Earthquake Potential in California-Nevada Implied
by Correlation of Strain Rate and Seismicity
, Mark D. Petersen
, and Zheng-Kang Shen
U.S. Geological Survey, Golden, CO, USA,
Department of Earth, Planetary, and Space Sciences, University of California, Los
Angeles, CA, USA
Rock mechanics studies and dynamic earthquake simulations show that patterns of seismicity
evolve with time through (1) accumulation phase, (2) localization phase, and (3) rupture phase. We
observe a similar pattern of changes in seismicity during the past century across California and Nevada. To
quantify these changes, we correlate GPS strain rates with seismicity. Earthquakes of M > 6.5 are collocated
with regions of highest strain rates. By contrast, smaller magnitude earthquakes of M ≥ 4 show clear
spatiotemporal changes. From 1933 to the late 1980s, earthquakes of M ≥ 4 were more diffused and broadly
distributed in both high and low strain rate regions (accumulation phase). From the late 1980s to 2016,
earthquakes were more concentrated within the high strain rate areas focused on the major fault strands
(localization phase). In the same time period, the rate of M > 6.5 events also increased signiﬁcantly in the high
strain rate areas. The strong correlation between current strain rate and the later period of seismicity
indicates that seismicity is closely related to the strain rate. The spatial patterns suggest that before the late
1980s, the strain rate ﬁeld was also broadly distributed because of the stress shadows from previous large
earthquakes. As the deformation ﬁeld evolved out of the shadow in the late 1980s, strain has refocused on
the major fault systems and we are entering a period of increased risk for large earthquakes in California.
California hosts one of the most prominent transform boundary faults of the world, the San Andreas Fault
system, which has ruptured in numerous magnitude 7 to 8 events, causing substantial damage and loss to
society. These earthquakes result from stick-slip ruptures on faults in response to nearly constant plate tec-
tonic loading. Because of the nature of stick-slip friction combined with the complex evolution of the fault
network, the occurrence of earthquakes is uncertain and the short-term prediction is difﬁcult.
Averaged over many cycles, however, a complex pattern of earthquake occurrence emerges. The evolution of
patterns through a single seismic cycle are deﬁned by (1) accumulation, (2) localization, and (3) rupture
phases, which have been observed in rock mechanics studies (Hamiel et al., 2004; Lockner et al., 1992) and
model simulations (Ben-Zion & Lyakhovsky, 2002; Lyakhovsky & Ben-Zion, 2009; Mori & Kawamura, 2005).
Seismic observations suggest that following a large earthquake, elastic stresses are dissipated near the
rupture sources causing a stress shadow (Simpson & Reasenberg, 1994) where smaller earthquakes are
distributed broadly (accumulation). With time, stresses in the region evolve out the shadow and reorganize
along the master fault along with corresponding changes in the distribution of seismicity (localization).
When the fault reaches a critical threshold, another rupture occurs and the process repeats (rupture).
Given these time-dependent behaviors and other spatiotemporal clustering patterns, there are certain
predictable features that we should be able to identify to improve earthquake forecasts and mitigate seismic
risk. That is, rather than assuming that the occurrences of earthquakes are time-independent (Poissonian),
these patterns can be analyzed to forecast the rates and sizes of future earthquakes incorporated in the
national seismic hazard maps, which are used to inform public safety on a shorter time scale (Field et al.,
2014; Petersen et al., 2014).
Here we correlate the temporal behavior of seismicity with strain rate models for California and Nevada to
glean information on the earthquake loading (strain rate) and unloading (seismicity) processes. Following
Helmstetter et al. (2007) and Shen et al. (2007), we use strain rates as a predictor of local seismicity.
Evaluation studies of these and other models (Schorlemmer et al., 2010; Zechar et al., 2013) of the
Regional Earthquake Likelihood Model experiment conducted by the Collaboratory for the Study of
Earthquake Predictability have shown that the strain rate-based model by Shen et al. (2007) was one of
ZENG ET AL. 1778
Geophysical Research Letters
• Earthquakes of M > 6.5 occurred in
regions of highest strain rate over the
past century. Its rate has increased
signiﬁcantly since late 1980s
• From 1933 to the late 1980s,
earthquakes of M ≥ 4 were more
diffused and broadly distributed in
both high and low strain rate regions
• From the late 1980s to 2016,
earthquakes were more concentrated
within the high strain rate areas
focused on the major fault strands
• Supporting Information S1
Zeng, Y., Petersen, M. D., & Shen, Z.-K.
(2018). Earthquake potential in
California-Nevada implied by correla-
tion of strain rate and seismicity.
Geophysical Research Letters, 45,
Received 16 OCT 2017
Accepted 23 JAN 2018
Accepted article online 29 JAN 2018
Published online 19 FEB 2018
©2018. American Geophysical Union.
All Rights Reserved.
This article has been contributed to by
US Government employees and their
work is in the public domain in the USA.