Abstract

The Laguna Salada Basin in northeastern Baja California, México, is an active half-graben with subsidence principally controlled by two major faults along the eastern basin margin—the Cañada David detachment fault and the dextral oblique Laguna Salada fault. Active-source, seismic-reflection data constrain the geometry of the active detachment fault and indicate two structural domains. The south domain is a supradetachment basin controlled by the Cañada David detachment fault. Two seismic profiles indicate the detachment fault dips 17°–20° west, has a minimum of 10.1 km of slip, and accumulates a sedimentary wedge more than 2.5 km thick in the west-central part of this basin domain. This estimation indicates that the subsurface portion of the Cañada David detachment accommodates 24% of extension in the western main plate boundary zone. The north domain is a dilatational stepover (or pull apart) controlled by the northwest-trending, west-dipping, dextral-oblique Laguna Salada fault and the north-trending, dip-slip Cañón Rojo fault, which defines the south boundary of the pull-apart basin domain. The Cañón Rojo fault accumulates more than 2 km of subsidence, but geometric considerations indicate that the basement in the hanging wall of the Laguna Salada fault projects to a depth of ∼3.8 km and intersects the 70° west-dipping Laguna Salada fault. Several faults cut the west margin of the floodplain lagoon and the hanging wall of both the Cañada David detachment and the Laguna Salada fault. The largest fault is west dipping and produces ∼500 m of vertical offset. Its location projects south of the Cañón Rojo fault, and we speculate these two faults may correlate. Seismic facies reflect its sedimentary environment and processes. Seismic facies 1 is high-amplitude, laterally continuous reflectors that represent flooding and prolonged lacustrine conditions. Seismic facies 2 is high- to low-amplitude, laterally discontinuous reflectors also representing flooding conditions. Seismic facies 3 is low-amplitude, poorly contrasted continuous to discontinuous reflectors interpreted as subaerial distal fan sandstone deposits. Seismic facies 4 is high-amplitude, discontinuous, imbricated to a chaotic pattern of reflectors. We interpret facies 4 as coarse-grained, high-energy alluvial fan deposits prograding over the basin floor from the west in the range front of Sierra Juarez. Seismic facies 1 and 2 predominate in the east and central portions of seismic profiles where the depocenter accumulates thicker sedimentary sequences.

INTRODUCTION

The Laguna Salada Basin (LSB) in northeastern Baja California, México, is an ∼20-km-wide, ∼100-km-long tectonic depression at the northwestern side conterminous with the Gulf of California rift system (Fig. 1). The LSB is structurally separated from the Salton Trough in southern California by the northern extension of the Laguna Salada fault (LSF), which splits in both left-stepping and right-stepping shear strands (Isaacs, 1987) and produces basement ridges up to ∼660 m (Fig. 1). Southward the flat topography of the basin contrasts with the steep relief of bounding ranges of Sierra Juárez to the west and Sierra Cucapah and El Mayor to the east (Fig. 1). Southward the modern basin becomes narrower and connects through a ∼5–10-km-wide inlet with the modern delta plain of the Colorado River and the tidal flats of northern Gulf of California (Fig. 1). Seasonal flooding of the Colorado River inundates the Laguna Salada Basin and produces intermittent estuarine conditions now rarely observed due to dams in the upper Colorado River (Cohen and Heges-Jeck, 2001).

On the basis of Quaternary fault scarp along the Laguna Salada fault and gravimetric and magnetic surveys of Kelm (1972), Mueller and Rockwell (1991) interpreted LSB as a “pull-apart” basin controlled by the NW-oriented, dextral-oblique Laguna Salada fault. However, structural studies in Sierra El Mayor (Siem and Gastil, 1994) and the Sierra Juárez range front (Romero-Espejel, 1997) introduced the concept of rift segmentation controlled by active low-angle normal faults and coeval strike-slip faults (Axen, 1995; Axen and Fletcher, 1998; Axen et al., 1999). García-Abdeslem et al. (2001) interpreted a strong positive gradient of the Bouguer gravimetric anomaly along the eastern margin as related to a major structural boundary of crustal rocks with density contrasts caused by the dextral oblique Laguna Salada fault and the Chupamirtos dextral oblique fault (Fig. 2). Two-dimensional (2D) modeling of gravity data (García-Abdeslem et al., 2001; Martín-Atienza 2001; Cortés-Arroyo, 2011) and stratigraphic studies in the Cerro Colorado basin and in three exploratory wells of Comisión Federal de Electricidad (CFE) (Vázquez-Hernández et al., 1996; Dorsey and Martín-Barajas, 1999; Martín-Barajas et al., 2001) are consistent with a northwest-trending, strongly asymmetric depocenter with a maximum sedimentary fill of ∼3 km adjacent to the Laguna Salada fault in the northwest (Fig. 2).

Industry seismic data collected by Petróleos Mexicanos (PEMEX) during the early 1980s and three exploratory wells drilled by CFE provide a unique opportunity to further investigate the structure and stratigraphy of LSB, particularly the geometry of an active detachment fault in depth (Axen et al., 1999; Fletcher and Spelz, 2009). Although only a few seismic lines were collected by PEMEX in the Laguna Salada Basin (Figs. 1 and 2), the available seismic data provide important geometric constraints for the architecture and evolution of this active supradetachment basin (Dorsey and Martín-Barajas, 1999).

Our study comprises the processing and interpretation of ∼115 km of seismic-reflection profiles (Figs. 1 and 2) and the correlation of the seismic sequences with stratigraphy described in three exploratory wells of CFE (Fig. 3). We first present the principal structural and stratigraphic characteristics to gain insight about the architecture and the amount of subsidence and extension principally in the southern LSB. We then discuss the distribution of seismic facies as related to distinct depositional environments within the basin.

TECTONIC AND STRUCTURAL BACKGROUND

Receiver function data (Lewis et al., 2001) and gravity data modeling (García-Abdeslem et al., 2001) suggest that crust-mantle interface for the Laguna Salada region is at ∼25 km depth. This contrasts with the ∼35–42 km crustal thickness beneath the peninsular ranges (e.g., Sierra Juárez) (Lewis et al., 2001). However, a large negative Bouguer gravity anomaly straddles the main gulf escarpment and the central part of the LSB. This suggests either a basin depocenter above homogeneous lithology of crustal rocks or a deep crustal root that flexurally supports topography (García-Abdeslem et al., 2001).

Historical seismicity (Ellsworth, 1990; Doser, 1994; García-Abdeslem et al., 2001; Hough and Elliot, 2004) and paleoseismological studies (Mueller and Rockwell, 1991, 1995; Fletcher and Spelz, 2009; Fletcher et al., 2014) demonstrate that LSB is controlled by the active LSF and the Cañada David detachment fault. These two major faults interact with other faults and accommodate transtensional shearing in the easternmost plate boundary zone (Fletcher et al., 2016). At this latitude, the active plate boundary zone is 33 km wide between the Imperial and the Laguna Salada faults and is 26 km wide between the Cerro Prieto fault and the Cañada David detachment (Fig. 1). Several intervening faults accommodate the transtensional strain.

The historical seismicity in the LSB is low in the NW sector (Fig. 1), and only two major earthquakes have been previously located within the eastern basin margin—the 1892, Mw > 7, and the 1934, Mw 6.5 (Ellsworth, 1990) (Fig. 1). Uncertainty remains regarding the epicentral location and the magnitude of the earthquake that occurred February 23, 1892, because its location was calculated based on the intensity distribution in southern California (Strand, 1980). Nevertheless, this earthquake originated due to a rupture in the LSF as indicated by fault scarps that reached up to 3–4 m of vertical scrolling and suggested a magnitude of 7.1 (Mueller and Rockwell, 1995). Furthermore, Hough and Elliot (2004) reported a magnitude of 7.2 using a method based on the distance decay of modified Mercalli intensity (MMI) values for earthquakes in western North America (Fig. 1).

Microearthquake and tectonic studies within the LSB region were completed in two surveys (Fig. 1) (García-Abdeslem et al., 2001). The first survey, which was carried out in 1991, was with five seismic stations located in the northern part of the basin recording from July 8 to December 17. In the second survey, these five stations were placed south, recording from April 27 to September 17, 1992. The seismic networks registered ∼582 microearthquakes with magnitudes lower than 3.6 chiefly concentrated along the LSF and the mountain front of Sierra Juárez toward the southeast (Fig. 1). The seismic activity reported by the seismological network of the northwestern México (RESNOM-CICESE) before the El Mayor–Cucapah earthquake (EMC) of April 10, 2010 (Mw 7.2) (Hauksson et al., 2010) indicates larger seismic activity in the study area concentrated along the western basin margin, after the seismic activity focused mainly along the eastern basin and continues in this sector (December, 2015).

The EMC earthquake had little influence in the subsidence of LSB but revealed the existence of a previously unidentified fault system in the southwest part of the delta of the Colorado River, west of the Cerro Prieto fault, which was considered the main plate boundary (Figs. 1 and 2). This earthquake produced a complex rupture that involved multiple major faults shearing the crustal block of Sierra Cucapah (Fletcher et al., 2014). The rupture propagated north along a complex dextral-oblique fault zone parallel to Laguna Salada fault through the Sierra Cucapah (Fletcher et al., 2014; Terán et al., 2015). Southeast, this earthquake and its aftershocks ruptured previously unidentified faults in the delta plain south of the epicenter and evidenced a wider plate boundary zone at this latitude (Hauksson et al., 2011; Chanes-Martínez et al., 2014).

GEOLOGICAL EVOLUTION OF THE LAGUNA SALADA BASIN

Early extension and subsidence in the LSB likely started in Late Miocene as indicated by faulted ca. 10.5 Ma volcanic deposits across the southern range front of Sierra Juárez (Mendoza-Borunda et al., 1998). Crustal extension in the Laguna Salada segment is principally accommodated by the low-angle Cañada David detachment fault (CDD) synchronously with the Laguna Salada strike-slip fault system (Siem and Gastil, 1994; Axen, 1995; Axen and Fletcher, 1998). The lowermost stratigraphic unit in Laguna Salada crops out in the footwall block of the Cañon Rojo fault, a N-S–trending, high-angle normal fault forming a dilatational stepover along the Laguna Salada fault (Fig. 2). The stratigraphic sequence exposed in the footwall block of the Cañon Rojo fault constitutes the Cerro Colorado basin (Dorsey and Martín-Barajas, 1999) (Fig. 2). The late Neogene sedimentary sequence overlies in fault contact Paleozoic and Mesozoic metamorphic and granitic intrusives that form the footwall block of the Cañada David detachment (Siem and Gastil, 1994). The lower unit is early Pliocene silty-clayey yellow-green marine mudstone (Siem and Gastil, 1994; Vázquez-Hernández et al., 1996). This unit also includes metric to sub-metric evaporite deposits and locally derived conglomerate and breccia. Overall, the lower unit in the Cerro Colorado basin has similar lithological and chronostratigraphic characteristics to units of the Imperial Group in the southwestern Salton Trough (Winker and Kidwell, 1996; Dorsey et al., 2011). Basin-margin alluvial and marine conglomerate and breccia locally crop out in the north and northwestern foothills of Monte Blanco dome (Siem and Gastil, 1994; Vázquez-Hernández et al., 1996).

The arrival of the Colorado River into the rift depression in the Lower Pliocene dramatically increased the sediment supply and deltaic progradation into early marine basins (Martín-Barajas et al., 2001; Pacheco et al., 2006; Helenes et al., 2009; Dorsey et al., 2011). The Imperial deposits in the Cerro Colorado basin grade upwards into reddish, quartzose non-marine siltstone-sandstone deposits of the Palm Spring Group (Vázquez-Hernández et al., 1996; Winker and Kidwell, 1996). During the Pliocene, uplifting of the mountain ranges of Sierra Cucapah up to 700 m progressively isolated Laguna Salada from the delta plain and from the southwestern Salton depression (Axen et al., 2000; Martín-Barajas et al., 2001) (Figs. 1 and 2). This new structural configuration progressively formed the semi-closed basin with an entrance from the south end connecting LSB with the delta plain and tidal flats of the northern Gulf of California. The modern lake basin is bordered by a 5–15-km-wide belt of coalescing alluvial fans derived from Sierra Juárez in the west and by a narrower 0.5–3-km-wide belt of alluvial fans fed from the Sierra Cucapah and Sierra El Mayor in the east. These two crystalline blocks are composed of Late Cretaceous granitic rocks and pre-Cretaceous high-grade metamorphic rocks (Barnard, 1968; Gastil et al., 1974; Siem and Gastil, 1994; Axen et al., 2000). Tertiary volcanic rocks locally overlay the crystalline basement at the edges of both of these ranges, principally in the southeast in Sierra Las Tinajas (Fig. 2). This implies that pre-rift Miocene volcanic rocks may be present in depth within the LSB. In summary, the basin fill is predominantly composed of marine to deltaic fine-grained sediments funneled by the Colorado River into the northern Gulf of California and delta plain. Upward sandy deltaic deposits progressively alternate with locally derived, coarse-grained sandstone and conglomerate from local source alluvial fans from Sierra Juárez and Sierra Cucapah and El Mayor. Up to ∼700 m of Late Pliocene to present lacustrine-estuarine deposits interfingered by progradational and retrogradational alluvial wedges represent the modern structural and sedimentary setting in LSB (Figs. 1 and 2).

MAJOR FAULTS IN THE LAGUNA SALADA BASIN

The Laguna Salada fault (LSF) and the low-angle Cañada David detachment fault (CDD) control the basin architecture and subsidence and represent two distinct basin domains (Siem and Gastil, 1994; Axen, 1995; Mueller and Rockwell, 1995; Axen et al., 1999; Spelz et al., 2010). The northern domain is controlled by dextral-oblique LSF trends ∼N45°W and dips 60°–75° to the SW. The Cañada David detachment fault controls the south basin domain. Its low-angle (<20°) and Quaternary fault scarps display a curvilinear trace extending ∼55–60 km along the western mountain front of Sierra El Mayor (Fig. 2) (Siem and Gastil, 1994; Fletcher and Spelz, 2009). Geometric analyses of the fault scarps along the eastern basin margin suggest that the CDD acquires a high angle within 10–16 km away from the Sierra Cucapah–El Mayor, and the anti-listric geometry defines the location of the depocenter (Fletcher and Spelz, 2009). The LSF and the CDD fault are kinematically linked by the Cañon Rojo and Chupamirtos faults (Mueller and Rockwell, 1991) (Fig. 2) forming a releasing stepover. The Cañon Rojo fault is responsible for the abandonment of the northern synformal megamullion of the CDD and defines the position of the modern range front (Mueller and Rockwell, 1995; Dorsey and Martín-Barajas, 1999; Fletcher and Spelz, 2009).

The main gulf escarpment in Sierra Juárez contains a northwest-trending west- and east-dipping normal faults array, each with relatively small offset (Romero-Espejel, 1997; Mendoza-Borunda et al., 1998). The Sierra Juarez fault zone is ∼30 km long and ∼5 km wide and produces a vertical relief of >1000 m in northern Sierra Juárez, although two E-dipping faults accommodate ∼700 m of the vertical relief and are likely antithetic faults of the west-directed detachment fault (Axen and Fletcher, 1998) (Fig. 2). Clusters of microseismicity located in the south escarpment of Sierra Juárez and Laguna Salada are likely related to the Sierra Juárez fault zone.

SEISMIC-REFLECTION DATA ANALYSIS

In the present study, we have processed and interpreted multichannel two-dimensional (2D) seismic-reflection data collected by PEMEX in Laguna Salada Basin from the “Delta del Colorado” prospect. The seismic source for the acquisition of the seismic data was dynamite. A pattern of 1150–300–0-300–1150 m of shots was recorded on 48 channels; each receiver was spaced every 50 m; record time was 6 seconds; and sample interval was 2 ms.

Seismic processing at CICESE included: (1) edition of traces, (2) assignment of geometry, (3) correction of static due to elevation, (4) direct wave attenuation, (5) ground roll attenuation, (6) deconvolution, (7) frequency-wavenumber (FK) filter, (8) order of traces by common depth point (CDP), (9) velocity analysis, (10) normal moveout (NMO) correction, (11) stacking, (12) spherical divergence, (13) time-variable filter, (14) automatic gain control (AGC), (15) migration, and (16) depth conversion. Subsequently, the data were interpreted using the technique of Badley (1985). Processing and interpretation of the seismic data make use of the ProMax anpd SeisWorks software of Landmark™ and OpendTectTM.

The analysis and interpretation were conducted in five seismic-reflection profiles (Figs. 1 and 2). Profiles 4973, 4965, 4957, and 4949 have lengths of 14, 11, 9, and 7 km, respectively, and cross the LSB in a northeast to southwest direction (Figs. 1 and 2). Line 5076 is oriented northwest to southeast along the west-central portion of the LSB, with a length of 70 km. Profile 5076 comprises three segments (a, b, and c) with lengths of 27 km, 30 km, and 13 km, respectively, and crosses all four transversal profiles (Figs. 1 and 2). The seismic transversal profiles are separated ∼20 km from each other and distributed throughout the basin from north to south (Figs. 1 and 2).

Seismic lines were migrated in time and converted to depth using a stacking velocity model (see Supplemental Figures1). The seismic resolution in the upper part of seismic lines (e.g., <300 m) is of lower quality because industrial interest in the depth structure commonly filters the high-frequency signal during acquisition. Wells ELS-2 and ELS-3 reached the crystalline basement at 1.5 km and 0.75 km, respectively, whereas well ELS-1 located near the LSF cut the 2.4 km of sediments and did not reach the basement (Fig. 3). The stratigraphy in ELS wells provided stratigraphic and seismic velocity constraints (Álvarez-Rosales and González-López, 1995; Martín-Barajas et al., 2001). Well ELS-2 is closer to intersection of seismic lines 4957 and 5076-a, and wells LS-1 and LS-3 are 10–15 km away from the nearest seismic line and provide indirect lithological and stratigraphic constraints (Fig. 3).

RESULTS

Acoustic Basement and Basin Configuration

The acoustic basement is a distinctive, laterally continuous, high-amplitude reflector in most profiles, except in sectors where the reflectors are rather chaotic and not distinguished from seismic noise (cf. south of profile 5076-b). The two southernmost transversal profiles (4973 and 4965) clearly show the low-angle fault that controls the basement ramp and the wedge-shaped sedimentary basin fill in the hanging wall of the CDD (Fig. 4). The acoustic basement in the hanging wall deepens to a maximum of ∼2100 m in profile 4973 and to ∼2500 m in profile 4965 (Figs. 4A and 4B). No coherent reflectors are observed beneath the low-angle fault plane in the footwall block.

The acoustic basement beneath the hanging wall is disrupted by faults with tents to a few hundred meters of offset. The most obvious faults are observed in profile 4973 to form a ∼2-km-wide sag in the acoustic basement (Fig. 4A). This sag is likely controlled by a west-dipping fault and at least two east-dipping antithetic faults. Another prominent relief in the acoustic basement occurs at the northwest end of profile 5076-b (Fig. 5A). There, a major fault produces a vertical offset of more than 500 m along a horizontal distance of ∼2 km (cdp ∼6550–6750). Near the south end of profile 5076-a, a series of faults with small vertical displacement disrupt the acoustic basement (Fig. 5B). The depth of basement in well ELS-2 coincides with the depth of basement in profile 5076-b (Fig. 5A) and likely maintains a similar depth as in line 5076-a (Fig. 5A). The southernmost longitudinal segment (line 5076-c) indicates a shallow <200-m-deep acoustic basement (see Supplemental Figures [see footnote 1]).

The acoustic basement has a prominent vertical offset in the southern part of profile 5076-b (Fig. 5B) where the basement is clearly imaged at ∼1700 m near cdp 6250. The vertical difference in depth is depicted at the south end of this profile, where coherent reflectors indicate basement at ∼600–700 m (Fig. 5A). Between the two ends, basement loses its distinctive high- amplitude characteristic and passes southward into a zone of chaotic reflectors (from cdp 5750–6150). Furthermore, the change in depth to basement in profile 5076-b is confirmed at the west end of line 4973 where it crosses the longitudinal seismic line 5076-b at its south end. Farther south, line 5076-c confirms that crystalline basement is very shallow in the east flank of Sierra Las Tinajas. The ∼1000 m vertical difference of depth to basement in the south end of profile 5076-b occurs at a distance of ∼10 km (Fig. 5A), but the seismic image lacks the resolution to interpret any fault that may control this basement relief. For the south basin domain, we interpret that basement is 2.5 km deep in a depocenter located 7 km west of the breakaway fault on the detachment west of Sierra El Mayor.

The depth to basement at intersection of lines 5076-a (Fig. 5A) and 4957 (Fig. 6A) agrees with well ELS-2, where the granitic basement was cut at 1590 m deep (Fig. 3). Near this intersection, line 5076-a presents a series of parallel, high-amplitude reflectors at depths from 1000 to 1300 m that we interpret as due to the lithological contrast in sediments (Fig. 5A). Below 1300 m, a ∼200-m-thick interval of low-amplitude reflectors named here the “white unit,” overlies chaotic reflectors that we interpret as the basal conglomerate unit and granitic basement reported in well ELS-2 (see Fig. 3). The acoustic basement in the crossing of line 4957 (Fig. 6A) matches the depth to crystalline basement of well ELS-2 and supports this interpretation. Continuous high-amplitude reflections at ∼1 km deep are observed in both seismic lines. These seismic reflectors are cut by at least three faults dipping northwest with small vertical offset in line 5076-a; elsewhere, these reflectors are undisrupted and laterally continuous across most of the seismic image (Fig. 5B). The white unit pinches out against the acoustic basement to the northwest.

The two northern transversal profiles 4957 and 4949 (Figs. 6A and 6B) indicate deepening of acoustic basement toward the east; from ∼250 m to more than 1200 m in profile 4949, and to >1600 m at the eastern end of profile 4957. Profile 4957 captures only half basin width, but depth to basement is more than 2.4 km farther east as indicated in well ELS-1 (Fig. 3), and thus basement is likely deeper in the north basin domain near the Laguna Salada fault.

Stratigraphy and Seismic Facies

The stratigraphic units in seismic lines in Laguna Salada are interpreted on the basis of seismic facies and stratigraphic sequences limited by sequence boundaries. Although seismic lines are medium to poor quality, lines 4973, 4965, and 4957 offer the possibility to interpret sedimentary sequences below 300 m depth (Figs. 4 and 6). We hereby distinguish four seismic facies. Facies 1 is characterized by a pattern of parallel high-amplitude, laterally continuous reflections. Facies 2 is medium- to low-amplitude, laterally continuous imbricated to subparallel reflections. Facies 3 is low-amplitude, discontinuous wavy reflections (e.g., white intervals), and facies 4 is defined as discontinuous, high- to low-amplitude, imbricated to chaotic pattern of reflectors. A description of facies is shown in the Supplemental Figures (see footnote 1).

We observe a systematic lateral facies change across the three transversal profiles (Figs. 4 and 6). Basin-wide continuous seismic reflections of facies 1 distinctively represent stratigraphic sequence boundaries in the LSB. In profile 4965 (Fig. 4B), we interpret five stratigraphic sequences with basal boundaries defined by these continuous reflectors of facies 1. Stratigraphic sequences include intervals of facies 2 and 3 above sequence boundaries defined by intervals of continuous reflectors of facies 1. Horizon 1 is the first laterally continuous reflector across the basin. Intervals of low to medium amplitude laterally wedge out westward and interfinger with chaotic reflectors of facies 4. Unit 1 includes upwards a thick interval of laterally long and continuous reflectors (sequence 1). Shorter, high-amplitude continuous reflectors that lap on the west distinctively form the lower part of this interval and form local angular unconformities within a smaller (2–3-km-wide) depocenter (Fig. 4B). Upwards, facies 1 is laterally continuous both east and west and expands over a broader depocenter.

An eastward migration of the depocenter is depicted upward in profile 4965 (Fig. 4B). Up section, the concave shape of reflectors shift toward the east, and sedimentary sequences thicken in the east and central parts of each depocenter. Depocenters define lens-shaped deposits that laterally wedge out and terminate in onlap against the continuous reflectors of facies 1 below. A new continuous reflector of facies 1 covers the lens-shaped deposit and defines a new sequence. However, the low resolution and low number of seismic lines prevent a detailed interpretation and correlation of most sequence boundaries. Nevertheless, an important observation is the eastward thickening of the two lowermost sequences adjacent to CDD in profile 4965, whereas the three upper sequences are symmetric lenses, and the thicker intervals are located 6–8 km west of the breakaway fault of the detachment (Figs. 4A and 4B).

A 200–300-m-thick interval of low-amplitude reflectors (cf. facies 3) is located above the crystalline basement in profile 5076-a (Fig. 5A). The southeast end of this profile presents a series of high-amplitude and continuous reflectors (facies 1) above the ∼200-m-thick “white unit” interval below. The acoustic basement underlies the white unit at ∼1600 m as defined in well ELS-2 (Fig. 5). Furthermore, the “white unit” is wedge shaped in seismic line 4957 and pinches out toward the northwest.

Facies 4 predominates west where the basement is shallower and closer to the Sierra Juárez range front. Chaotic and diffuse reflectors characterize facies 4, which laterally passes into subparallel and continuous, high-amplitude reflectors (facies 1 and 2). In seismic lines 4973 and 4965 (Fig. 4), facies 4 is progradational eastward and is absent or poorly expressed in the eastern side of transversal profiles, where facies 1 and 2 dominate. An independent evidence of eastward progradation of facies 4 is depicted in the northwestern half of longitudinal profile 5076-b (Fig. 5A). High-amplitude continuous to discontinuous reflectors (facies 1 and 2) above basement alternate at intervals tens to a few hundred meters thick. Above ∼1000 m deep, seismic facies of type 4 predominate, as well as in most of the south part of this profile (Fig. 5B).

DISCUSSION

Geometry of the Detachment Fault at Depth

The most important result is the direct evidence of the Cañada David detachment fault beneath a 2–2.5-km-thick sedimentary wedge in the south domain of LSB. Here the thickest basin fill corresponds to the site where the acoustic basement in the hanging wall intersects the acoustic basement in the footwall block. This intersection represents the minimum amount of subsidence in LSB controlled by the detachment fault. We also estimate the minimum displacement along the fault plane and its vertical and horizontal components (Fig. 7). From profile 4965 (Fig. 4B), the fault plane was projected to the surface (dotted red line) up to the height of 285 m above sea level, which is the elevation of the lower ridge in the western flank of Sierra El Mayor and corresponds to the upward projection of the CDD with an angle of dip of 160 (Figs. 4B and 7). In this calculation, we do not consider the maximum height of the mountain range to the east (∼700 m) or erosion in the footwall block of the detachment. The minimum displacement along the fault plane is ∼10.1 km, and the minimum horizontal displacement is ∼9.7 km. The vertical component (e.g., subsidence) of this geometric analysis is of ∼2.8 km. The 9.7 km of extension (e.g., horizontal displacement) represents 53% of the basin width from the range front of Sierra Juárez in the west to the range front of Sierra El Mayor in the east and represents 24% of extension across the 40-km-wide zone of extension from Sierra El Mayor (285 m above mean sea level [amsl]) to the summit of Sierra Juárez (1596 m amsl). The ∼10.1 km minimum displacement along the buried fault plane of CDD can be added to 14–18 km of extension reported in the lower plate of the CDD across Sierra El Mayor and Monte Blanco dome, respectively (Axen and Fletcher, 1998). This yields nearly 25–29 km of displacement in the CDD.

The basal nonconformity of sediments over the acoustic basement observed in the seismic lines constitutes, up to now, the most reliable piercing point to estimate the minimum amount of extension and the ∼2.8 km of subsidence controlled by the CDD in the south domain of Laguna Salada Basin.

A distinctive feature in the two seismic images of the CDD at depth is the eastward shift of the depocenter through time. Interestingly, the two lower sequences are wedge shaped with a maximum thickness adjacent to the fault plane (Figs. 4A and 4B). The upper three units are quasi-symmetric in shape and thicken in the central synform. We interpret that depocenters in the two lower sequences developed closer to the fault plane likely with a higher fault dip. As the detachment fault becomes low angle, the horizontal component increases and displaces the depocenter basinward as proposed by Fletcher and Spelz (2009). Furthermore, in the southernmost seismic image (profile 4973, Fig. 4B), the CDD has an anti-listric shape as proposed by Fletcher and Spelz (2008) as inherent to the development of a rolling hinge during footwall uplift. Profile 4973 (Fig. 4B) shows an apparent steeper angle of the CDD, but the calculation of the fault dip using the same procedure as in profile 4965 yields 17° for the CDD, which is similar to the 15° dip angle of detachment in profile 4965 (Fig. 4A).

The microseismic activity, according to García-Abdeslem et al. (2001), is located along the Laguna Salada fault, Cerro Prieto geothermal field, and the eastern front of Sierra Juárez, and only 17 events lie within the Laguna Salada Basin. These last events are concentrated in the northwestern basin domain, and no correlation with faults antithetic to the CDD is observed.

In profile 4973 (Fig. 4B), the lowermost stratigraphic unit is lenticular, and seismic reflectors are parallel to the acoustic basement. The top of this lower unit is an erosional unconformity underlying a narrow basin depocenter. Above the lower sequence boundary, the sedimentary sequences have a quasi-symmetric synform shape (cf. from ∼1300–2000 m in profile 4973). This depocenter broadens upwards and probably represents a broader zone of subsidence and/or an increase in sediment supply. We interpret that sequence 1 in profile 4973 (Fig. 4B) was partially eroded by inflow along the estuarine channel. Sequences 3, 4, and 5 maintain their thickness across the seismic profile 4973, and seismic reflections gently dip toward the east, whereas sequence 2 is wedge shaped and nearly 1000-m-thick sediments juxtapose the CDD.

The CDD likely includes synthetic and antithetic faults that merge at depth into the master fault. The two deeper synthetic faults are likely inactive and do not propagate upwards (Fig. 4), and they do not offset a thick interval of high-amplitude continuous reflectors observed at ∼600 m deep. Above ∼400 m, the poor resolution prevents further seismic interpretation, and the activity of faults located farther east in the sedimentary wedge is not imaged (cf. Fletcher and Spelz, 2009).

The north domain is controlled by the high-angle, dextral oblique Laguna Salada fault and the high-angle, dip-slip Cañón Rojo fault. Both control the modern depocenter and subsidence in the northern half of the basin. The Laguna Salada fault is high angle (∼60° to 70°) and forms a releasing stepover in the Cañón Rojo fault (Fig. 8). Mueller and Rockwell (1991) proposed that the Cañón Rojo fault is a dilatation stepover in a pull apart bounded by the dextral oblique Chupamirtos fault. The Chupamirtos fault apparently constitutes the hard link between the Laguna Salada and the active portion of CDD. The Chupamirtos fault bounds the Cerro Colorado basin along the west-southwest and likely intersects the CDD north of seismic profile 4965 (Fig. 8). The Chupamirtos fault likely represents the structural boundary between two basin domains. South of the Chupamirtos fault, Laguna Salada Basin is an active supradetachment basin, whereas north of the Chupamirtos fault, the Cañón Rojo stepover produces a ∼10-km-wide pull-apart basin laterally controlled by the Laguna Salada and Chupamirtos faults.

Profile 5076-b (Fig. 5B) contains a west-dipping fault that produces ∼500 m vertical offset of the acoustic basement in the hanging wall of the CDD. Although fault orientation is not defined in the seismic line, the Cañón Rojo fault projects south into the position of the largest fault in profile 5076-b (Fig. 5B). The correlation of the Cañón Rojo fault and the largest fault in profile 5076-b (Fig. 8) although speculative implies that the Cañón Rojo fault would have a smaller vertical offset to the south. Near its intersection with the Laguna Salada fault, the Cañón Rojo fault has a vertical offset of ∼1.3 km measured from the stratigraphic thickness of Plio-Pleistocene deposits, including the locally derived Red Beds, the Grey Gravel units, and the Palm Spring and Imperial deposits (Dorsey and Martín-Barajas, 1999). The stratigraphic thickness of the Cerro Colorado basin is a minimum of vertical offset in the Cañón Rojo fault. Additional ∼700-m-thick lacustrine deposits cut in well ELS-1 suggest that the vertical slip of the Cañón Rojo fault may attain ∼2000 m near its intersection with the Laguna Salada fault. Southward, the Cañón Rojo fault likely loses vertical displacement because it transfers part of the slip into the Chupamirtos fault.

The minimum depth to depocenter in the northern domain of the Laguna Salada Basin is well ELS-1, which drilled ∼2.4 km of deltaic, lacustrine-estuarine and alluvial fan sedimentary deposits (Martín-Barajas et al., 2001). This well did not reach the Imperial marine mudstone unit inferred to lie at a depth as recorded in stratigraphy of the Cerro Colorado basin (Vázquez-Hernández et al., 1996). The 2D gravity model of García-Abdeslem et al. (2001) indicates that the basin fill adjacent to the LSF is ∼3 km thick, which is a reasonable estimate of basement depth. An independent estimate of depth to basement is the eastward projection of acoustic basement in profile 4957. This geometric projection suggests that basement in the hanging wall intersects the Laguna Salada fault at ∼3.8 km below the surface, assuming that the Laguna Salada fault maintains a ∼70° dip to the west (Fig. 9). This amount of subsidence would be ∼3.5 km, if LSF dips 60° to the west. For these estimates, we infer that the acoustic basement is a flat ramp that deepens at an angle of ∼20° similar to average basement dip in profile 4957 (Fig. 6A). This calculation suggests a somewhat deeper depocenter as proposed by the 2D gravity modeling.

The contrast in structural style and the amount of subsidence among the north basin domain controlled by the dextral oblique Laguna Salada fault and the south basin domain controlled by the CDD fault requires a structural boundary likely in the Chupamirtos fault (Mueller and Rockwell, 1991). We propose that the Chupamirtos fault separates the active supradetachment basin domain in the south from the pull-apart basin domain controlled by the Cañón Rojo and Laguna Salada faults.

Several faults cut the basement along the west side of LSB (Figs. 4–6), and only the principal fault in this sector in each of these profiles is presented in Figure 8 (yellow mark). We interpret that these faults are a clear expression of several synthetic and antithetic faults cutting the hanging wall of the detachment and probably accommodating significant amounts of basin subsidence. Curiously, these faults roughly follow the west shoreline of the lake, and Figure 8 shows the direction of apparent dip in each of these faults in the seismic sections as indicated by yellow marks of apparent strike and dip.

Magnitude of Extension and Subsidence

The two seismic images of the supradetachment basin domain capture the hanging-wall basement ramp that subsided ∼2.7 km below sea level, whereas the detachment fault has accumulated a minimum of ∼10 km of finite extension. The amount of subsidence is also a minimum because mechanical compaction reduces the original porosity and sedimentary thickness and underestimates the original volume of sedimentary deposits and the amount of subsidence (Giles, 1997). Nevertheless, a crude estimate of the rates of extension and vertical subsidence suggests a ratio of 3:1, respectively. The lower sedimentary unit reported in Laguna Salada is the Imperial mudstone unit, which may correlate to either the Latrania Formation dated 5.2–6.1 Ma or to the lower part of Deguyinos Formation (5.1–4.2 Ma) (Dorsey et al., 2011) based on similar distinctive lithology and paleodepths (Vázquez-Hernández et al., 1996). Although this unit overlies crystalline basement in fault contact, and a slip in the detachment may have started synchronously with marine deposition. If we conservatively assign 7 Ma for the onset of extension, the 9.7 km of horizontal slip estimated in the CDD represents an extensional rate of ∼1.4 mm/yr. The 2.8 km of basin fill observed in seismic profile 4965 represents a minimum subsidence rate of ∼0.4 mm/yr. This is 25% of the 1.5 mm/yr calculated for the Cañón Rojo fault (Dorsey and Martín-Barajas, 1999) and sedimentation rates estimated at ∼1.5 mm/yr from spectral analysis of gamma ray log in well ELS-1 (Contreras et al., 2005). These estimates are consistent with a faster subsidence in the Laguna Salada and Cañón Rojo dilatational stepover compared to subsidence in the supradetachment basin domain controlled by the Cañada David detachment.

Mio-Pliocene detachment faults and coeval strike-slip faults occur south of Laguna Salada in Sierra San Felipe (Bryant, 1986; Seiler et al., 2010) and the Altar basin in northwestern Sonora (Pacheco et al., 2006, González-Escobar et al., 2013). Offshore, the Angel de la Guarda detachment coexisted during dextral shearing in the Tiburon and Amado dextral oblique faults that controlled the early separation of the Baja California peninsula from mainland México (Martín-Barajas et al., 2013). Elsewhere, examples of concurrent strike-slip faults and low-angle normal faults are reported in Mormon Mountains–Tule Springs Hills, Nevada (Wernicke, 1995), Panamint Valley in California (Wernicke, 1995; Numelin et al., 2007; Haines et al., 2014; among others). It seems that detachment faults and coeval strike-slip faults constitute a common and efficient way to partition oblique strain in the northern Gulf of California (Axen and Fletcher, 1998). Laguna Salada is unique among these examples because it is the only documented site of coeval active deformation. It is possible that detachment faults initiated during the early phase of transtension and produced a broader supradetachment depocenter that was subsequently overprinted by the Laguna Salada fault. The Cañón Rojo and Chupamirtos faults produced the abandonment of the Cerro Colorado synformal domain of the Cañada David detachment and reduced in ∼25% its original length (Siem and Gastil, 1994; Fletcher and Spelz, 2008). The CDD and Laguna Salada faults are, thus, a common example of coexistence of two fundamental modes of deformation and strain partition in the northern Gulf of California rift.

Seismic Facies and Sedimentary Sequences

Transversal profiles in southern Laguna Salada Basin domain show that seismic facies 4 dominate the western portion of the seismic images, and facies 1 and 2 dominate the eastern part, where depocenters define the thicker sedimentary fill. We interpret that facies 4 is produced by anastomosing channels and bars of alluvial fan deposits from Sierra Juárez (Figs. 4 and 6). Profiles 4965 (Fig. 4A) and 4957 (Fig. 6B) clearly show that laterally continuous reflectors (seismic facies 1 and 2) penetrate westward and interfinger with facies 3 and 4 produced by alluvial fan deposits. In profile 4957, the modern lakeshore is ∼7 km from the mountain front, and lacustrine facies 1 and 2 are found beneath the modern distal fan where eolian, alluvial, and lacustrine deposits interfinger. Due to their lateral continuity, facies 1 and 2 are interpreted to represent flooding and prolonged lacustrine conditions produced by the Colorado River entering Laguna Salada Basin. We speculate that the prolonged lake condition must have occurred during major sea level highstands, similar to the present time. This condition reduces the length and topographic relief for the fluvial runoff in the delta plain, and fluvial discharges reach supratidal flats and the delta front closer to the southern inlet that connects Laguna Salada and the northern Gulf of California. In contrast, during lowstands (>100 m), the delta front shifts south, and a large portion of the submarine delta (pro-delta) is exposed. Fluvial channels acquire a steeper profile, preventing Laguna Salada to retain water during prolonged periods of time. During lowstand sea level, Laguna Salada is intermittently dry, and playa-lake deposits must enhance progradation of alluvial fan and eolian deposits into the basin floor. These conditions might be similar to the modern situation in LSB produced by damming the Colorado River since the early part of the twentieth century. In 1984, flooding in the delta occurred due to the release of excess water in the river dam system. The event produced estuarine conditions in Laguna Salada for nearly five years and then dried out by ca. 1990 (Cohen and Henges-Jeck, 2001).

Two main processes likely cause the broad belt of alluvial fan deposits along the western margin. One is the strong asymmetric subsidence that maintains depocenters along the eastern margin near the Laguna and Cañón Rojo faults. The other is a higher coarse-grained sediment input due to higher topography and higher runoff in the range front of Sierra Juárez. Lacustrine deposits likely prevail for a longer time along the eastern basin margin due to higher subsidence rates and intermittent flooding of the Colorado River and locally from the Sierra Juarez mountain range. During flooding events, mostly silt and clay are transported in suspension into the lake basin and continuously accumulate over larger areas in the Laguna Salada. Climatic forcing and changes in sea level likely control the shift from estuarine conditions (flooding) to hyper-arid playa lake conditions (e.g., sea level lowstand) in Laguna Salada (Contreras et al., 2005).

The older, deeper, and narrower depocenters depicted in seismic lines 4965 and 4973 (Figs. 4A and 4B) show erosional features probably related to lateral shifts of estuarine channels during flooding. The most important erosional feature is observed in profile 4973. The lower lens shape sequence is ∼4 km wide and ∼500 m thick, and the seismic reflections are parallel to the acoustic basement (Fig. 4). Sequences 2 and 3 unconformably overlie the lower lenticular sequence and have a strong asymmetric thickness controlled by the detachment fault (Fig. 4). In both 4965 and 4973 profiles, the upper sedimentary sequences have a broader distribution and uniform thickness across the basin, although slightly thicker to the east. We interpret the eastward shift of the depocenter as related to the widening of the basin and to a lower angle in the detachment fault.

The low-amplitude wavy reflectors of seismic facies 3 are commonly observed above facies F1 to lateral interfingering on intervals of facies F2 and F1. This seismic facies implies a small lithological contrast among strata and probably represents sandstone-siltstone facies, as indicated in profile 4957 located 1.7 km to the south of well ELS-2 (Fig. 6). The white interval matches unit 3 in well ELS-2 and consists of an ∼200-m-thick sandstone that underlies a thick interval of mudstone with subordinate siltstone and sandstone (unit 4) (Martín-Barajas et al., 2001). This lithological change in well ELS-2 also corresponds to a change in seismic facies from low-amplitude, poorly contrasted seismic reflections of facies F3 to high-amplitude and continuous reflectors of facies F1. Horizon B–A likely corresponds to the boundary between units 3 and 4 in well ELS-2 (Martín-Barajas et al., 2001).

The south part of longitudinal profile 5076-b (Fig. 5B) shows nearly 11 km of chaotic reflectors (facies F4) that we also interpret as alluvial fan deposits close to the north end of Sierra Las Tinajas (Fig. 4B). Modern alluvial fans progradate and narrow the flood plain and channel in the southernmost part of the basin. However, interpretation of seismic sequences and facies distribution is limited due to low resolution and low number of seismic lines and is beyond the scope of this paper. Nevertheless, it is clear that the sedimentary record in LSB responds to both tectonic and climatic controls and constitutes an important archive yet to be explored in detail.

CONCLUSIONS

The Laguna Salada Basin is an active asymmetric depression structurally controlled by the Laguna Salada fault and the Cañada David detachment fault. The processing and interpretation of five seismic profiles indicate that these two master faults define two distinctive basin domains. The south domain is a supradetachment basin controlled by the Cañada David detachment fault. Two seismic profiles indicate the detachment fault dips 16°–20° west and has a minimum of 10.1 km of total slip. The supradetachment basin domain accumulates a sedimentary wedge more than 2.5 km thick in the west-central part of the basin, and the subsurface portion of the Cañada David detachment represents 24% of extension in the western main plate boundary zone. The north domain is a pull apart controlled by the northwest-trending, west-dipping, dextral-oblique Laguna Salada fault. The pull-apart forms a dilatational stepover with the north-south–trending, dip-slip Cañón Rojo fault, which defines the southern boundary of the pull-apart basin domain. The Cañón Rojo fault accumulates more than 2 km of subsidence, but seismic profiles in the north domain indicate the acoustic basement is an east-dipping ramp in the hanging wall of the Laguna Salada fault. Geometric considerations indicate the basement in the hanging wall projects to a depth of ∼3.8 km to 3.5 km and intersects the west-dipping Laguna Salada fault. This estimate assumes that the Laguna Salada fault maintains its surface dipping angle of 60° to 70° west, respectively.

We recognize four seismic facies representing the dominant sedimentary environments. Facies 1 and 2 are high-amplitude, laterally continuous reflectors that represent flooding and prolonged lacustrine conditions. Facies 3 is low-amplitude, poorly contrasted continuous to discontinuous reflectors interpreted as distal alluvial fan sandstone deposits, whereas facies 4 is high-amplitude, discontinuous, imbricated to a chaotic pattern of reflectors. We interpret facies 4 as high-energy, alluvial-fan coarse-grained deposits prograding over the basin floor from the west in the range front of Sierra Juarez. Seismic facies 1 and 2 predominate in the east and central portions of seismic profiles where the depocenter accumulates thick, fine-grained sedimentary sequences.

We want to thank Consejo Nacional de Ciencia y Tecnología (the National Council of Science and Technology [CONACYT]), México, for financial support, to PEMEX Exploration and Production for allowing the use of seismic data and Halliburton/Landmark, OpendTect, and Google Earth Pro for the use of their software through the University Grant Program to Centro de Investigación Científica y de Educación Superior de Ensenada (CICESE). We thank Sergio Arregui for technical support, Martín Pacheco and Ramón Mendoza-Borunda for fruitful discussion on interpretation. Constructive comments and suggestions by reviewer Dr. Gary Axen improved this manuscript.

1Supplemental Figures. Seismic profiles in two-way travel time (TWTT) and their velocity models used for depth conversion. Please visit http://dx.doi.org/10.1130/GES01261.S1 or the full-text article on www.gsapubs.org to view the Supplemental Figures.