In most extensional terrains such as the Rio Grande rift, alluvial fans and bajadas cover faults and terraces as extension progresses, thus limiting the faults and terraces as useful records of uplift. However, in the Franklin Mountains of western Texas and southern New Mexico (USA), rapid aggradation of basin floors by extensive playa lakes and floodplain deposits of the Rio Grande during the Pliocene buried the irregular mountain-front fans, thus creating a low-gradient surface. This originally planar surface was subsequently uplifted and deformed during faulting, providing a record of the Pliocene–Holocene extensional deformation in the southern Rio Grande rift. Deformation and uplift of the Franklin Mountains in the southern Rio Grande rift was estimated by measuring the elevation of late Pliocene terraces that are adjacent to range-bounding faults. The uplifted terraces are exposed along both sides of the Franklin Mountains, and lie as much as 130 m above their original elevation. Together the uplifted terraces form an anticlinal arch that mimics the profile of the range crest of the mountains. Three important conclusions can be drawn from the similarity of profiles among the terraces and mountain crest. First, the observation that the terraces mimic the range crest implies that the present-day topography of the mountains is likely tectonic in origin. Second, the east-side terraces are higher than the west-side terraces, suggesting rotation of the mountains during deformation. Estimated rotation since the Pliocene is ∼5% of the total rotation. Third, fault throw rate calculations indicate differential slip along the length of the eastern boundary fault zone. The fault profile and throw rate calculations along the eastern margin of the range are skewed to the south, suggesting that the southern segment of the Franklin Mountains has accumulated a majority of the slip during this time frame. These observations, coupled with geophysical data highlighting buried faults beneath the El Paso (Texas)–Juárez (Mexico) metropolitan region, suggest that normal faults related to uplift of the Franklin Mountains have been growing in length toward the south over the last several million years.
Normal fault systems develop through incremental slip events (earthquakes) that, over time, result in the final displacement profile of the fault. Extended terranes are also characterized by multiple normal fault systems that overlap to form relay ramps or intersect as extension progresses and individual faults grow in length (e.g., Peacock and Sanderson, 1991; Childs et al., 1995; Peacock, 2002; Nicol et al., 2010). These characteristics are true in cases of simple lithology and sandbox experiments (e.g., McClay and Ellis, 1987; McClay, 1990; Childs et al., 1995) but are also applicable to regions of diverse rock types and heterogeneous crustal features (e.g., Nicol et al., 2005). Although these relationships appear to be generally true for a wide range of extensional faults regardless of fault size, amount of offset, or lithology (e.g., Schlische et al., 1996), geologic evidence documenting sequential fault growth is commonly lacking.
In models of idealized, isolated normal faults, displacement profiles should preserve maximum displacement at the center of the fault, and displacement should decrease to zero at the fault tip lines (Barnett et al., 1987; Walsh and Watterson, 1987). However, displacement patterns are almost always segmented and irregular, and are also commonly asymmetrical, where maximum fault throw is centered closer to one fault tip than the other (e.g., Walsh and Watterson, 1987; Childs et al., 1995; Mansfield and Cartwright, 1996; Schlagenhauf et al., 2008). As slip accumulates, fault tips typically propagate in two directions and fault length increases. This process results in a power-law relationship, where fault length scales linearly with fault displacement (e.g., Pickering et al., 1995; Schlische et al., 1996). As fault length increases, adjacent faults eventually overlap and link, which has been shown to be an important fault growth mechanism in extensional settings (e.g., Cartwright et al., 1995; Finch and Gawthorpe, 2017; Whipp et al., 2017). Alternatively, some fault systems develop through a constant-fault-length model, where they approach their maximum length early and fault tips become fixed in space as fault displacement continues to accumulate (e.g., Nicol et al., 2005; Amos et al., 2010; Mouslopoulou et al., 2012; Curry et al., 2016).
The Rio Grande rift (southwestern North America) is a region of lithospheric extension along a narrow north-south zone from northern Colorado to southern New Mexico, western Texas, and northern Mexico (Fig. 1A). It remains a remarkable natural setting to investigate styles of faulting, evolution of normal faults, and fault linkage in a modest-extension environment. Differential slip along the length of normal faults, as described above, is common in extensional terranes such as the Rio Grande rift. Previous studies indicate that fault propagation and linkage may be an important process governing rift development. For example, in the Española Basin in northern New Mexico, multiple studies have been undertaken to decipher fault displacement profiles along the Pajarito fault system (Carter and Winter, 1995; Lewis et al., 2009; Goteti et al., 2013). Carter and Winter (1995) suggested that individual fault segments are generally bell shaped in the direction of dip, although they are asymmetric and displacement abruptly decreases at fault tips. Lewis et al. (2009) built upon these previous results and used the 1.25 Ma (Phillips, 2004) Tshirege Member of the Bandelier Tuff as a datum to construct a fault throw profile for the Pajarito fault system. Their results indicate that the Pajarito fault system is composed of multiple shorter segments. Along some of these fault segments, the maximum fault throw is near the center of the fault trace, while in others it is shifted to one side, producing an asymmetrical fault throw profile. In addition, each of these shorter fault segments merges and overlaps with other segments in a series of monoclinal folds and distributed small-offset faults (Lewis et al., 2009). These observations are consistent with the results of three-dimensional finite-element modeling of the evolution of the Pajarito fault system, which predicts the development of overlapping fault segments and relay ramps (Goteti et al., 2013).
To build upon these previous studies documenting sequential growth of normal faults through time in the Rio Grande rift, this paper focuses on uplifted Pliocene terraces preserved along the flanks of the Franklin Mountains, an extensional rift-flank uplift in the southern Rio Grande rift of far west Texas and south-central New Mexico. As described below, uplift of these terraces since deposition records recent movement along main rift-related normal faults. Their current elevations are therefore a valuable record of the recent evolution of these fault networks. These observations are used to compare results from the southern rift to fault profiles that have been studied to the north (Carter and Winter, 1995; Lewis et al., 2009) and to discuss fault propagation and linkage of these faults through time during continued development of the southern Rio Grande rift.
REGIONAL SETTING AND GEOLOGICAL DESCRIPTIONS
The Rio Grande rift exists as a narrow continental rift extending from northern Colorado to Socorro, New Mexico. South of Socorro, the rift abruptly widens and continues to at least southern New Mexico and western Texas, and probably into northern Mexico (e.g., Chapin and Cather, 1994) (Fig. 1A). In southern New Mexico, the rift is a wide feature and is physiographically more similar than the northern segments of the rift to the highly extended Basin and Range province to the west. Changing patterns in sedimentation, magmatism, and style of deformation all suggest that by 30–32 Ma, the area was subjected to early extension (Morgan et al., 1986), although the main phase of extension likely did not begin until ca. 25 Ma (Kelley et al., 1992; Chapin and Cather, 1994; Kelley and Chapin, 1997; Ricketts et al., 2016). This period of rapid extensional deformation lasted until ca. 10 Ma and resulted in a series of north-south–trending, linked half-grabens that alternate polarity (e.g., Mack and Seager, 1990; Chapin and Cather, 1994). Although extension has slowed since 10 Ma, paleoseismic investigations and GPS studies indicate the rift is still active today (McCalpin, 2005; McCalpin et al., 2011; Olig et al., 2011; Berglund et al., 2012).
The Franklin Mountains are a north-south–trending range extending from El Paso, Texas, into southern New Mexico (Fig. 1) (Harbour, 1972). Uplift and westward tilting of the range has exposed a rich assemblage of igneous, metamorphic, and sedimentary rocks, which document a complex history. The oldest rocks include Mesoproterozoic carbonates, siliciclastic sediments, and volcanic rocks and tuffs. These are preserved as roof pendants within the 1120 ± 35 Ma Red Bluff Granite suite, resulting in amphibolite-grade contact metamorphism (Thomann, 1981; Shannon et al., 1997). These rocks are in turn overlain by rhyolitic ignimbrites, porphyritic trachytes, and volcaniclastic sediments of the Thunderbird Group, which are interpreted to be the eruptive equivalent of the Red Bluff Granite suite (Thomann, 1981). A 2717-m-thick succession of Paleozoic sedimentary rocks are separated from the Precambrian rocks by a major nonconformity (LeMone, 1988). Uplift of the Franklin Mountains was primarily accomplished through slip along the eastern boundary fault zone (EBFZ), which marks the eastern edge of the range (Fig. 1B). Trenching studies of the EBFZ, shown in Figure 1B, identified multiple rupture events over the last 16,400 yr, with an average slip rate of 0.175 mm/yr (Keaton and Barnes, 1995; McCalpin, 2006).
The Franklin Mountains uplifted block separates the Mesilla Basin to the west from the Hueco Basin to the east (Fig. 1). The Mesilla Basin is ∼35 km wide by 100 km long. Mack et al. (1997) also considered the Mimbres Basin west of the Mesilla basin to be part of the Rio Grande rift, and combined, the total width of the Mesilla and Mimbres Basins spans ∼100 km in southern New Mexico.
Beginning in the Miocene, sediment carried by the Rio Grande from the mountains of Colorado and northern New Mexico modified the landscape in ways important to this study. The southern part of the rift was the depositional center of this extended axial stream system; the preexisting topography was largely buried under rapidly aggrading fluvial and lacustrine sediments (Mack et al., 2006). The first evidence of the Rio Grande in the southern rift near the Franklin Mountains has been dated at 5 Ma by using paleomagnetic reversals and dating of volcanic rocks interbedded within the sediments (Mack et al., 1998b; 2006). Fluvial deposition had shifted to the east side of the Franklin Mountains by 2 Ma when the ancestral Rio Grande spilled through Fillmore Gap between the Organ and Franklin Mountains (Seager et al., 1984; Mack et al., 2006) (Fig. 2). By the early Quaternary (2 Ma), the ancestral Rio Grande had bifurcated north of the Franklin Mountains near Fillmore Gap and had begun filling both the Hueco and Mesilla Basins with river sediments (Hawley et al., 1969; Gustavson, 1991; Hawley and Kennedy, 2004). The terraces mapped in this study are buried by these sediments in Fillmore Gap, indicating a terrace age significantly older than 2 Ma.
Deposition on the basin floors ended ∼700,000 yr ago (Vanderhill, 1986; Mack et al., 1993, 1998a, 1998b; Hawley et al., 1969; Lucas et al., 1999; Gile et al., 2007). The 640 ka Lava Creek B ash is found in the oldest inset sediments, demonstrating that initial Mesilla Valley downcutting occurred before 640 ka (Izett and Wilcox, 1982; Seager et al., 1984; Gile et al., 1981). The final depositional surface forms the lower La Mesa surface (LMS) (Figs. 1, 2) (Gile et al., 1981). The LMS of today is a broad flat slope that dips southward at ∼0.001. The LMS has not been heavily deformed and is therefore thought to preserve the southerly gradient of the paleo–Rio Grande. The LMS is offset by north-south–trending Quaternary faults and accumulation of eolian dunes. However, overall southerly gradients of 0.001–0.002 are preserved on the lower LMS flanking the Rio Grande, with higher slopes in the northern part of the basin and lower slopes to the south. Dunes and erosion result in high-frequency noise in elevation profiles, but the surfaces exhibit well-defined gradients. These gradients are similar to those of the modern Rio Grande, which has a modern gradient of 0.001 through the Mesilla Basin.
Eastern Boundary Fault Zone
The Franklin Mountains are separated from the Hueco Basin to the east by the eastern boundary fault zone (EBFZ) (Fig. 1). The EBFZ is an east-dipping structure that trends north-south along the entire length of the Franklin Mountains for a total distance of 53 km (Machette et al., 1998). Although in some locations the surface expression of the fault has been concealed or destroyed by urban development, the fault appears to be a single structure along most of its length. The fault can be mapped south to the southern edge of the Franklin Mountains, where it becomes concealed beneath the urbanized regions of El Paso (Texas) and Ciudad Juárez (Mexico). Rocks exposed in the footwall of the EBFZ include the Precambrian Red Bluff granitic suite and associated Thunderbird rhyolite, which are overlain by Paleozoic strata. The hanging wall is composed of late Oligocene Santa Fe Group rift fill capped by Camp Rice Formation (Gile et al., 1981; Collins and Raney, 1991; Hawley et al., 2009). Basin subsidence in this region occurred from ca. 10 to 2 Ma when the majority of fill was deposited (Hawley et al., 2009), suggesting that main slip along the EBFZ also occurred at this time.
Recent activity along the EBFZ and associated normal faults within the Hueco Basin is expressed through numerous fault scarps that cut the surface of the basin (Collins and Raney, 1991). Ramberg et al. (1978) documented en echelon faults along the east side of the Franklin–Organ–San Augustin–San Andres chain of mountains. They estimated 9.6 m of Holocene to late Quaternary vertical slip based on a 35-m-long fault scarp on the east side of the north Franklin Mountains, and about 7 m of slip near the south end of the Franklin Mountains. Collins et al. (1996) estimated an average slip rate of 0.1 mm/yr and noted that Quaternary scarps were up to 60 m high.
The EBFZ is the more active of the boundary fault zones. Studies near White Sands Missile Range, New Mexico, to the north of the EBFZ indicate that the last movements on the network of faults extending from the Organ Mountains to the east Franklin Mountains occurred within the last 4000–5000 yr (Seager, 1980). The date was determined by comparing soil development on the oldest unfaulted fan to the soil on the youngest faulted fan (Seager, 1983). Movement along the Artillery Range fault between the north Franklin and Organ Mountains may have blocked the ancestral Rio Grande’s course into the Hueco Basin at least 650,000 yr ago (Mack et al., 2006).
A study of the EBFZ by McCalpin (2006) focused on a trench at the mouth of Hitt Canyon, 3 km south of the Texas–New Mexico state line (Fig. 1). McCalpin (2006) found five normal faults, though there was uncertainty associated with the three oldest faults. He was able to date three slip episodes using 14C and compared them with infrared-stimulated luminescence done on four fine inorganic silt samples. The mean throw was measured at 3.5 m, although two paleoearthquakes exhibited displacement of ∼3 m. A mean slip rate of 0.175 mm/yr was estimated as a minimum slip rate from the last 16,400 yr from the total displacement of 11.2 m.
Western Boundary Fault Zone
The western boundary fault zone (WBFZ) is a poorly exposed structure with uncertain history. This fault may extend along the entire length of the western margin of the Franklin Mountains and possibly as far north as the Organ, San Augustin, and San Andres ranges (Lovejoy, 1975). However, it is partially covered by possible landslides, especially adjacent to the Franklin Mountains. Alternatively, Harbour (1972) saw no evidence of this fault north of Avispa Canyon (Fig. 1). The WBFZ cuts a major Laramide thrust fault and numerous smaller faults, suggesting that latest movement occurred during the Neogene (Scharman, 2006). The WBFZ appears to be inactive today because nowhere does it disrupt Pliocene terraces (Scharman, 2006). Instead, the active fault along the western side of the Franklin Mountains is the Mesilla Valley fault, which lies along the Rio Grande corridor and west of the terraces of this study (Henry et al., 1985).
The sedimentary record of the southern Rio Grande rift system includes Miocene strata and the Plio-Pleistocene Fort Hancock and Camp Rice Formations of the Santa Fe Group (Hawley, 1975; Mack et al., 1998a, 1998b; Hawley and Kennedy, 2004; Hawley et al., 2009) (Fig. 2). Along most of the east flank of the Franklin Mountains, Fort Hancock sediments consist of coarse gravels derived from the Franklin Mountains that intertongue with basin-floor playa mudstones. These units are typically exposed within a few meters above bedrock. Along the western side of the mountain, fluvial sediments of the Camp Rice Formation containing exotic clasts not derived from the Mesilla Basin form the upper 750 m of the basin fill. These sediments are intermittently exposed within a few tens of meters horizontally from the western front of the Franklin Mountains, and have been interpreted as deposits of the Rio Grande derived from upstream sources (Chapin and Seager, 1975; Collins and Raney, 1991).
Dating of the basin fill of the Hueco Basin is problematic due to lack of suitable material and difficulty in correlation with other basins. However, nearby basins began to fill ca. 12–15 Ma, and the majority of the basin fill is probably Miocene and Pliocene (Cather et al., 1994; Langford et al., 1999). Hawley et al. (2009) noted a wedge of upper Santa Fe Group sediments that expands horizontally toward the EBFZ that they tentatively correlated with Pliocene and early Quaternary sediments derived from the Rio Grande.
The fill of the Mesilla Basin has been characterized by several authors (Hawley and Lozinsky, 1992; Sellepack, 2003; Hawley and Kennedy, 2004), and is summarized in Figure 2. The Camp Rice and Fort Hancock Formations, which form the upper Santa Fe Group, represent the influx of Rio Grande sediments into the region. Mack and Seager (1990) placed the arrival of the Rio Grande in the Mesilla Basin at 5 Ma. However, far-traveled quartzite clasts first arrived in the Mesilla Basin by 3.58 Ma based on the proximity of the Camp Rice Formation to the Gilbert-Gauss geomagnetic polarity boundary (Mack et al., 1993; Repasch et al., 2017), suggesting that the Rio Grande had reached the Mesilla Basin by this time. In the Hueco Basin, Albritton and Smith (1968) and Gustavson (1991) dated the appearance of the Rio Grande sometime before 2.06 Ma based on the Huckleberry Ridge ash that is interbedded with Camp Rice sediments. However, Hawley et al. (2009) suggested an earlier appearance of the Rio Grande in the Hueco Basin, generally coincident with its appearance in the Mesilla Basin.
DEFORMED TERRACES OF THE FRANKLIN MOUNTAINS
Unlike most of the uplifts associated with the Rio Grande rift, the Franklin Mountains preserve a set of uplifted terraces that flank both sides of the range (Lovejoy 1971, 1975). The terraces overlie Rio Grande channel fills on the west flank of the range and therefore are younger than 5 Ma (Mack et al., 2006). The ancestral Rio Grande left cross-bedded and horizontally laminated pebbly sandstones and mudstones (Mack et al., 1998b) that formed the surface across which the terraces were initially deposited. The terraces are older than 2 Ma based on K-Ar dating of volcanic deposits (Chapin and Seager, 1975; Seager et al., 1984), vertebrate fauna (Strain, 1969; Metcalf, 1969), and studies of soil development on the upper beds of the Camp Rice Formation (Seager, 1980; Gile et al., 1981). When compared to the undeformed early Pleistocene LMS (Gile et al. 1981), these terraces are likely middle Pliocene in age.
The terraces at the same elevation along the eastern side of the Franklin Mountains are underlain by interbedded playa mudstones and alluvial fan gravels of the Fort Hancock Formation along the southern portion of the range, and Camp Rice Formation fluvial sandstones along the northern portion. The terraces are capped by thick alluvial fan gravels. During deposition of the Fort Hancock and Camp Rice Formations, uplift of the mountain block began to lift parts of this low-relief surface above the aggrading basin floors. This caused deposition to cease and defined the uplifted terrace surfaces that are the subject of this study. Continued uplift of the Franklin Mountains also uplifted these terraces within the footwall of the EBFZ (Harbour. 1972; Lovejoy, 1971) so that the terraces on both flanks of the range record overall deformation of the range after their deposition.
The terraces flank large portions of both sides of the Franklin Mountains, although they have been dissected by erosion (Figs. 1B, 1C). To constrain deformation of these surfaces, a Trimble Geo Xh 2005 Series Pocket PC was used to acquire 109 GPS-located points along both flanks of the Franklin Mountains as well as the LMS. The elevation data were generally more accurate than ±30 cm and commonly better than ±15 cm of true elevation above mean sea level. At many locations, alluvial fan deposits blanket the terraces within close proximity to the mountain front. Therefore, to estimate the true elevation of these terraces, elevations were measured every ∼21 m in an east-west direction across 22 of the terraces to establish profiles along the crest of each terrace from the alluvial fan to the toe of the terrace (Fig. 3). From these data, the elevations of the terrace toes are interpreted to be the true elevation at each location. In areas where the terraces were inaccessible such as the firing range at Fort Bliss military base and the LMS north of 32°N latitude, elevations were estimated from U.S. Geological Survey 7½-minute topographic maps or digital elevation models with accuracy of ±1.5 m on 14 terraces. This resulted in a total of 59 elevation measurements along the entire eastern and western flanks of the Franklin Mountains that were used to estimate deformation (Fig. 1).
Terrace Morphology and Stratigraphy
A typical uplifted terrace in the study area is a relatively low-gradient peninsula of higher topography extending from the mountains toward the surrounding basins (Fig. 4). Uplifted terraces are typically 85–150 m across and extend 140–350 m from the steeper alluvial fans to the toes (Fig. 4B). Crests along the length of the uplifted terraces are low gradient with slopes of 5% to 10% from the mountain front. The flanks of the uplifted terraces are steeper, with slopes of 35% to 50% into the surrounding arroyos. Many of the larger uplifted terraces are associated with canyons in the mountains that carve into terrace surfaces. The canyons debauch into arroyos that are diverted along the mountain front for a short distance before continuing between the uplifted terrace remnants to the basins. This suggests that early deposition of alluvial fan gravel on the upper terraces protected them from erosion. The more exposed areas flanking the canyons have thinner gravels and were more vulnerable to incision.
Two processes modified the terraces after deposition and altered data acquisition. First, the margins of the terraces were eroded. The distal ends of the terraces were reduced in elevation and the terraces were incised by arroyos exiting the Franklin Mountains, which cut the surface into a series of fingers extending from the mountain (Figs. 1C, 4A). Additionally, younger alluvial fans prograded over many of the uplifted terraces. These deposits have steeper slopes than the terraces (Figs. 3, 4). An oblique aerial view commonly gives some indication of the extent to which alluvial fans prograde over the uplifted terraces. The surfaces of these alluvial fans are composed of coarse, angular clasts derived from the mountain slopes. The transition from uplifted terrace to alluvial fan is gradual and can be difficult to identify in the field. At the toe of the uplifted terrace, the surface cover ranges from thin desert soil with a scattering of medium to coarse gravel to a well-cemented cobble surface. The varying thickness of alluvial fan gravels and the transition to terrace toes create noise in the interpreted terrace elevations (Fig. 3). Slope angles up the crests of terraces range from 0° to 10°.
The internal structure and stratigraphy of the uplifted terraces are usually concealed by indurated Pleistocene or Holocene calcic soils, except where exposed by arroyo incision or city development. Although investigating the internal stratigraphy of terraces in the Franklin Mountains is difficult in many places, terrace remnants are preserved in the Robledo Mountains 60 km to the northwest of the study area and 15 km north of Las Cruces, New Mexico (Fig. 1A), and are a useful analogue. In the Robledo Mountains, uplifted terraces are incised through their length, and the internal structure of terraces is better exposed than in the Franklin Mountains. These terraces consist of fine- to medium-grained sands of the Camp Rice Formation that were measured with clinometers to dip 2°–4° into the Robledo Mountains block. This tilting is similar to, but at lower angles than, the overall tilt of the range and is ascribed to tilting of the hanging-wall block during extensional deformation. Similar strata are exposed along the eastern and western flanks of the Franklin Mountains, but the limited exposure precludes estimates of the dip of the bedding, although dips are likely similar to bedding dips measured in the Robledo Mountains terraces.
In the incised terraces flanking the Robledo Mountains, the basin-floor strata are overlain by layers of coarse gravel that dip gently basinward. These beds lap onto the basin-floor sediments and are truncated at the top, and are inferred to represent the progradation of alluvial fan gravels across the Camp Rice sediments prior to erosion that occurred during uplift of the hanging wall. The terrace surface overlying these inclined strata is much flatter and appears to be an almost horizontal finger extending from the range front toward the basin.
Along-Strike Geometry of Terraces
Beginning ca. 5 Ma, deposition of the Camp Rice and Fort Hancock Formations began to form an initially planar surface that filled the Hueco and Mesilla Basins (Strain, 1980; Gile et al., 1981; Mack et al., 2006). Subsequent faulting related to continued uplift of the range deformed the originally planar surface, providing a datum from which to estimate displacement along the length of main normal faults. To estimate fault throw, we used the nearly horizontal LMS as an approximation of the original geometry of terrace deposits. The LMS forms a low-gradient surface with a 0.06% slope and preserves the southerly gradient of the paleo–Rio Grande. Calculated fault throw along the length of the faults, described below, is assumed to be minimum because the Camp Rice and Fort Hancock deposits in the Mesilla and Hueco Basins continued to aggrade during uplift of the range until 570,000 yr ago (Mack et al., 2006).
The uplifted terraces on both sides of the range are at different elevations above the modern LMS (Fig. 5). On the eastern side of the range, the highest terrace elevations are 1370 m, ∼118 m higher than the LMS. These terraces are ∼4 km south of South Franklin Mountain. At the north end of the range, the terrace elevations are much lower and only ∼22 m above the LMS. The terrace surface is underlain by fluvial channel sands of the Camp Rice Formation (Hawley and Kottlowski, 1969). Lava Creek B ash, which erupted at 0.64 Ma (Lanphere et al., 2002), is preserved in the uppermost strata west of Fillmore Gap, suggesting that uplift of terraces in this region occurred within the last 0.64 m.y. The Fillmore Gap segment is also not deformed like the uplifted terraces to the south. This leads to the inference that the Fillmore Gap is a younger terrace that formed on sediment that buried the older deformed terrace. In this area, deposition in the Mesilla and Hueco Basins was faster than the rate of uplift of the terraces (Mack et al., 2006).
The terrace elevations, when plotted along strike of the Franklin Mountains, show a broad 40-km-long asymmetrical anticlinal arch (Fig. 5). At the northern end of the uplift, terraces are preserved at a height of ∼30 m above the LMS. The height increases dramatically to the south, reaching maximum heights of ∼118 m near South Franklin Mountain along the eastern profile. Height above the LMS surface then decreases to the south. The southernmost terrace along the eastern profile is ∼79 m above the LMS, while the southernmost terrace along the western profile is ∼7 m above the LMS (Fig. 5).
Minimum and maximum uplift rates can be calculated using reasonable estimates of the total time over which uplift occurred and the total uplift magnitude. The maximum age of the terraces is 5 Ma, based on the date of the entry of the Rio Grande into the Mesilla Basin (Mack et al., 1998b, 2006). Terraces are older than the 700,000-yr-old upper LMS, west of Las Cruces, New Mexico (Vanderhill, 1986; Mack et al., 1993). A reasonable minimum age would be 3 Ma, due to the magnitude of uplift relative to the LMS. Using these estimates, fault throw rates were calculated for each terrace location along the eastern and western lengths of the Franklin Mountains (Fig. 6). These throw rates were calculated using assumed times of deposition of 5 Ma, 4 Ma, and 3 Ma. Minimum throw rates of ∼0.001–0.002 mm/yr are found along the western profile at the southern tip of the Franklin Mountains where they intersect with the city of El Paso, Texas. Maximum throw rates of ∼0.02–0.04 mm/yr are located along the eastern margin south of Fusselman Canyon and north of McKelligon Canyon. These values are an order of magnitude lower than those of McCalpin (2006), who estimated Quaternary fault slip rates of ∼0.18 mm/yr at a trench site along the EBFZ near the Texas–New Mexico border (Fig. 1). One possible explanation for this discrepancy is that throw rates calculated in this study are averaged over a much longer interval of time than the throw rates calculated by McCalpin (2006).
Terraces along the western edge of the Franklin Mountains have been uplifted by various amounts relative to the LMS since deposition, and show an anticlinal arch similar to, but at lower elevations than, that of the eastern flank of the range (Fig. 5). Minimum uplift of ∼7–8 m is at the southernmost terrace location, and maximum uplift of ∼90 m is located near Fusselman Canyon. As discussed previously, the WBFZ is likely inactive because it is not observed to offset Pliocene terraces (Scharman, 2006). Therefore, uplift of these terraces was likely controlled by the EBFZ along the eastern flank of the range. The observed uplift of these terraces suggests that, in addition to rotation of the Franklin Mountains during slip movement along the EBFZ, the entire range was also uplifted during continued extension. This deformation resulted in the observed terrace profile in Figure 5, and accounts for the tilting and rotation of the range first documented by Lovejoy (1971).
The geometry of the fault throw profile along the length of the EBFZ delineated by the uplifted terraces suggests that displacement has been concentrated to the south during the last 5 m.y., with maximum fault throw estimates of ∼118 m south of Fusselman Canyon. In contrast, the northern segment of the fault has accumulated a total of ∼30 m of throw during the same time (Fig. 5). At the southern tip of the range, the EBFZ has been uplifted ∼79 m above the LMS, while the southernmost location along the western margin has only been uplifted ∼7 m and extends farther south (Fig. 5).
Based on the geometry of uplifted terraces, the EBFZ appears to conform with normal fault growth models described above. To the north, the 30 m of uplift at the northern tip of the range (Fig. 5) suggests that over the last 5 m.y., the EBFZ has extended to the north, by either tip line migration, linkage of initially isolated faults, or some combination of the two. The 79 m of fault throw adjacent to the EBFZ at the southern tip of the Franklin Mountains suggests that this tip line has also been migrating south during the last 5 m.y. This interpretation is supported by gravity data that suggest the presence of multiple faults that have been obscured by city development; these faults trend north-south and extend into Ciudad Juárez (Marrufo, 2011; Avila et al., 2015). In particular, while the EBFZ is difficult to trace south of the Franklin Mountains, Avila et al. (2015) noted a steep north-south–trending gradient in gravity data and infer that the EBFZ extends for as much as 30 km south of its mapped southern extent, bisecting the downtown El Paso and Ciudad Juárez populated areas and terminating along the eastern edge of the Sierra de Juárez uplift in Mexico. This is also supported by water well information for the area (Keaton et al., 1995; Hawley et al., 2009). The presence of buried faults beneath El Paso and Juárez (Marrufo, 2011; Avila et al., 2015) and the observation that maximum uplift and throw rates are concentrated to the south all suggest that the EBFZ is growing in this direction, either by tip line migration or fault linkage. As a whole, these data are consistent with a model in which the EBFZ has been growing to the north and to the south during the last 5 m.y.
The asymmetrical nature of the fault throw profile can also give some indication of the fault’s past and possible future behavior. Skewed displacement profiles, where maximum displacement is not centered along the fault, suggest elastic interactions with neighboring faults (e.g., Peacock and Sanderson, 1991; Peacock, 2002). In these situations, faults can be hard linked (intersect) or soft linked (faults overlap and form relay ramps, but do not intersect). Fault interaction strongly affects the final displacement profile and is characterized by steep displacement gradients and asymmetrical profiles (Peacock, 2002), similar to what is observed along the southern edge of the displacement profile for the EBFZ. Coupled with the geophysical arguments for faults extending beneath the El Paso and Juárez urbanized regions (Marrufo, 2011; Avila et al., 2015), the displacement profile geometry for the EBFZ suggests that the EBFZ has been lengthening to the south through linkage with other fault systems.
Migration of the EBFZ to the south through fault linkage, as we postulate here, would have significant implications for future seismic hazards. Although there have not been any historic earthquakes along the EBFZ, two magnitude 2.5 earthquakes occurred in the Hueco Basin in 2012 (U.S. Geological Survey, 2017), most likely along intrabasinal faults. The most recent earthquake ruptures along the EBFZ were in the early to mid-Holocene (Keaton and Barnes, 1995), and seismic data are too sparse to reconstruct fault growth; this manuscript is the first to document growth of the fault zone. Based on trenching studies, McCalpin (2006) suggested that the EBFZ is capable of producing magnitude 7 earthquakes, and that at least four earthquakes have occurred along this fault within the past 64,000 yr. As faults in this region continue to grow and interact, the probability of an earthquake jumping from one fault to the next increases (Scholz and Gupta, 2000). This would increase the surface area of the ruptured fault, and therefore the seismic moment and magnitude of the earthquake compared to a rupture along the southern EBFZ alone.
Uplifted terraces exposed along the flanks of the Franklin Mountains preserve a record of differential normal fault slip along the EBFZ over the past 3–5 m.y. The low-gradient LMS west of the Rio Grande is composed of the same formations that make up terraces flanking the Franklin Mountains, and provides a reference from which to measure relative uplift of the terraces. Terrace elevations mimic the topography of the Franklin Mountains, suggesting that uplift of the terraces is tectonic in nature. The terrace elevation profile along the western margin of the uplift is similar to, but at lower elevations than, terrace elevations along the eastern margin, suggesting overall uplift and tilting of the entire Franklin Mountains since the Pliocene. Maximum throw rates for the EBFZ range from 0.02 to 0.04 mm/yr since 3–5 Ma. In addition, uplifted terraces suggest that the EBFZ has been growing in length to the north as well as to the south during this time interval. The asymmetric terrace profile, when coupled with geophysical and well-log data for the southern tip of the Franklin Mountains, suggests that continued growth of the EBFZ to the south is viable through linkage with other buried faults.
Constructive reviews by W.S. Baldridge, Associate Editor G.D.M. Andrews, Science Editor S. de Silva, and an anonymous reviewer helped to strengthen the arguments in this paper.