Abstract

Volcanic rocks in the Sonoma volcanic field in the northern California Coast Ranges contain heterogeneous assemblages of a variety of compositionally diverse volcanic rocks. We have used field mapping, new and existing age determinations, and 343 new major and trace element analyses of whole-rock samples from lavas and tuff to define for the first time volcanic source areas for many parts of the Sonoma volcanic field. Geophysical data and models have helped to define the thickness of the volcanic pile and the location of caldera structures. Volcanic rocks of the Sonoma volcanic field show a broad range in eruptive style that is spatially variable and specific to an individual eruptive center. Major, minor, and trace-element geochemical data for intracaldera and outflow tuffs and their distal fall equivalents suggest caldera-related sources for the Pinole and Lawlor Tuffs in southern Napa Valley and for the tuff of Franz Valley in northern Napa Valley. Stratigraphic correlations based on similarity in eruptive sequence and style coupled with geochemical data allow an estimate of 30 km of right-lateral offset across the West Napa-Carneros fault zones since ∼5 Ma.

The volcanic fields in the California Coast Ranges north of San Francisco Bay are temporally and spatially associated with the northward migration of the Mendocino triple junction and the transition from subduction and associated arc volcanism to a slab window tectonic environment. Our geochemical analyses from the Sonoma volcanic field highlight the geochemical diversity of these volcanic rocks, allowing us to clearly distinguish these volcanic rocks from those of the roughly coeval ancestral Cascades magmatic arc to the west, and also to compare rocks of the Sonoma volcanic field to rocks from other slab window settings.

INTRODUCTION

The rocks of the Sonoma volcanic field (Fig. 1) are part of a linear belt of exposures of Cenozoic volcanic rocks that progressively young to the northwest (Fox et al., 1985a) and have been disrupted by dextral faults of the San Andreas fault system from their original depositional locations (Fox, 1983; McLaughlin et al., 1996; Wakabayashi, 1999; Graymer et al., 2002a, 2002b). In the San Francisco Bay region, these volcanic rocks include the Quien Sabe Volcanics, volcanic rocks in the Berkeley Hills, the Sonoma Volcanics, the Tolay Volcanics, the Burdell Mountain Volcanics, and the Clear Lake Volcanics, using the nomenclature of Fox et al. (1985a) and Graymer et al. (2002b) (Fig. 1). Like other exposures of volcanic rocks in the northern California Coast Ranges, the Sonoma volcanic field contains heterogeneous assemblages of compositionally diverse lava flows, pyroclastic deposits, and local ash-flow tuffs (Fig. 2) (Fox, 1983; Graymer et al., 2002a).

Although the Sonoma volcanic field is the largest of the volcanic fields in the northern California Coast Ranges, developing an understanding of the stratigraphy, volcanology, and geochemical evolution of this complex has been difficult because of the limited erupted volumes from individual volcanic centers and consequent lack of lithologic continuity of many of the units. Extensive structural disruption throughout the western half of the volcanic field also contributes to the complexity of the regional correlation of units (Fox, 1983). This study delineates for the first time volcanic centers primarily within the eastern, less deformed half of the Sonoma volcanic field (Fig. 3) based on a combination of geologic mapping, geophysical signature, and geochemical and petrographic criteria to tie eruptive products to specific source areas. These combined techniques have led to our recognition of distinct eruptive styles that vary between the different centers and to estimates of offset along the West Napa-Carneros fault zones.

This study defines minor- and major-element geochemical trends within each volcanic center, building upon previous analyses of a relatively small number of samples from the Sonoma volcanic field (Johnson and O'Neil, 1984; Whitlock, 2002). We define geochemical trends of intracaldera and outflow tuffs and distal fall equivalents from individual volcanic centers. We use major-, minor-, and trace-element geochemical data to assist in correlation of volcanic units and to define caldera-related sources for regionally important tuffs, such as the Pinole and Lawlor Tuffs and the tuff of Napa, whose source areas have previously been only generally outlined.

The volcanic fields in the California Coast Ranges north of San Francisco Bay are temporally and spatially associated with the northward migration of the Mendocino triple junction (Dickinson and Snyder, 1979; Furlong, 1984; Fox et al., 1985a; Dickinson, 1997). The northward younging of volcanism has been attributed to the transition from subduction and associated arc volcanism to a slab window tectonic environment (or “slabless window,” using the terminology of Liu and Furlong, 1992) along the western margin of the North American plate (Dickinson and Snyder, 1979; Johnson and O'Neil, 1984; Fox et al., 1985a). Recent work (Cousens et al., 2008) has defined a Miocene–Pliocene Ancestral Cascades arc that was active at about the same time and roughly the same latitude as the volcanic centers of the Sonoma volcanic field. In this paper, we discuss the geochemical and tectonic setting of volcanic rocks of the Sonoma volcanic field in the context of the northward migration of the Mendocino Triple Junction (Johnson and O'Neil, 1984; Dickinson, 1997) and also compare these rocks to the generally contemporaneous rocks of the Ancestral Cascades volcanic arc (Cousens et al., 2008).

GEOLOGIC, GEOCHEMICAL, AND GEOPHYSICAL METHODS

The regional volcanic stratigraphy, identification of the location of volcanic centers (Fig. 3), and delineation of the spatial distributions of individual volcanic units were accomplished through extensive field observation by the authors. Our fieldwork supplemented and augmented geologic mapping being conducted concurrently by the California Geological Survey (Bezore et al., 2004, 2005; Clahan et al., 2004, 2005; Wagner et al., 2003, 2004, 2006) and by other investigators at the U.S. Geological Survey (Graymer et al., 2002a, 2007; McLaughlin et al., 2004, 2008). Where possible, we used new and existing paleomagnetic (Mankinen, 1972), geochronologic (Fox et al., 1985a; McLaughlin et al., 2004, 2008), and tephrochronologic (Sarna-Wojcicki, 1976; Sarna-Wojcicki et al., 1979, 1984) data to synthesize the regional volcanic stratigraphy. Stratigraphic correlation of volcanic units was accomplished through comparison of phenocryst assemblage and mineralogy, pumice and lithic content, and geochemical “fingerprinting” using major-, minor-, and trace-element geochemistry augmented by petrography.

Whole-rock samples of various volcanic rock types were collected at outcrops throughout the Sonoma volcanic field (SVF) for geochemical analysis. Care was taken to sample unaltered outcrops; hydrothermal alteration could generally be identified in the geochemical results as a loss of alkalis. Samples were trimmed of obvious lithic inclusions, but individual fiamme or pumice were not sampled or analyzed. All whole-rock samples were analyzed for 10 major oxides, determined by wavelength dispersive X-ray fluorescence spectrometry (WDXRF). Techniques and standards used for WDXRF analysis of major elements have been given by Taggart and Siems (2002). The detection limit for all elements including loss of ignition (LOI) is 0.01%. When compared with replicate analysis of internal reference materials, determined values are within 1%–5% of proposed or certified values for elements at >1 wt% and are within 5%–10% of proposed or certified values for elements at <1 wt% abundance (Taggart and Siems, 2002). All whole-rock samples were analyzed for 55 trace elements using inductively coupled plasma–atomic emission spectrometry (ICP-AES). Techniques and standards for the ICP-AES method for a smaller suite of elements have been described by Briggs (2002). ICP-AES precision and accuracy based on replicate analysis of internal basalt standards, duplicate analyses, and method blanks are ±5%–10% for most elements, ±10%–15% for Nb. Chemical analyses of a representative set of volcanic rocks from the Sonoma volcanic field are presented in Table 1, and all geochemical analyses are listed in Supplemental Table 11.

Gravity and aeromagnetic data reflect density and magnetization contrasts within the upper and middle crust; we use these data to help define the spatial extent and the three-dimensional geometry of the volcanic rocks and define the location of structures. The regional gravity data (Langenheim et al., 2006a; Langenheim et al., 2010) were gridded to produce an isostatic residual gravity map of the study region (Fig. 4). We used the method of Jachens and Moring (1990) to separate the isostatic gravity field into that component produced by variations in basement density (“basement gravity field”) and that caused by thick sedimentary and volcanic deposits, which is then inverted for basin thickness (Fig. 5).

Aeromagnetic anomalies are produced by a variety of sources of variable size and depth, typically related to the presence of Mesozoic basement rocks (ophiolitic rocks including serpentinite, gabbro, and basalt) and magnetic rocks within the Tertiary volcanic section. To separate those short-wavelength anomalies caused by shallow sources (e.g., Tertiary volcanic rock) from long-wavelength anomalies (e.g., serpentinite or ophiolite) caused by deeply buried pre-Cenozoic rocks, a match filter was applied (Phillips, 2001) to the aeromagnetic data (Langenheim et al., 2010). Match filtering separates the data into different wavelength components by modeling the observed anomalies as a sum of anomalies from distinct equivalent source layers at increasing depths (see Phillips, 2001). In order to assist in the interpretation of magnetic anomalies potentially caused by volcanic rocks or subvolcanic intrusions, we created two maps that portray the resulting separated fields produced by the dipole equivalent-source layers at 0.4141- (Fig. 6) and 1.546-km (Fig. 7) depths.

DEFINITION OF VOLCANIC CENTERS

We define for the first time volcanic source areas for many parts of the Sonoma volcanic field based on a combination of geologic mapping, geophysical signature, and geochemical and petrographic criteria. Table 2 presents synoptic data from many of the volcanic centers within the SVF, describing eruptive style, spatial distribution of erupted rocks, volcanic facies changes with distance from an interpreted center, geochemical data, and interpretation of geophysical data. Within the paper we describe in detail selected volcanic centers where we have concentrated our mapping and geochemical sampling within the eastern, less deformed half of the field. The list of volcanic centers presented in Table 2 is not comprehensive; there are a number of other known vents and sources for local flows that are not specifically described. In addition, there are volcanic units for which a specific source area has not been identified. Our geochemical database and field mapping is not exhaustive; we did not map nor analyze samples from several regionally significant volcanic units, including the Huichica Tuff and the Putah Tuff, and thus cannot comment on their possible correlation or source areas.

Previous Work in Defining Volcanic Centers

In the vicinity of Napa Valley, early mapping generally subdivided the volcanic rocks into an upper rhyolitic member and an underlying more heterogeneous sequence of rocks (Weaver, 1949; Kunkel and Upson, 1960). Subsequent geologic mapping (Fox et al., 1973) and supporting geochronology (Fox et al., 1985b) allowed the volcanic rocks to be generally subdivided into upper and lower members largely on the basis of age, rather than rock type—a considerable stratigraphic improvement. Both the upper and lower members were subdivided into a series of informal units (Fox et al., 1985b), each of these units still included considerable lithologic variability and did not explicitly link volcanic units to eruptive vents or spatial distribution. More recently, the U.S. Geological Survey (USGS) has produced digital geologic compilations that include the SVF at a regional scale (Blake et al., 2002; Graymer et al., 2002a, 2007), while the California Geological Survey (CGS) conducted 1:24,000-scale mapping within the volcanic field (e.g., Bezore et al., 2004, 2005; Clahan et al., 2004, 2005; Wagner et al., 2003, 2004, 2006). The USGS and CGS maps provide greater delineation of the variability of volcanic rock type, but do not identify volcanic source areas or correlate specific volcanic units.

Geochemical characterization of vitric pyroclastic rocks from within and surrounding the volcanic field has been employed to facilitate stratigraphic correlation of volcanic units whose source is presumed to be the SVF (Sarna-Wojcicki, 1976; Sarna-Wojcicki et al., 1979, 1984). Analyses are of volcanic glass from tephra layers, which include ash fall, pumice fall, ash flow, and water-reworked tuff deposits. These tephrochronologic analyses have been conducted at sites where the correlation of ash units would assist stratigraphic mapping and structural interpretation (McLaughlin et al., 2004, 2008), but have not been used globally within the field to establish stratigraphy nor have the results been tied to major-element whole-rock analyses of nonglassy rocks within the field. These analyses have established that certain distal fall deposits that appeared to have their source within the Sonoma volcanic field were regionally important (Sarna-Wojcicki, 1976; Sarna-Wojcicki et al., 1979, 1984), but individual source vents were only generally defined (McLaughlin et al., 2005).

General Description of Volcanic Centers

We define five major volcanic centers adjacent to Napa Valley (Fig. 3) including, from north to south, the Mount St. Helena (MSH), Calistoga Domes (CD), Wildlake (WL), Stags Leap (SL), and Cup and Saucer (CS) centers (Table 2 and Fig. 3). The Mount St. Helena volcanic center (MSH, Table 2 and Fig. 3) includes the caldera source of the 2.85 Ma tuff of Franz Valley (Table 2), and is the youngest and last active center in the Sonoma volcanic field (Fox et al., 1985a; McLaughlin et al., 2005). This center contrasts with many of the other eruptive centers within the Sonoma volcanic field in having volumetrically more abundant, thick siliceous ash deposits that include air fall, ash flow, lahar, and reworked water-transported ash deposits. This center was constructed on top of older domes and associated flows and tuffs of the Calistoga dome field (CD, Table 2 and Fig. 3) and a thick section of andesitic flows and lahars of the Wildlake volcanic center (WL, Table 2 and Fig. 3). In the south-central part of Napa Valley we define the Stags Leap volcanic center (SL, Table 2 and Fig. 3) as a large constructional composite volcano with a more than 350-m-thick sequence of basaltic andesite flows and lahars. At the south end of Napa Valley we define the Cup and Saucer volcanic center (CS, Table 2 and Fig. 3) where a lower sequence of interlayered mafic flows, volcanic agglomerate, and lithic tuff is overlain by a 150-m-thick section of 5.4 to <4.70 Ma ash-flow tuffs, tephra, and dacite lava flows that are associated with two nested calderas (Fig. 3). Rocks of the Mount George center (MG, Table 2 and Fig. 3) unconformably overlie rocks of the Cup and Saucer center and include local late-stage rhyolite flows and a welded tuff. On the geochemical diagrams, rocks from this center are combined with those from the Cup and Saucer volcanic center.

For the purposes of geochemical classification, we identify four volcanic centers in the vicinity of the town of Sonoma and in the mountains to the north: the Sonoma (SON), Arrowhead (AH), Bismarck Knob (BM), and Sugarloaf (SG) centers (Table 2 and Fig. 3). Additional local eruptive centers are described by Wagner et al. (2011). In contrast to the large volcanic centers adjacent to Napa Valley, the mountains to the north of the town of Sonoma contain a complex assemblage of volcanic units that appear to have erupted from numerous small volcanic centers (Wagner et al., 2011). The volcanic centers become progressively younger northward from the 7.5 Ma Arrowhead volcanic center (Table 2) at the southern end of the range to the 5.65–4.81 Ma volcanic rocks (Table 2) within the Sugarloaf Ridge volcanic center at the northern end of the range. The largest-volume and most stratigraphically complex center is the Sonoma center (SON, Table 2 and Fig. 3), which consists of a deformed sequence of basaltic andesite and andesitic flows and breccias with subordinate tuffs that appear to have erupted from a number of volcanic vents. The Sugarloaf Ridge and Arrowhead centers (SG, AH, Table 2 and Fig. 3) adjoin it to the north and south, respectively, and form more localized volcanic accumulations. The Sugarloaf Ridge volcanic center features near-vent, nonwelded to welded lithic rhyolitic tuffs that contain clasts up to 1 m in diameter and an arcuate alignment of basaltic andesite vents. The Bismarck Knob center (BM, Table 2 and Fig. 3) is defined to include local eruptive centers and late-stage rhyolite flows that unconformably overlie rocks of the Sonoma center.

The elongated exposure of volcanic rocks to the west of Sonoma Valley, including Sonoma Mountain, Bennett Mountain, and Taylor Mountain (Fig. 2) is the part of the volcanic field first identified as “Sonoma Volcanics” (Weaver, 1949). Our sampling in this area is extremely limited, so for the purposes of plotting geochemical data we group the heterogeneous and structurally complex volcanic rocks in the Sonoma Mountains and Taylor Mountain as the Sonoma Mountain-Taylor Mountain group (ST, Table 2), but we do not define specific eruptive centers. Overlying this sequence is the relatively minor Annadel center to the east of the City of Santa Rosa (AD, Table 2 and Fig. 3) (McLaughlin et al., 2008), consisting predominantly of localized eruptions of late-stage rhyolite lava flows.

Descriptive Geochemistry of Volcanic Centers

We present 343 new major- and trace-element analyses of whole-rock samples from lavas and tuff from the Sonoma volcanic field (Table 1 and Supplemental Table 1 [see footnote 1]). Volcanic rocks of the Sonoma volcanic field span the full compositional range from basalt and basaltic andesite through andesite, dacite, and rhyolite (LeBas et al., 1986) (Fig. 8). Individual volcanic centers tend to have a broad compositional range even though they are dominated volumetrically by a single rock type. For example, the Stags Leap center is comprised predominantly of andesitic composition flows and lahars, but contains rocks that nearly span the compositional range of the entire field (Fig. 8). Most of the SVF rocks are subalkaline; however, a small but important subset of samples from the Sugarloaf Ridge, Sonoma, and Cup and Saucer volcanic centers are compositionally distinct and trend toward alkaline compositions (Fig. 8A). SVF basalts, basaltic-andesites, and andesites have medium-K calc-alkaline compositions, although its more silicic rocks have high-K calc-alkaline compositions (Fig. 8B).

For silica contents between 45% and 70% (basalt to dacite), Ba is strongly positively correlated with SiO2 (Fig. 9), as are Rb and Th (not shown on Fig. 9). These correlations are much weaker among SVF rhyolites, which have widely variable concentrations of these elements (Table 1). Zr and Ba abundances are positively correlated with silica content between 50% and 70% SiO2, but are inversely correlated above 70% SiO2 (Fig. 9). Samples of the Cup and Saucer, Sugarloaf Ridge, and Sonoma volcanic centers that trend toward alkaline compositions have Zr abundance greater than 350 ppm, which are only typical among alkaline rocks. Sr increases slightly with silica up to ∼60% SiO2 but shows a general decrease with increasing silica content above 60% SiO2 (Fig. 9). Most SVF rocks have TiO2 concentrations of <1.5% (Table 1 and Fig. 9). Basalts and basaltic andesites have highly variable TiO2 concentrations (from 0.75% to 3%), whereas the TiO2 concentrations of more silicic rocks decrease monotonically with silica content (Fig. 9). MgO concentrations decrease monotonically with silica content, whereas those of K2O increase with increasing silica (Table 1, Figs. 9 and 10).

Incompatible element abundances vary by as much as a factor of 10 among samples from the various volcanic centers (Table 1 and Fig. 10). SVF basalts, basaltic andesites, and andesites from the Sonoma volcanic field are enriched in large-ion lithophile elements (LILE) such as K, Rb, Sr, Ba, and Cs and are distinctly depleted in high-field-strength elements (HFSE) such as Nb, Ta, and to a lesser extent Ti (Fig. 10). The magnitude of the Nb-Ta depletion is smallest in the most mafic SVF rocks and increases in more felsic SVF rocks (Fig. 10). Rare earth elements (REE) (La through Lu) are depleted with respect to LILE and strongly enriched with respect to HFSE (Fig. 10). For basalts and basaltic andesites, fractionation between LREE (La, Ce) and HREE (Y and Lu) is minimal and the REE pattern is relatively flat (Fig. 10). Incompatible element patterns normalized to the primitive mantle values of Sun and McDonough (1989) (Fig. 10) are moderately steep with relatively flat patterns in the middle to heavy rare earth elements. Andesites display clear negative Eu anomalies, whereas this is less apparent in more mafic samples (Fig. 10).

For selected tuffs from the Sonoma volcanic field, we have major-, minor-, and trace-element geochemistry for intracaldera and outflow tuffs and their distal fall equivalents. Intracaldera and outflow tuff geochemistry are whole-rock results from this study; distal fall material is from glass analysis of tephra (Sarna-Wojcicki, 1976; Sarna-Wojcicki et al., 1979). We recognize that the whole-rock analyses can be subject to unintended contamination from lithic fragments, or subject to a variety of postdepositional alteration processes that can induce scatter in the geochemical plots. Even given these uncertainties, the glass and whole-rock analyses show a remarkable degree of consistency where samples from a single tuff, defined on the basis of stratigraphic position and lithology, in addition to geochemistry, are compared (Fig. 11). We plot Zr against the high-field-strength element Ti (as TiO2) and use the element ratios Zr/Ba and Rb/Sr to geochemically differentiate intracaldera, outflow, and distal fall deposits for selected ash-flow tuffs from the Sonoma volcanic field (Fig. 11). Each eruptive unit tends to define its own geochemical trend; intracaldera, outflow, and fall deposits do not plot at the same place but tend to define somewhat linear trends on the geochemical plots (Fig. 11). Specific tuffs are discussed below as part of the volcanic center they are inferred to be associated with.

VOLCANIC CENTERS ON THE EAST SIDE OF NAPA VALLEY

Mount St. Helena Volcanic Center

The Mount St. Helena volcanic center (Table 2) consists of volcanic rocks associated with the formation of the Mount St. Helena caldera (Fig. 3). We interpret that the eruption of the tuff of Franz Valley at 2.85 Ma (McLaughlin et al., 2004) resulted in formation of the caldera. This outflow facies tuff consists of a single cooling unit comprised of both lithic- and pumice- (up to 30 cm) rich, nonwelded to partially welded ash-flow tuff (Table 2). Rocks from the Mount St. Helena volcanic center that are associated with the Mount St. Helena caldera range from high silica rhyolite, primarily represented in the outflow facies, to low silica rhyolite in the intracaldera tuff.

The caldera collapse is defined by a >900-m-thick sequence intracaldera facies ash-flow tuff and collapse breccias that comprise the upper part of Mount St. Helena (Sweetkind et al., 2005). The base of the intracaldera sequence is nonwelded and consists of beds of collapse breccia composed of rocks derived from the caldera wall in a matrix of ash-flow tuff. The upper part of the intracaldera section consists of a single cooling unit composed of partly to densely welded lithic-rich ash-flow tuff. The intracaldera tuff has a phenocryst assemblage of quartz and feldspar similar to that of the Franz Valley tuff and an 40Ar/39Ar age of 2.83 ± 0.08 Ma (Weiss and Livermore, 1996) that is identical with the tuff of Franz Valley, from which we infer that the Franz Valley tuff represents the outflow facies of the intracaldera tuff deposited within the Mount St. Helena caldera. Using the areal extent and the thickness of volcanic rocks, we estimate that the volume of material erupted from the Mount St. Helena center may be at least 40 km3.

The geomorphic expression of the caldera has been highly modified as a result of caldera resurgence, uplift of the eastern side of Napa Valley, and erosion of the older rocks that formed the walls of the caldera (Sweetkind et al., 2005). Hydrothermal alteration of both wall rocks and intracaldera tuffs on the west and southwest side of the caldera has contributed to erosion and destruction of the topographic expression of the caldera (Sweetkind et al., 2005). As a result intracaldera rocks form a prominent topographic high. The floor of the caldera is exposed on the northeast side of the caldera where intracaldera tuff rests on serpentinite (Fig. 3).

Older rhyolitic air fall and ash-flow tuffs were erupted from the vicinity of the Mount St. Helena caldera but the source region for these eruptions was destroyed during the eruption of the younger tuff of Franz Valley and formation of the Mount St. Helena caldera. These older tuffs include the ash-flow tuff of Petrified Forest (3.34–3.35 Ma, K-Ar, Sarna-Wojcicki, 1976; Sarna-Wojcicki et al., 1979), and the tuff of the Pepperwood Ranch (3.19 Ma, 40Ar/39Ar, McLaughlin et al., 2004). Also underlying the Mount St. Helena volcanic center is the Calistoga Dome center, which consists of broadly distributed, mostly silicic tuff apron and rhyolite dome complexes that form a coalescing dome field (Table 2). Assuming that the deep, linear trough in the modeled depth-to-basement surface (Fig. 5) beneath Diamond Mountain (to the south of Calistoga, Fig. 2) is filled with silicic volcanic rocks of the Calistoga Dome field, we estimate a total erupted volume of material from the Calistoga Dome center to be ∼100 km3.

Post–Mount St. Helena caldera eruption of high silica rhyolite occurred outside the south margin of the Mount St. Helena caldera. Eruption of postcollapse andesite was widespread outside and to the southwest of the Mount St. Helena caldera. Numerous flows of andesite cap the outflow facies of the Franz Valley tuff in the Mark West Springs quadrangle (McLaughlin et al., 2004). The vent areas for these flows are not well exposed except for one on the east central part of the Mark West Springs quadrangle (McLaughlin et al., 2004). To the east and northeast of Mount St. Helena, widespread late-stage basalts cap the stratigraphic sections.

Stags Leap Volcanic Center

The Stags Leap Volcanic Center (SL, Table 2 and Fig. 3) is well exposed to the east of Napa Valley, where a more than 350-m-thick sequence of basaltic andesite flows and lahars form a constructional composite volcano. Proximal to the inferred center of the volcano, lahar breccias comprise 40%–50% of the volcanic section (Figs. 12A and 12B); the volcanic section thins and becomes dominated by andesitic flows to the north and south of the inferred vent area (Fig. 12C). The volcanic sequence is cut by north-striking rhyolite dikes and is capped by nonwelded rhyolitic tuff overlain by two rhyolite lava flows (Fig. 12A) (Fox et al., 1973).

The age of the volcanic center ranges from 4.35 Ma to 4.3 Ma (40Ar/39Ar, Table 3), based on samples of the stratigraphically lowest and highest exposed andesite flows, respectively (Fig. 12). The uncertainty associated with the dates means that these two dates are essentially identical; it is clear that the rocks from this volcanic center were erupted over a very short period of time. Granitic rocks associated with the volcanic center (Fig. 12A) are some of the few outcrops of intrusive rocks associated with the Sonoma volcanic field. We suggest that these granitic rocks occupy the vent area for the Stags Leap volcanic center and may be time-equivalent to rhyolite dikes that intrude the andesitic pile and to the extrusive rhyolite flows that cap the volcanic sequence (Fig. 12).

To the north of the Stags Leap center on the east side of Napa Valley, the Wildlake center (Fig. 3) is also dominated by andesites but is much less flow-dominated than the Stags Leap center. The Wildlake center is mostly composed of small-volume andesitic pyroclastic eruptions with a smaller component of andesitic flows (Table 2). Using the areal extent of the volcanic center, the exposed thickness of the volcanic section, and the estimated total thickness from the depth-to-basement map (Fig. 5), we have estimated the erupted volume of material of the Stags Leap center to be ∼150 km3 and the Wildlake center to be ∼100 km3. For comparison, these volumes are similar to those of smaller stratovolcanoes in the modern Cascades arc, but smaller than the largest of the Cascade volcanoes such as Medicine Lake (600 km3) or the Newberry volcano (450 km3).

Cup and Saucer Volcanic Center

The Cup and Saucer volcanic center (CS, Table 2 and Fig. 3) includes volcanic rocks exposed near the city of Napa and in the surrounding hills; volcanic rocks associated with this center extend from near Carneros Valley on the west to Green Valley on the east (Figs. 2 and 3). Volcanic rocks from this center unconformably overlie Eocene sedimentary rocks and Great Valley sequence to the south of the city of Napa (Graymer et al., 2002a). Rocks from this center are unconformably overlain by volcanic rocks from the Stags Leap center on the north and the Mount George center in the hills to the east of Napa (Fig. 13). Rocks from this center include a lower section of andesite and minor interbedded tuff that is at least 300 m thick overlain by a 150-m-thick section of andesite, dacite, and rhyolite ash-flow tuffs that are interbedded with generally coeval small-volume extrusive domes and lithic nonwelded tuffs with a similar compositional range (Table 2). The erupted volume from the Cup and Saucer center is difficult to estimate given the great uncertainty in the thickness of the intracaldera pile and the relative lack of map data concerning distribution of outflow tuff.

Calderas and Intracaldera Rocks

On the basis of geologic, geochemical, and geophysical data, we interpret at least two nested calderas within the Cup and Saucer volcanic center (Figs. 3 and 13) that appear to be the principal eruptive vents for three regionally distributed ash-flow tuff and tephra deposits (see stratigraphic columns in McLaughlin and Sarna-Wojcicki [2003] and McLaughlin et al. [2005]). We define these calderas based on the presence of a semicircular topographic low with an interior topographic high, the presence in this vicinity of the coarsest pumice breccia and greatest abundance of lithic clasts in outflow tuffs (Sarna-Wojcicki, 1976), the presence within the interpreted calderas of two compositionally distinct types of megabreccia supported by a tuff matrix, and geophysical evidence for multiple arcuate structural margins (Table 2; see also Langenheim et al., 2010). A portion of one of the calderas was portrayed by Bezore et al. (2004), but the relative lack of outflow facies ash-flow tuff in the vicinity of the “Cup and Saucer” (the central topographic high, Fig. 13A) and lack of definitive ties between regionally distributed fall material and ash-flow tuffs at the volcanic center have hindered previous interpretations of this volcanic center.

Intracaldera rocks lie within the arcuate topographic depression formed by the hills that ring the town of Napa, and are well-exposed in the “Cup and Saucer,” which we interpret as a resurgent dome (Fig. 13A). Intracaldera rocks are a complex assemblage of rhyolite breccia, generally consisting of clasts of welded rhyolite in a nonwelded tuffaceous matrix. The intracaldera breccia is crudely stratified and consists of three phases that are spatially, lithologically, and geochemically distinct: an andesite to dacite breccia that is exposed to the south of the main topographic prominence, a rhyolitic phase that comprises most of the breccia outcrops, and a separate capping rhyolite breccia (Fig. 13A).

Andesite to Dacite Breccia

The andesitic to dacitic intracaldera breccia (63% SiO2) is a monolithologic breccia consisting of dark andesitic fragments up to 0.3 m in a matrix of yellow nonwelded to perlitic tuff. Matrix and clasts have nearly identical major- and trace-element chemistry, with distinctly low Zr concentrations and Rb/Sr ratios and high concentrations of TiO2 relative to the other phases of intracaldera breccia at the Cup and Saucer (Fig. 11). This breccia was mapped as a distinct unit called “the Andesite of Tulocay Creek” by Fox et al. (1985b).

Main Rhyolitic Breccia

The main rhyolitic type of intracaldera breccia (Fig. 13A) forms the bulk of the highland within the town of Napa and was mapped as the “tuff breccia of Napa” (Fox et al., 1985b). In places this type of intracaldera breccia is a monolithologic breccia with angular, aphyric rhyolite clasts up to 0.4 m within a rhyolitic tuff matrix. Elsewhere, the breccia clasts are flow-banded dacite commonly 20–40 cm but up to 3 m in diameter within an altered tuff matrix. This main breccia phase is rhyolitic (71%–76% SiO2) with low concentration of TiO2, low Zr/Ba, and high Rb/Sr relative to the other phases of intracaldera breccia at the Cup and Saucer (Fig. 11). In one location, vitrophyre grades downward into partly vitric, densely welded tuff and then downward into intracaldera breccia with a densely welded tuff matrix. The presence of vitrophyre strongly suggests that the tuff and breccia were emplaced as a hot mass during the caldera-forming eruption, rather than as a cold “mega-landslide” block as envisioned by some previous workers (Howell and Swinchatt, 2003). We interpret breccia bodies containing flow-banded dacite clasts to be collapse-related deposits within the caldera; as such dacite is common outside of the inferred structural margin of the caldera.

Capping Rhyolite Breccia

The third breccia phase is a texturally distinct rhyolite breccia that overlies the main breccia phase (Fig. 13A). This is also a rhyolite breccia (71%–73% SiO2) that is distinguished by the high Zr concentrations, high Zr/Ba, and low concentration of TiO2 relative to the other phases of intracaldera breccia at the Cup and Saucer (Fig. 11).

Correlative Outflow Tuffs

The three phases of intracaldera breccia from the Cup and Saucer volcanic center are similar in whole-rock major- and minor-element geochemistry to three known outflow tuffs, the Pinole Tuff (5.2–5.4 Ma, K/Ar, Sarna-Wojcicki, 1976), the Lawlor Tuff (4.84 ± 0.02 Ma, 40Ar/39Ar, McLaughlin et al., 2005), and the tuff of Napa (<4.70–4.71 Ma, 40Ar/39Ar, McLaughlin et al., 2005) (Fig. 11; see also stratigraphic columns in McLaughlin and Sarna-Wojcicki [2003] and McLaughlin et al. [2005]).

Pinole Tuff

In the vicinity of southern Napa Valley, the Pinole Tuff is a lithic-rich tuff with distinctive scoriaceous black rounded pumices 2–4 cm in size set in a dark brown andesitic partly welded tuff matrix (Fig. 14C). Angular lithic blocks up to 10 cm constitute 5%–10% of the rock mass; the scoriaceous pumice typically makes up ∼20% of the rock. The tuff can be mapped as a reasonably continuous band along the lower eastern slopes of Mount George, to the east of the city of Napa (Fig. 13A). Whole-rock major- and minor-element geochemistry of this tuff bears a distinct similarity to the andesitic phase of intracaldera breccia (Fig. 11).

As originally defined in exposures near Pinole and Rodeo along the south side of San Pablo Bay (Fig. 1), the Pinole Tuff consists of a series of lapilli and vitric-crystal-lithic tuff beds with an aggregate thickness of ∼50 m, including a basal white, massive tuff, succeeded by lithic pumiceous ashy tuff in layers up to a few meters thick (Lawson, 1914; Clark, 1912; Vitt, 1936; Chesterman, 1956). Parts of this section were identified by these early workers as having an andesitic composition on the basis of lack of quartz and presence of augite and hypersthene. The volcanic section originally defined as Pinole Tuff along the south side of San Pablo Bay is really a stack of distal fall and tephra related to many units in the Sonoma volcanic field. The exposed sections are geochemically heterogeneous and stratigraphically complex, clearly belonging to several distinct eruptive and/or fall units (Fig. 14A). Outcrops are stratigraphically layered consisting of falls and tephra of varying pumice and lithic content (Fig. 14A). Samples collected from multiple units in this section have a relatively wide geochemical spread, only some of which correspond to the outflow Pinole Tuff (Fig. 11). Whole-rock analyses of samples from single beds of pumiceous andesitic tuff with distinctive black scoriaceous pumice in exposures near Pinole and Rodeo (Figs. 1 and 14B) correspond to the whole-rock chemistry of outflow Pinole Tuff near Mount George (Fig. 14C) and to the whole-rock chemistry of the andesitic phase of the intracaldera megabreccia in the Cup and Saucer (Fig. 11). We infer that the Pinole Tuff was erupted from a caldera within the Cup and Saucer volcanic center, that the tuffs at Mount George represent the proximal outflow facies, and the scoria-bearing unit in the Rodeo and Pinole sections represents a more distal, predominantly fall, part of the unit.

Lawlor Tuff

Regionally distributed samples of fall deposits and reworked tuff identified as the Lawlor Tuff (Sarna-Wojcicki, 1976) are similar in major- and minor-element geochemistry to the main rhyolitic phase of intracaldera breccia, including the exposures of vitrophyre, in the Cup and Saucer area (Fig. 11). Geochemical data from glass analysis of tephra identified as Lawlor Tuff (Sarna-Wojcicki, 1976) fall in a very tight cluster (Fig. 11). The Lawlor Tuff is interpreted to have been erupted from the southern part of the Sonoma volcanic field on the basis of coarsening and thickening of plinian tephra deposits toward the volcanic field (Sarna-Wojcicki, 1976). Coarse plinian tephra of the Lawlor Tuff crops out along Monticello Road, just to the east of the Cup and Saucer complex (Fig. 13A) (Sarna-Wojcicki, 1976). Few exposures of welded outflow are known for the Lawlor Tuff. One possible correlative outflow tuff is a thick rhyolitic ash-flow tuff exposed in the Syar Industries quarry to the south of the town of Napa (Fig. 13A); this tuff is at least 60 m thick with a well-developed basal vitrophyre. Although we analyzed only a single sample, it is a good geochemical match for the rhyolitic phase of the intracaldera breccia and to the fall deposit and reworked tuff identified as the Lawlor Tuff (Fig. 11). We tentatively suggest that this tuff, which directly overlies the lower, dominantly andesitic part of the volcanic section within the Cup and Saucer volcanic center, may be a proximal outflow facies of the Lawlor Tuff.

Tuff of Napa

The tuff of Napa is a clast-supported, dacitic to rhyolitic pumice lapilli tuff breccia and pumice block breccia that has been extensively mined for pumice to the east of the Cup and Saucer area (Fig. 13) (Chesterman, 1956). Typical outcrops consist of massive beds (0.6–6 m thick) of unsorted, angular pumice fragments with minor lithic clasts of andesite, rhyolite, and dacite (Fig. 14D). Pumice clasts are coarse, averaging 1.3–5 cm in diameter, with blocks as much as 40 cm long; the size of pumice blocks and thickness of deposit are evidence that this unit is proximal to the eruptive vent (Sarna-Wojcicki, 1976). The pumice deposits are broadly distributed in the hills to the east of the town of Napa (Fig. 13B); correlative pumiceous tuff is common in the vicinity of Glen Ellen and Bennett Valley, south and west of the town of Kenwood (Delattre et al., 2007). Above Milliken Canyon (Fig. 13A) there is a partly welded tuff that has similar chemistry and may represent outflow tuff from the same pumice-producing eruption. The capping rhyolitic part of intracaldera megabreccia in the Cup and Saucer is similar in major- and minor-element geochemistry to the tuff of Napa (Fig. 11). Pumice-rich outflow tuff has a broad range in major- and minor-element geochemistry, perhaps in part due to loss of alkalis from the porous material. We infer that the tuff of Napa was erupted from a caldera within the Cup and Saucer volcanic center, the principal eruptive products being pumice breccias and flows deposited in the Mount George area and farther to the west. A number of pumiceous tephra beds in exposures near Pinole and Rodeo along the south side of San Pablo Bay (Fig. 1) are geochemically similar to the tuff of Napa (Fig. 11) and may represent a more distal, predominantly ash fall, part of the unit.

DISCUSSION OF VOLCANO-TECTONIC ENVIRONMENT

Age Progression and Dextral Offset of Volcanic Centers

Neogene volcanic rocks of the northern San Francisco Bay region are part of a linear belt of volcanic fields that are progressively younger to the northwest (Fox et al., 1985b). Mirroring the regional trend, the volcanic centers herein defined on the east side of Napa Valley become progressively younger to the northwest (Fig. 15) from 5.4 to <4.7 Ma within the Cup and Saucer volcanic center at the south end of the valley to the 4.35 to 4.3 Ma Stags Leap volcanic center, through the 3.2 to 2.8 Ma Wildlake and 2.85 Ma Mount St. Helena volcanic centers (citations for ages given in Table 2). Farther to the north, 2.2–0.09 Ma volcanic rocks of the Clear Lake volcanic field (Donnelly-Nolan et al., 1981; Schmitt et al., 2003a, 2003b) are the youngest Neogene rocks in the northern California Coast Ranges (Fig. 15). Volcanic rocks of the SVF show a variety of eruptive styles that are spatially variable and specific to an individual eruptive center, but do not follow the northward age progression of the volcanic centers (Fig. 15).

The volcanic centers on the east side of Napa Valley are largely intact, existing within a structural block bounded on the west by the Carneros and West Napa faults and on the east by the Green Valley fault (Fig. 15). However, volcanic rocks in the western half of the volcanic field have been displaced from their original depositional positions by dextral faults of the San Andreas system (Fox, 1983; Fox et al., 1985b; McLaughlin et al., 1996; Dickinson, 1997; Wakabayashi, 1999; Graymer et al., 2002b). Based on similarity in age, eruptive style, and geochemical traits, we suggest that the Sugarloaf Ridge volcanic center, located to the north of Sonoma Valley (Fig. 15), is an example of such dextral offset, being displaced from the Cup and Saucer volcanic center. The two centers are similar in age; volcanic rocks within the Sugarloaf center are 5.65–4.81 Ma, rocks within the Cup and Saucer center are 5.4 to <4.70 Ma (Table 2). The centers share geochemical traits in having some rocks that are alkaline to peralkaline (Fig. 10). The style of volcanism is somewhat similar in both fields; although we cannot document any caldera structures in the vicinity of the Sugarloaf Ridge volcanic center, we do recognize small-volume fragmental volcanic material associated with mafic vents. Langenheim et al. (2010) use offset gravity and magnetic anomalies to infer a combined 25–30-km right-lateral displacement on the Carneros and West Napa faults. Restoring the fault offset would place the Sugarloaf Ridge volcanic center to the southwest of the Cup and Saucer volcanic center, putting these two similar centers in proximity (Fig. 15). Langenheim et al. (2010) correlate deep-source gravity and magnetic anomalies across the West Napa fault to suggest as much as 40 km of right-lateral offset of Mesozoic features, potentially indicating a slip history that predated eruptions from the Cup and Saucer and Sugarloaf volcanic centers. Restoration of the Sugarloaf Ridge volcanic center is an example of how the Sonoma volcanic field may be reconstructed using a variety of geologic criteria. We have not attempted to reconstruct fault blocks to the west of the Rodgers Creek fault (Fig. 15) given our limited sampling and mapping and the necessity to constrain offset on various fault strands.

Tectonic Setting and Reconstruction

A slab window forms where the subducting slab is removed from beneath the overriding plate (Dickinson and Snyder, 1979; Thorkelson, 1996) (Fig. 16A). Removal of a subducting slab typically results in cessation or reduction of arc volcanism, which is often supplanted by volcanism of different eruptive style and composition (Johnson and O'Neil, 1984; Cole and Basu, 1995). Slab window formation can result in an increase in heat flow as hot asthenosphere that previously existed beneath the slab (subslab lithosphere) or in the lithospheric mantle wedge above the subducting slab (supraslab) fills the region previously occupied by the subducting slab (Liu and Furlong, 1992; Furlong and Schwartz, 2004) (Fig. 16A).

The transition from subduction and associated arc volcanism to a slab window tectonic environment along the western margin of the American plate began in the Miocene as the triple junction first impinged on the North American plate (Dickinson and Snyder, 1979; Johnson and O'Neil, 1984; Fox et al., 1985a). As the Mendocino Triple Junction migrated northward, the boundaries of the slab window were defined at the north end by the southern edge of the Gorda plate and to the west by the San Andreas fault and the edge of the Pacific plate (Furlong and Schwartz, 2004) (Fig. 16B). The east-west width of the window was controlled by the dip of the descending slab and the resultant shape of the overlying lithosphere of the North American plate (Thorkelson, 1996). In general, the migratory volcanic centers of north-central California with clear northward-younging trends follow the northward progression of the Mendocino Triple Junction and are in some way associated with thermal perturbations associated with the transition from subduction setting to a strike-slip setting and the development of a slab window tectonic environment (Fox et al., 1985a; Furlong, 1984; Liu and Furlong, 1992; Furlong and Schwartz, 2004; Dickinson, 1997). Neogene volcanic rocks in north-central California are roughly coeval with the inactive subduction-related ancestral Cascade magmatic arc in the Bodie-Lake Tahoe area (Cousens et al., 2008) (Fig. 16B), however, volcanic rocks of the Sonoma and other volcanic fields erupted considerably to the west of the magmatic arc and therefore are unlikely to be related to arc magmatism.

Of critical importance are tectonic reconstructions that attempt to pinpoint the location of the volcanic fields with respect to the paleoposition of the Mendocino Triple Junction (Fox et al., 1985b; Dickinson, 1997). Many reconstructions of the northwestwardly younging volcanic rocks in the northern California Coast Ranges are portrayed with a static North American reference frame such that all of the volcanic rocks are plotted as they exist today and the position of the triple junction is shown through time (Johnson and O'Neil, 1984; Fox et al., 1985b). In these diagrams, there is considerable uncertainty in the exact location of the Mendocino Triple Junction with respect to the location of a specific volcanic field, and thus, there exists some uncertainty as to the relation of these volcanic rocks to the slab window forming in the wake of the triple junction. Recent kinematic reconstructions of the western United States (McQuarrie and Wernicke, 2005; Wilson et al., 2005; McCrory et al., 2009) integrate seafloor magnetic anomaly data with regional extension and block rotations across the western United States in addition to correlations across fault blocks to arrive at dynamic tectonic reconstructions through time. We use the reconstruction of Wilson et al. (2005) because in the northern California Coast Ranges, in addition to the San Andreas fault, the Rodgers Creek and the Calaveras faults (for our purpose, approximating the Green Valley fault at the latitude of the Sonoma volcanic field) are incorporated into the reconstructions, allowing for the volcanic rocks of the Sonoma volcanic field to be placed within a discrete fault block (Fig. 17). In addition, this reconstruction used the volcanic fields as a constraint on the location of fault blocks with respect to the slab edge.

On Figure 17 we show the tectonic reconstruction of Wilson et al. (2005) at 5.9 Ma, which is roughly coincident with the early phase of the Cup and Saucer volcanic center. The reconstruction is fixed with respect to the Pacific plate. Superimposed on the fault-block reconstruction are the volcanic centers that we have identified along the east side of Napa Valley. We have restored 30 km of combined offset on the Carneros and West Napa faults such that the Sugarloaf Ridge volcanic center is adjacent to the Cup and Saucer volcanic center at this time. All of these volcanic centers fall between the Rodgers Creek fault to the west and the Calaveras fault to the east (Fig. 1). We superimpose on this reconstruction the successive position of the southern edge of the Gorda Plate (north edge of the slab window) from 8 Ma to present, following Atwater and Stock (1998).

Volcanism associated with the slab window migrated northward in response to northward migration of the triple junction while the south end of Cascades arc volcanism also migrated northward to its present location at Lassen volcanic field (Cousens et al., 2008) (Fig. 17). Between 5 and 10 Ma, when the oldest rocks within the Sonoma volcanic field rocks were forming, arc volcanism was ongoing along the Ancestral Cascades magmatic arc at approximately the same latitude. Arc volcanism migrated northward at a slightly faster rate than did volcanism within the Sonoma volcanic field, such that at the time of the Mount St. Helena eruptions at 2.85 Ma, coeval ancestral Cascades volcanism is ∼50 km farther north. This disparity increases for arc rocks younger than 2.6 Ma, which are situated well north of the generally coeval Clear Lake volcanic field.

Dickinson (1997) and Atwater and Stock (1998) differ significantly concerning the location of the reconstructed edge of the slab. For older time steps, such as 6 Ma, Dickinson's (1997) reconstruction puts the Mendocino Triple Junction much farther to the south. As a result, Dickinson's (1997) reconstruction results in the older fields having formed closer to the slab edge, with ever-increasing distances to the slab edge for the younger fields. In contrast, the Atwater and Stock (1998) reconstruction results in a relatively constant 90–110 km distance between the volcanic fields and the slab edge (Fig. 17). This constant difference is consistent with modern heat-flow measurements and thermal modeling (Liu and Furlong, 1992; Furlong and Schwartz, 2004) that suggests a time (and thus distance) lag between the passage of the triple junction and the generation of the largest heat-flow anomalies associated with movement of asthenospheric material into the slab window.

DISCUSSION OF VOLCANIC-ROCK GEOCHEMISTRY

Volcanic rocks from slab window settings have a variety of geochemical signatures; the chemistry of the erupted material depends on the tectonic setting, thermal history, and chemical characteristics of the mantle source, as well as the degree of crustal interaction and the pathway taken by the rising melt (Thorkelson, 1996). Slab windows interpreted to be related to ridge-trench collision may result in depleted, mantle-derived melts that ultimately produce tholeiitic- to alkalic-composition volcanism (Johnson and O'Neil, 1984; Cole and Basu, 1995). Slab window settings where mid-oceanic-ridge basalt–like rocks with alkaline affinities have been identified include the central and southern California coast (Johnson and O'Neil, 1984; Cole and Basu, 1995), Baja California, Mexico (Bellon et al., 2006; Benoit et al., 2002), Costa Rica and Panama (Johnston and Thorkelson, 1997), and the Antarctic Peninsula (Hole, 1988, 1990). Slab window volcanism can also produce rocks whose geochemical signatures are related to enriched mantle sources similar to oceanic-island basalts (OIB), such as in Patagonia (Gorring and Kay, 2001). With elevated heat flow the trailing edge of the slab at the margin of the slab window may melt, resulting in the eruption of sodic rocks including adakites (Defant and Drummond, 1990; Martin et al., 2005).

Although we do not have trace-element modeling results or isotopic data to constrain petrogenetic interpretation, here we use trace-element ratios to compare the volcanic rocks of the Sonoma volcanic field to rocks from other slab window settings and to generally coeval rocks of the ancestral Cascades. Chemical analyses of basalts and basaltic andesite are useful in defining melt sources and processes because these rocks are least likely to be affected by crustal contamination or fractionation. In Figure 18 we compare our geochemical data from 20 basaltic (<52% SiO2) lavas and dikes and 33 basaltic andesite lavas (52%–57% SiO2) with similar composition rocks from the Clear Lake volcanic field, the ancestral Cascade magmatic arc, and rocks from various slab window-related volcanic fields. For the purposes of geochemical classification, volcanic centers from the eastern half of the SVF are grouped into three general classes (Fig. 18): (1) volcanic centers that range from mafic to silicic, are explosive in eruptive style, and include some rocks that trend toward alkalic compositions (Cup and Saucer and Sugarloaf Ridge centers); (2) dominantly basaltic to andesitic fields (Stags Leap and Wildlake centers); and (3) dominantly silicic volcanic centers at the north end of the SVF (Mount St. Helena and Calistoga Domes centers). These groupings allow the geochemical characteristics of the individual centers to be assessed relative to geographic location (south to north) and age (oldest to youngest). All other mafic rocks, predominantly from the western half of the volcanic field, are denoted on Figure 18 as “other.”

Mafic rocks of the Sonoma and Clear Lake volcanic fields generally have medium-K calc-alkaline series compositions, with generally increasing total alkali content with silica (Fig. 18A). In this sense only, SVF rocks are similar in composition to the ancestral Cascade magmatic arc rocks (Fig. 18A). Some samples from the Sugarloaf Ridge and Cup and Saucer volcanic centers trend toward alkaline compositions, but in general mafic rocks of the Sonoma volcanic field appear distinct from intraplate alkalic rocks with OIB affinities from some slab window settings (Hole, 1990; Gorring and Kay, 2001) (Fig. 18A). Total alkali content varies subtly but systematically with age and spatial position of the SVF volcanic centers (Fig. 18A). The older, more southerly volcanic centers such as the Cup and Saucer and related Sugarloaf Ridge volcanic center are more sodic and have higher total alkali content than younger, more northerly centers such as CD-MSH and the Clear Lake field (Fig. 18A). Sonoma volcanic field basalts are characterized by primitive compositions, including moderate to high MgO (4.2%–13.1%) and Cr (80–400 ppm), and Mg# that ranges from 40 to 66.

Mafic rocks from the Sonoma volcanic field and other slab window-derived rocks are clearly distinguished from rocks of the ancestral and modern Cascades magamatic arc in having elevated TiO2 concentrations and low LILE/HFSE ratios (here, Ba/Nb) (Fig. 18B). A small number of samples from the Sonoma and Clear Lake volcanic fields display a “subduction component” and overlap with the data from the ancestral and modern Cascades magamatic arc (Fig. 18B). Basalt and basaltic andesite from the Sonoma volcanic field centers are similar in composition to slab window tholeiites from southern Baja (Benoit et al., 2002), but are enriched in Ba and have elevated Ba/Nb ratios relative to mafic rocks from ridge collision slab window environments that have OIB affinities such as Patagonia (Gorring and Kay, 2001) and the northern Baja Peninsula (Luhr et al., 1995) (Fig. 18B).

The plot of LREE/HREE (La/Yb) versus LILE/HFSE (Ba/Zr) separates the rocks of the Sonoma volcanic field from rocks associated with magmatic arc settings and from the more alkali rocks from slab window environments (Fig. 18C). Basalt and basaltic andesite rocks from the ancestral Cascades magmatic arc have high Ba/Zr ratios and moderately sloping REE patterns indicated by moderate to low values of La/Yb (Fig. 18C). Rocks from OIB-related slab window environments are distinguished by strong HREE depletions (high La/Yb) and lack incompatible element enrichment (low Ba/Zr) (Fig. 18C). Mafic rocks from the Sonoma volcanic field have relatively flat REE patterns (low La/Yb) and low LILE/HFSE (low Ba/Zr) (Fig. 18C). Mafic rocks from the Sonoma volcanic field centers are perhaps most similar in composition to slab window tholeiites from southern Baja (Benoit et al., 2002) (Fig. 18C, points labeled “Th”); Sonoma volcanic field rocks have much flatter REE patterns (low La/Yb) than slab window-related rocks from other tectonic settings in southern Patagonia (Gorring and Kay, 2001) and the Antarctic Peninsula (Hole, 1990). Incompatible trace-element patterns of mafic rocks from the Sonoma volcanic field centers show some similarities to the various lavas from southern Baja (Benoit et al., 2002) but are relatively depleted in Ba and have flatter REE patterns (Fig. 10). In general, mafic rocks from the Sonoma volcanic field centers do not precisely correspond to any of the general compositional categories defined by Benoit et al. (2002) and instead occupy their own space on the incompatible trace-element plots (Fig. 10) and ratio plots (Fig. 18).

Sonoma volcanic field rocks are also distinguished from magmatic arc rocks and rocks from OIB-associated slab window settings when the ratio Ce/Yb is plotted against Ba/Ce (Fig. 18D). Again, the Sonoma volcanic field centers are most similar in composition to slab window tholeiites from southern Baja (Benoit et al., 2002). Schmidt et al. (2008) defined four main segments of the modern Cascade arc on the basis of tectonic setting and mantle melting regimes; the southernmost segment is directly north of the Gorda slab edge. High values of Ba/Ce in the southern segment rocks were viewed by Schmidt et al. (2008) as indicative of subducted slab–derived fluids and decreasing Ce/Yb as indicative of increased degree of melting (Fig. 18D), using the trace-element ratios defined by Cameron et al. (2002). The distinct fluid-rich chemistry of the southern Cascades segment, similar to many of the Sonoma volcanic field rocks, was attributed to proximity to the Gorda plate edge and the slab window (Schmidt et al., 2008) (Fig. 18D).

CONCLUSIONS

Volcanic rocks in the Sonoma volcanic field contain heterogeneous assemblages of a variety of compositionally diverse volcanic rocks. We have used field mapping and whole-rock geochemistry in combination with new and existing age determinations to define for the first time discrete eruptive centers. Along the eastern side of Napa Valley, volcanic centers become progressively younger northwards, mirroring the regional age progression of volcanic rocks in the northern California Coast Ranges (Fox et al., 1985a) and the northward migration of the Mendocino Triple Junction (Dickinson and Snyder, 1979; Johnson and O'Neil, 1984; Fox et al., 1985a). Volcanic rocks of the Sonoma volcanic field show a broad range in eruptive style that is spatially variable, specific to an individual eruptive center, and does not correspond to the general northward age progression. Measured heat flow in the northern California Coast Ranges (Lachenbruch and Sass, 1980) and numerical models of the thermal effects of slab window migration (Liu and Furlong, 1992; Furlong and Schwartz, 2004; Groome and Thorkelson, 2009) indicate abrupt transient thermal effects associated with the northward passage of the triple junction as upwelling asthenospheric material replaces the Gorda slab beneath North America. The rapid temporal and spatial variation in eruptive style and volcanic rock composition would thus seem to indicate that slab window volcanism results from a transitory pulse of elevated heat flow over a relatively small geographic area.

Using tectonic reconstructions that locate the Sonoma volcanic field with respect to the paleoposition of the Mendocino Triple Junction, we surmise that (1) Sonoma volcanic field rocks were erupted in a position some 90–110 km south of the Gorda plate edge; and (2) the Miocene–Pliocene ancestral Cascades magmatic arc was active at about the same time and latitude as the volcanic centers of the Sonoma volcanic field. Our trace-element analyses of whole-rock samples from lavas and tuff from the Sonoma volcanic field highlight the geochemical diversity of volcanic rocks from the slab window setting, allow us to clearly distinguish these volcanic rocks from those of the roughly coeval magmatic arc to the west, and also to compare rocks of the Sonoma volcanic field to rocks from other slab window settings. Most of the mafic rocks of the Sonoma volcanic field have the greatest similarity to tholeiitic slab window-derived rocks in southern Baja California (Benoit et al., 2002). However, some mafic SVF rocks from the Cup and Saucer and Sugarloaf Ridge centers have alkaline compositions, very high Zr content, and relatively small LILE enrichment.

1
Supplemental Table 1. Excel file of geochemical analyses, Sonoma volcanic field. If you are viewing the PDF of this paper or reading it offline, please visit http://dx.doi.org/10.1130/GES00625.S1 or the full-text article on www.gsapubs.org to view Supplemental Table 1.