Abstract

The Miocene Caliente-Enterprise zone (CEZ) in southwestern Utah (USA) is a 20–50-km-wide east-northeast–trending left-lateral transfer zone that displaces north-south–trending crustal blocks of the eastern Basin and Range Province to the west. Previous paleomagnetic results from the central and western CEZ show significant counterclockwise vertical axis rotations of strike-slip–bounded fault blocks, and these rotation estimates vary in magnitude both across and along the strike of the zone. Results of recent detailed geologic mapping and new geochronologic data in the area east of previous studies allow us to extend paleomagnetic studies into the easternmost CEZ. New paleomagnetic results include data from 4 regionally extensive latest Oligocene to early Miocene (ca. 24–22 Ma) ignimbrite sheets and from 3 ca. 22–20 Ma Iron Axis laccoliths. These data reveal significant magnitudes and similar spatially variable components of counterclockwise vertical axis rotation. Rotation estimates from the ignimbrites are assessed relative to what we interpret to be a nonrotated to only minimally rotated reference section just north of the Colorado Plateau (Grass Valley) and from several low-extension areas in southeast Nevada. Accepted tilt-corrected paleomagnetic data from sites away from the reference areas are discordant in declination from the expected individual ash-flow tuff directions, with rotation (R) and flattening (F) estimates that range from R = –2° to –84° and F = +15° to –14°. Many of the rotation estimates are large and statistically significant. For example, site P-18 from the Bauers Tuff yields an R = –61.1° ± 5.3° and F = –0.6° ± 5.0°. Relative to the expected Miocene direction, in situ paleomagnetic data from the Iron Axis laccoliths, specifically the Three Peaks laccolith, yield a mean that is discordant in declination, with estimated R = –22.2° and F = –8.8° values. These rotation and flattening estimates, although consistent with the overall data set from volcanic rocks, must be considered of lesser quality, as we are unable to accurately correct these data for possible effects of local tilting. If the rotation estimate is viable, then we suggest that this component of deformation involves much of the upper crust, and we furthermore propose that the boundary of the eastern CEZ extends farther east than previously envisioned, to within a few kilometers of the breakaway with the Colorado Plateau. The transitional zone between the eastern CEZ and Colorado Plateau is therefore abrupt and occurs within a narrow zone near Cedar City, Utah.

INTRODUCTION

Since the middle Cenozoic, continental lithospheric extension in the southwestern United States has resulted in the relative westward transport of crustal material away from unextended stable crust of the Colorado Plateau (Fig. 1). In general, the transition zone separating the extended and unextended regions trends north-south; however, in southwestern Utah the transition zone swings nearly 90° and trends roughly east-west, forming a major displacement transfer system known as the Caliente-Enterprise zone (CEZ) (Anderson and Mehnert, 1979; Hudson et al., 1998; Fig. 2A). Here we define the CEZ after Axen (1998) as the region encompassing all transverse structures whether or not they include evidence for counterclockwise vertical axis rotation. Displacement transfer systems, or accommodation zones, are often subvertical fault systems that transfer displacement from one region of the crust to another (Moustafa, 1976; Bosworth, 1985, 1986; Lister et al., 1986; Rosendahl, 1987; Chapin, 1989; Faulds et al., 1990; Faulds and Varga, 1998). Transfer zones accommodate or partition strain into areas of heterogeneous extension, and often are not simple strike-slip or oblique-slip fault systems, but tend to be associated with diffuse belts of magmatism and, at times, significant vertical axis rotation of fault-bound structural blocks (Faulds and Varga, 1998; Hudson et al., 1998; Petronis et al., 2002, 2007, 2009). The CEZ is a sinistral transfer zone that has undergone vertical axis rotation of spatially variable, yet systematic magnitudes (Hudson et al., 1998). Emplaced within the CEZ are numerous 22–20 Ma laccoliths of the Iron Axis magmatic province, a northeast-trending igneous feature in southwestern Utah characterized by several early Miocene laccolithic intrusions, including the Pine Valley megalaccolith and extensive latest Oligocene to earliest Miocene regional ignimbrites (Hacker et al., 1996, 1999; Axen, 1998). Previous paleomagnetic studies in the CEZ concentrated on rocks from the central and western part of its extent, predominantly because those areas were already mapped (Hudson et al., 1998). This study focuses on the more recently mapped areas to the east and southeast (Hurlow, 2002; Rowley et al., 2006; Biek et al., 2009). We report paleomagnetic data from late Oligocene and Miocene volcanic and intrusive rocks from the eastern CEZ, which is the area thought to record the earliest phase of deformation within the zone (Hacker, 1998). New paleomagnetic data from the eastern part of the CEZ allow us to assess the amount and variation of vertical axis rotation where the CEZ interacts with the Iron Axis laccolith province along the western margin of the Colorado Plateau and to better define the eastern extent of the zone.

GEOLOGIC SETTING

The CEZ trends east-northeast across the approximately north-south structural grain of the eastern Basin and Range Province for ∼220 km west from Cedar City, Utah, into southeastern Nevada (Fig. 2A). This structural zone, manifested as a right-stepping jog of the Colorado Plateau, separates the northern Basin and Range Province from the narrower, but more highly extended, central Basin and Range Province (Wernicke et al., 1988) and coincides with the northern boundary of the so-called amagmatic corridor (Eaton, 1982) of the central Basin and Range Province. Throughout its Cenozoic history, the CEZ was a zone of distributed sinistral shear that grew longitudinally westward after the early Miocene. This has been interpreted to reflect increased crustal ductility beneath the zone as the result of the ca. 23–12 Ma magmatism within the Caliente caldera complex (Hudson et al., 1998), in addition to the extension that occurred to the north of the adjacent Colorado Plateau during the middle Miocene to Holocene (Axen, 1998). Two episodes of major Miocene to Holocene extension probably occurred north of the CEZ and led to the current arrangement of basins and ranges. To the south, the Miocene to Holocene extensional history was highlighted by a single, large-magnitude extensional event on three west-rooted detachment faults (Axen, 1998).

The eastern section of the CEZ is just south of the Escalante Desert, where it forms a tilt domain boundary (Fig. 2A). Large-magnitude counterclockwise vertical axis rotations have been recorded in the westernmost part of this eastern section of the CEZ (Hudson et al., 1998). Here, Miocene volcanic rocks located south of the Escalante Desert are intersected by north-northwest– to west-northwest–striking dextral fault systems that were probably active during vertical axis rotation, resulting in as much as 45°–85° of counterclockwise rotation, as evidenced by paleomagnetic data from the volcanic rocks in the area (Hudson et al., 1998). The central section of the CEZ overlaps the Caliente caldera complex, the eruption history of which lasted from ca. 23 to 12 Ma. The Caliente caldera is centered in eastern Nevada; the eastern edge of the caldera extends to within 25 km of Enterprise, Utah. Early eruptions appear to have occurred before major extension to the north or south during the Miocene, whereas the latter part of the caldera’s activity occurred coincident with large-magnitude extension and detachment faulting to the north and south (Axen, 1998). The western section of the CEZ is defined by the Pahranagat fault zone (Tschanz and Pampeyan, 1970), which is composed of three northeast-striking sinistral faults and includes the area north of the fault. These faults have been interpreted to accommodate varying degrees of extension that was transferred from the Dry Lake Valley–Delamar Valley area to the Desert Valley west of the Sheep Range, where discrete faults of the system end (Liggett and Ehrenspeck, 1974; Wernicke et al., 1984; Hudson et al., 1998). Areas north of the Pahranagat fault zone were also shown by Hudson et al. (1995, 1998) to have rotated >15° counterclockwise.

STRATIGRAPHY, LACCOLITHS, AND GEOCHRONOLOGY

Ignimbrite Stratigraphy

Cenozoic volcanic rocks, dominated by Oligocene to Miocene ash-flow tuffs and lava flows, as well as intrusive rocks, are abundant and well exposed in southeastern Nevada and southwestern Utah. The volcanic stratigraphy of the eastern segment of the CEZ is, in part, the result of Iron Axis magmatic activity that produced an assortment of ash-flow tuffs, lava flows, and allochthonous gravity slide masses related to the rapid emplacement and subsequent eruption of some of the Iron Axis laccoliths (Hacker, 1998; Hacker et al., 2002). In addition, far-traveled ash-flow tuffs from the Caliente caldera that predated and postdated Iron Axis magmatism were deposited in the area, and bimodal volcanism continued throughout the late Cenozoic and was accompanied by regional extensional faulting. To assess the regional pattern of vertical axis rotation, we focused our sampling on four readily identifiable ignimbrites: the Leach Canyon Tuff, the Bauers Tuff Member of the Condor Canyon Formation, the Harmony Hills Tuff, and the ash-flow tuff member of the rocks of Paradise. The ignimbrites are lithologically distinctive and of the appropriate age to have undergone the main phases of Cenozoic deformation in the area (Table 1). The Leach Canyon and Bauers Tuffs originated from the Caliente caldera complex more than 100 km to the northwest; the Harmony Hills Tuff originated from the nearby Bull Valley Mountains; and the rocks of Paradise tuff erupted from the Pinto Peak intrusion (Hacker et al., 2002).

Iron Axis Laccoliths

The Iron Axis laccolithic group consists of a series of early Miocene calc-alkaline hypabyssal intrusions and associated volcanic rocks. The three laccoliths from which we collected paleomagnetic data (Three Peaks, Granite Mountain, Iron Mountain) are part of a latest Oligocene to earliest Miocene laccolith swarm that straddles the western edge of the Colorado Plateau in southwestern Utah and are exposed ∼30 km west of Cedar City (Fig. 2A). Prior to laccolith emplacement, several large-volume, regional Oligocene calc-alkaline andesite to rhyolite ash-flow tuffs of the Wah-Wah Springs Formation (ca. 30 Ma), Isom Formation (ca. 27 Ma), and Quichapa Group (ca. 24–22.5 Ma) were erupted over the area from sources to the northwest and west. This sequence of pre–Iron Axis volcanic rocks overlies fluvial and lacustrine sedimentary rocks of the late Paleocene–Oligocene Claron Formation, which unconformably overlies Cretaceous and Jurassic sedimentary rocks deformed during the Sevier orogeny. The igneous rocks are part of the middle Cenozoic calc-alkaline igneous sequence that spans much of the western United States. These intrusions are considered to be associated with oblique convergence during subduction of oceanic lithosphere beneath western North America that produced large fluxes of mantle-derived mafic magma intruded into the overlying continental lithosphere (Johnson, 1991; Nelson and Davidson, 1998; Rowley, 1998). Emplacement of the Iron Axis laccoliths was focused along one or more Sevier-age thrust faults that displace Mesozoic and Cenozoic sedimentary strata at upper crustal levels in this area (Mackin, 1960; Blank and Mackin, 1967; Blank et al., 1992; Rowley et al., 1995; Rowley, 1998). The Three Peaks, Granite Mountain, and Iron Mountain laccoliths were emplaced into gently tilted limestone strata of the Jurassic Carmel Formation or the Temple Cap Formation. Emplacement into carbonate strata of the Carmel Formation produced several replacement magnetite, hematite, hemo-ilmenite, and ilmeno-hematite deposits that were extensively mined and together constitute the largest iron ore production in the western United States (Mackin, 1960; Blank and Mackin, 1967; Barker, 1995). The studied intrusions are quartz monzonite to granodiorite porphyries with phenocrysts of plagioclase (andesine-labradorite), biotite, hornblende, and/or pyroxene (diopsidic augite), and magnetite in a groundmass (∼33%–50% total volume) of very fine grained quartz and potassium feldspar (Petronis et al., 2004). Mafic nodules are rare and generally <5 cm along their long axis. The intrusions yielded K-Ar mineral dates and 40Ar/39Ar mineral age spectra determination values of 22–20 Ma (Armstrong, 1972; Hacker et al., 1996; McKee et al., 1997; Rowley et al., 2006), indicating a 2 m.y. period of Iron Axis magmatic activity (Petronis et al., 2004).

Magnetic Oxide Petrology of the Iron Axis Laccoliths

The magnetic oxide mineralogy of the Iron Axis laccoliths provides important information on the cooling rate, temperature, and chemistry of the magma and can reflect the igneous and hydrothermal history as revealed by their composition, morphology, and textural relationships to the silicate mineral phases (Lindsley, 1991; Speer and Becker, 1992, and references therein). In turn, combining these observations with detailed rock magnetic data provides a powerful tool to fully characterize the cooling history and postemplacement alteration of the intrusions (e.g., Alva-Valdivia et al., 2001; Lagrou et al., 2004; Uehara et al., 2010; Petronis et al., 2011). The distributions of opaque grains in the Iron Axis laccoliths are somewhat typical of many felsic to intermediate intrusions where subequant grains (50–100 μm) occur predominantly as interstitial material in the silicate framework in volume concentrations of 0.5%–2.0% (Fig. 3A). The opaque grains are typically Fe-Ti oxides and closely associated spatially with biotite and hornblende crystals, a textural relationship seen in other shallow igneous intrusions (Speer and Becker, 1992; Stevenson et al., 2007; Petronis et al., 2011). The grains occur as roughly equidimensional crystals to 100 μm in size and also as very fine material distributed throughout the groundmass (Figs. 3A, 3B). The Fe-Ti oxides include an abundance of titanomagnetite and titanomaghemite that in some cases have altered to hemo-ilmenite or ilmeno-hematite, and rare unoxidized magnetite (Fig. 3C). Many relatively coarse magnetite grains show evidence of apparent high-temperature oxidation exsolution (Buddington and Lindsley, 1964) and the formation of abundant ilmenite lamellae with a classic trellis oxidation exsolution pattern (e.g., similar to the Widmanstatten texture often seen in iron meteorites) (Figs. 3C, 3D). The oxy-exsolved lamellae show sharp, well-defined contacts with their titanomagnetite hosts and appear smooth in outline. The ilmenite lamellae are typically concentrated along cracks, around silicate inclusions, and along the titanomagnetite grain boundaries (Fig. 3D). These textural features are often interpreted as evidence that low-temperature subsolidus oxidation and not magmatic oxidation exsolution is responsible for the formation of ilmenite lamellae from a primary titanomagnetite phase (Lindsley, 1991, and references therein).

Petrographic analysis, paleomagnetic, and rock magnetic data indicate there are at least two separate magnetic mineral assemblages that contribute to the remanent magnetization of each intrusion; it is interesting that one of the magnetic mineral assemblages dominates at each site. The first assemblage is characterized by nearly pure magnetite to low-Ti titanomagnetite. The minerals associated with magnetite-dominated sites are hematite, plagioclase, quartz, biotite, epidote, and hornblende. Magnetite grains typically follow grain boundaries, especially as a product of oxidation of hornblende (Fig. 3B). The amount of hematite replacement varies from site to site. In some cases, there is almost no hematite present, yet at some sites it constitutes as much as 50% of the oxide area. The second magnetic mineral assemblage is dominated by ilmeno-hematite (Fig. 3C). The minerals characteristic of the ilmeno-hematite–dominated sites include plagioclase, quartz, biotite, white mica, pyroxenes, chlorite, and hornblende. The ilmeno-hematite crystals contain lamellae of hemo-ilmenite (ferroan ilmenite) that are mostly aligned parallel with the long axis of the silicate minerals that are host to these grains, most commonly biotite and hornblende (Fig. 3C). The presence of ilmeno-hematite with hemo-ilmenite lamellae is unusual for rocks of this composition, possibly indicating slow cooling (Carmichael, 1961), which is unexpected considering the field evidence for very shallow emplacement of the laccoliths (e.g., abundant drusy or miarolitic cavities). The coarse size of the ilmeno-hematite phase indicates an exsolution temperature between 570–600 °C, just above the Curie temperature of phases of this composition, likely allowing the ilmeno-hematite grains to acquire a thermoremanent magnetization (TRM) and retain a remanence that is comparable in direction with those sites dominated by magnetite-titanomagnetite assemblages (Carmichael, 1961).

Geochronology

Historically, K/Ar age determinations have provided the main age constraints on the Cenozoic volcanic stratigraphy and Iron Axis magmatism of this region. To improve our understanding of the timing of large magnitude pyroclastic volcanism in the area, we obtained 40Ar/39Ar incremental release age determinations on plagioclase from two tuff units previously dated by the K/Ar method. Separates were analyzed using standard procedures at the University of Nevada, Las Vegas Center for Geochronology (methods described and analytical data reported in Cornell, 2005; Cornell et al., 2001). The age of the Harmony Hills Tuff (Quichapa Group), a key unit deformed by all intrusions and emplaced across the CEZ, has been only poorly defined by 6 previously reported K/Ar dates ranging from ca. 24.4 to 20.3 Ma (Best et al., 1989). We obtained a well-defined plateau age (8 steps constituting 91% of the total 39Ar released) of 22.03 ± 0.15 Ma (Fig. 4). This date is indistinguishable from a 21.93 Ma 40Ar/39Ar biotite date reported for the immediately overlying ash-flow tuff member of the rocks of Paradise (Hacker et al., 1996, 2002). We also obtained a plateau age of 21.83 ± 0.17 Ma (4 steps, 55% total 39Ar) and an associated concordant isochron age of 21.46 ± 0.40 Ma on the Rencher Formation (Fig. 4). This new age is consistent with a 21.93 Ma 40Ar/39Ar age reported by Hacker et al. (1996) for the rocks of Paradise tuff directly underlying the Rencher Formation. Collectively, these data better define the age of key units sampled for this study (Table 1) and show that widespread explosive magmatism took place in the study area over a short time interval.

MOTIVATION AND REFERENCE LOCATION

The purpose of this study is to expand upon the research of Hudson et al. (1998) in a part of the CEZ in southwestern Utah. Until recently, this area was poorly studied due to the limited availability of geologic maps documenting field relations in the easternmost part of the zone. Hudson et al. (1998) suggested that the CEZ did not extend much farther than ∼10 km east of Newcastle (Fig. 2A). Many of our sampling sites, however, are located along the far eastern margin of the zone in an area that exhibits a pattern of faulting very similar to that of the central and western part of the CEZ (Butler et al., 2001). Hudson et al. (1998) documented some counterclockwise vertical axis rotation within the Harmony Mountains and Desert Mount areas between Newcastle and Cedar City and reported localized counterclockwise rotations of 20°–45° ∼10 km south of Cedar City (Fig. 2A). This study spans an ∼35 km × 70 km area north of the early Miocene Pine Valley megalaccolith in the eastern CEZ. The Pine Valley intrusion is within the stable unextended Colorado Plateau, and therefore serves as a marker for the southern boundary of the CEZ in this region.

Paleomagnetism Reference Section

The ignimbrites sampled in this study were erupted across the Pine Valley region and adjacent areas and were subsequently faulted, with the formation of numerous structural blocks during Cenozoic extension and the development of the CEZ. Observed paleomagnetic declinations that deviate or are discordant from some expected value can provide rotation estimates in either an absolute or a relative framework. Absolute rotation determinations require a result that adequately samples the geomagnetic field over a sufficiently long time. For example, a slowly cooled, large volume intrusion or a sequence of numerous basalt flows showing multiple polarity zones and directional dispersion that is consistent with an adequate sampling of paleosecular variation is typically regarded as capable of averaging paleosecular variation. If significant time has elapsed, the result is then compared with a sufficiently robust estimate of the time-averaged field (paleomagnetic poles) based on independent paleomagnetic data from the respective craton. Relative rotation determinations, however, require sampling a single, laterally extensive datum such as a regionally extensive ash-flow tuff (Wells and Hillhouse, 1989; Byrd et al., 1994; Petronis and Geissman, 2008; Sussman et al., 2011). Paleomagnetic investigations of ignimbrites reveal that individual cooling units typically yield an instantaneous record of the geomagnetic field (e.g., Reynolds, 1977; Geissman et al., 1982; Wells and Hillhouse, 1989; Byrd et al., 1994). Thus, combining results from a small number of discrete ash-flow tuffs will not provide an adequate long-term average of the geomagnetic field. Consequently, our analysis of the data from the ignimbrites we studied involved comparing results from each tuff to data obtained from what we interpret to be internally coherent reference sections. Two reference areas were chosen (Figs. 1 and 2A). One area encompasses the southern Delamar and Meadow Valley Mountains and farther north in Condor Canyon in southeastern Nevada (Fig. 1) (Hudson et al., 1995, 1998). This area provides reference locations to base vertical axis rotation estimates for the Harmony Hills and Bauers Tuffs. The locations were selected because the southern Delamar and Meadow Valley Mountains are within a zone of mild extension (Wernicke et al., 1988; Scott et al., 1995) and Condor Canyon, southeast of Pioche, Nevada, provides exposures of numerous Oligocene to early Miocene ash-flow tuffs. Hudson et al. (1998) showed that the combined results from the Delamar and Meadow Valley Mountains and Condor Canyon provide a time-averaged paleomagnetic result that indicates no significant rotation with respect to the North American craton. Our second reference area is located in Grass Valley, just west of the Stoddard Mountain laccolith (Fig. 2A). We selected Grass Valley as a reference area for the rocks of Paradise tuff and Leach Canyon Tuff because they were available for sampling in close proximity to each other and because the nearby Stoddard Mountain laccolith shows little evidence of vertical axis rotation relative to the expected Miocene field direction (Petronis et al., 2004). Paleomagnetic data from the Stoddard Mountain laccolith yielded a time-averaged remanence direction (declination, D = 351.1°, inclination, I = 57.8°, α95 = 3.7°) that is statistically indistinguishable from a late Miocene expected field direction (D = 358°, I = 58°, α95 = 6.3°; Mankinen et al., 1987; Fig. 5) implying little, if any, vertical axis rotation of the reference area (rotation, R = –6.9° ± 11.3°; flattening, F = 0.2° ± 7.1°). In Grass Valley, the average dip of the ignimbrite deposits, based on the orientation of eutaxitic fabrics and contacts between units, is ∼5° (Rowley et al., 2006). Given the low dips of these units and the proximity of this section to the stable Colorado Plateau margin, we assume that these rocks underwent minimal vertical axis rotation and that the current dip of the deposits is related to recent extension responsible for the present physiography of the region.

All paleomagnetic data from the ignimbrites, following correction for local tilt based on orientation of eutaxitic structures, contacts between volcanic units, or stratification in overlying or underlying volcaniclastic sedimentary deposits, are compared to data from these reference sections (Fig. 5). The Grass Valley reference directions for this study were established in the Leach Canyon Tuff and the rocks of Paradise tuff and the Bauers and Harmony Hills Tuffs are referenced to the Delamar and Meadow Valley Mountains and Condor Canyon (Hudson et al., 1998). The reference directions for each tuff were computed by averaging several mean directions for sites within the reference area (Tables 2A, 2B; Fig. 5). With the exception of the Leach Canyon Tuff, each ignimbrite at the reference section has a unique characteristic remanent magnetization (ChRM) that is well defined and of low dispersion. The average reference directions, following correction of dip, for the ignimbrites are as follows. The Harmony Hills Tuff yields a mean direction of D = 352.1°, I = 50.3°, the Bauers Tuff D = 349.3°, I = 44.3°, and the rocks of Paradise tuff D = 198.3°, I = 20.1° (Table 2C; Fig. 5). The ChRM of the Harmony Hills and Bauers Tuffs are interpreted to be of normal polarity. The rocks of Paradise tuff is interpreted to be of intermediate polarity as the virtual geomagnetic pole (VGP) derived from the ChRM mean direction is >45° from the North American paleomagnetic pole positions for early Miocene time. The well-defined ChRM of these three ignimbrites affords the opportunity for high-precision estimates of declination discrepancies across the eastern CEZ; any deviation of the ChRM at individual sites compared to the reference section, when inclinations are essentially identical, is interpreted as reflecting vertical axis rotation. The Leach Canyon Tuff yields a mean direction of D = 066.0°, I = –10.7°, also consistent with an intermediate polarity state. The directional data from the Leach Canyon Tuff are highly variable across the study area and we place little confidence in the overall results from this ignimbrite (see following). All paleomagnetic data from the Three Peaks, Granite Mountain, and Iron Mountain laccoliths are compared to the middle Miocene expected field direction (D = 358°, I = 58°, α95 = 6.3°; Mankinen et al., 1987) and provide an absolute rotation estimate relative to this direction.

METHODS

Field Methods

Samples were drilled at each site over an area of ∼10 m2, and for stratified materials at least 5 strike and dip measurements were obtained for each flow. All samples were collected using a portable gasoline-powered drill with a nonmagnetic diamond tip drill bit (Fig. 2A). Paleomagnetic data from 61 ignimbrite sites were collected along 25–35-km-wide (i.e., varying width along the traverse) north-south transverse north of the Pine Valley Mountains (Fig. 2B). Paleomagnetic data from the intrusions were collected from sites across the best exposed and accessible parts of the laccoliths (Fig. 2C). The site mean results for the ignimbrites and intrusions are shown in Figures 6 and 7, respectively, and associated statistics are listed in Tables 3 and 4, respectively. All laboratory and rock magnetic methods are shown in Appendix 1 and Appendix 2 and the rock magnetic results are presented in Appendix 3.

Ignimbrites

Of the 61 sites, 49 yield readily interpretable results. These include 13 sites from the Harmony Hills Tuff, 16 sites from the Bauers Tuff, 11 sites from the Leach Canyon Tuff, and 9 sites from the rocks of Paradise tuff. Twelve sites were eliminated from further consideration because they did not yield stable end-point behavior and/or had high within-site dispersion (see following). Of the 49 sites considered acceptable for further analysis, seven were established in the Grass Valley reference locality (Fig. 2B; Table 3); thus 42 sites in the ash-flow tuffs provide relative rotation estimates with respect to the reference locations across the CEZ.

Laccoliths

A total of 82 paleomagnetic sites were established in the Three Peaks, Granite Mountain, Iron Mountain laccoliths (Fig. 2C). The site mean results are shown in Figure 7 and associated statistics are listed in Table 4. These include 46 sites from the Three Peaks laccolith, 14 sites from the Granite Mountain laccolith, and 22 sites from the Iron Mountain laccolith. Of the 82 sites collected, 53 sites were used to calculate a grand mean for the 3 intrusions (Table 4). From the 46 sites analyzed in the Three Peaks laccolith, 39 yield readily interpretable results with 35 sites used to calculate a group mean result. Seven sites were rejected as they did not yield stable end-point behavior and were likely lightning struck. The remaining four sites excluded from the group mean have well-defined site mean directions, but they are separated by greater than two angular standard deviations from the overall group mean of the other 35 sites. For the Granite Mountain sites, eight sites yield interpretable data with seven sites used to calculate a group mean for the intrusion. The six rejected sites did not yield stable end-point behavior and showed evidence of significant hydrothermal alteration. The remaining one site rejected yielded a well-defined reverse polarity direction, but upon inversion through the origin, plotted greater than two angular standard deviations from the overall group mean. Of the 22 sites from the Iron Mountain laccolith, 16 yield stable end-point behavior with well-defined magnetization directions with 11 sites used to calculate a group mean direction. The six rejected sites did not yield stable end-point behavior and the five sites excluded from the group mean calculation were considered outliers because they were greater than two angular standard deviations from the overall group mean for the intrusion. Collectively, the grand mean calculated for the Iron Axis laccoliths differs from the Miocene age expected direction for this area based on the paleomagnetic pole estimates for North America for this time period (Mankinen et al., 1987).

Rock Magnetism

To characterize the magnetic mineralogy, we conducted a suite of standard rock magnetic experiments with the principal goal of identifying the mineral phases that carry the overall remanence and the quantity, composition, domain state, and grain size of the magnetic phases present. These tests included: (1) analysis of low-field susceptibility versus temperature, (2) alternating field (AF) demagnetization of anhysteretic remanent magnetization (ARM), (3) direct current (DC) acquisition of a saturation isothermal remanent magnetization (IRM), (4) DC demagnetization of the saturation IRM to yield a backfield IRM (coercivity of remanence), and (5) AF demagnetization of saturation isothermal remanent magnetization (SIRM). The rock magnetic methods are discussed in Appendix 2.

RESULTS

General Demagnetization Behavior

Demagnetization response during progressive AF demagnetization for most samples is generally well behaved, depending on rock type, and characterized by high-quality results (Fig. 8). Of the 61 sites established in the ignimbrites, 49 sites yield readily interpretable demagnetization data (Table 3). The rejected sites had high within-site dispersion (α95 > 15°, k < 15) and did not yield interpretable results; these sites are not discussed further. In general, most samples contain a single ChRM that is well grouped at the site level, but some samples also contain additional magnetizations that are readily removed by 20 mT or by 300 °C; this behavior is generally restricted to the Iron Mountain and Granite Mountain laccoliths. We interpret these low-coercivity, randomly oriented magnetization components as viscous overprints (VRM). After removing the VRM, the ChRM, which we interpret as the primary TRM, decays along a roughly univectorial path to the origin with <10% of the natural remanent magnetization (NRM) intensity remaining after treatment (Fig. 8A). For the Iron Axis laccoliths, 82 sites were analyzed with 53 sites used to calculate grand mean directions for each of the laccoliths. Some sites responded well to AF demagnetization, whereas other sites required a combination of AF and thermal (TH) demagnetization to isolate the ChRM. The sites that responded to AF demagnetization contain magnetite-titanomagnetite as the principal magnetic phase carrying the remanence of these rocks, consistent with the rock magnetic results (Appendix 3). In contrast, those sites that required both AF and TH demagnetization to isolate the ChRM are dominated by ilmeno-hematite with hemo-ilmenite lamella as the principal magnetic phase (Fig. 8B). Typical maximum laboratory unblocking temperatures for the ilmeno-hematite-dominated sites are ∼555 °C, with a range of maximum laboratory unblocking temperatures of 530–570 °C.

Ash-Flow Tuffs

The Leach Canyon Tuff yields a low-coercivity, generally single component magnetization that is well grouped at the within-site level (Fig. 8A). The ensemble of estimated site mean directions is highly variable, in particular in inclination values, across the sampled region and the pattern of estimated site mean directions does not reveal a readily identifiable trend (Fig. 6). Median destructive field values range from 10 to 20 mT with <25% of the NRM intensity remaining after AF treatment to 120 mT (Fig. 8A; Appendix 3). Thermal demagnetization of representative specimens to 500 °C did not isolate any additional magnetization components and reduced the intensity to <10% of the NRM value. The Harmony Hills Tuff in the reference area yields well-grouped, low-coercivity north-northeast declination and steep positive inclination demagnetization data that decay along a roughly linear path to the origin (Fig. 8A). The data are well grouped at the site level with <25% of the NRM intensity remaining after AF treatment to 120 mT. For higher coercivity samples, TH demagnetization to 585 °C yielded no change in direction data. The estimated site mean data vary in direction systematically across the sampled region (Fig. 6). The Bauers Tuff yields high-coercivity north-northwest declination and steep positive inclination data in the reference area, although >50% of the NRM intensity remains after AF treatment to 120 mT (Fig. 8A). Selected treatment of high-coercivity specimens with TH demagnetization to 575 °C results in a roughly univectorial decay path to the origin and does not isolate any additional magnetization components. The estimated site mean data vary in direction systematically across the sampled region (Fig. 6). The rocks of Paradise tuff in the reference location yields a predominantly low-coercivity northeast declination and shallow positive inclination that is readily removed by 10 mT, leaving a characteristic southwest declination and shallow positive inclination direction. Median destructive field ranges from 10 to 50 mT (Fig. 8A). The remaining component, defined by ∼30% of the NRM intensity, decays along a near linear path to the origin and is well grouped at the site level with <10% of the NRM intensity remaining after AF treatment to 120 mT. TH demagnetization to 500 °C of duplicate specimens did not isolate any additional magnetization components.

Three Peaks, Granite Mountain, and Iron Mountain Laccoliths

The paleomagnetic behavior of the Iron Axis laccoliths can be readily divided into two types: (1) a well-defined response to AF demagnetization over a broad range of peak fields, and (2) a very limited response to AF demagnetization, thus requiring TH demagnetization to isolate the ChRM. In sites that responded well to AF demagnetization, the principal mineral assemblage carrying the ChRM is dominated by magnetite and titanomagnetite that are partially oxidized to hematite, as shown by backscatter electron microprobe images (Figs. 3A, 3B, 3D). The second group of sites, all from parts of the Three Peaks laccolith, is characterized by a very limited response to progressive AF demagnetization, yet complete laboratory unblocking in thermal demagnetization by ∼555 to 580 °C, although a few sites required TH demagnetization to 680 °C to fully isolate the ChRM. In sites characterized by very high coercivities and that required TH demagnetization, the principal mineral phase is ilmeno-hematite with hemo-ilmenite lamella (Fig. 3C).

Field Tests

Where field exposures allowed, several baked contact tests (Everitt and Clegg, 1962) were attempted between the host Jurassic rocks and the intrusions. A positive test is characterized by the remanence direction of the igneous rock being acquired by the host rock and progressively dominated by the ChRM of the host rock (if any) as a function of proximity to the igneous contact. All contact tests failed to show a gradational change between a magnetization consistent with an expected Miocene direction for the laccoliths and that of the Jurassic host rocks. We attribute this result to the poor paleomagnetic character of the host rock. Sites in the host rock located well away from intrusion contacts did not yield stable end-point behavior, and we interpret this result to indicate that the Jurassic host rocks do not preserve a geologically stable remanent magnetization. Host rock obtained within a few centimeters of any contact with intrusive rock carries a magnetization having a direction similar to that of the laccolith. The demagnetization response of the host rock, however, is erratic, lacking any stable end-point behavior, for any samples obtained beyond a few centimeters from the contact. We interpret these results, although somewhat tenuously, to indicate that the laccoliths preserve a geologically stable remanence and that host rock at least immediately adjacent to the intrusion margin acquired a magnetization parallel to that of the laccolith.

Three Peaks Results

Paleomagnetic data from the 35 accepted sites normal polarity yield an in situ group mean direction of northwest declination and moderate positive inclination (D = 332.9°, I = 51.7°, α95 = 3.4°, k = 52.9). If we can assume that the magnetization characteristic of the Three Peaks laccolith adequately averages a Miocene geomagnetic field, the group mean result differs, in both declination and inclination, from the expected Miocene direction (D = 358°, I = 58°) (Fig. 7A) with an inferred rotation (R) and flattening (F) estimate (Beck, 1980; Demarest, 1983) of R = –22.2° ± 6.4° and F = 8.8° ± 4.3° relative to expected Miocene age directions based on the synthetic pole and North American paleomagnetic pole for the Miocene (Mankinen et al., 1987; Besse and Courtillot, 2002). Using paleosecular variation (PSV) models (e.g., Merrill and McElhinny, 1983) for the average paleolatitude of the study area (∼37.7°N), the VGP angular standard deviation is predicted to be ∼15.8°. The VGP angular standard deviation of the 35 sites in the Three Peaks laccolith is 13.1°, which is within error of the predicted PSV dispersion range and is consistent with the sampling of sufficient field variations. The lack of reverse polarity magnetizations in the Three Peaks laccolith may simply reflect the fact that the early Miocene (22–20 Ma) is characterized by several normal and reverse polarity chrons (e.g., Cande and Kent, 1995) and magnetization blocking took place during a single normal polarity chron. We argue that the magnetization characteristic of the Three Peaks laccolith provides an adequate average of the paleomagnetic field.

Granite Mountain and Iron Mountain Results

Of the 36 sites analyzed from the Iron Mountain (N = 22) and Granite Mountain (N = 14), 24 yield readily interpretable results. Paleomagnetic data from the 11 sites used to calculate a group mean from the Iron Mountain laccolith yield a mean direction of west-northwest declination and moderate positive inclination (D = 303.6°, I = 60.6°, α95 = 12.9°, k = 14.9, N = 4R, 7N) that is statistically distinct from the expected Miocene direction with an inferred R of –54.4° ± 28.4° and F = –2.6° ± 14.3° (Fig. 7B). The dispersion of the VGPs from Iron Mountain is 26.6°, which is well beyond the predicted range for the site (see following). The Iron Mountain data do not pass the reversal test at 95% confidence (McFadden and McElhinney, 1990). Paleomagnetic data from the 7 accepted sites from the Granite Mountain laccolith yield a group mean direction of north-northwest declination and moderate positive inclination (D = 353.2°, I = 45.2°, α95 = 14.1°, k = 22.5, N = 7N) that is statistically indistinguishable from the expected Miocene direction with an inferred R of –4.8° ± 21.6° and F = 12.8° ± 15.3° (Fig. 7C). The dispersion of the VGPs from Granite Mountain is 16.3°, which is within error of the predicted range for the site (see following).

DISCUSSION

Interpretation of the Paleomagnetic Data

With the exception of the Leach Canyon Tuff, each tuff in the study area has a characteristic magnetization that is distinct in direction and, overall, the inclination of the characteristic magnetization maintains a consistent value with respect to the paleohorizontal, an observation that is consistent with previous observations of the paleomagnetics of ash-flow tuffs (e.g., Geissman et al., 1982; Wells and Hillhouse, 1989; Byrd et al., 1994; Best et al., 1995; Hudson et al., 1994), and is considered a requirement in using paleomagnetic data from ignimbrites to estimate magnitudes of vertical axis rotation. Therefore, the data from dispersed sites from a regionally extensive ignimbrite, when accurately referenced to the paleohorizontal, can be compared to estimate relative vertical axis rotations with respect to the reference location (Fig. 1). In principle, data from all tuffs, regardless of their age, exposed at the same locality should yield similar relative rotation estimates if rotation postdated their emplacement. The paleomagnetic data from the ignimbrites (excluding the Leach Canyon Tuff) across the eastern CEZ reveal an orderly and predictable pattern of discordant directional data with respect to the reference location sites (Fig. 9). The rock magnetic data and highly dispersed paleomagnetic results from the Leach Canyon Tuff indicate that this unit is not reliable to use to estimate vertical axis rotation across the study area. All VGPs determined from directions from the Leach Canyon Tuff are more than 45° from North American paleomagnetic poles for the early Miocene, and it is likely that this deposit recorded rapid changes of the geomagnetic field during a transition or other form of high-amplitude field event (Verosub and Banerjee, 1977; Merrill and McFadden, 1994). In addition, the low coercivity of the remanence typical of each site (Appendix 3) further indicates that the results from the Leach Canyon Tuff are not reliable (e.g., Geissman et al., 1982; Hudson et al., 1998).

Typically, the data from the three tuffs sampled in stratigraphic succession at the same general locality yield internally consistent declination and inclination results as well as similar magnitudes of rotation, although there are some inconsistencies, largely restricted to the Leach Canyon Tuff (Fig. 8). In the southern part of the study area, which includes the Pine Valley and central regions, the site mean data are all essentially concordant with reference directions for each ash-flow tuff unit, thus showing little evidence of vertical axis rotation. In contrast, site mean data from tuffs exposed to the north and east, including the Pinto, Holt Canyon, and Newcastle Reservoir localities, exhibit a progressive northward increase in the magnitude of counterclockwise rotation (Fig. 9). The paleomagnetic data from all locations north and northwest of the Grass Valley reference area are discordant in declination from the individual tuff expected directions (Table 3). The rotation and flattening estimates (Beck, 1980; Demarest, 1983) range from R = +38° to –84° and F = +15° to –14° (most F values are considerably lower and not statistically significant). Of the 49 site mean directions, 10 yield rotation estimates that we consider to be of relatively low confidence. The low-confidence sites include seven from the Leach Canyon Tuff, two from the Bauers Tuff, and one from Harmony Hills Tuff. We consider low-confidence rotation estimates to be associated with sites that yield a statistically significant F estimate that is 10° greater than the delta F error (see Petronis et al., 2009) (Table 3). The F values for data from the identical ignimbrite should be ∼0°, assuming that the ChRMs at each site are accurately referenced to the paleohorizontal and have been adequately isolated during progressive demagnetization (e.g., Wells and Hillhouse, 1989). One explanation for flattening values that are significantly greater than their associated error is that the data are inadequately referenced to the paleohorizontal. Other sources of error include not fully isolating the ChRM of the individual samples at a site, lightning strikes, unusual rock magnetic properties of a particular phase (Geissman et al., 1982), or inclination flattening associated with the development of a strong compaction foliation (e.g., Uyeda et al., 1963; Gattacceca and Rochette, 2002). An additional explanation for this phenomenon is that the ignimbrite acquired a remanence over a protracted period of time, possibly while the geomagnetic field direction was rapidly changing. It is likely that a combination of these factors contributes to the low confidence estimates for these 10 sites (Table 3), especially those in the Leach Canyon Tuff. With the exception of the Leach Canyon Tuff, for the ignimbrites sampled, after structural correction, the overall dispersion of site mean directions is far greater in declination than in inclination (Fig. 6), as would be expected if a single, sheet-like deposit, with an overall uniformly directed magnetization, is fragmented and distorted by different magnitudes of vertical axis rotation (e.g., Hillhouse and Wells, 1991; Petronis et al., 2009). The paleomagnetic data from the ignimbrites, when compared with respective data from the reference sections, indicate an organized pattern of vertical axis rotation (Fig. 9; Table 3), in that rotation magnitudes vary from <1° for sites located in the south to a maximum of –84° for sites located in the north, near Newcastle (Fig. 9). Four sites yield clockwise vertical axis rotations, when compared with respective data from the reference section. Three of these sites are from the Leach Canyon Tuff, and we question the overall reliability of this pyroclastic deposit to provide any form of representative geomagnetic field direction. The clockwise rotation estimate from the Bauers Tuff is statistically insignificant (R = 8.0° ± 6.6°), yet we cannot entirely discount the possibility that some areas may have undergone localized clockwise or very minimal rotation, and these could reflect local crustal heterogeneities and/or unusual fault geometry, although we suspect this conclusion to be tenuous (e.g., Lamb, 1987).

The paleomagnetic data from 53 accepted sites in the Iron Axis laccoliths, all located northeast of the Newcastle Reservoir sampling locality, yield a combined overall in situ grand mean direction of northwest declination and moderate positive inclination (D = 331.3°, I = 53.4°, α95 = 4.3°, k = 21.6). This approach to evaluating the paleomagnetic data from the three laccoliths we studied is likely inappropriate, but serves as an initial examination. Several individual site results from the Iron Mountain and Granite Mountain laccoliths are less robust, given the pervasive alteration of these laccoliths, and we place less confidence in the rotation estimates based on average mean directions from these intrusions. Therefore, we base our rotation estimate for the area northeast of the Newcastle Reservoir exclusively on the results from the Three Peaks laccolith.

The lesser quality of the paleomagnetic results from Granite Mountain and Iron Mountain intrusions may reflect the fact that each intrusion is pervasively hydrothermally altered. Most sites contain multiple magnetization components that are isolated across a spectrum of peak laboratory unblocking temperatures and applied fields. A few sites, however, respond to AF demagnetization and often yield a single component magnetization that decays linearly to the origin with <10% of the NRM remaining after treatment to 120 mT peak fields after a weak VRM is removed by 20 mT (Fig. 8B). These sites yield directional data similar to the results from the Three Peak laccolith and Curie point estimates consistent with magnetite and/or titanomagnetite as the dominant magnetic phase (Table 5; Appendix 3). We argue that the sites that respond well to AF demagnetization and yield directional data consistent with the Three Peaks laccolith carry a ChRM that is likely a primary TRM. Samples from sites that require a combination of AF and TH demagnetization often reveal the presence of a high-coercivity phase (likely ilmeno-hematite or hematite) that decays to the origin at peak temperatures ranging from 590 to 660 °C and yields directions similar to those from the Three Peaks intrusion. It is probable that for these sites, we are successfully isolating the primary ChRM of the rocks, which has been partially overprinted by a magnetization associated with late-stage hydrothermal alteration. Unfortunately, many of the sites analyzed from the Granite Mountain and Iron Mountain laccoliths yielded a complex demagnetization response reflecting overlapping coercivity spectra and mixed polarity results at the site level. We interpret this demagnetization response to indicate that these sites have complex magnetization acquisition histories and no longer contain a primary ChRM component.

The 35 accepted sites from the Three Peaks laccolith yield an in situ group mean direction of northwest declination and moderate positive inclination (D = 332.9°, I = 51.7°, α95 = 3.4°, k = 52.9, N = 35N) that is discordant to the expected Miocene direction with an inferred R = –22.2° ± 6.4° and F = 8.8° ± 4.3°. This result is consistent with previously published rotation estimates from nearby ash-flow tuffs (e.g., estimates of R = –10° to –20° from the Bauers and Harmony Hills Tuffs near Desert Mound, although some sites are not rotated) (Figs. 7A and 9; Hudson et al., 1998). Efforts to estimate vertical axis rotation based on paleomagnetic data from intrusions are compromised because of the inherent inability to accurately reference the data to the paleohorizontal, considering the possibility that postmagnetization acquisition local tilting has changed the original direction of magnetization in the pluton (e.g., Beck, 1980). In the case of intrusions that have a relatively uniform cooling history, and do not represent substantially deep crustal levels, paleomagnetic data that are discordant from expected field directions of comparable age can be reconciled in an infinite combination of rotations about axes of any orientation. Realistically, however, the tilting of crustal materials generally is accommodated by rotation about horizontal to subhorizontal tilt axes. To investigate the possibility of whole-scale tilting of the Three Peaks laccolith, we considered structures in the general area that could have accommodated a possible post–22 Ma tilt of the structural blocks in the area and thus the likely tilt direction and plausible magnitude of tilt.

The Three Peaks laccolith is bounded on the east by the East fault, an east-side-down normal fault, and on the west by the Northwest Intrusive fault, a west-dipping normal fault (Rowley et al., 2006). Motion along the Northwest Intrusive fault likely tilted the footwall, composed of Cretaceous Iron Springs Formation strata that strike northeast and dip northwest at 20°–30°. It is likely that the stratal orientation reflects a combination of pre–22 Ma deformation-associated Sevier-age shortening and late Cenozoic Basin and Range extension. Strata adjacent to the intrusion show steep (>45°) west dips likely reflecting deformation of the country rock associated with intrusion of the Three Peaks laccolith. The Three Peaks laccolith could also have been tilted to the northwest along with adjacent Cretaceous strata; however, we believe this unlikely, considering that the laccolith is located in a structural horst (Rowley et al., 2006). Stratified materials from other structural horst blocks in the region show no preferred tilt orientation and dip magnitudes that are modest (<30°). In the northern Bald Hills, ∼7 km northwest of the Three Peaks laccolith, Cenozoic strata typically have a north-northeast strike with gentle east dips (10°–15°). To the south in the Bald Hills, closer to the northwest margin of the Three Peaks laccolith, a major northwest-striking structure, the Hole-in-the-Wall fault, cuts across the structural grain of the Bald Hills (Rowley et al., 2006). South of this structure is a complexly faulted zone with fault-bounded structural blocks of varying orientations, with strikes ranging from northwest, north, to east-west and dips ranging from 20° to 45°. To the west in the Antelope Range, the Leach Canyon and Harmony Hills Tuffs in different fault blocks show strikes ranging from north-south to east-west and dips as much as 35° with dispersed directions. Approximately 6 km southwest of the New Castle Reservoir, Mesozoic and Cenozoic rocks are variably faulted and there is considerable variation in the orientation of fault blocks (Rowley et al., 1988, 2006). We argue that the area surrounding and including the Three Peaks laccolith is variability faulted, and that there is no compelling field evidence for a regional, post–22 Ma tilt affecting the area. Nonetheless, we admit that tilting may contribute to the discordant paleomagnetic data from the Iron Axis laccoliths and that many forms of tilt correction applied to one or all of the data sets from the Iron Axis laccoliths may result in an erroneous estimate of vertical axis rotation of crustal blocks including the laccoliths.

Kinematic Indicators

The orientation of small-scale structures as well as kinematic indicators on major and minor structures in the study area would, ideally, aid in the understanding of how the paleomagnetically observed vertical axis rotations in the region have been accommodated. Unfortunately, very few data exist on the kinematics of the structures in the area. Michel-Noel et al. (1990) provided abundant kinematic data on faults in the Rainbow Canyon area within the CEZ. Rainbow Canyon is located southwest of Caliente, in eastern Nevada, ∼60 km west of the study area. The data are from slip indicators collected on the Hiko Tuff (18.2 Ma) and the tuff of Etna (14 Ma). Here, the dominant set of faults strikes northwest, and striae for this fault set have dual modes of nearly pure dextral and normal sense. Oblique striae are rare. Hudson et al. (1998) interpreted the fault kinematic data to indicate that dextral-sense displacement along northwest-striking faults was responsible for counterclockwise vertical axis block rotation; they argued that the separation of fault modes into dip-slip and strike-slip end members implies that the declination discordance and strata dips in this area may have developed separately in response to multiple phases of vertical axis rotation and tilt, respectively. Timing relations between dip-slip and strike-slip faulting (Michel-Noel et al., 1990) and dip and vertical axis rotation development suggests that these deformation events probably overlapped in time and space and may have alternated on a short time scale (Hudson et al., 1998). Hudson et al. (1998) pointed out that it is unclear if this conclusion can be generalized for the entire CEZ. We note that fault geometries in our study area are similar to those in the Rainbow Canyon area.

Previous Paleomagnetic Results

Hudson et al. (1998) sampled lithologically distinct early Miocene ash-flow tuffs of wide spatial distribution that were involved in the main phase of Cenozoic deformation within the CEZ; their paleomagnetic data show varying, yet systematic degrees of counterclockwise vertical axis rotations throughout the CEZ, and the results that we have obtained in this study are similar in overall pattern (Fig. 9). The inferred progressive increase in the magnitude of counterclockwise rotation from south to north across the CEZ was interpreted by Hudson et al. (1998) as the result of accommodating an increase in the magnitude of extension in areas directly north and south of the zone. Hudson et al. (1998) evaluated vertical axis rotations relative to undeformed reference areas in the southern Delamar Mountains, Meadow Valley Mountains, and Condor Canyon in southeastern Nevada by between-site, within-unit comparisons. Their study documented rotation estimates ranging from +18° clockwise to as high as –125° counterclockwise, at 95% confidence; they found that all of the ash-flow tuffs examined had been rotated to some degree, with the largest rotations in the east-central part of the CEZ and a systematic decrease to the west. Data from several sections of the Bauers and Harmony Hills Tuffs, in the eastern part of the CEZ, show considerable variation in rotation magnitudes over a distance of 40 km (Fig. 9). Rotation estimates range from large counterclockwise values (–70° to –98°) near Newcastle, to moderate values (–20° to –40°) on the eastern side of the Antelope Range, to minor values (<–20°) near Desert Mound. Local rotations (–20° to –45°) in the Bauers and Harmony Hills Tuffs are identified in the South Hills, 10 km south of Cedar City (Fig. 9). Hudson et al. (1998) interpreted these rotation estimates to reflect local drag adjacent to a buried fault west of the Hurricane fault system (Anderson and Mehnert, 1979). Based on the small rotation estimates of a number of Bauers Tuff and Harmony Hills Tuff sites collected from the Harmony Mountains and Desert Mound, Hudson et al. (1998) proposed that the CEZ terminates ∼10 km east of Newcastle, despite the existence of normal, dextral and sinistral faults of dominantly northwest to northeast strike (Fig. 10). The new paleomagnetic data we report from the Iron Axis laccoliths and the associated ignimbrites from the surrounding area serve to modify this conclusion. We argue here that the termination of the eastern CEZ extends farther east and north of Cedar City.

Eastern Boundary of the CEZ

Geologic fault patterns in the CEZ reveal a distinct trend where northwest-southeast–oriented strike-slip faults are more prominent in areas that have undergone greater magnitudes of counterclockwise rotation (e.g., Newcastle area), with rotation estimates decreasing as fault strikes are more north-south (e.g., Pine Valley; Hurlow, 2002; Rowley et al., 2006; Biek et al., 2009; Fig. 10). The results from this study are consistent with the data of Hudson et al. (1998) in that the magnitude of counterclockwise rotations increases to the north, from Pine Valley to Newcastle (Fig. 9). To the north, rotation estimates diminish to <10° in the Escalante Desert and define the northern boundary of the shear zone, and to the east and northeast, data from the Antelope Range and Desert Mound reveal a decreasing, yet statistically significant, amount of rotation from west to east (see Hudson et al., 1998). The data from this study allow for a more accurate delineation of the eastern and southern margin of the CEZ in that they reveal how counterclockwise rotations increase from zero at the southern edge of the study area (north of the Pine Valley laccolith) to a maximum of –80° near Newcastle (Fig. 9).

Northeast of Desert Mound, sites in the Three Peak laccolith yield an in situ grand mean direction (D = 332.9°, I = 51.7°) that is discordant, with inferred estimates of R = –22.2° ± 6.4° and F = 8.8° ± 4.3°. The results from the Three Peak laccolith are consistent with the rotation estimates (–10° to –20°) from the Bauers and Harmony Hills Tuffs near Desert Mound (Fig. 9), yet we emphasize that the overall reliability of this result is compromised by our inability to accurately reference the data to the paleohorizontal. Where the Stoddard Mountain and Pine Valley laccoliths are exposed, Cenozoic faults strike roughly north-northwest. In the area exposing the laccoliths we studied, ∼30 km to the northeast, fault traces trend northwest to west-northwest, more typical of the zones that have been rotated within the CEZ. North of Stoddard Mountain, Cenozoic faults change strike over a few kilometers from north-south to north-northwest (Fig. 10). Clearly the zone between the Stoddard Mountain and Three Peaks, Iron Mountain, and Granite Mountain laccoliths is marked by an abrupt change in fault orientation. We propose that the southern boundary of the CEZ is just north of the Stoddard Mountain laccolith (Fig. 11), and attribute the change in fault strike to reflect the southernmost boundary of the zone of vertical axis rotation within the eastern CEZ. We project this zone to the northeast beyond the Three Peaks laccolith to where it disappears into the alluvial fill of the Escalante Desert, where no Miocene igneous rocks are exposed. We argue that the boundary of the eastern CEZ extends east to the breakaway with the Colorado Plateau (Fig. 11), and hypothesize that the transitional zone between the eastern CEZ and Colorado Plateau is a narrow (<∼10 km wide) zone northwest of Cedar City.

The available paleomagnetic data imply that deformation in the eastern CEZ progressively fragmented the crust north of the Pine Valley and Stoddard Mountain laccoliths, resulting in components of vertical axis rotation and tilting associated with regional left-lateral transtensional deformation, a conclusion similar to that reached by both Hudson et al. (1998) and Axen (1998). Considering a rotating block model (e.g., McKenzie and Jackson, 1986), the southern boundary of the CEZ may be pinned against the stable and thick Colorado Plateau margin. To the north, the blocks became unpinned and allowed crust in the region to rotate in an unrestricted fashion. This scenario implies that the Pine Valley laccolith is at the western margin of the Colorado Plateau, whereas the Stoddard Mountain laccolith is positioned within the southern extent or rather the transition zone of the eastern CEZ, and the minimal observed rotation of the area reflects its southern boundary being pinned against the stable Colorado Plateau block. As proposed by others (e.g., Hudson et al., 1998; Axen et al., 1998), we interpret the southern boundary of the eastern CEZ to mark a fundamental change in crustal properties and to represent a crustal boundary that Cenozoic extension exploited as the crust extended to the west away from the Colorado Plateau margin.

The inception of counterclockwise rotation during formation of the CEZ is only broadly determined to have begun after 18 Ma and continued until after 14 Ma, based on ages of rotated tuffs within the western and central CEZ (Hudson et al., 1998). Formation of the Caliente caldera complex in the central CEZ was partly coeval with the formation of the CEZ, leading previous workers to speculate on the relation between magmatism and CEZ deformation (Hudson et al., 1998; Axen, 1998). Hudson et al. (1998) attributed the significant width of the CEZ (compared to other accommodation zones in the Basin and Range Province) to enhanced ductility of the crust associated with coeval magmatism. In the eastern part of the CEZ, an en echelon pattern of dextral faults and laccoliths defines the Iron Axis magmatic province (Rowley et al., 1995) (Fig. 9). The spatial distribution of the array of variable composition Iron Axis intrusions is partly controlled by older northeast-striking, southeast-verging Sevier-age (Late Cretaceous to early Cenozoic) thrust faults. Iron Axis magmatism preceded formation of the eastern CEZ, as suggested by rotations of dated tuffs and the northern smaller laccoliths. Axen (1998) postulated that the emplacement of the Iron Axis laccoliths in part hindered the initiation of the CEZ by forming a barrier that forced the zone to widen significantly away from (west of) the laccoliths. Based on our new data, the two southernmost and largest of the Iron Axis laccoliths may have affected the deformational response of the crust either through pluton strengthening or stitching of preexisting crustal weaknesses. We speculate that these large laccoliths represent the shallow-crustal expression of a more voluminous, deep-seated batholith perhaps near the mid-crustal strength maximum. In contrast, the small laccoliths to the north likely do not have deep plutonic roots and were too small to inhibit deformation, and instead were deformed and rotated a magnitude similar to that of Cenozoic volcanic rocks (Figs. 9 and 11).

CONCLUSIONS

Paleomagnetic data from volcanic and shallow intrusive rocks in the eastern CEZ reveal an orderly pattern of rotation of crustal blocks progressing from near 0° along the southern boundary to –84° near the northern boundary. These data are consistent with previous paleomagnetic studies that reveal similar and spatially variable patterns of deformation in the central and western CEZ. We extend the eastern zone of the CEZ to within a few kilometers of the breakaway of the Colorado Plateau, east and north of Cedar City, Utah. We argue that the Pine Valley and Stoddard Mountain laccoliths, given their size and proximity to the stable Colorado Plateau margin, did not undergo deformation related to the development of the CEZ, whereas smaller intrusions located farther north likely had similar magnitudes of counterclockwise rotation as their near age-equivalent volcanic rocks, although the resolution of our rotation estimates for the plutons is limited because of an inability to accurately restore the data to the paleohorizontal. The southernmost part of the CEZ remained rigid during Neogene crustal extension and may represent a major crustal strength boundary, possibly inherited from the early development of the western margin of North America, in Neoproterozoic and Paleozoic time.

We thank Bevan Killpack and the Pine Valley Ranger District of Dixie National Forest for logistical assistance, and Adam Brister, Brian O’Driscoll, Greg Peacock, David B. Hacker, Jr., and Stumpy for assistance with sample collection. We thank Gary Axen and Mark Hudson for constructive reviews that greatly improved the manuscript. Partial funding for this project was provided by National Geographic Society Grants-In-Aid of Research Program award 8106-06 (Petronis), a New Mexico Highlands University Faculty Research Committee Grant, the Kent State University Research Council, and the Utah Geological Survey.

APPENDIX 1. PALEOMAGNETIC METHODS

Remanent magnetizations and standard rock magnetic analyses were conducted either at the New Mexico Highlands University (NMHU) Paleomagnetic-Rock Magnetism Laboratory using an AGICO JR6A dual speed spinner magnetometer within a Magnetic Measurements Ltd. MMLFC shielded room or at the University of New Mexico (UNM) Paleomagnetism laboratory with a 2G Enterprises Model 760R three-axis superconducting rock magnetometer with an integrated alternating field (AF) demagnetization system in a magnetically shielded room. Typically, between 6 and 14 specimens from each site were progressively AF demagnetized in 10 to 18 steps to a maximum field of 120 mT using an ASC Scientific D-tech 2000 AF demagnetizer at the NMHU lab or the integrated 2G Enterprises AF demagnetization system at the UNM lab. Samples of high coercivity were further treated with thermal (TH) demagnetization until <10% of natural remanent magnetization (NRM) remained; typically to 590 °C for the ash-flow tuffs and to 660 °C for some intrusive igneous rocks. To compare AF behavior, duplicate specimens of some samples were thermally demagnetized using an ASC Scientific TD48 or Schonstedt TSD-1 thermal demagnetizer. No difference in the directional data between AF and TH demagnetization experiments was observed (Fig. 8). For most samples, a single line could be fit to the demagnetization data and, typically, involved 10–15 data points. Sample MAD values for linear data were typically <4°, but ranged from <1° to 5°. In some specimens, low-coercivity, random magnetization vectors, interpreted as viscous remanent magnetization (VRM) overprints, were readily removed by about 20 mT. Following removal of the VRM, the characteristic remanent magnetization, interpreted to be the primary thermoremanent magnetization, decays along a generally univectoral path to the origin after treatment in 120 mT fields, or by 660 °C. For <10% of the samples, demagnetization vectors were anchored to the origin. The demagnetization data were plotted on orthogonal demagnetization diagrams (Roy and Park, 1974; Zijderveld, 1967) and linear segments were fit using principal component analysis (Kirschvink, 1980). Site mean directions and associated statistics followed the methods of Fisher (1953). Data from individual specimens were considered outliers and rejected from the site mean calculation if the angular distance between the specimen mean direction and the estimated site mean direction was >18°. For a few sites in the Three Peak laccolith best fit lines could be established through the demagnetization data. For these sites (Table 4), remagnetization circle analysis was used to isolate a best fit direction (McFadden and McElhinny, 1988). The technique supposes that at times a sample may have multiple magnetization components with overlapping blocking temperatures or coercivity spectra. The data plot as arcs or remagnetization circles during progressive demagnetization, and often the remagnetization circles intersect at the direction of one of the NRM components (Butler, 1984). All site mean directions from stratified materials were corrected for structural tilt by rotation about individual strike axes of strata at each site location whereas in situ results are reported for all intrusive rocks (Tables 3 and 4).

APPENDIX 2. ROCK MAGNETISM

All susceptibility experiments were measured with an AGICO MFK1-A kappabridge susceptibility meter with a CS-4 high-temperature attachment; remanence measurements were conducted with an AGICO JR6A dual-speed spinner magnetometer, and isothermal magnetizations were imparted using an in-house built static 3 T impulse magnet at the New Mexico Highlands University laboratory. Continuous low-field susceptibility versus temperature measurements were carried out in a stepwise heating and cooling fashion from 25 °C to 700 °C to 40 °C in an argon atmosphere using a CS-4 furnace attachment for MFK1-A. These experiments allow for an evaluation of the magnetic mineral composition based on Curie point estimates and assist with revealing mixtures of magnetic phases within a given sample. Anhysteretic remanent magnetization (ARM) was imparted in a DC field of 0.1 mT and a peak alternating field (AF) field of 95 mT, and then AF demagnetized in ∼11 steps to 95 mT. Isothermal remanent magnetization (IRM) acquisition experiments were conducted by exposing the specimens to a strong magnetic field with an impulse magnet from 0.005 T to saturation. Following demagnetization of the ARM, IRM acquisition experiments were conducted where samples were treated in a stepwise fashion to higher fields until saturation. IRM acquisition and backfield IRM acquisition are measured to estimate the coercivity of remanence of the samples and to provide information on magnetic grain size and composition. ARM and saturation isothermal remanent magnetization (SIRM) demagnetization curves that are characterized by a high coercive force and high medium destructive field (MDF) values indicative of single domain (SD) magnetic behavior (Dunlop and Özdemir, 1997). Alternatively, rapid decay of natural remanent magnetization (NRM), ARM, and SIRM combined with low coercive force values generally indicate multidomain (MD) behavior. Backfield IRM experiments were conducted by applying specimens to an increasing field to 1.33 T along the –Z axis until the +Z remanence was reduced to zero (i.e., the sign changed). Finally, the samples were again saturated in a 1.33 T field and AF demagnetized. The modified Lowrie-Fuller test (Johnson et al., 1975) can be used to estimate the magnetic domain state. This experiment was conducted on the ash-flow tuffs only. The results of all rock magnetic experiments are provided in Appendix 3.

APPENDIX 3. ROCK MAGNETISM

Low-Field Susceptibility versus Temperature Experiments

Continuous low-field susceptibility versus temperature measurements from room temperature to 700 °C were conducted on representative samples from the ignimbrites and the intrusive rocks. Curie points were estimated using either the inflection point (Tauxe, 1998) or Hopkinson peak methods (Moskowitz, 1981). Pure magnetite has a Curie point of 580 °C, which decreases nearly linearly with increasing Ti substitution to ∼–150 °C for pure ilmenite. Curie or Neel points of other common minerals include hematite (675 °C), pyrrhotite (320 °C), and greigite (∼330 °C) (Dunlop and Özdemir, 1997). To avoid oxidation and chemical alteration of the magnetic phases, all experiments were conducted in an inert Ar atmosphere; although it is likely that residual O2 clinging to the grains results in some degree of oxidation (Petronis et al., 2011).

Low-field continuous susceptibility versus temperature experiments from the Three Peaks and Granite Mountain laccoliths yield a very narrow spectrum of response and reveal a slight increase in room temperature susceptibility of ∼20% to ∼30% after the complete cycling to 700 °C, with most samples fully reversible (Appendix Fig. 1A). Using the inflection point method of Tauxe (1998), these rocks yield inferred Curie points from 548 to 579 °C, temperatures consistent with a low-Ti titanomagnetite to pure magnetite phase with little evidence of a second magnetic phase at low temperatures (Table 5). Results from the Iron Mountain laccolith, in contrast, yield a spectrum of low-field continuous susceptibility versus temperature data indicating the presence of more than one magnetic phase in the majority of sites analyzed (Appendix Fig. 1A). All samples reveal a minor to significant increase in room temperature susceptibility ranging from <0% to ∼50% after the complete cycling to 700 °C, with only two samples showing fully reversible results (Appendix Fig. 1A). Using either the inflection point (Tauxe, 1998) or Hopkinson peak method (Moskowitz, 1981), samples yield inferred Curie point estimates that range from 555 to 638 °C, with an average temperature of 570 °C, temperatures consistent with moderate- to low-Ti titanomagnetite and fine-grained maghemite (Table 5). Several samples show an increase in susceptibility (bump) on heating between ∼220 and 350 °C that is not present on the cooling curve, suggesting the presence of a second magnetic phase (Appendix Fig. 1A). The bump may indicate that the oxidation of a Fe-Ti oxide phase (likely titanomaghemite) is inhomogeneous, and during the heating experiment the mineral in some fashion homogenizes to a more susceptible phase, as reflected by the modest increase in susceptibility (e.g., Kropáček and Pokorná, 1973; Hrouda et al., 2006). Alternatively, it is possible that some samples might contain a small quantity of pyrrhotite, with a Curie point that is between 200 and 400 °C.

Low-field continuous susceptibility versus temperature experiments from the four tuffs yield a spectrum of responses that varied by rock type. The Leach Canyon Tuff yields a slight increase (∼10%) in susceptibility after complete cycling to 700 °C with fully reversible curves (Appendix Fig. 1B). Using the Hopkinson peak method (Moskowitz, 1981), the samples yield inferred Curie point estimates from 534 to 565 °C, a temperature range consistent with moderate- to low-Ti titanomagnetite. All curves show an increase in susceptibility (bump) on heating between ∼220 and 350 °C that is not present on the cooling curve, suggesting the presence of a second magnetic phase. The Harmony Hills Tuff yields little to no increase in susceptibility after complete cycling to 700 °C with fully reversible curves in most cases (Appendix Fig. 1B). Using the inflection point method (Tauxe, 1998), the samples yield inferred Curie point estimates from 566 to 578 °C, consistent with low-Ti titanomagnetite to nearly pure magnetite. The rocks of Paradise tuff shows a moderate (50%) to significant (80%) decrease in susceptibility after complete cycling to 700 °C with nonreversible behavior (Appendix Fig. 1B). Using the inflection point method (Tauxe, 1998), the samples yield inferred Curie point estimates that range from 562 to 579 °C, consistent with low-Ti titanomagnetite to nearly pure magnetite, although it is likely that the magnetic mineralogy changed during the heating experiment. All curves show an increase in susceptibility on heating near 330 °C that is not present on the cooling curve, suggesting the presence of a second ferromagnetic phase. The decrease in susceptibility may reflect that the magnetic fraction is being altered by heating, and this alteration produces a phase with a lower susceptibility. Unfortunately, the nature of this alteration is poorly understood (see Hrouda, 2003), yet these results are typically interpreted to indicate that the primary magnetic mineral is, in all cases titanomaghemite. The lower Curie temperatures observed during the cooling cycle may result from (1) the inversion of titanomaghemite to titanium-poor magnetite and ilmenite during heating and (2) the subsequent reduction of these phases to form titanomagnetite. Experiments by Hammond and Taylor (1982) argued that the lower Curie point that is observed during cooling, like the higher Curie point observed during heating, are simply an artifact of the measurement procedure and are controlled by the oxygen fugacity and maximum temperature of the laboratory oven. The Bauers Tuff shows a moderate (50%) to significant (80%) increase in susceptibility after complete cycling to 700 °C with nonreversible behavior (Appendix Fig. 1B).

Isothermal Remanent Magnetization (IRM) Acquisition Curves and Backfield IRM

IRM and backfield IRM data from select specimens of each of the tuff units and the Iron Axis laccoliths all show a very narrow range of responses (Appendix Figs. 2A, 2B). All curves show steep acquisition with near complete saturation by 0.15 T and no evidence of high-coercivity phases in applied fields to 2.5 T. These results are consistent with low-Ti titanomagnetite phase of a restricted magnetic grain size, likely single domain (SD) to pseudo–single domain (PSD). Coercivity of remanence values in all units is <0.05 T (Appendix Figs. 2A, 2B) and distributed unblocking temperatures to 580–600 °C (Fig. 8) indicate that a Fe-Ti oxide phase, likely low-Ti magnetite, is the main remanence carrier in the samples.

Lowrie-Fuller Test

Following the modified Lowrie-Fuller tests (Johnson et al., 1975), the alternating field (AF) decay of normalized natural remanent magnetization (NRM), anhysteretic remanent magnetization (ARM), and saturation isothermal remanent magnetization (SIRM) was compared for representative samples from the ignimbrites (Appendix Fig. 3). This test is based on experimental observations that normalized AF demagnetization curves of weak-field thermoremanent magnetization (TRM) (i.e., NRM) and strong-field TRM (i.e., SIRM) have different relationships for SD and MD grains of magnetite (Dunlop and Özdemir, 1997). Laboratory investigations typically use weak-field ARM as an alternative for TRM, and we also used that procedure. In large MD grains, SIRM requires larger destructive fields than ARM to reach the same normalized level, meaning the medium destructive field (MDF) of SIRM is larger than the MDF of ARM. Because small MD grains exhibit a mixture of SD-like (high remanence) and MD-like (low coercivity) behavior, the resistance of ARM to demagnetization relative to SIRM suggests the presence of PSD grains of magnetite (Appendix Fig. 3). The results of the test indicate that the dominant magnetic mineral in all ignimbrites is likely a ferrimagnetic phase, probably SD to PSD low-titanium magnetite, although some variability was observed. It is probable that that maghemite, which may display maximum laboratory unblocking temperatures above magnetite, contributes to the remanence (Dunlop and Özdemir, 1997).