I) AU(I) UNITS
1 1.00E-02 1.00E+00 1.50E+03
- 79 -
Chapter 4
2 1.50E-01 1.00E+00 4.00E+03

3 3.00E-01 1.00E+00 3.30E+03
4 5.00E-01 1.00E+00 2.20E+03
5 1.00E+00 1.00E+00 1.30E+03
6 2.00E+00 1.00E+00 7.50E+02
7 3.00E+00 1.00E+00 6.00E+02
8 4.00E+00 1.00E+00 4.00E+02
9 5.00E+00 1.00E+00 3.00E+02
TIME FRAMES
I TF(I)
1 5.00E+01
INITIAL GRID OF POINTS

File of list of sites: C:\Crisis 2008 Extra\Pruebas\Mxico\CD_3Ciudades.TXT
THE INITIAL GRID WAS MODIFIED WITH THE FOLLOWING POLYGONS:

Number of polygons: 1
Polygon 1. Number of vertex: 16

LONG LAT
-117.7324 32.7306
-116.4633 27.6343
-113.8458 24.3694
-110.0386 21.7417
-104.1692 17.0435
- 80 -
Chapter 4
-97.5065 14.1769
-89.8128 13.7787
-86.1642 18.9546
-85.6883 23.2546
-90.7646 22.9361
-94.3339 21.0250
-95.8409 27.3157
-100.5999 31.1380
-105.8348 32.6509
-111.9422 33.4472
-117.0185 33.9250
ATTENUATION MODELS
--------------------------------------------------------------------------------------------------------
Model: 1 ATCOSTAm_Trunc
Of class: Crisis2008.NewAttenuation.AttenuationClasses.AttenuationTable
Brief description: Not available
Original units: Not available
Dimension: Acceleration
Spectral period range: 0.005 to 6
Valid distance range: 5 to 500
Valid magnitude range: 4 to 8.5
Type of distance metric: Focal
Residuals distribution: LogNormal
Parameter: Units coefficient=1
Parameter: Attenuation Table=C:\Crisis 2008 Extra\Pruebas\Mxico\ATCOSTAm_Trunc.ATN
--------------------------------------------------------------------------------------------------------
- 81 -
Chapter 4
--------------------------------------------------------------------------------------------------------
Model: 2 Abrahamson y Silva No HW S/S
Of class: Crisis2008.NewAttenuation.AttenuationClasses.AbrahamsonAndSilva97
Brief description: Model by Abrahamon and Silva...
Original units: cm/s/s
Valid distance range: 0.1 to 200
Type of distance metric: Rrup
Parameter: Site is in the hanging wall=False
Parameter: Sigma truncation=-1
Parameter: Soil Type=0
Parameter: Style of fault=2
--------------------------------------------------------------------------------------------------------
--------------------------------------------------------------------------------------------------------
Model: 3 NormalDaniel_RRup_5_Trunc_CR2007
Of class: Crisis2008.NewAttenuation.AttenuationClasses.AttenuationTable
Brief description: Not available
Original units: Not available
Valid distance range: 0.01 to 500
Type of distance metric: Focal
- 82 -
Chapter 4

Parameter: Attenuation Table=C:\Crisis 2008 Extra\Pruebas\Mxico\NormalDaniel_RRup_5_Trunc_
CR2007.atn
--------------------------------------------------------------------------------------------------------
PROPERTIES OF THE SOURCES
----------
REGION: 1 Baja California intraplaca norte
Gutenberg-Richter
SOURCE IS ACTIVE
Base Attenuation model: Abrahamson y Silva No HW S/S
Area source
Number of vertex : 10
Long Lat Depth(km)
-119.2500 34.5600 10.0000
-118.1250 33.0600 10.0000
-117.0000 31.5600 10.0000
-115.8750 30.0600 10.0000
-114.7500 28.5600 10.0000
-113.5800 29.4400 10.0000
-114.7050 30.9400 10.0000
-115.8300 32.4400 10.0000
-116.9950 33.9400 10.0000
-118.0800 35.4400 10.0000
----------
- 83 -
Chapter 4
REGION: 2 Baja California intraplaca sur

Gutenberg-Richter
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-114.7500 28.5600 10.0000
-113.6250 27.0600 10.0000
-112.5000 25.5600 10.0000
-111.3750 24.0600 10.0000
-110.2500 22.5600 10.0000
-109.0800 23.4400 10.0000
-110.2050 24.9400 10.0000
-111.3300 26.4400 10.0000
-112.4550 27.9400 10.0000
-113.5800 29.4400 10.0000
----------
REGION: 3 Baja California interplaca norte
Gutenberg-Richter
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-117.0000 33.9900 7.0000
-116.3050 33.0680 7.0000
-115.6100 32.1450 7.0000
- 84 -
Chapter 4
-114.9150 31.2230 7.0000

-114.2200 30.3000 7.0000
-113.5300 30.8200 7.0000
-114.2230 31.7430 7.0000
-114.9150 32.6650 7.0000
-115.6080 33.5880 7.0000
-116.3000 34.5100 7.0000
----------
REGION: 4 Baja California interplaca centro
Gutenberg-Richter
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-114.2200 30.3000 10.0000
-113.3600 29.1550 10.0000
-112.5000 28.0100 10.0000
-111.8100 28.5300 10.0000
-112.6700 29.6750 10.0000
-113.5300 30.8200 10.0000
----------
REGION: 5 Baja California interplaca sur
Gutenberg-Richter
SOURCE IS ACTIVE
Area source
- 85 -
Chapter 4
Long Lat Depth(km)
-108.0000 22.0000 10.0000
-107.1000 23.0000 10.0000
-108.6500 25.0000 10.0000
-110.5500 27.6100 10.0000
-111.8600 28.4800 10.0000
-112.5000 28.0000 10.0000
-110.2500 25.0000 10.0000
----------
REGION: 6 Sierra Madre Occidental
Gutenberg-Richter
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-110.0000 29.0000 20.0000
-110.5500 27.6100 20.0000
-108.6500 25.0000 20.0000
-106.0000 25.0000 20.0000
-106.0000 29.0000 20.0000
----------
REGION: 7 Cuencas y Sierras
Gutenberg-Richter
SOURCE IS ACTIVE
- 86 -
Chapter 4
Area source
Long Lat Depth(km)
-110.0000 34.0000 20.0000
-110.0000 29.0000 20.0000
-106.0000 29.0000 20.0000
-106.0000 25.0000 20.0000
-104.0000 25.0000 20.0000
-104.0000 29.0000 20.0000
-105.0000 34.0000 20.0000
----------
REGION: 8 Cuenca de Burgos
Gutenberg-Richter
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-104.0000 27.5000 20.0000
-104.0000 22.0000 20.0000
-99.5000 22.0000 20.0000
-99.5000 27.5000 20.0000
----------
REGION: 9 Interfaz Pacfico-Rivera
Gutenberg-Richter
SOURCE IS ACTIVE
- 87 -
Chapter 4
Area source
Long Lat Depth(km)
-110.0000 18.5000 5.0000
-106.0000 18.5000 5.0000
-106.0000 20.0000 5.0000
-108.0000 20.0000 5.0000
-108.0000 22.0000 5.0000
-110.0000 22.0000 5.0000
----------
REGION: 10 Sismicidad difusa 1
Gutenberg-Richter
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-105.0000 34.0000 20.0000
-104.0000 29.0000 20.0000
-104.0000 27.5000 20.0000
-99.5000 27.5000 20.0000
-99.5000 22.0000 20.0000
-104.0000 22.0000 20.0000
-104.0000 25.0000 20.0000
-108.6500 25.0000 20.0000
-106.0000 21.5000 20.0000
-105.5000 20.0000 20.0000
-105.0000 21.0000 20.0000
- 88 -
Chapter 4
-100.2800 20.5400 20.0000

-96.5000 20.2000 20.0000
-97.0000 34.0000 20.0000
----------
REGION: 11 Sismicidad difusa 2
Gutenberg-Richter
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-116.3000 34.5100 20.0000
-113.5300 30.8200 20.0000
-111.8100 28.5300 20.0000
-110.5500 27.6100 20.0000
-110.0000 29.0000 20.0000
-110.0000 34.0000 20.0000
----------
REGION: 12 Centroamrica
Gutenberg-Richter
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-88.2000 15.8000 5.0000
-85.3000 15.8000 5.0000
- 89 -
Chapter 4
-85.3000 17.0000 5.0000

-88.2000 17.0000 5.0000
----------
REGION: 13 Subduccin Chiapas
Characteristic model
SOURCE IS ACTIVE
Base Attenuation model: ATCOSTAm_Trunc
Area source
Long Lat Depth(km)
-94.0180 14.5270 15.0000
-92.6670 13.6200 15.0000
-92.3010 14.0690 30.0000
-93.6130 15.1000 30.0000
----------
REGION: 14 Subduccin Brecha de Tehuantepec
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-95.0000 15.1970 15.0000
-94.0180 14.5270 15.0000
-93.6130 15.1000 30.0000
-93.9870 15.3920 30.0000
-95.0000 15.9100 30.0000
- 90 -
Chapter 4
----------
REGION: 15 Subduccin Oaxaca Este
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-96.3490 15.5260 15.0000
-96.0000 15.5000 15.0000
-95.0000 15.1970 15.0000
-95.0000 15.9100 30.0000
-96.2670 16.2570 30.0000
----------
REGION: 16 Subduccin Oaxaca 1
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-97.3540 15.6700 15.0000
-96.3490 15.5260 15.0000
-96.2670 16.2570 30.0000
-97.2100 16.4430 30.0000
----------
- 91 -
Chapter 4
REGION: 17 Subduccin Oaxaca 2

SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-97.8770 15.7750 15.0000
-97.3540 15.6700 15.0000
-97.2100 16.4430 30.0000
-97.6550 16.5280 30.0000
----------
REGION: 18 Subduccin Oaxaca Oeste
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-98.2420 15.8730 15.0000
-97.8870 15.7750 15.0000
-97.6550 16.5280 30.0000
-98.0350 16.5910 30.0000
----------
REGION: 19 Subduccin Ometepec
SOURCE IS ACTIVE
- 92 -
Chapter 4

Area source
Long Lat Depth(km)
-99.6630 16.3430 15.0000
-98.2420 15.8730 15.0000
-98.0350 16.5910 30.0000
-99.4380 17.0100 30.0000
----------
REGION: 20 Subduccin San Marcos
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-100.0620 16.4730 15.0000
-99.6630 16.3430 15.0000
-99.4380 17.0100 30.0000
-99.8360 17.1430 30.0000
----------
REGION: 21 Subduccin Guerrero
SOURCE IS ACTIVE
Area source
- 93 -
Chapter 4
Long Lat Depth(km)

-101.0050 16.7970 15.0000
-100.0620 16.4730 15.0000
-99.8360 17.1430 30.0000
-100.8100 17.4370 30.0000
----------
REGION: 22 Subduccin Petatln
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-101.7010 17.0440 15.0000
-101.0050 16.7970 15.0000
-100.8100 17.4370 30.0000
-101.4990 17.7090 30.0000
----------
REGION: 23 Subduccin Michoacn
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-103.1350 17.7010 15.0000
-101.7010 17.0440 15.0000
- 94 -
Chapter 4
-101.4990 17.7090 30.0000

-102.9160 18.3930 30.0000
----------
REGION: 24 Subduccin Colima 1
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-103.8680 18.3000 15.0000
-103.1350 17.7010 15.0000
-102.9160 18.3930 30.0000
-103.6120 18.7610 30.0000
----------
REGION: 25 Subduccin Brecha de Colima
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-104.4570 18.7170 15.0000
-103.8680 18.3000 15.0000
-103.6120 18.7610 30.0000
-104.1820 19.1730 30.0000
- 95 -
Chapter 4
----------
REGION: 26 Subduccin Jalisco
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-104.4570 18.7170 15.0000
-104.1820 19.1730 15.0000
-105.0000 20.0000 15.0000
-105.1300 20.2560 15.0000
-105.5000 21.0000 30.0000
-106.0000 21.0000 30.0000
-105.7000 20.0000 30.0000
-105.0000 19.1000 30.0000
----------
REGION: 27 Jalisco nuevo
Gutenberg-Richter
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-106.0000 21.0000 15.0000
-105.7000 20.0000 15.0000
-105.0000 19.1000 15.0000
-104.0000 18.4000 15.0000
- 96 -
Chapter 4
-103.7300 18.8400 30.0000

-104.0000 19.0000 30.0000
-105.0000 20.0000 30.0000
-105.5000 21.0000 30.0000
----------
REGION: 28 Gro-Mich nuevo
Gutenberg-Richter
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-104.0000 18.4000 15.0000
-103.0000 17.6000 15.0000
-102.0000 17.1500 15.0000
-101.0000 16.8000 15.0000
-100.0000 16.4500 15.0000
-99.0000 16.1000 15.0000
-98.8200 16.8100 30.0000
-100.0000 17.2000 30.0000
-101.0000 17.5000 30.0000
-102.0000 17.9000 30.0000
-103.7300 18.8400 30.0000
----------
REGION: 29 Oaxaca nuevo
Gutenberg-Richter
SOURCE IS ACTIVE
- 97 -
Chapter 4

Area source
Long Lat Depth(km)
-99.0000 16.1000 15.0000
-98.0000 15.8000 15.0000
-97.0000 15.6000 15.0000
-96.0000 15.5000 15.0000
-95.0000 15.2000 15.0000
-95.0000 15.9000 30.0000
-96.0000 16.2000 30.0000
-97.0000 16.4000 30.0000
-98.0000 16.6000 30.0000
-98.8200 16.8100 30.0000
----------
REGION: 30 Chiapas nuevo
Gutenberg-Richter
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-95.0000 15.2000 15.0000
-92.6700 13.6200 15.0000
-92.3000 14.0800 30.0000
-94.0000 15.4000 30.0000
-95.0000 15.9000 30.0000
- 98 -
Chapter 4
----------
REGION: 31 Prof. Interm Oeste nueva
Gutenberg-Richter
SOURCE IS ACTIVE
Base Attenuation model: NormalDaniel_RRup_5_Trunc_CR2007
Area source
Long Lat Depth(km)
-105.5000 21.0000 30.0000
-105.0000 20.0000 30.0000
-104.0000 19.0000 30.0000
-102.0000 17.9000 30.0000
-101.0000 17.5000 30.0000
-100.0000 17.2000 30.0000
-99.0000 16.8500 30.0000
-99.0000 19.1000 100.0000
-100.0000 19.2000 100.0000
-101.0000 19.1000 100.0000
-102.0000 19.2000 100.0000
-103.3000 19.3000 100.0000
-104.0000 19.8000 100.0000
-105.0000 21.0000 100.0000
----------
REGION: 32 Prof. int. centro nueva
Gutenberg-Richter
SOURCE IS ACTIVE
Area source
- 99 -
Chapter 4
Long Lat Depth(km)
-99.0000 16.8500 30.0000
-98.0000 16.6000 30.0000
-97.0000 16.4000 30.0000
-96.0000 16.2000 30.0000
-95.0000 15.9000 30.0000
-95.0000 17.2000 100.0000
-96.0000 18.3000 100.0000
-97.0000 18.8000 100.0000
-98.0000 18.9000 100.0000
-99.0000 19.1000 100.0000
----------
REGION: 33 Prof. int. Este nueva
Gutenberg-Richter
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-95.0000 15.9000 30.0000
-94.0000 15.4000 30.0000
-92.3000 14.0800 30.0000
-91.5000 14.9000 100.0000
-94.0000 16.5000 100.0000
-95.0000 17.2000 100.0000
----------
- 100 -
Chapter 4
REGION: 34 Petrolera
Gutenberg-Richter
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-92.5000 18.5000 30.0000
-92.0000 19.0000 30.0000
-93.0000 19.1250 30.0000
-94.0000 19.2500 30.0000
-96.0000 19.5000 30.0000
-96.0000 18.5000 30.0000
-95.5000 18.0000 30.0000
-95.0000 17.5000 30.0000
-94.0000 17.5000 30.0000
-93.0000 17.5000 30.0000
----------
REGION: 35 Golfo
Gutenberg-Richter
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-96.5000 21.0000 15.0000
-96.5000 20.2000 15.0000
-96.0000 19.5000 15.0000
- 101 -
Chapter 4
-92.0000 19.0000 15.0000

-91.0000 21.0000 15.0000
-94.0000 21.0000 15.0000
----------
REGION: 36 Eje volcnico
Gutenberg-Richter
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-105.5000 20.0000 15.0000
-103.1500 18.5000 15.0000
-99.0000 18.5000 15.0000
-96.0000 18.5000 15.0000
-96.0000 19.5000 15.0000
-96.5000 20.2000 15.0000
-100.2800 20.5400 15.0000
-105.0000 21.0000 15.0000
----------
REGION: 37 Intraplaca
Gutenberg-Richter
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
- 102 -
Chapter 4
-103.1500 18.5000 15.0000

-97.0000 16.0000 15.0000
-95.0000 16.2000 15.0000
-94.0000 16.0000 15.0000
-93.0000 17.5000 15.0000
-96.0000 18.5000 15.0000
-99.0000 18.5000 15.0000
----------
REGION: 38 Chiapas Volcn
Gutenberg-Richter
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-93.0000 17.5000 15.0000
-94.0000 16.0000 15.0000
-91.8000 14.0000 15.0000
-89.0000 14.0000 15.0000
-88.0000 16.0000 15.0000
-90.2500 17.2500 15.0000
----------
REGION: 39 Profundos Chiapas
Gutenberg-Richter
SOURCE IS ACTIVE
Area source
- 103 -
Chapter 4
Long Lat Depth(km)
-96.0000 18.3000 100.0000
-95.0000 17.2000 100.0000
-94.0000 16.5000 100.0000
-91.9000 15.1500 100.0000
-91.6000 15.5000 200.0000
-93.3000 16.7500 200.0000
-95.0000 18.0000 200.0000
----------
REGION: 40 Motagua 1
Gutenberg-Richter
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-93.0000 15.2000 5.0000
-92.4000 14.4500 5.0000
-90.0000 14.5000 5.0000
-89.0000 14.5000 5.0000
-88.0000 15.1000 5.0000
-88.0000 16.3000 5.0000
-89.5500 15.2000 5.0000
----------
Gutenberg-Richter
- 104 -
Chapter 4
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-93.0000 15.2000 10.0000
-92.4000 14.4500 10.0000
-90.0000 14.5000 10.0000
-89.0000 14.5000 10.0000
-88.0000 15.1000 10.0000
-88.0000 16.3000 10.0000
-89.5500 15.2000 10.0000
----------
Gutenberg-Richter
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-93.0000 15.2000 15.0000
-92.4000 14.4500 15.0000
-90.0000 14.5000 15.0000
-89.0000 14.5000 15.0000
-88.0000 15.1000 15.0000
-88.0000 16.3000 15.0000
-89.5500 15.2000 15.0000
- 105 -
Chapter 4
----------
REGION: 43 Polochic 1
Gutenberg-Richter
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-94.0000 16.0000 5.0000
-93.0000 15.2000 5.0000
-89.5500 15.2000 5.0000
-88.0000 16.3000 5.0000
----------
Gutenberg-Richter
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-94.0000 16.0000 10.0000
-93.0000 15.2000 10.0000
-89.5500 15.2000 10.0000
-88.0000 16.3000 10.0000
----------
Gutenberg-Richter
- 106 -
Chapter 4
SOURCE IS ACTIVE
Area source
Long Lat Depth(km)
-94.0000 16.0000 15.0000
-93.0000 15.2000 15.0000
-89.5500 15.2000 15.0000
-88.0000 16.3000 15.0000
SEISMICITY
Gutenberg-Richter sources
Source name M0 Lambda0 E(Beta) c(Beta) E(Mu) D(Mu)

Baja California intraplaca nor 4.500 1.14E+00 0.970 0.097 5.800 0.000
Baja California intraplaca sur 4.500 1.21E+00 0.933 0.036 5.800 0.000
Baja California interplaca nor 4.500 2.51E+00 1.782 0.093 7.700 0.300
Baja California interplaca cen 4.500 7.26E-01 1.637 0.168 7.400 0.400
Baja California interplaca sur 4.500 2.09E+00 1.674 0.082 7.200 0.600
Sierra Madre Occidental 4.500 1.16E-01 2.880 0.030 5.600 0.000
Cuencas y Sierras 4.500 2.69E-01 2.880 0.030 5.600 0.000
Cuenca de Burgos 4.500 1.87E-01 2.880 0.030 5.600 0.000
Interfaz Pacfico-Rivera 4.500 3.41E+00 1.736 0.088 7.200 0.000
Sismicidad difusa 1 4.500 6.58E-01 2.880 0.030 5.600 0.000
Sismicidad difusa 2 4.500 1.80E-01 2.880 0.030 5.600 0.000
Centroamrica 4.500 4.97E-01 1.942 0.180 7.700 0.300
Jalisco nuevo 4.500 2.01E+00 1.827 0.110 7.200 0.000
- 107 -
Chapter 4
Gro-Mich nuevo 4.500 4.79E+00 1.547 0.077 7.200 0.000

Oaxaca nuevo 4.500 6.72E+00 1.847 0.063 7.200 0.000
Chiapas nuevo 4.500 1.89E+01 2.059 0.037 7.200 0.000
Prof. Interm Oeste nueva 4.500 2.16E+00 1.699 0.097 7.800 0.200
Prof. int. centro nueva 4.500 1.71E+00 1.576 0.110 7.900 0.200
Prof. int. Este nueva 4.500 2.78E+00 1.761 0.087 7.800 0.200
Petrolera 4.500 6.05E-01 3.050 0.209 6.700 0.500
Golfo 4.100 1.05E-01 2.704 0.459 6.500 0.500
Eje volcnico 4.500 2.49E-01 1.884 0.223 7.200 0.300
Intraplaca 4.500 1.44E+00 1.889 0.124 6.500 0.500
Chiapas Volcn 4.500 1.61E+00 2.005 0.119 7.000 0.200
Profundos Chiapas 4.500 2.52E+00 2.207 0.093 7.500 0.300
Motagua 1 5.000 2.77E-01 2.234 0.309 7.800 0.000
Motagua 2 5.000 2.77E-01 2.234 0.309 7.800 0.000
Motagua 3 5.000 2.77E-01 2.234 0.309 7.800 0.000
Polochic 1 5.000 1.20E-01 2.187 0.105 7.800 0.000
Polochic 2 5.000 1.20E-01 2.187 0.105 7.800 0.000
Polochic 3 5.000 1.20E-01 2.187 0.105 7.800 0.000
Characteristic model sources
Source name Med(T) T0 D F sM M0 Mu

Subduccin Chiapas 18.700 20.000 7.500 0.000 0.270 7.000 8.400
Subduccin Brecha de Tehuantep 24.700 200.000 7.500 0.000 0.270 7.000 8.400
Subduccin Oaxaca Este 24.800 26.000 7.500 0.000 0.270 7.000 8.400
Subduccin Oaxaca 1 39.400 13.000 7.500 0.000 0.270 7.000 8.400
Subduccin Oaxaca 2 77.900 63.000 7.500 0.000 0.270 7.000 8.400
Subduccin Oaxaca Oeste 104.700 23.000 7.500 0.000 0.270 7.000 8.400
- 108 -
Chapter 4
Subduccin Ometepec 26.700 9.000 7.500 0.000 0.270 7.000 8.400

Subduccin San Marcos 89.900 29.000 7.500 0.000 0.270 7.000 8.400
Subduccin Guerrero 39.700 80.000 7.500 0.000 0.270 7.000 8.400
Subduccin Petatln 52.600 12.000 7.500 0.000 0.270 7.000 8.400
Subduccin Michoacn 25.600 6.000 7.500 0.000 0.270 7.000 8.400
Subduccin Colima 1 47.500 18.000 7.500 0.000 0.270 7.000 8.400
Subduccin Brecha de Colima 56.700 183.000 7.500 0.000 0.270 7.000 8.400
Subduccin Jalisco 19.800 57.000 7.500 0.000 0.270 7.000 8.400
**********************************************************************
WARNINGS ABOUT MAGNITUDE-DISTANCE RANGES VALIDITY
Region 3, Baja California interplaca norte: The maximum magnitude in the region is larger than the
maximum valid magnitude of GMPE Abrahamson y Silva No HW S/S (7.7 > 7.5)
Region 12, Centroamrica: The maximum magnitude in the region is larger than the maximum valid
magnitude of GMPE Abrahamson y Silva No HW S/S (7.7 > 7.5)
Region 40, Motagua 1: The maximum magnitude in the region is larger than the maximum valid magnitude
of GMPE Abrahamson y Silva No HW S/S (7.8 > 7.5)
Region 43, Polochic 1: The maximum magnitude in the region is larger than the maximum valid magnitude
The integration distance Rmax is greater than the maximum valid distance of GMPE Abrahamson y Silva
No HW S/S (500 > 200)
**********************************************************************
- 109 -
Chapter 4
4.1 Example SMX file

The following table shows results for a single computation site:
Pct 50.00 Pct 84.134 Pct 97.725 Pct 99.865
Site: -96.4024 19.7250
IT Sa Mag R Src Sa Mag R Src Sa Mag R Src Sa Mag

1 4.250E+02 6.800 0.00 5 7.138E+02 6.800 0.00 5 1.199E+03 6.800 0.00 5 2.014E+03 6.800 0
2 4.791E+02 6.800 0.00 5 8.102E+02 6.800 0.00 5 1.370E+03 6.800 0.00 5 2.317E+03 6.800 0
3 5.712E+02 6.800 0.00 5 9.753E+02 6.800 0.00 5 1.666E+03 6.800 0.00 5 2.844E+03 6.800 0
4 7.051E+02 6.800 0.00 5 1.217E+03 6.800 0.00 5 2.101E+03 6.800 0.00 5 3.626E+03 6.800 0
5 8.055E+02 6.800 0.00 5 1.388E+03 6.800 0.00 5 2.394E+03 6.800 0.00 5 4.125E+03 6.800 0
6 9.191E+02 6.800 0.00 5 1.569E+03 6.800 0.00 5 2.679E+03 6.800 0.00 5 4.572E+03 6.800 0
7 9.225E+02 6.800 0.00 5 1.576E+03 6.800 0.00 5 2.694E+03 6.800 0.00 5 4.602E+03 6.800 0
8 7.176E+02 6.800 0.00 5 1.237E+03 6.800 0.00 5 2.132E+03 6.800 0.00 5 3.675E+03 6.800 0
9 6.000E+02 6.800 0.00 5 1.032E+03 6.800 0.00 5 1.777E+03 6.800 0.00 5 3.057E+03 6.800 0
10 2.860E+02 6.800 0.00 5 5.089E+02 6.800 0.00 5 9.058E+02 6.800 0.00 5 1.612E+03 6.800 0
11 1.598E+02 6.800 0.00 5 2.904E+02 6.800 0.00 5 5.281E+02 6.800 0.00 5 9.598E+02 6.800 0
12 7.734E+01 6.800 0.00 5 1.451E+02 6.800 0.00 5 2.721E+02 6.800 0.00 5 5.104E+02 6.800 0
13 3.152E+01 6.800 0.00 5 5.907E+01 6.800 0.00 5 1.107E+02 6.800 0.00 5 2.075E+02 6.800 0
Structural
period Results for Pct 50.00 Results for Pct 84.134 Results for Pct 97.725 Results for Pct 99
index
For each structural period, the table is showing the characteristics of the earthquake that produces the
highest accelerations for a given percentile (or epsilon value). Sa is the highest acceleration associated to the
appropriate percentile, Mag is the earthquake magnitude of the largest earthquake, R is the distance to the
site and Src is the index of the seismic region to which the dominating earthquake belongs.
- 110 -
Chapter 5
5. Seismicity models used

Speaking generally, CRISIS expects to have the seismicity described by means of the probabilities of
having 1, 2, ..., Ns earthquakes of given magnitudes, in a given location, during the next Tf years. As can be
noted, this is the most general description of seismicity that can possibly be given.
In order to get this information, CRISIS admits three types of seismicity models. The first two were already
contained in older CRISIS versions, and both are related to Poissonian occurrences, although they differ in
the way in which they define the earthquake magnitude exceedance rate. The third model is generalized
non-Poissonian model that gives explicitly the required probabilities. See details of the three models in the
following links.
Modified Gutenberg-Richter model
Characteristic-earthquake model
Generalized non-Poissonian model
Generalized Poissonian model
5.2 Characteristic earthquake

This model is associated to Poisson occurrences, so the probability of exceeding intensity level a in the next
Tf years, given that an earthquake with magnitude M took place at a distance R from the site, is given by:
(1)
where p1(a|M,R) is the exceedance probability of intensity level a, given that a magnitude M event occurred
at a distance R from the site, and Dl (M) is the Poissonian magnitude exceedance rate associated to the
magnitude range (also called magnitude bin) characterized by magnitude M. Note that p1 (a|M,R) depends
only on magnitude and site- to- hypocenter distance. This probability does not depend on earthquake
occurrence probabilities.
In turn, Dl(M) can be computed as
(2)
where it is implicit that the magnitude bin characterized by magnitude M goes from M- DM/2 to M+D M/2.
For the Characteristic-earthquake model, the earthquake magnitude exceedance rate is given by:
- 111 -
Chapter 5
(3)
where (.) is the standard normal cumulative function, M 0 and M u are the minimum and maximum
characteristic magnitudes, respectively, and EM and s are parameters defining the distribution of M. EM can
be interpreted as the expected value of the characteristic earthquake and s as its standard deviation. 0 is the
exceedance rate of magnitude M0.
In addition, an slip-predictable behavior can be modeled assuming that EM grows with time elapsed since
the last characteristic event, T00, in the following fashion:
(4)
Of course, if F is set to zero, then EM=D, independently of time elapsed.
5.3 Generalized Non-Poissonian model

This type of seismicity description allows for direct specification of the required probabilities, that is, the
probabilities of having 1, 2, ..., Ns earthquakes of given magnitudes, in a given location, during the next Tf
years.
This information is given to CRISIS by means of a binary file, with extension nps (non-Poisson seismicity),
which has the following format:
Generalized seismicity file

Description Variable Type Length Comments
Number of point sources TotSrc Integer 4
Number of magnitude bins Nbin Integer 4
Number of time frames Nt Integer 4
Maximum number of events
Ns Integer 4
for which Prob(i,j) is given
Magnitude representative of
M(1) Double 8 Magnitude values are useful onl
bin 1
parametric attenuation models a

used. They are not used in
Magnitude representative of generalized attenuation models
M(Nbin) Double 8
bin Nbin
Time frame 1 Tf(1) Double 8
- 112 -
Chapter 5

Time frame Nt Tf(Nt) Double 8
Seismicity
Seismicity record for source 1 Seis(1) 8+8*Ns*Nt*Nbin
record
Seismicity
Seismicity record for source 2 Seis(2) 8+8*Ns*Nt*Nbin
record

Seismicity record for Seismicity
Seis(TotSrc) 8+8*Ns*Nt*Nbin
source TotSrc record
Seismicity record
Variable Type Length
Probability of Prob(1,1,1) Double 8
having 1, Prob(1,1,2) Double 8
2,,Ns Block associated
...
events of to Magnitude 1
Prob(1,1,Ns) Double 8
magnitude 1
Probability of
in time frame Prob(1,2,1) Double 8
having
1 1,
2,,Ns
events of
magnitude 1
in time frame
2
- 113 -
Chapter 5
Seismicity record
Variable Type Length
Prob(1,2,2) Double 8
...
Prob(1,2,Ns) Double 8
... ... ... ...
Probability of Prob(1,Nt,1) Double 8
having 1,
2,,Ns Prob(1,Nt,2) Double 8
events of
magnitude 1 ...
in time frame
Nt Prob(1,Nt,Ns) Double 8
... ... ... ... ...
Probability of Prob(Nbin,1,1) Double 8
having 1, Prob(Nbin,1,2) Double 8
2,,Ns ... ... ...
events of
Prob(Nbin,1,Ns) Double 8
magnitude
Probability
Nbin in timeof Prob(Nbin,2,1) Double 8
having
frame 11, Block associated
2,,Ns Prob(Nbin,2,2) Double 8 to Magnitude
events of Nbin
magnitude ... ... ...
Nbin in time
frame 2 Prob(Nbin,2,Ns) Double 8
... ... ...
Probability of Prob(Nbin,Nt,1) Double 8
having 1,
2,,Ns Prob(Nbin,Nt,2) Double 8
events of
magnitude ... ... ...
Nbin in time
frame Nt Prob(Nbin,Nt,Ns) Double 8
5.3 Generalized Poissonian model
In this option, included by suggestion of Dr. Ramn Secanell, seismicity is given by means of a non-
parametric description of the activity (or occurrence) rates of earthquakes of given magnitudes at one or
various sources.
Seismicity information is given to CRISIS in a text file, with extension *.gps (generalized Poisson
seismicity), which has the following format:
- 114 -
Chapter 5
Description Comments
ID Header A line of text used only for identification purposes
Number of different sources whose seismicity is described
NumSources
in this file
Number of magnitude bins in which the seismicity curve is
NumBins
discretized
Magnitude 1 Central point of magnitude bin 1
Magnitude 2 Central point of magnitude bin 2
... ...
Magnitude NumBins Central point of magnitude bin NumBins
Occurrence rate of earthquakes with magnitude 1 in source
Dl(1,1)
1
Dl(2,1)
1
... ...
Occurrence rate of earthquakes with magnitude NumBins
Dl(NumBins,1)
in source 1
Dl(1,2)
2
... ...
Dl(NumBins,2)
in source 1
... ...
Dl(NumBins,NumSources)
in source NumSources
The format of the *.gps file allows for the use of ":" as a separator: everything written before the separator is
ignored by CRISIS. The following table gives an example of a *.gps file, describing the seismicity of four
sources using 9 magnitude bins (please recall that everything written before the ":" sign is ignored by
CRISIS):
Example of *.gps file

Four ModifiedGR sources with M0=4, Mu=8, Beta=2, Lambda0=1
NumSources : 4
NumBins : 9
Magnitude 1: 4.2222
Magnitude 2: 4.6667
Magnitude 3: 5.1111
Magnitude 4: 5.5556
Magnitude 5: 6.0000
Magnitude 6: 6.4444
Magnitude 7: 6.8889
Magnitude 8: 7.3333
- 115 -
Chapter 5
Magnitude 9: 7.7778
Source 1 M=4.22222222222222 : 0.589085325604199
Source 1 M=4.66666666666667 : 0.242180217513314
Source 1 M=5.11111111111111 : 0.0995632639374275
Source 1 M=5.55555555555556 : 0.0409316814876875
Source 1 M=6 : 0.0168275173307138
Source 1 M=6.44444444444445 : 0.00691799919337915
Source 1 M=6.88888888888889 : 0.00284407449411698
Source 1 M=7.33333333333333 : 0.0011692339796495
Source 1 M=7.77777777777778 : 0.00048068645951254
Source 2 M=4.22222222222222 : 0.589085325604199
Source 2 M=4.66666666666667 : 0.242180217513314
Source 2 M=5.11111111111111 : 0.0995632639374275
Source 2 M=5.55555555555556 : 0.0409316814876875
Source 2 M=6 : 0.0168275173307138
Source 2 M=6.44444444444445 : 0.00691799919337915
Source 2 M=6.88888888888889 : 0.00284407449411698
Source 2 M=7.33333333333333 : 0.0011692339796495
Source 2 M=7.77777777777778 : 0.00048068645951254
Source 3 M=4.22222222222222 : 0.589085325604199
Source 3 M=4.66666666666667 : 0.242180217513314
Source 3 M=5.11111111111111 : 0.0995632639374275
Source 3 M=5.55555555555556 : 0.0409316814876875
Source 3 M=6 : 0.0168275173307138
Source 3 M=6.44444444444445 : 0.00691799919337915
Source 3 M=6.88888888888889 : 0.00284407449411698
Source 3 M=7.33333333333333 : 0.0011692339796495
Source 3 M=7.77777777777778 : 0.00048068645951254
Source 4 M=4.22222222222222 : 0.589085325604199
Source 4 M=4.66666666666667 : 0.242180217513314
Source 4 M=5.11111111111111 : 0.0995632639374275
Source 4 M=5.55555555555556 : 0.0409316814876875
Source 4 M=6 : 0.0168275173307138
Source 4 M=6.44444444444445 : 0.00691799919337915
Source 4 M=6.88888888888889 : 0.00284407449411698
Source 4 M=7.33333333333333 : 0.0011692339796495
Source 4 M=7.77777777777778 : 0.00048068645951254
Note that the values given in this file are the occurrence rates of earthquakes with magnitudes contained in a
magnitude bin. In other words, CRISIS expects, for a magnitude bin between M1 and M2, with M2>M1, the
- 116 -
Chapter 5
number of earthquakes per unit time that this source generates with magnitudes between M 1 and M 2 . For
instance, if these occurrence rates were to be computed from a usual exceedance rate curve, l (M), the
occurrence rate of earthquakes in the mentioned magnitude bin would be l(M1)-l(M2).
For hazard computation purposes, earthquakes generated in this source will have only the magnitudes given
in the file as the central points of the various bins. Therefore, it is the responsibility of the user to give a
magnitude discretization that is dense enough.
This option was originally created specifically to be applied with the smoothing method developed by Woo
(1996). Therefore, this option is frequently used to describe the seismicity of numerous point sources whose
geometrical properties (location, rupture planes) are given by means of an SSG (smoothed seismicity
geometry)file. In this case, CRISIS interprets that each source described in this seismicity file corresponds
to a point source described in the ssg file.
However, this Generalized Poisson model can be used to describe, in a non- parametric manner, the
seismicity of area or line geometries. In these cases, CRISIS will interpret that the occurrence rates given in
the *.gps file are associated to the whole area/line, so the program will uniformly distribute the occurrence
rate across/along the area/line.
5.3 Gridded seismicity

This model is used to describe seismicity with a rapid geographic variation. Seismicity is assumed to be of
Modified Gutenberg-Richter type, defined as follows:
where M stands for magnitude, l ( M) is the earthquake magnitude exceedance rate, M 0 is the threshold
magnitude, l0 is the exceedance rate at M=M0, Mu is the maximum possible magnitude, and b is a
parameter.
M 0 is assumed to be constant across the seismicity region, but l 0 , b and M u can have an arbitrary
geographical variation, defined by means of one grid for each one of these three parameters.
These three parameters are given to CRISIS in three files (for instance, L0.grd, EB.grd and MU.grd) with
format *.grd (Surfer 6 ASCII or Binary formats)
- 117 -
Chapter 5
In general, this seismicity model can be used only with grid sources. The limiting coordinates
([Xmin,Ymin],[Xmax,Ymax]) of the seismicity grids must coincide with the limits of the source geometry
grid. The number of rows and columns in the seismicity grids may be equal or smaller than those of the
source geometry grid. The number of rows and columns may be different in the three seismicity files.
5.3 Modified Gutenberg-Richter model

This model is associated to Poisson occurrences, so the probability of exceeding intensity level a in the next
Tf years, given that an earthquake with magnitude M took place at a distance R from the site, is given by:
(1)
where p1(a|M,R) is the exceedance probability of intensity level a, given that a magnitude M event occurred
at a distance R from the site, and Dl (M) is the Poissonian magnitude exceedance rate associated to the
magnitude range (also called magnitude bin) characterized by magnitude M. Note that p1 (a|M,R) depends
only on magnitude and site- to- hypocenter distance. This probability does not depend on earthquake
occurrence probabilities.
In turn, Dl(M)can be computed as
(2)
where it is implicit that the magnitude bin characterized by magnitude M goes from M- DM/2 to M+D M/2.
For the modified Gutenberg-Richter model, the earthquake magnitude exceedance rate is given by:
(3)
- 118 -
Chapter 5
where l 0 is the exceedance rate of magnitude M 0 , b is a parameter equivalent to the "b-value" for the
source (except that it is given in terms of the natural logarithm) and Mu is the maximum magnitude for the
source.
CRISIS can account for uncertainty in both b and Mu . In this case, in order to handle uncertainty in b the
user must give its expected value and coefficient of variation of and give also parameters that describe the
uncertainty in the maximum magnitude.
5.4.1. Uncertainty in the maximum magnitude

CRISIS regards M u , the maximum magnitude, as an unknown quantity. We assign to this variable a
uniform probability distribution between M1 and M2, which are informed to CRISIS in terms of two values:
the expected value of the maximum magnitude, E(Mu), and DM, such that M1 and M2 are given by:
M1=E(Mu)-DM (1)
M2=E(Mu)+DM (2)
Thus, maximum magnitude is considereder equally likely for all values between M1 and M2.
5.4.1. Handling uncertainty on seismicity parameters l and b

0
As it can be seen in the definition of the modified Gutenberg-Richter model used by CRISIS, parameters
l 0 , b and M u completely define the earthquake magnitude occurrence rate. This topic deals with how
CRISIS handles the uncertainty in the first two parameters.
Using a Bayesian framework, CRISIS regards these two parameters as independent random -unknown-
variables. Moreover, it assumes that uncertainty in b is correctly described with a gamma probability
distribution and, for the reasons that we will describe in the following paragraphs, it disregards uncertainty
in l0.
In order to show the soundness of this treatment, we will start assuming the following commonly accepted
hypothesis:
a. Occurrences are Poissonian
b. The probability distribution of magnitudes follow a Gutenberg- Richter relation that is
unbounded by the right-hand side. This amounts to say that the maximum possible magnitude, M u,
is much larger than M0, the cut-off magnitude of the hazard analysis
A consequence of assumption a) is that the times between earthquakes with magnitude M M 0 , t, are
independent, equally distributed random variables with exponential distribution. Thus, its associated
probability density function is:
(1)
where l0 is an unknown parameter. Also, it follows from assumption a) that times of earthquake occurrence
and their corresponding magnitudes are independent from each other. From assumption b) we gather that
magnitudes are also independent, equally distributed random variables with a shifted exponential
distribution. Therefore, their probability density function is:
- 119 -
Chapter 5
(2)
where b is also an unknown parameter. It can be verified that equation 1 integrates to unity in the range t0
while equation 2 integrates to 1 in the range M M0 (remember that, up to now, M is unbounded).
Consider the observation of an event consisting in the occurrence of N earthquakes, with times between
them t i and magnitudes M i , i=1.. N. According to our assumptions, the likelihood of this event, given
unknown parameters q= (l0, b) can be written as:
(3)
Or, in other words,
(4)
From equation 4, the classic maximum likelihood estimators for l0 and b can be found:
(5)
(6)
where is the total observation time in the catalog.
Following with the Bayesian approach, l 0 and b are regarded as random variables whose probability
distributions are fixed a priori and then updated in the light of the observations.
A common approach is to use as prior distributions the natural conjugates of the process. In this case, an
examination of the likelihood function in equation 4 shows that the following likelihood (the kernel of the
probability function) is the natural conjugate of the process:
(7)
It can be seen that the selected prior is the product of two gamma distributions. Then, we apply Bayes'
theorem, thus finding the posterior distribution of our unknown parameters:
(8)
It can be clearly seen that, a posteriori, both l 0 and b are gamma distributed but, more relevant for our
current discussion, we observe that, a posteriori, they are independent from each other since the joint
posterior likelihood of q is simply the product of the likelihoods of l0 and b.
The result is perhaps unexpected for those not familiar with the use of Bayesian methods (now the reader
can see that the maximum likelihood approach is a particular case of the more general Bayesian method),
but it is intuitively correct. Let us say that one is estimating l0 and b with the maximum likelihood method
(equations 5 and 6). Now say that, after a first estimation round one discovers that one of the magnitudes in
the sample was wrong. This new information, as can be appreciated in equations 5 and 6, would change of
course the estimation of b, but it would not change our estimation of l0.
Equation 8 justifies two CRISIS features:
- 120 -
Chapter 5
1. Treating l 0 and b as if they were independent (provided, of course, that they have been
estimated by Bayesian means or, at least, with the maximum likelihood method);
2. Treating the uncertainty in b assuming that this variable is gamma distributed.
Equation 8, by the way, also gives clues as to the size of the uncertainty in b: a posteriori, its coefficient of
variation is 1/(N+k-1)1/2 so, if the prior information is not overwhelmingly large (that is, if r<<N, meaning
that the sample size is reasonably large) then its coefficient of variation is of the order of 1/N1/2.
Now, we will remove the restriction that M u >>M 0 . We will assume that, as CRISIS does, magnitude
exceedance rate follows a modified Gutenberg-Richter law, given by:
(9)
where M0MMu. In this case, the probability density function of M is the following:
(10)
Replacing equation 10 into equation 4, having in mind that nothing has changed regarding occurrence
times, we find:
(11)
Now, the maximum likelihood estimators cannot be determined analytically (although, in general, they do
not differ much from those obtained with equations 5 and 6). But, if we continue with the Bayesian process,
we can find that, although b is not gamma distributed anymore (although its distribution is not far from a
gamma if M 0 and M u are not too close), l0 and b remain independent a posteriori due to the fact that l0 is
not present in the b - related term of the event likelihood, so the posterior joint likelihood of q is, again,
simply the product of the likelihoods of l0 and b.
We will proceed now to illustrate the reasons why CRISIS disregards uncertainty in l0 . Consider the basic
seismic hazard equation, expressed in terms of intensity exceedance rates (but a similar analysis could be
performed for exceedance probabilities in given time frames), for a single point-source located at distance R
from the site of analysis:
(12)
where n(a|l0,b) is the exceedance rate of intensity a given that l 0 and b are known. Replacing equation 10
into equation 12 we find that:
(13)
To remove the conditionality in n(a) we integrate with respect to the joint probability density function of the
unknown parameters (l0 and b in this case), which amounts to computing its expected value with respect to
these parameters:
(14)
- 121 -
Chapter 5
Since we already had established that l0 and b are independent random variables, it follows that:
(15)
and, since the distribution of b does not depend on l0,
(16)
Therefore,
(17)
where E b(.) denotes expected value with respect to b . It is clear from equation 17 that the first moment of
the exceedance rate (the quantity usually reported as the exceedance rate) is insensitive to uncertainty in
l0 but, since n(a) depends on the probability distribution assigned to b (we need this distribution to compute
the expected value with respect to b), it definitively depends on the uncertainty in b.
In summary, in order to compute the expected value of the exceedance rates, CRISIS solves equation 17 for
point-sources that come from the subdivision of the sources originally given by the user, using a gamma
distribution to describe the uncertainty in b . Since exceedance rates are additive, so are their expected
values; hence, disregarding uncertainty in l0 for the purpose of computing the first moment of the intensity
exceedance rate is rigorously justified.
Note, from equation 17, that disregarding uncertainty in b would be equivalent to replacing the probability
density function assigned to this parameter with the following Diracs delta function:
(18)
In this case, equation 17 would take the following form:
(19)
which is the classic hazard equation (see equation 13) when parameters l0 and b are deterministically equal
to their respective expected values. In general, however, equation 17 must be considered only a first-order
approximation to the true value of the intensity exceedance rate.
Clearly, if higher-order moments of n(a) were required, a correct answer could only be given accounting for
the uncertainty in l 0. But, in its present version CRISIS reports only the expected value of the intensity
exceedance rates, so there is no need to know how uncertain l0 is.
- 122 -
Chapter 5
5. Site Effects
CRISIS permits inclusion of local site effects in hazard computations. Site effects are given to CRISIS in
terms of amplification factors that depend on site location, structural period and ground-motion level (in
order to account for soil non-linearity).
Amplification factors are interpreted by CRISIS in the following way. Suppose that during the hazard
computations, CRISIS requires to compute the intensity at structural period T that would take place at site S
due to an earthquake of magnitude M originating at hypocenter H. We will denote this intensity as
I(S,T,M,H).
Normally, I(S,T,M,H) is computed using the attenuation relation or ground motion prediction model that the
user has selected for the source to which H belongs, or using the special attenuation model that the user has
assigned to the source-site combination to which S and H belong.
The value so computed is interpreted by CRISIS as the median intensity without site effects. But if site
effects are given, then the median intensity that CRISIS will use for the hazard computations, I S , is the
product of I(S,T,M,H) and the amplification factor given by the user, which depends on site location,
structural period and ground motion level, I0. We will denote this amplification factor as A(S,T,I0). In other
words,
IS(S,T,M,H)=I(S,T,M,H) A(S,T,I0) (1)
Clearly, if no site effects are present, then A(S,T,I0)=1.

On the other hand, the uncertainty in the intensity after site effects is computed in the following way. If the
user has furnished not only amplifications factors but also an optional file of sigma values, the uncertainty
measure will be extracted from the file of sigma values. If a file of sigma values has not been furnished by
the user, the standard deviation of the acceleration after site-effects will have the same value than the one it
had before site-effects.
The user has to give CRISIS means to obtain the amplification factors A(S,T,I0).and, optionally, uncertainty
values s (S,T,I 0 ).These factors are given to CRISIS by means of two (or three) binary files that will be
described in the following paragraphs. These files must have the same base name, but different extensions.
1) Predominant period file
This is a binary grid file in Surfer 6 binary format (*.grd). The main purpose of this file is to locate in space
the grid for which amplification factors are given, as well as to give the grid's resolution. This grid contains
as "z-values" the predominant ground periods associated to each point of the grid. Points with positive
periods are interpreted as part of the area for which site effects are known. Points with negative periods are
interpreted as outside the area for which site effects are known. Hence, for these points, the amplification
factor will always be 1, regardless of period and ground motion level. For these points, the uncertainty will
be that of the acceleration computed without site-effects.
Extension grd is required for this file. For instance, MySiteEffects.grd.
2) Amplification factors file
This is also a binary file, with extension ft. For instance, MySiteEffects.ft.
- 123 -
Chapter 5
This file contains the amplification factors themselves. As we have indicated, amplification factors depend
on site location, structural period and ground-motion level. Dependence on ground-motion level is included
to account for non-linear soil behavior. In view of this, amplification factors are given by means of a 4-
index matrix.
The first two indexes are used to sweep through space, that is, rows and columns of a grid; please note that
the size and location of the grid of amplification factors are exactly the same than for the grid of
predominant periods. The third index sweeps through structural periods, while the fourth index sweeps
through ground motion levels.
In principle, amplification factors for a given site and period can be different depending on the size of the
ground motion. In general, CRISIS uses as an indicator of this size the intensity for the shortest period
available for the GMPM that is used to compute the intensity without site effects. In most of the cases (but
not always) this intensity corresponds to peak ground acceleration.
The format in which amplification factors must be given is described in the following table:
Block Variable Size Co

Th
A number 1 Integer res
fut
If N
the
Number of ground motion levels, NL Integer
beh
Header ass
Number of periods, NT Integer
Ground motion level 1 Double

Ground motion level NL Double
Period 1 Double
- 124 -
Chapter 5

Period 2 Double
Double
Period NT Double
Th
am
Amplification function for ground-motion level 1 NT doubles
For site 1,1 fun
ag
and
mo
is a
col
NT
num
NT doubles
one
stru
per
firs
Amplification function for ground-motion level
NT doubles is a
NL
to P
and
- 125 -
Chapter 5

Amplification function for ground-motion level 1 NT doubles Th
Amplification function for ground-motion level 2 NT doubles the
the
NT doubles
tha
Amplification function for ground-motion level ass
NT doubles
NL pre
perI
For site 1,2 tha
wo
sta
the
d
cor
fo
adv
the
firs
cou
se
the
co
Nx
are
For site Nx,Ny num
gri
alo
axi
(co
and
num
gri
NT doubles
alo
axi
giv
ass
NT doubles pre
NL
per
file
The first column of the following table presents an example of the contents of a site- effects file with
extension ft. We recall, however, that this file must be in binary format.
Value Comments
1 A number 1, reserved for future use
3 3 ground motion levels
5 5 different structural periods
20 First ground motion level
100 Second ground motion level
300 Third ground motion level
- 126 -
Chapter 5
Value Comments
0.0 First period for which amplifications are given
0.2 Second period for which amplifications are given
0.5 Third period for which amplifications are given
1.0 Fourth period for which amplifications are given
2.0 Fifth period for which amplifications are given
Five amplifications factors, one for each structural period, for ground-motion
1.3 1.5 2.3 1.0 0.9
level 1
1.2 1.4 2.2 0.9 0.8
level 2
1.1 1.3 2.1 0.7 0.7
level 3
2.3 2.5 3.3 2.0 1.9
level 1
2.2 2.4 3.2 1.9 1.8
level 2
2.1 2.3 3.1 1.7 1.7
level 3
...
2.3 2.5 3.3 2.0 1.9
level 1
2.2 2.4 3.2 1.9 1.8
level 2
2.1 2.3 3.1 1.7 1.7
level 3
3) Sigma file
This is also an optional binary file, with extension sig. For instance, MySiteEffects.sig. If this file is not
given, then the uncertainty after site effects will be exactly the same as uncertainty before site-effects-
This file contains the values of the uncertainty parameter that will be used instead of that provided for the
attenuation relation before site-effects. Sigma values depend on site location, structural period and ground-
motion level. Dependence on ground-motion level is included to account for non-linear soil behavior. In
view of this, sigma values are given by means of a 4-index matrix which has exactly the same structure as
the matrix than contains the amplification factors (see Table above)
5.4 Site Effects

CRISIS permits inclusion of local site effects in hazard computations. Site effects are given to CRISIS in
terms of amplification factors that depend on site location, structural period and ground-motion level (in
order to account for soil non-linearity).
- 127 -
Chapter 5
Amplification factors are interpreted by CRISIS in the following way. Suppose that during the hazard
computations, CRISIS requires to compute the intensity at structural period T that would take place at site S
due to an earthquake of magnitude M originating at hypocenter H. We will denote this intensity as
I(S,T,M,H).
Normally, I(S,T,M,H) is computed using the attenuation relation or ground motion prediction model that the
user has selected for the source to which H belongs, or using the special attenuation model that the user has
assigned to the source-site combination to which S and H belong.
The value so computed is interpreted by CRISIS as the median intensity without site effects. But if site
effects are given, then the median intensity that CRISIS will use for the hazard computations, I S , is the
product of I(S,T,M,H) and the amplification factor given by the user, which depends on site location,
structural period and ground motion level, I0. We will denote this amplification factor as A(S,T,I0). In other
words,
IS(S,T,M,H)=I(S,T,M,H) A(S,T,I0) (1)
Clearly, if no site effects are present, then A(S,T,I0)=1.

On the other hand, the uncertainty in the intensity after site effects is computed in the following way. If the
user has furnished not only amplifications factors but also an optional file of sigma values, the uncertainty
measure will be extracted from the file of sigma values. If a file of sigma values has not been furnished by
the user, the standard deviation of the acceleration after site-effects will have the same value than the one it
had before site-effects.
The user has to give CRISIS means to obtain the amplification factors A(S,T,I0).and, optionally, uncertainty
values s (S,T,I 0 ).These factors are given to CRISIS by means of two (or three) binary files that will be
described in the following paragraphs. These files must have the same base name, but different extensions.
1) Predominant period file
This is a binary grid file in Surfer 6 binary format (*.grd). The main purpose of this file is to locate in space
the grid for which amplification factors are given, as well as to give the grid's resolution. This grid contains
as "z-values" the predominant ground periods associated to each point of the grid. Points with positive
periods are interpreted as part of the area for which site effects are known. Points with negative periods are
interpreted as outside the area for which site effects are known. Hence, for these points, the amplification
factor will always be 1, regardless of period and ground motion level. For these points, the uncertainty will
be that of the acceleration computed without site-effects.
Extension grd is required for this file. For instance, MySiteEffects.grd.
2) Amplification factors file
This is also a binary file, with extension ft. For instance, MySiteEffects.ft.
This file contains the amplification factors themselves. As we have indicated, amplification factors depend
on site location, structural period and ground-motion level. Dependence on ground-motion level is included
to account for non-linear soil behavior. In view of this, amplification factors are given by means of a 4-
index matrix.
- 128 -
Chapter 5
The first two indexes are used to sweep through space, that is, rows and columns of a grid; please note that
the size and location of the grid of amplification factors are exactly the same than for the grid of
predominant periods. The third index sweeps through structural periods, while the fourth index sweeps
through ground motion levels.
In principle, amplification factors for a given site and period can be different depending on the size of the
ground motion. In general, CRISIS uses as an indicator of this size the intensity for the shortest period
available for the GMPM that is used to compute the intensity without site effects. In most of the cases (but
not always) this intensity corresponds to peak ground acceleration.
The format in which amplification factors must be given is described in the following table:

Th
A number 1 Integer res
fut
If N
the
Number of ground motion levels, NL Integer
beh
ass
Number of periods, NT Integer
Header Ground motion level 1 Double

Ground motion level NL Double
Period 1 Double
Period 2 Double
Double
Period NT Double
- 129 -
Chapter 5

Th
am
For site 1,1 fun
ag
and
mo
is a
col
NT
num
NT doubles
one
stru
per
firs
NT doubles is a
NL
to P
and
- 130 -
Chapter 5

Amplification function for ground-motion level 1 NT doubles Th
Amplification function for ground-motion level 2 NT doubles the
the
NT doubles
tha
Amplification function for ground-motion level ass
NT doubles
NL pre
perI
For site 1,2 tha
wo
sta
the
d
cor
fo
adv
the
firs
cou
se
the
co
Nx
are
For site Nx,Ny num
gri
alo
axi
(co
and
num
gri
NT doubles
alo
axi
giv
ass
NT doubles pre
NL
per
file
The first column of the following table presents an example of the contents of a site- effects file with
extension ft. We recall, however, that this file must be in binary format.
Value Comments
1 A number 1, reserved for future use
3 3 ground motion levels
5 5 different structural periods
20 First ground motion level
100 Second ground motion level
300 Third ground motion level
- 131 -
Chapter 5
Value Comments
0.0 First period for which amplifications are given
0.2 Second period for which amplifications are given
0.5 Third period for which amplifications are given
1.0 Fourth period for which amplifications are given
2.0 Fifth period for which amplifications are given
1.3 1.5 2.3 1.0 0.9
level 1
1.2 1.4 2.2 0.9 0.8
level 2
1.1 1.3 2.1 0.7 0.7
level 3
2.3 2.5 3.3 2.0 1.9
level 1
2.2 2.4 3.2 1.9 1.8
level 2
2.1 2.3 3.1 1.7 1.7
level 3
...
2.3 2.5 3.3 2.0 1.9
level 1
2.2 2.4 3.2 1.9 1.8
level 2
2.1 2.3 3.1 1.7 1.7
level 3
3) Sigma file
This is also an optional binary file, with extension sig. For instance, MySiteEffects.sig. If this file is not
given, then the uncertainty after site effects will be exactly the same as uncertainty before site-effects-
This file contains the values of the uncertainty parameter that will be used instead of that provided for the
attenuation relation before site-effects. Sigma values depend on site location, structural period and ground-
motion level. Dependence on ground-motion level is included to account for non-linear soil behavior. In
view of this, sigma values are given by means of a 4-index matrix which has exactly the same structure as
the matrix than contains the amplification factors (see Table above)
- 132 -
Chapter 6
6. Source geometry
In general, sources are the portions of the Earth in which it is possible that earthquakes take place.
CRISIS accepts source geometries of the following three types:
1) Areas (polygons)
2) Faults (polylines)
3) Points
4) Grid
5) Area Planes
6.1 Area source

This option defines the active source as an area source. In general, area sources are polygons defined by the
3D coordinates of their vertex. In the example below we have a 3D polygon with 8 vertexes simulating a
dipping plate with varying dip angle. Vertical planes are allowed.
In the case of area sources, in order to perform the spatial integration, CRISIS will divide the polygon into
triangles. It first checks if triangulation can be made in the XY plane. Numbering of the vertex of the
polygon must be done counter-clockwise in this plane when looked from above the surface of the Earth. If
there are vertical planes, CRISIS will try to triangulate the area in the XZ plane, so numbering of vertex
must be done counterclockwise in this plane. Finally, CRISIS will try to triangulate in the YZ plane. There
are some bizarre source geometries that cannot be well resolved by CRISIS; for instance, an L-shaped
vertical plane. In these cases, an error will be reported.
- 133 -
Chapter 6
6.1.1.
6.1.2. Behavior options
Normal
This is the original behavior in CRISIS area sources. In general, the rupture areas are circles (that, is,
ellipses with aspect ratio equal to 1), whose area is related to magnitude through parameters K1 and K2.
The rupture areas are contained in the plane of the source area itself. If the source area is a horizontal plane
(that is, all its vertex have the same depth) then the rupture planes will be horizontal. If the area source is a
vertical plane, then the circles that represent the ruptures will be contained in a vertical plane. If the area
geometry is complex (that is, it is a non-planar area), then the rupture plane will be that of the triangle in
which the corresponding hypocenter is contained. Note that for this behavior, the aspect ratio of the ruptures
will always be 1, regardless of the aspect ratio selected by the user (see Fault aspect ratio later in this topic).
In this option, CRISIS allows the rupture area to get out of the source area. If this behavior is not considered
correct, then the option Treat as Fault must be selected.
Treat as fault
Since CRISIS 2008, we introduced a new attribute of area sources that indicates that the area should be
treated as a fault. The difference between a normal area source and one that is treated as a fault is that in
the later case CRISIS does not allow rupture areas to extend outside the boundaries of the area source itself,
while CRISIS has not this restriction in a regular area source (i.e., rupture areas can go out of the sources
boundaries). This difference is relevant only in the cases in which Rrup or RJB are used as distance
measures and rupture areas are larger than 0 (that is, parameters K1 and K2 > 0).
So far, for an area source to be treated as a fault the following conditions must be met:
1) It must have exactly 4 vertex
2) All vertex must roughly be in the same plane (there are tolerances)
3) All internal angles of the polygon must be roughly straight (there are tolerances)
In this case the rupture areas will be ellipses with aspect ratio equal to the value given by the user in the
Fault Aspect Ratio textbox, with an area related to magnitude through parameters K1 and K2. The rupture
areas are contained in the plane of the source area itself. If the source area is a horizontal plane (that is, all its
vertex have the same depth) then the rupture planes will be horizontal. If the area source is a vertical plane,
then the circles that represent the ruptures will be contained in a vertical plane. If the area geometry is
complex (that is, it is a non-planar area), then the rupture plane will be that of the triangle in which the
corresponding hypocenter is contained.
Breaks always
This option means that, regardless of magnitude, the area will break completely for each earthquake. This
option is normally used for characteristic earthquakes which, by hypothesis, will completely fill up the
rupture area, regardless their magnitudes. In view of this, there is only one hypocenter associated to the area.
This hypocenter is the point within the source closest to the computation site. Again, this is only relevant
when RJB or Rrup are being used as distance measures. In this case, the values of K1 and K2 are irrelevant.
- 134 -
Chapter 6
6.1.3. Thickness and slices

For this option, an area source is interpreted as a "slice". Hence, if parameter Slices equals 1, then we are
dealing with a common area source. If Slices =N, N>1, then CRISIS will construct N area sources whose
vertex have the same (X,Y) coordinates but Z coordinates (that is, depths) that depend on parameter
Thickness. In general, if zj is the depth of the j-th vertex of the original area source (that is, the value given
in the DataGrid as Depth in the corresponding vertex), the depth of the same vertex but for slice number k
will be z jk=z j+(k-1)*DZ, where DZ=Thickness/(N-1). In this case, the seismicity will be evenly divided
among the N slices. The option is intended to simulate an even distribution of seismicity with depth.
6.1.4. Parameters K1 and K2

Parameters K1 and K2 relate earthquake magnitude with rupture size.
6.1.5. Fault Aspect Ratio

Indicate in this textbox the fault aspect ratio you want elliptical ruptures to have if the option "Treat as fault"
has been selected. In this case, the aspect ratio, Ar is defined as
Ar=Dx/Dy (1)
where Dx is the dimension of the fault in the X direccion and Dy is the dimension of the fault in the Y
direction. It must be recalled that, in order to be treated as a fault, an area source must have exactly four
vertex that form a rectangle that lies in a single plane. By definition, the X direction is the one that joins
vertex 1 and 2 of the area source, while the Y direction is the one that joins vertex 2 and 3.
6.1 Area Planes geometry

This option defines the active source as an area source in which the rupture planes have an orientation given
by the user. They are different from the old area sources because in the area source geometry type the
rupture planes were the planes formed by the area itself, while in Area Planes geometry the rupture planes
have a constant orientation given by the user.
6.1.1. Vertex
In general, this source is defined by the 3D vertex of a polygon. In the example below we have a 3D
polygon with 8 vertexes.
- 135 -
Chapter 6
In order to perform the spatial integration, CRISIS will divide the polygon into triangles. It first checks if
triangulation can be made in the XY plane. Numbering of the vertex of the polygon must be done counter-
clockwise in this plane when looked from above the surface of the Earth. If there are vertical planes,
CRISIS will try to triangulate the area in the XZ plane, so numbering of vertex must be done
counterclockwise in this plane. Finally, CRISIS will try to triangulate in the YZ plane. There are some
bizarre source geometries that cannot be well resolved by CRISIS; for instance, an L-shaped vertical plane.
In these cases, an error will be reported. This subdivision will define the location of the potential
hypocenters but not he orientation of the rupture plane.
6.1.2. Orientation of rupture plane

An orientation of the rupture planes of the sources can be assigned by means of a unit vector, n=(ux,uy ,uz),
normal to the rupture plane. In the following figure examples are presented of some plane orientations with
their corresponding normal vector, as understood by CRISIS:
- 136 -
Chapter 6
6.1.3. Parameters that define rupture size

A magnitude-dependent size of the rupture plane can be assigned using parameters K1 and K2, This choice
is relevant only in the cases in which Rrup or RJB are used as distance measures.
6.2 Line source

This option defines the active source as a fault (line) source. Line sources are, in general, polylines defined
by the 3D coordinates of their vertexes. The example below shows a fault source of 4 vertexes, located in
the XZ plane with varying depth.
Breaks always
This option means that, regardless of magnitude, the line will break completely for each earthquake. This
option is normally used for characteristic earthquakes which, by hypothesis, will completely fill up the
rupture length, regardless their magnitudes. In view of this, there is only one hypocenter associated to the
line. This hypocenter is the point within the line closest to the computation site. Again, this is only relevant
when RJB or Rrup are being used as distance measures. In this case, the values of K1 and K2 are irrelevant.
6.2 Grid Source

This option defines the active source as a collection of point sources located at the nodes of a rectangular
grid that is parallel to the surface of the Earth, that is, a grid in which all the nodes have the same depth.
Each one of the nodes is a potential hypocenter. The nodes of the grid are the only hypocenters that CRISIS
will consider in the calculations as point sources. If the grid is not sufficiently dense, the calculation sources
may be too far apart.
- 137 -
Chapter 6
The grid is defined by the following parameters :
6.2.1. Geographic limits and density of the grid

Longitude Latitude
Origin (usually the SW corner) Xmin Ymin
End (usually the NE corner) Xmax Ymax
Number of lines N M
The total number of nodes in the grid will therefore be N*M
6.2.2. Source depth

The source depth is uniform for the entire grid and it is given in kilometers.
6.2.3. Delimitation polygon (optional)

The grid can be delimited by a polygon or group of polygons given in a Shape file *. Shp
- 138 -
Chapter 6
Only the grid nodes that lie within at least one of the polygons will be considered active point sources.
6.2.4. Orientation of rupture plane (optional)

An orientation of the rupture planes of the point sources can be assigned by means of the "Angle grids".
These grids contain information that permits definition of a unit vector, n=(ux ,uy,u z), normal to the rupture
plane. In the following figure examples are presented of some plane orientations with their corresponding
normal vector, as understood by CRISIS:
These vectors are given to CRISIS by means of three files that contain the X, Y and Z components,
respectively, of the normal vectors that define the plane orientations. These files must be in *.grd format (see
Surfer 6 ASCII and Surfer 6 Binary formats).
The names of the files are fixed:
- 139 -
Chapter 6
l NormalVector_X.grd
l NormalVector_Y.grd
l NormalVector_Z.grd
The path of the folder containing these files must be provided to CRISIS. If angle grids are not given to
CRISIS horizontal rupture planes (dip 0) are assumed. Angle grids must have the same origin, end and
node size than the main source grid:
The inclusion of angle grids is relevant only in the cases in which Rrup or RJB are used as distance
measures and rupture areas are different from 0, that is, parameters K1 and K2 > 0 (see next item).
6.2.5. Orientation of rupture plane (optional)

A magnitude-dependent size of the rupture plane can be assigned using parameters K1 and K2, This choice
is relevant only in the cases in which Rrup or RJB are used as distance measures.
6.3 Point (SSG) sources

This option defines the active source as a collection of point sources, in which each vertex will be a point
source.
Each point is a potential earthquake hypocenter, and is defined, in the newer versions of CRISIS, in terms
of the following parameters:
1) Longitude, latitude and depth (in Km) of the point.
2) A unit vector normal to the rupture plane associated to each point source. This unit vector is relevant only
when the ground motion prediction model associated to this source uses distance measures for which the
rupture area is relevant (that is, RRUP or RJB; see the definition of source-to-site distances in this link).
- 140 -
Chapter 6
Since point sources are generally used to geometrically describe potentially thousands of hypocentral
locations, information about this type of source is given to CRISIS by means of an ASCII file, usually with
extension ssg, with the following structure:
Point geometry file

Description Variable Type
ID Header Header String
Number of point sources TotSrc Integer
Geometry record for source 1 Geom(1) Geometry record
Geometry record for source 2 Geom(2) Geometry record

Geometry record for source TotSrc Geom(TotSrc) Geometry record
The following table describes the structure of a geometry record:
Geometry record
Description Variable Comments
h.X in degrees
Hypocentral location h.Y in degrees
h.Z in Km, always positive
Unit vector describing the orientation e1.x
of the fault plane These three values describe a unit vector normal to the
e1.y
fault plane. X is longitude, Y is latitude and Z is depth
e1.z
The following table gives and example of a point- source geometry file, where 16 point sources are
geometrically described:
Line in file Comment

Griglia Marzocchi Header line for identification purposes
16 16 points are described
- 141 -
Chapter 6
Line in file Comment

5.550 44.950 11 0 0 0
5.550 45.050 11 0 0 0
5.550 45.150 11 0 0 0
5.550 45.250 11 0 0 0
5.650 44.650 11 0 0 0 These lines give the longitude, latitude and depth (11 Km in this
5.650 44.750 11 0 0 0 case, always positive) for the 16 sources.
5.650 44.850 11 0 0 0
Note that the coordinates of the unit vector normal to the fault plane
5.650 44.950 11 0 0 0 are given as 0, 0, 0. This means that they are unknown, or perhaps
5.650 45.050 11 0 0 0 irrelevant. They would be irrelevant, for instance, if an attenuation
5.650 45.150 11 0 0 0 model based on focal distance is going to be used for the hazard
5.650 45.250 11 0 0 0 computations. If the unit vector normal to the fault plane is described
5.650 45.350 11 0 0 0 with (0,0,0) a horizontal plane will be the default.
5.650 45.450 11 0 0 0
5.650 45.550 11 0 0 0
5.750 44.450 11 0 0 0
5.750 44.550 11 0 0 0
- 142 -
Chapter 6
6. Tools Menu
l GMPE branch constructor. This tool allows construction of a set of CRISIS hazard models that
collectively constitute a logic tree, in which each hazard model, represented by a *.dat file, is one of the
branches of the tree.
l Map comparer. This tool allows compare two maps of seismic hazard that were calculated with
different considerations.
l Site-effects files conversion. This tool allows converting a text file with a grid of amplification factors
into a binary file.

l Non-Poisson files conversion. Use this tool to convert generalized ASCII non-Poisson seismicity files
(nps extension) into the binary format required by CRISIS or vice-versa (from binary to ASCII).
6.3 GMPM logic-tree branch constructor

This tool allows construction of a set of CRISIS hazard models that collectively constitute a logic tree, in
which each hazard model, represented by a *.dat file, is one of the branches of the tree.
The logic tree that can be constructed with this tool is one in which the geometry and seismicity
characteristics are fixed -therefore, they are the same for all branches of the logic tree- but each branch of
the tree represents a different combination of ground motion prediction models (GMPM).
The following information must be given to the logic-tree constructor:
6.3.1. Base Hazard Model

Give here, by clicking in the (...)button, the name of the *.dat file that will be the basis for the construction
of all branches. This file must be a valid CRISIS model, from which seismicity, geometry and general
information will be taken, and will be made part of all branches. The logic-tree constructor will also take
from the base hazard model the number of attenuation models that are present, NMOD, as well as the
sources to which each model is assigned. As we will see later, for each one of the NMOD GMPM, the user
can give several competing options. The logic tree constructor will construct *.dat files that will describe all
the combinations of GMPM.
6.3.2. Branch data

....
6.3.3. GMPM and Weights data grids

Each row of the data grid represents one of the attenuation models given in the base hazard model. The first
row is attenuation model 1, and so on. For instance, if the base hazard model was constructed with four
different models, these data grids will have four rows. The user cannot change the number of rows.
For a given row, say the i-th, each non-empty column indicates the possible values that the i-th attenuation
model in the base hazard model could take. For instance, if for row 1 there are three non-empty columns,
- 143 -
Chapter 6
this means that attenuation model 1 can take three possible values. Click in any cell of the data grids to
change the selected GMPM or the assigned weight.
6.3.4. Add or Delete columns

Use these buttons to add or delete columns to the data grids.
6.3.5. Clear selected cell

Clears the selected cell.
6.3.6. Create branches

Once the user has given all the required information regarding the GMPM that will be used and their
corresponding weights, use this button to create all the *.dat files that represent the logic-tree branches. This
button will also cretae an *.ltc file, which contains the names of the *.dat files corresponding to the
branches, as well as the weights associated to each branch. Load this *.ltc file in the logic-tree computation
screen in order to make the hazard computations for each branch and the logic-tree combination.
6.3 Non-Poisson seismicity text to binary

Use this tool to convert generalized ASCII non-Poisson seismicity files (nps extension) into the binary
format required by CRISIS or vice-versa (from binary to ASCII). Simply double-click into the text box in
order to read a file name and then use the appropriate button to go from ASCII to binary or vice-versa.
6.3 Site-effects text to binary

Use this tool to convert ASCII files that describe site-effects into the binary files required by CRISIS. The
detailed format of site-effect files can be found here.
6.3 Map Comparer

Use this tool to compare two hazard maps already computed by CRISIS. The tool allows visualization of
absolute or relative differences between two hazard maps defined by the hazard model that originated them
(the associated *.dat file), a fixed level of intensity or exceedance probability, an intensity measure and a
given time frame.
This tool was inspired by a similar one developed by F. Martinelli and C. Meletti for CRISIS Web.
- 144 -
Chapter 7
7. File Formats
The following is a description of the various file formats in use by CRISIS.
7.5 Cities file

The city file is in ASCII format and must contain the following data:
Number of cities
Name of the state, name of the city, longitude and latitude of the city (1 line for each city).
Example:
2 Number of cities
GUERRERO, Acapulco, -99.900, 16.850
AGUASCALIENTES, Aguascalientes, -102.300, 21.883
State City Longitude Latitude
7.6 Data Types

Type of stored variables has the following codes:
Name Value Comments

Byte 1
Short 2
Integer 3
Corresponds to character "D". This code is used
Single 68
for compatibility with binary Surfer 6 format
Double 5
Long 6
7.1 Format of modGRN grids

modGRN format is an extension of binary Golden Surfer 6 format.
While binary Golden Surfer 6 format consists on a header and a succession of values stored as 4-byte single
numbers, modGRN format allows to store the values with other types of variables.
The only difference between modGRN and binary Golden Surfer 6 formats is the first byte in the header,
that in the case of the modGRN format indicates what type of variables are being stored.
The structure of a modGRN binary file is the following:
Name Variable type Description

Header GridHeader Grid header
- 145 -
Chapter 7
Name Variable type Description

The type is that indicated by grid First grid row Each row has constant Y
header byte ID. It can be byte, value. First row corresponds to Ylo and
z11, z12, ...
short, integer, single, double or the last row corresponds to Yhi. Within a
long. row, values are ordered form Xlo to Xhi.
The type is that indicated by grid
z21, z22, ... Second grid row
header byte ID
The type is that indicated by grid
z31, z32, ... Third grid row
header byte ID
The type is that indicated by grid The remaining rows until reaching that
...
header byte ID corresponding to Yhi
7.2 Grid Header

This is the header of modGRN files. Its structure is very similar to that of the header of binary Surfer 6 files,
except that the first byte indicates what type of variable will be stored.
Name Type Length Description Comments

The first byte indicates what
type of variables will be stores.
IDD Byte 4 4 identification bytes
See codes in the following
table
Nx Short 2 Number of columns in the X direction
Ny Short 2 Number of rows in the Y direction
Xlo Double 8 Minimum longitude of the grid
Xhi Double 8 Maximum longitude of the grid
Ylo Double 8 Minimum latitude of the grid
Yhi Double 8 Maximum latiitude of the grid
Zlo Double 8 Maximum Z value in the grid
Zhi Double 8 Minimum Z value in the grid
The first byte indicates what type of variables will be stores, according to the following codes:
Name Value Comments

Byte 1
Short 2
Integer 3
Corresponds to character "D". This code is used for com-
Single 68
patibility with binary Surfer 6 format
- 146 -
Chapter 7
Name Value Comments

Double 5
Long 6
7.3 Map file

CRISIS recognizes two format files to generate the map: 1) shape format, and 2) ASCII format.
The map in a shape format corresponds to a standard ESRI shape file.
The map file in ASCII format must contain the following data:
l Number of polygons
l For each polygon:
l Name of the polygon
l Number of vertex of the polygon
l For each vertex: Latitude and longitude
Example:
2 Number of polygons
State 1 Name of polygon 1
6 Number of vertex of polygon 1
13 10 Coordinates of the six vertex of polygon 1
12 11
11 10
10 10
10 8
12 8.5
State 2 Name of polygon 2
6 Number of vertex of polygon 2
10 10 Coordinates of the six vertex of polygon 2
11 10
12 11
13 10
13 12.4
12 13.1
7.4 SHP files

A shapefile stores non-topological geometry and attribute information for the spatial features in a data set.
The geometry for a feature is stored as a shape comprising a set of vector coordinates.
An ESRI shapefile consists of a main file, an index file, and a dBASE table. The main file is a direct access,
variable-record-length file in which each record describes a shape with a list of its vertices. In the index file,
each record contains the offset of the corresponding main file record from the beginning of the main file.
- 147 -
Chapter 7
The dBASE table contains feature attributes with one record per feature. The one- to- one relationship
between geometry and attributes is based on record number. Attribute records in the dBASE file must be in
the same order as records in the main file.
(This description was taken from: ESRI Shapefile Technical Description: an ESRI White PaperJuly
1998. See this report for further information on shapefiles)
7.5 Surfer 6 ASCII Grid Format

This section has been taken from SURFER 8 Help File
DSSA grid files contain five header lines that provide information about the size and limits of the grid,
followed by a list of Z values. The fields within DSSA files must be space delimited.
The listing of Z values follows the header information in the file. The Z values are stored in row-major
order starting with the minimum Y coordinate. The first Z value in the grid file corresponds to the lower left
corner of the map. This can also be thought of as the southwest corner of the map, or, more specifically, the
grid node of minimum X and minimum Y. The second Z value is the next adjacent grid node in the same
row (the same Y coordinate but the next higher X coordinate). When the maximum X value is reached in
the row, the list of Z values continues with the next higher row, until all the rows of Z values have been
included.
The general format of a DSSA grid file is:
Value Comment
Id The identification string DSAA that identifies the file as an ASCII grid file
nx ny nx is the integer number of grid lines along the X axis (columns)
ny is the integer number of grid lines along the Y axis (rows)
xlo xhi xlo is the minimum X value of the grid
xhi is the maximum X value of the grid
ylo yhi ylo is the minimum Y value of the grid
yhi is the maximum Y value of the grid
zlo zhi zlo is the minimum Z value of the grid
zhi is the maximum Z value of the grid
grid row 1 These are the rows of Z values of the grid, organized in row order. Each row has a constant
Y coordinate. Grid row 1 corresponds to ylo and the last grid row corresponds to yhi.
Within each row, the Z values are arranged from xlo to xhi
grid row 2
grid row 3
The following example grid file is ten rows high by ten columns wide. The first five lines of the file contain
header information. X ranges from 0 to 9, Y ranges from 0 to 7, and Z ranges from 25 to 97.19. The first Z
value shown corresponds to the lower left corner of the map and the following values correspond to the
increasing X positions along the bottom row of the grid file. This file has a total of 100 Z values.
- 148 -
Chapter 7
DSAA
10 10
0.0 9.0
0.0 7.0
25.00 97.19
91.03 77.21 60.55 46.67 52.73 64.05 41.19 54.99 44.30 25.00
96.04 81.10 62.38 48.74 57.50 63.27 48.67 60.81 51.78 33.63
92.10 85.05 65.09 53.01 64.44 65.64 52.53 66.54 59.29 41.33
94.04 85.63 65.56 55.32 73.18 70.88 55.35 76.27 67.20 45.78
97.19 82.00 64.21 61.97 82.99 80.34 58.55 86.28 75.02 48.75
91.36 78.73 64.05 65.60 82.58 81.37 61.16 89.09 81.36 54.87
86.31 77.58 67.71 68.50 73.37 74.84 65.35 95.55 85.92 55.76
80.88 75.56 74.35 72.47 66.93 75.49 86.39 92.10 84.41 55.00
74.77 66.02 70.29 75.16 60.56 65.56 85.07 89.81 74.53 51.69
70.00 54.19 62.27 74.51 55.95 55.42 71.21 74.63 63.14 44.99
7.6 Surfer 6 Binary Grid file format

This section has been taken from SURFER 8 Help File
Surfer 6 grid files [.GRD] use a layout similar to the ASCII Surfer 6 file format. The only difference is in
the identification string and that Surfer 6 grid files are binary. Data types used in Surfer 6 grid files include:
Type Description
char single byte
short 16 bit signed integer
float 32 bit single precision floating point value
double 64 bit double precision floating point value
The Surfer 6 format has the following layout:

Element Type Description
4 byte identification string DSBB which identifies the file as a Surfer 6
id char
binary grid file.
nx short number of grid lines along the X axis (columns)
ny short number of grid lines along the Y axis (rows)
xlo double minimum X value of the grid
- 149 -
Chapter 7
Element Type Description

xhi double maximum X value of the grid
ylo double minimum Y value of the grid
yhi double maximum Y value of the grid
zlo double minimum Z value of the grid
zhi double maximum Z value of the grid
first row of the grid. Each row has a constant Y coordinate. The first row
z11,
float corresponds to ylo, and the last row corresponds to yhi. Within each row, the
z12,
Z values are ordered from xlo to xhi.
z21,
float second row of the grid
z22,
z31,
float third row of the grid
z32,
float all other rows of the grid up to yhi
- 150 -
Chapter 7
7. Appendix: GMPM currently supported by

CRISIS
Class name Short name Tectonic regime
1 AbrahamsonAndSilva97 Abrahamson and Silva (1997) Active_Shallow_Crusta
2 AkkarBommer07 Akkar and Bommer (2007) Active_Shallow_Crusta
3 AkkarBommer10 Akkar and Bommer (2010) Active_Shallow_Crusta
4 Arroyoetal10 Arroyo et al. (2010) Subduction
5 AtkinsonBoore03 Atkinson and Boore (2003) Subduction
6 AtkinsonBoore06 Atkinson and Boore (2006) Stable_Continental
7 AtkinsonBoore2008 Atkinson and Boore (2008) Stable_Continental
8 AttenuationTable Attenuation Table CRISIS_Models
9 BCHydro2012 BCHydro Subduction
10 BindiGMPM Bindi et al, 2011 Active_Shallow_Crusta
11 BooreAtkinson08 Boore and Atkinson (2008) NGA Active_Shallow_Crusta
12 Campbell03 Campbell (2003) Stable_Continental
13 Campbell2008 Campbell and Bozorgnia (2008) NGA Active_Shallow_Crusta
14 CampbellBozorgnia04 Campbell and Bozorgnia (2003) Active_Shallow_Crusta
15 CauzziFaccioli08 Cauzzi and Faccioli (2008) Active_Shallow_Crusta
16 CauzziFaccioli08Simple Cauzzi and Faccioli (2008; simple version) Active_Shallow_Crusta
17 CauzziFaccioli08Vertical Cauzzi and Faccioli (2008; vertical SA) Active_Shallow_Crusta
18 Chiou2008 Chiou and Youngs (2008) NGA Active_Shallow_Crusta
19 DGarcia05 Garca et al. (2005) Subduction
20 General General Attenuation Model CRISIS_Models
21 Idriss2008 Idriss2008 Active_Shallow_Crusta
22 Jaimes2005 Jaimes 2005, Mexico City Other
23 Kanno08 Kanno08 Subduction
24 Lin2008 Lin and Lee (2008) Subduction
25 McVerry2006 McVerry2006 Other
26 PasoliniEtAl08 Pasolini et al. (2008) Active_Shallow_Crusta
27 PeerSadigh97 PEER Sadigh 97 Active_Shallow_Crusta
28 Reyes1999 Reyes 1999 Mexico City Other
29 SabettaPugliese96EpicDist Sabeta and Pugliese (1996; epicentral Distance) Active_Shallow_Crusta
30 SabettaPugliese96FaultDist Sabetta and Pugliese (1996; fault distance) Active_Shallow_Crusta
31 Sadigh97 Sadigh et al. (1997) Active_Shallow_Crusta
32 SEA99 Spudich et al. (1999; SEA99) Active_Shallow_Crusta
33 ToroEtAl97 Toro et al (1997) Stable_Continental
34 TP05 Tavakoli and Pezeshk (2005) Stable_Continental
35 Youngs97 Youngs et al. (1997) Subduction
- 151 -
Chapter 7
Class name Short name Tectonic regime

36 ZhaoEtAl2006 ZhaoEtAl2006 Subduction
- 152 -
Chapter 7
7. Validation Tests
The verification of a seismic hazard computer code is crucial for ensuring the user that the calculations
performed with it are reliable.Delete this text and replace it with your own content.
The numerical verification process of CRISIS was carried out by adopting as tools the tests developed in a
project sponsored by the Pacific Earthquake Engineering Research Center (PEER) and documented in
Verification of Probabilistic Seismic Hazard Analysis Computer Programs by Thomas et al. (2010). In
that project, several seismic hazard codes were tested, namely:
FRISK88M Version 2.0 (developer: G. Toro)
EZ-FRISK (developer: R. McGuire)
XDC52, HAZ50, TREE50 (developer: B. Youngs)
HazFX v3.f, hazgridX v3.f, filtrate.peer.f (developers: A. Frankel and S. Harmsen)
ALEAS (developer: J. Savy Abramowitz)
EQ-Elements (developer: A. Mendez)
GP-Haz (developer: P. Tan)
NEWHAZ (developer: M. Stirling)
mrs 3.1 (developer: R. LaForge)
OpenSHA (developer: N. Field)
HAZDIR (developer: B. Rowshandel)
The verification exercise consisted of two sets of test problems aimed at testing fundamental aspects of the
codes: treatment of fault sources, recurrence models and rates, attenuation relationships and their associated
uncertainties. For the simplest cases analytical solutions were also provided.
The following document gives the details of the validation process carried out with CRISIS 2008.
7.6 PEER Validation Tests

Detailed description of the validation process carried out with CRISIS 2008 according to PEER reference
(Thomas et al 2010).
- 153 -
- 154 -

CRISIS Manual PDF

Uploaded by

Copyright:

Available Formats

CRISIS Manual PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

CRISIS Manual PDF

Uploaded by

Copyright:

Available Formats

Chapter 1

CRISIS USER'S MANUAL

1.1 Brief description

1.1.1. Earthquake occurrence

1.1.2. Source geometry

1.1.3. Attenuation models

1.1.4. Spatial integration procedure

1.2 What is new in CRISIS (since CRISIS 2013 up to CRISIS

l Sigma grids in site-effects

l Generalized Poissonian (new seismicity model)

l Gridded seismicity (new seismicity model)

l Grid source (new geometry model)

l Area Planes (new geometry model)

l Various forms of writing results

l Event set generator

l Map file is back

l Disaggregation file (*.des) is back

l Some corrections were made to GMPM Atkinson and Boore, 2006

l New collection of GMPM

l A small correction in computation of areas in Area and AreaPlanes

l New options in Area geometries

file will progressively substitute the old *.dat file

1.2 What is new (from CRISIS2015 v1.2 on)

results, which are presented here.

2. Building a data file

Open a *.dat or *.xml data file

Save a data file

Give the names of the map and cities reference files

Define the computation-sites grid

Give the geometry of the sources

Give data on spectral ordinates

Give seismicity of the sources

Give data about GMPM

Give site-effects information

Give global general parameters

Validate, save, and start execution

See hazard maps

Show disaggregation chart

Give batch disaggregation parameters

CAPRA seismic scenario generation

Compute event-set for a selected computation site

Show a screen with various useful tools

Show a screen to compute optimum spectra

2.2.1. File Menu

l New. This option allows building a new data file.

2.2.1. Input Menu

the hazard analysis.

2.2.1. GMPM Viewer

Give in this tab page:

Give in this tab page:

2.2.1. Run Menu

2.2.1. Hazard Menu

CAPRA scenario generation

2.2.1. Tools Menu

into a binary file.

2.2.1. Help Menu

l Index. This option show you the help file.

2.1 Open an existing input file

2.2 Save data to a file

Open a .dat or .xml data file