We integrate new geochronological, petrographic, and geochemical data with previously published (Sierra Crest–Little Walker volcanic center) and new (Ebbetts Pass volcanic center) structural and stratigraphic data to describe two deeply dissected, spectacularly well-exposed Walker Lane pull-apart basins in the ancestral Cascades arc. The Miocene (ca. 12–5 Ma) Sierra Crest–Little Walker arc volcanic center and the Miocene–Pliocene (ca. 6–4.6 Ma) Ebbetts Pass arc volcanic center formed in pull-apart basins in the Walker Lane, a NNW-trending zone of dextral strike-slip and oblique normal faults at the western edge of the Basin and Range Province. The Sierra Crest–Little Walker arc volcanic center is a transtensional arc volcanic field that is as areally extensive (∼4000 km2) and as long-lived (∼3 m.y.) as the Pliocene to Holocene Long Valley transtensional rift volcanic field, which is also in the Walker Lane. The Ebbetts Pass volcanic center is the next major volcanic center to the north within the ancestral Cascades arc. It formed within a smaller pull-apart basin. Its estimated original volume (∼270 km3) is comparable to the estimated volume for the Lassen volcanic center in the last 825 k.y. (200 km3), a major magmatic focus in the northernmost Walker Lane pull-apart basin. Postvolcanic faulting and erosion provide spectacular three-dimensional time-integrated views of the structure and stratigraphy of the Miocene to Pliocene volcanic centers described herein. The volcanic stratigraphy of this region is divided into four unconformity-bounded sequences: (1) Oligocene silicic ignimbrites erupted in central Nevada (Valley Springs Formation), (2) Miocene arc volcanic and volcaniclastic rocks (Relief Peak Formation), (3) ca. 11.5–9 Ma high-K arc volcanic rocks (Stanislaus Group), and (4) Miocene–Pliocene arc volcanic rocks (Disaster Peak Formation). We present 19 new 40Ar/39Ar ages and integrate them with previous 40Ar/39Ar ages to erect a detailed stratigraphic and intrusive chronology. For the first time, we present a systematic description of all of the volcanic and intrusive map units, using outcrop photos, photomicrographs, modal analyses, and geochemical data.
Geochemistry and modal analyses show that volcanism at the younger, smaller Ebbetts Pass volcanic center is more evolved than that of the Sierra Crest–Little Walker volcanic center. The Sierra Crest–Little Walker volcanic center is dominated by Stanislaus Group basalt, trachybasaltic andesite, trachyandesite, and andesite. Climactic eruptions at its southern end produced the Little Walker caldera and its Stanislaus Group trachydacite ignimbrites. In contrast, the Ebbetts Pass volcanic center erupted much less basalt, and an approximately equal mixture of basaltic andesite, andesite, dacite, and rhyolite. By virtue of its higher SiO2, on average, it shows geochemical evidence for higher extents of crustal contamination. All of the volcanic rocks except the Stanislaus Group are compositionally identical to Mesozoic plutonic rocks of the Sierra Nevada Batholith, indicating a similar petrogenesis by shallow-level fractional crystallization with no evidence for deep-seated pyroxenite cumulates. Only the high-K2O Stanislaus Group requires a significant degree of fractionation of clinopyroxene and Fe-Ti oxides. The major-oxide trends and clinopyroxene-rich modes of the Stanislaus Group may be the volcanic expression of processes that produced pyroxenite cumulates. The basal and most voluminous unit of the Stanislaus Group, the Table Mountain Latite, is a likely parent to succeeding units of the Stanislaus Group (Eureka Valley Tuff and Dardanelles Formation), but this particular liquid line of descent is otherwise rare in the Mesozoic and Cenozoic history of the Sierra Nevada.
Volcano-tectonic activity continued for a longer period (ca. 12–6 Ma) in the northern half of the Sierra Crest–Little Walker volcanic center than it did in the southern half (ca. 12–9 Ma), overlapping for ∼1 m.y. with the onset of volcano-tectonic activity at the Ebbetts Pass volcanic center to the north. This provides local-scale confirmation of the regional-scale interpretation that a transtensional rift tip propagated northward with time within the arc axis, in concert with northward migration of the Mendocino triple junction. Transtensional rift magmatism followed in its wake and continues today at various points along the length of the Walker Lane. “Tectono-stratigraphic recycling” was a key geologic process throughout the development of these ancestral Cascades arc pull-apart basins, and it consisted of the transfer of megaslide slabs, up to 2 km long and tens to hundreds of meters thick, from footwall to hanging-wall blocks in the transtensional rift. A time-slice series of block diagrams is used to illustrate the structural controls on arc volcanism in the early stages of Walker Lane transtensional faulting (ca. 12–4.6 Ma).
The Sierra Nevada (California) forms an asymmetrical block-faulted mountain range with a gentle western slope and a steep fault-bounded range front on the east. The steep eastern range front was previously interpreted to be the result of Basin and Range extension (Bateman and Wahraftig, 1966; Slemmons, 1953, 1966). However, it is now known to form the western boundary of the Walker Lane, an ∼100-km-wide, NNW-trending zone of dextral strike-slip and oblique normal faults that lies along the western edge of the Basin and Range Province (e.g., Faulds and Henry, 2008; Jayko and Bursik, 2012; Busby, 2013). The Walker Lane accommodates ∼20%–30% of the right-lateral motion between the Pacific and North American plates (Argus and Gordon, 1991; Dixon et al., 2000; Oldow, 2003; Unruh et al., 2003; John et al., 2012). This region has been described as the northernmost extension of the Gulf of California transtensional rift, where the process of continental rupture has not yet been completed, and rift initiation can be studied on land (Faulds and Henry, 2008; Jayko and Bursik, 2012; Busby, 2013). The northern and southern segments of the Walker Lane are dominated by translation along long NNW-trending dextral strike-slip faults (Faulds and Henry, 2008; Jayko and Bursik, 2012; Busby, 2013). In contrast, the central Walker Lane (Fig. 1) shows evidence of distributed dextral shear accommodated by oblique-normal faults, block rotations, and partitioning of oblique deformation between subparallel normal and strike-slip faults (e.g., Wesnousky, 2005; Dong et al., 2014; Bormann et al., 2016).
Faults of the central Walker Lane strike approximately NE, as well as approximately N-S (Fig. 1). The Carson domain, along the northern boundary of the central Walker Lane, has numerous ENE-striking faults with oblique sinistral-normal slip, and clockwise block rotations of up to 44°–90° measured on volcanic rocks (Cashman and Fontaine, 2000; Faulds and Henry, 2008; Kreemer et al., 2009). The Mina deflection, along the southern margin of the central Walker Lane, is a regional-scale releasing bend consisting of ENE-striking sinistral-oblique, sinistral, and normal faults (Ryall and Priestly, 1975; Oldow, 1992, 2003; Wesnousky, 2005; Faulds and Henry, 2008; Oldow et al., 2008; Lee et al., 2009a, 2009b; Jayko and Bursik, 2012). Series of basins have opened up along the Mina deflection from the middle Miocene to the present day (Oldow et al., 1994, 2001, 2009; Stockli et al., 2003; Lee et al., 2009a; Ferranti et al., 2009; Tincher and Stockli, 2009). Like the Carson domain, the Mina deflection shows paleomagnetic evidence of clockwise vertical-axis block rotations (Petronis et al., 2004, 2007, 2009, 2016; Gledhill et al., 2016). Clockwise block rotations of up to 55° also occur in the region between the Carson domain and the Mina deflection (Carlson et al., 2013; Carlson and Faulds, 2015, 2016). This includes clockwise rotations of ∼18° in volcanic rocks on the Sierran frontal fault zone just north of the Little Walker caldera (Figs. 1 and 2A), described herein and by Pluhar et al. (2009) and Carlson et al. (2013). Geodetically determined rates of rotation (Bormann et al., 2016) appear to underestimate the contribution of vertical-axis rotation determined by paleomagnetic studies, perhaps because fault-bounded, rotated blocks are significantly smaller than the regional-scale blocks presented in geodetic models (Carlson, 2017). Additionally, the importance of vertical-axis rotations along the western boundary of the central Walker Lane has been severely underevaluated (Carlson, 2017); the stratigraphic and structural results described herein are ideal for future studies on this.
The Carson domain and Mina deflection represent very large right steps in the central Walker Lane, ∼60 km and 90 km, respectively, with vertical-axis rotations described above. In this paper, we describe smaller right steps along the western edge of the central Walker Lane (Fig. 1 and 2A): an ∼15 km right step, referred to herein as the Sierra Crest–Little Walker pull-apart, and an ∼5 km zone of right-stepping faults, referred to herein as the Ebbetts Pass pull-apart. This paper describes Oligocene through Pliocene volcanic sections that will be appropriate for much-needed paleomagnetic assessment of the importance of vertical-axis rotations. In this paper, we demonstrate the controls of these faults on two large volcanic centers that formed in the pull-apart basins.
The central Walker Lane (Fig. 1) has more abundant Cenozoic volcanic rocks than the northern and southern Walker Lane, probably because it lies at a higher rift angle relative to the overall plate margin (Busby, 2013). The abundant volcanic rocks make the central Walker Lane ideal for detailed reconstructions of transtensional rifting and magmatism in a continental setting, because the volcanic strata are dateable by the 40Ar/39Ar method (e.g., Busby et al., 2008a, 2008b, 2013a, 2013b, 2016; Hagan et al., 2009; Nagorsen-Rinke et al., 2013; Fleck et al., 2015). Our studies have focused on the western margin of the central Walker Lane along the central Sierra Nevada range front and range crest (Fig. 1). There, Cenozoic volcanic and volcaniclastic rocks are extensive and are spectacularly well exposed in three dimensions over rugged topography with high structural relief. This provides the opportunity to study the structural controls on volcano-tectonic basins, vents, and plumbing systems as they relate to preserved eruptive units.
In this paper, we describe two volcanic centers sited in ancestral Cascades arc pull-apart basins in the central Walker Lane: (1) the ca. 12–9 Ma Sierra Crest–Little Walker volcanic center, and (2) the ca. 6–4.6 Ebbetts Pass volcanic center. The Miocene (ca. 12–9 Ma) Sierra Crest–Little Walker volcanic center is a transtensional volcanic system, and it is as areally extensive (∼4000 km2) and long-lived (∼3 m.y.) as the Pliocene to recent Long Valley transtensional rift volcanic center (Busby, 2012; Busby et al., 2013a, 2013b). The Sierra Crest–Little Walker volcanic center also forms the largest volcanic center recognized to date in the ancestral Cascades arc, and it records the birth of the Walker Lane “future plate boundary” within the arc at ca. 12–9 Ma (Busby, 2013). The Ebbetts Pass volcanic center is the next major ancestral Cascades arc volcanic center to the north, and we show herein that it formed within a smaller Walker Lane pull-apart basin in Miocene to Pliocene time. An estimated original volume of ∼270 km3 for the Ebbetts Pass volcanic center is comparable to that estimated by Hildreth (2007) for the Lassen transtensional arc volcanic center in the last 825 k.y. (200 km3). The main difference between the Miocene to Pliocene volcanic fields described herein and the more well-studied, active volcanic fields of the Walker Lane is that the Miocene to Pliocene volcanic fields are dissected to the ideal level for viewing the structural controls on volcanism and sedimentation over a period of ∼8 m.y. For the first time, this paper presents photomicrographs and modal analyses on rocks of the Sierra Crest–Little Walker volcanic center and integrates these with whole-rock geochemical data. We also present a new map, and new structural, stratigraphic, petrographic, and geochronologic data on the Ebbetts Pass volcanic center. These are integrated into time-slice block diagrams that describe the evolution of the central Sierra segment of the ancestral Cascades arc as a series of northward-migrating, transtensional, intra-arc pull-apart basins in the Walker Lane.
Our geochemical data show that arc volcanism was most primitive where transtension was areally extensive, in the Sierra Crest–Little Walker volcanic center, and of largest volume, in the southern part of that center. A higher degree of crustal contamination at the Ebbetts Pass pull-apart is attributed to its much smaller size, with probable lower extension rates on faults that may not have all penetrated the lithosphere, leading to lower rates of basalt intrusion. Magmatism and transtension gradually migrated northward in time from the southern half of the Sierra Crest–Little Walker volcanic center into its northern half, and then northward into the Ebbetts Pass volcanic center, as part of a regional-scale northward migration of the rift tip, in concert with northward migration of the Mendocino triple junction.
STRATIGRAPHIC FRAMEWORK FROM PALEOCHANNEL FILLS
Prior to our mapping shown in Figures 2A and 2B, the Cenozoic stratigraphy of the region described herein was best known from paleochannel fills in the unfaulted part of the Sierra Nevada west of the crest (Fig. 3). This was described in detail by Busby et al. (2016), and we refer the reader to dozens of references to previous work cited there. In this section, we provide a brief review of Sierran paleochannels, and we describe the relatively simple stratigraphy preserved in largely unfaulted paleochannels of the region west of the pull-apart basins (Fig. 2A). This provides a basis for describing the much more complex and faulted stratigraphy in the two volcanic centers described herein.
Paleochannels of the Sierra Nevada were first exploited for placer gold deposits during the California Gold Rush that began in 1848. Then, more than a century ago, they were mapped in surveys that showed that they flowed westward, like the modern rivers of the range (Ransome, 1898; Lindgren, 1911). These studies assumed that the heads of the paleochannels lay at the modern range crest. A first paradigm shift occurred about a half century ago, when it was recognized that at least some of the paleochannel fill was sourced from Nevada, showing that the range was younger than the paleochannels (younger than 6 Ma; Slemmons, 1953; Curtis, 1954; Bateman and Wahraftig, 1966; cf. Wakabayashi, 2013). More recent work has demonstrated that Sierran paleochannels are ancient features that were carved into the shoulder of a broad high uplift formed during Late Cretaceous to Paleocene crustal shortening (the Nevadaplano). Their headwaters lay in central Nevada prior to disruption of the Nevadaplano by Basin and Range extension (cf. Henry, 2008; Henry et al., 2012).
Sierran paleochannels in the area described here (Fig. 1) contain four unconformity-bounded stratigraphic sequences (Fig. 3). Sequence 1 consists of Oligocene silicic ignimbrites erupted from arc calderas in central Nevada (Valley Springs Formation). Sequence 2 consists of andesitic arc volcaniclastic and volcanic rocks (Relief Peak Formation). Sequence 3 consists of distinctive high-K2O arc lavas and ignimbrites (Stanislaus Group). These include Table Mountain Latite mafic to intermediate lavas, and the Eureka Valley Tuff, with three trachydacite ignimbrite members, and two lava members that range from basalt through trachydacite. Sequence 4 records a return to andesitic arc magmatism (Disaster Peak Formation). The Relief Peak Formation is defined as andesitic rocks underlying or older than the high-K2O Stanislaus Group, and the Disaster Peak Formation is defined as andesitic rocks overlying or younger than Stanislaus Group (Slemmons, 1953, 1966; Keith et al., 1982). This division is only feasible in the area containing Stanislaus Group, from the Sonora Pass region to the Ebbetts Pass region (Fig. 2).
As described in detail by Busby et al. (2016), three paleochannels crossed what is now the Sonora Pass to Ebbetts Pass region prior to their beheading or derangement by faults and volcanoes of the Sierra Crest–Little Walker volcanic center and the Ebbetts Pass volcanic center (Fig. 4). This paper focuses on volcanic and intrusive rocks of the Sierra Crest–Little Walker volcanic center and the Ebbetts Pass volcanic center, and their structural framework.
Cenozoic faults are herein identified and mapped by offsets in the Cenozoic volcanic rocks, or offsets of the contacts between volcanic rocks and underlying Mesozoic basement (Figs. 2, 4A, and 4B). Many of the faults described here were first mapped by Slemmons (1953), Keith et al. (1982), and Armin et al. (1984) on the basis of these features. However, the synvolcanic slip history of the faults described here was previously unrecognized because many of the faults were reactivated after volcanism ceased, cutting all volcanic units (Figs. 2A and 2B). Detailed volcanic lithofacies mapping was required to demonstrate synvolcanic faulting on the basis of stratal relations, such as abrupt thinning of strata onto footwalls, fanning dips, and lateral offsets across piercing points that die out up section (Busby et al., 2008a, 2013a, 2013b). In addition, Slemmons (1953, and unpublished mapping) inferred that straight canyons trending approximately N-S or NE-SW were also fault-controlled, even if they only occurred in basement rocks and were not constrained by offsets in volcanic strata. Some of those faults are also shown on Figure 2, but they are dashed and queried. Kinematic indicators are very few, because fault planes are only well exposed on the most resistant volcanic units, which are intrusions. Where faults pass from more resistant volcanic rocks into softer granitic basement, they are buried beneath stream boulders or forested areas. However, they follow straight traces and commonly reemerge to offset strata or volcanic-basement contacts along volcanic ridges.
Sierra Crest–Little Walker Volcanic Center
The structure of the Sierra Crest–Little Walker volcanic center was described in detail by Busby et al. (2013a, 2013b), so only a brief summary is provided here. Our previous work was the first to show evidence of synvolcanic faults and their controls on vents and ponded volcanic sections, in what we term the Sierra Crest–Little Walker volcanic center. The Stanislaus Group (Fig. 3) contains distinctive widespread units that act as regional strain markers, making it much easier to document the importance and nature of synvolcanic faulting.
The Sierra Crest–Little Walker volcanic center is dominated by the Sierra Crest graben-vent system (Fig. 4A; Busby et al., 2013a, 2013b). This has a very large full graben (the Sierra Crest graben) bounded by NNW-striking oblique normal-dextral faults, and a major NE-striking transfer zone that emanates from its northern half, with dominantly sinistral oblique normal faults (Fig. 4B). Synvolcanic transtensional faults ponded high-K2O lavas in grabens to thicknesses of >400 m. Vents occur along these faults as fissures and lesser fault-controlled point sources (red stars, Fig. 4A). These faults include (Figs. 2, 4A, and 4B):
(1) NNW-SSE faults that bound the 28-km-long, 8–10-km-wide Sierra Crest full graben. These include the East Fork Carson fault on the east and the Kennedy Creek–Seven Pines–Disaster Creek faults on the west. Vents lie in the hanging walls near the surface trace of these faults. This group also includes NE-SW–striking faults that accommodate right steps in the NNW-SSE graben-bounding faults.
(2) Series of small NNW-SSE half grabens on the west side of the Sierra Crest graben, including the Red Peak, Bald Peak, and Arnot Creek faults.
(3) The NE-SW transfer zone faults and grabens that extend ∼20 km along the northeast side of the Sierra Crest pull-apart basin, including the Poison Flat and Mineral Mountain fault zones. This zone is inferred to transfer transtensional slip from the Sierra Crest graben to a fault zone that lay east and northeast of the graben (Fig. 4B), the Antelope Valley fault (Fig. 1).
Ebbetts Pass Volcanic Center
The Ebbetts Pass volcanic center formed at a much smaller pull-apart basin than that of the Sierra Crest–Little Walker volcanic center (Figs. 4A and 4B). The basin is bounded to the west and northwest by a series of right-stepping, N-S–trending, dextral-oblique, down-to-the-east normal faults that lie on the modern Sierra crest and range front (Noble Canyon, Grover Hot Springs, and Silver Mountain faults; Figs. 2B and 4). It is bounded to the southeast and south by the down-to-the-northwest Wolf Creek fault (Figs. 2B and 4). Mesozoic basement is successively dropped down to the east across the right-stepping oblique faults, and silver mines are located along them (Fig. 2B), similar to the normal fault-localized silver deposits immediately to the east and southeast (John, 2001). Our mapping of these faults is new, so a description of each fault is given here.
Noble Canyon Fault
The Noble Canyon fault (Figs. 2B and 4) forms a dramatic 410 m (1340 ft) topographic escarpment in granitic rock on the west side of Noble Canyon, dropping volcanic rocks down on the east side of the fault. It was previously mapped by Keith et al. (1982) and Armin et al. (1984). Although it clearly cuts all of the volcanic rocks south of Highway 4 (Fig. 2B), we infer that it is also synvolcanic because the contact between granitic basement and overlying volcanic rock is dropped at least 120 m (400 ft) lower on the east side of the fault (Fig. 2B). The fault strikes 330°–350°, dips 75°–90° east, and is more than 12 km long. The fault ends northward in a right step to the Grover Hot Springs fault (Figs. 2 and 4). It ends southward at the Wolf Creek fault, but it connects to the Disaster Creek fault, which together with the Arnot Creek fault forms the western boundary of the northern Sierra Crest graben (Figs. 2 and 4). The Disaster Creek fault connects southward to the Seven Pines fault, which in turn transfers westward to the Kennedy Creek fault; these form the western boundary of the southern Sierra Crest graben (Figs. 2A and 4). As discussed in the following, the faults on the west boundary of the Sierra Crest graben became active during deposition of the Relief Peak Formation, but there is no evidence for activity along the Noble Canyon until late in the deposition of the Disaster Peak Formation.
Grover Hot Springs Fault
The Grover Hot Springs fault is a long fault that extends 15 km northward past the trace shown on Figure 4 (Armin et al., 1984). It dips steeply east and strikes 340°, dropping undifferentiated andesitic volcaniclastic rocks down to the east against the Valley Springs Formation, which in turn rests on granitic basement (Fig. 2B). The fault appears to end in the south near the contact between altered basal fill and overlying unaltered 4.6–4.5 Ma volcanic rocks of the Ebbetts Pass pull-apart structure. The next fault to the east (Silver Mountain fault) also dies out southward (Fig. 2B), so we infer that the faults stepped right in a releasing stepover (Fig. 4B).
Grover Hot Springs, the Silver Peak mine, and the Pennsylvania mine are all situated along the Grover Hot Springs fault. Because Mesozoic basement does not crop out on the hanging wall, we can only say that vertical separation along the fault is >180 m (600 ft). The fault zone is well exposed for 1 km north of Highway 4, where it forms a 5-m-wide gouge zone. Within 3 m of the fault zone, the Valley Springs Formation ignimbrite (which is normally flat lying) is rotated into the fault zone, with a strong vertical foliation striking 350°. The undifferentiated volcaniclastic rocks (Tvu, Fig. 2B) are strongly altered along the fault.
Silver Mountain Fault
The Silver Mountain fault is parallel to the Grover Hot Springs and Noble Canyon faults (striking 340°; Figs. 2B and 4), but it dips more shallowly (55°–60° east). Armin et al. (1984) mapped the Silver Mountain fault, but only south of Highway 4 (highway shown on Fig. 2). Multiple silver mines are sited along this fault, including the Rippon, Gould and Curry, IXL, Lady Franklin, and Exchequer mines. This fault also passes through the “Silver Mountain site,” the location of a mining town during the 1850–1870s (Curtis, 1954; Clark and Evans, 1977). The undifferentiated volcaniclastic rocks (Tvu; Fig. 2B) on both sides of the fault lack features that can be used to determine displacement.
Wolf Creek Fault
The Wolf Creek fault is a normal fault that separates the Ebbetts Pass pull-apart basin to the northwest from the Sierra Crest–Little Walker pull-apart basin to the south (Figs. 2 and 4). South of the Wolf Creek fault, there is a >1-km-thick section of Relief Peak Formation and Stanislaus Group (Fig. 2). North of the fault, altered andesitic volcaniclastic rocks form the basal fill of the Ebbetts Pass pull-apart (Tvu and Tva) basin, overlain by unaltered volcanic rocks of the lower silicic section (dacite lava [Tdpld], rhyolite welded ignimbrite [Tdpwi], and rhyolite lava [Tdplr]), and the Ebbetts Pass stratovolcano (Tdpeps; Fig. 2B).
The basal fill of the Ebbetts Pass pull-apart basin consists of undifferentiated and altered volcanic rocks (Tvu and Tva; Fig. 2B). It is a crumbly blue-gray volcanic-volcaniclastic rock that is too strongly altered to clays, micas, quartz, and pyrite to allow 40Ar/39Ar dating. Stratification is faintly visible. The basal fill was mapped as Relief Peak Formation (Trp) by Armin et al. (1984), but there is no evidence that it is older than the Stanislaus Group, since there are no dates on it, nor are there any stratigraphic relations with the Stanislaus Group. We infer that the basal fill is largely volcanic debris-flow deposits that were catastrophically deposited into the pull-apart basin as it began to form, and these were extensively altered by fluids during emplacement of the overlying unaltered volcanic section and its coeval intrusions. The basal fill is thus most likely Disaster Peak Formation (Tdp) dropped down against Relief Peak Formation along the Wolf Creek fault (Fig. 2B), although we show it as undifferentiated (Tvu, Tva) because it is undated.
The thick fill of the Sierra Crest graben-vent system ends abruptly against the Wolf Creek fault, suggesting it continues northward at depth beneath the Ebbetts Pass pull-apart basin (i.e., it was down dropped beneath it along the Wolf Creek fault). A splay of the Wolf Creek fault drops 4.6–4.5 Ma Ebbetts Pass stratovolcano basin fill down against ca. 6.4 Ma basin fill (Fig. 2B). This strand is shown as a queried normal fault that brings up a dacite lava in its footwall (Tdpld, dated at 6.367 ± 0.017 Ma; Fig. 5). The fault is partially buried by a younger rhyolite lava (Tdplr, dated at 4.73 ± 0.03 Ma; Fig. 5), where it is shown dotted through that unit. All strata north (basinward) of the queried fault are 4.6–4.5 Ma (Tdpeps, Ebbetts Pass stratovolcano; Fig. 4A) and are inferred to lie in the hanging wall of the queried normal fault strand. The fault strand is queried because the dacite lava (Tdpld) could alternatively represent a megaslide block, but its large size (>2.5 km long and >225 m thick) and the absence of associated smaller slide blocks make that seem unlikely.
Summary of Structural Framework
The Sierra Crest graben-vent system as a whole forms a very large right releasing stepover (Fig. 4B), with increasing transtensional fault offset southward (Busby et al., 2013b). This controlled the siting of the Little Walker caldera in the southeast corner of the volcanic field (Fig. 4A), which produced ignimbrites of the Eureka Valley Tuff (Fig. 3). Extreme transtension in the very large Sierra Crest–Little Walker pull-apart is inferred to have triggered rapid ascent of low-degree partial melts (Putirka and Busby, 2007), causing outpouring of high-K2O “andesite flood lavas” (Putirka and Busby, 2007; Busby et al., 2013b). The NE-striking fault zone in the northeast part of the Sierra Crest–Little Walker volcanic center mimics the right steps and NE-striking faults of the Carson domain and Mina deflection but on a smaller scale (Fig. 1). Three models have been proposed to explain fault-slip transfer across the Mina deflection, summarized by Nagorsen-Rinke et al. (2013): (1) normal faults at a dextral strike-slip fault stepover (e.g., see panel 2, Fig. 4B), (2) oblique-slip (sinistral and normal) faults, or (3) clockwise block rotations between sinistral faults (e.g., see Fig. 3; Nagorsen-Rinke et al., 2013). Model 2 applies to the transfer zone mapped herein (Fig. 4B, panel 3), but with faults that dip more steeply (80°–90°) than those portrayed in model 2 of Nagorsen-Rinke et al. (2013). We predict that future paleomagnetic work along the NE-striking fault zone in the Sierra Crest–Little Walker volcanic center will show evidence for clockwise block rotations, similar to those determined in the Carson domain and Mina deflection (described earlier herein). This area would be ideal for determining if block rotations are present, and whether or not they are progressive, because some structural blocks contain units that range from ca. 10.5 Ma (Table Mountain Latite), through ca. 9.54 Ma (Eureka Valley Tuff), to ca. 7.25 Ma (basalt lavas; Figs. 2B and 5).
The Ebbetts Pass pull-apart represents a much smaller right step than the Sierra Crest–Little Walker pull-apart, along a series of small stepovers (Fig. 4B). It differs from the simple pull-apart model shown in panel 2 of Figure 4B by having through-going normal faults along its southern boundary, and a volcanic fill that thins northward (Figs. 2 and 4B).
FIELD AND PETROGRAPHIC CHARACTERISTICS AND NEW 40Ar/39Ar GEOCHRONOLOGY
In this section, we present the first systematic description of all the volcanic and intrusive map units in the Sierra Crest–Little Walker and Ebbetts Pass volcanic centers, using outcrop photos, photomicrographs, modal analyses, and geochemical data, in conjunction with 40Ar/39Ar geochronology (Figs. 5–13). Previous papers have presented small geochemical data sets in stratigraphic context (Ransome, 1898; Noble et al., 1976; Putirka and Busby, 2007; Busby et al., 2008a, 2013b; Koerner et al., 2009), or used those data sets as part of a regional-scale analysis (Putirka et al., 2012; du Bray et al., 2014). However, this is the first paper to present a large geochemical data set in stratigraphic and intrusive context (Table 1), plotted on maps (Fig. 2) and located by global positioning system (GPS; see Supplemental Table S11). Although abundant field photographs and descriptions have been previously published (Slemmons, 1953, 1966; Noble et al., 1974; Busby et al., 2008a; Busby and Putirka, 2009; Gorny et al., 2009; Busby et al., 2013a, 2013b, 2016), this is the first paper to link outcrop photos (Figs. 6 and 10) with photomicrographs (Figs. 9 and 13) to document textures and mineralogy. Published modal analyses were previously restricted to a very small part of the area (Busby et al., 2013b), but they are herein presented for all map units across the two volcanic centers (Figs. 7, 8, 11, and 12; raw counts in Table S2 [footnote 1]).
We also report new 40Ar/39Ar geochronology on 19 samples, summarized in Figure 5, with plots presented in the Supplemental Files (footnote 1). These analyses were performed on mineral separates and matrix material from lavas and intrusions, and minerals from ignimbrites, using methods and facilities described previously (Busby et al., 2013b). Ages are based on 28.02 Ma for the Fish Canyon sanidine (FCs) standard (Renne et al., 1998) and the decay constants of Steiger and Jäger (1977), to facilitate direct comparison with previous studies (e.g., Fleck et al., 2015). More recent calibrations (e.g., Renne et al., 2011) yield slightly older ages, which we consider to be more accurate.
The relationships of volcanic and intrusive rocks to structures in the Sierra Crest–Little Walker and Ebbetts Pass pull-apart basins evolved with time. A series of block diagrams are provided in Figure 14 to show the evolution of the pull-apart basins and volcanic centers.
Sierra Crest–Little Walker Volcanic Center and Pull-Apart Basin
Relief Peak Formation
Figure 14A reconstructs the paleogeography of the region immediately prior to the onset of Walker Lane transtension, during deposition of Relief Peak Formation deposits in paleochannels. The Relief Peak Formation in paleochannels consists largely of stratified andesitic debris-flow and fluvial deposits, with lesser basaltic andesite to andesite block-and-ash-flow tuffs (Busby et al., 2016). Relief Peak Formation lavas in the paleochannel fills are restricted to: (1) a section of basalt lavas below The Dardanelles (samples DC001 and DC067; Table 1), and (2) a single basaltic trachyandesite lava in the Stanislaus paleochannel (sample 8–19–08; Table 1). Mafic lavas are fluid and could have been sourced from tens of kilometers to the east. Igneous rocks on Relief Peak interpreted by Roelofs (2004) to be Relief Peak Formation lavas instead are younger intrusions (Busby et al., 2016). For these reasons, we infer that arc front volcanoes lay upslope to the east (i.e., the study area lay in a forearc position), and paleochannels between the arc front and what is now the Sierra were not yet disrupted by faulting (Fig. 14A).
Relief Peak Formation rocks also occur as basal graben fill in the Sierra Crest–Little Walker pull-apart basin (Fig. 14B). The basal graben fill contrasts with the older paleochannel fill by consisting of mass-transport deposits and lacking fluvial deposits. Buried paleochannels are preserved below the graben fills in a few areas (Figs. 2 and 14B) but were mostly destroyed by faulting and avalanching. The Stanislaus paleochannel was beheaded by faulting, but the Cataract paleochannel, although faulted, was not dismembered at this time (Busby et al., 2016). The Mokelumne paleochannel lay north of the faults (Fig. 14B).
Relief Peak Formation debris-avalanche and debris-flow deposits occur as basal fill in all grabens in the southern half of the Sierra Crest–Little Walker pull-apart basin (Fig. 14B). They were redeposited from the two paleochannels that crossed that region, the Stanislaus and Cataract paleochannels (Figs. 4 and 14B). Relief Peak Formation debris-avalanche and debris-flow deposits are absent from the north part of the Sierra Crest graben (Fig. 2), because no paleochannels crossed it there (Figs. 14A and 14B).
The N-S grabens have no in situ primary volcanic rocks of Relief Peak Formation, only megaslide blocks. In contrast, the NE transfer zone basins along the Wolf Creek to Poison Flat fault zones (WCF and PFFZ in Fig. 14B) contain in situ, thick sections of massive (nonstratified) debris-flow deposits with in situ interstratified primary volcanic rocks, including andesite block-and-ash-flow tuff (sample JHEP-90) and andesite to basaltic andesite lavas (samples JHEP-89, MG-08-02, 920-5, 920-6, and 921-2; Table 1). We infer that the NE transfer zone basins acted as a funnel that extended toward the arc front, which still lay to the east during Relief Peak Formation time. This funnel captured primary eruptive products from the arc and rapidly filled the basins with massive volcanic debris-flow deposits.
The Relief Peak Formation primary volcanic rocks in the NE transfer zone basins differ from primary volcanic rocks in the younger Disaster Peak Formation by being dominated by clinopyroxene and orthopyroxene instead of hornblende (Fig. 7A). Representative photomicrographs of basaltic andesite lava show unaltered plagioclase (with sieve textures) and two pyroxenes in a glassy groundmass, with embayment and rimming of plagioclase, and clinopyroxene rimming orthopyroxene (Fig. 9A). One andesite block-and-ash-flow tuff yielded a 40Ar/39Ar hornblende age of 11.33 ± 0.03 Ma (Fig. 5), which is consistent with the age of the base of the overlying Table Mountain Latite (10.41 ± 0.08 Ma; Fig. 5).
No intrusions or vent facies deposits of Relief Peak Formation age have been found in the area depicted by Figure 14B, except for an andesite sill that intrudes graben fill near the southeastern margin of the Sierra Crest–Little Walker pull-apart (Fig. 14B; Busby et al., 2013b). This has a 40Ar/39Ar hornblende age of 12.15 ± 0.04 Ma (Fig. 5; Busby et al., 2013b), providing a minimum age for the onset of faulting in the Sierra Crest–Little Walker pull-apart. Avalanching continued up to the onset of eruption of Table Mountain Latite “flood andesites,” as shown by the overlap in ages (within error) between block-and-ash-flow tuff in the youngest megaslide block (10.39 ± 0.18 Ma) and the base of the Table Mountain Latite at Sonora Peak (10.41 ± 0.08 Ma; Fig. 5). Thus, the onset of Sierra Crest–Little Walker transtensional tectonism appears to have preceded the onset of voluminous arc axis volcanism of the Sierra Crest–Little Walker center by at least 1.8 m.y.
The Stanislaus Group consists of three formations, the Table Mountain Latite, Eureka Valley Tuff, and Dardanelles Formation, with the Eureka Valley Tuff divided into five members (ignimbrites and lavas). These are all represented in paleochannels (Fig. 3), but they are much thicker and more extensive in the Sierra Crest–Little Walker volcanic center and pull-apart basin, where they were all erupted (Figs. 2 and 4A).
Detailed field descriptions of the Stanislaus Group in the Sierra Crest–Little Walker volcanic center were given in Busby et al. (2013a, 2013b). We herein provide a generalized field description, linking outcrop characteristics with new petrographic and geochemical data (Figs. 6, 7, 8, and 9; Table 1).
Table Mountain Latite. The Table Mountain Latite forms one of the most extensive and perhaps the most voluminous lavas in California, second only after the 16 Ma Lovejoy flood basalt described by Garrison et al. (2008). Its preserved volume is ∼200 km3, with lavas forming sections up to 500 m thick (Fig. 6A), and single flows up to 100 m thick (Fig. 6B). One of the Table Mountain Latite trachyandesite lavas is paleomagnetically correlated to the Sierra Nevada Foothills (Pluhar et al., 2009), forming a single flow unit of >25 km3 (Gorny et al., 2009; Busby et al., 2013a, 2016). Flow-top breccias are common, but some flow tops are marked only by a vesicular horizon. Stretched vesicles are stretched parallel to the approximately N-S grabens (Busby et al., 2013b), in contrast with stretched E-W vesicles in the E-W paleochannels (Koerner et al., 2009). A section of 23 flows was measured on Sonora Peak in the Sierra Crest graben (Busby et al., 2008a), and a section of 20 flows was measured in an approximately N-S half graben on the Sierra Nevada range front just north of the Little Walker caldera (Pluhar et al., 2009; Busby et al., 2013b). In another approximately N-S range-front half graben, the Table Mountain Latite lavas have an interbedded megaslide slab of hornblende andesite lava with a 40Ar/39Ar age of 12.95 ± 0.09 Ma, derived from the Relief Peak Formation (Busby et al., 2013b; see Fig. 5). This indicates ongoing landsliding from active faults during the eruption of Table Mountain Latite.
Table Mountain Latite lavas emanate from fissure vents along faults or in the hanging walls near the surface trace of faults (Figs. 2 and 4). The vents are marked by fissure vent deposits up to 300 m thick that form ramparts up to 8 km long, formed largely of scoria bombs and lapilli, with scattered dense blocks up to 6 m in diameter (Fig. 6D). Because these are very large-volume lavas relative to other intermediate-composition lavas described in the literature, and because they were fissure fed (Figs. 2, 4, and 6D), they are referred to as “flood andesites” (Busby et al., 2013a).
Lavas in the southern part of the Sierra Crest–Little Walker pull-apart are basalt, basaltic trachyandesite, and trachyandesite, while those in the northern part also include basaltic andesite and andesite (Table 1). The amount of extension increases southward toward the Little Walker caldera (Busby et al., 2013b), which may explain this difference.
The Table Mountain Latite is easily recognized by its distinctive large sieve-textured plagioclase phenocrysts (Figs. 6C, 9B, 9C, and 9D), which are present in most, but not all, lavas (Fig. 9E). Sieve textures may be present in plagioclase cores or rims (Fig. 9B and 9D) or throughout the crystal (Fig. 9B). Many lavas have two pyroxenes, although clinopyroxene is more common and more obvious on outcrop because it forms larger phenocrysts (Figs. 7B and 9B–9E). Clinopyroxene is commonly twinned (Figs. 9B and 9D) and may occur in glomerocrysts with orthopyroxene and opaque oxides (Fig. 9B). Basalt, basaltic trachyandesite, and basaltic andesite lavas and dikes of Table Mountain Latite have phenocrysts of olivine with or without pyroxene or plagioclase (Figs. 7B and 9E), and trachyandesite lavas commonly have groundmass olivine.
The Table Mountain Latite is the most mafic stratigraphic unit of the Stanislaus Group, and it lacks silicic lavas entirely (Table 1). The only other stratigraphic unit this mafic is the youngest unit of the Stanislaus Group, the Dardanelles Formation, which is a single, small-volume, nearly aphyric shoshonite lava (Table 1). Existing ages on the base and top of the Table Mountain Latite overlap within analytical error (10.41 ± 0.08 and 10.36 ± 0.06 Ma; Figs. 3 and 5), so the flood lavas may have erupted in a short time. In contrast, the members of the Eureka Valley Tuff represent a longer time span: ca. 670–530 kyr for all members (including Basal Lava Flow Member), and ca. 280–120 kyr for the three ignimbrite members (Figs. 3 and 5).
Eureka Valley Tuff. The name “Eureka Valley Tuff” should ideally be modified to “Eureka Valley Tuff Formation,” because it includes lava members as well as “tuff” (ignimbrite) members (Fig. 3), as first recognized by Priest (1979) and Brem (1977). However, due to historical precedent, it continues to be referred to as the Eureka Valley Tuff (King et al., 2007; Koerner et al., 2009; Pluhar et al., 2009; Carlson et al., 2013). Ignimbrites of the Eureka Valley Tuff were inferred to have been erupted from the Little Walker caldera of Priest (1979), and this was confirmed by magnetic anisotropy studies of caldera outflow ignimbrites by King et al. (2007). However, the caldera fill is too altered to determine if Eureka Valley Tuff ignimbrites are present there. Lavas were not only erupted from the Little Walker caldera margins (Priest, 1979; Brem, 1977), but also along the length and width of the Sierra Crest–Little Walker volcanic center (Figs. 2 and 14C).
The Eureka Valley Tuff consists of three trachydacite ignimbrite members as well as two lava members (Table 1; Figs. 2, 3, and 14C). The two lava members differ from the Table Mountain Latite by having silicic lavas (trachydacite), in addition to mafic to intermediate lavas. The Eureka Valley Tuff marks the first appearance of hydrous phenocryst phases in the Stanislaus Group: Some of the trachydacite lavas and all of the trachydacite ignimbrites contain biotite, and some have hornblende in addition to pyroxenes (Figs. 7C, 9F, 9I, and 9J; also see modal analyses of Tollhouse Flat and By Day Members inBusby et al., 2013b). Like the Table Mountain Latite, nearly all of the ignimbrites and lavas have sieve-textured plagioclase (Figs. 9F, 9G, and 9H). Ignimbrites are massive and poorly sorted, with mixtures of pumice, glass shards, and accidental volcanic rock fragments (Figs. 6E, 6F, 6H, 6J, and 7C). Pyroclastic fall consists largely of pumice and is stratified and well sorted by size and density, with an absence of rock fragments (Figs. 6I, 7C, and 9J); it could alternatively be called pumice fall or Plinian fall. Broken crystals occur in both ignimbrites and pyroclastic fall deposits (Figs. 9I and 9J).
The lowest ignimbrite of the Eureka Valley Tuff, the Tollhouse Flat Member (Figs. 2 and 3), is by far the most voluminous of the three ignimbrites. It is also the most spatially extensive unit of the Stanislaus Group due to the much greater mobility of pyroclastic flows relative to lavas (for its distribution, see fig. 3 inBusby et al., 2016). It forms one cooling unit throughout the Sierra Crest–Little Walker volcanic center and beyond, including the reference section several kilometers north of the Little Walker caldera measured by King et al. (2007). It has a 40Ar/39Ar biotite age of 9.54 ± 0.04 Ma (Fig. 5). The second ignimbrite, By Day Member, looks similar to the Tollhouse Flat Member in the field, with a black basal vitrophyre passing upward into dark-gray, largely welded ignimbrite (Figs. 6E–6H). However, it is easily distinguished from the Tollhouse Flat ignimbrite by the presence of larger phenocrystic biotite in the Tollhouse Flat Member (Noble et al., 1974; see also Fig. 7C). The By Day Member forms one cooling unit in the Sierra Crest–Little Walker volcanic center and also in the Cataract paleochannel to the west, but at the reference section measured by King et al. (2007), it has three cooling units. It has a 40Ar/30Ar biotite age of 9.4 ± 0.3 Ma (Fig. 5). The third and uppermost ignimbrite (Upper Member) is by far the least extensive (Fig. 2). In the Sierra Crest–Little Walker volcanic center, the Upper Member ignimbrite is white and nonwelded (Figs. 6H and 6I), and it has interstratified pyroclastic fall deposits (Figs. 6I and 7C). At the reference section measured by King et al. (2007), the ignimbrite is welded and forms two cooling units. It has a 40Ar/30Ar biotite age of 9.34 ± 0.04 Ma (Fig. 5). All three ignimbrites have abundant volcanic rock fragments (Figs. 7C and 9I).
(1) Basal Lava Flow Member, which is underlain by the Table Mountain Latite and locally overlain by the Tollhouse Flat Member: It is distinguished from Table Mountain Latite by its trachydacite composition and presence of hornblende and biotite (Figs. 7C and 9F; Table 1). Our new 40Ar/30Ar hornblende age of 9.94 ± 0.03 Ma (Fig. 5) shows that the Basal Lava Flow Member is intermediate in age between the Table Mountain Latite and the basal ignimbrite of the Eureka Valley Tuff. The dated locality lies along the NE-trending Poison Flat fault where a trachydacite intrusion with vertical flow banding cuts the Table Mountain Latite and passes upward into a 70-m-thick trachydacite lava with flow banding parallel to lavas in underlying Table Mountain Latite (Fig. 2). This vent area is indicated by a red star in Figures 2B and 4.
(2) Lava Flow Member, which lies between Tollhouse Flat and By Day Member ignimbrites: Many of the lavas of the Lava Flow Member are indistinguishable from Table Mountain Latite (except by their stratigraphic position). Most lavas have basalt, basaltic trachyandesite, trachyandesite, and basaltic andesite compositions (Table 1), with large sieve-textured plagioclase phenocrysts (Fig. 9H), and many have clinopyroxene phenocrysts, which may be abundant to sparse (Figs. 7C, 9G, and 9H). However, trachydacite and trachydacite-andesite lavas also occur in the Lava Flow Member (unlike Table Mountain Latite; Table 1). Lava Flow Member rocks have large sieve-textured plagioclase (Fig. 9H), and variable large (Fig. 9G) to small or absent clinopyroxene (Fig. 9H). Vent facies deposits also include trachydacite (sample 831-2; Fig. 7C). The Lava Flow Member also includes a well-preserved basalt cinder cone intruded by olivine basalt and passing laterally into olivine basalt lavas (Fig. 6G).
The Lava Flow Member lavas are tabular and concordant with bedding in underlying and overlying strata, but they appear chaotic in one small area, on the ridge between Golden Canyon and Murray Canyon (Fig. 2B). There, the Lava Flow Member shows inconsistent bedding orientations and rapid lateral juxtaposition of lavas with varying textures and compositions. A trachydacite lava with flow breccia from that section yielded a new 40Ar/30Ar plagioclase age of 11.66 ± 0.07 Ma (sample JHEP-64; Fig. 2B). Because this lava lies up section from the 9.54 ± 0.04 Ma Tollhouse Flat Member (Fig. 2B), and because the section is chaotic, this sample is interpreted to be from a megaslide block within the Lava Flow Member (Fig. 5), indicating ongoing faulting during eruption of the Lava Flow Member. The megaslide slab overlaps in age with our dates on andesites of the Relief Peak Formation (Fig. 5), and we have found no in situ high-K2O rocks older than 10.41 ± 0.08 Ma (basal Table Mountain Latite on Sonora Peak; Fig. 5). High-K2O rocks were erupted at this time in the Bodie Hills (John et al., 2012), but that is 55 km away from the location of the dated slide block. The source of the megaslide slab is thus unknown.
We have found no lavas between the By Day and Upper Members. The error on the 40Ar/39Ar date on the By Day Member ignimbrite is large (Figs. 3 and 5), overlapping in age with the overlying Upper Member, but it is at least ∼0.17 m.y. younger than the underlying Tollhouse Flat Member (Fig. 3). Perhaps the Lava Flow Member records a greater time gap between the first two ignimbrites than the second and third ignimbrites. However, it seems more likely that the absence of lavas between the By Day and Upper Member ignimbrites records waning high-K2O effusive volcanism across the Sierra Crest–Little Walker volcanic center, as indicated by initial rapid flood andesite eruptions (Table Mountain Latite) through eruption of progressively smaller volumes of ignimbrite alternating with small-volume lava effusions (Eureka Valley Tuff). The uppermost high-K2O unit, the Dardanelles Formation, occurs in the type section of Stanislaus Group designated by Slemmons (1953), but it is the least-extensive and least-volumetric unit (discussed next).
Dardanelles Formation. The Dardanelles Formation consists of a single, distinctive, black glassy, nearly aphyric shoshonite lava, up to 60 m thick (top eroded), and two plugs along the western margin of the Sierra Crest graben (Busby et al., 2013a). Previous workers thought it had multiple lavas because they confused it with the Lava Flow Member (see summary in Koerner et al., 2009). The lava is discontinuously preserved for ∼15 km westward down the Cataract paleochannel to Dardanelles Cone (Busby et al., 2016). One could argue that it should be included in Eureka Valley Tuff, but the other lava members include silicic (trachydacite) lavas, and we have found none above the Upper Member, so we follow the historical precedent and retain the formational name, even though it is only one lava.
We report a new 40Ar/39Ar whole-rock age on the single Dardanelles Formation shoshonite lava of 9.137 ± 0.017 Ma (Fig. 5). This is consistent with its stratigraphic position above the Upper Member of Eureka Valley Tuff (Fig. 3; Busby et al., 2013a). It is ∼150–260 k.y. younger than the Upper Member, which makes it closer in age to the ignimbrites than the Basal Lava Flow Member, which is ∼470–510 k.y. older than the basal ignimbrite (Tollhouse Flat Member).
The Dardanelles Formation is the youngest high-K2O unit; thus, high-K2O volcanism of the Stanislaus Group spanned ∼1.2–1.4 m.y.
Disaster Peak Formation
The ca. 9–5 Disaster Peak Formation in the Sierra Crest–Little Walker volcanic center marks the end of high-K2O arc magmatism in the arc axis (Fig. 14D), suggesting that the supply of low-partial melt magmas was exhausted, or that their release by transtension slowed. The southern half of the Sierra Crest–Little Walker volcanic center became magmatically and tectonically inactive, except for a few small hornblende andesite intrusions that cut the Stanislaus Group, a hornblende-bearing ignimbrite and lava preserved as erosional remnants on the ridge crest north of the Kennedy Creek fault, and a quartz dacite sill intruded along the Chango Lake fault on the range front (Busby et al., 2008a, 2013a, 2013b, 2016).
Disaster Peak Formation strata in the northern part of the Sierra Crest–Little Walker pull-apart basin form nearly half of the up to 1300 m thick fill (Figs. 2B and 14D). There, the Disaster Peak Formation is ∼25% primary volcanic rock (basalt to andesite lavas and andesite to dacite block-and-ash-flow tuffs; Fig. 2B; Table 1), and the rest is debris-flow and fluvial deposits with volcanic clasts. Disaster Peak Formation in the northern part of the Sierra Crest–Little Walker pull-apart basin differs from Relief Peak Formation in the following ways: (1) It contains abundant reworked silicic pyroclastic debris in the form of pumice, crystals, and ash in debris-flow and fluvial deposits (Fig. 6K), while the Relief Peak Formation only has megaslide slabs of silicic pyroclastic rocks of the Oligocene Valley Springs Formation. The 600-m-high white cliffs on the north and east faces of Arnot Peak (Fig. 2B) are dominated by reworked silicic pyroclastic debris, although in most places, it is subordinate to andesitic lithic clasts. (2) It has far more abundant lavas, at all stratigraphic positions. (3) Abundant basaltic andesite, andesite, and dacite intrusions are closely associated spatially with the northern part of the Sierra Crest–Little Walker volcanic center and die off away from it into granitic basement (Fig. 2). These characteristics show that the ancestral Cascades arc axis lay within the Sierra Crest–Little Walker pull-apart basin during Disaster Peak Formation time (Fig. 14D), rather than east of it, as it did during Relief Peak Formation time (Figs. 14B and 15A). The reworked silicic pyroclastic debris records silicic explosive eruptions penecontemporaneous with deposition of the fluvial and debris-flow deposits, indicating that the arc was becoming more evolved. The Disaster Peak Formation also has more hornblende and less pyroxene than the Relief Peak Formation (Figs. 7A and 7D), although this is not a reliable way to distinguish between the two formations.
In the N-S Sierra Crest graben, the Disaster Peak Formation forms a N-S paleochannel cut into the Stanislaus Group (Fig. 14D), and this is filled with interstratified fluvial and debris-flow deposits, with clast imbrication and cross-bedding indicating northward transport. This paleochannel (“deranged Cataract paleochannel”; Fig. 4A) lies perpendicular to the E-W Nevadaplano paleochannels (Figs. 4A and 14A) and parallel to modern Walker Lane drainages (Fig. 1), indicating tectonic reorganization of the landscape at ca. 9–5 Ma (Busby et al., 2016). We infer that it fed into the Mokelumne paleochannel, which remained undisrupted in this time frame (Fig. 14D). This section includes a hornblende orthopyroxene andesite block-and-ash-flow tuff (JHEP-55; Fig. 2B; Table 1), which yielded a 40Ar/39Ar hornblende age of 4.96 ± 0.05 Ma (Fig. 5). In contrast, the NE transfer zone basins lack fluvial deposits and are dominated by massive debris-flow deposits, similar in appearance to Relief Peak Formation; however, they differ from Relief Peak Formation by having far more interstratified lavas (Figs. 2B and 14D).
Lavas in the Sierra Crest graben include clinopyroxene basaltic andesite lavas (samples 830-3, 830-2, and 918-1; Fig. 7D; Table 1) and a basalt lava with plagioclase and trace olivine (sample 906-1; Figs. 2B and 7D; Table 1). Samples 827-1 and 827-2 (andesite and basaltic andesite, respectively; Table 1) are tentatively mapped as intrusions on Figure 2B (Tdpia and Tiba, east of Disaster Peak), but they may instead be lavas. Basal contacts on lavas within the deranged Cataract paleochannel locally exhibit complex mixing with the sedimentary rock below, indicating eruption onto wet and unconsolidated sands (peperites). Block-and-ash-flow tuffs in the Sierra Crest graben are generally too lenticular to map individually within Disaster Peak Formation, but one mappable body lies at the base of the Disaster Peak Formation along the Paradise Valley trail south-southeast of Arnot Peak (Tdpba; Fig. 2B), consisting of an andesite block-and-ash-flow tuff (sample 824-3; Table 1). Smaller lenses of block-and-ash-flow tuff within Disaster Peak Formation include (Fig. 2B; Table 1): (1) an andesite block-and-ash-flow tuff along the ridge north of Murray (sample 911-4); (2) an andesite-dacite block-and-ash-flow tuff northwest of Arnot Peak, with phenocrysts of clinopyroxene, orthopyroxene, and hornblende in roughly equal proportions (sample JHEP-94; Fig. 7D; Table 1), and local fine-grained clots of pyroxene intergrown with plagioclase (Fig. 9K); and (3) an andesite block-and-ash-flow tuff south of Disaster Peak with hornblende and minor orthopyroxene (sample JHEP-55; Fig. 7D). It yielded a new 40Ar/39Ar hornblende age of 4.96 ± 0.05 Ma (Fig. 5).
Lavas in the NE transfer zone basins include a section of seven olivine basalt lavas along the Mineral Mountain–Poison Flat fault zone (sample JHEP-83, and samples 817–1, 2, 3, 4, 5; Fig. 2B; Table 1). These have plagioclase and trace olivine (Fig. 7D).
Tectonic recycling continued during deposition of Disaster Peak Formation in the northern Sierra Crest–Little Walker volcanic center, as shown by the presence of a dacite lava megaslide slab in a transfer zone basin (sample JHEP-70; Fig. 2B). This lies up section from the Stanislaus Group (Fig. 2), but it yielded a 40Ar/39Ar hornblende age of 11.05 ± 0.031 Ma (Fig. 5). It is therefore recycled from Relief Peak Formation.
Numerous plugs are closely associated with the Disaster Peak Formation in the northern half of the Sierra Crest–Little Walker volcanic center and pull-apart (Fig. 2). These are mineralogically and geochemically similar to the lavas and block-and-ash-flow tuffs (Figs. 7 and 8; Table 1) and probably represent feeders for them. Some of them are localized along faults or buried faults (e.g., Jones Canyon fault; Fig. 2B).
In summary, the northern Sierra Crest–Little Walker pull-apart remained magmatically active from ca. 9 to ca. 5 Ma, with basalts, basaltic andesites, and andesites at all stratigraphic levels throughout the pull-apart, accompanied by emplacement of basalt to dacite intrusions. The Ebbetts Pass pull-apart basin and volcanic center began to form by 6.4 Ma (Figs. 5 and 14D), thus overlapping in time with the northern part of the Sierra Crest–Little Walker volcanic center and pull-apart by ∼1.4 m.y. (Fig. 14D).
Ebbetts Pass Volcanic Center and Pull-Apart Basin
The basal fill of the Ebbetts Pass pull-apart basin (Tvu and Tva) consists of andesitic debris-flow deposits that are hydrothermally altered (Figs. 2B and 14E). In contrast, the overlying volcanic rocks in the Ebbetts Pass pull-apart basin are unaltered (Fig. 13). We infer that the altered basal debris-flow deposits record catastrophic sedimentation during initial subsidence of the pull-apart basin, similar to mass-transport deposits at the base of the Sierra Crest–Little Walker pull-apart basin. The basal fill was then altered by convective flow of heated groundwater due to emplacement of intrusions below the volcanic center that grew above it. Unaltered intrusions in the altered basal fill probably represent the last stages of this process. These include a two-pyroxene andesite intrusion (sample MG-08-13) and a two-pyroxene hornblende biotite quartz dacite intrusion (sample MG-08-10; Fig. 12; Table 1). The dacite intrusion has large quartz and plagioclase phenocrysts, hornblende with groundmass cores, and small twinned pyroxene phenocrysts or glomerocrysts set in a microcrystalline groundmass (Fig. 13F). Additional unaltered intrusions include a small unmapped andesite body (sample MG-08-11; Table 1) and a large mapped andesite body with a sample at each end (Tdpia, samples MG-08-13 and 920-2; Fig. 2B). The unaltered intrusions in the altered basal fill are inferred to represent feeders to unaltered volcanic rocks and intrusions in the overlying part of the section, because they are similar in chemistry and mineralogy.
We divide the unaltered volcanic rocks of the Ebbetts Pass pull-apart basin into a lower silicic section and an upper section that we refer to as the Ebbetts Pass stratovolcano (Fig. 5).
Lower Silicic Section
Mappable bodies of the lower silicic section include an older (Miocene) dacite lava (Tdpld), and younger (Pliocene) rhyolite lavas (Tdplr) and strongly welded rhyolite ignimbrite (Tdpwi; Fig. 2B) that are ∼1.6 m.y. younger than the Miocene lavas (Fig. 5).
The dacite lava forms a fault-bounded body >2.5 km long and >225 m thick in the hanging wall of a strand of the Wolf Creek fault, along the southern margin of the basin (Tdpld; Fig. 2B). This splay is partly buried by the Pliocene rhyolite lava (Tdplr), described later, and is crosscut by the undated andesite intrusion (Tdpia) described earlier (Fig. 2B). The dacite lava has a new 40Ar/39Ar hornblende age of 6.367 ± 0.017 Ma (sample MG-08-12; Fig. 5). We interpret it to be part of a suite of andesite to dacite lavas emplaced along the margins of the incipient Ebbetts Pass pull-apart at 6.4–6.2 Ma (Fig. 14D); further evidence for this includes a Miocene andesite megaslide slab, 25 m thick and >200 m long, with a hornblende 40Ar/39Ar age of 6.203 ± 0.011 Ma (Fig. 5; sample MG-08-08; Fig. 2B). This slab was shed from the basin margin into strata of the Pliocene Ebbetts Pass stratovolcano, described later. The Miocene dacite lava varies from plagioclase hornblende dacite (sample 920-3; Fig. 11A) to plagioclase clinopyroxene orthopyroxene hornblende biotite quartz dacite (sample MG-08-12; Fig. 11A). Although most of the body is dacite (samples 920-3, 920-4, and MG-08-12; Table 1), the base of the lava is an andesite (sample 920-1; Table 1), and mafic enclaves are present near the western margin of the lava. The mineralogical and compositional heterogeneity in the dacite lava indicates that it is composed of multiple flows or records magma mingling. It differs from the Pliocene silicic volcanic rocks and intrusions by lacking sanidine (Fig. 11A).
Our new dates on the Pliocene rhyolite lavas and the rhyolite welded ignimbrite are very similar in age, perhaps within 60 k.y. of each other using the analytical errors (Fig. 5), and they lie along strike of each other (Tdplr and Tdpwi, respectively; Fig. 2B).
The Pliocene rhyolite welded ignimbrite (Tdpwi; Fig. 2B; Table 1) has a new 40Ar/39Ar sanidine single-crystal age of 4.636 ± 0.011 Ma (Fig. 5). It consists largely of sintered bubble-wall shards and collapsed pumice (Fig. 13A), with 30% phenocrysts, including plagioclase, biotite, sanidine, and quartz ± hornblende (Fig. 11A). Quartz and plagioclase are strongly embayed, and the mafic minerals are fresh, with no inclusions or rims (Fig. 13A). The rhyolite ignimbrite has very strongly welded domains that appear banded, alternating with more weakly welded domains that are clearly vitroclastic. The orientation of pumice compaction foliation is variable, and fiamme may be attenuated to >4 m in length and deformed into complex flow folds, typical of very strongly welded (“lava-like” or “rheomorphic”) ignimbrites. The glassy part of the ignimbrite has perlitic fracturing, and in the upper vapor phase-altered part, the fiamme are altered. The Pliocene rhyolite lava map unit (Tdplr; Fig. 2B) has a new 40Ar/39Ar hornblende age of 4.73 ± 0.03 Ma (Tdplr, sample JHEP-44; Fig. 4). It forms a stubby map unit, >60–135 m (200–440 ft) thick and 1 km long, typical of silicic lavas. Its mineralogy is identical in two samples collected close to each other (200 m apart, JHEP-43 and JHEP-44; Figs. 2B, 11A), with plagioclase, hornblende, biotite, and sanidine. However, a third sample collected 1 km away from those (JHEP-53; Fig. 2B) also has clinopyroxene and quartz (Fig. 11A), although it is otherwise identical in the field. Thus, the rhyolite lava probably represents a composite of two or more flows, or comingled magmas. The lava is pervasively flow-banded, with gray glassy bands (Fig. 13B) and pale-red stony/devitrified bands (Fig. 13C).
The Pliocene rhyolite welded ignimbrite and rhyolite lava are overlain by the Ebbetts Pass stratovolcano (Tdpeps; Fig. 2B). The base of the Ebbetts Pass stratovolcano is marked by basaltic andesite lavas that dip away from its center and downlap onto underlying strata. It is this downlap that forms the unconformity recognized by Armin et al. (1984).
Ebbetts Pass Stratovolcano
Our Ebbetts Pass stratovolcano map unit (Tdpeps) largely corresponds to the unit that Armin et al. (1984) mapped as “Miocene Raymond and Silver Peak Andesites of Wilshire (1957) (Trs).” However, it is Pliocene (Fig. 5) in age and varies in composition from basaltic andesite through rhyolite, with one peperite basalt intrusion in the western flank, described below (Table 1). Armin et al. (1984) described what we term the “Ebbetts Pass stratovolcano” as “thick-bedded pyroxene and hornblende andesite lahars,” but we find it is dominated by lavas, flow breccias, and block-and-ash-flow tuffs, with lesser “lahars” (i.e., volcanic debris-flow deposits). They mapped this unit in the hanging wall (east of) the Noble Canyon fault, and in its footwall to the northwest, north of Ebbetts Pass in the Reynolds Peak–Raymond Peak area, and we concur (Fig. 2B). In addition, we recognized a small erosional remnant of it in the footwall of the Noble Canyon fault south of Ebbetts Pass (Fig. 2B).
The Ebbetts Pass stratovolcano is eroded, but its cone-shaped morphology and quaquaversally dipping beds are obvious in photos (Figs. 10A and 10B) and map pattern (Fig. 2B). It grew within the Ebbetts Pass pull-apart basin, and its center is occupied by the intrusions on Highland Peak and Silver Peak (Fig. 2B). It eventually filled the pull-apart and built up and out westward across the western basin-bounding fault (Noble Canyon fault), onto the footwall of the fault, where strata dip west (Fig. 2B).
The base of the Ebbetts Pass stratovolcano is a sequence of black olivine two-pyroxene basaltic andesite to basaltic trachyandesite lavas and flow breccias (samples JHEP-47 and JHEP-9; Table 1; Figs. 11A and 13D). This sequence lies largely within the pull-apart basin in the hanging wall of the Noble Canyon fault, but it is also preserved in the small erosional remnant on the footwall of the Noble Canyon fault south of Ebbetts Pass, which is close to the center of the stratovolcano. The 40Ar/39Ar groundmass age of the footwall sample is 4.6 ± 0.7 Ma, whereas the 40Ar/39Ar plagioclase age of the hanging wall sample is 4.90 ± 0.02 Ma. Because the hanging-wall lavas rest upon the welded rhyolite ignimbrite (Tdpwi), which has a very precise 40Ar/39Ar sanidine date (4.636 ± 0.014 Ma on 14 crystals; Fig. 5), we prefer the matrix age for the basal sequence. The black basaltic andesite lavas on the footwall overlie a white clinopyroxene hornblende orthopyroxene quartz silicic lava (sample JHEP-10; Figs. 2B, 10C, and 11A), and a nearby biotite rhyolite dike yielded a plagioclase 40Ar/39Ar age of 4.626 ± 0.017 Ma (sample JHEP-8; Figs. 2B, 5, and 12). The age of the rhyolite dike overlaps with the age of the Pliocene Ebbetts Pass volcanic center, so it is inferred to be a feeder for the stratovolcano. The small footwall erosional remnant south of Ebbetts Pass also has a narrow (<20-m-wide) channel filled with pumiceous fluvial sandstone (Fig. 10D).
In the hanging wall of the Noble Canyon fault (i.e., within the pull-apart basin), the stratovolcano section is 1 km thick (top eroded), and the basal sequence of black basaltic andesite lava passes upward into a much more varied section of basaltic andesite to rhyolite lavas and block-and-ash-flow tuffs (Table 1), with minor interstratified volcanic debris-flow deposits. The strata show steep depositional dips and stratal discordances caused by slumping and synvolcanic faulting (Fig. 10E). One sample traverse within the pull-apart basin includes (in stratigraphic order; Fig. 2B): (1) a 50-m-thick section of olivine pyroxene basaltic trachyandesite lavas and flow breccias (sample 814-3; Table 1); (2) an 80-m-thick section of plagioclase-bearing block-and-ash-flow tuffs, breccia, and columnar-jointed lava, where the lava is trachyandesite in composition (sample 814-4; Table 1); (3) a 50-m-thick perlitic flow-banded biotite dacite lava (sample 814-5; Table 1), which passes laterally into a nonwelded silicic ignimbrite; (4) a 3-m-thick columnar-jointed biotite hornblende trachyandesite lava (sample 814-6; Table 1); and (5) a 10-m-thick mafic lava overlain by a 20-m-thick white quartz biotite rhyolite block-and-ash-flow tuff (sample 814-7; Table 1) with large blocks (Figs. 10F and 10G). Lesser volcanic debris-flow deposits intervene between the sampled primary volcanic units. A second sample traverse within the pull-apart basin includes (in stratigraphic order, Fig. 2B): (1) a 30-m-thick two-pyroxene andesite lava with trace hornblende (sample MG-08-08; Table 1; Figs. 11A and 13E), and (2) a hornblende two-pyroxene andesite lava with flow breccia (sample MG-08-08; Table 1). The latter is an ∼80-m-thick, 300-m-long megaslide block with an age (6.203 ± 0. 011 Ma; Fig. 5) that is much older than strata that lie down section from it.
Silicic intrusions in the core of the Ebbetts Pass stratovolcano in the pull-apart basin (on Silver Peak and Highland Peak [Tdpis]; Fig. 2B) are the youngest dated rocks in the stratovolcano, with a 40Ar/39Ar sanidine age of 4.478 ± 0.007 Ma (JHEP-49; Figs. 2B and 5). Although the three closely spaced samples on the flank of Highland Peak are all dacites (JHEP-48, 49, 50; Table 1), modal analyses show they differ, with sample JHEP-49 containing biotite and quartz (Fig. 13G), which are absent from sample JHEP-48 (Fig. 12). Other samples contain sanidine (Fig. 13H). The cliffs on the two peaks also show some variation in color, which may reflect variations in composition. However, the cliffs are too steep to allow detailed mapping of the intrusions on Highland Peak, and Silver Peak is completely inaccessible. For these reasons, they are generalized as silicic intrusions (Tdpis; Fig. 2B) rather than dacite intrusions.
In the footwall of the Noble Canyon fault (i.e., west of the pull-apart basin), in the Reynolds Peak–Raymond Peak area, strata dip west and are up to 500 m thick (top eroded; Figs. 2B and 10H). This section was deposited when the stratovolcano filled the pull-apart basin, and it represents the youngest preserved strata in the stratovolcano (Fig. 14E). Strata dip uniformly westward south of Reynolds Peak (Fig. 10H), but these pass laterally (northward) into slump folded strata at Reynolds Peak (Fig. 10I). The slump folding indicates sector collapse, perhaps triggered by movements on the Noble Canyon fault. In the part of the section that is not slumped, the lower part of the section is dominated by a reddish black andesite block-and-ash-flow tuff (Fig. 10J), and the upper part of the section is dominated by a light-gray andesite block-and-ash-flow tuff (Fig. 10K; samples CB09-4 and CB09-5, respectively; Table 1). The slumped section contains a basalt peperite intrusion (sample CB09–3; Table 1), which shows very complex mixing with its host volcaniclastic rocks, indicating intrusion while the host was unconsolidated and wet. We infer that this basalt is a flank intrusion that was plumbed up the nearby basin-bounding fault (Noble Canyon fault; Figs. 2B and 14E). This may have contributed to the slumping there, or perhaps it was intruded during a seismic episode on the Noble Canyon fault.
Lavas preserved in erosional remnants of the Mokelumne paleochannel ∼6–12 km west of the Ebbetts Pass pull-apart basin (Tdpeps; Fig. 2A) are interpreted to be lavas that were erupted from the Ebbetts Pass stratovolcano and flowed down the relict paleochannel, which had become beheaded by the pull-apart structure (Fig. 14E). Outcrops north of Highway 4 (Fig. 2B) were first mapped as “Miocene Raymond and Silver Peak Andesites of Wilshire (1957)” by Armin et al. (1984), and we concur with this correlation. Additionally, we herein interpret these outcrops as lavas. These lavas are similar to a lava newly mapped as Ebbetts Pass stratovolcano in the Mokelumne paleochannel south of Highway 4 (Fig. 2A; samples CB09-1 and CCB09-2; Table 1). The lava south of Highway 4 forms an ∼30-m-high north-facing cliff called Bull Run (fig. 11G inBusby et al., 2016). The basal ∼5 m section of the lava has unusual thin, wispy, crystal-rich accumulations with both horizontal and vertical orientations, although they are irregular (Figs. 10L, 10M). These wispy accumulations may indicate that crystal-rich magmas mingled with crystal-poor magmas, although the major-element chemistry of the crystal-rich sample (CB09-2; Table 1) is the same as that of the crystal-poor sample (CB09-1; Table 1). It is not understood why the crystal-rich accumulations are oriented in vertical as well as horizontal planes, nor why they are restricted to the basal sixth of the lava.
Intrusions in and around the Ebbetts Pass volcanic center differ from those in and around the Sierra Crest–Little Walker volcanic center by having biotite ± sanidine ± quartz, which are absent in the Sierra Crest–Little Walker volcanic center, except for biotite in the Eureka Valley Tuff (Figs. 8 and 12). This is consistent with the more-evolved compositions of volcanic rocks in the Ebbetts Pass volcanic center (Table 1). Many of the intrusions in and around the Sierra Crest–Little Walker and Ebbetts Pass volcanic centers are not dated, although maximum ages are provided for most samples by crosscutting relationships (Figs. 8 and 12). However, because the mineralogy of the intrusions matches that of volcanic rocks in each center (Figs. 7, 8, 11, and 12), and the two centers differ, we infer that the intrusions were largely emplaced coeval with volcanism in each of the respective centers.
We tentatively estimate the minimum original volume of the Ebbetts Pass volcanic center at ∼270 km3. It is difficult to estimate original volumes of Quaternary Cascades volcanoes, due to glacial erosion (cf. Hildreth, 2007), and the Ebbetts Pass volcanic center has also undergone Alpine erosion, and is even older (Pliocene to late Miocene). To make our estimate, we assumed that the Ebbetts Pass volcanic center completely filled the pull-apart basin, in order for the Ebbetts Pass stratovolcano to build out beyond it, as shown by erosional remnants preserved on the peaks 6–7 km to the NW of the center. The basin fill consists of a lower, sheet-like sequence of lavas and ignimbrites (lower silicic unit) probably deposited along the length and width of the basin, overlain by the Ebbetts Pass stratovolcano, the quaquaversal dips of which suggest it was conical, although it seems likely that some of its volcanic material was funneled northward along the basin (not preserved). Within the pull-apart basin, the central silicic intrusions form the highest units, so only a minimum thickness of the volcanic fill can be estimated there (∼1 km). However, the NW flank (top eroded) adds another 500 m to the total minimum thickness, and the center presumably lay hundreds of meters higher than the flank. We therefore estimate a minimum thickness of 2 km. The northward-widening pull-apart basin is an average of 9 km wide (E-W), and the N-S extent is ∼15 km. An estimated original volume of ∼270 km3 is comparable to that estimated for the Lassen volcanic center in the last 825 k.y. (200 km3; cf. Hildreth, 2007). The Lassen volcanic center is also a good modern analog, because it is a major arc magmatic focus at the rift tip (i.e., in the northernmost Walker Lane pull-apart basin; Busby, 2013; Smith et al., 2016).
Rift Volcanic Rocks
Small basalt and basaltic andesite plugs, most containing olivine phenocrysts (Fig. 13I), are scattered around the region (Fig. 2) but are not associated with mapped faults. We only dated one of these intrusions, which has a 40Ar/39Ar plagioclase age of 3.65 ± 0.03 Ma (Fig. 5). The youngest dated mafic volcanic rock in the region is the Columns of the Giants lava on the south side of Highway 108 (mapped as unit Qb on Fig. 2A). This yielded a 40Ar/39Ar age of 177 ± 7 ka (Farmer et al., 2013). We discovered a basalt feeder dike for this (sample Qdike 8–19–08; Table 1). These mafic rocks form part of the regional suite of postsubduction magmas (Farmer et al., 2013), which have been interpreted as continental rift magmas (Lange and Carmichael, 1996; Cousens et al., 2011, 2012; Putirka et al., 2012; John et al., 2015b; Putirka and Platt, 2012).
While we lack mineral composition data, which would otherwise yield more precise answers, two mineral textures in particular provide some clues as to how magmas might have been processed.
(1) In the now-classic experiments by Tsuchiyama (1985), the sieve (or in his paper “dusty”) textures of Figure 9A were demonstrated to be a product of magma mixing: His work showed that in nearly all cases, the sieve areas have high Ca compared to inner, nonsieved cores, and so the texture can be formed by mixing a low-An crystal from a felsic magma into a more mafic, or at least higher-Ca magma, with which the crystal then attempts to equilibrate (Tsuchiyama, 1985). The more coarse “skeletal” plagioclase textures were not reproduced in those experiments, however, and Tsuchiyama (1985) concluded that such skeletal plagioclase represents mixing under conditions of supercooling, as proposed by Kuo and Kirkpatrick (1982). However, in more recent experiments, Iezzi et al. (2014) were able to obtain skeletal plagioclase textures at relatively rapid undercooling with no implicit need for magma mixing, and at the Tequila volcanic field in Mexico, Frey and Lange (2011) were able to link skeletal-type textures to undercooling initiated by magma degassing (rather than magma mixing). Clearly, some of our plagioclase textures are due to magma mixing, but some might also represent near-closed system cooling, or degassing.
(2) Many of our samples are saturated with two pyroxenes, and clinopyroxene (Cpx) often rims orthopyroxene (Opx). This Cpx-on-Opx relationship can come about in different ways, depending upon bulk composition. If these minerals originated from a mafic basaltic parent (>10 wt% MgO), then the Cpx-on-Opx (Fig. 9A) may indicate fractional crystallization of that parent magma at ≥8 kbar pressures. Experimental work by Grove et al. (2003) showed a crossover in phase relationships for such mafic magmas, where Opx crystallization only preceded that of Cpx (by ∼50 °C) at high pressures (their fig. 2), and this phase relationship held over a wide range of H2O contents (0–6 wt% H2O; their fig. 4). Another possibility, though, is that Opx and Cpx derive from different basaltic magmas. Experimental data indicate that at low pressures, Opx either follows Cpx (Grove et al., 2003) or does not saturate at all (e.g., Barclay and Carmichael, 2004). So another possibility is that Opx grains were derived from basaltic magmas at elevated pressure ranges and mixed with lower-MgO basalts (<8 wt%; e.g., Barclay and Carmichael, 2004) that were near or at Cpx saturation. Yet another possibility is that the parent magmas were more felsic: Blatter and Carmichael’s (2001) experiments on magmas having 62–64 wt% SiO2 (and <6% MgO) could be cosaturated with both Opx and Cpx, where Opx preceded Cpx saturation in the pressure range 0.001 kbar (1 atm) to 3.5 kbar by anywhere from 20 to 100 °C, depending upon bulk composition and pressure. As with plagioclase textures, there are strong hints of magma mixing, and mineral compositions will tell us much about where and when mixing occurred and the kinds of magmas that interacted with one another.
Whole-rock major-element compositions from rocks in the Sierra Crest–Little Walker and Ebbetts Pass volcanic centers show that these systems are quite similar, with the notable exception of the Stanislaus Group (Fig. 15), which includes the Table Mountain Latite (lavas), the Eureka Valley Tuff (both ignimbrites and lavas), and the Dardanelles Formation (a single lava). Other volcanic and intrusive units are essentially identical to one another, being mostly calc-alkalic, and ranging from basalt to rhyolite in composition (Fig. 15A), with whole-rock compositions that fall on remarkably straight trends in SiO2 versus CaO (Fig. 15D), TiO2 (Fig. 15E), and K2O (Fig. 15F) diagrams. Such linear trends are indicative of magma mixing, although changes in slope in SiO2 versus Al2O3 (Fig. 15H), for example, require at least three different mixing end members—or, more likely, mixing and fractional crystallization were both important for generating magmatic diversity. At both the Sierra Crest–Little Walker and Ebbetts Pass volcanic centers, the majority of volcanic rocks trace back to similar basaltic parental magmas (Figs. 15A–15F), having ∼49% SiO2 and 12% MgO. However, the parental magmas are certainly not singular in composition, as they would appear to have varying amounts of total Fe2O3 (8–11 wt%, projected to 12% MgO) and TiO2 (1%–2.5% at 49% SiO2), and K2O contents vary from 0.7% to 2.0% at 49% SiO2, and are >5 wt% for the Table Mountain Latite. However, beyond K2O, the mafic compositions at each volcanic center overlap, indicating that similar parental magmas were delivered to the Sierra Crest–Little Walker and Ebbetts Pass volcanic centers.
The Stanislaus Group samples are a notable exception, as these samples are nearly all trachytic (basaltic trachyandesite to trachydacite), and their trachytic character is mostly a result of higher K2O (Fig. 15F). These are not the first high-K2O volcanic rocks to be observed in the Sierra Nevada, and Farmer et al. (2002) presented an interesting model whereby high-K2O volcanism in the southern Sierra Nevada is a response to the removal of continental lithosphere during “delamination” (e.g., Ducea and Saleeby, 1998). In this model, lithosphere removal allows the heating of specially enriched parts of the lower crust, producing the enriched high-K2O lavas. We reject this model for several reasons. Perhaps the most important reason is that the first-order response to lithosphere removal, and consequent asthenosphere upwelling, should be an increase in the degree of mantle partial melting, as the lithosphere is replaced by convective mantle, which is the source of mid-ocean-ridge basalt (MORB). The predicted volcanic expression, then, is an increase in the total volume of erupted volcanic materials, with geochemical trends that involve greater melt fractions and an increase in MORB-like (convective) geochemical signatures. However, none of the high-K2O lavas exhibits these characteristics: Both the central Sierra Nevada Stanislaus Group and the southern Sierra Nevada high-K2O suites are the products of very low degrees of melt fraction of typical continental mantle lithosphere, rather than a special enriched source (Putirka and Busby, 2007; Putirka et al., 2012). Worse still, lavas erupted after the presumed delamination event—including entrained mantle xenoliths (e.g., the Big Pine volcanic field)—have isotopic signatures that require an ancient continental mantle lithosphere source (Putirka et al., 2012). As shown by Putirka et al. (2012, their fig. 14), the convective mantle does indeed make its appearance, but only to the south and east, beneath the Coso volcanic field, where the north-migrating Mendocino triple junction has had sufficient time to affect the removal of continental mantle lithosphere. We would partially agree with Farmer et al. (2013) that the lithosphere beneath the Sierra Nevada is being heated, but we argue that the lithosphere is degraded slowly, and that high-K2O volcanism marks the earliest stages of lithosphere degradation, not its removal, which comes much later (e.g., Putirka et al., 2012, their fig. 16).
Our new data support not just a distinct source for the Stanislaus Group, but also a distinct differentiation history compared to other magmas of the Sierra Crest–Little Walker and Ebbetts Pass volcanic centers. As noted in earlier studies (Putirka and Busby, 2007; Putirka et al., 2012), the Table Mountain Latite samples have high concentrations of incompatible trace elements, which can be attributed to low degrees of partial melting. However, the Table Mountain Latite and Eureka Valley Tuff also exhibit a distinct break in slope in TiO2 (Fig. 15E) and trend to higher Al2O3 contents and lower CaO contents (Figs. 15G and 15H). These trends indicate significant clinopyroxene fractionation (Figs. 15G and 15H; see Priest 1979), and the low CaO and lower Al2O3 contents of the Eureka Valley Tuff can be produced by adding plagioclase to the crystallization assemblage, at ∼58% SiO2. The major-oxide trends (Fig. 15) also indicate that the Eureka Valley Tuff lavas and Dardanelles Formation lava can be obtained by Table Mountain Latite fractionation, if both plagioclase and Fe-Ti oxides reach saturation in Table Mountain Latite liquids at 60% SiO2. In contrast, TiO2 decreases with increasing silica across the entire range of SiO2 contents (48%–76%) for non–Stanislaus Group samples from the Sierra Crest–Little Walker and Ebbetts Pass volcanic centers; this might not necessarily reflect later Fe-Ti oxide saturation for Stanislaus Group lavas, but instead more complete mixing of the former, or perhaps the preferential eruption of unmixed lavas in the Stanislaus Group. If the contrasts in Fe and Ti are not due to delayed Fe-Ti oxide saturation, then the Stanislaus Group might better capture the crystallization-fractionation processes by which the Sierra Crest–Little Walker and Ebbetts Pass volcanic centers were produced.
The Stanislaus Group magmas appear to be cosanguineous, which supports the interpretation that they represent a single magmatic system, even though lavas of the Stanislaus Group erupted over a much larger area than the caldera-forming ignimbrites (Figs. 2 and 14). We surmise that all these magmas shared a similar magma plumbing system, although this hypothesis is by no means required: Separate plumbing systems may well deliver magmas to similar depths, which then fractionate over similar temperature intervals. In any case, the Table Mountain Latite and Eureka Valley Tuff fall close to a simple fractionation trend, involving only clinopyroxene for the most mafic samples, and clinopyroxene + plagioclase + Fe-Ti oxides for trachyandesites. In contrast, the remaining (non–Stanislaus Group) Sierra Crest–Little Walker volcanic center samples and all of the Ebbetts Pass volcanic center samples fall on remarkably linear trends, similar to one another but distinct from the Table Mountain Latite and Eureka Valley Tuff. Those linear trends are not matched by the naturally curved trends resulting from fractional crystallization alone, and so they almost certainly represent magma mixing to a degree that was absent for the Table Mountain Latite and Eureka Valley Tuff.
We also find that Tertiary volcanic rocks (excepting the Stanislaus Group) match the major-element compositions of Cretaceous granitic rocks. This match indicates no fundamental difference in the petrologic origin of the Cretaceous plutonic and Tertiary volcanic suites. In such a case, the volcanic systems may indeed provide a better understanding of the genesis of the Sierra Nevada Batholith. For example, Lee et al. (2006) have suggested that Cretaceous plutons may be derived by clinopyroxene + garnet fractionation in the lower crust. However, we see no evidence for the involvement of garnet in the crystallization history of Tertiary Sierra Nevada volcanic rocks. Instead, olivine is found in the most primitive of Tertiary volcanic samples, and clinopyroxene and plagioclase dominate the more felsic samples. In addition, fractionation of olivine + clinopyroxene + plagioclase + Fe-Ti oxides, combined with magma mixing, would explain Tertiary volcanic major-oxide variations. Clearly, whatever processes produced the Tertiary volcanic rocks were capable of producing the Sierran granitic plutons as well.
CONCLUSIONS: FEATURES AND PROCESSES COMMON TO TRANSTENSIONAL ARC-RIFT SETTINGS
We conclude with a summary of the evolution of the Sierra Crest–Little Walker and Ebbetts Pass volcanic centers and pull-apart basins (Fig. 14), describing characteristics that should be recognized in other arc-rift settings, and some future directions for research.
A key geologic process in the ancestral Cascades arc pull-apart basins is herein referred to as “tectono-stratigraphic recycling”: the transfer of megaslide slabs from footwall to hanging-wall blocks in the transtensional arc rift (Figs. 5 and 14). These are referred to as “slabs” rather than “blocks” because they commonly are thin and long (aspect ratios of ∼1:5–1:20), with internal layering subparallel to the strata that enclose them. This indicates that the slabs broke off the parent outcrop along bedding planes and slid, relatively intact, rather than tumbling and breaking into equant masses of variable orientation. In some cases, these slabs are relatively easy to recognize; for example, megaslide blocks of white rhyolite welded ignimbrite of the Oligocene Valley Springs Formation are obvious within dark-tan to brown andesitic debris-flow deposits of the Miocene Relief Peak Formation. More difficult to recognize is recycling of Relief Peak Formation andesitic volcaniclastic deposits into Relief Peak Formation debris-avalanche deposits. Most difficult to recognize is the recycling of lava megaslide slabs; these are commonly first identified by anomalously old 40Ar/39Ar ages and then confirmed as megaslide slabs through the details of their field relations, including chaotic bedding, with smaller (i.e., meter- to decameter-scale) blocks surrounding the megaslide slabs.
Pull-apart basins (e.g., Fig. 4B) subside more rapidly than any other basin type (Nilsen and Sylvester, 1995). They are steep-walled and deep, with uplifted blocks interspersed with basins in very close proximity, and with rapidly shifting sites of uplift and subsidence (Nilsen and Sylvester, 1995; Mann, 2007). Megaslide blocks in arc-related pull-apart basins are common due to the abundance of tabular, igneous stratigraphic units (lavas and welded ignimbrites) that commonly rest on easily undermined volcaniclastic deposits (debris-flow and fluvial deposits).
Exploitation of Arc Axis by Transtensional Rifting
It has long been recognized that arcs have evolved into rifts in a wide variety of crustal settings (e.g., Yamaji, 1990; Murphy et al., 1990; Lawton and McMillan, 1999; Martín-Barajas et al., 1995; Busby et al., 2006; Centeno-Garcia et al., 2011). This is due to localization of extension or transtension in the thermally weakened arc axis, where rifting succeeds much more rapidly (in millions of years) than it does in intracontinental rifts (which take tens of millions of years; Umhoefer, 2011). The Sierra Crest–Little Walker volcanic center records the onset of Walker Lane transtensional rifting at ca. 12 Ma, through exploitation of the arc axis at a large volcanic center.
Controls of Pull-Apart Size and Magnitude of Extension on Arc Magma Compositions
The formation of the very large Sierra Crest–Little Walker pull-apart in the ancestral Cascades arc triggered rapid ascent of low-degree partial melts at the onset of Walker Lane transtension (Stanislaus Group), causing the outpouring of high-K2O lava (shoshonite, trachyandesite, and trachybasaltic andesite) as well as basalt. On a smaller scale, more basalt and high-K2O lava erupted from the southern part of the Sierra Crest–Little Walker pull-apart where transtension was greatest, relative to the northern part, which erupted more andesite and basaltic andesite. This is consistent with the interpretation that the magnitude of extension is important in eruption of low-degree partial melts and primitive melts. The Ebbetts Pass pull-apart, in contrast, is much smaller than the Sierra Crest–Little Walker pull-apart, consistent with the observations that it has only one basalt, few high-K2O rocks, and a much greater proportion of dacite, and it erupted rhyolites, which are not present in the Sierra Crest–Little Walker pull-apart. We attribute this to a higher degree of crustal contamination due to slower ascent of magmas in a smaller pull-apart structure. This is consistent with studies in the nearby Bodie Hills (Fig. 1), where minimal extension was accompanied by eruption of intermediate to silicic rocks, few basalts, and only one small-volume high-K2O lava (John et al., 2012).
All of the ancestral Cascades arc volcanic rocks described here, except for the Stanislaus Group, are compositionally identical to the unconformably underlying Mesozoic plutonic rocks of the Sierra Nevada Batholith; this indicates a similar petrogenesis, by shallow-level fractional crystallization, with no need for deep-seated pyroxenite cumulates. However, the petrogenesis of the Stanislaus Group requires a greater degree of fractionation of clinopyroxene and Fe-Ti oxides, and their major-oxide trends and clinopyroxene-rich modes might be the volcanic expression of processes that lead to pyroxenite cumulates. Such compositions are rare elsewhere in the Cenozoic ancestral Cascades arc (du Bray and John, 2011; du Bray et al., 2013) and the Mesozoic Sierra Nevada Batholith. In the ancestral Cascades arc, these compositions occur within a large pull-apart basin within the arc (Stanislaus Group). A similar example occurs in the Ryukyu arc, where voluminous (∼560 km3) late Miocene to Pleistocene high-K “flood andesites” erupted from fissures and filled the Hiatsu “volcano-tectonic depression” (Nagao et al., 1995, 1999; Miyoshi et al., 2010). This distinctive volcanic and geochemical style may be used to infer extreme extension or transtension within ancient arcs where the structural setting might otherwise not be known.
Still unknown, however, is how these magmatic plumbing systems respond to an evolving transtensional field. For example, we have argued previously that the high-K2O lavas represent the transport of deep-seated, low-degree partial melts (e.g., Putirka et al., 2012). However, most of the high-K2O lavas have just 3–4 wt% MgO, and so these are clearly not mantle-derived magmas (see Putirka, 2017). If the parent magmas contain >12 wt% MgO, then there has been considerable amounts of fractionation before these lavas reached the surface. Also, mafic rocks are similarly rare in the volcanic suites that preceded and followed the pulses of high-K2O volcanism. Because the lavas from the various suites are replete with clinopyroxenes, amphiboles, and plagioclase feldspars, they are a perfect target for thermobarometry, and so by measuring their mineral compositions, we can estimate the staging depths (and temperatures) from which these magmas were erupted. Our current model is that the evolutionary pattern of fractional crystallization was shallow-deep-shallow, as the system erupted lavas that transition from low-K2O to high-K2O and then back to low-K2O compositions. However, the evolution of a transtensional system may well impose a permanent change on the plumbing system or in some way affect how magmas are transported to the surface, or perhaps even eruption-triggering mechanisms (e.g., Putirka, 2017). In Putirka et al. (2012), we also showed that subduction-related geochemical signatures linger well past the end of subduction processes. The post-transtension system may similarly establish shallow crustal reservoirs that did not exist prior to the onset of transtension, but that remain connected to deep-seated conduits that continue to feed thermal energy to shallow depths. Mineral compositions will allow us to determine whether transtensional stress regimes continue to influence volcanic activity long after their stress regimes have passed.
Unzipping of a Ribbon Continent by Rift Tip Propagation in the Arc Axis
In some cases, a continental rift may exploit the length of the arc axis simultaneously, as in the case of the Gulf of California (Lizarralde et al., 2007). There, stalling of a large subducting (Farallon) microplate offshore of Baja California resulted in plate capture at ca. 12 Ma (Stock and Lee, 1994). A distinctly different geologic record has been created by subduction zone microplate interactions offshore of California. There, the Walker Lane transtensional rift tip has propagated northward with time in the ancestral Cascades arc axis since ca. 12 Ma, in concert with northward migration of the Mendocino triple junction (Busby, 2013; Smith et al., 2016). The study presented herein, of two pull-apart basins in the ancestral Cascades arc, provides local-scale confirmation of this regional-scale interpretation. In ancient arc-rift terranes, regional age patterns of the arc-to-rift transition (simultaneous vs. time-transgressive) may be used to help reconstruct the geologic record of microplate subduction.
Structural Controls on Volcanic Plumbing Systems and Coeval Surface Processes
Much attention has been paid to magmatic plumbing systems at active volcanoes by using geophysical and geochemical methods. However, the geologic contexts of continental arcs have been neglected relative to the study of their geochemical characteristics (Hildreth, 2007). Our geologic observations reveal the structural controls on magmatic plumbing systems at eroded extinct volcanoes, and their relation to coeval surface processes (Fig. 14). This required construction of detailed geologic maps, using the geochemical, petrographic, and geochronological data presented herein. We hope that the study presented herein convinces future workers of the importance of integrating laboratory data with detailed geologic mapping, which is the only way we might ever understand the interplay among tectonics, volcanic styles, and magma chemistry.
This research was supported by National Science Foundation (NSF) grants EAR-01252 (to Busby, P.B. Gans, and I. Skilling), EAR-0711276 (to K. Putirka and Busby), EAR-0711181 (to Busby), and EAR-1347901 (to Busby and S.M. DeBari), as well as U.S. Geological Survey EDMAP program awards 03HQAG0030, 06HQA061, and 09HQPA0004 (to Busby), and funds from the University of California, Santa Barbara, Academic Senate (to Busby). Formal reviews by Pat Cashman and Michael Petronis are gratefully acknowledged, and we are very thankful for extensive reviews and edits of the manuscript by David John. We also thank Geosphere Science Editor Shan da Silva for his patience.