Modern physiography suggests an ancient subduction zone
At the latitude of San Francisco, California has three main physiographic provinces. From the Pacific coast on the west to more inland positions in the east, these are: the Coast Ranges, the Great Central Valley, and the Sierra Nevada. In spite of their subsequent geologic histories, these are all remnants of California’s former history as a convergent margin.
Today, the region is famous as a transform boundary, where the Pacific Plate and the North American Plate slide laterally past one another along the San Andreas Fault.
But the modern plate boundary in Califronia is not the same as the ancient plate boundary. During the Mesozoic, there was a third plate offshore: the Farallon Plate, which was in between the Pacific and the North American Plates. The relative motion between the Farallon and North America was quite different. Back then, coastal California was a convergent plate boundary, and the Farallon was subducting beneath western North America.
The boundary between the Pacific Plate and the Farallon Plate was an oceanic ridge, the site of seafloor spreading. In other words, it was a divergent plate boundary between two plates of oceanic lithosphere. The newly-minted Farallon traveled a short distance eastward, and then subducted beneath continental lithosphere of the North American Plate, producing key features of a convergent plate boundary: the accretionary wedge (focus of this case study) but also partial melting of the overlying mantle wedge. This melting was triggered by the release of water from the subducted plate. The resulting magma rose to produce a continental magmatic (and volcanic) arc. Weathering and erosion of the arc’s mountains produced sediment that was transported downhill, and accumulated in both forearc and back-arc basins.
The rocks of the Coast Ranges preserve a record of their past. Let’s take a look at both rock types and structures. To make the story somewhat simpler, geologists divide the rocks of the region up into packages of rocks that have seen relatively similar stories of formation, deformation, metamorphism, and exhumation. We call these packages terranes. On the peninsula of San Francisco itself, east of the modern plate boundary (i.e., the San Andreas Fault), there are three distinct terranes, separated by two map-scale bodies of sheared-out rocks called mélange:
Note the difference in orientation of the terrane contacts (each of them a manifestation of one instance of the plate boundary between North America and the Farallon Plate) and the modern plate boundary. In the Mesozoic, the plate boundary was a subduction zone dipping to the east under North America, and that shape is preserved as powerful evidence of the fundamental geometry of that boundary. The modern transform plate boundary, in contrast, is vertical.
Further to the north and east, there are more terranes still. Each terrane is a sliver-shaped package of rocks that was subducted (some rather shallowly, some quite deeply) and then transferred from the subducting Farallon Plate to the overriding North American plate.
Rocks & structures associated with the subduction of oceanic crust
The original seafloor rock types that were fed into the subduction zone included:
- Mantle peridotite
- Oceanic crustal rocks: Basalt and gabbro
- Deep sea sediments: chert + shale
- Turbidites: shale + graywacke
- Continental margin sediments: Sandstone + conglomerate
Once subducted, these rocks metamorphosed to various degrees, making new rocks:
- Greenstone (greenschist where foliated)
Besides metamorphosing, being subducted wasn’t easy on these rocks. They accrued several kinds of deformation as a result of the tectonic stresses they experienced:
- Chevron folds
- Thrust faults
Let’s examine the metamorphic transformations in turn: first the oceanic lithosphere itself, and then the sedimentary cover. We’ll look at the protolith, and then at the metamorphic rock (or rocks) it can become. Finally, we’ll examine the deformational structures that record evidence of the stresses of subduction.
The hard rocks: mantle and crust
MANTLE: Peridotite to serpentinite
The upper mantle is made of a rock called peridotite. Peridotite is dominated by the mineral olivine, though some peridotites also have plenty of pyroxene and plagioclase in them. Peridotite formed long ago when the early Earth cooled off and solidified in its uppermost reaches. Thought it is solid, the mantle still flows slowly. Warm mantle peridotite rises due to its low density, while cold peridotite sinks due to its higher density. This relative motion sets up convection related “cells” in the mantle that are thought to be one of the important driving forces that power the surficial movement of lithospheric plates, including the manifestation of convergent motion that we call “subduction.”
Not all of the lithospheric mantle necessarily gets recycled during subduction. The process of “underplating” can cause slivers of the subducting plate (including lithospheric mantle) to be ripped off and added to the underside of the overriding plate. This ‘subduction accretion’ builds up the accretionary wedge from below, and is the major source of mantle rock introduced into the subduction complex.
When the peridotite is broken off and mixed into the accretionary wedge, it can come into contact with water, and metamorphose. This creates serpentinite, a metamorphic rock dominated by the mineral serpentine. Serpentine is a hydrated magnesium silicate. Serpentinite is the California ‘state rock,’ though because some of the minerals in serpentinite (such as chrysotile) have an asbestiform habit, it makes lawyers itchy, and there are periodic proposals to replace its official ‘state rock’ legal status.
OCEANIC CRUST: Basalt & its three metamorphic descendants
When the hot mantle is decompressed, it is capable of partial melting, which derives a mafic magma form the ultramafic source rock. Estimates are that mafic magmas derive from somewhere between 1% and 25% partial melting of source peridotite.
The resulting magma could cool at depth to generate a gabbro, or could be erupted on the seafloor to make basalt. When lava erupts at the bottom of the ocean, the high heat capacity of the surrounding seawater wicks heat away efficiently. This chills the edge of the lava instantly, forming a solid crust resembling a loaf of bread.
The interior of the lava “loaf” is still molten, and if it is under sufficiently high fluid pressure, it will crack open the crust and squirt another lobe of fluid lava out into the water, repeating the process. Played out time and time again over geological time, this results in a pile of “pillow” shaped features. These “pillow lavas” are a therefore a primary volcanic structure; They are the signatures of seafloor eruptions of new oceanic crust.
When basalts are subducted, they serve as particularly good recorders of the pressure and temperature they encounter. If they are shallowly subducted, they will make greenstone. If they are subducted a bit deeper still (with high pressure), they will transform into blueschist, and if they are subducted really deep (with the highest pressure), eclogite will form.
Greenstone is the lowest grade of subduction related metamorphism. Very few of the basalts seen today in the Franciscan complex are still basalt. Many/most have been metamorphosed to make greenstone. When the minerals that make up basalt are subjected to depths around 20 to ~35 km in a subduction zone, the pressure and temperature are sufficient to make the plagioclase and pyroxene react and generate the minerals chlorite and actinolite and sometimes (if water is present), epidote. The jet black basalt is now turned a pale green in color, and this is quite a common sight in the California Coast Ranges. If tectonic shear stresses are low, large-scale primary features such as lava pillows may be preserved, despite the metamorphic recrystallization that has reorganized the rock on smaller scales.
Blueschist is the rock that results when basaltic protoliths are driven deeper into the subduction zone, to perhaps ~35 to 50 km depth. The rate of subduction is greater than the rate of thermal equilibration; as a result the subducted basalt soon finds itself at conditions where the pressure is quite high, but the temperature is still relatively cold. These are the conditions at which the minerals lawsonite and glaucophane form. Metabasalts in the Angel Island Terrane have seen these conditions, and where basalt was present, it has been recrystallized to make blueschist.
It’s important to note that blueschist cannot form from “just any rock.” It must have basalt (or greenstone, which is of course made of the same blend of elements) as its protolith. If a sandstone is subducted to the same depths and temperatures, it cannot make a blueschist, as it doesn’t have the right starting ingredients. Such a sandstone is said to have been metamorphosed “at blueschist facies,” but it is neither blue nor a schist. We can still find crystals of lawsonite in it if we examine it in thin section (under a petrographic microscope), but to the naked eye, it frequently looks just like a normal sedimentary sandstone!
Eclogite represents the highest set of temperatures and pressures; the protolith basalt’s atomic ingredients reorganize to form distinctive “Christmas tree” colored minerals: the rich green of omphacite (a kind of pyroxene), and the cranberry-red of garnet (the variety called pyrope). Eclogites are astonishingly dense, which is perhaps not surprising considering they form to be stable at pressures corresponding to depths of at least ~50 km!
Eclogite is very dense. Depending on the location sampled, the density of eclogite ranges from 3.45 to 3.75 g/cm3. Recall that the mantle’s peridotite clocks in at around 3.3 g/cm3, so this means that once eclogite forms, it can act as a dead weight, pulling on the rest of the subducted plate. This downward pull is thought to be a primary driver of “slab pull,” a force partially responsible for the motions of lithospheric plates.
The sedimentary journey
On top of the oceanic crust is where sediments can be deposited.
ABYSSAL DEPTHS: Chert & shale
In the offshore realm, far out in the Pacific Ocean (but atop the Farallon Plate), deep sea sediments of chert and clay shale accumulated. The source of the material for the chert was the dead bodies (“tests”) of radiolaria, which rained down through kilometers of sea water as “marine snow.” Unlike calcite, which tends to dissolve in the deep sea below the carbonate compensation depth, the silica that makes up the radiolarian skeletal material tends to be stable.
We don’t fully understand the alternation between chert and shale that characterizes the Marin Headlands Terrane, but it may be due to annual or seasonal cycles, or Milankovitch cycles in Earth’s orbit. Some workers have suggested it’s a diagenetic separation of the two sediment types: that is to say, the radiolaria skeletal remains and the clay are deposited at the same time, and only later separated once buried, by diagenetic processes.
Because chert is made of silica, it is both chemically stable and rather hard, so it stands up well to weathering and erosion. As a result, chert can be found underlying some of the highest landforms in the Coast Ranges.
TURBIDITES: sand & mud via underwater avalanches
Closer to shore, the weathering and erosion of the continent brings a lot of sediment into the ocean. Along the continental slope, turbidity currents deliver slurries of sand and mud in suspension into the deep water. Though they come from surface waters, they are dense due to all their entrained sediment, and they tend to hug the bottom as they advance. Occasionally, they will travel along submarine channels, and spread out into complexes of abyssal fans. As these turbulent currents slow down, they drop their sedimentary load, with larger (heavier) particles dropping out first, and finer (lighter-weight) particles staying suspended in the water column for longer, eventually dropping out last. The resulting graded beds are signatures of this mode of deposition. We call the overall packages of graywacke sandstone and black shale deposited by turbidity currents “turbidites.”
NEARSHORE FACIES: River-borne sand & gravel
Closer to the coast, and in shallower waters of the continental shelf, rivers debouching into the sea drop deltaic deposits. For the most part, this is sand and mud, but in many places pebbles and cobbles (gravel) make it all the way to the shore too. These sedimentary deposits result in rocks such as shale (former mud), sandstone (former sand), and conglomerate (former gravel). Sometimes plant fragments are preserved in these rocks as little scraps of coal.
The ideal rock sequence for the rocks making up the subducted seafloor is therefore something like this:
METASEDIMENTS: schist and metaconglomerate
When they metamorphose, these sedimentary rocks transform into phyllite, schist or metaconglomerate. Mudrocks like shale can recrystallize to make shiny, foliated phyllite, which will turn to schist with higher grades of metamorphism. Sandstones turn into “metasandstone,” which looks much like sandstone does (but contains key metamorphic minerals when examined in thin section). Some metasandstones can also develop a schistocity (scaly foliation, as seen in the example here).
Conglomerates made of a variety of clasts can demonstrate an interesting response to shearing stresses: the weaker rock types will smear out into ribbon-like shapes, while the stronger rock types will resist deformation as rigid chunks or else break into pieces. These “stretched pebble conglomerates” record some deformation, but not so much as to qualify the metaconglomerate as mélange.
Deformation of these rocks into geological structures
Convergent plate boundaries are stressful places: the rocks and sediments fed into the subduction zone are subject to compression and shearing, and the frequently deform in response to these stresses. Examples of deformation typical of subduction zone are well recorded in the rocks of California’s Coast Ranges: folding, faulting, and extensive zones of shearing that result in “mixed-up” rocks called mélange. Let’s examine each in turn.
CRUMPLED LAYERS: Folding
Folding is particularly obvious in the regularly-layered cherts of the Marin Headlands Terrane, though it can also be spotted where bedding can be detected in the sandstone and shale layers in other adjacent terranes. Mechanically differentiated rock layers, such as chert/shale interbeds, are particularly susceptible to folding when compressed parallel to their layering: they buckle and crumple. Because the shale is weak and the chert is relatively stiff, layers of chert slide relative to one another on a “geological banana peel” of shale. This flexural slip allows much more folding than a homogeneous body of chert would be able to accommodate.
Folding occurs on many scales in California’s Coast Ranges. As the example below shows, millimeter-thin layers can be folded, but so too can entire terranes on the scale of kilometers. For instance, the Angel Island Terrane is near the axis of a regional synform, with more recently accreted terranes on either side of it as well as wrapping around below it.
SLIP ALONG FRACTURES: Faulting
Another form of deformation that is common in the accretionary wedge is faulting. Through faulting, new slivers of the oceanic section are slathered onto the bottom of the accretionary wedge complex, building it up from below.
At Marin Headlands, north of the Golden Gate Bridge, there are no fewer than 17 thrust faults that have placed deeper rocks of the oceanic crust on top of near-shore and off-shore sediments, repeating the overall oceanic section again and again.
Sometimes these faults are discrete and crisp, where two vastly different rock units have been brought into contact along the slip surface. An example of this can be seen at right, where former mantle rock (serpentinite of the Hunter’s Point Shear Zone) has been emplaced atop seafloor sediments (shale and sandstone mainly) of the Marin Headlands Terrane. During the Mesozoic, this fault was for a while the plate boundary between North America and the Farallon Plate. As subduction proceeded, new faults formed within the subducting plate, and the relative motion between the two plates transferred from a shallower fault to a deeper one. Newly subducted sediments were shoved underneath the mantle sliver, as the Farallon moved downward relative to North America. Later, another new fault formed, becoming the new plate boundary, and this fault was abandoned, left inert in the geologic record, with the footwall rock below it now also accreted to the North American Plate.
THE MIXED-UP ROCK: tectonic shearing to make mélange
Other rocks may be sheared out, deforming in a mass. Serpentinites are particularly weak rocks under shear stress, and break into thousands of tiny faults, giving the overall rock a “scaly fabric,” which looks at first like a cleavage. However, many of the little flakes shows slickenlines on their surfaces, indicating they are little faults that have seen some slip. A large body of serpentinite can thus be transformed to a broad shear zone. If other rock types get mixed into the sheared-out serpentinite, a serpentinite mélange will result. Mélange is a term for tectonically-mixed-up rock, where varying blocks or “lozenges” of varying rock types get entrained in a sheared matrix. We call this texture “block in matrix.” Most commonly the blocks are of coherent blocks of the matrix rock, but in other cases, all sorts of exotic blocks can be introduced. At Ring Mountain in Tiburon, for instance, a serpentinite matrix hosts blocks of blueschist, eclogite, amphibolite, peridotite, and meta-chert.
This pulpy mess is extraordinarily incompetent, and is the source of frequent landslides in coastal California. Shockingly, the south tower of the Golden Gate Bridge is anchored in serpentinite mélange of the Hunter’s Point Shear Zone.
Shale acts in the same slippery fashion as serpentinite, as the flakes of clay it contains slip relative to one another, creating a scaly fabric. Shale-based mélange is just as common as serpentinite-based mélange. In San Francisco itself, the Hunter’s Point Shear Zone is a wide swath of serpentinite mélange, which the City College Fault Zone has more of a shale basis to the scaly matrix. Near Pacifica, a town further south along the coast, the mélange has a distinctive mix of metavolcanic and metasedimentary source rocks, resulting striking mixtures of pale green and dark gray.
The end of subduction
Subduction of the Farallon Plate continued until the around 25 million years ago. By the point, the faster rate of subduction had caught up to the slower rate of seafloor spreading, which consumed the last bit of young, fresh Farallon crust, and then the oceanic ridge itself. This brought the North American Plate into contact with the Pacific Plate for the first time. This “moment” was the birth of the San Andreas Fault, the modern transform plate boundary between North America and the Pacific Plate. The shape of the Farallon Plate was such that it was narrowest in the middle, and wider both north and south of that. So the first spot the subduction zone gobbled up the spreading center was in the middle of the Farallon Plate. This broke it into two. Over time, the Farallon’s remnants (the Juan de Fuca Plate in the north, and the Cocos Plate in the south) continued to subduct, but they get smaller and smaller with each passing millennium. Between them, the San Andreas Fault continues to get longer and longer. Here’s an animation showing how this has played out:
That covers what happened horizontally after the end of subduction, but there were also vertical motions of the crust. With the cessation of the drag of the subducting slab, the accretionary wedge was no longer pulled downward. It was allowed to rebound upward, and eventually rose to poke up above sea level, making the California Coast Ranges. Similarly, the forearc basin uplifted to become the flat-as-a-pancake Great Central Valley, and erosion of the volcanoes of the continental volcanic arc revealed their deeper plumbing system: the plutons of granite and granodiorite that we now see in the Sierra Nevada range.
Anderson, Don (2007). “The Eclogite Engine,” Chapter 5 in New Theory of the Earth (book), Cambridge University Press.
Atwater, Tanya (1970). “Implications of plate tectonics for the Cenozoic tectonic evolution of western North America,” Bull. Geol. Soc. Amer., v. 81, p. 3513-3536. DOI: 10.1130/0016-7606(1970)81[3513:IOPTFT]2.0.CO;2
Kirkpatrick, Jamie; Rowe, Christie; Bentley, Callan; Blisniuk, Kim, and Wakabayashi, John (2019). Streetcar 2 Subduction, digital field trip guide the geology of the San Francisco Bay Area. American Geophysical Union.
Warhaftig, Clyde (1984). A Streetcar to Subduction, and other plate tectonic trips by public transportation in San Francisco. American Geophysical Union: book.