PRESENT-DAY LOADING RATE OF SOUTHERN SAN ANDREAS AND
EASTERN CALIFORNIA SHEAR ZONE FAULTS FROM GPS GEODESY
by
Joshua C. Spinler
A Prepublication Manuscript Submitted to the Faculty of the
DEPARTMENT OF GEOSCIENCES
In Partial Fulfillment of the Requirements
for the Degree of
MASTER OF SCIENCE
In the Graduate College
THE UNIVERSITY OF ARIZONA
2008 JOURNAL OF GEOPHYSICAL RESEARCH, VOL. ???, XXXX, DOI:10.1029/,
1
2
3
Present-day loading rate of southern San Andreas
and eastern California shear zone faults from GPS
geodesy
1
1
2
Joshua C. Spinler, , Richard A. Bennett, , Megan L. Anderson, , and Sigrún
1
Hreinsdóttir,
J.C. Spinler, Department of Geosciences, University of Arizona, Tucson, AZ 85721-0077, USA
(jspinler@email.arizona.edu)
1
Department of Geosciences, University of
Arizona, Tucson, Arizona, USA.
2
Department of Geology, Colorado
College, Colorado Springs, Colorado, USA.
D R A F T
August 13, 2008, 3:02pm
D R A F T
X -2
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
4
Abstract.
We present new results from a dense network of Global Po-
5
sitioning System (GPS) stations located in the eastern Transverse Ranges
6
Province (ETR), a transition zone between the southern San Andreas fault
7
(SSAF) and eastern California shear zone (ECSZ). We determined crustal
8
velocities at 68 sites using data from a new 24-station campaign network,
9
37 nearby continuous stations from the SCIGN and PBO networks, and 7
10
campaign sites from within the San Bernardino Mountains. The Joshua Tree
11
Integrated Geodetic Network (JOIGN) campaign data set consists of seven
12
multi-day measurement campaigns conducted each May, September, and Febru-
13
ary over the ∼2.0 year period from September 2005 to September 2007. We
14
combined the JOIGN and other campaign data from 2005 to 2007 with the
15
continuous data for the period of 1994 to 2007. We used the GPS-determined
16
site velocity estimates to constrain elastic loading rates for four fault-block
17
scenerios. For all models, we estimated relative block motions accounting for
18
elastic strain on locked faults using weighted least squares such that com-
19
mon mode rotation and translation attributable to reference frame uncer-
20
tainty contribute no power to the misfit. We tested models that represent
21
the Pinto Mountain and Blue Cut faults of the ETR, a new NNW striking
22
fault (the ”Landers-Mojave earthquake line”) that cuts obliquely across the
23
ETR and Mojave Desert faults, and models including various combinations
24
of these end-member fault-block models. All models produce fault loading
25
rate estimates for the SSAF that are consistent with the majority of estimates
26
for intermediate- and short-term fault slip rates based on tectonic geomor-
DRAFT
August 13, 2008, 3:02pm
DRAFT
X-3
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
27
phology and paleoseismology, respectively. However, our slip rate estimates
28
of 5-22 mm/yr are markedly different from the overall average (∼5 Ma) slip
29
rate of 30 mm/yr. ECSZ, in contrast, rate estimates based on geology and
30
geodesy disagree on all time-scales. Constructive and destructive interfer-
31
ence of strain fields associated with the major faults comprising complex fault
32
systems results in strain rate patterns that can equally well be explained by
33
minor or umapped faults if geodetically determined velocity gradients are
34
associated with patterns of seismic energy release independent of mapped
35
faults.
D R A F T
August 13, 2008, 3:02pm
D R A F T
X-4
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
1. Introduction
36
Potential for a magnitude ≥7.5 earthquake is generally regarded as high along the south-
37
ernmost San Andreas fault (SSAF) zone in southern California (Sykes and Seeber, 1985;
38
Working Group on California Earthquake Probabilities, 1995; Bennett et al., 1996; Fi-
39
alko, 2006). Simulated ground shaking assuming rupture of the Coachella Valley and San
40
Bernardino strands of the fault during a single great earthquake demonstrate significant
41
hazards for major population centers in San Bernardino and Los Angeles counties (Olsen
42
et al., 2006). However, two of the largest southern California earthquakes in the past
43
several decades, the 1992 Mw 7.3 Landers and 1999 Mw 7.1 Hector Mine earthquakes, oc-
44
curred instead in the Mojave Desert portion of the eastern California shear zone (ECSZ)
45
adjacent to the southern San Andreas fault. One interpretation of these events, and other
46
smaller events such as the 1947 Mw 6.5 Manix, 1948 Mw 6.0 Desert Hot Springs, 1986 Mw 5.6
47
North Palm Springs, and the 1992 Mw 6.1 Joshua Tree and Mw 6.5 Big Bear earthquakes
48
(Figure 1), is that they represent a natural, but rather diffuse component of the deforma-
49
tion cycle characterizing the plate boundary zone (e.g., Press and Allen, 1995; Seeber and
50
Armbruster, 1995). According to this view, this spatially-diffuse, temporally-intermittent
51
seismicity does not contribute appreciably to the long-term pattern of deformation which
52
is dominated by infrequent large to great earthquakes on the main plate boundary struc-
53
tures such as the SSAF. If correct, contemporary seismic hazard along the southern San
54
Andreas fault, which has not ruptured in recent time, would be considered high depending
55
on the rate at which elastic strain is accumulating on the fault. On the other hand, the
56
pattern of recent earthquakes has stimulated vigorous discussion of alternative models
D R A F T
August 13, 2008, 3:02pm
D R A F T
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
X-5
57
for plate boundary kinematics involving present-day reorganization of the SSAF-ECSZ
58
system (Nur et al., 1993; Du and Aydin, 1996; Li and Liu, 2006), consistent with a reduc-
59
tion of slip rate on the portion of the fault zone between Cajon Pass and Biskra Palms.
60
If new ECSZ faults are forming at the expense of this portion of the San Andreas fault
61
zone, the likelihood of a throughgoing rupture involving both the Coachella Valley and
62
San Bernardino segments may need to be reconsidered. Moreover, geological observations
63
pertaining to time scales of millions of years would have little bearing on contemporary
64
seismic hazards associated with the greater San Andreas fault system.
65
Important tests of these competing hypotheses can be obtained from a detailed un-
66
derstanding of the recent slip history of the SSAF and adjacent fault zones. Slip rate
67
is an important parameter for active tectonics in large part because the rate at which
68
slip accrues is a function of the frequency and magnitude of past–and by extrapolation
69
future–earthquakes. However, the extent to which past slip rates provide an accurate
70
characterization of the present and future fault behavior is complicated by the fact that
71
the upper crustal portions of seismogenic faults are generally everywhere locked except
72
during very brief periods when accumulated elastic strain is recovered by seismic rupture
73
or aseismic sliding. That is, strictly speaking, present-day slip rates on the seismogenic
74
upper crustal portions of faults are almost always zero and thus of no practical value for
75
hazards assessment and geodynamics research.
76
In a spatially-complex rapidly-deforming plate boundary zone, which experiences fre-
77
quent large earthquakes, such as the Pacific-North America boundary zone of southern
78
California, determination of fault loading rate by geodetic methods is complicated by
79
(1) frequent coseismic displacements, (2) the complex geometry of upper crustal fault
D R A F T
August 13, 2008, 3:02pm
D R A F T
X-6
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
80
systems, and (3) the unknown rheological properties of the crust and upper mantle. Fre-
81
quent coseismic earthquakes hamper determination of precise interseismic site velocities
82
by displacing monuments, and thereby introducing additional parameters that need to be
83
determined from the data. When upper crustal faults lie in close proximity their elastic
84
strain fields are superimposed and thus difficult to discriminate from one another without
85
knowledge of detailed subsurface fault geometry. In addition, viscous relaxation of earth-
86
quake induced stress concentrations in the lower crust and/or upper mantle may modify
87
the secular strain rate field for years following a large earthquake, such that crustal mo-
88
tions associated with elastic loading are difficult to separate from motions associated with
89
ongoing relaxation (Savage, 1990).
90
The rate of interseismic elastic loading determined by geodesy provides an important
91
complement to direct measurement of past fault slip, with similar application to present-
92
day mechanics and hazards. Assuming that elastic strain within a fault system will
93
eventually be released by displacement along faults, geodesy provides an indication of
94
how far field plate motion will be partitioned as slip within the plate boundary fault
95
system. Moreover, the rate at which any particular fault is strained by loading of elastic
96
strain on the crust may be considered as a proxy for the rate of fault slip that will be
97
accrued over one or more future earthquake cycles.
98
However, competing hypotheses for present-day behavior of the SSAF zone are diffi-
99
cult to discriminate using existing slip rate data given the wide range of rate estimates
100
reported in the literature (Figure 2). Apparent differences can be explained by (1) mea-
101
surement errors not accounted for by reported ranges of uncertainty (e.g., Bird, 2007),
102
(2) temporal changes in slip such that geological and geodetic inferences recording aver-
D R A F T
August 13, 2008, 3:02pm
D R A F T
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
X-7
103
ages over different periods reveal different rates (e.g., Bennett, 2007), or (3) along strike
104
variation in slip rate, such that point estimates determined by geological methods ap-
105
pear to contrast with one another and/or with the spatially averaged rates determined
106
by geodesy. The first of these possibilities may be time dependent in the sense that new
107
and improved measurement and modeling techniques should result in new precise slip rate
108
estimates that supersede older, less precise estimates. Thus, published estimates may be
109
analyzed historically according to publication date, according to measurement technique
110
and associated time-scale, and/or according to location along strike.
111
In this paper we use data from a new dense GPS network in the eastern Transverse
112
Ranges Province (ETR), together with data from neighboring continuous GPS stations,
113
to investigate the pattern of elastic strain accumulation on southern San Andreas and
114
ECSZ faults, with the goal of exploring along strike variation in fault loading rates along
115
the SSAF. In section (2), we provide a review of the data and arguments upon which the
116
various working hypotheses for southern California fault kinematics are based. In section
117
(3), we describe the GPS data set that forms the basis of our study. In section (4) we
118
provide a detailed description of our fault modeling procedure. In section (5) we discuss
119
the results of our modeling and their implications for fault kinematics and seismic hazards
120
in southern California.
2. Slip Rate Debates
121
Initial slip rate estimates for the time period from the latest Pleistocene through to the
122
present for the northern segments of the southern SAF indicate rates near Biskra Palms
123
(Figure 1) ranging from 23-35 mm/yr (Keller et al., 1982) and for the San Bernardino
124
strand to slip at 23.5 ± 3.5 mm/yr (Weldon and Sieh, 1985) based on dating of offset
D R A F T
August 13, 2008, 3:02pm
D R A F T
X-8
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
125
alluvial fan deposits. Harden and Matti (1989) suggest a slip rate of 14-25 mm/yr for
126
the San Bernardino strand of the SAF over the past 14,000 years through dating of soil
127
development on offset alluvial surfaces. Early geodetic determinations of the slip rate for
128
the SSAF (e.g., Bennett et al., 1996) were comforting and suggested that the slip rate
129
along strike was constant, and that elastic loading was recovered by slip on the fault.
130
Refined estimates for fault slip rate on the SSAF zone based on GPS and InSAR geodetic
131
measurement are slightly lower though in agreement to within errors with the earlier
132
geodetic estimates: 23.0 ± 8 mm/yr (Becker et al., 2005), 23.5 ± 0.5 mm/yr (Meade and
133
Hager, 2005), 25 ± 3 mm/yr (Fialko, 2006), and 22.3 ± 0.7 mm/yr (Fay and Humphreys,
134
2005). Estimates for SSAF slip rate based on strain rates derived from trilateration,
135
however, span a larger range from about 15 to 25 mm/yr (Lisowski and Savage, 1991;
136
Johnson et al., 1994; Anderson et al., 2004). Part of this discrepancy may be related to
137
the way in which line length changes were interpreted. Lisowski and Savage (1991) infer
138
velocities such that fault perpendicular motions are minimized. Similarly, Anderson et al.
139
(2004) use subnetworks to infer spatially-averaged strain rates.
140
Geodetic slip rate estimates for the San Bernardino strand of the San Andreas fault
141
have been determined in fewer studies. Meade and Hager (2005) and Becker et al. (2005)
142
determine right lateral slip rates in the range of about 5-6 mm/yr, significantly lower than
143
the previous rates. Subsequent reanalysis of the Biskra Palms site (van der Woerd et al.,
144
2006) revealed that the rate provided by Keller et al. (1982) was an overestimate. The
145
differences between slip rate values for different locations along the SAF can be explained
146
either by along strike variation in slip rate, temporal variation in slip rate, or by errors in
147
one of more of the data sets. Slip rate variations through time have been inferred for the
D R A F T
August 13, 2008, 3:02pm
D R A F T
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
X-9
148
San Andreas fault over long time periods (Bennett et al., 2004), but slip rate variation
149
over shorter periods of time remain an open question. However, Matmon et al. (2005)
150
investigate a rich slip rate data set for the central San Andreas fault finding that rates
151
estimated for a common location on the fault are all consistent over time scales of 10 kyr
152
to 100 kyr. If slip rate variations on the central San Andreas fault are constant over late
153
Pleistoene time scales, then why not the SSAF?
2.1. ECSZ
154
The geology geodesy slip rate debate isn’t just limited to the southern SAF, but can be
155
investigated for the ECSZ. Trilateration data for the ECSZ prior to the large earthquakes
156
of the 1990s indicated rates of shear strain that were unexpectedly high (Dokka and Travis,
157
1990; Savage et al., 1990; Sauber et al., 1994). Block modeling of geodetic data for the
158
same region show a high dependence on the spacing of the model faults used (Peltzer et
159
al., 2001; Oskin and Iriondo, 2004; Meade and Hager, 2004; Meade and Hager, 2005),
160
while Rockwell et al., (2000) find Holocene slip rates that are an order of magnitude lower
161
than those interpreted from geodetic data. There is also debate on the duration and radius
162
of the effect of postseismic strain accumulation. Freed et al., (2007) find long distance
163
transients associated with the Hector Mine earthquake, whereas Argus et al. (2005) find
164
no long term change in continuous GPS stations at a regional distance scale. The present
165
era of spatially clustered earthquakes in the Little San Bernardino Mountains-Mojave
166
Desert region, and possibly similar clusters in the past (e.g., Rockwell et al., 2000), as
167
well as accelerated present-day crustal deformation relative to long-term geological rates
168
(Peltzer et al., 2001; Oskin and Iriondo, 2004; Oskin et. al., 2008) may also be understood
D R A F T
August 13, 2008, 3:02pm
D R A F T
X - 10
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
169
in terms of stress transfer following these and other recent large earthquakes (e.g., Toda
170
et al., 2005).
3. GPS data
3.1. Campaign GPS
171
To help constrain the rates of slip on the SSAF, particularly the region of the fault be-
172
tween Biskra Palms and Cajon Pass, as well as the faults of the southernmost ECSZ, we
173
established a new dense campign-style GPS network in the eastern Transverse Range
174
Province (ETR). The Joshua Tree Integrated Geodetic Network (JOIGN) consists of
175
12 new stations located within Joshua Tree National Park. We have also adopted sev-
176
eral existing stations within the ETR. Since September 2005, we have performed seven
177
campaign surveys of the network roughly equally spaced in time between September
178
2005 and September 2007. We analyzed these campaign data together with continu-
179
ous GPS data from the southern California region (Figure 3, Table 1, Table 2) using the
180
GAMIT/GLOBK software (King and Bock, 2005; Herring, 2005). Time series of station
181
position estimates in the Stable North America Reference Frame (SNARF 2000, Blewitt
182
et al., 2006) are shown in Figure 4. There is little indication of seasonal variation, and
183
the motion of each site through time is well approximated by a straight line. The slopes
184
of the best fit lines provide an estimate of secular site velocity. Figure 5 shows a subset of
185
the newly installed JOIGN network, for which we compare velocity results obtained for
186
various subsets of the total dataset. We determined site velocities for each of the three sea-
187
sons for which data has been collected to date, February, May, and September. Variation
188
amongst the velocity estimates record potential fluctuations and measurement errors in
189
the amplitudes for each season, independent of perfectly periodic seasonal deformations.
D R A F T
August 13, 2008, 3:02pm
D R A F T
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
X - 11
190
Our results agree with one another quite well, with a WRMS of ∼1 mm/yr and NRMS ∼1
191
(Table 3). Despite there being one less campaign each for February and May compared to
192
September (two, two, and three respectively), each velocity is virtually indistinguishable,
193
within error, from both the total velocity determined for each site and from the mean
194
velocity calculated from the individual seasonal values. Positional estimates for the full
195
JOIGN network are also in good agreement with one another with a WRMS value of ∼2
196
mm and a NRMS of ∼1 (Table 4).
197
In addition to the JOIGN campaign dataset, we analyzed campaign sites located within
198
the San Bernardino Mountains to the northwest (Figure 3). These sites have been observed
199
between 1 and 6 times over the period from 2003 to 2007 (McGill, pers. comm., 2007) .
200
We reobserved a select set of these sites in October 2007.
3.2. Continuous GPS
201
We have selected a number of CGPS sites within our study area, in an effort to better
202
constrain the rates of slip on the faults. We have analyzed GPS data from 37 continuous
203
sites for the time period 1994-2007, with 23 sites from the SCIGN network and 14 sites
204
from the Plate Boundary Observatory (PBO) GPS network. The resulting velocities are
205
shown in Figure 3. The formal uncertainties for these continuous stations are usually
206
≤1 mm/yr, with the majority of the older SCIGN sites having uncertainties # 1 mm/yr
207
(Table 2). Time series for a subset of the PBO GPS network sites show little to no seasonal
208
variation in the signal, and the individual sites velocities can be well approximated using
209
a straight line fit to the data (Figure 6).
4. Fault Sip Estimation
D R A F T
August 13, 2008, 3:02pm
D R A F T
X - 12
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
4.1. Block models
210
In order to investigate crustal motions along the SSAF and into the ECSZ, we adopted
211
an elastic half-space block model, similar to the one used by Bennett et al. (1996). In this
212
approach, the Earth is represented as a series of infinitely deep, elastic (Poisson) blocks in
213
contact along their edges, or ”faults”, dividing up an elastic half-space. Each block moves
214
independently of the other blocks, and undergo temporally uniform, horizontal motion.
215
The fault-parallel component of velocity between each block is set to be continuous from
216
the surface to a pre-described locking depth, below which it is allowed to be discontinuous.
217
The fault-perpendicular velocity can be discontinuous across the fault boundary. This
218
results in the velocity field for each block describing rigid-body translation, or ”block
219
rotation”, plus spatially variable deformation that decays to 0 as depth increases. We
220
chose our faults to be simplified versions of mapped, and in some cases unmapped, faults
221
within the SSAF and ECSZ systems. Each fault is described as a continuous, vertical
222
planar surface, while all locking depths were assigned to be 15 km, which is in the range
223
of published values for the region (Fialko, 2006; Meade and Hager, 2005; Bennett et al.,
224
1996). Known earthquakes of magnitude > 7 in the ETR have ocurred on at least one
225
of the faults in our models, whereas earthquakes in block interiors have magntude < 5
226
(Figure 7).
227
Previous models for strain accumulation in this region (Bennett et al., 1996; Meade and
228
Hager, 2005; Becker et al., 2005; McCaffrey et al., 2005; Bennett et al., 2007; Pollitz et al.,
229
2008) are likely not sufficiently parameterized to account for the new high-precision, high-
230
resolution velocity field we are developing. None of the previous models have realistically
231
accounted for mapped ETR faults located to the south of the Pinto Mountain fault.
D R A F T
August 13, 2008, 3:02pm
D R A F T
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
X - 13
232
Through our block models, relative block motions are automatically resolved onto the
233
faults that form the boundaries. Each of the GPS sites input into the model are automati-
234
cally assigned to the block in which they are located. From here, we have the ability to let
235
each of the blocks rotate. This is done through a parameterization of the individual block
236
motions into local translations and rotations about a vertical axis, which is equivalent to
237
an Euler pole. In our parameterization, we are able to constrain the rotational component
238
of each individual block, which is typically the most poorly known component of an Euler
239
vector, through the rotation about the vertical axis through the centroid of each block.
240
The axis orthogonal to the Euler pole describes the translation associated with each block.
241
The individual block motions are then determined by removing the contribution of the
242
locked faults from the calculated model velocities (Okada, 1985).
243
We chose to model the SSAF-ECSZ system by selecting four different fault-block sce-
244
narios (Figures 8, 9, 10, and 11) that incorporate known faults in the region. Model 1
245
includes the mapped Blue Cut, Sheephole, and Chiriaco faults, and the blocks of the ETR
246
are allowed to rotate about a vertical axis based on the transrotation model of Dickinson
247
(1996). We used published gravity data (Langenheim et. al., 2007) (Figure 12) to better
248
constrain our fault locations within the ETR, and in particular the Joshua Tree National
249
Park region. Model 1 contains 7 crustal blocks with 14 planar fault segments (Figure 8a).
250
Model 2 includes only 5 planar fault segments, and 4 crustal blocks (Figure 9a), and serves
251
as the opposing end member to Model 1. The ETR and ECSZ faults in Model 2 were de-
252
termined by patterns of seismicity over the past ∼60 years (Figure 7), and primarily follow
253
the rupture associated with the Joshua Tree and Landers earthquakes of 1992. The two
254
new faults used in this model represent the recent development of a through-going, NNW
D R A F T
August 13, 2008, 3:02pm
D R A F T
X - 14
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
255
trending, right-lateral strike-slip fault determined by the pattern of seismicity resulting
256
from the 1992 Landers, and associated earthquakes, referred to as the Landers-Mojave
257
line by Nur et al., 1993. Model 3 is similar to Model 1, in that is uses mapped faults
258
in the ETR and ECSZ and the ETR crustal blocks are allowed to rotate, but shifts the
259
easternmost location of fault slip westward to the Pisgah fault system of the ECSZ, and
260
an umapped set of faults that continue from the termination of the Pisgah in the south,
261
through eastern Joshua Tree NP (Figure 10a). This model tests the recent westward shift
262
in tectonism proposed by Dokka and Travis (1990). Model 4 is a combination of the
263
transrotation model of Dickinson (1996) shown in Model 1 and the Landers-Mojave line
264
model of Nur et al. (1993) shown by Model 2 (Figue 11a). This model is designed to best
265
match both the mapped faults and gravity anomalies of Model 1, while also allowing for
266
the presence of a newly developing through-going fault along the Landers/Joshua Tree
267
rupture lines.
268
We have chosen to not model the San Jacinto Fault in this investigation because our
269
GPS network does not include any sites that are located to the south and west of the San
270
Jacinto Fault. Therefore, if the fault was modeled in its mapped location, we would have
271
no constraints on the block to the southwest of the fault.
4.2. Estimation results
272
4.2.1. Transrotation Model
273
The transrotation model (after Dickinson, 1996) can account for the majority of the
274
observed velocity field (Figure 8b), with a mean residual velocity, which is found by
275
subtracting the model velocities from that of the observed velocities at each point, of
276
0.81 ± 0.01 mm/yr. The model velocity estimates have a χ2 /DOF = 9.15, where χ2 is
D R A F T
August 13, 2008, 3:02pm
D R A F T
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
X - 15
277
the residual sum of squares and DOF is the number of degrees of freedom. 26.5% of the
278
residual velocity rates are smaller than their 2-σ uncertainty estimates (Figure 8c). The
279
largest residual velocities are found in the eastern part of the ETR blocks, as well as a
280
couple of stations located very near to faults. The large residuals in the eastern portions
281
of the ETR blocks indicate that additional faults or blocks may be needed to produce a
282
better fit to the data.
283
The results show a northward along-strike reduction in the slip rates along the different
284
strands of the SSAF (Table 5, Figure 13), from 21.5 ± 0.1 to 19.0 ± 0.1 mm/yr of right-
285
lateral motion along the segments of the Coachella Valley strand, to 6.2 ± 0.1 mm/yr on
286
the Little San Bernardino strand, and 3.9 ± 0.1 mm/yr for the Mill Creek strand. The
287
ECSZ faults to the north of the Pinto Mountain fault also display right-lateral motion,
288
with the Emerson fault accommodating 12.7 ± 0.1 mm/yr and 5.2 ± 0.1 mm/yr for the
289
Pisgah fault. The east-west trending Pinto Mountain fault displays left-lateral motion
290
(7.5 ± 0.1 mm/yr) along the segment closest to its intersection with the SSAF. The sense
291
of motion estimated for the other two sections of the Pinto Mountain fault show right-
292
lateral motion increasing in rate to the east. This reverse in sense of motion along the
293
Pinto Mountain fault can be explained by the rotation of the ETR blocks. The Blue
294
Cut fault is accommodating 12.1 ± 0.2 mm/yr of left-lateral motion, while the Chiriaco
295
fault to the south is only slipping right-laterally at 1.0 ± 0.1 mm/yr. The Sheephole
296
fault segments, along with the Ludlow fault to the north, show a left-lateral sense of
297
motion, opposite to the sense of motion of the other NW-SE trending strike-slip faults
298
of the SSAF and ECSZ. However, the left-lateral sense of motion is explained through
299
the rotation about an Euler pole of the ETR crustal blocks. Rotation of the ETR blocks
D R A F T
August 13, 2008, 3:02pm
D R A F T
X - 16
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
300
(Figure 8e) fits well with the observed data in the western half of the blocks, yet doesn’t
301
match the sparse data points located to the east.
302
4.2.2. Landers-Mojave Line Model
303
The block model for the Landers-Mojave earthquake line, proposed by Nur et al. (1993)
304
accounts for the observed velocity field about equally well as the Transrotation model
305
discussed previously (Figure 9b), with a mean residual velocity of 0.94 ± 0.1 mm.yr. This
306
model produces a χ2 /DOF = 9.76, and 27.9% of the residual velocity estimates are smaller
307
than their 2-σ uncertainty estimates (Figure 9c).
308
This model also produces a reduction in slip rates along strike in the SSAF, as well
309
as a partitioning of that slip onto the San Andreas strands between Biskra Palms and
310
Cajon Pass, as well as the ECSZ faults of the Mojave desert (Table 6, Figure 13). The
311
southernmost segment of the San Andreas (Coachella Valley strand) is slipping right-
312
laterally at a rate of 21.4 ± 0.1 mm/yr. In the vicinity of Biskra Palms, the faults
313
separates into 2 branches, one heading NW towards Cajon Pass, and the other heading
314
primarily northwards, into the ECSZ. The rates calculated for these two sections show the
315
majority of the right-lateral motion is being accommodated by the ”Joshua Tree” fault,
316
(14.7 ± 0.1 mm/yr) and then onto the Landers-Mojave line fault (14.2 ± 0.1 mm/yr),
317
while the Little San Bernardino and Mill Creek strands of the SSAF only slip with rates
318
of 7.6 ± 0.1 and 5.3 ± 0.1 mm/yr, respectively. In this fault orientation, the 3 crustal
319
blocks to the west of the SSAF-LML line are all moving in a northward direction relative
320
to North America (Figure 9e). The largest residuals in this model are located in the ETR
321
and ECSZ regions. These large residual values east of the through-going LML, indicate
D R A F T
August 13, 2008, 3:02pm
D R A F T
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
X - 17
322
that the crustal motions in this region cannot be defined in this simple case. It appears
323
that more blocks are required in order to best fit the observed data to the predicted rates.
324
4.2.3. Modified Transrotation Model
325
Model 3 is a modified version of Model 1, with the eastern boundary of active deforma-
326
tion shifted westward within the ECSZ, as proposed by Dokka and Travis (1990). This
327
model fits the observed velocity field (Figure 10b) only slightly worse than Model 1, with
328
mean residual estimates of 0.89 ± 0.1 mm.yr (Figure 10c), with the majority of the largest
329
residuals located along the eastern Blue Cut fault. By bringing the easternmost edge of
330
deformation to the west, χ2 /DOF reduces to 8.77, compared to Model 1’s value of 9.15.
331
It also had 26.5% of its residual velocities smaller than their 2-σ uncertainty estimates.
332
This model predicts clockwise rotation of the ETR range blocks. This rotation can be
333
observed in the velocity field near the SAF, while there are no data points that predict
334
a southwestward sense of motion located along the eastern boundary of the ETR blocks
335
(Figure 10e).
336
This model predicts slightly lower slip rates along the various strands of the SSAF (Table
337
7, Figure 13), ranging from 21.0 ± 0.1 and 16.4 ± 0.1 mm/yr right-lateral motion for the
338
Coachella Valley strands, to 5.7 ± 0.1 and 3.2 ± 0.1 mm/yr for the Little San Bernardino
339
and Mill Creek strands, respectively. The reduction in slip rates along the SSAF is shown
340
to be taken up by the two NW-SE trending ECSZ faults, with the Emerson and Pisgah
341
fault groups accommodating 13.3 ± 0.1 and 2.6 ± 0.1 mm/yr of right-lateral motion
342
respectively. The Pinto Mountain fault segment (PMT1, Figure 10a), Blue Cut, and
343
Chiriaco faults all slip left-laterally. The eastern Pinto Mountain fault segment (PMT2,
344
Figure 10a) is slipping with a right-lateral sense of motion (0.3 ± 0.1 mm/yr). This switch
D R A F T
August 13, 2008, 3:02pm
D R A F T
X - 18
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
345
in sense of motion is caused by the differential rotation of the block to the south of the
346
fault (Figure 10e). In this fault orientation, the Blue Cut fault in the ETR iare slipping
347
left-laterally at a rate of 10.9 ± 0.2 mm/yr, with the Chiriaco fault to the south having
348
2.3 ± 0.1 mm/yr of left-lateral slip. Unlike model 1, where the entire set of NW-SE
349
trending faults along the eastern boundary showed left-lateral motion, the easternmost
350
faults of this model aren’t accommodating slip in the same sense of motion across the
351
Pinto Mountain fault. In this model, the ”Eastern Joshua Tree” faults are determined to
352
have 1.0 ± 0.1 and 0.7 ± 0.1 mm/yr of left-lateral slip, north to south, respectively.
353
4.2.4. Transrotation with the Landers-Mojave Line Model
354
Model 4 combines the east-west trending faults of the ETR from the transrotation model
355
of Dickinson (1996), while also allowing for the presence of the northerly Landers-Mojave
356
earthquake line of Nur et al. (1993). This orientation of the faults produced the lowest
357
mean residual velocities, 0.64 ± 0.1 mm/yr, highest percentage of the residuals less than
358
their 2-σ uncertainty values, 29.4%, and the lowest χ2 /DOF value (7.50) of all the models
359
tested (Figure 11c). By adding the three faults comprising the Landers-Mojave line (JT1,
360
JT2, and LML, Figure 11a), it significantly reduced both the residual velocites, but also
361
reduced the amount of rotation needed by the crustal blocks of the ETR to best fit the
362
observed data (Figure 11e).
363
This model predicts the highest slip rates for the SSAF, with rates of 21.4 ± 0.1, 18.8 ±
364
0.1, 10.6 ± 0.1, and 5.6 ± 0.1 mm/yr of right-lateral motion from south to north (Table 8,
365
Figure 13). The higher rates along the northern segments of the SSAF cause a reduction
366
in slip available for the faults of the LML, with rates of 7.6 ± 0.2, 6.7 ± 0.2, and 10.1
367
± 0.1 mm/yr for the JT1, JT2, and LML fault sections. In this model, the primarily
D R A F T
August 13, 2008, 3:02pm
D R A F T
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
X - 19
368
east-west trending faults of the ETR have relatively small slip rates, compared to the
369
previous models (Figure 11e). The Blue Cut and Chiriaco faults to the south are both
370
slipping with a left-lateral sense of motion at rates of 3.2 ± 0.2 and 0.5 ± 0.1 mm/yr,
371
respectively. Similarly to Model 1, all of the faults that comprise the boundary between
372
the ETR and ECSZ with ”stable North America” are slipping left-laterally, with small
373
rates ∼1-2 mm/yr for each section.
5. Discussion
374
We use both statistical analysis as well as an evaluation of the slip rates for each model to
375
determine model ”fit” to the observed data. The Bayesian information criterion (BIC) is a
376
useful measure of the ”goodness” of each model. In this approach, the BIC is determined
377
as
BIC = n ln
!
χ2
n
"
+ k ln(n),
(1)
378
where n is the number of observations, χ2 is the residual sum of squares, and k is equal to
379
the number of free parameters, which can be found by subtracting the degrees of freedom,
380
or independent model parameters, from the number of observations. The results for the
381
BIC calculations for each model are shown in Table 9. Given any number of models,
382
the one with the lowest BIC is the one that is preferred to best fit the observations.
383
The two models that include the Landers-Mojave earthquake line are favored by the BIC
384
investigation, with values ∼5-6% lower than that of the transrotation model, and a ∼2%
385
lower than that of the modified transrotation. Model 4 had the lowest χ2 /DOF value by
D R A F T
August 13, 2008, 3:02pm
D R A F T
X - 20
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
386
∼16-30% compared to the other models, while also having the lowest mean residual rates
387
(Table 9).
388
We also investigated the rates calculated for the individual fault segments to compare
389
what their resulting geologic expression should be as well as to compare them to previ-
390
ously published geodetically determined rates. Rotation of the ETR blocks creates large
391
extensional and compressive rates for the faults that bound those blocks (Figure 14, Ta-
392
bles 5, 6, 7, and 8), for both the transrotation and modified transrotation models. For
393
Model 1, the northernmost Sheephole fault is predicted to be opening at a rate of 14.0 ±
394
0.2 mm/yr. As there are no apparent structures in this location that could account for
395
this rate of extension, we question the validity of this model. The rotation of the ETR
396
blocks also results in some other suspect slip rates, as seen by the 12.1 ± 0.2 mm/yr of
397
left-lateral motion calculated for the Blue Cut fault. At this rate, the Blue Cut fault
398
should be the major accommodator of slip compared to that of the presumed major E-
399
W trending structure that serves as the boundary between the ETR and the ECSZ, the
400
Pinto Mountain fault (Meade and Hager, 2005). The same argument can be applied to
401
the modified transrotation model, which has the northern Sheephole fault opening at a
402
rate of 8.6 ± 0.1 mm/yr and producing 10.9 ± 0.2 mm/yr of left-lateral motion on the
403
Blue Cut fault. The models that include the Landers-Mojave earthquake line produce
404
moderate to large opening rates for the LML of 6.4 ± 0.1 and 4.7 ± 0.1 mm/yr (LML
405
and Transrotation plus LML models, respectively). These large rates of opening can be
406
reduced by changing the strike of the fault to a more northwesterly direction, as shown
407
by the rates calculated for the Emerson fault group in the other models (0.6 ± 0.1 and
D R A F T
August 13, 2008, 3:02pm
D R A F T
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
X - 21
408
0.9 ± 0.1 mm/yr of opening for the transrotation and modified transrotation models,
409
respectively).
410
The strike-slip rate of the Blue Cut fault in the transrotation with LML model is
411
comparable to that of the Pinto Mountain fault, as opposed to being ∼2 times larger for
412
the models without the LML. All models predict 3.0-4.8 mm/yr of closing for the Mill
413
Creek strand of the SAF, which compares well to rates determined through cosmogenic
414
dating of uplift for the San Bernardino Mountains (Spotila et al., 2001).
415
All of the models show a northwestward reduction in the amount of accommodated slip
416
for the SAF, from values of ∼21 mm/yr near the Salton Sea, ∼16-19 mm/yr in Coachella
417
Valley, ∼6-11 mm/yr near Biskra Palms, and ∼3-6 mm/yr for the Mill Creek strand.
418
These rates fit within the ranges of published slip rates for various segments of the southern
419
SAF (Meade and Hager, 2005; Becker et al., 2005; Fialko, 2006), but are calculated for
420
much smaller fault segments. Differences between our results and the previously published
421
geodetically-determined results can be attributed to the lack of modeled faults within the
422
ETR and the slip that is accommodated along those faults, as well as the use of the data
423
from the new JOIGN network.
6. Conclusions
424
The recent development of a new campaign GPS network within the eastern Transverse
425
Ranges over the past ∼2 years allows for better determination of slip rates along the
426
southern San Andreas fault, as well as how that slip is partitioned amongst the many
427
mapped faults throughout the ETR and eastern California shear zone to the north. The
428
results from the JOIGN network, when analyzed with CGPS data from the surrounding
429
regions, provides an excellent dataset from which we determine fault slip rates through
D R A F T
August 13, 2008, 3:02pm
D R A F T
X - 22
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
430
the use of elastic half-space block modeling. We proposed four separate models, with
431
different fault scenarios within the ETR and ECSZ. The models that are purely based on
432
the transrotation model of Dickinson (1996) grossly overestimate the amount of opening
433
along the eastern margins of the ETR, due to the large rotational nature of the ETR
434
blocks. Another set of models incorporate the development of a new through-going fault,
435
the Landers-Mojave earthquake line, in which a new fault is forming along the line of
436
seismicity produced by the Joshua Tree and Landers earthquakes of 1992. The two models
437
with the LML give reduced Bayesian information criterion (BIC) values compared to those
438
of the purely transrotational mechanism. A model that combines the transrotation with
439
the new LML produces the lowest mean residuals to the observed data, the lowest χ2 /DOF
440
value, as well as having one of the lowest BIC values. This model produced reasonable
441
tensile rates found for the eastern ETR as well as strike-slip rates within the range of
442
previously published values for slip rates along the SSAF. All models we investigated
443
showed a northwestward reduction in slip rates along the SSAF, ranging from 21 mm/yr
444
for the southern Coachella Valley segment to ∼5 mm/yr along the Mill Creek strand.
445
This reduction in slip taken up by block rotation in the ETR and right-lateral slip within
446
the faults of the ECSZ, allows us to conclude that the new Landers-Mojave earthquake
447
line faults are forming, given our best model features that fault orientation.
448
Acknowledgments. We thank Serkan Arca, Beth Bartel, Ryan Bierma, Dana Brodie,
449
Goran Buble, Eric Flood, Austin Holland, Suncana Laketa, Dave Mendel, Angel Otarola,
450
Kalyca Spinler, Sarah Thompson, and Chris Walls for providing valuable assistance with
451
the GPS campaigns. We thank Robert Powell for an inspiring field trip and for stimulating
452
discussion of southern California tectonics. We thank Sally McGill and all the students
D R A F T
August 13, 2008, 3:02pm
D R A F T
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
X - 23
453
who assisted her with collection of the San Bernardino GPS data. We also thank Jonathan
454
Matti, Chris Menges, and Luc Lavier for thought-provoking discussions. GPS equipment
455
was provided by the UNAVCO Facility. GPS field work and data analysis was funded by
456
USGS NEHRP grants 06HQGR0008, 07HQGR0055, and 08HQGR0023 and the University
457
of Arizona.
References
458
M. Anderson, J. Matti, and R. Jachens. Structural model of the San Bernardino basin,
459
California, from analysis of gravity, aeromagnetic, and seismicity data. Journal of
460
Geophysical Research-Solid Earth, 109(B4), 2004.
461
D.F. Argus, M.B. Heflin, G. Peltzer, F. Crampe, and F.H. Webb. Interseismic strain
462
accumulation and anthropogenic motion in metropolitan Los Angeles. Journal of Geo-
463
physical Research-Solid Earth, 110(B4), 2005.
464
T.W. Becker, J.L. Hardebeck, and G. Anderson. Constraints on fault slip rates of the
465
southern California plate boundary from GPS velocity and stress inversions. Geophysical
466
Journal International, 160(2):634–650, 2005.
467
468
R.A. Bennett. Instantaneous slip rates from geology and geodesy. Geophysical Journal
International, 169(1):19–28, 2007.
469
R.A. Bennett, A.M. Friedrich, and K.P. Furlong. Codependent histories of the San An-
470
dreas and San Jacinto fault zones from inversion of fault displacement rates. Geology,
471
32(11):961–964, 2004.
472
R.A. Bennett, W. Rodi, and R.E. Reilinger. Global positioning system constraints on fault
473
slip rates in southern California and northern Baja, Mexico. Journal of Geophysical
D R A F T
August 13, 2008, 3:36pm
D R A F T
X - 24
474
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
Research-Solid Earth, 101(B10):21943–21960, 1996.
475
P. Bird. Uncertainties in long-term geologic offset rates of faults: General principles
476
illustrated with data from California and other western states. Geosphere, 3(6):577–
477
595, 2007.
478
G. Blewitt, D. Argus, R. Bennett, Y. Bock, E. Calais, M. Craymer, J. Davis, T. Dixon,
479
J. Freymueller, T. Herring, D. Johnson, K. Larson, M. Miller, G. Sella, R. Snay, and
480
M. Tamisiea. A Stable North American Reference Frame (SNARF): First Release. In
481
UNAVCO/IRIS Annual Meeting, 2005.
482
W.R. Dickinson. Kinematics of transrotational tectonism in the California Transverse
483
Ranges and Its contribution to cumulative slip along the San Andreas transform fault
484
system. Geological Society of America Special Paper, (305), 1996.
485
R.K. Dokka and C.J. Travis. Role of the eastern California shear zone in accommodating
486
Pacific-North-American plate motion. Geophysical Research Letters, 17(9):1323–1326,
487
1990.
488
489
Y.J. Du and A. Aydin. Is the San Andreas big bend responsible for the Landers earthquake
and the eastern California shear zone? Geology, 24(3):219–222, 1996.
490
N.P. Fay and E.D. Humphreys. Fault slip rates, effects of elastic heterogeneity on geode-
491
tic data, and the strength of the lower crust in the Salton Trough region, southern
492
California. Journal of Geophysical Research-Solid Earth, 110(B9), 2005.
493
494
Y. Fialko. Interseismic strain accumulation and the earthquake potential on the southern
San Andreas fault system. Nature, 441(7096):968–971, 2006.
495
A.M. Freed, R. Burgmann, and T. Herring. Far-reaching transient motions after Mojave
496
earthquakes require broad mantle flow beneath a strong crust. Geophysical Research
D R A F T
August 13, 2008, 3:36pm
D R A F T
X - 25
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
497
Letters, 34(19), 2007.
498
V.A. Frizzell, J.M. Mattinson, and J.C. Matti. Distinctive Triassic megaporphyrutuc
499
monzogranite - evidence for only 160 km offset along the San-Andreas fault, southern
500
California. Journal of Geophysical Research-Solid Earth and Planets, 91(B14):14080–
501
14088, 1986.
502
J.W. Harden and J.C. Matti. Holocene and Late Pleistocene slip rates on the San-Andreas
503
fault in Yucaipa, California, using displaced alluvial-fan deposits and soil chronology.
504
Geological Society of America Bulletin, 101(9):1107–1117, 1989.
505
T.A. Herring. Documentation for the Global Kalman Filter VLBI and GPS Analysis
506
Software (GLOBK), Version 10.1. Massachusetts Institute of Technology, Cambridge,
507
Massachusetts, 2005.
508
H.O. Johnson, D.C. Agnew, and F.K. Wyatt. Present-day crustal deformation in southern
509
California. Journal of Geophysical Research-Solid Earth, 99(B12):23951–23974, 1994.
510
E.A. Keller, M.S. Bonkowski, R.J. Korsch, and R.J. Shlemon. Tectonic geomorphology
511
of the San-Andreas fault zone in the southern Indio Hills, Coachella Valley, California.
512
Geological Society of America Bulletin, 93(1), 1982.
513
514
R.W. King and Y. Bock. Documentation for the GAMIT GPS processing software release
10.2. Massachusetts Institute of Technology, Cambridge, Massachusetts, 2005.
515
V.E. Langenheim, S. Biehler, D.K. McPhee, C.A. McCabe, J.T. Watt, M.L. Anderson,
516
B.A. Chuchel, and P. Stoffer. Preliminary isostatic gravity map of Joshua Tree National
517
Park, southern California. U.S. Geological Survey Open-File Report, (2007-1218), 2007.
518
Q.S. Li and M. Liu. Geometrical impact of the San Andreas Fault on stress and seismicity
519
in California. Geophysical Research Letters, 33(8), 2006.
D R A F T
August 13, 2008, 3:36pm
D R A F T
X - 26
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
520
M. Lisowski, J.C. Savage, and W.H. Prescott. The velocity-field along the San-Andreas
521
fault in central and southern California. Journal of Geophysical Research-Solid Earth
522
and Planets, 96(B5):8369–8389, 1991.
523
A. Matmon, D.P. Schwartz, R. Finkel, S. Clemmens, and T. Hanks. Dating offset fans
524
along the Mojave section of the San Andreas fault using cosmogenic Al-26 and Be-10.
525
Geological Society of America Bulletin, 117(5-6):795–807, 2005.
526
R. McCaffrey. Block kinematics of the Pacific-North America plate boundary in the
527
southwestern United States from inversion of GPS, seismological, and geologic data.
528
Journal of Geophysical Research-Solid Earth, 110(B7), 2005.
529
530
B.J. Meade and B.H. Hager. Viscoelastic deformation for a clustered earthquake cycle.
Geophysical Research Letters, 31(10), 2004.
531
B.J. Meade and B.H. Hager. Block models of crustal motion in southern California con-
532
strained by GPS measurements. Journal of Geophysical Research-Solid Earth, 110(B3),
533
2005.
534
535
536
537
A. Nur, H. Ron, and G.C. Beroza. The nature of the Landers-Mojave Earthquake Line.
Science, 261(5118):201–203, 1993.
Y. Okada. Surface deformation due to shear and tensile faults in a half-space. Bulletin of
the Seismological Society of America, 75(4):1135–1154, 1985.
538
K.B. Olsen, S.M. Day, J.B. Minster, Y. Cui, A. Chourasia, M. Faerman, R. Moore,
539
P. Maechling, and T. Jordan. Strong shaking in Los Angeles expected from southern
540
San Andreas earthquake. Geophysical Research Letters, 33(7), 2006.
541
Working Group on California Earthquake Probabilities. Seismic hazards in southern
542
california: Probable earthquakes, 1994 to 2024. Bulletin of the Seismological Society of
D R A F T
August 13, 2008, 3:36pm
D R A F T
X - 27
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
543
544
545
America, 85(2):379–439, 1995.
M. Oskin and A. Iriondo. Large-magnitude transient strain accumulation on the Blackwater fault, Eastern California shear zone. Geology, 32(4):313–316, 2004.
546
M. Oskin, L. Perg, E. Shelef, M. Strane, E. Gurney, B. Singer, and X. Zhang. Elevated
547
shear zone loading rate during an earthquake cluster in eastern California. Geology,
548
36(6):507–510, 2008.
549
550
G. Peltzer, F. Crampe, S. Hensley, and P. Rosen. Transient strain accumulation and fault
interaction in the Eastern California shear zone. Geology, 29(11):975–978, 2001.
551
F.F. Pollitz, P. McCrory, J. Svarc, and J. Murray. Dislocation models of interseismic
552
deformation in the western United States. Journal of Geophysical Research-Solid Earth,
553
113(B4), 2008.
554
555
556
557
R.E. Powell and R.J. Weldon. Evolution of the San-Andreas fault. Annual Review of
Earth and Planetary Sciences, 20:431–468, 1992.
F. Press and C. Allen. Patterns of seismic release in the southern California region.
Journal of Geophysical Research-Solid Earth, 100(B4):6421–6430, 1995.
558
T Rockwell, C Loughman, and P Merifield. Late Quaternary rate of slip along the San-
559
Jacinto fault zone near Anza, southern California. Journal of Geophysical Research-Solid
560
Earth and Planets, 95(B6):8593–8605, 1990.
561
T.K. Rockwell, S. Lindvall, M. Herzberg, D. Murbach, T. Dawson, and G. Berger. Paleo-
562
seismology of the Johnson Valley, Kickapoo, and Homestead Valley faults: Clustering of
563
earthquakes in the eastern California shear zone. Bulletin of the Seismological Society
564
of America, 90(5):1200–1236, 2000.
D R A F T
August 13, 2008, 3:36pm
D R A F T
X - 28
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
565
J. Sauber, W. Thatcher, S.C. Solomon, and M. Lisowski. Geodetic slip rate for the eastern
566
California shear zone and the recurrence time of Mojave Desert earthquakes. Nature,
567
367(6460):264–266, 1994.
568
J.C. Savage, M. Lisowski, and W.H. Prescott. An apparent shear zone trending north-
569
northwest across the Mojave-Desert into Owens-Valley, eastern California. Geophysical
570
Research Letters, 17(12):2113–2116, 1990.
571
L. Seeber and J.G. Armbruster. The San-Andreas fault system through the Transverse
572
Ranges as Illuminated by earthquakes. Journal of Geophysical Research-Solid Earth,
573
100(B5):8285–8310, 1995.
574
J.A. Spotila, K.A. Farley, J.D. Yule, and P.W. Reiners. Near-field transpressive defor-
575
mation along the San Andreas fault zone in southern California, based on exhuma-
576
tion constrained by (U-Th)/He dating. Journal of Geophysical Research-Solid Earth,
577
106(B12):30909–30922, 2001.
578
579
L.R. Sykes and L. Seeber. Great earthquakes and great asperities, San-Andreas fault,
southern-California. Geology, 13(12):835–838, 1985.
580
S. Toda, R.S. Stein, K. Richards-Dinger, and S.B. Bozkurt. Forecasting the evolution
581
of seismicity in southern California: Animations built on earthquake stress transfer.
582
Journal of Geophysical Research-Solid Earth, 110(B5), 2005.
583
J. van der Woerd, Y. Klinger, K. Sieh, P. Tapponnier, F.J. Ryerson, and A.S. Meriaux.
584
Long-term slip rate of the southern San Andreas Fault from Be-10-Al-26 surface ex-
585
posure dating of an offset alluvial fan. Journal of Geophysical Research-Solid Earth,
586
111(B4), 2006.
D R A F T
August 13, 2008, 3:36pm
D R A F T
X - 29
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
587
R.J. Weldon and K.E. Sieh. Holocene rate of slip and tentative recurrence interval for
588
large earthquakes on the San-Andreas fault, Cajon-Pass, southern California. Geological
589
Society of America Bulletin, 96(6):793–812, 1985.
D R A F T
August 13, 2008, 3:36pm
D R A F T
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
X - 29
Table 1. Velocity estimates for campaign sites in a stable North America reference frame.
Velocity
Site
Longitude
11131
244.036
11141
244.757
1
8252
243.750
BLOY1
244.066
2,3
CHER
243.048
DESO1
244.600
2
DIVD
243.031
F7261
244.001
1
JT01
243.966
JT021
243.922
1
JT03
243.867
JT041
243.836
1
JT05
243.818
JT061
243.671
1
JT07
243.743
JT081
243.851
1
JT09
243.694
JT101
243.651
1
JT11
243.776
JT121
244.324
1
JTRE
244.236
KEY21
243.809
LUCS1,2,3
243.118
1,2,3
MEAD
243.172
MEEK2,3
243.383
2,3
ONYX
243.291
PB151
244.152
1
PB21
244.318
PITS1,2
243.074
1,3
RICU
243.531
WARR1
243.593
1
2
JOIGN, CSU-SB, 3 SCEC
D R A F T
Latitude
East
North
33.677
33.681
33.886
33.925
34.003
33.715
34.078
33.974
34.074
34.025
34.026
33.973
34.025
33.916
34.088
34.108
34.011
34.029
33.978
33.874
33.834
34.083
34.440
34.178
34.258
34.193
33.906
33.972
34.161
34.264
34.055
-8.52
-3.87
-5.43
-4.61
-19.80
-3.29
-16.42
-4.92
-4.11
-4.79
-5.72
-5.57
-6.98
-9.27
-5.22
-4.05
-7.33
-6.46
-4.52
-4.37
-3.63
-8.59
-11.23
-13.66
-7.77
-10.44
-5.89
-2.40
-15.30
-5.66
-10.93
8.23
1.44
11.14
4.59
17.26
1.74
18.42
4.42
4.14
4.22
6.65
8.09
7.51
12.11
9.21
6.45
8.88
9.32
10.94
3.10
7.71
5.22
11.69
15.19
12.28
14.54
2.60
3.45
10.86
12.97
12.30
Uncertainty
East
August 13, 2008, 2:04pm
North
Correlation
0.79
0.78
-0.013
0.75
0.76
-0.019
1.11
1.16
0.005
1.45
1.52
-0.004
5.71
3.83
0.213
0.88
0.85
-0.060
0.71
0.71
0.018
1.85
1.93
0.006
1.22
1.25
0.005
1.15
1.18
0.006
1.13
1.16
0.004
1.12
1.15
0.000
1.21
1.24
-0.000
1.13
1.17
0.004
1.43
1.45
0.001
1.43
1.46
0.001
1.14
1.16
-0.002
1.17
1.20
-0.002
3.27
3.37
-0.002
2.55
2.66
0.035
1.22
1.27
0.008
2.19
2.30
-0.033
0.33
0.37
0.014
2.07
2.18
0.067
0.37
0.37
0.033
0.97
0.88
-0.016
1.79
1.87
0.002
1.93
1.99
-0.019
1.49
1.49
-0.041
0.28
0.28
0.044
2.30
2.41
-0.015
All rates given in mm/yr
D R A F T
X - 30
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
Table 2. Velocity estimates for CGPS sites in a stable North America reference frame.
Velocity
Site
Longitude
Latitude
East
North
BBRY
243.116
1
BEMT
244.002
BMHL1
243.947
1
BMRY
243.015
CACT1
244.010
1
COTD
243.613
CTMS1
243.630
1
HNPS
244.365
IMPS1
244.855
1
KYVW
243.827
LDES1
243.567
1
OPBL
244.082
OPCL1
243.695
1
OPCP
243.917
OPCX1
243.851
2
P490
243.574
P4912
243.773
2
P504
244.234
P5112
244.704
2
P584
243.048
P5852
243.454
2
P598
243.290
P5992
243.463
2
P600
243.788
P6012
243.920
P6062
243.120
2
P607
244.179
P6092
243.107
2
P610
244.236
PIN11
243.542
1
PIN2
243.542
PSAP1
243.506
1
SDHL
243.721
SGPS1
243.304
1
TMAP
234.840
WIDC1
243.608
1
WWMT
243.346
4
5
SCIGN, PBO-GPS
34.264
34.001
34.251
33.963
33.655
33.732
34.124
33.705
34.158
33.925
34.267
34.370
34.428
34.367
34.430
33.523
33.575
33.516
33.887
33.893
34.019
34.192
34.217
33.866
33.959
34.462
33.741
34.063
34.426
33.612
33.612
33.819
34.255
33.913
33.641
33.935
33.955
-11.18
-5.59
-2.95
-14.00
-9.30
-14.54
-5.80
-4.61
-2.44
-7.73
-5.24
-2.34
-5.11
-1.87
-1.95
-15.92
-13.95
-7.20
-2.69
-15.54
-9.26
-10.44
-8.04
-8.71
-5.41
-9.85
-5.48
-11.01
-1.66
-15.82
-15.82
-12.97
-5.30
-13.49
-13.24
-10.05
-13.00
13.51
4.65
2.92
17.94
7.18
16.49
9.58
1.78
0.39
6.65
9.13
1.56
5.44
2.78
2.52
17.25
14.49
4.41
0.66
18.92
12.25
14.54
11.91
9.82
5.42
10.59
3.40
16.28
0.73
16.66
16.66
15.51
6.75
16.45
12.67
11.77
16.39
1
D R A F T
Uncertainty
East
North
Correlation
0.04
0.04
0.035
0.09
0.09
0.017
0.03
0.03
0.021
0.03
0.03
0.020
0.06
0.06
0.029
0.06
0.06
0.025
0.03
0.03
0.020
0.04
0.04
0.004
0.03
0.03
0.018
0.10
0.10
0.003
0.03
0.03
0.019
0.04
0.04
0.023
0.04
0.05
0.018
0.05
0.06
0.020
0.03
0.03
0.017
0.35
0.35
0.009
0.31
0.32
0.003
0.25
0.25
0.003
0.29
0.29
0.003
0.32
0.32
0.005
1.07
1.08
0.000
0.97
0.88
-0.016
1.54
1.55
0.000
0.50
0.50
0.006
0.78
0.78
0.001
0.96
0.97
0.004
0.42
0.42
0.002
0.63
0.63
-0.001
1.45
1.46
-0.001
0.03
0.03
0.017
0.03
0.03
0.017
0.03
0.03
0.020
0.11
0.12
0.019
0.04
0.04
0.018
0.05
0.05
0.021
0.04
0.04
0.021
0.14
0.14
0.008
All rates given in mm/yr
August 13, 2008, 2:04pm
D R A F T
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
X - 31
Table 3. WRMS and NRMS values calculated from seasonal velocities.
WRMS (mm/yr)
NRMS
D R A F T
SITE
East
North
East
North
JT01
JT02
JT03
JT04
JT05
JT07
JT08
JT09
JT10
MEAN
4.15
1.50
1.93
1.19
1.26
0.33
0.82
0.14
0.66
1.33
1.61
0.45
0.65
0.01
1.40
1.26
1.66
0.78
0.83
0.96
2.92
1.33
2.00
1.01
1.04
0.30
0.73
0.13
0.52
1.11
1.15
0.41
0.69
0.01
1.19
1.14
1.50
0.75
0.67
0.83
August 13, 2008, 2:04pm
D R A F T
X - 32
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
Table 4. WRMS and NRMS values calculated for positional estimates for each campaign site
WRMS (mm)
D R A F T
NRMS
SITE
East
North
East
North
8252
BLOY
F726
JT01
JT02
JT03
JT04
JT05
JT06
JT07
JT08
JT09
JT10
JT11
JT12
JTRE
KEY2
PB15
PB21
WARR
MEAN
2.73
2.68
1.73
2.84
2.56
3.53
3.04
3.17
2.42
2.76
3.13
2.86
3.39
2.49
2.01
3.89
1.64
2.40
2.75
1.14
2.66
2.91
2.59
1.49
1.79
1.72
2.07
1.92
1.82
1.57
2.44
2.24
1.99
1.96
1.83
2.26
2.56
2.12
2.01
2.45
2.02
2.09
0.88
0.92
0.52
1.03
0.95
1.31
1.04
1.17
0.87
0.93
1.04
1.01
1.22
0.95
0.70
1.22
0.54
0.77
0.93
0.41
0.92
0.93
0.89
0.45
0.66
0.65
0.78
0.67
0.68
0.57
0.82
0.76
0.71
0.72
0.70
0.78
0.80
0.69
0.64
0.82
0.72
0.72
August 13, 2008, 2:04pm
D R A F T
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
X - 33
Table 5. Modeled fault slip rate results for Transrotation model. All rates are averaged along
the length of the fault. Positive rates indicate right-lateral (strike-slip) and compressive (tensile)
motion.
Fault Segments
Strike-Slip Tensile-Slip
rate
rate
Blue Cut
-12.1 ± 0.2
1.7 ± 0.1
Chiriaco
1.0 ± 0.1
-1.4 ± 0.1
Emerson
12.7 ± 0.1
-0.6 ± 0.1
Ludlow
-1.8 ± 0.1
1.4 ± 0.1
Pinto Mountain #1
-7.5 ± 0.1
-0.7 ± 0.1
Pinto Mountain #2
0.5 ± 0.1
2.5 ± 0.1
Pinto Mountain #3
3.6 ± 0.1
-4.6 ± 0.2
Pisgah
5.2 ± 0.1
0.4 ± 0.1
SAF - Coachella #1 19.0 ± 0.1
-1.6 ± 0.1
SAF - Coachella #2 21.5 ± 0.1
3.9 ± 0.1
SAF - Little SB
6.2 ± 0.1
-1.1 ± 0.1
SAF - Mill Creek
3.9 ± 0.1
4.7 ± 0.1
Sheephole #1
-7.0 ± 0.2 -14.0 ± 0.2
Sheephole #2
-2.6 ± 0.1
1.3 ± 0.1
All rates given in mm/yr
D R A F T
August 13, 2008, 2:04pm
D R A F T
X - 34
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
Table 6. Modeled fault slip rate results for the through-going Landers-Mojave line model. All
rates are averaged along the length of the fault. Positive rates indicate right-lateral (strike-slip)
and compressive (tensile) motion.
Fault Segments
Strike-Slip Tensile-Slip
rate
rate
Joshua Tree
14.7 ± 0.1
Landers-Mojave Line 14.2 ± 0.1
Pinto Mountain
-1.8 ± 0.1
SAF - Coachella
21.4 ± 0.1
SAF - Little SB
7.6 ± 0.1
SAF - Mill Creek
5.3 ± 0.1
All rates given
D R A F T
-1.5 ± 0.1
-6.4 ± 0.1
-1.2 ± 0.1
3.0 ± 0.1
-1.2 ± 0.1
3.0 ± 0.1
in mm/yr
August 13, 2008, 2:04pm
D R A F T
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
Table 7.
X - 35
Modeled fault slip rate results for the modified transrotation model. All rates
are averaged along the length of the fault. Positive rates indicate right-lateral (strike-slip) and
compressive (tensile) motion.
Fault Segments
Strike-Slip Tensile-Slip
rate
rate
Blue Cut
-10.9 ± 0.2
Chiriaco
-2.3 ± 0.1
E Joshua Tree #1
-1.0 ± 0.2
E Joshua Tree #2
-0.7 ± 0.1
Emerson
13.3 ± 0.1
Pinto Mountain #1
-8.6 ± 0.1
Pinto Mountain #2
0.3 ± 0.1
Pisgah
2.6 ± 0.1
SAF - Coachella #1 16.4 ± 0.1
SAF - Coachella #2 21.0 ± 0.1
SAF - Little SB
5.8 ± 0.1
SAF - Mill Creek
3.2 ± 0.1
All rates given
D R A F T
0.1 ± 0.1
-2.1 ± 0.1
-8.6 ± 0.1
-1.1 ± 0.1
-0.9 ± 0.1
-0.5 ± 0.1
3.1 ± 0.1
0.6 ± 0.1
-0.8 ± 0.1
2.7 ± 0.1
-1.6 ± 0.1
4.8 ± 0.1
in mm/yr
August 13, 2008, 2:04pm
D R A F T
X - 36
Table 8.
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
Modeled fault slip rate results for the translation with the Landers-Mojave line
model. All rates are averaged along the length of the fault. Positive rates indicate right-lateral
(strike-slip) and compressive (tensile) motion.
Fault Segments
Strike-Slip Tensile-Slip
rate
rate
Blue Cut
-3.2 ± 0.2
Chiriaco
-0.5 ± 0.1
Joshua Tree #1
7.6 ± 0.2
Joshua Tree #2
6.7 ± 0.2
Ludlow
-2.0 ± 0.1
Landers-Mojave Line 10.1 ± 0.1
Pinto Mountain #1
-1.3 ± 0.1
Pinto Mountain #2
-2.5 ± 0.1
Pinto Mountain #3
1.5 ± 0.1
Pisgah
5.6 ± 0.1
SAF - Coachella #1 18.8 ± 0.1
SAF - Coachella #2 21.4 ± 0.1
SAF - Little SB
10.6 ± 0.1
SAF - Mill Creek
5.6 ± 0.1
Sheephole #1
-1.0 ± 0.2
Sheephole #2
-1.1 ± 0.1
All rates given
D R A F T
1.0 ± 0.1
-1.4 ± 0.1
-3.5 ± 0.1
-2.0 ± 0.1
0.0 ± 0.1
-4.7 ± 0.1
-3.6 ± 0.1
0.8 ± 0.1
1.5 ± 0.1
-0.7 ± 0.1
0.5 ± 0.1
2.8 ± 0.1
0.4 ± 0.1
4.0 ± 0.1
-2.7 ± 0.2
-0.1 ± 0.1
in mm/yr
August 13, 2008, 2:04pm
D R A F T
X - 37
SPINLER ET AL.: SOUTHERN SAN ANDREAS LOADING RATE
Table 9. BIC calculation for each of the proposed models
Model
Transrotation
Landers-Mojave Line
Modified Transrotation
Transrotation + LML
D R A F T
Number of
Observations
χ2
136
136
136
136
1089.41
1240.04
1060.57
877.37
Number of
χ2 /DOF
Free Parameters
17
9
15
19
August 13, 2008, 2:04pm
9.15
9.76
8.77
7.50
BIC Value
366.50
344.81
353.02
346.88
D R A F T
35˚N
1999 Hector Mine
1992 Big Bear
SA
F
1992 Landers
C.P.
1992 Joshua Tree
34˚N
1986 North
Palm Springs
B.P.
F
SA
33˚N
117˚W
115˚W
118˚W
116˚W
Figure 1. Holocene fault map of southern California. The grey circles display the seismicity for the last ~40 years. The major ( > 5.0 magnitude) earthquakes for that time period are shown with their corresponding focal mechanism plots. The major earthquakes, the 1986 North Palm Springs, 1992 Joshua Tree, 1992 Big Bear, 1992 Landers, and 1999 Hector Mine earthquakes are labeled. SAF = San Andreas Fault; B.P. = Biskra Palms; C.P. = Cajon Pass. Rate (mm/yr)
A
30
20
10
0
1985
1990
1995
2000
2005
Year of publicaiton
Rate (mm/yr)
B
30
20
10
0
Rate (mm/yr)
40
30
80
120
160
Along strike distance (km)
C
G:CV
G:SG
P:MC
TG:BP
TG:CP
TG:SB
20
10
0
0
20
40
60
Averaging interval (Kyr)
80
Figure 2. Fault slip rates for locations along the southern San Andreas fault. Rates are plotted by (a) year of publication, (b) along‐strike distance from south to north, and (c) time‐period for which the rate was estimated. Estimates were determined through tectonic geomorphology (TG), geodesy (G), and paleoseismology (P) for different locations on the SSAF; CV=Coachella Valley strand, SG = San Gorgonio, MC = Mill Creek, BP = Biskra Palms, CP = Cajon Pass, SB = San Bernardino strand. (Data from Bennett et al., 1996; Frizzell et al., 1986, Harden and Matti, 1989; Powell and Weldon, 1992; Rockwell et al., 1990; Van der Woerd and Klinger, 2003, personal commun.; Weldon and Sieh, 1985). 34˚30'N
0
10 mm/yr
km
10 20
34˚15'N
34˚00'N
33˚45'N
117˚00'W
116˚30'W
116˚00'W
115˚30'W
33˚30'N
115˚00'W
Figure 3. GLOBK veloctiy solution for southern California. Red squares are the locations of the JOIGN campaign sites that we have developed since September, 2005. Yellow squares are continuous sites in the area that we have incorporated into our analysis. Cyan squares are locations of a new campaign network that we have begun occupying in cooperation with Dr. Sally McGill (Cal State San Bernardino). All velocities (blue arrows) are in a Stable North America Reference Frame (SNARF 2000, Blewitt et al., 2006). Error ellipses represent the 95% confidence interval. Black lines represent select Holocene faults. Squares without plotted velocites are sites where we have yet to collect enough data to determine a velocity. JT01
80
JT02
JT03
Velocity (mm)
40
JT04
JT05
0
JT07
JT08
-40
JT09
JT10
-80
2005.5
2006.0
2006.5
2007.0
Date
2007.5
80
North
2008.0
JT01
JT02
Velocity (mm)
40
JT03
JT04
0
JT05
-40
JT08
JT07
JT09
-80
JT10
East
2005.5
2006.0
2006.5
2007.0
Date
2007.5
2008.0
Figure 4. Position time series from a selection of the JOIGN campaign sites spanning ~2 years (September 2005 to September 2007) and seven campaigns. Scatter amongst the data is 1‐2 mm in the North component and 3‐4 mm in the East component, comparable to CGPS stations (Figure 6). The normalized RMS (Table 4), are ≤ 1 for both components, indicating that the formal estimates provide a conservative measure of uncertainty. 34˚10'N
Pinto M
t
JT08
JT07
34˚00'N
JT10
JT05
JT09
JT03
JT02
JT04
February
May
September
GLOBK
F-M-S Rate Mean
5mm/yr
116˚30'W
JT01
Blue Cut
116˚15'W
116˚00'W
Figure 5. Velocity plot for different subsets of the JOIGN network. Velocities are plotted by the season in which they were collected (February, May, and September). Also plotted are the full solution GLOBK velocity, as well as the mean rate for the three different subsets. Scatter amongst the seasonal‐based estimates (Table 3) is ~1 mm/yr with the normalized RMS values near 1, indicating that the scatter of seasonal datasets is consistent with that of the overall solution, implying little to no seasonal signal within the data (Figure 4). 80
60
40
P600
North (mm)
P599
P585
20
0
-20
-40
-60
2004
P584
P491
2005
P490
2006
Time
2007
2008
Figure 6. North position time series for selected PBO CGPS stations included in the analysis. Observations range in length from ~4 years (P584) to ~1 yr (P585). The scatter amongst the data is consistent with those of the campaign sites (Figure 4). 34˚30'N
1992 Landers
34˚00'N
117˚00'W
116˚30'W
116˚00'W
115˚30'W
33˚30'N
115˚00'W
Figure 7. Patterns of faulting, seismicity, and the locations of the Landers‐Mojave line model faults. Grey circles show earthquake epicenters in southern California for the period of 1975 to 2007. Intense seismicity along the western boundary of JTNP initiated with the 1992 Joshua tree earthquake. Aftershocks of this and the subsequent Landers earthquake highlight a narrow belt of deformation, which connects the southern San Andreas fault to the ECSZ. This belt of seismicity coincides with a recently identified zone of small offset faults. White lines indicate modeled fault segments. The location of the ETR and ECSZ faults were chosen to follow this recent narrow belt of earthquake activity. EM
A
OW
DL
LU
N
A
SG
PI
SO
ER
H
PMT2
D
PMT3
T1
PM
SAF-MC
34˚30'N
SHEEPHOLE1
34˚00'N
BLUE CUT
SAF-LSB
SHEEPHOLE2
C
F-
SA
V1
CO
CHIRIA
SAF-CV2
117˚00'W
116˚30'W
116˚00'W
115˚30'W
33˚30'N
115˚00'W
B
E
C
F
Computed
Observed
20 mm/yr
Figure 8. Fault‐block model for the Transrotation model of Dickinson (1996). Red triangles represent sites with observed velocities. White lines indicate fault block boundaries used in the model. Black lines represent select Holocene faults, for comparison with selected modeled faults. (a) Index to fault names selected for each segment. PMT = Pinto Mountain fault; SAF‐MC = San Andreas fault, Mill Creek strand; SAF‐LSB = San Andreas fault, Little San Bernardino strand; SAF‐CV = San Andreas fault, Coachella Valley strands. (b) Comparison of the observed velocites (yellow vectors with 95% confidence error ellipses) and the velocities predicted by the fault‐block modeling (blue vectors). (c) Residual velocites (observed ‐ computed) and their 95% confidence error ellipses. (d) Modeled velocity field on a grid, which is the combination of the individual block motions (e) and the elastic strain associated with the locked faults (f). LML
A
34˚30'N
D
T
SAF-M
PM
C
34˚00'N
SAF-LSB
JTREE
SA
F-
C
117˚00'W
116˚30'W
V
116˚00'W
115˚30'W
33˚30'N
115˚00'W
B
E
C
F
Computed
Observed
20 mm/yr
Figure 9. Same as Figure 8, except using the Landers‐Mojave earthquake line model. JTREE = Joshua Tree fault; LML = Landers‐Mojave earthquake line; PMT = Pinto Mountain fault; SAF‐LSB = Little San Bernardino segment of the SAF; SAF‐MC = Mill Creek strand of the SAF; SAF‐CV = Coachella Valley strand of the SAF. N
AH
SG
PI
SO
ER
EM
A
34˚30'N
D
PMT2
SAF-M
C
T1
PM
T
EJ
34˚00'N
1
BLUECUT
SAF-LSB
EJT
2
C
FSA
V1
CHIRIA
CO
SAF-CV2
117˚00'W
116˚30'W
116˚00'W
115˚30'W
33˚30'N
115˚00'W
B
E
C
F
Computed
Observed
20 mm/yr
Figure 10. Same as Figure 8, except using the modified translation model. EJT = Eastern Joshua Tree faults; PMT = Pinto Mountain faults; SAF‐LSB = Little San Bernardino segment of the SAF; SAF‐MC = Mill Creek strand of the SAF; SAF‐CV = Coachella Valley strands of the SAF. D
DL
AH
SG
OW
PMT2
T1
PM
SAF-M
C
34˚30'N
LU
PI
LML
A
PMT3
SHEEPHOLE1
JT2
34˚00'N
BLUECUT
SAF-LSB
JT1
SHEEPHOLE2
FSA
V1
C
CO
CHIRIA
SAF-CV2
117˚00'W
116˚30'W
116˚00'W
115˚30'W
33˚30'N
115˚00'W
B
E
C
F
Computed
Observed
20 mm/yr
Figure 11. Same as Figure 8, except using the combination of the transrotation and Landers‐Mojave line model. JT = Joshua Tree faults; LML = Landers‐Mojave line; PMT = Pinto Mountain fault; SAF‐LSB = Little San Bernardino segment of the SAF; SAF‐MC = Mill Creek strand of the SAF; SAF‐CV = Coachella Valley strand of the SAF. 34˚00'N
33˚45'N
115˚45'W
115˚15'W 115˚30'W
Figure 12. Contours of gravity anomalies for the ETR (Langenheim et. al., 2007). The locations of the faults for the transrotation model were chosen by selecting locations that best matched the gravity data as well as the mapped Holocene faults (Figure 3). The strands of the SSAF, as well as the Pinto Mountain strands, Blue Cut, and Chiriaco modeled faults follow the steep gravity contours (see Figure 8a for a reference to fault labels). 116˚30'W
116˚15'W
116˚00'W
A
34˚30'N
B
34˚00'N
117˚00'W
116˚30'W
116˚00'W
115˚30'W
33˚30'N
115˚00'W
D
C
mm/yr
-24
-12
0
12
24
Fault slip rate (mm/yr)
Figure 13. Strike‐slip rates for the fault segments of the four different models. Reds indicate a right‐lateral sense of motion, with blues indicating left‐lateral motion. Rates are given in Tables 5, 6, 7, and 8. Models are (a) transrotation, (b) Landers‐Mojave line, (c) modified transrotation, and (d) transrotation with the Landers‐Mojave line. Individual block motions are plotted for each case. A
34˚30'N
B
34˚00'N
117˚00'W
116˚30'W
116˚00'W
115˚30'W
33˚30'N
115˚00'W
D
C
mm/yr
-20
-10
0
10
20
Fault slip rate (mm/yr)
Figure 14. Tensile‐slip rates for the fault segments of the four different fault‐block models. Reds indicate closing and blues indicate opening. Each fault section is broken into 4 segments of equal length in order to better determine the rates, especially in areas experiencing block rotations. Average rates for each of the labeled faults are given in Tables 5, 6, 7, and 8. Models are (a) transrotation, (b) Landers‐Mojave line, (c) modified transrotation, and (d) transrotation with the Landers‐Mojave line. Individual block motions are plotted for each case.