Annual Meeting - SCEC.org

Report | SCEC Research Accomplishments
Figure 16. Example of CVM evaluation, using the 2008
Chino Hills earthquake. (top) Map-based GOF for SA-2s,
and (bottom) bias of spectral accelerations at 65 stations
for periods 2-10s.
regional seismic hazards assessment.
54 | Southern California Earthquake Center
other SCEC CVM's. Multiple CVM's are evaluated at one time, so that
performance of a new CVM can be compared to existing, or prior,
CVM versions. The GOF method includes a series of metrics, such as
PGA, PGV, and spectral accelerations (SA) at various periods (Olsen
and Mayhew 2010).
Finally, we recognize that state-of-the-art velocity models do not
resolve small-scale amplification effects in the near-surface sediments,
possibly introducing bias in earthquake ground motion simulations,
as frequencies increase. Preliminary analysis shows that even simple
and rather weak fractal stochastic inhomogenities imply significant
variations in ground motion amplifications (up to a factor of six or
more) as well as de-amplification, including bands of strong
amplification aligned along the average ray path from a 0-2 Hz
horizontally propagating SH-wave source. Simulations with
vertically incident planar SH-wave sources show that the small-scale
heterogeneities included in the upper ~100m of the sediment column
contribute more to the site effects, as compared to small-scale
heterogeneities buried deeper in the sediments. Thus, the USR Focus
Area is supporting efforts to explore how to establish a realistic
stochastic model of near-surface inhomogeneities by comparison of
earthquake ground motion simulations to data and by directly
mapping the statistical properties of shallow Vs estimates. If
incorporated in the CVMs, such models may improve deterministic
ground motion prediction as supercomputers allow the highest
frequency to increase.
Community Fault Model (CFM)
The SCEC Community Fault Model (CFM) is an object-oriented, 3D
representation of more than 140 active faults in Southern California,
and includes direct contributions from more than twenty SCEC
investigators (Plesch et al., 2007). The model consists of triangulated
surface representations (T-surfs) of major faults, which are defined by
surface geology, seismicity, well logs, seismic reflection profiles, and
geologic cross sections. These 3D fault representations are intended to
support SCEC research efforts in fault system modeling and
earthquake rupture propagation, as well as to serve as a basis for
This past year, we completed a major update of the model (CFM v.4.0), which incorporates improvements in 3D fault
representations, a detailed fault surface trace layer, and a new naming and numbering scheme for individual 3D fault models
that allows for direct connections to the USGS/CGS Quaternary Fault database (Qfaults) and other SCEC data sets (Nicholson
et al., 2011). Fault representations in CFM are now referenced to the modern WG84 datum and the new surface layer in CFM
allows 3D fault models to be registered to the more detailed Qfaults digital trace maps. Systematic revision of CFM 3D fault
segments was required to ensure compatibility between some previous CFM fault representations and the newer Qfaults
surface traces, as well as by the availability of extensive catalogs of relocated earthquake hypocenters to define better the
subsurface geometry of active faults. New 3D fault representations for major fault zones include the San Andreas from San
Gorgonio Pass to the Salton Sea, the Mecca Hills, San Jacinto, Elsinore-Laguna Salada (including El Mayor-Cucapah), and an
Fernando/Sierra Madre fault systems. In addition, a new SCEC fault database hierarchical naming and numbering scheme is
implemented that provides unique identifiers (number, name, abbreviation) for each level of the fault hierarchy under which a
particular fault segment is classified. Levels of fault hierarchy include Fault Area, Fault Zone or System, Fault Section, Fault
Name, Fault Strand or Model, and Fault Component. These additional fault hierarchical levels allow for more flexible