Paleogeographic reconstruction of regional accretionary complex architecture, Franciscan Complex, northwestern San Francisco Bay Area, California, USA

The Mesozoic–Cenozoic-age Franciscan Complex of western California, USA, is considered by many to be the archetypal subduction accretionary complex (Fig. 1; e.g., Ernst, 2011; Wakabayashi, 2015; Raymond, 2018). Studies dating from just after the development of plate tectonics theory to the present document a record of continental-margin accretion of Franciscan rocks along a subduction zone on the western margin of North America (e.g., Hamilton, 1969; Ernst, 1970, 2016; Blake and Jones, 1974; Dickinson et al., 1982; Wakaba yashi, 1992, 2015, 2021; Raymond, 2018). Accretion at the plate margin produced an architecture in the subduction accretionary complex consisting of an array of faulted masses. Each faulted mass is designated here as an accretionary unit (AU). An AU is defined as a major mappable, lithologically distinctive, and fault-bounded rock-body component of a subduction accretionary complex that consists of lithologic or tectonostratigraphic units truncated at the body margin by boundary-long (major) faults (Raymond, 2016, 2018; Raymond et al., 2020). The individual AUs generally consist of rock masses of partial to complete ocean plate stratigraphy, which are defined as the tripartite sequence of (1) ocean crustal ultrabasic–basic and associated rocks; (2) overlying pelagic sediment, reefal limestone, mudrocks, or a combination of these; and (3) sequence-capping masses of siliciclastic sedimentary rock (Isozaki et al., 1990; and see Chipping, 1971; Wahrhaftig, 1984; Kusky et al., 2013; Raymond, 2017b). AU stacks commonly consist of a younging downward group of AUs, each of which is dominated by sedimentary rocks of the trench and ocean plate stratigraphy (e.g., McLaughlin et al., 2000; Wakabayashi, 2015; Raymond, 2018; Bero et al., 2020; Raymond et al., 2020). Although these broad and general observations of Franciscan accretionary complex structure and history are widely accepted, the details of Franciscan Complex architecture are described in diverse ways and are debated (e.g., compare the cross sections of Prohoroff et al., 2012, and Blake et al., 2000; compare Prohoroff et al., 2012, and Raymond and Bero, 2015; and see Raymond, 2018, figs. 6 and 17). If the Franciscan Complex is to serve as the type example of an accretionary complex, it is essential to clarify the details of local- to- regional Franciscan architecture.

The current subduction accretionary complex architecture of western California varies geographically. North of the greater San Francisco Bay Area, the Franciscan architecture in the Northern Coast Ranges (Fig. 1), where major strikeslip faulting is not extensive within the complex, has been outlined via small–medium-scale reconnaissance studies (e.g., Worrall, 1981; Blake et al., 1999; McLaughlin et al., 2000). In that northern area, there is a broad regional and general downward (westward) younging of accreted units (McLaughlin et al., 2000; Dumitru et al., 2015). In contrast, the Franciscan architecture in the greater San Francisco Bay Area and the Diablo Range of central California has been complicated by an overprint of extensive late Cenozoic deformation, particularly faulting (Figs. 1 and 2; e.g., Buising and Walker, 1995; Page et al., 1998; Wakabayashi, 1999, 2015; McLaughlin et al., 2012; Raymond, 2018, fig. 8). In the San Francisco Bay Area, the AUs and overlying Cenozoic rocks occur on a group of major crustal blocks bounded by highangle, late Cenozoic faults of the San Andreas (transform) dextral strikeslip fault system (Fig. 2). Cenozoic thrust faulting and folding affected these blocks as well (e.g., Wagner et al., 2002; Wagner et al., 2011; McLaughlin et al., 2012). The history of accretion and subsequent Cenozoic deformation resulted in an assemblage of crustal blocks, each with Franciscan Complex AUs as basement (the structurally lowest crustal rock mass) and in most cases overlain by late Cenozoic to modern sediments and volcanic rocks (Figs. 2 and 3; Bailey et al., 1964; Ernst, 1970, 2011; Berkland et al., 1972, McLaughlin et al., 2012; Wakabayashi, 2015; Raymond, 2018).

In the greater San Francisco Bay Area, numerous local-to-regional reconnaissance studies have highlighted details of the Franciscan Complex geology (e.g., Berkland, 1969; Schlocker, 1974; Wahrhaftig, 1984; Blake et al., 2000; McLaughlin et al., 2001; Graymer et al., 2007, 2018; Prohoroff et al., 2012; Bero, 2014; Wakabayashi and Rowe, 2015; Bero et al., 2020). Yet, overlapping studies have led to strikingly different structural and architectural pictures (for visual images, compare the cross sections of Blake et al., 2000, and Prohoroff et al., 2012, or the maps of Berkland, 1969, and Rice et al., 2002). In part, these conflicting images of the architecture result from (1) perceptions based on models envisioned for the architecture, and (2) the fact that long distance, late Cenozoic block movements that disrupted the original (primary) accretionary architecture are not accounted for in architectural reconstructions. As a result, a clear understanding of Franciscan accretionary complex architecture in the San Francisco Bay Area and consensus on architectural details do not exist.

The Franciscan Complex architecture of the northwestern San Francisco Bay Area, as is the case regionally in the greater San Francisco Bay Area, consists of numerous regional- to localscale rock masses (the larger of which are AUs), which are bounded by AU or blockbounding faults (Fig. 3; Raymond, 2018; Raymond et al., 2020; Bero et al., 2020; this report). Across the northwestern San Francisco Bay Area, from southwestern Sonoma County to southern Marin County, two of the large blocks—the Mt. Tamalpais Block on the southwest and the Novato Block on the northeast—dominate the structural framework. The Tamarancho Shear Zone separates these blocks (Figs. 2 and 3). Two smaller Franciscan-based blocks, the Marin Headlands Block and the Nicasio Reservoir Block, overlie parts of the Mt. Tamalpais Block and are bounded at the base by thrust faults (Fig. 3; Wahrhaftig, 1984; Prohoroff et al., 2012; Bero et al., 2020). These two blocks lack an exposed upper contact. One smaller additional highangle, fault-bounded block, the Tiburon Block, is exposed on Tiburon Peninsula and Angel Island in the southeastern portion of the northwestern San Francisco Bay Area (Fig. 2). The entire group of Franciscan rock–based blocks here is bounded on the southwest by the San Andreas fault (sensu stricto) and on the northeast by the Petaluma Valley–Point Richmond–Silver Creek fault (Fig. 3). Our work has been concentrated within these blocks in the central portion of the northwestern San Francisco Bay Area.

The purposes of this paper are multiple. We (1) describe the Tamarancho Shear Zone of the northwestern San Francisco Bay Area; (2) briefly describe and compare the tectonostratigraphies of the Mt. Tamalpais and Novato blocks; (3) show how paleogeographic/palinspastic reconstruction and the radically different tectonostratigraphies of the two major blocks of the northwestern San Francisco Bay Area Franciscan Complex require that the two blocks be separated along and across strike in the original accretionary complex architecture, revealing significant variations in accretionary history; and (4) briefly discuss how current models of the structure and architecture of the northwestern San Francisco Bay Area Franciscan Complex do not reflect the impacts of Cenozoic deformation, and hence are not accurate representations of the primary accretionary architecture.

The broadscale Franciscan structure and architecture of the northwestern San Francisco Bay Area have been cast in terms of terranes and nappes over the past several years (e.g., Blake et al., 1984, 2000; Wakabayashi, 1992, 2015). Hence, individual structural units have been assigned names that presumably reflect similar properties of rock masses while implying certain details of accretionary architecture. As discussed by Raymond (2018), the nappe rubric is inappropriate for subduction accretionary units because it denotes horizontal overthrust movements of thin sheets versus the underthrusting typical of subduction zones (e.g., Dennis, 1967; Tollman, 1987). Most terrane designations for rock masses of the northwestern San Francisco Bay Area do not meet the defining criteria for individual terranes, notably because the terranes contain units, particularly chert units, which are correlative with units in other rock masses, and this feature is prohibited by the terrane definition (Howell and Jones, 1984; Raymond, 2018).

Instead of designating rock masses as terranes, we designated the major fault-bounded units of the northwestern San Francisco Bay Area as AUs (Raymond, 2016, 2018; Raymond et al., 2020; Bero et al., 2020; Fig. 3). Each AU is characterized by specific rock assemblages of particular ages and histories. For example, sandstones and shales of both distal and proximal submarine-fan facies (facies A–E), which have been dismembered by postsedimentation deformation, characterize the Novato Quarry AU (fan facies used herein are those of Mutti and Ricci Lucchi, 1972, 1978). Detrital zircons from likely correlative exposures to the east suggest deposition during the Late Cretaceous at ca. 83 Ma (Snow et al., 2010). In contrast to the Novato Quarry rocks, the Alcatraz AU primarily has fan-channel deposits of submarine-fan facies A or B and C, plus overbank facies E submarine-fan sandstone-shale facies, which although broken, do not appear to be predominantly dismembered. Alcatraz AU rocks were deposited between ca. 100 Ma and 140 Ma, as indicated by fossil content (e.g., Schlocker, 1974; Armstrong and Gallagher, 1977). The Bolinas Ridge AU stack is an assemblage of downward-younging, diverse submarine-fan facies AUs deposited between ca. 122 Ma and 82 Ma (Bero et al., 2020, 2021). (Note that the ages here may differ from those reported. Inasmuch as individual workers use different methods of calculating maximum depositional age [MDA], to enable comparisons from different studies we recalculated MDAs as Y3Z ages, i.e., ages representing the average of the three youngest zircons. The Y3Z ages are reported herein.) The Yolla Bolly AU is characterized by metamorphosed rock masses containing glaucophane, lawsonite, and jadeitic pyroxene—blueschist-facies indicator minerals. The Tiburon Ridge and Angel Island AUs similarly contain blueschist-facies metamorphic rocks. The Nicasio Reservoir AU is a deformed mass of pillow basalts plus minor gabbro and Lower Cretaceous chert, whereas the Marin Headlands Block consists of a stacked set of AU thrust sheets, with each AU composed of fragments of ocean plate stratigraphy (e.g., Wahrhaftig, 1984; Raymond et al., 2020). Each Marin Headlands AU is separated from adjoining AUs by thrust faults, mélange zones, or both. Rocks range from Early Jurassic to Late Cretaceous in age (Murchey, 1984; McPeak et al., 2015).

The Tamarancho Shear Zone was named by Berkland (1969) as a zone of deformed rock in the southwestern corner of the Novato 7.5′ quadrangle (Location [Loc.] 1 in Fig. 3). Here, Berkland’s map shows that the Tamarancho Shear Zone consists of an ~300- m- wide zone marked with shear-zone symbols, which hosts several natural springs and contains three exotic blocks of rock on the northeast—one a serpentinite, another a silica carbonate rock, and the third a chert mass. Additional rock types present within the sheared matrix of the zone include metabasite (“greenstone”) and schist (Berkland, 1969).

The Tamarancho Shear Zone has been variously depicted, and in some cases, simply ignored, in depictions of the geology of the northwestern San Francisco Bay Area. In addition to Berkland’s description, mapping by Gluskoter (1969) in an area located to the northwest of the Novato quadrangle depicted the northern end of the Tamarancho Shear Zone as a single dashed line, which he called a “tectonic zone.” He indicated that near Nicasio Reservoir (Loc. 3 in Fig. 3), the zone is up to one mile (1.6 km) wide. The zone in the Nicasio Reservoir area is described as containing block- in- matrix structure (Gluskoter, 1969), which is also characteristic of mélanges. Our mapping now suggests that it is likely that mélange rocks were mistakenly included by Gluskoter (1969) in the Tamarancho Shear Zone at Nicasio Reservoir during his mapping, which was done before Hsü (1967, 1968) introduced the mélange concept to Franciscan studies.

Blake et al. (1982) produced a terrane map of California. On that map and in a successor report in 1984, the majority of the Franciscan rocks of the north-western San Francisco Bay Area located east of the San Andreas fault were depicted as Central Terrane mélange rocks, replacing an earlier Central Belt designation (Blake and Jones, 1978; this older “Belt” terminology is reviewed in Raymond, 2018). These mélange rocks were shown to enclose masses of Nicasio Reservoir, Novato Quarry, and Yolla Bolly terrane rocks. The revised map published in 1984 (Blake et al., 1984) similarly showed the dominant unit as Central Terrane, but in addition to masses of the three terranes included earlier, the Central Terrane mélange is also shown to include masses of Marin Headlands and San Bruno Mountain Terrane rocks, as well as “blueschist blocks.” The Tamarancho Shear Zone is not depicted on either map, but Blake et al. (1984) mention the zone, calling it a major boundary. They indicate that the Tamarancho Shear Zone separates more westerly “oceanic” rocks from more easterly “continental” rocks, and they depict it in a cross section. On a more recent smallscale terrane map, Blake and two other coauthors (Blake et al., 2000) show a thrust fault partly along the line of Gluskoter’s “tectonic zone” at Nicasio Reservoir and lying within the Central Terrane mélange. The northern extension of the thrust fault (in Blake et al., 2000) continues for several kilometers north, beyond the northern limits of Gluskoter’s (1969) map. On their more detailed reconnaissance geologic map, however, Blake et al. (2000) do not show the fault zone, and they do not show it in their cross section.

In the same year that Blake et al. (1982) released their first terrane map, Wright (1982) completed his dissertation in which he depicted the Tamarancho Shear Zone as a distinct boundary within the northwestern San Francisco Bay Area. He showed the zone separating “undifferentiated” Franciscan Terrane of the “Central Belt”—including northeastern masses of Angel Island Terrane (the Yolla Bolly Terrane of Blake et al., 1982, 1984) and Novato Quarry Terrane—from several other units, including a “Central Marin Terrane,” the Nicasio Reservoir Terrane, another undifferentiated Central Belt unit, and the Bolinas Ridge Terrane (corresponding somewhat to the San Bruno Mountain Terrane of Blake et al., 1984). Later, he (Wright, 1984) relocated the position of the Tamarancho Shear Zone on one but not all of his maps of that year, showing the fault line substantially north of the location in the southwestern corner of the Novato quadrangle, where it had been mapped by Berkland (1969).

On their regional map of this area, Prohoroff et al. (2012) do not label the Tamarancho Shear Zone, but they do show an unlabeled fault in their cross section, likely the Tamarancho Shear Zone, separating their eastern and western mélanges. The location appears to align with Berkland’s (1969) placement of the zone. Thus, the Tamarancho Shear Zone is described as a wide zone, shown as a distinct fault, plotted in different locations, and is entirely ignored on some maps.

The Tamarancho Shear Zone was not mapped by Berkland (1969) beyond the limits of the quadrangle, although he noted that he found evidence of it for at least a mile (in part, beyond the quadrangle boundaries). As we discuss below, Berkland (1969) and we consider the Tamarancho Shear Zone to be a major structure in the region. Using Berkland’s (1969) description as a basis for recognizing the fault zone, we describe some fundamental features of the Tamarancho Shear Zone, noting topographic features and local contrasts in rock types and structure east and west of the zone. Northwest of the southwestern corner of the Novato quadrangle, where Berkland (1969) first identified the Tamarancho Shear Zone, we examined the zone in the area of Roy’s Redwoods Open Space Preserve (Loc. 2 in Fig. 3, see Fig. 4) and at Nicasio Reservoir located farther to the northwest (Loc. 3 in Fig. 3, see Fig. 5). Between these two localities, the Tamarancho Shear Zone is concealed by alluvium, covered by reservoir waters, or lies within generally inaccessible private properties.

At Roy’s Redwoods (Fig. 4), we located a 220- m- wide zone of sheared rock—a tectonic mélange—along strike of the Tamarancho Shear Zone of Berkland (1969). The rocks of the zone are much like those of an ocean-plate stratigraphy mélange (OPS mélange, defined as a mélange containing fragments and blocks of oceanplate stratigraphy rocks), but the zone also contains at least one glaucophane schist block in addition to blocks of metabasite, metachert, and serpentinite breccia (Figs. 6B–6D). All blocks occur within a dominant matrix of shear-fractured lithic metasandstone and metashale (Fig. 6A). Bedding (S₀) is commonly obscure in the metasandstones, which suggests thick beds of submarine-fan facies A and B. Fan facies C and E beds occur locally. Multiple fracture sets (S₁, S₂, etc.) are pervasive. Where S₀ is clearly present, S₁, an apparent shear-fracture fabric, commonly is approximately parallel to S₀. Additional fracture sets are both subparallel and at high angles to S₀, where the bedding is visible.

Northeast of the Tamarancho Shear Zone mélange zone at Roy’s Redwoods is a section of dismembered formation (which by definition is pervasively sheared at the mesoscopic scale) consisting of epiclastic to semischistose sandstone-metasandstone and metamudrock (Fig. 7A). The unit is informally designated here as the Dismembered Formation of Roy’s Redwoods (Kfrr in Fig. 4). The dominant rocks are lithic sandstones/metasandstones. Close to shear planes, the rocks experienced microscopic- to mesoscopic-scale deformation. Coarser-grained rocks locally contain granule- to sand-sized, dark gray mudrock clasts. Submarine-fan facies observed include A(?), B, C, and E. Outcrops commonly consist of yellow-brown weathered, massive facies A or B sandstone or metasandstone pervaded by two to four fracture sets. A few pebble- to cobble-sized fragments of chert occur in colluvium overlying the metasandstones in the northeast. Along the fire road near the northern edge of the map, one green chert clast is exposed within a dark gray metasand-stone layer (Fig. 7B).

Southwest of the Tamarancho Shear Zone, a dismembered sandstone/ metasandstone-dominated unit is exposed that consists of (meta)sandstone and (meta)mudrock. The dominant rock is unmetamorphosed to slightly metamorphosed, weakly semischistose lithic arenite containing rounded, elongate sand- to granule-sized clasts of black mudrock (Fig. 8C). Slight metamorphism is reflected by textural and mineralogical changes. Slight to moderate dynamic metamorphism produced very weak semi-schistose to moderate schistose fabrics in sandstones via recrystallization of matrix materials, framework grains, and lithic clast components. Lowgrade mineralogic metamorphism is represented by neoblastic growth at the micron-scale level (1–25 µm), with acicular phases (possibly pumpellyite) nucleated on grain boundaries, grain inclusions, and at independent nucleation centers within quartz and feldspar grains. These rocks are exposed in road cuts along the Nicasio Hill portion of Nicasio Valley Road and in exposures to the west (Figs. 4 and 8). In the unit, locally deformed and slightly metamorphosed sandstone layers appear to be more continuous than the metasandstone layers present within the Tamarancho Shear Zone. Bedding ranges from thin to massive, and submarine-fan facies A, B, C, D, E, and G are locally recognizable. The lithology of the unit is similar to that of the Kfu5 (metalithic arenite) unit of Bero et al. (2020), which is exposed to the south at western Mt. Tamalpais, but the unit here is informally called the Dismembered Formation of Nicasio Hill and is designated as “Kfnh” in Figure 4. This western unit appears to represent a belt of rock that, perpendicular to the strike of the Tamarancho Shear Zone, is slightly more than half a kilometer wide. To the southwest, it is succeeded by an unnamed mélange unit containing blocks of blueschist-facies metamorphic rocks (an uncolored unit in Fig. 4). Many exposures of the Kfnh are like those of the Dismembered Formation of Roy’s Redwoods (Kfrr) in being massive with multiple joint sets.

Our mapping reveals lithofacies essentially identical to those of Kfnh near the westcentral part of the Roy’s Redwoods area map (Fig. 4). Here, between the unnamed southwestern mélange unit and the Tamarancho Shear Zone, and capping the ridge followed by a fire road (dashed line near the label “Private Property” in Fig. 4), Blake et al. (1974, 2000) depicted a blueschistfacies “Yolla Bolly” terrane unit that is composed of semischistose to schistose rocks and surrounded by Central “Terrane” mélange on its east, south, and west (Fig. 9). Yet, west of the Nicasio Valley Road (the generally northtrending road marked with red and white dashes in Fig. 4) and across the location of the eastern-bounding fault of this proposed “Yolla Bolly” unit, we found no semischistose to schistose, blueschist-facies rocks, but rather found Kfnh-like rocks. Hence, our map does not depict a “Yolla Bolly” body here, as discussed in more detail in the Tectonostratigraphic Sections on the Novato and Mt. Tamalpais Blocks section below.

Our second map of the Tamarancho Shear Zone is a map of the northwestern Nicasio Reservoir area (Fig. 5). Here, Gluskoter (1969) had considered the Tamarancho Shear Zone tectonic zone to be concealed, at least in part, by a western arm of the reservoir. The Novato Block is located to the northeast, whereas to the southwest there are rocks of both the Mt. Tamalpais Block and the structurally overlying, dominantly volcanic Nicasio Reservoir AU (Weaver, 1949; Gluskoter, 1969; Bero et al., 2020; Fig. 3). We concur generally with Gluskoter’s (1969) placement of the Tamarancho Shear Zone, but we find the zone here to be substantially narrower than the 1.6 km width he suggested. Here, the Tamarancho Shear Zone contains highgrade blocks of glaucophane-schist breccia (“highgrade” as used by Coleman and Lanphere, 1971) and related actinolite schist and diablastite, as well as blocky serpentinized peridotite, chert, metachert, chlorite metabasite, and meta basite breccia (Figs. 5, 10, and 11)—the main rock types noted in the zone to the southeast at Roy’s Redwoods and the Novato quadrangle—with some additional rock types. At Nicasio Reservoir, the Tamarancho Shear Zone is largely concealed by water and colluvium, the latter occurring abundantly on both the central reservoir peninsula/island and the western shore of the reservoir. We estimate the width of the zone at ~350 m. The width is difficult to determine due to limited exposures and because the zone cuts a mélange unit containing rocks similar in character to those of the zone. Northeast of the Tamarancho Shear Zone, mélange exposed on the central island/peninsula within the reservoir is an OPS mélange, apparently one of several mélanges and olistolithic sandstoneshale units of the Novato Block. Two mélange layers in the north containing blocks similar to some in the Tamarancho Shear Zone are partially exposed along the southern peninsula of a small island that is transected by the Petaluma–Point Reyes Station Highway (dashed purple highway in the northern part of the map).

We found almost no exposures of the actual sheared rock of the fault zone at Nicasio Reservoir. Across the island/peninsula, neither sheared tectonic mélange matrix of the Tamarancho Shear Zone nor mélange matrix of the Novato Block mélanges are exposed, except very locally in areas along block margins (e.g., see the orientation of fabric on the margin of the greenstone block on the small northern island in Fig. 5). The block-margin exposures of matrix reveal shear-fracture fabrics, but fabric orientations are not reflective of the inferred overall fabric of the mélanges because they vary in orientation along the margins of the enclosed blocks, and some strikes are oriented at high angles to the trend of the zone. As is expected in mélanges, the orientations of bedding attitudes within blocks of the mélanges vary widely.

There are two promontories on the southwestern side of the island/peninsula within the Reservoir (Fig. 5). On the southern promontory, a chert exposure on a southwestern peninsular point that appears clearly on satellite imagery lies within the Tamarancho Shear Zone. The block exhibits shear planes (Fig. 11A), but their orientations are N35°W versus the approximately N51°W trend of the Tamarancho Shear Zone at this locality. The shear plane exhibits slickenlines and grooves (Figs. 11B and 11C) that plunge vertically, suggesting (late?) dipslip faulting, with the rocks to the southwest moved down relative to the chert and mélange to the northeast. These data from Nicasio Reservoir suggest a polyphase history of faulting.

On available local- to- regional maps (e.g., this report; Gluskoter, 1969; Wright, 1984), the Tamarancho Shear Zone is depicted as having a rather straight trace across the hilly topography, which suggests that it dips steeply. The variable but significant width of the zone, however, may conceal indications of a shallower dip if the fault had such a dip during an earlier phase of faulting.

Overall, the Tamarancho Shear Zone separates Novato Block Franciscan rocks from Mt. Tamalpais Block Franciscan rocks. A faulted tectonostratigraphic section of Franciscan rocks of the Mt. Tamalpais Block was recently described by Bero et al. (2020) and consists of a generally westward- and downward-younging sequence of accretionary units (Figs. 3 and 12). At Mt. Tamalpais, highangle faults mark many unit boundaries, which suggests post-accretion structural fragmentation of the section. Even though high-angle faults currently separate units that likely accreted along low-angle faults, the tectonostratigraphic units range in Y3Z maximum depositional age from 122 Ma at the top of the inferred stack to 86 Ma at the bottom. Most of the units are dismembered prehnitepumpellyite–facies metasandstone-metamudrock units. Preliminary work underway shows that some of these units may extend as far north as Nicasio Reservoir. The only highgrade rocks (blueschist, amphibolite, or eclogite facies) presently known within the Mt. Tamalpais Block constitute blocks that occur within mélanges (or landslides derived from mélanges).

Franciscan tectonostratigraphic sections of the Novato Block, as well as other blocks located to the northeast of the Tamarancho Shear Zone and the Petaluma Valley–Point Richmond–Silver Creek fault (which, within the northwestern San Francisco Bay Area bounds the Novato Block on the northeast), notably contain tectonostratigraphic units of blueschist-facies rock (e.g., called variously “Yolla Bolly Terrane,” “Angel Island Terrane,” “metamorphosed Franciscan rocks,” or schistose metawacke of Ebabias Creek; Fig. 12). Possibly equivalent rocks, exposed on adjoining Cenozoic fault blocks, have detrital zircon ages that suggest deposition of sedimentary precursors of some of the blueschist-facies rocks in late Early Cretaceous time, between 100 Ma and 110 Ma (Dumitru et al., 2018; Apen et al., 2021). Typically, the blueschist-facies rocks occur either below serpentinized peridotites of interpreted Coast Range Ophiolite or below local, thin accretionary units of serpentinite-matrix mélange that underlie the Coast Range Ophiolite (e.g., Berkland, 1969; Raymond and Bero, 2015; and see Bero, 2014, for a similar tectonostratigraphy on the Tiburon Block to the southeast). Note that the tectonostratigraphic sections of the Novato Block Franciscan Complex are incomplete, in that both above and beneath the blueschist-facies AUs, many of the accretionary units or rocks of their ages that are present on the Mt. Tamalpais Block are missing. Above the blueschist-facies rocks, units older than 110 Ma are missing (compare the Western Mt. Tamalpais section to the Novato area section in Fig. 12). Beneath the blueschist-facies rocks and typically beneath the underlying mélange, the youngest unit of the northwestern San Francisco Bay Area Franciscan Complex—the Novato Quarry AU, with a depositional age of 85 Ma (Y3Z)—is the next broken to dismembered tectonostratigraphic unit. Thus, clastic units between 100 Ma and 85 Ma MDA do not occur in the Novato Block sections. Clearly, the tectonostratigraphies of the Novato Block and the Mt. Tamalpais Block, which lie on opposite sides of the Tamarancho Shear Zone, are strikingly different in composition and architectural history.

Because a thrust slice of “Yolla Bolly Terrane” rocks depicted by Blake et al. (1974, 2000; also noted in the Maps of the Tamarancho Shear Zone and Bounding Units section above) lies west and northwest of Roy’s Redwoods, the structure associated with the trace of the Tamarancho Shear Zone in this region appears to be both complicated and inconsistent with the tectono stratigraphic differences between blocks noted here and depicted in Figure 12. The “Yolla Bolly Terrane” body of rock is shown by Blake et al. (1974, 1984, 2000), Wright (1982, 1984), and Prohoroff et al. (2012) at a location southwest of that at which we mapped the Tamarancho Shear Zone (Fig. 9). Wright (1982) implied that “Yolla Bolly” blueschist-facies rocks are associated with rocks masses located northeast of the Tamarancho Shear Zone by plotting the trace of the Tamarancho Shear Zone west of and curving around this “Yolla Bolly Terrane” mass. Thus, he included the mass in the northeastern Novato Block on his terrane map. If the Tamarancho Shear Zone is relatively straight, however, as appears to be the case along most of its strike length, this block would lie west of the Tamarancho Shear Zone and on the Mt. Tamalpais Block (Figs. 3 and 9).

Most of the proposed blueschist “Yolla Bolly” block is depicted in an area located on private, inaccessible land. On public land located west of Roy’s Redwoods, however, the rocks that we examined lying within the proposed boundary of this “Yolla Bolly” block do not support the existence of a body of foliated blueschist-facies rock west of the Tamarancho Shear Zone. The rocks here retain sedimentary bedding and textures in less massive outcrops (Figs. 13A and 14), rather than the strongly semischistose to schistose (“TZ-2”) fabric of Blake et al. (1967) that is typical of “Yolla Bolly” rocks. West of the proposed eastern “Yolla Bolly” contact, where strongly semischistose to schistose blueschist-facies rocks are depicted by Blake et al. (1974, 1984, 2000), Wright (1982, 1984), and Prohoroff et al. (2012), we instead found sandstones and metasandstones like those of the Kfnh unit (Figs. 13 and 14; and compare Figs. 8B to 13B and 8C to 13C). These largely unmetamorphosed to weakly metamorphosed sandstones lack the strong “TZ-2” foliation and instead contain relatively unflattened mudrock grains of sand to granule sizes (Figs. 13C and 14). Second, in continuous exposures that extend for more than 100 m along the fire road (dashed line in Fig. 4), extending northwest from Nicasio Valley Road, there is no clear evidence of the eastern boundary fault of the “Yolla Bolly” Terrane depicted in Blake et al. (2000). Here, the proposed fault should separate foliated “Yolla Bolly” rocks from Central “Terrane” mélange. Neither mélange nor foliated blueschists are present here, and no fault separates such rocks. Third, the proposed blueschist-facies “Yolla Bolly” body does not include any interbedded chert, a feature characteristic of the “Yolla Bolly Terrane” (Blake et al., 1984). Neither the maps of Blake et al. (1974, 1984, 2000) nor our mapping reveal any chert. Fourth, thin sections and photomicrographs of rocks from the area of the proposed Yolla Bolly Terrane reveal that the rocks lack the glauco-phane, lawsonite, and jadeitic pyroxene reported for that terrane (Fig. 14). The southwestern boundary of the proposed “Yolla Bolly” unit coincides approximately with the dismembered formation-mélange boundary mapped in this study. Hence, overall, the currently available evidence does not support the existence of a strongly foliated blueschist-facies rock body located west of the Tamarancho Shear Zone, at least in the area located west of Roy’s Redwoods.

In summary, the Tamarancho Shear Zone separates the southwestern and northeastern portions of the Franciscan (accretionary) Complex of the northwestern San Francisco Bay Area. The blocks on either side of the Tamarancho Shear Zone display distinctly different stacking sequences (Fig. 12; Berkland, 1969; Wright, 1984; Prohoroff et al., 2012; this report). As such, the Tamarancho Shear Zone is recognized as a major structure bounding major crustal blocks that are interpreted to have been formed by late Cenozoic faulting and a structure that may have played a role in dismemberment of the original Franciscan accretionary complex. Later faulting on the Tamarancho Shear Zone clearly juxtaposed blocks with Franciscan rocks formed in two different parts of the accretionary complex.

Constraints on the timing and amounts of dextral slip on the Tamarancho Shear Zone are limited and general. In terms of timing, none of the available maps showing the Tamarancho Shear Zone depict it as cutting any Cenozoicage rocks. It is overlain south of the Novato area by Quaternary alluvial sediments, which sets a young (or lower) age limit of ~2.6 m.y. on movement (Walker et al., 2018). The Tamarancho Shear Zone is shown by Berkland (1969) to cut the Novato Quarry AU, which has a proposed Y3Z MDA of 85 Ma, i.e., middle Late Cretaceous. This sets the upper age limit.

The Tamarancho Shear Zone has a general trend of ~N50°W. While this trend is at an angle to the trend of the San Andreas master strike-slip fault located in the western part of the northwestern San Francisco Bay Area (Fig. 3), the trend is within the 40–60° trend-range of a group of modern zones of seismicity and late Cenozoic San Andreas System (high-angle, strike-slip) faults mapped in the broader region (e.g., Berkland, 1969; Wong, 1991; Wagner et al., 2011; McLaughlin et al., 2012; Bero, 2014). Thus, we conclude that the TRZ is likely a late Cenozoic fault that is a member of that group.

The nature of the Tamarancho Shear Zone, as we have described it, is that of a high-angle (vertical?) fault zone that may or may not have had an earlier history involving faulting that produced a fault of lower dip. Late faulting of dip-slip character, at least locally, occurred on vertical fault planes. To date, we have not discovered slickenlines on fault planes that reflect strike-slip movements. In addition, inasmuch as we have not discovered, to date, any “piercing points” on opposing sides of the Tamarancho Shear Zone that could be used determine the amount of slip, the total amount of movement of the rocks west of the Tamarancho Shear Zone relative to rocks to the east remains unknown. In fact, displacement amounts on faults bounding the Marin Headlands, Nicasio Reservoir, Mt. Tamalpais, and Tiburon blocks, like those on the Novato Block, are all largely unconstrained at the time of this writing.

While determination of the exact separation on the Tamarancho Shear Zone between the Novato and Mt. Tamalpais blocks is not possible, determination of some minimal amounts of displacement or separation for all rocks west of the Petaluma Valley–Point Richmond–Silver Creek fault is possible. Using the best radiometric- and detailed mapping-based data from works such as Sarna-Wojcicki (1992), Wakabayashi (1999), Langenheim et al. (2010), and Raymond (2018) for San Francisco Bay Area blocks east of the Petaluma Valley–Point Richmond–Silver Creek fault, we estimate that the Novato–Mt. Tamalpais blocks moved north on San Andreas System faults from their point of formation in the Mesozoic subduction accretionary complex by a minimum of 125 km and may have been displaced by as much as 200 km. We expand on this in the Paleogeographic Reconstruction and Accretionary History section below.

To attempt to reconstruct even a generalized original, pre–San Andreas faulting paleogeography and accretionary architecture of the northwestern San Francisco Bay Area Franciscan Complex, it is essential to remove late Cenozoic strike-slip separations among blocks. By doing so, the various blocks can be restored to pre-faulting positions relative to one another. Such a reconstruction, however, cannot be constrained like a balanced cross section at this point in time because few “pin points” or piercing points exist in the San Francisco Bay Area Coast Ranges that would allow a link to be recognized between one block and another to establish pre–late Cenozoic faulting positions. Thus, the lack of limiting data makes reconstruction an exercise in the application of a set of preferred choices of block movement. Block movements here are defined as movements along major faults. Preferred block movements—hence, displacements—are based on a few data points, some estimates of offsets on faults, and general considerations. Multiple alternative configurations exist (e.g., compare Wakabayashi, 1999; Raymond, 2018; and Apen et al., 2021).

Several studies attempt to address the amount of separation that has occurred on one or more of the major San Andreas System faults of the San Francisco Bay Area and adjoining areas (e.g., McLaughlin et al., 1988; Sarna-Wojcicki, 1992; Powell, 1993; McLaughlin et al., 1996; Wakabayashi, 1999; Graymer et al., 2002; Langenheim et al., 2010; Wagner et al., 2011; McLaughlin et al., 2012; Chapman et al., 2016; Raymond, 2018). The best data for such analyses are data that include a combination of lithology, lithostratigraphy, and radiometric dating of Cenozoic volcanic rock bodies on opposing sides of faults (e.g., Fox et al., 1985; Sarna-Wojcicki, 1992; McLaughlin et al., 1996; and see the review in Table 1 of Wakabayashi, 1999). As is the case with other types of pin points for Coast Range rocks, there are few compelling correlations of this kind. One apparent correlation is that of the “Tolay Volcanics” with the “Berkeley Hills Volcanics” on opposing sides of the Rodgers Creek–Hayward fault (Fig. 2), which suggests that the Petaluma–East Shore Block has moved north on the order of ~49 ± 12 km (37–61 km) along the fault (Fox et al., 1985; Wakabayashi, 1999; and see McLaughlin et al., 2012). The Rodgers Creek–Hayward fault and all of the major block-bounding faults are dominantly right-lateral, strike-slip faults, so such a movement suggests that the Novato and Mt. Tamalpais blocks also moved north by at least this amount. The use of map-based separations from studies to the east reveals an additional minimum northward movement of 76 km for a total of 125 km.

In many analyses of central California strike-slip fault movements, the arguments used to constrain separations between adjoining blocks are based on anecdotal, debatable correlations between rock masses. These correlations are subject to alternative interpretations. The result is that the paleogeo-graphic reconstructions presented below and discussed here are generalized and just two of several possible reconstructions. We use these generalized reconstructions of the paleogeography as a basis for a presentation of the Franciscan Complex architecture at a time before San Andreas System faulting occurred. Our new reconstructions replace existing published architectures.

Several studies over the past four decades have presented regional structural and architectural interpretations of the Franciscan Complex of the northwestern San Francisco Bay Area. In general, the interpretations—primarily by Blake et al. (1984, 2000), Prohoroff et al. (2012), and Wakabayashi (1990, 1992, 2015)—are depicted in map or cross-section format, or both. Notably, in some cases these interpretations included reconstructions of relationships among the various AUs, which at the time were called terranes, nappes, or both. These previous interpretations, particularly those developed before the year 2000, were based on earlier understandings of subduction accretionary complex architecture and were developed prior to the use of detrital zircon U-Pb dating that has proved to be a powerful tool for understanding temporal relationships and the provenance of units within the complexes. All of these depictions, of course, present the architecture and relationships among units largely as they presently exist, and in doing so, several suggest a continuity that does not exist across major strike-slip faults separating several major modern crustal blocks (depicted in Fig. 2). None presents a restored section that accounts for offsets on the Tamarancho Shear Zone and other strike-slip faults.

Figure 15 presents three cross sections across the northwestern San Francisco Bay Area and reveals contrasts among architectural models—our current working model and two earlier constructs. The architecture of Blake et al. (2000; Fig. 15A) consists of a set of fault blocks bounded by high-angle dipslip and strike-slip faults. This architectural model neither addresses the major differences between the blocks now juxtaposed across the Tamarancho Shear Zone nor depicts a large mélange matrix enveloping various “terranes” as described in their text. The cross section presented by Prohoroff et al. (2012; Fig. 15B) depicts the Tamarancho Shear Zone (unnamed on their cross section) as separating a set of three western mélanges, which are equivalent to rocks of our Bolinas Ridge AU stack on the Mt. Tamalpais Block (Fig. 3), from rocks of the Novato Block, including those of three eastern mélanges and the Novato Quarry unit (Kfnq in Figs. 3, 5, and 15). Clearly, in their cross section, the Tamarancho Shear Zone separates blocks of different tectonostratigraphy, a fact that is emphasized by our cross section of the region (Fig. 15C) and by the tectonostratigraphic stacks shown in Figure 12. The crustal blocks of Prohoroff et al. (2012) display folds and faults formed largely during Neogene time that affected thin rock layers that are presumably accreted layers of the accretionary complex. In spite of the presence of the Tamarancho Shear Zone, in showing similar folding in rocks on both sides of the fault and depicting all mélanges (east and west) in the same color, the Prohoroff et al. (2012) cross section seems to suggest some continuity across the two juxtaposed blocks, just as did an earlier presentation of the architecture by Blake et al. (1984). The architectural presentations of Wakabayashi (1990, 1992, 1999, 2015) (not shown in Fig. 15) changed over time in terms of the location of cross-section lines on maps and in unit thicknesses within the cross sections. The 2015 presentation of Wakabayashi includes two cross sections of the northern to central San Francisco Bay Area. One is drawn along a cross-section line across the northwestern San Francisco Bay Area and is a modification of the cross section of Prohoroff et al. (2012). In this revised architectural interpretation, a number of changes are made, including depiction of Novato Quarry rocks west of the Tamarancho Shear Zone, which suggests a continuity of units across the Tamarancho Shear Zone that minimizes the importance of the fault. The second cross section of Wakabayashi (2015) is a three-part composite section that crosses parts of seven separate blocks and likewise suggests a continuity of units across various faults.

None of the earlier interpretations presents the rocks or the tectonostratigraphic stacks of the Novato and Mt. Tamalpais blocks (or those of the Tiburon, Bay, and Petaluma-East Shore blocks) as markedly different (as we have shown in Figs. 12 and 15), and hence as having been formed in, and derived from, different parts of the accretionary complex. In addition, no major separations are suggested on the San Andreas System fault segments in the Bay Area. Yet, work on faulting in the San Francisco Bay Area clearly shows that different blocks have moved northwestward by different and significant amounts from their points of origin in the primary accretionary complex to the south (e.g., Wakabayashi, 1999; Raymond, 2018).

To reconstruct the subduction accretionary complex geometry and paleogeography for pre-San Andreas System time, the significance of the metamorphic differences between the Novato and Mt. Tamalpais blocks, and the constraints imposed by similarities or differences in depositional ages and the provenance of sediment for sedimentary protoliths of the various units, must be acknowledged. Recall that blocks located northeast of the Tamarancho Shear Zone are characterized by at least one blueschist-facies accretionary unit (e.g., Berkland, 1969; Blake et al., 2000; Bero, 2014). These jadeitic pyroxene-bearing rocks have been subducted to depths of more than 23 km (based on analyses of similar rocks; e.g., Patrick and Day, 1989; Ernst, 1993; Ernst and McLaughlin, 2012; Raymond, 2018). The blueschist-facies, rock-bearing tectonostratigraphies of the Novato Block, as well as those of other northeastern blocks, have similarities with tectonostratigraphic stacks in regions of the central–southern Diablo Range, notably the southern Diablo Range areas of Panoche Road–Panoche Pass and nearby Ortigalita Peak, which are characterized by significant blueschist-facies units (e.g., Briggs, 1953; Ernst, 1965, 1971; Echeverria, 1980; Wakabayashi and Dumitru, 2007). Hence, some Franciscan rocks of the blocks located northeast of both the Tamarancho Shear Zone and the Mt. Tamalpais Block, including the Novato and Tiburon blocks, were deeply subducted, accreted, and perhaps underplated like rocks in the central–southern (as well as northern) Diablo Range (cf. Ernst, 1975; Kimura et al., 1996; and see below).

Blocks located southwest of the Tamarancho Shear Zone, in contrast to the northeastern blocks, contain Franciscan Complex rocks that have experienced no more than minor prehnite-pumpellyite–facies metamorphism (e.g., Wahrhaftig, 1984; Swanson and Schiffman, 1979; Meneghini and Moore, 2007; Bero et al., 2020). These rocks were subducted to depths of less than 13 km (depths are based on pressures; see Schiffman and Day, 1999; Wakabayashi, 2015; Raymond, 2018; and see review in Raymond, 2007, Chapter 22).

Some AUs in the Mt. Tamalpais Block and other blocks located to the southwest of the Tamarancho Shear Zone were deposited during the same general late Early Cretaceous time period (100–115 Ma) as the blueschistfacies AUs of the Tiburon Block and perhaps the Novato Block, both of which are located to the northeast (Bero et al., 2020, 2021; Apen et al., 2021). Kernal density estimate (KDE) plots (Fig. 16) and analyses of the detrital zircons for some rocks of the Tiburon, Novato, and Mt. Tamalpais blocks and rocks of the southern Diablo Range show MDAs of less than 115 Ma and dominant zircon populations of 115 Ma to 140 Ma, which reflects the deposition of sediments in Early Cretaceous time with primary sources with rock ages in the 115–130 Ma time range. Secondary peaks on some KDE diagrams for both the Tiburon and Mt. Tamalpais blocks occur at ca. 140 Ma, which supports an interpretation that rocks of both blocks received sediment from the same general provenance. Some similarities also exist among these Franciscan-sediment KDE profiles and those of forearc Great Valley Group sediments (e.g., the 104 Ma MDA Tiburon Ridge TB-4 profile of Apen et al., 2021, compared to that of the 97 Ma MDA Grabast unit of DeGraaff-Surpless et al., 2002; Fig. 16). The ages are consistent with a Sierra Nevada sedimentary source for the southern Great Valley forearc and the Franciscan accretionary complex sedimentary protoliths of the southern Diablo Range, the Tiburon Block blueschist-facies rocks, and some Mt. Tamalpais prehnite-pumpellyite–facies metasedimentary rocks (e.g., DeGraaff-Surpless et al., 2002; Dumitru et al., 2015; Bero et al., 2020, 2021; Apen et al., 2021).

Clearly, although the rocks located northeast and southwest of the Tamarancho Shear Zone have similar ages and sediment provenances, there is a sharp contrast and major differences in subduction and post-subduction history across the Tamarancho Shear Zone. Those differences are revealed by metamorphic history and stacking sequences. The rocks composing the Mt. Tamalpais Block (to the southwest) were subducted to shallow depths of structural burial, whereas some of those to the northeast on the Novato and Tiburon blocks were subducted to greater depths. In addition, the absence of AUs with MDAs of 120–110 Ma and 92–100 Ma on the Novato and Tiburon blocks, combined with the occurrence of blueschist-facies rocks structurally overlying lower-grade, weakly metamorphosed rocks, suggest a history of post-accretion, out-of-sequence faulting to bring the blueschist-facies rocks up to shallower levels of the subduction accretionary complex. The out-of-sequence faulting likely occurred before the Novato and Tiburon blocks were juxtaposed against the Mt. Tamalpais Block, either during or later in the accretionary phase of subduction-driven, accretionary complex formation, inasmuch as blueschist-facies units are missing from the eastern Mt. Tamalpais Block.

Given the similarity of age and provenance of the sediments and differences in metamorphism, there are at least two possible structural configurations to explain the differences in subduction history (Fig. 17). In either case, there was a blanket of sediment washed from the Sierra Nevadan source and spread across a large expanse of forearc, trench, and incoming plate. One structural configuration that provides an explanation for the differences in history involves a transform fault separating northern and southern segments of the subducting plate (the faulted plate model in Fig. 17A). In this structural model, the northern plate segment would have had a steeper and deeper subduction configuration than the more southerly plate segment. Clearly, development of such a contrasting but simultaneous configuration of subduction is enigmatic. In this faulted plate model, the 100–115 Ma sediment on the northern plate segment (precursor to Novato and Tiburon blocks and blueschist-facies Franciscan rocks) would be subducted in part or completely to depths greater than 23 km. Rocks and sediment on the more southerly plate segment (precursor to Mt. Tamalpais Block Franciscan rocks), which was undergoing relatively flat, shallow subduction, would accrete at depths of less than 13 km, yielding a lower-grade metamorphic imprint. Post-accretion, out-of-sequence faulting would bring blueschist-facies rocks from a zone of deep accretion or under-plating up to a high level in the northern area of the subduction accretionary complex of this region (Fig. 18C). Subsequently, during late Cenozoic time, the blueschist-bearing rocks and others from the Novato and Tiburon Blocks would be geographically juxtaposed with the prehnite-pumpellyite–bearing rocks of the Mt. Tamalpais Block by faulting on the Tamarancho Shear Zone.

The second structural configuration that could explain the development of contrasting blueschist-facies and prehnite-pumpellyite–facies rocks of the same age and provenance involves the poorly understood possibility of a temporary dual subduction zone (Fig. 17B). Such zones exist in a variety of configurations, none of which matches exactly what would be required to explain the Franciscan rocks discussed here (e.g., Burg et al., 2006; Mishin et al., 2008; Jagoutz et al., 2015; Ji et al., 2017; Hasegawa and Nakajima, 2017; Pusok and Stegman, 2019; Király et al., 2021). One example with some similarities was suggested by Moores and Twiss (1995; Fig. 9.8C), who modified the earlier work of Bird (1978) on the Zagros Mountains, to depict, as do we, a relatively shallow, dual subduction-like geometry with an outer splay of a subduction-zone fault complex as the leading edge of an imbricate stack of AU-like masses. The inboard boundary of the imbricate stack is a primary subduction boundary. In depicting a similar geometry to the one we envision (but not necessarily one with simultaneous subduction, or one in a non-collisional setting), Moores and Twiss (1995) provide a partial analogue for the dual subduction we suggest as a possibility for the accretion of the northwestern San Francisco Bay Area rocks.

In a dual subduction model, as was the case for the faulted plate model, sediments of the same general age would be deposited in, and would extend from, the trench for a substantial distance onto the incoming plate. For the Franciscan Complex in the southern Diablo Range region, we envision a temporary and localized process involving dual subduction zones, with the outer transient zone resulting from a relatively localized plate weakness. As subduction and accretion progressed, inboard parts of the sediment sheet would be subducted to greater depths than outboard parts, which simultaneously would be subducted and accreted at shallower depths. The inboard rocks/sediment of the future Novato and Tiburon blocks would be subducted more deeply, accreted at greater depths, and metamorphosed to blueschistfacies grade (Fig. 18A). The similarly aged sediment/rock of the future Mt. Tamalpais Block, however, would experience shallow subduction and accretion from the lower plate of the dual subduction zone. The resulting rocks would accrete and be metamorphosed in a position that would be outboard with respect to the dual subduction-fault interface (Figs. 18A and 18B). As in the case of the faulted plate model, later out-of-sequence faulting would bring blueschist-facies rocks up to shallower levels of the accretionary complex (Fig. 18C). Lesser amounts of late Cenozoic strike-slip faulting (and trench-perpendicular shortening) would be required to juxtapose Novato-Tiburon Block rocks with Mt. Tamalpais rocks in the dual subduction model.

If dual subduction occurred, it was apparently localized because the currently available evidence does not reveal outboard rocks like those of the Mt. Tamalpais Block opposite the northern Novato Block, and it is unclear whether Novato Block–Tiburon Block tectonostratigraphies extend to the south, where they would occur inboard of the Franciscan rocks of the Bay Block. Whatever the case may be to the south, there appears to have been differences in accretionary history along strike or across strike in the subduction zone.

Note that shallow subduction yields a wide accretionary zone. In both the faulted plate and dual subduction models, significant shortening across the accretionary complex would be required to arrive at the modern configuration in which rocks of the Novato and Tiburon blocks lie in close proximity to rocks of the Mt. Tamalpais Block (e.g., Sample and Fisher, 1986). Current data do not allow us to define the amount or details of shortening in the paleogeographic reconstruction (Fig. 19).

We reconstructed two general paleogeographies shown in Figure 19 using both the few compelling correlations available and the generalized estimates of block separations for most block movements, to move blocks in retrograde fashion back to the south. As noted, most of these movements are not constrained by pin points. In part, the estimates are based on those of Wakabayashi (1999), but we make modifications based on our own studies and those of Fox (1983), Fox et al. (1985), Sarna-Wojcicki (1992), Langenheim et al. (2010), and McLaughlin et al. (2012). Our result for the dual subduction model (Fig. 19A) shows various San Francisco Bay Area rock masses in broadly similar positions as those attained in the previous studies of Wakabayashi (1999), Raymond (2018), and Apen et al. (2021), in that blocks such as the Mt. Tamalpais and Marin Headlands blocks are moved south on the order of 160–220 km. A general principle we used for block movements is that more westerly blocks generally were displaced farther north during faulting (and thus are moved farther south in the reconstruction) than were more easterly blocks (cf. Wakabayashi, 1999). This principle was used to move the Mt. Tamalpais Block farther south than the Novato and Tiburon blocks along the Tamarancho Shear Zone and Richmond Bay faults. In the faulted plate model, to completely separate the Mt. Tamalpais Block from the Tiburon and Novato blocks (so that the former is not outboard of the latter), the Mt. Tamalpais Block and those associated with it (the Nicasio Reservoir and Marin Headlands blocks) must be moved a minimum of 255 km south of their present positions (Fig. 19B).

We emphasize that the paleogeographic reconstructions are generalized. While the retrograde movements allowed us, as was the case in previous studies, to relocate northwestern San Francisco Bay Area Franciscan rocks at the time of accretion to new locations west of the central to extreme southern Diablo Range (Figs. 19A and 19B), the lack of critical information—such as detrital zircon dates for MDAs for all sedimentary rock masses, limited knowledge of the details of the geology of the various AUs of the San Francisco Bay Area, and the lack of lithostratigraphic correlations of units dissected and offset along the major faults—currently prohibits a more detailed and accurate paleogeographic rendering of the setting and architecture. One important conclusion of this paleogeographic reconstruction, however, is that the blocks and the rocks therein, within the accretionary complex, occupy different positions laterally, relative to one another, than is currently the case.

During late Cenozoic deformation of the Coast Ranges, in addition to being moved along strike-slip faults, each of the Cenozoicaged, fault-bounded blocks experienced shortening due to thrust faulting and folding (e.g., Page et al., 1998; McLaughlin et al., 2012; Raymond and Bero, 2015; Raymond, 2019; Bero et al., 2020). In most cases, studies have not revealed the amounts of shortening from these events, but the few studies available, and our work in the Mt. Tamalpais Block, suggest that a minimum of 20–30% of shortening occurred (Page et al., 1998; Raymond and Bero, 2015; Raymond, 2019). This means that the accretionary complex located west of the central to southernmost Diablo Range must be widened by this minimum amount to attain its original configuration. Substantially greater shortening likely occurred during accretion.

On the paleogeographic base maps of Figures 19A and 19B, we have constructed generalized sketch maps of the positions of northwestern San Francisco Bay Area crustal blocks within the primary accretionary complex as it might have existed during development southwest of the modern Diablo Range and before the San Andreas System faults formed (Fig. 19). Later juxtaposition of the Novato and Mt. Tamalpais blocks along the Tamarancho Shear Zone, after accretion, was simultaneous with or followed the northward transport of all of the blocks of the southwestern Diablo Range region to the northwestern San Francisco Bay Area. The Tamarancho Shear Zone became inactive, but blocks of basement rock comprising the future northwestern San Francisco Bay Area AUs were moved north to their present locations and shuffled in position relative to one another, in part by movements associated with faulting on such faults as the Rodgers Creek–Hayward fault, which remains active.

Different tectonostratigraphic stacking sequences in Franciscan Complex rocks of the northwestern San Francisco Bay Area require that the Franciscan accretionary complex rocks within Cenozoic fault blocks of the area be formed in different settings. The structural history of the rocks was complex. We envision a localized, transient dual subduction zone or a faulted subducting plate west of the modern southern Diablo Range during Mesozoic accretion to explain the contrasting tectonostratigraphies within northwestern San Francisco Bay Area fault blocks that formed by subsequent late Cenozoic faulting. In addition, out-of-sequence faulting of blueschist-facies rocks, juxtaposition of blocks of rock of different histories, and northward transport of accretionary complex rocks were all a part of the tectonic history of the northwestern San Francisco Bay Area.

Faulting on Neogene strike-slip faults both dissected the primary Franciscan accretionary architecture and later juxtaposed Franciscan Complex fragments from different parts of the accretionary complex. Clearly, the Tamarancho Shear Zone is a major structure along which fragments of the accretionary complex containing differing tectonostratigraphies were juxtaposed after the major phase of Franciscan accretion in Late Cretaceous time. It likely was an early splay of a San Andreas System fault. Major splays such as the Petaluma Valley–Point Richmond–Silver Creek fault and the Rodgers Creek–Hayward fault dissected the primary accretionary complex architecture and transported crustal block fragments northward to the northwestern San Francisco Bay Area, where they were assembled to form the modern geologic architecture.

Both detailed mapping, as presented here, and regional analyses are needed to broadly characterize the Franciscan accretionary complex. This is true with regard to many subduction accretionary complexes. Paleogeographic reconstruction is an essential element in the process of revealing the primary accretionary architecture, an architecture that is not revealed by either individual or composite cross sections of modern architectures in polydeformed regions (like the northwestern San Francisco Bay Area). Maps such as those presented in Figure 19 and schematic cross sections like those presented in Figure 18 depict a primary post-accretion, pre–San Andreas regional architecture for the Franciscan Complex of the northwestern San Francisco Bay Area. Maps and cross sections of presentday geology, like those presented in Figures 3, 4, 5, and 15, which reveal the details of some modern local Franciscan accretionary complex architecture, are keys to the past and are necessary for understanding accretionary complex architecture. Similar presentations of accretionary complex architecture are needed from other regions beyond the northwestern San Francisco Bay Area for a broader understanding of the architecture and history of the long north-to-south Franciscan Complex.

The wider significance of this work is that it suggests that understanding the architecture and history of subduction accretionary complexes requires application of a wide range of geological tools to unravel the details of local-to-regional structure. Distinguishing mélanges within the subduction accretionary complex from zones of strike-slip faulting of similar character that cut the accretionary complex requires detailed mapping. Understanding sedimentation histories and depositional settings requires the application of detrital zircon U-Pb dating and sedimentary facies analyses. Removing the effects of post-accretion deformation by paleogeographic reconstruction and the unfolding of folded rocks—facilitated by metamorphic, detrital zircon, and structural studies—allows for a much clearer picture of accretionary complex structure that can be compared to images of modern, active subduction accretionary complexes. Here, we add to the studies of ancient and modern subduction accretionary complexes that together reveal the diversity of structures and histories developed at subduction zone, accretionary plate margins.

Sonoma State University’s Department of Geology provided laboratory facilities. John Wakabayashi and Russell Graymer engaged in discussions with one of us (L. Raymond) about various aspects of local-to-regional geology specific to this study. We thank Darrel Cowan, Terry Pavlis, and John Wakabayashi, who provided valuable reviews and suggestions for improvement of the manuscript.