PUBLICATIONS
Journal of Geophysical Research: Solid Earth
RESEARCH ARTICLE
10.1002/2014JB011244
Key Points:
• Map of surface rupture scarp and
measurements of surface deformation
• Rupture segment linkage and growth
• Rupture kinematics are inferred
from the observed rupture
branching pattern
Correspondence to:
M. Suter,
SuterMax@alumnibasel.ch
Citation:
Suter, M. (2015), Rupture of the
Pitáycachi Fault in the 1887 Mw 7.5
Sonora, Mexico earthquake (southern
Basin-and-Range Province): Rupture
kinematics and epicenter inferred from
rupture branching patterns, J. Geophys.
Res. Solid Earth, 120, 617–641,
doi:10.1002/2014JB011244.
Received 3 MAY 2014
Accepted 16 DEC 2014
Accepted article online 19 DEC 2014
Published online 26 JAN 2015
Rupture of the Pitáycachi Fault in the 1887 Mw 7.5 Sonora,
Mexico earthquake (southern Basin-and-Range
Province): Rupture kinematics and epicenter
inferred from rupture branching patterns
Max Suter1
1
Instituto de Geología, Universidad Nacional Autónoma de México Estación Regional del Noroeste, Hermosillo, Sonora, Mexico
Abstract During the 3 May 1887 Mw 7.5 Sonora earthquake (surface rupture end-to-end length: 101.8 km),
an array of three north-south striking Basin-and-Range Province faults (from north to south Pitáycachi, Teras,
and Otates) slipped sequentially along the western margin of the Sierra Madre Occidental Plateau. This
detailed field survey of the 1887 earthquake rupture zone along the Pitáycachi fault includes mapping the
rupture scarp and measurements of surface deformation. The surface rupture has an endpoint-to-endpoint
length of ≥41.0 km, dips ~70°W, and is characterized by normal left-lateral extension. The maximum surface
offset is 487 cm and the mean offset 260 cm. The rupture trace shows a complex pattern of second-order
segmentation. However, this segmentation is not expressed in the 1887 along-rupture surface offset profile,
which indicates that the secondary segments are linked at depth into a single coherent fault surface. The
Pitáycachi surface rupture shows a well-developed bipolar branching pattern suggesting that the rupture
originated in its central part, where the polarity of the rupture bifurcations changes. Most likely the rupture
first propagated bilaterally along the Pitáycachi fault. The southern rupture front likely jumped across a step
over to the Teras fault and from there across a major relay zone to the Otates fault. Branching probably
resulted from the lateral propagation of the rupture after breaching the seismogenic part of the crust, given
that the much shorter ruptures of the Otates and Teras segments did not develop branches.
1. Introduction
North-south striking, west dipping Basin-and-Range Province normal faults and associated half-grabens
are located along the >300 km long western edge of the Sierra Madre Occidental Plateau in northeastern
Sonora, Mexico. On 3 May 1887, the largest historical earthquake of the southern Basin-and-Range
tectonic-physiographic province occurred within this discontinuous fault array and produced the world’s
longest recorded normal fault surface rupture in historic time [dePolo et al., 1991; Yeats et al., 1997]. Field
observations indicate that three first-order range-bounding normal faults (from north to south: Pitáycachi,
Teras, and Otates; Figure 1) ruptured in this 1887 Sonoran earthquake [Suter and Contreras, 2002]. These
are the only Basin-and-Range Province faults in Mexico with known historical coseismic surface rupture.
The rupture has a ~70°W dipping plane of movement, a maximum displacement (net slip) of 5.2 m, and is
composed (from north to south) of the Pitáycachi, Teras, and Otates rupture segments (Figures 1 and 2 and
Table 1). Including two isolated minor segments to the north of the Pitáycachi segment, the cumulative
1887 rupture trace is 86.7 km long, and the distance between the rupture trace tips is 101.8 km. Empirical
scaling relations between surface rupture length and moment magnitude for normal faults [Wells and
Coppersmith, 1994] estimate a magnitude Mw of 7.5 ± 0.3 for this earthquake.
In this paper, I present a detailed structural analysis of the 1887 earthquake surface rupture along the
Pitáycachi fault. Earlier papers dealt with the structure of the 1887 surface rupture along the Teras [Suter,
2008a] and Otates [Suter, 2008b] faults. This study is based on geological mapping (Figure 3) and structural
and morphological field measurements. I also summarize here the trace, rupture parameter values, and
surface offset profile for the entire 1887 rupture along all three faults. Knowledge of the rupture geometry,
slip distribution, and fault zone properties is necessary to understand the process of earthquake rupture
[e.g., Shaw, 2006; Scholz, 2007] and required to assess the regional seismic ground-shaking hazard
[e.g., International Atomic Energy Agency, 2010; U. S. Nuclear Regulatory Commission, 2012]. Furthermore,
the amount of slip expected at a given site on the Pitáycachi fault is required to estimate the local
SUTER
©2014. American Geophysical Union. All Rights Reserved.
617
Journal of Geophysical Research: Solid Earth
10.1002/2014JB011244
Figure 1. Shaded relief map of northeastern Sonora and northwestern Chihuahua (Mexico) and the adjacent U.S. border region showing in red the 1887 rupture
trace (barbs on hanging wall; P, Pitáycachi segment; T, Teras segment; O, Otates segment) and in black the traces of interpreted Basin-and-Range Province faults.
SB, San Bernardino Basin; AB, Arizpe Basin; RB, El Rodeo Basin. GC, Guadalupe Canyon fault; LE, Los Embudos fault; LC, La Cabellera fault. Black rectangles: region
covered by Figure 3. The lower hemisphere equal-area stereoplots represent mean slickenline orientations measured on the rupture surface and earthquake focal
mechanisms (P quadrants shaded in grey; parameters in Table 4). Tectonic-physiographic provinces marked on the location map are SM, Sierra Madre Occidental; SB,
southern Basin-and-Range; GP, Great Plains; RG, Rio Grande rift; CP, Colorado Plateau; CB, central Basin-and-Range; and BC, Baja California.
surface fault displacement hazard [e.g., La Pointe et al., 2002; Youngs et al., 2003; Wesnousky, 2008;
Boncio et al., 2012].
The 1887 surface rupture along the Pitáycachi fault is an excellent natural laboratory for making structural
and morphological field observations because the rupture scarp is still well exposed in this arid environment
(Figure 4). Here I use detailed new data to explore several topics related to coseismic rupture dynamics: First,
the Teras and Otates rupture segments measure between 18 and 20 km, a typical length for normal fault
segments [Jackson, 2002], and are structurally simple (Figure 1 and Table 1). The Pitáycachi surface rupture
segment, on the other hand, is ≥41 km long and structurally complex (Figure 3). What explains the structural
complexity of this rupture segment? Second, what is the slip history of the Pitáycachi fault? Did previous
ruptures have similar along-strike variations of coseismic slip? Are the 1887 rupture dimensions characteristic
SUTER
©2014. American Geophysical Union. All Rights Reserved.
618
Journal of Geophysical Research: Solid Earth
10.1002/2014JB011244
Figure 2. Historic 1887 photograph by Camillus S. Fly of the 1887 earthquake rupture in a stand of ocotillo (Fouquieria
splendens) between Arroyo Pitáycachi and Cañón de los Embudos, most likely at 109.135° longitude west/31.088° latitude
north (site 124 in Figure 3). The view is from the northwest. The scarp is subvertical, composed of alluvial gravel cemented by
caliche, and its height is estimated here as 4.7 m. The two persons are standing in a fissure that developed along the scarp. The
ground surface is not rotated toward the scarp, which suggests a simple planar near-surface geometry of the rupture.
for ruptures along this fault? Third, the 1887 rupture along the Pitáycachi fault exhibits a bipolar branching
pattern. Can we infer from this observation the rupture kinematics of this preinstrumental earthquake, such
as the location of the epicenter and the direction of rupture propagation or even make inferences about slip
partitioning and the physics of the rupturing process?
2. Tectonic Setting
The Pitáycachi fault is part of the fault network consisting of north-south striking, mostly west dipping
subparallel Basin-and-Range Province normal faults that delimit the western edge of the Sierra Madre
Occidental Plateau in northeastern Sonora, Mexico (Figure 1) and also deform the western parts of the Boot
Heel (Peloncillo and Animas Mountains) and Mogollon-Datil (Mule and Mogollon Mountains) volcanic fields
farther north, in southeastern Arizona and southwestern New Mexico [Machette et al., 1986; Mack, 2004].
Compared to the Rio Grande rift, adjacent in the northeast (Figure 1), this region has lower heat flow and a
thicker, less extended crust [Keller et al., 1990; Baldridge et al., 1995; Keller, 2004].
The region near the 1887 surface rupture has been extended along subvertical faults by ~10% [Suter, 2008b].
This extension initiated circa 25 Ma (late Oligocene), synchronous with the formation of the Rio Grande rift
[Chapin and Cather, 1994], as indicated by the age of basalt flows intercalated with the lowermost fill of
nearby extensional basins [McDowell et al., 1997; Paz Moreno et al., 2003; González-León et al., 2010], which are
mostly half-grabens. Extension was associated with 1.0 to 1.5° of clockwise vertical-axis rotation of the
Colorado Plateau (Figure 1), relative to the stable craton, from late Oligocene to at least the late Miocene
[Chapin and Cather, 1994]. Late Oligocene to early Miocene deposits form the bulk of the basin fills. For
example, subsidence of the Arizpe Basin (Figure 1), located ~100 km southwest of the Pitáycachi fault,
occurred mostly between 25 Ma and 21 Ma, at a rate of ~0.5 mm/a. Similarly, the subsidence of the El Rodeo
Basin (Figure 1), located ~100 km southwest of the Otates fault, occurred between 25 Ma and 18 Ma
[González-León et al., 2010]. There is no clear understanding of the Plio-Quaternary tectonic development
within the Basin-and-Range Province of northeastern Sonora.
The focal depth of well-recorded microearthquakes in the region of the 1887 earthquake rupture, located
from local and regional body wave arrival times of several seismic networks [Natali and Sbar, 1982; Castro
et al., 2010], indicate that the seismogenic thickness of the crust is 10 to 15 km. Coulomb stress modeling
suggests that the microearthquakes are concentrated where the 1887 rupture caused an increase in static shear
SUTER
©2014. American Geophysical Union. All Rights Reserved.
619
SUTER
—
—
0.06
0.08
0.06
—
a
Average repeat time of 1887 sized ruptures.
b
Arithmetic mean.
c
Based on integration of regression function.
d
During the Quaternary based on Bull and Pearthree
e
Víbora-North
Víbora-South
Pitáycachi
Teras
Otates
Entire Rupture
[1988].
Since no slickensides were observed on the free face of the Teras rupture segment, the average dip measured on the Teras fault plane is given as a proxy.
—
—
d
100–200
15–26
30–42
—
—
—
2.77
1.27
1.89
2.11
—
—
5.16
2.08
2.70
5.16
b
0.46
0.49
4.87
1.84
2.50
4.87
—
—
70
e
62
68
69
0.8
2.1
41.0
20.0
18.2
101.8
Segment
109.149
109.173
109.159
109.212
109.184
109.149
31.270
31.242
31.194
30.810
30.516
31.270
109.151
109.168
109.180
109.259
109.168
109.168
31.263
31.223
30.825
30.635
30.352
30.352
15.6
13.3
2.0
12.0
5.8
0.1
0.39
b
0.43
c
2.60
c
1.12
c
1.75
1.97
Average
Slip (m)
Maximum
Surface
Offset (m)
Average
Dip (°W)
Strike
(deg)
Latitude
(°N)
Longitude
(°W)
Latitude
(°N)
Longitude
(°W)
Length
(km)
Southern
Endpoint
Northern
Endpoint
Table 1. Structural Parameters of the 1887 Sonora Earthquake Surface Rupture
Average
Surface
Offset (m)
Maximum
Slip (m)
Recurrence
a
Interval (kyr)
Long-Term
Slip Rate
(mm/yr)
Journal of Geophysical Research: Solid Earth
10.1002/2014JB011244
stress in the surrounding crust [Suter and Contreras,
2002]. These microearthquakes have been interpreted
as aftershocks of the 3 May 1887 Mw 7.5 main shock
[Castro et al., 2010]. The long aftershock duration can be
explained by the unusually large magnitude of the main
shock and by the low slip rates and long main shock
recurrence times of the faults that ruptured in 1887
(Table 1). Geodynamic models by Li et al. [2007] show
that the strain energy released by a large intraplate
earthquake will migrate to the surrounding region
and may dominate the local strain energy budget for
thousands of years following the main shock, in contrast
to interplate seismic zones, where strain energy is
dominated by tectonic loading [Stein and Liu, 2009].
Significant east-west extension with a strain rate of
1.06 ± 0.2 × 10 9 yr 1 has been documented from GPS
measurements over a recent 4 year period in the
southernmost part of the Rio Grande rift along
latitude 32.5°N [Berglund et al., 2012]. Assuming a
linear strain distribution, this corresponds to an
extensional deformation rate of ~0.12 mm/yr per
100 km of east-west distance. Slip vectors from
earthquake focal mechanisms [Suter and Contreras,
2002] and slickenlines on the exposed 1887 rupture
surface along the Otates fault [Suter, 2008b] also
indicate east-west orientation of maximum horizontal
extension, which, on a regional scale, can also be
inferred from borehole breakouts recorded farther
east, in Chihuahua [Suter, 1987, 1991].
3. Stratigraphy
The Pitáycachi fault separates a major horst in its
footwall to the east from the San Bernardino Basin
in its hanging wall to the west (Figures 1, 3, and 4).
Lithostratigraphic units exposed in the footwall of
the Pitáycachi fault include Permian carbonates,
Mesozoic sediments (Glance Conglomerate, La Morita
Formation, and Mural limestone), a Laramide (Late
Cretaceous to Eocene) granite stock and related
dikes, Oligocene volcanic rocks of the Sierra Madre
Occidental, and Pleistocene to Recent piedmont
alluvial fan deposits and terrace fill deposits (Figure 3).
Lithostratigraphic units exposed in the hanging wall of
the fault (San Bernardino Basin) include a conglomerate
probably belonging to the late Oligocene to early
Miocene Báucarit Formation, Pliocene to Recent
axial-fluvial deposits, relict piedmont alluvial fan
deposits, terrace fill deposits, and Pleistocene basalts
of the Gerónimo volcanic field [Kempton and Dungan,
1989]. The lithology and age of the subsurface basin
fill is as yet unknown. Its thickness proximal to the
Pitáycachi fault is estimated as ~3000 m by Sumner [1977]
based on gravimetric measurements and modeling.
©2014. American Geophysical Union. All Rights Reserved.
620
Journal of Geophysical Research: Solid Earth
10.1002/2014JB011244
Figure 3. Geologic map showing in red the rupture trace of the 1887 earthquake along the Pitáycachi fault. Red star is the estimated location of the 1887 epicenter.
Open areas are unmapped. Projection: Universal Transverse Mercator (zone: 12 N, datum: NAD1927-Mexico). The dashed line marks the road connecting Morelos
with Agua Prieta and Bavispe, and the dotted lines are drainages.
The Pliocene to early Pleistocene (2 Ma–750 ka) deposits include siltstone, sandstone, and conglomerate and
are typical of braided streams [Biggs et al., 1999]. These deposits crop out in either extremities of the basin,
west of Mesa La Víbora in the north and in the Morelos region in the south, but are covered by piedmont
alluvial fan deposits near the center of the basin (Figure 3). The extent of the axial-fluvial deposits northeast of
Morelos (Figure 3) suggests that the upper (eastern) part of the Bavispe River had drained into the San
Bernardino Basin prior to Pleistocene headward erosion and stream capture by the lower (western) part of
the Bavispe River (Figures 1 and 3).
The maximum thickness of the piedmont alluvial fan deposits derived from the footwall of the Pitáycachi
fault is 200–300 m. The typical inclination of the well-preserved relict fan surfaces is 0.7–1.0° (Figure 3). The
large relict alluvial fan surfaces of Mesa La Víbora and north of Arroyo Los Embudos (Figure 3) are estimated
by Bull and Pearthree [1988] and Pearthree et al. [1990] to be of early to middle Pleistocene (1.5 Ma–125 ka)
age based on soil profile characteristics.
The constructional surfaces of these piedmont alluvial fans as well as the ones of late Pleistocene (125–10 ka)
infill terraces along some tributaries of the San Bernardino River, ~20–50 m above the modern flood plain, are
SUTER
©2014. American Geophysical Union. All Rights Reserved.
621
Journal of Geophysical Research: Solid Earth
10.1002/2014JB011244
Figure 4. Surface rupture (low white ribbon) along the central part of the Pitáycachi rupture segment, seen from the WNW.
The scarp height is approximately 4 m. Background: ignimbrites of the Sierra Mare Occidental; middle ground (foothills of
the Sierra Madre): Cretaceous Mural limestone; foreground: alluvial fan deposits of the San Bernardino valley. Photograph
taken in 1999 by the author.
displaced by the 1887 earthquake surface rupture (Figure 3). These constructional surfaces are excellent
geomorphic markers; their ages provide constraints on the fault activity. The terraces along Arroyo Los
Embudos are estimated to be of middle to late Pleistocene (750–10 ka) age by Pearthree et al. [1990].
Chronologically better constrained surfaces of broadly equivalent infill terraces along the Casas Grandes River
valley in northwestern Chihuahua formed at 14 ka [Nordt, 2003] and along the Rio Grande valley in the
Albuquerque region at ~47–40 ka (late Pleistocene) [Cole et al., 2007].
4. Structural Configuration of the Pitáycachi Fault
The west dipping Pitáycachi normal fault bounds the San Bernardino Basin in the east (Figures 1 and 3). The
mountain ranges in the footwall (Figure 4) have elevations >2000 m above sea level (asl); the highest
location, prominent Cerro Pitáycachi, reaches ~2280 m asl. The footwall is more back eroded and embayed
than the footwalls of the Teras and Otates faults, which form pronounced linear fault escarpments. For that
reason, the trace of the Pitáycachi fault is morphologically less conspicuous, and there are only a few outcrops
of the fault plane. The fault trace is only well defined where the 1887 rupture displaces Pleistocene alluvial fan
and terrace fill deposits against bedrock (Figure 3), which places a lower limit of 31 km on the length of the
Pitáycachi fault. Farther north, the fault plane is concealed beneath these Pleistocene surficial deposits and its
structure and exact location remain uncertain. However, the fault is likely to extend at least as far as the 1887
rupture, which is at least 41.0 km but possibly up to 50.2 km long (see below). The maximum throw of the
Pitáycachi fault (~4000 m) can be estimated by adding the relief of Cerro Pitáycachi above the alluvial fan
surface (1080 m) to the thickness of the San Bernardino Basin fill proximal to the fault, estimated as ~3000 m by
Sumner [1977] based on gravimetric measurements and modeling. The fault strikes approximately north-south,
and its average dip at the surface is 72°W (mean from measurements at eight outcrops; Figure 3). An
identical dip can be inferred from the composite focal mechanisms of microearthquakes located west of
the Pitáycachi fault trace (Figure 1) [Natali and Sbar, 1982], which suggests that the fault has a constant
steep dip down to midcrustal levels.
The footwall of the Pitáycachi fault is characterized by two previously unreported WNW-ESE striking major
normal faults, which both influenced the segmentation of the 1887 earthquake rupture, as documented
below. Normal faults of the same orientation also exist in the footwall of the Otates fault, and one of them
coincides with the southern end of the 1887 rupture (Figure 1) [Suter, 2008b].
The Los Embudos fault is the northern of these two major cross faults and crops out on the south side of the
valley drained by Arroyo Los Embudos (Figure 3). The fault is >12 km long (Figure 1), strikes between east-west
and N60°W, exhibits >900 m of down-to-the-north throw, and juxtaposes felsic volcanic rocks against basalt.
The scarp of the Los Embudos fault is much less embayed than that of the Pitáycachi fault. However, the fault is
SUTER
©2014. American Geophysical Union. All Rights Reserved.
622
Journal of Geophysical Research: Solid Earth
10.1002/2014JB011244
onlapped in the west by alluvial fan deposits (Figure 3), which precludes fault activity after the middle
Pleistocene. A transverse intrabasin high or basement ridge, inferred in this part of the San Bernardino Basin by
Sumner [1977] from gravity measurements, is likely to correspond to the footwall of the Los Embudos fault
beneath the basin fill.
The La Cabellera fault is the southern of the two major cross faults and bounds the Sierra La Cabellera
mountain range (Figures 1 and 3). Most of the fault trace is buried beneath alluvial fan deposits, implying
that the fault has been inactive at least since the middle Pleistocene. Farther east, beyond Figure 3, in the
valley of Arroyo El Púlpito (Figure S2 in the supporting information in Suter [2008b]), the fault dips 58°S
and juxtaposes volcanic bedrock against conglomerate of the late Oligocene to early Miocene Báucarit
Formation. The La Cabellera fault strikes N60°W and has a throw >1250 m, inferred from the relief between
the top of the felsic volcanic rocks at the Bavispe River (Figure 3) and the erosional surface of the same
stratigraphic unit in the Sierra La Cabellera.
Both the Los Embudos and the La Cabellera faults have the same order-of-magnitude length and throw as
the north-south striking Basin-and-Range Province normal faults of the study area (Figure 1) and were active
after deposition of the Oligocene mafic lava flows. Alternating activity between the two normal fault systems
would require permutations of the horizontal principal stresses, which suggests that these are of similar
magnitudes. Alternatively, the two systems may have been active concurrently. The Guadalupe Canyon fault,
an east-west striking normal fault like the Los Embudos fault, is located 16 km north of the Los Embudos fault,
in southernmost Arizona (Figure 1) and forms 10 to 20 m high scarps on the remnants of Pliocene to early
Pleistocene alluvial fans [Machette et al., 1986]. Slip along the nearly orthogonal Pitáycachi and Guadalupe
Canyon normal faults during Pleistocene time suggests that intermittent permutations of the least and
intermediate tectonic principal stresses have extended into the Pleistocene.
The San Bernardino Basin is a half-graben bound on its eastern margin by west dipping normal faults.
Reconnaissance mapping along the road that leads from Colonia Morelos (Figure 3) in northwestern direction
to kilometerpost 47 of the highway connecting Agua Prieta with Fronteras (Figure 1) indicates that there is no
major conjugate normal fault on the west side of the basin. The stratigraphic sequence at the western
basin margin consists of basal felsic volcanic rocks overlain by a poorly stratified tilted conglomerate without
basalt components, probably belonging to the late Oligocene to early Miocene Báucarit Formation. This
conglomerate is overlain (in this sequence) by basalt, lake deposits, and relict alluvial fan deposits. Because it
pinches out toward the west, the conglomerate probably represents syntectonic basin fill related to early
activity of the Pitáycachi fault.
The width of the San Bernardino half-graben, between this conglomerate and the trace of the Pitáycachi
fault, is 16 km. The maximum width of half-graben structures is thought to broadly correlate with the
seismogenic thickness of the crust [Scholz and Contreras, 1998; Jackson, 2002], which can be inferred from
the focal depth of aftershocks. The thickness of the seismogenic layer estimated from the 16 km width of
the San Bernardino half-graben is in broad agreement with the 10 to 15 km focal depth of the recorded
microseismicity [Natali and Sbar, 1982; Castro et al., 2010].
The geologic slip rate of the Pitáycachi fault can be estimated from the Pliocene to early Pleistocene (2 Ma–750 ka)
braided-stream deposits southeast of Arroyo La Cabellera (Figure 3), which are vertically displaced by up to 60 m,
indicating a slip rate in the range between 0.03 and 0.08 mm/yr. These values are comparable to the geologic
slip rates of the Teras and Otates faults (0.06 and 0.08 mm/yr, respectively, Table 1) inferred from their structural
geometry (average dip and throw of the limit between the Oligocene felsic and mafic volcanic rock units,
which are mappable across the fault) and the assumption that the regional extension initiated ~25 Ma ago
[Suter and Contreras, 2002]. The more recent geologic slip rate of the Pitáycachi fault since Quaternary time, on
the other hand, has been only 0.02 mm/yr based on the estimated age of soils formed on alluvial surfaces
displaced by the fault [Bull and Pearthree, 1988; Pearthree et al., 1990]. This rate is at the very lowest end of the
normal fault slip rates documented across the Great Basin and the southern Basin-and-Range Province, which
range over four orders of magnitude [McCalpin, 1995; dePolo and Anderson, 2000].
5. The 1887 Earthquake Rupture Along the Pitáycachi Fault
Here I provide a detailed account of the surface rupture of the 1887 earthquake along the Pitáycachi fault
based on field observations. The 125 year old down-to-the-west rupture scarp is still well exposed (Figure 4).
I mapped the scarp and surrounding geology (Figure 3) on 1:50,000 scale stereographic black-and-white aerial
SUTER
©2014. American Geophysical Union. All Rights Reserved.
623
Journal of Geophysical Research: Solid Earth
10.1002/2014JB011244
photographs supplied by Instituto
Nacional de Estadística, Geografía e
Informática. The 1887 rupture coincides
with the base of a compound rupture
scarp. Exceptions are some secondary
rupture strands and where the rupture
crosses the late Pleistocene infill
terraces of some tributaries of the
San Bernardino River.
Figure 5. Parameters used to characterize the 1887 surface rupture scarp
(modified from Bucknam and Anderson [1979] and McCalpin [2009]). The
scarp width, scarp height, and the height of the free face were determined
with a level and tape measure at 96 sites along the Pitáycachi fault (Figure 3
and Table 2). The surface offset was calculated from these measurements,
whereas the amount of dip slip was obtained from the surface offset and the
dip of the rupture plane.
The length of the surface rupture along
the Pitáycachi fault is 41.0 km (from
endpoint to endpoint of mapped
scarp; Figure 3 and Table 1). This is a
minimum value, since surface rupture
was observed by both Goodfellow
[1888] and Aguilera [1888] farther south,
on the northern bank of the Bavispe
River (Figure 3), which adds another
1.2 km to the length of this rupture
segment. Moreover, according to
Goodfellow [1887, 1888], the rupture
continues south for another 8 km
beyond the Bavispe River, which I have
not been able to confirm. This would
bring the length of the Pitáycachi
rupture segment up to 50.2 km.
5.1. Rupture Scarp Morphology and Scarp-Bounding Fissure
The 1887 rupture scarp along the Pitáycachi fault was studied at 96 sites (including seven sites on the Víbora
segments), marked with reference numbers in Figure 3, where I measured the scarp width, scarp height,
and the preserved height of the free face (Figure 5 and Table 2). I also noted at each site the material exposed
in the footwall and, in the case of alluvial fan deposits, the clast size, and whether the material was cemented
or not. The alluvial fan gravel displaced by the surface rupture is coarse and poorly stratified (Figure 2).
Where the footwall exposes bedrock, the scarp head (Figure 5) is commonly covered by bedrock colluvium,
which facilitates measurement of the scarp height and width. The surface offset (vertical separation of the
inclined ground surface) was calculated from the scarp height and the slope angle of the ground surface
above and below the scarp (Figure 5). Scarp heights measure up to 595 cm, and surface offsets reach up to
487 cm (Table 2). Estimates for surface offset remain constant at any location along the fault trace, whereas
the scarp width and height increase as the scarp crest erodes back and the scarp broadens with time.
However, this increase does not cause major differences between scarp height and surface offset (Table 2),
since the original surface slopes are low. This contrasts with the situation along the Otates rupture segment,
where the surface offset deviates significantly from the scarp height because of the steep, up to 31° inclined
colluvial slopes adjacent to the 1887 rupture [Suter, 2008b].
Aguilera [1888] reports 1887 rupture scarp inclinations of 75 to 90° along the Pitáycachi fault, whereas the
scarp inclinations in uncemented material (Table 2) are now ≤32° (angle of repose). Fault scarp segments
steeper than the angle of repose (free faces, Figure 5) still exist at 42 of the 96 observation sites (Table 2); they
are best preserved where the rock debris is cemented.
According to both Aguilera [1888] and Goodfellow [1888], a continuous fissure (Figure 2) formed along
the base of the entire rupture scarp along the Pitáycachi fault. Based on my field observations the fissure
is only preserved at sites 72 and 73 in Figure 3, in the southern part of the Pitáycachi rupture segment.
In contrast, this basal fissure is still visible at many sites along the Teras and Otates segments [Suter,
2008a, 2008b]. This suggests that the downslope debris flow across the rupture scarp of the Pitáycachi
segment is diffusive and has resulted in fissure filling. In contrast, the debris flow across the 1887 free
SUTER
©2014. American Geophysical Union. All Rights Reserved.
624
Journal of Geophysical Research: Solid Earth
10.1002/2014JB011244
a
Table 2. Scarp Measurements Along the Surface Rupture of the Pitáycachi Fault
Site
Longitude
(°W)
Latitude
(°N)
179
178
104
103A
103
101
102
89
88
87
86
85
91
92
93
94
95
96
97
98
99
109
110
113
112
111
114
115
100
116
117
108
118
107
106
105
119
120
121
122
123
124
125
127
128
129
130
131
132
133
134
135
136
137
139
142
143
146
145
109.149
109.150
109.172
109.171
109.170
109.169
109.168
109.157
109.155
109.150
109.145
109.143
109.140
109.140
109.138
109.136
109.135
109.136
109.134
109.133
109.128
109.121
109.122
109.138
109.138
109.136
109.138
109.136
109.125
109.130
109.129
109.128
109.133
109.130
109.130
109.130
109.129
109.131
109.130
109.134
109.135
109.135
109.135
109.137
109.141
109.144
109.152
109.154
109.155
109.158
109.160
109.162
109.161
109.161
109.162
109.166
109.164
109.163
109.164
31.269
31.266
31.237
31.234
31.230
31.226
31.223
31.193
31.192
31.190
31.185
31.181
31.175
31.173
31.171
31.167
31.165
31.163
31.156
31.153
31.146
31.148
31.146
31.150
31.144
31.142
31.139
31.135
31.136
31.131
31.131
31.132
31.129
31.125
31.118
31.112
31.101
31.099
31.098
31.095
31.092
31.088
31.081
31.076
31.066
31.056
31.048
31.045
31.038
31.035
31.030
31.022
31.015
31.012
31.005
30.990
30.986
30.983
30.979
SUTER
Cumulative
Distance
(km)
b
b
c
c
c
c
c
0
0.3
1.0
1.6
2.0
2.8
3.0
3.4
3.6
4.0
4.2
5.0
5.4
6.2
d
d
e
e
e
e
e
7.4
e
e
8.0
d
8.6
9.4
10.2
11.2
d
11.4
11.9
12.2
12.6
13.6
14.1
15.0
16.3
17.2
17.7
18.2
18.6
19.2
20.1
20.9
21.2
22.0
23.7
24.2
e
25.0
Elevation
(m)
Scarp
Height
(cm)
Scarp
Width
(cm)
Scarp
Slope
(deg)
Free
Face
(cm)
Surface
Slope
(deg)
Surface
Offset
(cm)
1250
1245
1200
1200
1195
1100
1185
1180
1185
1190
1195
1195
1165
1155
1180
1200
1205
1200
1200
1200
1205
1250
1240
1185
1185
1190
1190
1195
1235
1215
1220
1225
1205
1215
1200
1195
1195
1185
1190
1180
1180
1180
1190
1180
1175
1165
1185
1175
1210
1210
1215
1215
1205
1210
1220
1190
1190
1200
1195
60
65
40
50
55
50
50
55
125
160
220
180
215
205
325
250
170
235
255
220
120
65
120
130
240
215
195
285
90
175
105
125
110
550
400
355
210
85
310
175
295
470
415
405
360
405
430
595
215
335
340
355
515
375
360
395
530
130
255
930
940
450
450
450
375
430
210
465
465
1250
1255
460
580
690
570
460
460
605
720
450
400
460
870
460
450
295
520
310
450
270
450
450
1145
900
800
400
450
750
500
670
840
735
710
795
900
670
1,085
500
820
900
980
970
650
770
700
1060
500
700
3.7
4.0
5.1
6.3
7.0
7.6
6.6
14.7
15.0
19.0
10.0
8.2
25.1
19.5
25.2
23.7
20.3
27.1
22.9
17.0
14.9
9.2
14.6
8.5
27.6
25.5
33.5
28.7
16.2
21.3
21.3
15.5
13.7
25.7
24.0
23.9
27.7
10.7
22.5
19.3
23.8
29.2
29.5
29.7
24.4
24.2
32.7
28.7
23.3
22.2
20.7
19.9
28.0
30.0
25.1
29.4
26.6
14.6
20.0
0
0
0
0
0
0
0
0
0
0
80
50
100
0
0
100
80
0
100
0
30
0
0
0
100
0
100
100
0
0
0
0
0
100
0
100
90
0
100
60
200
200
?
?
100
0
180
200
150
100
0
0
180
150
0
200
300
0
0
1.7
1.4
0.8
0.8
0.8
0.8
0.8
0.8
0.8
1.0
1.0
1.0
0.7
0.7
2.1
2.9
2.9
3.3
4.6
3.3
3.3
3.8
3.8
1.6
1.9
1.9
1.9
1.9
5.7
2.3
2.3
1.9
2.3
7.6
7.6
3.3
3.8
1.1
2.9
2.5
2.9
2.9
3.8
2.9
2.3
5.7
5.7
5.7
4.6
3.8
5.7
5.7
6.0
3.8
4.6
4.6
5.7
5.7
3.8
31
46
34
44
49
45
44
58
119
152
199
159
210
198
300
222
147
209
207
179
94
38
89
105
225
200
185
268
59
157
94
110
92
397
280
309
183
76
273
158
273
442
366
370
328
315
363
487
175
280
250
257
404
332
298
339
424
80
208
©2014. American Geophysical Union. All Rights Reserved.
625
Journal of Geophysical Research: Solid Earth
10.1002/2014JB011244
Table 2. (continued)
Site
Longitude
(°W)
Latitude
(°N)
144
148
147
149
150
151
152
153
154
155
159
158
157
168
156
163
164
165
166
167
224
177
176
175
174
173
171
172
170
169
73
72
71
70
69
68
74
109.164
109.163
109.175
109.160
109.157
109.155
109.157
109.159
109.158
109.159
109.158
109.157
109.157
109.166
109.155
109.151
109.143
109.142
109.148
109.148
109.156
109.155
109.147
109.146
109.148
109.153
109.154
109.157
109.156
109.162
109.166
109.168
109.169
109.170
109.173
109.174
109.175
30.976
30.973
30.973
30.969
30.964
30.958
30.952
30.946
30.943
30.935
30.930
30.920
30.917
30.914
30.913
30.909
30.901
30.898
30.891
30.890
30.886
30.879
30.879
30.875
30.870
30.862
30.860
30.857
30.854
30.853
30.849
30.847
30.845
30.842
30.839
30.839
30.834
Cumulative
Distance
(km)
25.35
25.8
d
26.3
26.8
27.6
28.5
29.0
29.4
30.2
30.9
31.9
32.2
d
32.8
33.4
34.6
35.0
35.6
35.75
d
d
36.9
37.2
38.0
38.8
e
e
39.8
40.4
41.0
41.4
41.6
41.8
42.2
42.3
42.8
a
The sites are listed from north to south and
b
Víbora-North segment.
c
Víbora-South segment.
d
Independent secondary rupture segment.
e
Elevation
(m)
Scarp
Height
(cm)
Scarp
Width
(cm)
Scarp
Slope
(deg)
Free
Face
(cm)
Surface
Slope
(deg)
Surface
Offset
(cm)
1185
1185
1145
1195
1195
1180
1200
1160
1140
1120
1130
1150
1130
1105
1060
1010
1080
1080
1050
1040
1000
1020
1060
1070
1030
1020
1010
1000
1040
1000
1010
1005
1020
1020
980
980
970
240
330
55
390
270
410
425
480
365
335
285
435
270
50
200
215
210
330
55
175
115
100
170
275
360
330
140
170
230
300
235
220
165
200
160
220
185
1000
630
400
640
0
840
650
730
1,260
750
860
870
720
380
480
550
300
520
300
500
540
300
400
900
710
650
350
470
440
680
285
120
380
220
180
450
450
13.5
27.6
7.8
31.4
90.0
26.0
33.2
33.3
16.2
24.1
18.3
26.6
20.6
7.5
22.6
21.4
35.0
32.4
10.4
19.3
12.0
18.4
23.0
17.0
26.9
26.9
21.8
19.9
27.6
23.8
39.5
61.4
23.5
42.3
41.6
26.1
22.3
50
190
0
160
270
0
70
50
0
220
0
0
0
0
80
0
140
0
0
0
0
0
0
0
0
0
0
50
50
0
120
180
0
160
125
60
0
2.5
7.6
2.3
11.3
7.6
5.7
11.3
11.3
5.7
7.6
7.6
11.3
7.6
1.9
5.7
2.3
5.7
11.3
3.8
5.7
1.9
5.7
5.7
7.6
7.6
7.6
5.7
5.7
5.7
7.6
7.6
5.7
2.0
7.6
5.7
5.7
5.7
196
246
39
262
270
326
295
334
239
235
170
261
174
37
152
193
180
226
35
125
115
70
130
155
265
243
105
123
186
209
197
208
178
171
142
175
140
marked in Figure 3.
Secondary rupture segment branching off main rupture.
face of the Teras and Otates segments is nondiffusive and seems to be controlled by the growth of
small rills into the scarp.
Fault scarp-bounding fissures have been observed in other earthquake surface rupture studies [e.g., Kurushin
et al., 1997; Fenton and Bommer, 2006]. The existence of the fissure suggests that the rupture formed near
the surface as a mode I fracture (tension gash, tensile or opening mode). However, the vertical surface
displacement and slickenline striations observed within the plane of movement (documented below) clearly
indicate the rupture to be of mode II (in-plane shear). The fissure may have resulted from shaking-induced
ground failure. It also may have opened as a tension gash, within the tip zone of the shear rupture, before
the shear rupture broke through the surface [Klinger et al., 2005], a combination of modes I and II or mixed
mode fracturing [Broek, 1986]. On the other hand, I have not observed any rotation of the ground surface
toward the scarp (Figure 2), which could have explained the formation of the fissure by a curved geometry of
the underlying fault [Xiao and Suppe, 1992].
SUTER
©2014. American Geophysical Union. All Rights Reserved.
626
Journal of Geophysical Research: Solid Earth
10.1002/2014JB011244
5.2. Coseismic Surface Offset
I measured surface offsets at 89
locations along the Pitáycachi fault
(Table 2; vertical separation of the
ground surface in Figure 5). Of these
measurements, 71 were made on the
main rupture scarp (Figure 6a) and 18
were made on rupture branches and
secondary ruptures. On the main
rupture scarp, the maximum vertical
surface offset is 487 cm at site 131
(Figure 3), and the arithmetic mean is
232 cm. The distribution shows two
local minima (arrows in Figure 6a) that
are located where the slip is more
distributed and partitioned onto
several branches or secondary
ruptures (Figure 3).
If surface offset measurements on
branches and secondary ruptures are
projected onto the main rupture
(Figure 6b), these local minima are not
observed, there is less scatter, and the
mean surface offset is higher (251 cm). In
these aggregate vertical surface offsets,
I added the measurements (Table 2) of
sites 98 and 113; 99, 112, and 110; 100
and 115; 108, 117, and 116; 121 and 120;
144 and 147; 157 and 168; 166 and 224;
167 and 224; and 176 and 177.
A quadratic best fit to these aggregate
vertical ground surface offsets
(Figure 6b) indicates an average of
Figure 6. (a) Surface offsets calculated for 71 scarp measurement sites
260 cm. This value corresponds to
along the primary 1887 rupture scarp of the Pitáycachi segment. The
the area below the regression curve
mean surface offset is 232 cm. The maximum offset (487 cm) is located
near the center of the segment. The dashed line is a quadratic best fit. The divided by the difference between the
overall distribution is symmetric. Two local minima, indicated by arrows,
limits of integration, which is the
are located where the slip is partitioned onto several branches or
distance between the northernmost
secondary ruptures (Figure 3 and Table 2). The star marks the estimated
and southernmost measurement sites.
location of the 1887 epicenter. (b) If offset measurements on 11 branches
The average value of the regression
and secondary ruptures are added to offsets along the primary rupture
trace (see text for details), local slip deficits are not apparent, the value of function (260 cm) is greater than the
the regression curve increases slightly, and the mean surface offset
arithmetic mean of the individual
increases to 251 cm. Integrating the area below the regression curve leads
surface offsets (251 cm) because more
to a still higher and probably more representative mean surface offset of
measurements were taken at either
260 cm. The star marks the estimated location of the 1887 epicenter.
end of the surface rupture (Figure 6a),
where the surface offsets taper off. The
average value of the regression is therefore likely to be more representative of average offset than the mean
of the individual surface offsets and was for that reason included in Table 1.
The ground surface offset distribution is symmetric, suggesting a uniform stress drop and homogeneous
lithologic and remote stress conditions [Bürgmann et al., 1994], which is also supported independently by the
low dispersion in the orientation of the slip vectors measured on the 1887 rupture surface (Figure 7). The
maximum ground surface offset is located in the center of the segment (Figure 6) and close to where the
SUTER
©2014. American Geophysical Union. All Rights Reserved.
627
Journal of Geophysical Research: Solid Earth
10.1002/2014JB011244
rupture initiation is inferred from the rupture
branching pattern (see discussion below).
This is in agreement with observations
elsewhere (e.g., 1999 Mw 7.1 Hector Mine,
California earthquake) [Treiman et al., 2002]
that earthquake ruptures nucleate in or
close to regions of large slip, near the major
asperity that they will eventually break
[Manighetti et al., 2005].
The ground surface offset distribution
can be used to obtain the coseismic slip
distribution. The 1887 rupture plane dip is
used to transform from vertical surface
offset measurements (y) to dip slip (s)
(Figure 5). Based on the 70° average dip of
the 1887 rupture plane and the mean
ground surface offset obtained from
integrating the area below the regression
function (Figure 6b, 260 cm), I calculate a
maximum dip slip of 5.16 m and a mean of
2.77 m for the 1887 earthquake on the
Pitáycachi fault (Table 1). This maximum
approaches the maximum historical dip slip
documented for an earthquake surface
rupture in the Basin-and-Range Province,
which was 5.8 m during the 1915 Mw 7.1
Pleasant Valley, Nevada earthquake
[Wallace, 1984].
Figure 7. (a) Lower hemisphere equal-area stereoplot showing the
18 fault planes and slickenline striations measured on the
Pitáycachi segment of the 1887 free face (Table 3). The small open
circle and arrow on each great circle fault measurement represent
the orientation of each slickenline and the sense-of-slip direction on
the hanging wall block. The slickenlines generally indicate east-west
extensional dip slip with a minor left-lateral component. P
(shortening) and T (extension) kinematic axes for each measurement
are represented by closed and open circles, respectively. Average P
and T kinematic axes (black squares) and corresponding great circles
suggesting earthquake focal mechanism nodal planes (P quadrants
shaded in grey) are based on a Bingham Distribution model [Fisher
et al., 1987]. The corresponding average parameters are listed in
Table 4. The figure was generated with the program FaultKin 5, version
5.2, by R. W. Allmendinger. (b) Same data and projection as in Figure 7a.
The tectonic stress tensor (orientations and relative magnitudes of
the principal stresses) was modeled from these directional slip data
based on an inversion algorithm by Angelier [1990]. The dynamics of
this fault population is characterized by an ENE-WSW orientation of the
minimum horizontal stress. See text for further explanations.
SUTER
©2014. American Geophysical Union. All Rights Reserved.
The ground surface offset distribution
can be used as a proxy for the coseismic
slip distribution at depth. In many case
studies, such as the 1992 Mw 7.3 Landers
earthquake [Wald and Heaton, 1994] or the
2002 Mw 7.9 Denali fault, Alaska earthquake
[Haeussler et al., 2004], the surface rupture
measurements are consistent with the
subsurface slip distribution inferred from
the inversion of seismic or geodetic
data. In the case of the preinstrumental
1887 Sonora earthquake, the envelope
formed by the largest surface offsets
at nonconsecutive measuring points
(Figure 6b) probably best represents the
overall slip distribution at depth.
5.3. Directional Data of Coseismic Slip
and Tectonic Stress Tensor
At seven sites, 18 coseismic fault plane
and slickenline measurements were made
along the Pitáycachi segment of the 1887
surface rupture (Table 3 and Figure 7a).
Overall, the slickenline striations indicate
extensional dip slip with a minor left-lateral
628
Journal of Geophysical Research: Solid Earth
10.1002/2014JB011244
a
Table 3. Slickenlines Measured on the Rupture Plane of the Pitáycachi Segment
Site
Longitude
(°W)
Latitude
(°N)
Elevation (m)
Strike (deg)
Dip (°W)
Rake (deg)
Lithology
126
109.136
31.080
1190
109.159
109.161
31.034
31.009
1220
1210
141
143
109.167
109.164
30.991
30.986
1190
1190
152A
109.157
30.951
1195
153
109.159
30.946
1160
63
62
64
68
77
82
86
74
78
77
68
66
75
64
65
60
62
64
90
114
88
85
70
90
102
90
90
84
80
75
74
86
83
75
70
69
limestone
206
138
188
197
195
181
168
166
165
188
201
201
203
206
206
187
188
206
207
222
dike
caliche
granite
granite
rhyolite
rhyolite
Type of
Motion
N
N-RL
N
N-LL
N-LL
N
N-RL
N
N
N-LL
N-LL
N-LL
N-LL
N-LL
N-LL
N-LL
N-LL
N-LL
a
The sites are listed from north to south and marked in Figure 3; the slickenlines are graphed in Figure 7. N: normal; N-RL: normal right-lateral; N-LL: normal
left-lateral.
displacement component. The fault plane dips between 60° and 86°W (mean: 70°W). At several sites, the
measurements indicate a minor component of left-lateral strike slip but only at two sites a minor component
of right-lateral strike slip (Table 3). There is no obvious correlation of these outcrop-scale minor horizontal
components of motion with map-scale geometric irregularities of the rupture trace such as bends or en
echelon arrays. However, the results of a kinematic analysis of these directional slip data (Table 4 and
Figure 7a) correlate with the composite focal mechanisms for well-located microearthquakes by Natali and
Sbar [1982], which suggest dip slip with a right-lateral strike-slip component near the northern tip and normal
dip slip near the southern tip of the Pitáycachi rupture segment (Table 4 and Figure 1).
Furthermore, the focal mechanism obtained by Wallace and Pearthree [1989] for the 25 May 1989 M 4.2
earthquake (Table 4 and Figure 1), which probably had its source on the northern part of the Teras fault [Suter,
2008a], also suggests dip slip with a minor left-lateral strike-slip component on the 65°W dipping nodal
plane [Suter and Contreras, 2002, Figure 2]; slip vector measurements on the Teras fault plane (Table 4) [Suter,
2008a], a proxy for the 1887 slip vector on the Teras rupture segment, are practically identical to the slip vector
of the 25 May 1989 M 4.2 earthquake. Moreover, coseismic slickenlines measured on the Otates segment of the
1887 surface rupture [Suter, 2008b] indicate extensional dip slip without major lateral displacement (Figure 1).
The tectonic stress tensor (orientations and relative magnitudes of the principal stresses) was modeled from
the measurements of coseismic slickenlines on the Pitáycachi segment of the 1887 surface rupture. Based on
a striation inversion algorithm by Angelier [1990], the dynamics of this slickenline population is characterized
by an ENE-WSW orientation of the minimum horizontal stress (Figure 7b). The plunge of the maximum
principal stress S1 is nearly vertical, whereas the plunges of the intermediate principal stress S2 and the least
principal stress S3 are horizontal and nearly horizontal, respectively. The resulting stress ratio ϕ (0.39)
indicates that S2 is closer in magnitude to S3 than to S1. This low ϕ value may explain the intermittent
permutations of the horizontal principal stresses inferred from the Quaternary extension on the nearly
orthogonal Pitáycachi and Guadalupe canyon normal faults (Figure 1).
Summarizing these observations, the 1887 rupture is characterized at the surface and at focal depth by
extensional dip slip on 55° to 72°W dipping faults and additionally by a minor left-lateral strike-slip component
on the Teras and Pitáycachi segments (Table 4 and Figure 1). A likely explanation for these minor horizontal
slip components are subtle differences in strike of the three rupture segments (Figure 1), since the regional
extension is oriented ENE-WSW (Figure 7b). The Otates segment, exhibiting pure dip slip, strikes slightly west
of north, whereas the Pitáycachi and Teras segments, exhibiting each minor left-lateral slip, strike either
north-south or slightly east of north.
SUTER
©2014. American Geophysical Union. All Rights Reserved.
629
SUTER
N-RL
N-LL
N
N-LL
N-LL
N
158
88
90
58
82
94
70
69
72
65
55
68
148
193
215
216
204
170
31.178
c
31.013
30.872
30.823
c
30.670
c
30.437
109.177
c
109.162
109.240
109.389
c
109.250
c
109.164
1
2
3
4
5
6
a
Pitáycachi
Pitáycachi
d
Pitáycachi
e
Teras
f
Teras
Otates
b
The references are listed from north to south and marked in Figure 1, where the parameters are also graphed. N: normal; N-RL: normal right-lateral; N-LL: normal left-lateral. CFM: composite
focal mechanism; S: slickenlines; FM: focal mechanism.
b
Near northern end of Pitáycachi surface rupture segment.
c
Midpoint on the rupture trace between outermost slickenline measurement sites.
d
Near southern end of Pitáycachi surface rupture segment.
e
Focal mechanism for the 25 May 1989 M 4.2 earthquake, which possibly had its source on the Teras fault.
f
Slickenlines measured on the Teras fault plane are given here as a proxy, since no slickenlines were observed on the 1887 free face of the Teras rupture segment.
Natali and Sbar [1982]
This paper
Natali and Sbar [1982]
Wallace and Pearthree [1989]
Suter [2008a]
Suter [2008b]
CFM
S
CFM
FM
S
S
0
24
27
14
8
23
100
281
305
284
283
264
30
66
63
57
75
67
10
109
125
170
160
71
Plunge
(deg)
Plunge
(deg)
Trend
(deg)
Type of
Motion
Rake
(deg)
Dip
(°W)
Strike
(deg)
Latitude
(°N)
Longitude
(°W)
Rupture
Segment
Reference
Table 4. Slip Parameters on the 1887 Rupture Segments From Slickenlines and Focal Mechanisms
a
P Axis
Trend
(deg)
T Axis
Type of
Measurement
Source
Journal of Geophysical Research: Solid Earth
10.1002/2014JB011244
5.4. Structure of the Surface Rupture
and Its Segmentation
Here I describe the surface rupture zone of the
Pitáycachi segment from north to south. In its
northern part, the rupture trace passes entirely
within piedmont alluvial fan and terrace fill deposits,
whereas in its southern part, the trace passes
along the range front and separates mostly bedrock
from piedmont deposits (Figure 3). The rupture
shows a well-developed bipolar branching pattern
consisting of at least six north facing bifurcations in
the northern part of the segment and two south
facing bifurcations in its southern part (Figure 3).
Furthermore, on a map scale, the rupture trace shows
a complex pattern of second-order segmentation
internal to the Pitáycachi segment (Figure 3):
South of Arroyo Pitáycachi, two segments with
an across-strike separation of 2.0 km are linked by an
en echelon scarp array; where the surface rupture
traverses the La Cabellera cross fault, it is characterized
by a relay zone with multiple strands; and in its
southernmost part, the surface rupture is composed
of an en echelon array (Figure 3). The width of the
rupture zone is typically narrow (1–5 m) but can be
wider (up to 2 km) near the bifurcations and links
between second-order segments. No secondary
hanging wall deformation such as antithetic scarps or
back-tilted strata, commonly associated with normal
fault surface ruptures [McCalpin, 2009], was observed
along the 1887 rupture trace.
5.4.1. Víbora Segments
Two short segments, Víbora-South and Víbora-North,
exist to the north of the Pitáycachi rupture segment
(Figure 3 and Table 1). Both are located on Mesa La
Víbora, a large planar Pleistocene relict alluvial fan
surface, northwest of Cajón Bonito. The fault scarps
are easily detectable on aerial photographs and
satellite imagery because of being associated with a
strip of denser vegetation along them. Along a third
such feature, located west of the Víbora segments
(near western edge of Figure 3), the 1887 surface
offset was so small that it was not discrete enough
to be measured 125 years later. Víbora-North strikes
N16°E, has a length of 820 m, and a maximum
surface offset of 46 cm (Table 1). Similarly, Víbora-South
strikes N13°W, has a length of 2070 m, and a
maximum surface offset of 49 cm (Table 1). It does
not continue into the Pleistocene deposits of the
Cajón Bonito drainage.
5.4.2. Pitáycachi Segment
The northernmost part of the Pitáycachi segment
(north of site 97, Figure 3) is characterized by a
counterclockwise 20° bend of the rupture trace
from north to north-northwest. At its northern end
©2014. American Geophysical Union. All Rights Reserved.
630
Journal of Geophysical Research: Solid Earth
10.1002/2014JB011244
Figure 8. (left and right) Map (location outlined in Figure 3) and vertical aerial photograph showing detail of surface
rupture branching pattern and related slip partitioning in the northern part of the Pitáycachi segment, south of Cañón
de los Embudos. The 1887 fault scarps are expressed by a dark line of denser vegetation. The measured surface offsets
(values in Figure 8 (left), shown in meters) are a proxy for the near-surface slip along the rupture. See text for discussion.
(north of site 86), the southwest dipping rupture bends an additional 44° and terminates against the
projected trace of the WNW-ESE striking, NNE dipping Los Embudos cross fault, which crops out east of the
rupture trace, in the footwall of the Pitáycachi fault (Figures 1 and 3). It is likely that the Los Embudos fault
acted as a structural barrier and arrested the northward propagating 1887 rupture, as the along-rupture
surface offset decays abruptly toward the northern endpoint, at only 8.6 km distance from 4 m high scarps at
site 107 (Figures 3, 6, and 8 and Table 2).
A major branching pattern, composed of two or more bifurcations, is located between sites 97 and 107
(Figures 3 and 8). These bifurcations open to the north, suggesting that the rupture propagated toward the
north along this stretch of the fault trace. The individual branches formed on either side of the main rupture,
displace previously undeformed alluvial fan deposits, and are characterized by dense vegetation bands on
aerial photographs (Figure 8). The mapped length of the longest branch is 3.0 km. A fracture mechanics
interpretation of this branching pattern, based on its comparison with the results of analog experiments, is
given below.
Notable map-scale rupture trace characteristics south of site 107 are (from north to south) a shear lens
between sites 106 and 107, another north facing rupture bifurcation south of Arroyo La Cueva (sites 120
and 121, Figure 3), and a major west bend of the trace south of this branching point. The northernmost
outcrop where the rupture sheared alluvial fan deposits against bedrock is located north of site 125. Site 126
yielded the northernmost structural measurements on the rupture surface, which dips there 63°; slip was
normal, without any measurable horizontal component (Table 3). Two more north facing rupture bifurcations
were mapped directly north of Arroyo Pitáycachi (between sites 127 and 128); two branches, 300–400 m long,
splay off the main rupture on either side at angles between 20 and 40° (Figure 3).
Directly south of Arroyo Pitáycachi, a left-stepping en echelon rupture pattern, composed of six map-scale 1887
rupture segments (Figure 9), bridges the boundary between two regional scale, approximately north-south
striking normal fault segments with an across-strike separation of 2.0 km (Figure 3). This regional-scale, 2 km
wide right step over should have left-lateral motion, whereas the smaller, left step overs of the en echelon array,
each <100 m wide, should have right-lateral motion across each of these small zones. The only slickenlines
measured on the surface rupture within the en echelon array, at site 206, indicate a minor left-lateral strike-slip
component (Table 3).
SUTER
©2014. American Geophysical Union. All Rights Reserved.
631
Journal of Geophysical Research: Solid Earth
10.1002/2014JB011244
Figure 9. (left) Uninterpreted and (right) interpreted vertical aerial photograph showing the 1887 rupture trace (white
lines) south of Arroyo Pitáycachi, which forms a left-stepping en echelon array composed of six segments. Also note the
north facing rupture bifurcation in the lower part of the figure. On the uninterpreted photograph in Figure 9 (left), the 1887
rupture scarp segments are expressed by a dark line caused by a strip of denser vegetation. These features are best seen
looking at the rupture from the northwest. The location of this photograph is outlined in Figure 3.
Notwithstanding the oblique orientation of the overall slip vector within this relay zone, the largest vertical 1887
rupture surface offset (487 cm) was measured within this en echelon array, at site 131. The fact that the
along-strike surface offset is not tapering toward the relay zone but increases to the maximum observed offset
strongly suggests that the two regional-scale north-south striking fault segments form a single through-going
structure at depth and that the en echelon segments are branches of this upwardly bifurcating fault.
The southernmost north facing branching pattern was mapped at the southern end of this en echelon
array (Figure 9), between sites 134 and 135. In contrast, the two branching patterns following farther
south open to the south (Figure 3); the northern one of them being located only 5 km farther south, near
site 146. There, the eastern, NNW-SSE striking branch of the 1887 surface rupture is only 300 m long, even
though most of the cumulative throw prior to 1887 is taken up by this bedrock-bounding normal fault
(Figure 3). The western, approximately north-south striking rupture strand, on the other hand, is much
longer and offsets alluvium.
South of Arroyo El Salis, the surface rupture passes for about 5 km along the mountain front, is oriented
north-south, and separates bedrock from alluvial fan deposits (Figures 3 and 10). Here secondary ruptures,
too small to show in Figure 3, are visible south of site 157 on a vintage oblique aerial photograph (right
foreground in Figure 10); they could be either R1 Riedel shears or south facing rupture branches or both. A more
detailed inspection of the outcrop did not clarify this ambiguity; there is local fracture cleavage and
tectonic breccia but very minor displacement and no observable slickenlines. Coseismic Riedel shear
structures are typically related to strike-slip movement [Tchalenko and Ambraseys, 1970; Haeussler et al., 2004;
Lin and Nishikawa, 2011]. If the observed secondary ruptures are R1 Riedel shears, they indicate a component
of left-lateral horizontal strike-slip motion within the rupture zone. This interpretation is supported on a map
scale by the major right-stepping en echelon array of surface rupture segments between Arroyo de la Cabellera
and the Bavispe River (Figure 3), which also indicates a component of horizontal left-lateral motion.
Near Arroyo La Cabellera, the rupture trace gradually changes to a more NW strike, parallel to the strike of the
La Cabellera cross fault and comes to an end within alluvium, in the projected continuation of that fault
(Figure 3). The 1887 rupture was arrested by this major WNW-ESE striking cross fault, which bounds the Sierra
SUTER
©2014. American Geophysical Union. All Rights Reserved.
632
Journal of Geophysical Research: Solid Earth
10.1002/2014JB011244
La Cabellera mountain range and
has a throw >1250 m. This structural
configuration is very similar to the one at
the northern termination of the 1887
rupture at the Los Embudos fault. Across
the La Cabellera fault, the surface rupture
is characterized by a right step over with
an across-strike separation of 0.6 km
(Figure 3). Accordingly, the southernmost
part of the Pitáycachi surface rupture,
between the La Cabellera cross fault and
the Bavispe River (Figure 3), could be
considered a separate segment. However,
the profile of aggregate surface offsets
(Figure 6b) does not show a significant
trough at the latitude of this cross fault.
Figure 10. Looking north along the 1887 rupture scarp (white arrows) in
the central part of the Pitáycachi segment (modified from DuBois and
Smith [1980, Figure 4A]; photograph taken by Pete Kresan). The rupture
separates bedrock in the footwall from piedmont alluvium in the
hanging wall. In the foreground, the surface offset is 261 cm at site 158
and 174 cm at site 157 (Figure 3 and Table 2). Note the south facing
rupture bifurcation in the right foreground marked by an arrow. A,
piedmont alluvium; R, rhyolite (Cerro Pitáycachi); Mo, La Morita
Formation (east dipping siliciclastic rocks); G, granite; D, dike within La
Morita Formation (mapped in Figure 3). The distance between the
foreground and Cerro Pitáycachi is ~20 km.
South of site 166, the surface rupture
strikes north-south and crosses a ~3 km
wide relatively young and broad drainage
where the upper part of the Bavispe River
likely previously flowed into the San
Bernardino Basin. The Pliocene to early
Pleistocene (2 Ma–750 ka) braided-stream
deposits are displaced by the fault and dip
east in its footwall (Figure 3). Here the
topographic relief of the compound fault
scarp measures at most 60 m (Figure 11).
Farther south, the 1887 surface rupture
separates the fill of the San Bernardino
Basin from bedrock as cumulative throw
increases (Figure 3).
The southernmost part of the Pitáycachi
surface rupture segment shows a south
facing bifurcation near site 171 and
forms a map-scale right-stepping en
echelon array, composed of five segments, between site 173 and the Bavispe River. The relay ramps between
most of the segments are breached, which is expressed by sudden bends in the continuous fault trace
(Figures 3 and 11). The en echelon pattern indicates a minor left-lateral strike-slip component within this relay
zone, which is compatible with the relative sense of motion expected in a right step over between adjacent
normal fault segments.
6. Structural Interaction Between the Pitáycachi and Teras Rupture Segments
According to the historic observations by Goodfellow [1887], the Pitáycachi and Teras rupture segments
(Figure 1) overlap. Based on his rupture map, the southern end of the Pitáycachi segment continues across
the Bavispe River into the Sierra Pilares de Teras for at least 8 km. The same rupture configuration is shown on
the map by Aguilera [1888]. Facsimilar reproductions of these historic surface rupture maps by Goodfellow
[1887] and Aguilera [1888] can be found in Suter [2006]. According to Goodfellow [1888], the scarp heights
along the Pitáycachi rupture segment in the Sierra Pilares de Teras are ~0.9 m and do not exceed 1.5 m. Both
Goodfellow [1888] and Aguilera [1888] noticed the scarp height attenuate to near zero, approaching either
side of the Bavispe River Valley. Goodfellow [1888] observed local reverse faulting, 3.2 km south of the river. At
his southernmost observation point, 8 km south of the river, the scarp height was ≤30 cm and the rupture
branched. Because the rupture passes in the Sierra Pilares de Teras within volcanic rocks and is not expressed
SUTER
©2014. American Geophysical Union. All Rights Reserved.
633
Journal of Geophysical Research: Solid Earth
10.1002/2014JB011244
Figure 11. Historic 1887 photograph by Camillus S. Fly of the 1887 earthquake rupture in alluvial gravel east of Arroyo La
Cabellera. The view is toward the south-southeast from 109.187° longitude west/30.874° latitude north, ~150 m south of
site 175 (Figure 3). The height of the surface rupture scarp in the foreground is ~3 m and measures 3.6 m at site 174. The
two hills in the background are located in the footwall of the Pitáycachi fault and composed of Tertiary basalt. The arrow
points to an east-west oriented portion of the rupture scarp, between measurement sites 169 and 170 (Figure 3), which
connects the segments of a breached relay ramp. This caption corrects the location and description of the same photograph
in Suter [2006, Figure 7].
by a compound rupture scarp, I have not been able to confirm Goodfellow’s observations. A photograph
by C. S. Fly [Suter, 2006, Figure 10] documents the 1887 rupture trace directly south of the river, where it
separates Tertiary volcanic rocks from scree.
Based on my observations, a right step over separates the Pitáycachi and Teras segments of the 1887
earthquake surface rupture across the Bavispe River near Colonia Morelos (Figures 1 and 3) [Suter, 2008a]. The
across-strike separation measures 2.5 km. This step over coincides with a minimum in the vertical 1887
surface offset distribution (Figure 12). Within ~10 km of this first-order step over, both the Pitáycachi and
Teras rupture segments are characterized by several second-order right-stepping en echelon step overs
indicating a distributed left-lateral component of motion across this transfer zone. This array of the rupture
trace makes the existence of the Pitáycachi segment south of the Bavispe River reported by Goodfellow [1887]
questionable. Alternatively, the segment described by Goodfellow [1887] south of the river could be an
independent rupture segment or a northward bifurcating branch of the Teras rupture segment.
The local minimum in the vertical surface offset profile (Figure 12) suggests that Pitáycachi and Teras are
independent rupture segments that do not merge at depth [Suter and Contreras, 2002]. No map-scale
cross fault exists in the Tertiary volcanic rocks east of the step over (Figure 3). However, this step over is located
at the southern axial termination of the San Bernardino Basin and in the projected western continuation of the
southern margin of the Llanos de Carretas Basin (Figure 1), which is possibly fault bound. This suggests the
existence of an east-west striking cross fault in the basement, beneath the Tertiary volcanic rocks, in the step
over region of the 1887 rupture between the Pitáycachi and Teras segments.
7. Coseismic Surface Offset Profile and Segmentation of Entire 1887 Rupture
I aggregated the vertical surface offsets along the Pitáycachi segment with the ones along the Víbora-North,
Víbora-South, Teras, and Otates segments into a single profile (Figure 12). Disregarding the two minor
Víbora rupture segments, the rupture was arrested in the north as well in the south at major cross faults
(Figures 1 and 12), and the surface offset profile tapers rapidly toward these faults. The segmented structure
of the surface rupture is noticeable from the surface offset distribution as well as the rupture trace geometry.
Such segmented ruptures are typical for large (Mw > 7.0) historical earthquakes of the Basin-and-Range
SUTER
©2014. American Geophysical Union. All Rights Reserved.
634
Journal of Geophysical Research: Solid Earth
10.1002/2014JB011244
Figure 12. Fault trace map (top part of figure, with north to the left) and surface offset scatter diagram of the entire 1887 earthquake rupture. The star marks the
estimated location of the 1887 epicenter.
Province, such as the 1915 Pleasant Valley, Nevada; 1954 Fairview Peak, Nevada; or 1959 Hebgen Lake,
Montana, earthquakes [Doser and Smith, 1989; dePolo et al., 1991; Zhang et al., 1999]. Similarly, the Santa
Rita fault zone, located along the southeastern margin of the Tucson Basin in southeastern Arizona, is a
late Pleistocene example of a segmented Basin-and-Range Province fault rupture with an estimated
recurrence time of 100 ka and an estimated magnitude up to 7.3, which is based on the fault length
of 58 km and scarp heights up to 7 m [Pearthree and Calvo, 1987].
The boundary between the Pitáycachi and Teras segments is defined by a 2.5 km wide right step over and a
local minimum in the surface offset profile. The boundary between the Teras and Otates segments is not
noticeable from the profile but is defined by a 15 km long gap in the rupture trace (Figures 1 and 12). The
offset distributions of both the Teras and Otates segments are asymmetric, with the maximum being located
toward the neighboring rupture segment. This suggests that the Teras and Otates segments could be part of
a single continuous 1887 rupture that stepped across the structurally complex basement ridge between
them [Suter, 2008a, 2008b]. However, this suggestion is not supported by field observations.
The maximum surface offset for the entire 1887 rupture measures 4.87 m and the weighted average offset for
the entire rupture is 1.97 m (Table 1). The ratio between the maximum and average offsets and the overall
shape of the 1887 surface offset profile (Figure 12) are comparable to the ones of historical normal faulting
earthquakes in the western U.S., where the maximum slip typically is more than twice the mean and occurs
SUTER
©2014. American Geophysical Union. All Rights Reserved.
635
Journal of Geophysical Research: Solid Earth
Slip-to-Length Ratios of Individual Rupture Segments
6
5
Slip (m)
4
3
2
10.1002/2014JB011244
along a limited extent of the rupture.
Displacements within 20% of the
maximum characterize only about 5% of
the rupture length [Hecker et al., 2010]. In
the case of the 1887 Sonora earthquake,
the ratio between the maximum and
average offsets for the entire rupture is 2.5,
and surface offsets within 20% of the
maximum (3.9 m) extend over 18%
(18.2 km) of the rupture length.
1
8. Scaling Relation Between
Rupture Length and Slip
0
0
10
20
30
40
The average and maximum slip-to-length
ratios along the four major rupture
Figure 13. Scatter diagram showing the rupture length-to-slip ratios
segments, Otates, Teras, Pitáycachi, and
of the major individual 1887 rupture segments (from Table 1) and
Víbora-South (Table 1) both broadly
related linear regressions. Open circles: maximum slip; closed circles:
define a linear relationship (Figure 13). For
mean slip; solid line: mean slip-to-length regression; dashed line:
the Víbora-South segment, its average
maximum slip-to-length regression.
and maximum surface offset values were
taken as proxies for lacking slip values.
5
The slope of the mean slip-to-length regression curve, 5.9 × 10 , is similar to the 6.5 × 10 5 value obtained
for a larger set of intraplate earthquake rupture parameters by Scholz [2002, Figure 4.11] and the 8 × 10 5
value obtained by Ferrill et al. [2008] for a fault zone that slipped in the 1992 Landers earthquake.
Rupture Length (km)
In both regression models, the coefficient of determination (r2 in Figure 13) is close to one, which indicates that
independent of the length of the rupture segments, the individual slip-to-length ratios are fairly constant. These
constant ratios support the interpretation that Teras and Otates are separate rupture segments and not a
continuous segment as could be inferred from their along-strike surface offset distributions (Figure 12).
Since the slip-to-length ratios of the individual rupture segments are practically identical, it can be inferred
that the 1887 rupture segments are self-similar; their slip scales linearly with their length (Figure 13), which is
in agreement with fracture mechanics models of faulting such as the Dugdale-Barenblatt and the Critical
Fault Tip Taper models [Cowie and Scholz, 1992]. Furthermore, since the maximum slip-to-length ratio is a
strain that corresponds to the static stress drop [Scholz, 2007], it can be inferred that all the 1887 rupture
segments experienced a similar static stress drop.
9. Pleistocene Ruptures of the Pitáycachi Fault
In the northern part of the Pitáycachi segment, the 1887 rupture coincides with the base of a compound scarp
that displaces piedmont alluvial fan surfaces vertically up to 73 m (Figure 14). These surfaces are estimated by
Bull and Pearthree [1988] and Pearthree et al. [1990] to be of early to middle Pleistocene (1.5 Ma–125 ka) age
based on soil profile characteristics. Several topographic profiles of the displaced footwall fan surface (Figure 14a)
were constructed with an Abney level with clinometer and a telescopic leveling rod (see Figure 3 for profile
traces). The northernmost profile (D) is located on Mesa La Víbora, two profiles (A and B) are located between
Arroyo Los Embudos and Arroyo La Cueva, and the southernmost profile (C) directly south of Arroyo Pitáycachi.
No other vertically displaced piedmont alluvial fan surfaces could be observed in the footwall of the fault system
farther south along the Pitáycachi fault (with a possible exception east of La Cabellera ranch). No alluvial fan
deposits are observed in the footwall of the Teras or the Otates fault [Suter, 2008a, 2008b].
I plot the along-strike offset of the Pleistocene alluvial surface together with the 1887 along-strike
surface offsets (Figure 14b). These offset profiles are broadly proportional, indicating that events that
postdate the deposition of the Pleistocene surface, approximately 10 all together, had slip distributions
similar to 1887. This suggests that the 1887 rupture dimensions are characteristic for ruptures along
the Pitáycachi fault.
SUTER
©2014. American Geophysical Union. All Rights Reserved.
636
Journal of Geophysical Research: Solid Earth
10.1002/2014JB011244
10. Inferred Rupture
Nucleation and Rupture Path
Whereas the source and kinematics of
recent earthquakes have been
modeled by inverting seismic and
geodetic observations, the epicenter
and rupture propagation of this
preinstrumental earthquake are
inferred here from surface structural
observations such as rupture
branching and the slip distribution. As
described above, the rupture shows a
well-developed bipolar branching
pattern consisting of at least six north
facing bifurcations in the northern part
of the segment and two south facing
bifurcations in its southern part
(Figure 3). Branching probably resulted
from the lateral propagation of the
rupture, given that the much shorter
ruptures of the Otates and Teras
segments did not develop branches.
Since the rupture length (≥41 km) is
much longer than the rupture width
(~15 km), the rupture propagated only in
horizontal direction after having
breached the entire brittle thickness of
the crust, allowing for branching.
In analog experiments of branching
ruptures, the bifurcations are
consistently in the direction of rupture
Figure 14. (a) Topographic scarp profiles (traces marked in Figure 3) propagation [Bahat, 1982; Sharon et al.,
showing the displacement of a Pleistocene alluvial fan surface by the
1995; Sharon and Fineberg, 1999;
Pitáycachi (profiles A–C) and Víbora-South (profile D) faults. For each
Bouchbinder et al., 2010]. The polarity of
profile, the origin of the reference system is at the 1887 rupture trace.
the bifurcations can therefore be taken
(b) Along-strike offsets of the 1887 surface rupture (black circles) and of
Pleistocene alluvial fan surface (black diamonds). The displacement of the as an indicator of the direction of
fan surface increases toward the center of the Pitáycachi rupture segment rupture propagation, and the changes
and is broadly proportional to the 1887 surface offset.
in polarity may indicate where the
rupture initiated. This suggests that
the rupture of the Pitáycachi segment originated in its central part, where the polarity of the rupture
bifurcations changes (Figure 3). To locate the epicenter of the 1887 main shock, the point on the rupture
trace centered between the innermost branches of opposed polarity was extrapolated to a depth of 15 km
(base of the seismogenic crust) with the 80° dip of the rupture measured nearby. In this way, the epicenter
was located at 109.190° longitude west/31.005° latitude north (Figure 3); the maximum horizontal
uncertainty of the epicenter location in north-south direction is 5 km. The point of rupture nucleation is,
probably coincidentally, nearly at the midpoint between the surface rupture trace extremities.
The largest branching pattern, composed of two or more bifurcations, is located south of site 97 (Figures 3
and 8). The individual branches formed on either side of the main rupture, and all the bifurcations open to
the north. Based on the polarity of its branching pattern, the rupture is inferred to have propagated here from
south to north. The rupture front separated the ground surface vertically by 4 m at site 107 before
becoming unstable and distributing into several branches. Most of the branches die out after a relatively
short distance. Most of the slip is taken up by the 3 km long western branch. The through-going major
SUTER
©2014. American Geophysical Union. All Rights Reserved.
637
Journal of Geophysical Research: Solid Earth
10.1002/2014JB011244
rupture has a low surface offset where it passes parallel to and east of this western branch but gains again
more offset farther north, beyond the tips of the other branches (Figure 8).
Similar branching patterns have been observed in millimeter-scale analog experiments of the dynamic growth
of mode I fractures (tension gash, tensile or opening mode) in brittle materials such as Plexiglas [Sharon
et al., 1995]. A dynamic instability occurs as the speed of a crack increases beyond a critical value (typically
0.4 times the speed of Raleigh waves), at which a crack will change its topology as it sprouts short-lived side
branches. As in the observed branching pattern of the 1887 rupture (Figures 3 and 8), the experimental
branches, in general, do not appear as a single side branch but in bunches that progressively become more
complex as the mean crack velocity increases [Sharon et al., 1995]. Because a rupture is driven by the energy flux
to its tip, the formation of a new branch decreases the amount of energy flowing into the main rupture, thereby
reducing its velocity [Broek, 1986]. Conversely, the death of a branch causes the main rupture to accelerate,
as more energy is available to drive it. This results in fluctuations in the velocity of the rupture front, with a
timescale characteristic of a branch lifetime [Sharon and Fineberg, 1999]. Our observations in conjunction with
these millimeter-scale experimental results suggest that the properties of branching instabilities are length
scale independent and that the measured surface offsets (and implicitly slip) within the documented 1887
branching pattern (Figure 8) correlate with the fluctuations in rupture velocity.
Based on the branching pattern, the rupture first propagated bilaterally along the Pitáycachi fault, from
where the southern rupture front first cascaded across a 2.5 km wide step over to the Teras fault and then
across a structurally complex basement ridge to the Otates fault. The postulated rupture path is in agreement
with the intensity distribution of this earthquake [Aguilera, 1888; Sbar and DuBois, 1984], which indicates a
strong rupture directivity effect toward the south. Studies of the historical seismicity [Suter, 2001] and recent
microseismicity [Castro et al., 2010] of this region indicate that the area of aftershock activity has increased
southward with time along the same fault system. This may also be related to the southward direction of
the 1887 rupture propagation and the related rupture directivity effect.
11. Conclusions
In the 3 May 1887 Sonoran earthquake (surface rupture end-to-end length: 101.8 km; Mw: 7.5 ± 0.3) an array of three
major north-south striking half-graben bounding Basin-and-Range Province faults (from north to south Pitáycachi,
Teras, and Otates) ruptured sequentially along the western margin of the Sierra Madre Occidental Plateau.
Based on fault slickenline measurements and focal mechanisms of microseismicity attributed to related
aftershocks, the 1887 rupture is characterized at the surface and at midcrustal level by extensional dip slip on
these 65–70°W dipping faults and additionally by a minor left-lateral strike-slip component on the Teras and
Pitáycachi segments, which can be explained by the slightly oblique orientation of these faults with respect to
the ENE-WSW regional extension. The rupture was arrested in the north as well as in the south at major cross
faults, and the surface offset profile tapers rapidly where the 1887 rupture approaches these preexisting cross
faults. The 1887 surface rupture segments have a consistent maximum slip-to-rupture length ratio, which
indicates that the defined rupture segments are self-similar and experienced a similar static stress drop.
With a length of ≥41 km, the Pitáycachi segment, analyzed in detail in this paper, is a composite of second-order
fault segments internal to the Pitáycachi segment, which are mostly bridged by en echelon rupture
scarp arrays. This second-order segmentation is not detectable as local minima in the 1887 along-rupture
surface offset profile, which indicates that the segments form a single through-going structure at
depth, and the en echelon segments are branches of this upwardly bifurcating fault. Based on surface
deformation, maximum slip along the Pitáycachi fault in the 1887 earthquake was 516 cm and the average
slip was 277 cm. The maximum comes close to the maximum observed dip slip of a historic earthquake
surface rupture in the Basin-and-Range Province, which was 5.8 m during the 1915 Mw 7.1 Pleasant
Valley, Nevada, earthquake. The 1887 slip distribution for the Pitáycachi fault is symmetric, suggesting a
uniform stress drop and homogeneous lithologic and remote stress conditions, which is also supported
independently by the low dispersion in the orientation of the 1887 slip vectors measured on the surface
rupture. The 1887 along-strike surface offsets are broadly proportional to the 1 order-of-magnitude larger
along-strike offsets of a distinct Pleistocene alluvial fan surface, which suggests that the 1887 rupture
dimensions may be characteristic for ruptures along the Pitáycachi fault.
SUTER
©2014. American Geophysical Union. All Rights Reserved.
638
Journal of Geophysical Research: Solid Earth
10.1002/2014JB011244
The Pitáycachi surface rupture shows a well-developed bipolar branching pattern, which permits inferences
about slip partitioning, the rupture path, and the physics of the rupturing process. The branching pattern
consists of at least six north facing bifurcations in the northern part of the segment and two south facing
bifurcations in its southern part. Branching probably resulted from the lateral propagation of the rupture due
to the length of the fault, given that the much shorter ruptures of the Otates and Teras segments did not
develop branches. This branching pattern indicates that the 1887 rupture nucleated in the central part of this
segment, where the polarity of the rupture bifurcations changes, and propagated bilaterally. To locate the
epicenter of the 1887 main shock, a point on the rupture trace, centered between the innermost branches of
opposed polarity, was extrapolated to a depth of 15 km with the 80° dip of the rupture measured nearby.
Based on this technique, the epicenter is estimated to be at 109.190° longitude west/31.005° latitude north.
Our observations in conjunction with the results of millimeter-scale analog experiments [Sharon et al., 1995]
suggest that the properties of branching instabilities are length scale independent and that the measured
surface offsets (and implicitly slip) within the documented 1887 branching pattern correlate with the
fluctuations in rupture velocity.
Acknowledgments
All data necessary to understand,
evaluate, replicate, and build upon the
reported research are contained within
the paper. This project was made
possible in part by financial and logistic
support from Universidad Nacional
Autónoma de México (UNAM) and
Consejo Nacional de Ciencia y
Tecnología (CONACYT, grant G33102-T).
The historical photographs of the
surface rupture by Camillus S. Fly are
courtesy of the Karl V. Steinbrugge
Collection, National Information Service
for Earthquake Engineering, University
of California, Berkeley. I am thankful to
Phil Pearthree, an anonymous reviewer,
and especially Scott E. K. Bennett for
their detailed and constructive
evaluations of the manuscript. My
participation in an excursion to the
Pitáycachi fault in 1990, organized by
Phil Pearthree and William Bull during a
Cordilleran Section meeting of the
Geological Society of America,
fomented my interest in the structure of
the 1887 surface rupture. More recently,
my understanding of the 1887
rupture benefited from discussions with
several colleagues at the Swiss Nuclear
Regulatory Commission (ENSI) and at
the 2013 Palo Verde Nuclear Generating
Station Seismic Hazard Evaluation
workshop.
SUTER
References
Aguilera, J. G. (1888), Estudio de los fenómenos séismicos del 3 de mayo de 1887, Ann. Minis. Fomento de la República Mexicana, 10, 5–56.
Angelier, J. (1990), Inversion of field data in fault tectonics to obtain the regional stress. III: A new rapid direct inversion method by analytical
means, Geophys. J. Int., 103, 363–376.
Bahat, D. (1982), Extensional aspects of earthquake induced ruptures determined by an analysis of fracture bifurcation, Tectonophysics, 83, 163–183.
Baldridge, W. S., G. R. Keller, V. Haak, E. Wendlandt, G. R. Jiracek, and K. H. Olsen (1995), The Rio Grande rift, in Continental Rifts: Evolution,
Structure, Tectonics, Dev. in Geotectonics, vol. 25, edited by K. H. Olsen, pp. 233–275, Elsevier Science B. V, Amsterdam, Netherlands.
Berglund, H. T., A. F. Sheehan, M. H. Murray, M. Roy, A. R. Lowry, S. Nerem, and F. Blume (2012), Distributed deformation across the Rio Grande
rift, Great Plains, and Colorado Plateau, Geology, 40, 23–26, doi:10.1130/G32418.1.
Biggs, T. H., R. S. Leighty, S. J. Skotnicki, and P. A. Pearthree (1999), Geology and geomorphology of the San Bernardino valley, southeastern
Arizona, Arizona Geological Survey, Open File Rep. 99–19, 20 pp.
Boncio, P., P. Galli, G. Naso, and A. Pizzi (2012), Zoning surface rupture hazard along normal faults: Insight from the 2009 MW 6.3 L’Aquila,
central Italy, earthquake and other global earthquakes, Bull. Seismol. Soc. Am., 102, 918–935.
Bouchbinder, E., J. Fineberg, and M. Marder (2010), Dynamics of simple cracks, Ann. Rev. Condens. Matter Phys., 1, 371–395.
Broek, D. (1986), Elementary Engineering Fracture Mechanics, 4th revised ed., Martinus Nijhoff Publishers, Dordrecht, Netherlands.
Bucknam, R. C., and R. E. Anderson (1979), Estimation of fault-scarp ages from a scarp-height-slope-angle relationship, Geology, 7, 11–14.
Bull, W. B., and P. A. Pearthree (1988), Frequency and size of Quaternary surface ruptures of the Pitaycachi fault, northeastern Sonora Mexico,
Bull. Seismol. Soc. Am., 78, 956–978.
Bürgmann, R., D. D. Pollard, and S. J. Martell (1994), Slip distribution on faults: Effect of stress gradients, inelastic deformation, heterogeneous
host-rock stiffness, and fault interactions, J. Struct. Geol., 16, 1675–1690.
Castro, R., P. M. Shearer, L. Astiz, M. Suter, C. Jacques-Ayala, and F. Vernon (2010), The long-lasting aftershock series of the 3 May 1887 MW 7.5
Sonora earthquake in the Mexican Basin and Range Province, Bull. Seismol. Soc. Am., 100, 1153–1164.
Chapin, C. E., and S. M. Cather (1994), Tectonic setting of the axial basins of the northern and central Rio Grande rift, in Basins of the Rio
Grande Rift: Structure, Stratigraphy, and Tectonic setting, edited by G. R. Keller and S. M. Cather, Geol. Soc. Am. Spec. Pap., 291, 5–25.
Cole, J. C., S. A. Mahan, B. D. Stone, and R. R. Shroba (2007), Ages of Quaternary Rio Grande terrace-fill deposits, Albuquerque area, New
Mexico, N. M. Geol., 29(4), 122–132.
Cowie, P. A., and C. H. Scholz (1992), Physical explanation for the displacement-length relationship of faults using a post-yield fracture
mechanics model, J. Struct. Geol., 14, 1133–1148.
dePolo, C. M., and J. G. Anderson (2000), Estimating the slip rates of normal faults in the Great Basin, USA, Basin Res., 12, 227–240.
dePolo, C. M., D. G. Clark, D. B. Slemmons, and A. R. Ramelli (1991), Historical surface faulting in the Basin and Range province, western North
America: Implications for fault segmentation, J. Struct. Geol., 13, 123–136.
Doser, D. I., and R. B. Smith (1989), An assessment of source parameters of earthquakes in the Cordillera of the western United States, Bull.
Seismol. Soc. Am., 79, 1383–1409.
DuBois, S. M., and A. W. Smith (1980), The 1887 Earthquake in San Bernardino valley, Sonora: State of Arizona, Bureau of Geology and Mineral
Technology, Spec. Pap. 3, 112 pp.
Fenton, C. H., and J. J. Bommer (2006), The MW 7 Machaze, Mozambique earthquake of 23 February 2006, Seismol. Res. Lett., 77, 426–439.
Ferrill, D. A., K. J. Smart, and M. Necsoiu (2008), Displacement-length scaling for single-event fault ruptures: Insight from Newberry Springs
fault zone and implications for fault zone structure, in The Internal Structure of Fault Zones: Implications for Mechanical and Fluid-flow
Properties, edited by C. A. J. Wemberley et al., Geol. Soc. London Spec. Publ., 299, 113–122.
Fisher, N. I., T. Lewis, and B. J. J. Embleton (1987), Statistical Analysis of Spherical Data, Cambridge Univ. Press, Cambridge, U. K.
González-León, C. M., V. A. Valencia, M. López-Martínez, H. Bellon, M. Valencia-Moreno, and T. Calmus (2010), Arizpe sub-basin: A sedimentary
and volcanic record of Basin and Range extension in north-central Sonora, Mexico, Rev. Mex. Cienc. Geol., 27, 292–312.
Goodfellow, G. E. (1887), The Sonora earthquake, Science, 10(236), 81–82.
Goodfellow, G. E. (1888), The Sonora earthquake, Science, 11(270), 162–166.
Haeussler, P. J., et al. (2004), Surface rupture and slip distribution of the Denali and Totschunda faults in the 3 November 2002 M 7.9
earthquake, Alaska, Bull. Seismol. Soc. Am., 94, S23–S52.
Hecker, S., T. E. Dawson, and D. P. Schwartz (2010), Normal-faulting slip maxima and stress-drop variability: A geological perspective, Bull.
Seismol. Soc. Am., 100, 3130–3147.
International Atomic Energy Agency (2010), Seismic Hazards in Site Evaluation for Nuclear Installations, Safety Standard Series No. SSG-9,
International Atomic Energy Agency, Vienna, Austria.
©2014. American Geophysical Union. All Rights Reserved.
639
Journal of Geophysical Research: Solid Earth
10.1002/2014JB011244
Jackson, J. A. (2002), Using earthquakes for continental tectonic geology, in International Handbook of Earthquake and Engineering
Seismology, edited by W. H. K. Lee et al., part A, pp. 491–503, Academic Press, San Diego, Calif.
Keller, G. R. (2004), Geophysical constraints on the crustal structure of New Mexico, in The Geology of New Mexico, edited by G. H. Mack
and K. A. Giles, New Mexico Geol. Soc. Spec. Pub., 11, 439–456.
Keller, G. R., P. Morgan, and W. R. Seager (1990), Crustal structure, gravity anomalies and heat flow in the southern Rio Grande rift and their
relationship to extensional tectonics, Tectonophysics, 174, 21–37.
Kempton, P. D., and M. A. Dungan (1989), Geology and petrology of basalts and included mafic, ultramafic, and granulitic xenoliths of the
Geronimo volcanic field, southeastern Arizona, N. M. Bur. Mines Miner. Resour., Mem., 46, 161–185.
Klinger, Y., X. Xu, P. Tapponier, J. van der Woerd, C. Lasserre, and G. King (2005), High-resolution satellite imagery mapping of the surface rupture and
slip distribution of the MW ~7.8, 14 November 2001 Kokoxili earthquake, Kunlun fault, northern Tibet, China, Bull. Seismol. Soc. Am., 95, 1970–1987.
Kurushin, R. A., A. Bayasgalan, M. Ölziybat, B. Enhtuvshin, P. Molnar, C. Bayarsayhan, K. W. Hudnut, and J. Lin (1997), The surface rupture of the
1957 Gobi-Altay, Mongolia, earthquake, Geol. Soc. Am. Spec. Pap., 320, 143.
La Pointe, P. R., T. Cladouhos, and S. Follin (2002), Development, application, and evaluation of a methodology to estimate distributed slip on
fractures due to future earthquakes for nuclear waste repository performance assessment, Bull. Seismol. Soc. Am., 92, 923–944.
Li, Q., M. Liu, Q. Zhang, and E. Sandvol (2007), Stress evolution and seismicity in the central–eastern United States: Insights from geodynamic modeling,
in Continental Intraplate Earthquakes: Science, Hazard, and Policy Issues, edited by S. Stein and S. Mazzotti, Geol. Soc. Am. Spec. Pap., 425, 149–166.
Lin, A., and M. Nishikawa (2011), Riedel shear structures in the co-seismic surface rupture zone produced by the 2001 MW 7.8 Kunlun
earthquake, northern Tibetan Plateau, J. Struct. Geol., 33, 1302–1311.
Machette, M. N., S. F. Personius, C. M. Menges, and P. A. Pearthree (1986), Map showing Quaternary and Pliocene faults in the Silver City 1° x 2°
quadrangle and the Douglas 1° x 2° quadrangle, southeastern Arizona and southwestern New Mexico, U. S. Geol. Survey, Miscellaneous
Field Studies Map MF-1465-C, scale 1: 250,000, with pamphlet, 20 p.
Mack, G. H. (2004), Middle and late Cenozoic crustal extension, sedimentation, and volcanism in the southern Rio Grande rift, Basin and
Range, and southern Transition Zone of southwestern New Mexico, in The Geology of New Mexico, edited by G. H. Mack and K. A. Giles, New
Mexico Geol. Soc. Spec. Pub., 11, 389–406.
Manighetti, I., M. Campillo, C. Sammis, P. M. Mai, and G. King (2005), Evidence for self-similar, triangular slip distributions on earthquakes:
Implications for earthquake and fault mechanics, J. Geophys. Res., 110, B05302, doi:10.1029/2004JB003174.
McCalpin, J. P. (1995), Frequency distribution of geologically determined slip rates for normal faults in the western United States, Bull.
Seismol. Soc. Am., 85, 1867–1872.
McCalpin, J. P. (2009), Paleoseismology in extensional tectonic environments, in Paleoseismology, Int. Geophys., vol. 95, edited by
J. P. McCalpin, 2nd ed., pp., 171–269, Academic Press, Burlington, Mass.
McDowell, F. W., J. Roldán-Quintana, and R. Amaya-Martínez (1997), Interrelationship of sedimentary and volcanic deposits associated with
Tertiary extension in Sonora, Mexico, Geol. Soc. Am. Bull., 109, 1349–1360.
Natali, S. G., and M. L. Sbar (1982), Seismicity in the epicentral region of the 1887 northeastern Sonora earthquake, Mexico, Bull. Seismol. Soc.
Am., 72, 181–196.
Nordt, L. (2003), Late Quaternary fluvial landscape evolution in desert grasslands of northern Chihuahua, Mexico, Geol. Soc. Am. Bull., 115,
596–606.
Paz Moreno, F. A., A. Demant, J. J. Cochemé, J. Dostal, and R. Montigny (2003), The Quaternary Moctezuma volcanic field—A tholeiitic to alkali
basaltic episode in the central Sonoran Basin and Range province, México, in Tectonic Evolution of Northwestern México and the
Southwestern USA, edited by S. E. Johnson et al., Geol. Soc. Am. Spec. Pap., 374, 439–455.
Pearthree, P. A., and S. S. Calvo (1987), The Santa Rita fault zone: Evidence for large magnitude earthquakes with very long recurrence
intervals, Basin and Range Province of southeastern Arizona, Bull. Seismol. Soc. Am., 77, 97–116.
Pearthree, P. A., W. B. Bull, and T. C. Wallace (1990), Geomorphology and Quaternary geology of the Pitaycachi fault, northeastern Sonora,
Mexico, in Geologic Excursions Through the Sonoran Desert Region, Arizona and Sonora, edited by G. E. Gehrels and J. E. Spencer, Ariz. Geol.
Surv. Spec. Pap., 7, 124–135.
Sbar, M. L., and S. M. DuBois (1984), Attenuation of intensity for the 1887 northern Sonora, Mexico earthquake, Bull. Seismol. Soc. Am., 74,
2613–2628.
Scholz, C. H. (2002), The Mechanics of Earthquakes and Faulting, 2nd ed., Cambridge Univ. Press, Cambridge, U. K.
Scholz, C. H. (2007), Fault mechanics, in Treatise on Geophysics, edited by D. L. Turcotte, pp. 441–483, Elsevier Science B. V, Amsterdam,
Netherlands.
Scholz, C. H., and J. Contreras (1998), Mechanics of continental rift architecture, Geology, 26, 967–970.
Sharon, E., and J. Fineberg (1999), Confirming the continuum theory of dynamic brittle fracture for fast cracks, Nature, 397, 333–335.
Sharon, E., S. P. Gross, and J. Fineberg (1995), Local crack branching as a mechanism for instability in dynamic fracture, Phys. Rev. Lett., 74,
5096–5099.
Shaw, B. E. (2006), Initiation propagation and termination of elastodynamic ruptures associated with segmentation of faults and shaking
hazard, J. Geophys. Res., 111, B08302, doi:10.1029/2005JB004093.
Stein, S., and M. Liu (2009), Long aftershock sequences within continents and implications for earthquake hazard assessment, Nature, 462,
87–89.
Sumner, J. R. (1977), The Sonora earthquake of 1887, Bull. Seismol. Soc. Am., 67, 1219–1223.
Suter, M. (1987), Orientational data on the state of stress in northeastern Mexico as inferred from stress-induced borehole elongations,
J. Geophys. Res., 92, 2617–2626, doi:10.1029/JB092iB03p02617.
Suter, M. (1991), State of stress and active deformation in Mexico and western Central America, in Neotectonics of North America, edited by
D. B. Slemmons et al., Geol. Soc. Am., 1, 401–421.
Suter, M. (2001), The historical seismicity of northeastern Sonora and northwestern Chihuahua, Mexico (28–32°N, 106–111°W), J. South Am.
Earth Sci., 14, 521–532.
Suter, M. (2006), Contemporary studies of the 3 May 1887 MW 7.5 Sonora, Mexico (Basin and Range Province) earthquake, Seismol. Res. Lett.,
77, 134–147.
Suter, M. (2008a), Structural configuration of the Teras fault (southern Basin-and-Range Province) and its rupture in the 3 May 1887 MW 7.5
Sonora, Mexico earthquake, Rev. Mex. Cienc. Geol., 25, 179–195.
Suter, M. (2008b), Structural configuration of the Otates fault (southern Basin-and-Range Province) and its rupture in the 3 May 1887 MW 7.5
Sonora, Mexico earthquake, Bull. Seismol. Soc. Am., 98, 2879–2893.
Suter, M., and J. Contreras (2002), Active tectonics of northeastern Sonora, Mexico (southern Basin and Range Province) and the 3 May 1887
MW = 7.4 earthquake, Bull. Seismol. Soc. Am., 92, 581–589.
SUTER
©2014. American Geophysical Union. All Rights Reserved.
640
Journal of Geophysical Research: Solid Earth
10.1002/2014JB011244
Tchalenko, J. S., and N. N. Ambraseys (1970), Structural analysis of the Dasht-eBayaz (Iran) earthquake fractures, Geol. Soc. Am. Bull., 81, 41–60.
Treiman, J. A., K. J. Kendrick, W. A. Bryant, T. K. Rockwell, and S. F. McGill (2002), Primary surface rupture associated with the MW 7.1 16
October 1999 Hector Mine earthquake, San Bernardino County, Calif., Bull. Seismol. Soc. Am., 92, 1171–1191.
U. S. Nuclear Regulatory Commission (2012), Practical Implementation Guidelines for SSHAC Level 3 and 4 Hazard Studies, NUREG-2117 U. S.
Nuclear Regulatory Commission, Washington, D. C.
Wald, D. J., and T. H. Heaton (1994), Spatial and temporal distribution of slip for the 1992 Landers, California earthquake, Bull. Seismol. Soc.
Am., 84, 668–691.
Wallace, R. E. (1984), Fault scarps formed during the earthquakes of October 2 1915 in Pleasant Valley, Nevada, and some tectonic
implications, U. S. Geol. Surv. Prof. Pap. 1274-A, 1–33.
Wallace, T. C., and P. A. Pearthree (1989), Recent earthquakes in northern Sonora, Ariz. Geol., 19(3), 6–7.
Wells, D. L., and K. J. Coppersmith (1994), New empirical relationships among magnitude, rupture length, rupture width, rupture area, and
surface displacement, Bull. Seismol. Soc. Am., 84, 974–1002.
Wesnousky, S. G. (2008), Displacement and geometrical characteristics of earthquake surface ruptures: Issues and implications for seismic-hazard
analysis and the process of earthquake rupture, Bull. Seismol. Soc. Am., 98, 1609–1632.
Xiao, H., and J. Suppe (1992), Origin of rollover, AAPG Bull., 76, 509–529.
Yeats, R. S., K. Sieh, and C. R. Allen (1997), The Geology of Earthquakes, Oxford Univ. Press, New York.
Youngs, R. R., et al. (2003), A Methodology for Probabilistic Fault Displacement Hazard Analysis (PFDHA), Earthquake Spectra, 19, 191–219,
doi:10.1193/1.1542891.
Zhang, P., F. Mao, and D. B. Slemmons (1999), Rupture terminations and size of segment boundaries from historical earthquake ruptures in
the Basin and Range Province, Tectonophysics, 308, 37–52.
SUTER
©2014. American Geophysical Union. All Rights Reserved.
641