The Long Valley region of eastern California (United States) is the site of abundant late Tertiary–present magmatism, including three geochemically distinct stages of magmatism since ca. 3 Ma: Mammoth Mountain, the Mono-Inyo volcanic chain, and Long Valley Caldera. We propose two tectonic models, one explaining the Mammoth Mountain–Mono-Inyo magmatism and the other explaining the presence of Long Valley Caldera. First, the ongoing Mammoth Mountain–Mono-Inyo volcanic chain magmatism is explained by a ridge-transform-ridge system, with the Mono-Inyo volcanic chain acting as one ridge segment and the South Moat fault acting as a transform fault. Implicit in this first model is that this region of eastern California is beginning to act as an incipient plate boundary. Second, the older Long Valley Caldera system is hypothesized to occur in a region of enhanced extension resulting from regional fault block rotation, specifically involving activation of the sinistral faults of the Mina deflection. The tectonic models are consistent with observed spatial and temporal differences in the geochemistry of the regional magmas, and the westward progression of magmatism since ca. 12 Ma.
The Sierra Nevada microplate is bound on its eastern margin by Walker Lane and Eastern California shear zone, which together accommodate ∼25% of the motion between the Pacific and North American plates (Dixon et al., 2000; Dewey, 2002; Unruh et al., 2003; Wesnousky, 2005). Geodetic data indicate that the Walker Lane belt is transtensional, with motion partitioned between Sierra Nevada frontal normal faults and dextral strike-slip faults to the east of the normal faults (Oldow, 1992; Wesnousky and Jones, 1994; Unruh et al., 2003). Previous work (e.g., Faulds et al., 2005; Putirka and Busby, 2007) proposed that the Walker Lane belt is an incipient rift margin.
Current rift models for the Walker Lane belt do not utilize magmatism to accommodate the space created by the observed divergence, a common property of established rift margins. However, along the Walker Lane belt in the vicinity of Long Valley (California, United States) abundant volcanic activity has occurred for the past 3 m.y., including the initiation of Long Valley rhyolitic eruptions (ca. 2.2 Ma), eruption of the Bishop Tuff from Long Valley Caldera (760 ka), repeated eruptive events at Mammoth Mountain (110–57 ka), and the Mono-Inyo Craters (largely younger than 20 ka) (Bailey, 1989, 2004; Hill, 2006; Mahood et al., 2010). The anomalous volume and frequency of magmatism in the Long Valley area compared to other late Quaternary (e.g., Big Pine, Coso) volcanic fields along the eastern Sierra Nevada (Fig. 1) suggest that some tectonic process results in localized magmatic intensification.
We investigate whether the interplay between transtensional deformation and magmatism in the Long Valley region of eastern California can explain the location, timing, and distinct geochemical signature of the different phases of magmatism. We propose two tectonic models that are consistent with transtensional deformation along the eastern margin of the Sierra Nevada microplate, and that lead to concentrated magmatism in the vicinity of Long Valley. The first model, a ridge-transform-ridge system, addresses ongoing deformation and magmatism. In this model, the dextral, east-southeast–oriented South Moat fault zone separates a northern dike swarm (reflected at the surface by the Mono-Inyo Craters) from east-west–directed extension with volume expansion inferred from focal mechanisms. The second model addresses Long Valley magmatism, which is inferred to occur in an area of enhanced extension resulting from regional fault movement. In particular, the temporal association of magmatism in Long Valley with initiation of sinistral faults within the Mina deflection is used to suggest that differential block movement created a local tectonic gap in the region of Long Valley Caldera. These interpretations are consistent with both the evolving transtensional deformation in the Walker Lane–Eastern California shear zone systems and the different magma sources for Mammoth Mountain and Long Valley magmatism.
The Sierra Nevada microplate is a crustal block moving independently from both the Pacific plate to the west and the North American plate to the east (Fig. 2; Dixon et al., 2000; Unruh et al., 2003). The microplate formed ca. 12–11 Ma, during opening of the Gulf of California (Henry and Perkins, 2001; Putirka and Busby, 2007). The dextral San Andreas fault system defines the western margin of the microplate. The dextral, transtensional Walker Lane–Eastern California shear zone and Sierra Nevada frontal fault system define its eastern margin (Fig. 2; Unruh et al., 2003; Wesnousky, 2005; Faulds and Henry, 2008). For clarity, we use Walker Lane to indicate the series of discontinuous strike-slip faults north of ∼lat 38°00′N, and Eastern California shear zone for the strike-slip fault system south of ∼38°00′N (sensu Oldow et al., 2008). Although extension occurs throughout the Basin and Range Province (e.g., Oldow, 2003), we define points east of the Walker Lane–Eastern California shear zone system as stable North America.
The Sierra Nevada frontal fault system is a series of normal faults that demarcate the eastern escarpment of the Sierra Nevada (Fig. 2). The normal faults strike north-northwest–south-southeast, or oblique to the northwest-southeast Sierra Nevada–North America extension direction, and are segmented with a left-stepping, en echelon geometry (Bateman and Wahrhaftig, 1966). Previous interpretations suggested that the fault system accommodated Basin and Range extension and uplift of the Sierra Nevada (Bateman and Wahrhaftig, 1966), but new interpretations suggest that the normal faults also accommodate northwest-southeast–directed movement between the Sierra Nevada microplate and the North American plate (Sonder and Jones, 1999; Unruh et al., 2003).
The boundary between the Sierra Nevada microplate and stable North America near Long Valley is geographically divided based on deformation style into the Eastern California shear zone, central Walker Lane, and the transfer zone between the two portions (Mina deflection). In the Eastern California shear zone, several large-displacement dextral faults, including, but not limited to, the Owens Valley, Fish Lake Valley, Furnace Creek, Panamint Valley, and Hunter Mountain faults, currently accommodate the majority of dextral offset (Beanland and Clark, 1994; Reheis and Sawyer, 1997; Wesnousky, 2005; Kirby et al., 2008). The relative plate motion within the central Walker Lane is accommodated along a series of north-northwest–striking dextral faults (Ekren and Byers, 1984; Stewart, 1988; Faulds and Henry, 2008). The dextral faults are left stepping from east to west, and no single fault accommodates all of the deformation (e.g., Ekren et al., 1980; Faulds and Henry, 2008).
Slip is currently transferred from the Eastern California shear zone to the central Walker Lane along a major east-northeast–oriented, 80-km-wide right step termed the Mina deflection (Oldow, 1992), which consists of a series of 20–40-km-long east-northeast–striking sinistral strike-slip and dip-slip faults that form east-northeast–oriented crustal blocks (Ryall and Priestly, 1975; Oldow, 1992, 2003; Wesnousky, 2005; Faulds and Henry, 2008; Oldow et al., 2008). To the east, slip on faults of the Mina deflection is gradually lost onto the north-northwest–striking dextral faults of the central Walker Lane. No clear boundary exists on the western edge of the Mina deflection, and faults appear to be in continuity with a large crustal block (herein called the Adobe block) that terminates east of the Sierra Nevada range front (Fig. 2). This block is less faulted than surrounding regions, although it contains several minor fault zones (e.g., Bald Mountain, Bald Mountain–Clover patch, and Benton Range zones; Bailey, 1989).
The location and orientation of the Mina deflection are likely controlled by the basement structure, and geological studies suggest that such structures are originally Precambrian age (Stewart, 1988; Oldow, 1992; Oldow et al. 2008). Reactivation of the Mina deflection, however, is a relatively recent event. Prior to the Pliocene, dextral deformation and magmatism were located farther east. The Silver Peak–Lone Mountain extensional complex formed from 11 to 4 Ma as right-step motion between the Furnace Creek fault system and faults of the central Walker Lane (Oldow, 1992; Oldow et al., 1994, 2008). Crustal blocks within Silver Peak–Lone Mountain extensional complex have rotated ∼20°–25° clockwise (Petronis et al., 2002, 2007), indicating significant displacement on the bounding faults (Oldow et al., 2008). The abandonment of the Silver Peak–Lone Mountain extensional complex and subsequent activation of the Mina deflection does not correspond to any known change in plate motion between the Pacific and North American plates (e.g., Atwater and Stock, 1998), but it does correspond with the onset of high-K magmatism (Feldstein and Lange, 1999; Putirka and Busby, 2007). Putirka and Busby (2007) speculated that this magmatism may have triggered a new phase of tectonism. Regardless of the cause, the initiation of the Mina deflection was temporally coincident with onset of voluminous magmatism starting ca. 3 Ma, including that of Long Valley.
TECTONIC MODEL FOR MAMMOTH MOUNTAIN AND THE MONO-INYO CRATERS
Any model for recent (200 ka–present) deformation and magmatism in Long Valley must address these constraints, discussed in detail in the following.
1. Magmatism at Mammoth Mountain and the Mono-Inyo Craters are geochemically distinct (Hildreth, 2004).
2. The Mono-Inyo volcanic chain north of Mammoth Mountain defines an approximately north-south trend in volcanic activity. This trend of magmatism terminates north of the South Moat fault zone near its west end.
3. Dextral motion occurs along the South Moat fault zone, which reactivates a small portion of the ca. 760 ka Long Valley Caldera. The fault did not exist prior to the formation of the caldera, and it does not continue westward past Mammoth Mountain or eastward past the Sierra Nevada range front (Prejean et al., 2002; Hill, 2006).
4. Earthquakes southeast of Mammoth Mountain occur on north-south– to north-northeast–oriented trends. The earthquakes occur south of the South Moat fault zone and the trends intersect the South Moat fault zone near its eastern end. These earthquakes exhibit non-double-couple solutions or resolved sinistral movement parallel to their trends, and have not been observed to offset any surface features.
Mammoth Mountain Magmatism
Between 175 and 55 ka, basalt and dacite erupted west of Long Valley Caldera (Fig. 3; Hildreth, 2004; Mahood et al., 2010). Magmatism was focused at Mammoth Mountain, which is situated on the southwestern margin of the physiographic Long Valley Caldera. Mammoth Mountain rises >750 m above the caldera rim, and consists of a pile of ∼23 trachydacite and alkali rhyodacite lava domes and flows erupted between ca. 110 and ca. 55 ka, but mostly between 68 and 57 ka (Mahood et al., 2010). This stack of lavas, ∼7 km in diameter, is within a larger (∼16 km diameter), more dispersed field of mafic vents (175–8 ka), the ages of which bracket the eruptive ages of Mammoth Mountain.
The numerous vents for Mammoth Mountain and its periphery of mafic lavas are entirely outside the ring-fault zone of Long Valley Caldera (Hildreth, 2004). The four youngest rhyolite lavas (150–100 ka) of the Long Valley system are contemporaneous and contiguous with (but not within) the Mammoth array, but (like all Long Valley rhyolites, 2.2 Ma to 100 ka) they are subalkaline, in contrast to the alkaline lavas of the Mammoth system (fig. 4 of Hildreth, 2004). The eruptive volume of silicic products at Mammoth Mountain (110–55 ka) is ≤5 km3 and the volume of mafic lavas in its periphery (175–8 ka) is <10 km3; these modest volumes contrast with the rhyolite volume of >900 km3 erupted from the Long Valley magma reservoir between 2.2 Ma and 0.6 Ma (plus minor additions 0.5–0.1 Ma). The Mammoth Mountain trachydacites and rhyodacites and the surrounding mafic lavas are more alkalic than products of either Long Valley or the Mono-Inyo chain (Hildreth, 2004). Thus, based on geochronology and geochemistry, Mammoth Mountain had a magmatic plumbing system distinct from both adjacent volcanic systems (Hildreth, 2004).
With respect to other Pleistocene magmatism, Mammoth Mountain occupies a unique structural position. It is situated directly on top of the mid-Cretaceous, lithospheric-scale, dextral transpressional Rosy Finch shear zone (Fig. 3; Tikoff and Teyssier, 1992; Tikoff and Greene, 1997; Tikoff and Saint Blanquat, 1997). The Rosy Finch shear zone is part of the Sierra Crest shear zone system, which extends ∼300 km along the crest of the Sierra Nevada. The shear zone extends to the southeast of Mammoth Mountain, and dextral shearing extends to the northwest as the Gem Lake shear zone (Greene and Schweikert, 1995; Tikoff and Saint Blanquat, 1997).
Ongoing seismicity is also occurring on Mammoth Mountain. In 1989, a 6-month-long earthquake swarm began under Mammoth Mountain, consisting of thousands of low-energy (MW < 3) earthquakes (Hill and Prejean, 2005; Hill, 2006). Effusion of CO2 around Mammoth Mountain (Sorey et al., 1993; Hill, 1996) and long-period earthquakes (7–25 km below sea level) to the southwest near Devils Postpile (Pitt and Hill, 1994) accompanied the Mammoth Mountain swarm. These events were likely related to the presence and movement of basaltic magma at 10–25 km depth (Sorey et al., 1998; Hill and Prejean, 2005).
Mono-Inyo Volcanic Chain
The north-south chain of Mono-Inyo rhyolite domes and coulees is closely aligned with the Sierran range front fault system for ∼25 km, from the northwest corner of the caldera northward into Mono Basin. The chain is entirely younger than Mammoth Mountain and terminates ∼8 km north of it. The Mono Craters segment consists of ∼28 separate extrusions of compositionally similar, subalkaline high-silica rhyolite that form a continuous curvilinear chain of close-set vents that erupted episodically between ca. 0.7 ka and the present. The Inyo segment continues straight southward along the range front and consists of seven extrusions that erupted between the middle Holocene and A.D. 1350; several of them exhibit mixing of Mono rhyolite with Long Valley rhyolite and with other subordinate contributions (Hildreth, 2004). Vent alignments and episodes of concurrent extrusion of subsets of the Mono-Inyo lavas indicate repeated ascent of dikes parallel to the range front (Eichelberger et al., 1988; Bursik and Sieh, 1989).
South Moat Fault Zone
In 1978, the MW ≈ 5.8 Wheeler Crest earthquake occurred ∼15 km southeast of Long Valley Caldera. Prior to this event, no recorded earthquakes since about the 1940s (cf. Hill, 2006, for complete review) occurred in or around the caldera. Following this event, from 1978 to 1980, a series of smaller (MW < 4) earthquakes occurred that progressed northwest toward the caldera (Ryall and Ryall, 1980; Rundle and Hill, 1988). In 1980, 4 MW ≈ 6 events occurred just south of the caldera, and were followed by a series of aftershocks and smaller (MW ≈ 5) earthquakes until 1983 (Pitt and Cockerham, 1983; Ryall and Ryall, 1983; Rundle and Whitcomb, 1984; Savage and Cockerham, 1984; Hill, 2006).
The 1978–1983 seismicity delineated a zone of dextral slip along the southern structural margin of Long Valley Caldera, named the South Moat fault zone (Fig. 4; Rundle and Whitcomb, 1984; Savage and Cockerham, 1984; Hill et al., 1985; Hill, 2006). In mid-1997, seismicity in the South Moat region again increased in frequency, with >12,120 earthquakes occurring over 7 months and a cumulative seismic moment equal to a MW = 5.4 and again suggesting dextral motion (Hill, 2006). To better delineate the South Moat fault zone structure, Prejean et al. (2002) relocated >45,000 earthquakes between 1980 and 2000, and determined that the South Moat fault zone consists of several strands, including the western and eastern lobes. The motion on both the western and eastern lobes of the South Moat fault zone, as suggested by the relocated earthquakes, is dextral (Prejean et al., 2002).
Dextral motion on the South Moat fault zone is inconsistent with regional deformation. Sinistral motion occurring on similarly oriented faults, 40 km to the northeast within the Mina deflection (Oldow et al., 2008; Figs. 2 and 4). Furthermore, dextral motion on faults of Walker Lane–Eastern California shear zone occurs at nearly orthogonal angles. Consequently, we interpret that movement on this South Moat fault must be a local, and not regional, effect.
South of the South Moat Fault Zone
South of the South Moat fault zone, the bedrock consists dominantly of metamorphosed sedimentary and volcanic rocks of the Mount Morrison pendant (e.g., Greene et al., 1997). While no recent volcanism has occurred, this region has been the site of a series of north-south– to north-northeast–oriented earthquakes swarms, including the 1978–1983 earthquake swarms. Julian (1983) and Julian and Sipkin (1985) calculated that at least 3 of the large 1978–1983 earthquakes contained significant non-double-couple mechanisms, a result they attributed to tensile fracturing with fluid or magma intrusion. Wallace (1985) and Prejean et al. (2002) alternatively suggested that double-couple mechanisms from multiple faults (suggesting faulting rather than volumetric expansion) could explain the observed earthquakes, but did not rule out the possibility of volume expansion associated with the earthquakes. S-wave shadowing south of the caldera outlined possible shallow magmatic bodies (Ryall and Ryall, 1981, 1984; Sanders, 1984; Peppin et al., 1989). Although we focus on the non-double-couple earthquakes in this region, some earthquake swarms exhibit clear sinistral resolved movement parallel to the trend of the swarm (D. Hill, 2011, personal commun.).
Disagreement exists on the validity of the 1978–1983 non-double-couple focal mechanisms (e.g., Wallace, 1985; Prejean et al., 2002), but more recent earthquakes suggest that non-double-couple focal mechanisms occur around Long Valley. In 1997, another earthquake swarm occurred along the South Moat of the caldera and along a north-northeast trend to the south of the caldera (Hill, 2006). Although the majority of these earthquakes indicated dextral motion within the South Moat, a subset of the earthquakes had non-double-couple focal mechanisms. The isotropic component of these seismic moment tensor solutions indicated volume increase, likely related to magmatic fluid injection (Dreger et al., 2000; Prejean et al., 2002; Foulger et al., 2004).
We interpret the normal faulting and volume expansion from fluid movement to represent the development of the southern ridge in our ridge-transform-ridge model. This interpretation is consistent with previous interpretations of the non-double-couple events and normal faulting south of the caldera (e.g., Julian and Sipkin, 1985; Moos and Zoback, 1993; Prejean et al., 2002). In our model, the Mono-Inyo Craters are the well-established portion of the ridge system, whereby dike intrusion reaches the near surface. The southern area is less developed, and has not progressed to the clear north-south–trending magmatism stage of ridge development. Instead, the southern ridge is in its early stages, requiring a combination of both faulting and fluid injection (presumably magmatically derived).
Ridge-Transform Model for Mammoth Mountain and Dike Intrusion
We propose that the Mammoth Mountain area acts as a ridge-transform-ridge system (Fig. 5). Dikes associated with the Mono-Inyo Craters form the first, northern ridge. The dextral South Moat fault forms the transform between ridges, and the north-south trend of seismicity from the Mount Morrison pendant area forms the second, southern ridge.
Dikes associated with the Mono-Inyo Craters and the dikes suggested by recent seismicity may be accommodating extension along the Sierra Nevada microplate–North American plate boundary. Dikes can accommodate regional extension, as inferred for other regions including axial dike swarms in rift settings (Parsons et al., 1998; Teyssier and Tikoff, 1999). In our model (Fig. 5), the dextral movement on the South Moat fault zone is in response to the offset nature of extension and volume expansion to the north and south. This geometry explains many aspects of the regional geology: (1) there is a lack of continuation of the South Moat fault zone farther east or west than its current geometry; (2) the Mono-Inyo crater swarms end at the west end of the South Moat fault zone, while the Mount Morrison swarms end at the east end of the South Moat fault; (3) the dextral South Moat fault zone is not consistent with the regional geology, but exhibits internal consistency with our model.
The model also addresses the issue of the numerous mafic and intermediate alkalic eruptions around Mammoth Mountain. In a ridge-transform-ridge model, Mammoth Mountain is situated at the inside corner of the ridge-transform system. At oceanic rift margins, this is the area where extension attenuates the crust, forming oceanic core complexes (e.g., Tucholke and Lin, 1994; Tucholke et al., 1998; Fig. 5). In the ridge-transform-ridge model, we suggest that the repeated, focused events at Mammoth Mountain were likely facilitated by its location in an area of pronounced attenuation (Fig. 4). Alternatively, magmatism at this location could be facilitated by the Cretaceous Rosy Finch shear zone, over which Mammoth Mountain is situated. However, the presence of the preexisting shear zones and/or fault zones does not appear to have significant influence on dike injection north or south of Mammoth Mountain, because dikes largely crosscut these older structures. The presence of the Rosy Finch shear zone, however, indicates that Mammoth Mountain magmatism may have a different magma conduit, which explains the different geochemical signature between it and the Mono-Inyo crater magmatism.
TECTONIC MODEL FOR LONG VALLEY CALDERA
Eruptive History of Long Valley
The eruptive history was documented by Bailey (1989) and summarized and updated in Hildreth (2004). Mafic and dacitic precursory volcanism was widespread between 4.5 and 2.5 Ma, but since 2.2 Ma the Long Valley system has erupted only rhyolite, almost entirely high-silica rhyolite (74%–77% SiO2). Caldera formation accompanied eruption of 650 km3 of rhyolitic magma as the Bishop Tuff at 760 ka (Hildreth and Wilson, 2007). The caldera-forming eruption was preceded by >50 rhyolite eruptions (total ∼100 km3) that built the Glass Mountain complex (2.2–0.78 Ma) and was followed closely by eruption of ∼100 km3 of intracaldera early postcaldera rhyolite (760–650 ka) that has been uplifted and intricately faulted to form the resurgent dome. Subsequent activity (530–100 ka) has produced only ∼8 km3 of rhyolite, restricted to the annular moat between the resurgent uplift and the caldera wall. Following several large earthquakes in 1980, ∼80 cm of uplift centered on the resurgent dome was measured (Hill et al., 2002, 2003), attributed to renewed magmatic intrusion or to regional tectonic disturbance of the subcaldera hydrothermal system (Hildreth, 2004).
Long Valley is not the only concentration of Pleistocene magmatism along the eastern side of the Sierra Nevada Batholith: contemporaneous volcanic fields to the south include Big Pine, Coso, and Golden Trout (Fig. 1; e.g., Moore and Dodge, 1980; Manley et al., 2000; Farmer et al., 2002). These volcanic fields are all within the area of inferred upwelling asthenosphere that occurred due to the loss of the dense lithospheric root under the eastern Sierra Nevada Batholith (Jones et al., 1994; Ducea and Saleeby, 1996; Fliedner et al., 1996; Wernicke et al., 1996; Manley et al., 2000; Saleeby et al., 2003; Zandt, 2003). Despite temporal and spatial similarities, the Long Valley volcanic field has two distinctive features relative to the Pleistocene volcanic fields to the south. First, Pleistocene volcanism at the Long Valley volcanic field is significantly more voluminous than the volcanic fields to the south, especially considering the caldera-forming eruption of the 760 ka Bishop Tuff. Estimates of magmatism in the southern volcanic fields are: Big Pine (∼0.2 km3 rhyolite, ∼0.5 km3 basalt; Beard and Glazner, 1995), Coso (∼2 km3 rhyolite, ∼30 km3 basalt; Manley and Bacon, 2000), and Golden Trout (∼2 km3 rhyolite, minor basalt; Bacon and Duffield, 1981). Second, Long Valley Caldera is situated on the western edge of North American basement, as indicated by the 0.706 line, whereas the volcanic fields to the south are well within North American lithosphere (Fig. 1).
The numerous trachydacite domes of Mammoth Mountain and nearly all the mafic lavas peripheral to it are mildly alkaline, in contrast to the subalkaline rhyolites of Long Valley, Glass Mountain, and Mono Craters. Fundamentally, the distinction probably arises from smaller melt fractions extracted from mantle domains during basaltic magma generation beneath the Mammoth system in the past 200 k.y. The Long Valley rhyolitic magma output exceeded that of the Mammoth system by two orders of magnitude. Consequently, it can be inferred that mantle melt fractions were greater (thus less alkalic) and/or that basaltic magma flux from the mantle beneath Long Valley was so much larger (and longer sustained) that deep-crustal partial melts contributed a major proportion of the magma that ultimately evolved into the erupted rhyolites. The deep-crustal rocks beneath Long Valley are likely to be predominantly Mesozoic arc-intrusive suites and their differentiates and cumulates, and are thus subalkaline.
The westward progression of the magmatic focus from Glass Mountain (2.2–0.78 Ma) to the caldera (0.76 Ma) to the early rhyolites (0.76–0.65 Ma) entailed generation of subalkaline magmas in enormous volumes, the erupted rhyolites of which are merely some modest fraction. After 200 ka, when the next westward advance of the magmatic focus took place, magma generation diminished greatly in volume, mantle melt fractions became smaller (and thus more alkaline), and crustal melt contributions to the Mammoth system dacites became small relative to those that had contributed to the Long Valley rhyolites (cf. Hildreth, 2004).
Geometry of Fault Blocks in Walker Lane
Previous models have focused on Long Valley forming in a releasing bend between two normal faults along the Sierran range front (e.g., Bursik, 2009). Although consistent with faulting immediately to the north and south of Long Valley, this model ignores the greater significance that Long Valley is situated along a left step in the oblique-slip (right lateral) Sierran range front fault system. Therefore, based on kinematics in a left-stepping, right-lateral fault system, Long Valley should be in a constraining bend, not releasing bend, if the local fault patterns are assumed.
We utilize a simplified fault block model to explain the 3 Ma–200 ka tectonics near Long Valley resulting from transtensional deformation. This approach follows directly from other fault block models used successfully in California (e.g., Dickinson, 1996). An important aspect of the geometry of the Long Valley volcanic field involves its location. The right-lateral deformation steps ∼100 km to the east near Long Valley, from the Eastern California shear zone to Walker Lane along the sinistral Mina deflection (Ryall and Priestly, 1975; Hill, 2006; Oldow et al., 2008).
We define the following regions of interest, within a fixed North American plate reference frame (see Figs. 6A, 6C): (1) the region north of the Mina deflection; (2) the White-Inyo block; (3) crustal blocks within the Mina deflection; (4) the Adobe block; (5) the Owens Valley block; and (6) the Sierra Nevada block (Fig. 6). At the onset of our model, we adopt a simplified geometry of blocks that may have existed ca. 3–4 Ma when the Mina deflection initiated, and evolve the model based on the known fault kinematics and global positioning system (GPS) velocities of the region (Figs. 3 and 7A). All blocks, aside from those within the Mina deflection, are treated as rigid. The absolute shape and size of the blocks has been simplified in Figure 7, in order to provide a straightforward presentation.
Three blocks are assigned velocities with respect to a fixed North American plate. The White Mountain block is separated from a fixed North American plate by the Fish Lake Valley fault, which has estimates of slip rates predominantly between 4 and 10 mm/yr (Reheis and Sawyer, 1997; Dixon et al., 2000; Frankel et al., 2007). We use the Dixon et al. (2000) slip rate estimate of 8 mm/yr on the Fish Lake Valley fault to define the movement of the White Mountain block relative to a fixed North America. The Owens Valley region is translating northwest at 9.5 mm/yr and the Sierra Nevada block is translating northwest at 11.5 mm/yr, using estimates from GPS modeling (Dixon et al., 2000) and offset markers (Reheis and Sawyer, 1997; Oldow et al., 2001; Frankel et al., 2007). The movement and velocity of all other blocks are dependent on the three assigned velocities (Fig. 6B).
Using the assigned velocities, we observe the following kinematic relations between the blocks and between the blocks and a fixed North America as the system evolves (Fig. 6C). To the south of the Mina deflection, the White-Inyo block translates northward along the Fish Lake Valley–Furnace Creek fault system. The Owens Valley block translates northward slightly faster than the White-Inyo block, and slightly more westward, leading to oblique dextral faulting between the two blocks (Owens Valley fault).
White-Inyo Range as a Regional Indentor: Rotation of the Mina Deflection Blocks
The largest blocks in fault block models typically control the movement of the smaller blocks. Within the transtensional Walker Lane belt, the largest block is the Sierra Nevada microplate. The northwestern movement of the Sierra Nevada block, in isolation, would not create any tectonic voids in the vicinity of Long Valley. However, the interaction of the blocks in our model produces space for the formation of Long Valley Caldera. We hypothesize that the White-Inyo block acts as an indentor as it translates northward. The northward translation impinges on the crustal blocks of the Mina deflection, causing the crustal blocks to rotate clockwise, and sinistral faulting to occur between the blocks (Fig. 6C). A similar model of impingement was hypothesized by Lee et al. (2009) for the northern end of Owens Valley. Because the eastern end of the Mina deflection is pinned, it acts as a fulcrum for vertical axis rotation (pole of rotation). A simplified model is shown in Figure 6B. The pinned end of a line acts as an Euler pole (fulcrum in Fig. 6B). A critical point is that the White-Inyo Mountains are not on the western edge of the Mina deflection blocks, but rather impinge on the center of the blocks. If an indentor, such as the White-Inyo Mountains, pushes on the central portion of the line with some velocity, points farther from the fulcrum will move at a higher velocity than points near the indentor (Fig. 6B).
Figure 6C explores the consequence of this rotation on the rest of the model. The Adobe block, attached to the western edge of the fault blocks of the Mina deflection, swings northward with the western edge of those fault blocks. The Owens Valley block also translates north, but not as fast as the Adobe block. This differential motion creates a gap south of the Adobe block (northwest of the Owens Valley block) and southwest of the Mina deflection (Fig. 6C). If the Sierra Nevada block were moving with a velocity subparallel to the Owens Valley region, it could potentially fill the void as a result of having a westward jog in the Sierra Nevada range front at Mammoth Lakes. The Sierra Nevada block, however, has a greater westward component to its velocity, leading to normal faulting between it and the region immediately east of it (e.g., Adobe block, Owens Valley block).
Model Assumption and Predictions
The model presented in this study predicts that a gap opens between rigid blocks at the locality of Long Valley Caldera, based solely on the post-3 Ma kinematics of the area. We therefore suggest that the rotation of the Mina deflection blocks caused by northward translation of the White-Inyo block opened a space for magma ascent and intracrustal accumulation associated with Long Valley Caldera. The kinematics of the Mina deflection slip transfer, that not present to the south, facilitated magmatism at Long Valley, leading to far more voluminous eruptions relative to the volcanic fields to the south (e.g., Coso, Big Pine).
Our model accommodates the observed movement along most known faults (e.g., Owens Valley fault, Fish Valley–Furnace Creek fault, sinistral faults within the Mina deflection, Sierra Nevada frontal faults). A possible problem exists, however, between relative movement of the Adobe block and Sierra Nevada blocks. During Long Valley magmatism, the model predicts extension with a minor component of sinistral motion. The sinistral motion is required because the northward velocity of the Adobe block, due to the rotation of fault blocks in the Mina deflection, is greater than the northward velocity of the Sierra Nevada block; this difference effectively opens the tectonic gap. There are no reports, however, of sinistral movement along the eastern range front of the Sierra Nevada, adjacent to the Adobe block, after 4 Ma.
Several possibilities may account for this discrepancy. First, we may be overestimating the velocity due to rotation from the impinging White-Inyo block. The overestimation could result from an inexact east-west location of the White-Inyo block with respect to the Mina deflection, or the initial geometry of the system. Second, there may be some internal deformation occurring within the Adobe block, decreasing the differential motion between the Adobe block and the Sierra Nevada block. There are no geologic constraints on the strike-slip component in any of the fault zones within the Adobe block (e.g., Bald Mountain, Bald Mountain–Clover patch, and Benton Range fault zones). Third, sinistral offset along the range front may not be reported, because (1) it is so contrary to expectations in an overall right-lateral system, and/or (2) left-lateral slip occurs as an oblique-slip component on range-bounding faults. We used conservative velocity estimates to evolve the model, although we acknowledge that elastic strain accumulation makes it difficult to determine long-term geologic rates from geodetic signals. Assigning a larger velocity to the Sierra Nevada block could eliminate this discrepancy.
Initiation of a New Plate Boundary
The current magmatism in the Long Valley area can be reasonably considered as part of a ridge-transform-ridge system (Mono-Inyo Craters–South Moat fault–north-northeast–oriented inferred dike swarms) (Fig. 5). Dikes effectively accommodate a significant proportion of the regional extension; a similar conclusion was reached by Parsons et al. (1998)–while working in the Snake River Plain in southern Idaho. This ridge-transform-ridge model explains many puzzling aspects about the region. First, as noted by Bursik and Sieh (1989), it explains the lack of postglacial offset of faults along the Sierra Nevada range front north of Mammoth Mountain. Second, it explains the presence of the right-lateral South Moat fault in an east-west orientation that is kinematically highly improbable within an overall right-lateral plate boundary (as defined by the overall orientation of Walker Lane). Further, it explains the limited areal extent of this fault and why only the south margin of the caldera is reactivated.
The ridge-transform-ridge model explains the earthquake swarms that are frequent southeast of Long Valley Caldera. Focal mechanisms and stress inversions from the earthquake data suggest that in this area, the minimum horizontal stress is approximately north-northeast–southwest, or nearly perpendicular to the regional direction of northwest-southeast (Prejean et al., 2002). Prejean et al. (2002) modeled these data, and concluded that dextral faulting in the South Moat fault zone and normal faulting to the south can explain the northeast-southwest minimum horizontal stress directions. Although faulting may accommodate some of the deformation in this area, some of the earthquakes are likely a product of magma or magmatically derived fluids (Moos and Zoback, 1993; Prejean et al., 2002). Characterizing these earthquakes solely as a result of fault movement is problematic because of the presence of some non-double-couple seismic moment tensor solutions; the lack of ground rupture; the high-frequency, linear clustering; and temporal swarming, which is more indicative of fluid and/or dike related earthquakes. Our model, which does not disagree with the conclusion of Prejean et al. (2002) that faulting accommodates some of the local deformation, helps rectify these observations with the Prejean et al. (2002) conclusions. We suggest that a combination of normal faulting and volume expansion due to fluid and/or magma migration is a product of the ridge-transform-ridge tectonic setting.
The dike-dominated (ridge) segments do not appear to be affected by preexisting structures. To the north, the Mono-Inyo Craters chain is not parallel to the Sierra Nevada range front or to the Cretaceous Gem Lake shear zone. To the south, the inferred dike swarm does not intrude any fault or shear zones associated with Mesozoic deformation in the Sierran block (e.g., Greene et al., 1997; Tikoff and Saint Blanquat, 1997). In both areas, dike emplacement accommodates the motion of the Sierra Nevada block relative to the east side of the Sierra (Adobe and Owens Valley blocks).
The ridge-transform-ridge system is also diagnostic of rifted margins, and suggests that this zone represents the incipient rifting of the Sierra Nevada block from North America. This geometry of rifting is known to be inherited from continental breakup (i.e., the rupture initiates when the lithosphere is continental), as observed in the failed mid-continental rift in the upper midwestern United States (e.g., Chandler et al., 1989) and the East African Rift (e.g., Scott et al., 1992).
Changing Tectonic Models through Time
We have proposed two distinct tectonic models for the development of magmatism in the Long Valley, distinguished by different time periods (Fig. 7). The fault block model is appropriate in this region until ca. 200 ka (pre–Mammoth Mountain magmatism), and strike-slip motion still occurs in the Walker Lane region. Previous models for increased magmatic activity along the eastern Sierra Nevada have suggested that magmatism is a product of regional extension and delamination (e.g., Manley et al., 2000). The area of delamination suggested by xenoliths (Ducea and Saleeby, 1996, 1998) corresponds with areas of volcanic fields younger than 4 Ma, such as Long Valley, Coso, and Big Pine (Manley et al., 2000; Zandt, 2003). However, these regional models fail to address why the Long Valley volcanic field is orders of magnitude more voluminous than the other synchronous volcanic fields.
Incorporating our two tectonic models into previous studies, we note that over the past ∼12 m.y., there are three distinct styles of how extension has been created and accommodated near Long Valley. The first phase (12–3 Ma; Fig. 7A) accommodates the slip transfer at the right step through the formation of an extensional complex (Silver Peak–Lone Mountain) between the two stepped strike-slip faults (e.g., Oldow et al., 2008). The second phase (3–0.2 Ma; Fig. 7B) accommodates the slip transfer through block rotation within the Mina deflection, creating pronounced extension at the present-day location of Long Valley Caldera. The third phase (after 0.2 Ma–present; Fig. 7C) consists of a partitioned system, with regional extension being accommodated through normal faulting and dike intrusion along a ridge-transform-ridge system. In addition, we note the westward of magmatism since ca. 2.2 Ma, with a present focus at the Mono-Inyo volcanic chain and Mammoth Mountain, or even farther west at Red Cones (Hildreth, 2004).
Over time, magmatic activity changes from large, distributed magmatism (3–0.2 Ma) to focused, linear trends of magmatism (0.2–present), allowing relative motion to be accommodated by dike injection rather than faulting. This is a key difference in how magmatism and transtensional deformation are linked between the second and third phases. In the third phase, focused magmatic activity acts to accommodate regional transtension; in the second phase distributed magmatic activity does not accommodate regional transtension, but rather fills extended areas that do not reflect relative plate motions.
We propose that the distinct magmatic suites of Long Valley Caldera, Mammoth Mountain, and the Mono-Inyo Craters are all tectonically controlled, associated with ongoing transtensional deformation. The Mono-Inyo Craters, Mammoth Mountain magmatism, South Moat fault zone, and dike injection inferred from non-double-couple earthquakes south of Mammoth Mountain form the equivalent of a ridge-transform-ridge system. This geometry represents the incipient rifting of the Sierra Nevada block from North America, facilitated by the magmatism associated with delamination.
In contrast, the older Long Valley system formed in a major zone of focused local extension associated with fault block rotation in the central Walker Lane belt. Activation of the sinistral faults of the Mina deflection resulted in clockwise block rotations within the dextral Walker Lane system. We hypothesize that the fault movement in the Mina deflection is driven by the White-Inyo Mountains, which act as a local indentor. The rotation of the fault blocks of the Mina deflection creates a tectonic gap at Long Valley, providing room for the development of the large Long Valley magma reservoir.
The idea for this paper originated on Lookout Mountain in Long Valley Caldera during the “Assembly of Plutons” field forum in 2005 led by J. Bartley, D. Coleman, and A. Glazner. Seth Kruckenberg contributed important advice on figures. We thank an anonymous reviewer, David Hill, and Stephanie Prejean for helpful reviews, and C. Busby for editorial assistance. Tikoff was partially supported by the National Science Foundation Earth Sciences division.