33.1 Useful programs for hazard

Following is a list of programs that may be useful during a PSHA study. Most of these programs are described in more detail in different sections of the SEISAN manual.

SELECT: Select a subset of earthquake data according to given criteria. This can be useful both in connection with initial quality testing of earthquake catalogs and for selecting events within a given source zone. SE has a similar but more advanced function.
CATSTAT: Program to compute and plot the yearly, monthly and daily number of events in addition to the time-of-day distribution of events from a given catalogue.
NORHEAD: The program can move magnitudes form following header lines to empty space on first header line.
CAT_AGA: Program to reorder the hypocenter lines in a CAT-file according to hypocenter agency, in order to put the prime estimate in the beginning.
CLUSTER: Program that searches for dependent events, defined through time, distance and depth windows, in a given CAT file.
ASSO: Program to merge events which are close in time, magnitude and space in a database. This program is useful when merging catalogs to avoid double entries. ASSOCI will perform a similar search, but only based on the time difference between events. Furthermore, ASSOCI can only merge two events at a time. It is thus recommended to use ASSO rather than ASSOCI when merging catalogs for hazard assessment purposes.
EXFILTER: Identifies probable explosions, based on user-defined parameters involving time-of-day distribution and the mining locations. It can be used for catalogue clean-up and for discrimination between earthquakes and man-made explosions.
MAG: Magnitude regression and conversion program. Prepares a plot showing the data scatter and the best-fitted line for conversion between two magnitude types. Magnitude conversions can then be performed after a user defined priority list.
EPIMAP: Plots coastlines, national boundaries and earthquake epicenters. It is also possible to select a subset of earthquakes from a chosen polygon on the epicenter map.
BVALUE: Prepares magnitude-frequency of occurrence diagrams and computes a- and b-values with maximum likelihood and least square approximation. In addition, the threshold magnitude and the maximum observed magnitude can be obtained. It is generally recommended to use the Gutenberg-Richter or Weichert functions built into Seisan Explorer instead of BVALUE.
SPEC: Computes amplitude spectra for a given set of earthquake records and plots spectral ratios. It can be used to assess local site effects.
CRISIS2012: Computes seismic hazard in terms of the probability of exceedance vs. earthquake intensity measure. Any intensity measure can be provided through user defined ground motion prediction equations in addition to a large number of built-in relations. Epistemic uncertainty can be considered through the definition of a logic tree. Results can be provided as hazard maps or in terms of hazard curves for selected sites. The program can furthermore provide a deaggregation of the results. The program only runs under Windows and needs to be installed separately. Look for a MSI file in SUP.
SEISAN EXPLORER: A number of functions have been built into Seisan Explorer to aid the preparation of input to a PSHA:
- Function 'Gutenberg-Richter relation' allows for determining a- and b-values for a given catalog using different magnitude intervals and bin sizes (similar to the BVALUE program).
- Function 'Poisson distribution' allows for visually checking whether an earthquake catalog is Poisson distributed or not.
- Function 'Completeness check' provides a 'staircase plot' of the earthquake catalog, showing for different magnitude classes the cumulative number of events as a function of time. This plot allows for defining catalog completeness for different magnitude classes.
- Function 'Weichert method' allows for determining a- and b-values for a given catalog, accounting for varying catalog completeness for different magnitude classes.

Probabilistic seismic hazard computations are done using the Crisis2012 program. In addition, the programs listed above and a number of other programs that manipulate earthquake data within the SEISAN package are useful tools to assess the parameters that are needed to perform a seismic hazard analysis for an area of interest. Crisis2012 is developed by Mario Ordaz of the Institute of Engineering, UNAM (Ordaz, 1999,1991).

Step by step procedure for seismic hazard analysis

Following is a summary of the steps that need to be completed in order to produce a seismic hazard map.

1. Compile a catalogue for the area of interest from local, regional and global sources. Most commonly, several catalogs are merged. After quality control of the individual catalogs (SELECT, CATSTAT and some of the Seisan Explorer tools can be useful for this) the catalogs can be merged using SPLIT and ASSO.
2. Evaluate the preliminary catalogue completeness by using the 'Catalog completeness' function of Seisan Explorer.
3. Convert magnitudes into one uniform magnitude, preferably to moment magnitude MW. To do this, regression curves must be prepared for different magnitude scales. Program MAG can be used for this purpose.
4. Clean up the catalogue for dependent events (i.e. induced seismicity, non-earthquakes, foreshocks, aftershocks, earthquake swarms). Here a search has to be made for clusters of events both in time and space. Program CLUSTER can be used for this purpose. The probable explosions may be removed by using the program EXFILTER. Following, the 'Poisson distribution' function of Seisan Explorer can be used to confirm that the cleaned catalog is Poisson distributed.
5. The evaluation of the catalogue completeness is dependent upon the clean-up process and the magnitude unification. It is therefore necessary that steps 2-4 be repeated until a reliable catalogue is prepared.
6. Delineate earthquake source zones for the area of interest. The zonation can be based on a seismicity map with the clean catalog and additional information from geology, geophysics, seismotectonics, paleoseismology etc. A seismicity map can be prepared using EPIMAP or GMAP (with Google Earth). Both programs have zooming and area selection procedures which can be used to define the source zone polygons.
7. For each earthquake source zone select the subset of events that fall in the chosen area. This can be done by using the SELECT program. Alternatively, EPIMAP can be used to draw polygons interactively on the screen and put the subset of events within this polygon into a file.
8. The seismicity within each source zone is assumed to be uniform following a Poisson distribution. This can be checked using the 'Poisson distribution' function of Seisan Explorer. For each source zone, the following parameters then need to be defined:
   - The a and b-value of the Gutenberg-Richter relation: these can be defined using the 'Weichert method' or 'Gutenberg-Richter relation' functions of Seisan Explorer or from BVALUE. When applying these programs, magnitude interval and bin sizes must be chosen critically, taking into account the catalogue completeness and the detection threshold.
   - Maximum expected magnitude with its standard deviation: This is usually inferred through other available information, such as geology, palaeoseismicity, or subjective judgment of the scientist.
   - Maximum observed magnitude: This is the largest magnitude observed within the catalogue time span.
   - Threshold magnitude: The so-called lower bound magnitude, which is chosen, based on engineering considerations. Usually magnitudes less than 4.0 are not considered engineering significant.
9. It should be assessed whether there are characteristic earthquakes in the region of interest. This can be done through careful examination of the catalogue and the active faults in the area. If characteristic earthquake sources are included in the analysis, it is important that the seismicity rate is reduced correspondingly in any area sources overlapping with the characteristic fault sources.
10. Ground motion prediction equations (attenuation relations), describing the level of ground shaking as a function of magnitude and distance, must be defined for each source zone. Such relations are based on empirical ground motion observations. If no suitable relations are available for the chosen study area and cannot be derived based on existing data, relations can be adopted from tectonically similar regions. In that case it is important to check the predictions of the chosen relations against available recordings from the study area. Many relations are defined for different types of ground conditions (rock, stiff soil, soft soil) and it is important that appropriate relations are chosen.
11. Run the CRISIS2012 program in order to set up a hazard model with the parameters determined in steps 6-10. The model can be defined through the graphical user interface of CRISIS2012. Optionally, a logic tree can be set up to account for epistemic uncertainty. Results can be displayed as hazard maps or hazard curves directly, or exported as ASCII files. There is also an option for deaggregating the results.
12. The local site conditions should be considered for critical sites. SPEC program can be used to obtain the amplification factors due to unconsolidated sediments. These factors can be used later to adjust the response spectra.

Many of the programs mentioned above are described individually throughout this manual at different sections. In the following the programs that are directly relevant to hazard computations and not described in other sections of the manual are explained in detail.