BBP v22.4.0

The SCEC BBP 22.4 is a full release of the Broadband Platform and includes a significant number of new capabilities and improvements added since the last release of the BBP in June 2019. Below is a brief list of the major improvements and modifications included in this release of the Broadband Platform since version 19.4.0:

• Included multi-segment rupture support for GP, SDSU, Irikura 1, Irikura 2, and Song simulation methods. Currently provided multi-segment validation events are Landers, Ridgecrest 19C, and San Simeon.
• Updated GP low-frequency module (jbsim) to perform time-shifting in the time domain instead of frequency domain. This avoids calculating numerous FFTs and results in a significant processing speedup of 10x, and benefits all simulation modules that use the GP low-frequency code.
• Modified UCSB simulation method to include support for double corner frequency (DCF)
• Included FAS validation module that produces FAS Goodness-of-Fitness plots, for a better seismological interpretation of the results. This is in addition to the PSA validation that is the main metric for design in engineering.
• Added Green's Functions and validation events for the Southern Walker Lane region (Southern Sierra Nevada) of California. Validation events include the 2019 Ridgecrest sequence.
• Expanded the GP GoF module to include a Vs30 GoF plot (similar to the distance and map GoF plots) that can be used to show GoF bias related to stations' Vs30s.
• Updated the GP rupture generator to genslip-5.5.2 and the high-frequency code to hb_high_6.0.5.

All new Broadband Platform users should work with this new version of the Broadband Platform. Existing BBP users are encouraged to upgrade to this latest release of the SCEC Broadband Platform, unless they need to reproduce results produced by a previous version. For more information about the Broadband Platform, or details on the new capabilities of BBP 22.4.0, please refer to the Broadband Platform documentation available on BBP's GitHub wiki pages. Should you have any questions about the installation or operation of the BBP, please do not hesitate to contact the SCEC software development group at software@scec.org.

19.8.0-python3

This is the initial Python 3 version of the Broadband Platform. It has been tested and the results for all methods (except Irikura Recipe 2) are equivalent to results produced by the Python 2 based BBP v19.4. The Irikura Recipe 2 differences are due to changes in that method, and are not due to the change in Python versions. This Python 3 release is a minor release, without updated documentation.

19.4.0

Broadband Platform 19.4.0

This full release of the Broadband Platform includes several new features and bug fixes. Below is a summary of the improvements and modifications included in this release of the Broadband Platform.

Method Updates

• Added the Irikura Recipe Method 2 simulation method to the Broadband Platform. It uses the same rupture generator as the Irikura Recipe Method 1, developed by Arben Pitarka, followed by the GP low-frequency wave propagation code. It then uses a high-frequency module developed by NIED.

• Both Song and Irikura Recipe Method 1 methods now support multi-segment ruptures. These two methods create a version 2.0 SRF file that can be used directly by the GP wave propagation codes and allows a single Broadband Platform simulation to include all segments from a multi-plane rupture.

• Used the GP site response module in all California and Japan validation events.

• Upgraded the SDSU BBToolbox package to version 2.1, with time-domain merging of LF and HF seismograms.

• Re-calculated Greens Functions for California and Japan using 500m/s as reference Vs, bringing the computed synthetic seismograms closer to site conditions and requiring smaller corrections by the site amplification module.

Other Improvements

• Updated the GP rupture generator to genslip-5.4.1 and the GP high-frequency code to hb_high_6.0.3.

• Added Green's Functions and validation events for the Central coast region of California.

• Extended the GP site response module so it can work with all 7 simulation methods.

• Upgraded cluster simulation scripts for use with the Slurm job scheduler. The original collection of scripts using PBS is still available for reference.

17.3.0

Broadband Platform 17.3.0

This full release of the Broadband Platform includes the following features and bug fixes. Below is a summary of the improvements and modifications includes in this release of the Broadband Platform.

Method Updates

• Added the Irikura Recipe Method 1 simulation method to the Broadband Platform. It includes a new rupture generator module developed by Arben Pitarka to generate a SRF file. Then, it uses the GP method wave propagation codes to generate low- and high-frequency seismograms.

• Updated GP rupture generator code to genslip-5.2.2. This new version of genslip accepts three new parameters used for the simulation of multi-segment ruptures, where time series can be generated segment by segment and then added together before post-processing is done.

• Updated GP match.py module to improve the merging of low- and high-frequency seismograms.

General Improvements

• Added a FAS calculation module to the Platform. Work contributed by Jeff Bayless using David Boore's 'smc2fs2' and 'asc2smc' codes. The FAS module produces per-station plots with both N/S and E/W components, along with the smoothed EAS (effective amplitude spectrum) of the two horizontal components.

• Included the calculation of zeta parameter in the RZZ2015 module.

• Added the Central United States simulation region, contributed by Mehrdad Hosseini and Paul Somerville. This includes a new set of Green's functions, calculated up to 1800km.

• Fixed issue in the Anderson GoF and RZZ2015 codes that was causing time series to not align correctly. Thanks to Kim Olsen and Rumi Takedatsu for reporting this bug.

• Added two new scripts: 'merge_multisegment_validation.py' and 'merge_multisegment_scenario.py' to the utils/misc directory. These scripts can be used to combine time series from a number of separately-calculated segments, allowing for a multi-segment rupture to be simulated.

Cluster Improvements

• bbp_hpcc_validation.py and bbp_hpcc_scenario.py include option for user to override default walltime. This allows users to specify a short walltime if they know a job will run quickly, allowing the job to be potentially quickly scheduled by PBS.

• Added a '-s' option to both bbp_hpcc_validation.py and bbp_hpcc_scenario.py to enable the use of the site response module. Currently only the GP site response module is supported (it is used for all methods).

• Several modifications to the cluster scripts to enable multi-segment ruptures to be simulated. Added a '--segment' option to specify the segment number, a '--variation' option to enable the use of different sets of random seeds in the cluster. Also added a '--firstsegment' option to enable the scripts to find the first segment of a multi-segment run so that common seeds can be used for certain parameters across multiple segments (used by the GP method).

• Cluster scripts now save metadata file on top-level simulation directory including all command-line options used to generate the cluster simulation. This is useful to track simulation parameters.

• Added two scripts, 'bbp_merge_multisegment_validation.py' and 'bbp_merge_multisegment_scenario.py' to the utils/batch directory. These scripts work similarly to the ones in utils/misc but can be used to calculate multi-segment runs using segments calculated on the cluster.