Lecture 3: Subduction

1
Lecture 3 Subduction

• Questions
• What is subduction and why does it cause
  volcanism?
• What controls the differences in lava composition
  at arc volcanoes?
• What controls the geomorphology and eruption
  style of volcanoes?

1
2
Hydrothermal vents

• One of the coolest things about mid-ocean ridges
  is that the heat from the magma drives vigorous
  hydrothermal systems. This has many important
  consequences
• It supports ecosystems of microorganisms, plants,
  and animals that function in the total absence of
  photosynthesis. Some believe this is the
  environment where life originated and/or survived
  big impacts
• It has large effects on the chemical and isotopic
  composition of seawater
• It drives alteration of the oceanic crust and
  lithosphere, and hence the storage of water in
  materials that will be subducted. This water may
  drive arc volcanism at subduction zones or be
  recycled to the mantle.

2
3
Hydrothermal venting at mid-ocean ridges

• As new ocean crust cools, it fractures and
  becomes permeable to hydrologic flows. Seawater
  invades the crust, is heated, reacts with the
  rock, and is expelled from warm or hot vents.
• The chemistry of the hydrothermal fluid is
  controlled by exchange with the basalt, by phase
  separation into vapor and brine, by microbial
  activity, and by mixing on discharge with cold
  seawater

3
4
Hydrothermal venting at mid-ocean ridges

• These diagrams show chemistry of hydrothermal
  vent fluids with increasing time after a dike
  injection event that provided a new heat source
  and a new cycle of high-temperature venting. The
  component mixed with seawater evolved from a
  low-Cl, low-Fe vapor to a high-Cl, high-Fe brine
  as the temperature of venting decreased

4
5
Hydrothermal venting at mid-ocean ridges

• Looking at extinct oceanic crust preserved in
  ophiolite complexes, evidence of the hydrothermal
  alteration processes that occurred near the ridge
  is preserved in the crustal section. Besides the
  presence of sulfide deposits and hydrated
  minerals, one signature of water-rock interaction
  is the shift in the oxygen isotope ratio 18O/16O
  in the minerals.
• Seawater-rock interaction at low temperature
  leaves the rock enriched in 18O, whereas at high
  temperature it leaves the rock depleted in 18O.
  In the Oman ophiolite we can therefore see
  evidence of low-temperature circulation in the
  extrusive lavas and dikes, and high-temperature
  circulation in the gabbro section.

The fact that the average value in this whole
section is close to the primary mantle value led
Gregory and Taylor to propose that alteration at
mid-ocean ridges determines the value of seawater
18O/16O.
5
6
Subduction and associated magmatism
6

• Observation Most of the worlds volcanoes (and
  nearly all the hazardous ones) sit above
  subduction zones!
• The basic problem slabs are cold...why does
  subduction cause volcanism?

7
Subduction and associated magmatism

• Key observations
• Volcanic front is typically 100 km above the
  Wadati-Benioff zone, independent of subduction
  angle or velocity.
• Subduction zone magmas are wet, often several
  percent water (that is why arc-type volcanoes
  erupt explosively, more on this later).
• Geochemical tracers show signature of subducted
  material (sediment, altered basalt) coming up in
  the arc lavas.

7
8
Evolution of slab and mantle wedge

• Thermal environment is set by the temperature
  structure of the downgoing lithosphere, by
  induced convection in the mantle wedge, and by
  heat transport and generation within the system.
• Models show isotherms dragged far down the slab
  and predict moderately low temperature to great
  depth along the slab-wedge interface.

8
9
Evolution of slab and mantle wedge

• What materials are available to contribute to
  magmatic activity at subduction zones?
• Downgoing slab is made of sediments, altered
  oceanic crust, and depleted (serpentinized?)
  lithospheric mantle peridotite.
• The mantle wedge is made of peridotite, possibly
  modified first by melt extraction in the back-arc
  basin, then by addition of mobile components from
  the slab.
• The overriding plate is made of old oceanic or
  continental crust and earlier products of ongoing
  subduction processes.
• We can understand the possible environments for
  melt generation by mapping the phase
  relationships of these materials onto thermal
  models of slab-wedge thermal convection
• If we know what is there, and in the laboratory
  we measure where these materials melt or
  dehydrate, perhaps we can develop a sensible
  model for the unseen processes the explains the
  observed products.

9
10
Evolution of slab and mantle wedge
10

• First consider the metamorphism of
  hydrothermally altered basaltic rocks in the
  crust of the downgoing slab.
• This is our fist look at the metamorphic facies
  diagram.

It shows the generalized phase assemblages that
occur on metamorphism of hydrated basaltic rocks
at various Pressure and Temperature
conditions. The metamorphic conditions seen by
rocks of any composition are named by these
facies, but here it is convenient that the facies
are named for hydrated basaltic rocks, since we
are looking at subduction of altered oceanic
crust.
11
Evolution of slab and mantle wedge
11

• Hydrothermal alteration of the crust creates
  hydrous minerals and stores water in the rock.

Plotting the P-T path of slabs shows that
subducting basalt passes through blueschist into
eclogite facies. Eclogite is an anhydrous rock,
so by 40 km we have almost total dehydration of
basalt. We follow this water to understand the
melting process. Note the red curve is the
water-saturated solidus, not relevant if water is
dehydrated at lower T.
12
Evolution of slab and mantle wedge

• Water promotes melting of rocks at high
  pressure, where the solubility of water in the
  melts becomes large (think freezing point
  depression).
• This diagram shows the solidus curves for basalt
  and peridotite when dry (lines at high
  temperature) and with excess water (concave
  upwards curves at lower T).

• It also shows the stability limit of amphibole,
  the primary mineral able to store water and
  prevent it from enhancing melting.
• If all the water can be tied up by amphibole
  formation (2), the effective solidus is formed
  by the amphibole stability limit and the
  water-saturated solidus (hatched).

AmP
AmB
AP
WP
Note reversed sign on Pressure axis!
(Petrologists plot pressure increasing upwards,
geophysicists plot depth increasing downwards)
12
13
Evolution of slab and mantle wedge

• The next step is to take these phase boundaries
  and draw them on the P-T field generated by a
  thermal convection model of a subduction zone, to
  show the likely regions of dehydration and melt
  generation.

• In this model, the slab dehydrates (at AmB), but
  never melts.
• Peridotite melting is restricted to this area
  here, at higher P and T than the intersection of
  amphibole stability and the wet solidus of
  peridotite.

13
14
Evolution of slab and mantle wedge
14

• A more general, complete and intuitive
  qualitative version is shown in this diagram by
  Prof. Wyllie

• Things to note in this mess
• The key observation that the volcanic front sits
  100 km above the Benioff zone could be explained
  by the shape of the Amphibole stability curve.
• Shows possibility of slab melting if any water
  is retained to gt100 km. Probable only in
  Archaean or for very young slabs.
• Mantle melts likely to pond and evolve at base
  of low-density crust.

• Sorry about the reversed sense of
  subductionpeople draw it both ways.

15
Evolution of slab and mantle wedge
15

• The slab can lose its water in several ways
  dehydration and/or melting of sediments, basalt,
  and/or serpentinite.
• In most modern subduction zones, the basaltic
  component of the slab probably does not melt.
  This may have been different in the past when
  slabs were typically hotter.
• Instead there are three principal additions of
  fluid to the mantle wedge first, shallow
  sediment melting later, basalt dehydration
  finally, serpentinite dehydration.

Old model (Davies and Stevenson) transports water
as amphibole in peridotite parallel to the slab
until a higher pressure (AmP line), when it
returns to vapor that rises vertically across the
amphibole stability line and becomes bound again.
Eventually it could zig-zag across until it
crosses the wet peridotite solidus (WP or
location N in above figures). The problem with
this model it takes too long.
16
Evolution of slab and mantle wedge

• Timing of transport from slab to arc
• 10Be, a cosmogenic nuclide with a half-life of
  1.6 Ma, is subducted with sedimentary fluids. It
  turns up, still alive, in arc volcanics, which
  implies that the amphibolite zig-zag method of
  transport from the slab to beneath the volcanic
  front is, at least, not the only thing going on.

U-Th disequilibria in some arcs there are two
slab components a sediment melt, a delay of gt350
ka to return to U-Th secular equilibrium, then a
hydrous fluid from basalt dehydration, followed
within 30 ka by eruption For a subduction rate
of 10 cm a-1 and a slab dip of 60, this implies
dehydration of the basalt 30 km deeper than
melting of the sediment, and rapid transport to
melting site, melt extraction, and eruption.
16
17
More recent model of slab/wedge evolution

• Hebert et al. (2008) variable viscosity model
  with tracing of water in vapor, melt, hydrous
  minerals, and nominally anhydrous minerals

18
More recent model of slab/wedge evolution
19
Compositions of typical Arc volcanic rocks

• Arc volcanics are usually calc-alkaline suites
  with the differentiation sequence
  basaltandesitedaciterhyolite. Variants depend
  on K2O content.
• Low-K continental arc assemblages follow a
  sequence gabbrogabbrodioritetonalitelow-K
  granodiorite.
• High-K island arcs generate trachybasalt-trachyand
  esitephonolite.
• Broadly speaking there is a time progression in
  the evolution of an island arc from low-K
  (tholeiitic), to medium-K (calc-alkaline), to
  high-K (shoshonitic).

• This is best understood as an increase in the
  extent of contamination and assimilation of new
  primary magmas by older rocks of the same arc.

Low-K, normal, High-K
19
20
Mature arcs 2 MVB Cascades
21
Compositions of typical Arc volcanic rocks

• Low-K (or calcic or tholeiitic) series is
  dominated by basalt and basaltic andesite. Low-K
  dacites and rhyolites are rare. This is common in
  juvenile island arcs (isolated islands built
  directly on oceanic crust.
• Calc-alkaline suite is dominated by andesite,
  although the whole sequence from basalt to
  rhyolite is usually found. This suite dominates
  the building of mature island arcs (Japan,
  Indonesia), where islands have joined together
  and built a continuous arc of proto-continent
  above sea-level.
• Calc-alkaline andesites also dominate continental
  margin arcs like the Andes and Cascades.

• When the arc is mature, or subduction stops,
  high-K series become common. These are dominated
  again by more mafic rocks, basalt to basaltic
  andesite.

21
22
gt23700 rocks!
23
Compositions of typical Arc volcanic rocks

• The frequency of eruption of rocks at various
  stages of differentiation, as well as the
  occurrence of the different series, is best
  thought of in terms of density filtering.
• It is difficult for magma to rise though less
  dense country rock. Broadly speaking the density
  of rock decreases with increasing SiO2 content or
  decreasing FeOMgO content.
• When the overlying plate consists only of
  basaltic oceanic crust, arc-related basalts can
  rise directly and erupt with little
  differentiation and with little opportunity for
  assimilation.
• Thus, early in the life of an arc, the low-K
  series occurs and is dominated by mafic members.
• As the crust evolves and thickens, it becomes
  harder for basalt to rise through it. Instead
  primary basalts are likely to pond at the base of
  the crust and undergo differentiation and
  assimilation.
• The result is calc-alkaline andesite, dacite, and
  eventually rhyolite.
• High-K and shoshonite series presumably pond for
  long times and undergo extensive assimilation,
  but eruption of mafic members of the series is
  promoted by extensional tectonics that set in as
  subduction slows.

23
24
Eruption Styles, Volcano Morphology, Lava
Composition
Arc Volcanoes Steep cones, explosive eruptions
Help!
24
25
Eruption Styles, Volcano Morphology, Lava
Composition
Hot-spot Volcanoes Low, shield-profile cones,
effusive eruptions
Cool!
25
26
Eruption Styles, Volcano Morphology, Lava
Composition

• The way that a magma erupts is determined by its
  composition, by the flux of magma and by local
  tectonics
• The most important variables are viscosity and
  volatile content.
• More viscous magmas build steeper-sided cones.
• Viscosity is determined by composition,
  temperature, crystal content, and volatile
  concentration. Generally, more SiO2-rich lavas
  erupt at lower temperature, carry more
  phenocrysts, and are more viscous at equal
  temperature, so all these effects work together
  to make viscosity increase with SiO2 content.
• Basalts make shield volcanoes. Andesites and
  basaltic andesites make stratovolcanoes.
  Rhyolites make domes.
• More volatile-rich magmas yield more explosive
  eruptions, especially if they are viscous enough
  to retard bubble escape.
• Basalts are fairly dry and tend to experience
  effusive eruptions a wet basalt can fragment and
  make a cinder cone.
• Rhyolites, especially, are very volatile-rich and
  viscous, so bubbles can accumulate to the point
  of fragmentation, leading to explosive ash
  eruptions.

26
27
Volcano morphology basaltic

• Shield volcano, formed by repeated effusion of
  low-viscosity magma from a central vent and flank
  rift systems

Cinder cone, a small-scale feature (100 m high)
formed by fragmentation of wet basaltic magma.
27
28
Volcano morphology andesite-rhyolite

• Stratovolcano cone, formed by alternating
  effusion of lava flows and ejection of
  pyroclastic debris

A caldera, formed by collapse of the roof of a
shallow magma chamber after sudden eruption and
emptying of the chamber. Calderas range from 1
km to 50 km in diameter on Earth. Contrast a
caldera, formed by collapse, with a crater,
formed right at a surface vent by direct
expulsion of material at the surface.
28
29
Eruption Styles, Volcano Morphology, Lava
Composition

• Explosive eruptions are of two general types
• driven by magmatic fragmentation from exsolution
  of dissolved volatiles, or
• phreatomagmatic explosions resulting from the
  boiling of external water brought into contact
  with hot magma.
• If an explosive eruption generates a large column
  of ash that reaches the stratosphere, it is a
  plinian eruption.
• Explosive eruptions, instead of making lava
  flows, make pyroclastic flows and pyroclastic
  deposits.

• These include nuées ardentes, or glowing clouds,
  which are very fast moving incandescent masses of
  turbulent air-ash mixtures (example, Mt. Pelée on
  Martinique, 1902), and
• lahars or volcanic mudflows, essentially
  water-induced landslides of unconsolidated
  volcanic debris on steep slopes.

29
30
Volcanic DepositsLahar, or volcanic mudflow
30