Subduction lecture

1
Chris Goldfinger Burt 282 7-5214 gold_at_coas.ore
gonstate.edu
OCE 661 Plate Tectonics
Course notes at Http//activetectonics.coas.orego
nstate.edu
2
(No Transcript)
3
(No Transcript)
4
(No Transcript)
5
(No Transcript)
6
Arc/forearc type is directly related to the
interplay of forces at the plate boundary. Arc
extension vs compression is closely related to
plate age. Older plates sink and the hinge in
the slab rolls back, away from the arc. This
leaves almost a free surface a the the trench.
Gravitational extension of the arc, continent or
ocean basin of the overriding plate can result in
forearc extension, development of backarc basins,
and other extensional features.
7
This is an illustration of some of the
arc/forearc features expected across the range of
compressional and extensional regimes. It is
also a history of the Baja peninsula, formerly a
subducition zone. When the age of the subducting
plate changes with time, the regime will change
with it. This system underwent a change from
extension to compression as the subducting plate
age shifted toward younger with time as the ridge
approached north America. When this occurs,
normal faults are pushed into reverse, and
extensional features are closed up and converted
into compressional features.
8
Why would a normal arc system like this change to
a backarc basin system?
How about a shift in age of the subducting plate
from young to old?
9

• Thermal, seismologic, and velocity model of a
  subduction zone.
• What controls where earthquakes and deformation
  occur?
• Thermal structure and the brittle ductile
  transition at the downdip end
• Fluid pressure, thermal and tectonic dehydration
  reactions at the updip end.

300C
450C
Brittle ductile transition
From Wells et al. 2003

• How can you approach the problem of determination
  of the coupled zone?
• Thermal and geodetic modeling, shown here by the
  gray locked and transition bars over the forearc.
  
• The distribution of seismicity in the upper
  plate, shown here, marks the brittle-ductile
  transition. The weaker plate controls the limits
  of deformation.
• Comparison with other subduction zones. The
  predicted Rupture width from gravity anomalies
  (and forearc basin width). Maybe. Maybe not.

10
Great Earthquakes and locked plate
boundaries-Some of the observables and their
recent interpretations. Well focus on Sumatra
and Cascadia, with help from Nankai, South Chile,
and many other places.
11
Lets start with forearc basins. Every
subduction zone has one, most have two. The
Willamette Valley is one of two that Cascadia
has. The other one is offshore under the
continental shelf. Sumatra has one for the
most part. The Sumatran inner forarc high is
under theshelf, but is very large and close to
the arc, more like the Willammette Valley (more
properly the Puget-Willamette forearc basin.
12
Forearc basins are synclines, sort of. A
syncline is an active fold generally caused by
compression normal to the axis. Most forearc
basis show compression on their seaward sides,
and little or nothing in terms of deformation on
their arcward sides. A better description might
be a monocline with thrust faulting on one side.
This example from south sumatra, near the 2007
Bengkulu 8.4 earthquake is typical.
13
compression
Growth strata
compression
Forearc basins, even in compressional settings,
are more than just the place that sediments
accumulate between the arc and accretionary
wedge. They commonly exhibit both compressional
and extensional features. But fundamentally they
seem to be compressional, as shown by this
seismic profile of the Sunda margin. Notice the
growth faulting, units that thin toward the
margin of the basin, showing active growth of the
syncline. Notice also the compressional
features. The fault at left is a thrust fault in
many margins. Why would there be a basin in the
middle of a compressional accretionary wedge? Is
it extensional, or is it just there because the
active wedge at the left keeps getting larger?
14
Evolution of the Sunda Forearc wedge and basins
Its unclear in many cases whether the forearc
basin is there because of active subsidence,
active folding, or active folding and uplift of
just the seaward part of the fold. But overall,
many forearc basins seem to be the passive result
of uplift to the seaward side. Most have
compressional structures, rarely extensional
except where the forearc as a whole is extending,
as in subduction erosion.
15
Backarc basins are due most likely to counter
flow in the upper mantle, or rollback of the slab
and trench suction pulling the arc forward with
the slab The sedimentary sequences can be
straightforward, or reflect complex
volcano-clastic history with changes in backarc
type over time.
16
Forearcs can also be quite complex.
Fundamentally accretionary, they may also have
low sediment supply and can be mostly
crystalline, as in Peru. Why would you get a
situation like this? The entire forearc is
extensional, full of normal faults???
Old slab, itself heavily normal faulted causes
rollback and extension. Rough slab surface
causes basal erosion, i.e the chainsaw model.
17
Generally, if there is extension in a
compressional setting, it tends to be relatively
shallow and independent of the compression at
depth, as shown here.
18
Variations in structural style can include major
extensional and compressional features in the
same setting. This example shows a major set of
normal faults backtilting parts of an
accretionary wedge and controlling sedimentation
in the forearc basins. This also occurs on the
Washington margin.
19
Variations in structural style can include major
extensional and compressional features in the
same setting. This example shows a major set of
normal faults backtilting parts of an
accretionary wedge and controlling sedimentation
in the forearc basins. This also occurs on the
Washington margin.
20
Snickers bar tectonics-brittle extension above
a ductile melange.
The snickers model
21
This shallow extension above a detachment surface
is a bit unusual, and may be the result of a
previous episode of erosion that truncated the
outer wedge, allowing the upper wedge to slide
seaward. Such extension may the result of
adjustments to wedge taper to reach an
equilibrium critical taper.
22
(No Transcript)
23
Gutscher 1998
Interplay between sediment input and basal
friction on the basal thrust in subduction zones.
These factors are not independent. Why?
Rapid sediment input results in overpressuring
of the wedge as trapped pore water cannot escape
fast enough. High pore fluid pressure in turn
results in reduced basal friction as pore fluid
pressure approaches lithostatic. At near
lithostatic pressures, the wedge becomes
decoupled, and vergence of the wedge may change
to mixed or landward vergence (seaward dipping
imbricate thrusts).
Wilett et al., 2004
24
The most recent version, Wand and Hu (2006) is
known as the dynamic coulomb model. In a
nutshell, it includes the co-seismic strain and
fluid pulse expected on the decollement, allowing
the wedge to move during high porefluid pressure
conditions during the earthquake.
25
Wilett et al., 2004
26
Gutscher et al., 2001
A well known example of landward vergence in a
submarine wedge
27
A well known example of landward vergence in a
submarine wedge In Northern Oregon and
Washington. Experiment here shows landward
vergence as a result of backstop shape and strain
rate. This probably isnt the reason for
landward vergence in Washington, but is the case
in southern Oregon. What might be the reason?
Gutscher et al., 2001
28

• The two most common reasons for landward vergence
  are the two weve discussed
• Backstop shape
• High pore fluid pressure (reduction of basal
  shear stress)

Nitinat Fan Astoria Fan
Mud volcano
Landward vergence
Some other relevant information Backstops are
virtually always dipping seaward The fans are
Pleistocene in age, less than 1my. They are 1-3
km in thickness, and 20-80 km of wedge is
composed of re-accreted fans.
The mud volcano is primary evidence of
hydrofracturing due to lithostatic (or greater)
pore fluid pressures. This one is on the abyssal
plain 5 km seaward of the deformation front.
The combined evidence suggests the Washington
wedge is highly decoupled. This results in
landward vergence, and has implications for
earthquakes. What implications?
29
Plate tectonics tip for Today If you see this,
dont stop to take pictures
30
Ok, so can any of these observations be related
to earthquakes? Or be used to predict updip and
downdip limits, and plate resistance forces?
31
Lets look a the 2004 Sumatra earthquake to
visualize both the elastic response to the
earthquake, and the generation of the tsunami.
32
Elastic tsunami
Tsunami can be generated also from submarine
landslides, and from slow earthquakes. If the
impulse takes place in 15 minutes or less, the
period is compatible with tsunami wave
excitation.
33
Unlike any other forearc, Sumatra has islands
offshore in the forearc high, they are the
forearc high. These islands go up and down
according to where they sit relative to the
locked interface and co-seismic slip. This is
before the earthquake
34
This is after the earthquake notice the
stripped vegetation, and the white dead reef on
the west side that has been uplifted.
35
In the field, uplift and subsidence are very
obvious
Some of the slip was not seismic, Port Blair
continued to subside for several hours after the
tsunami had passed, suggesting very long period
slip continued. Courtesy C.P. and Kusala
Rajendran, CESS, Trivandrum India
36
Destroyed mangrove trees, land was uplifted by 1m
37
Line of barnacles 0.75 m above the present high
tide
38
The little boat that is generally available for
inter island short trips. Who said Kusala is
nervous!
39
Avis island was elevated by 1 m, no one lives
here. Coconuts for free, if you can climb the
trees
40
On the way to the interview Island. Note the new
sprouts in the square, the new high-tide level
(change by 1 m)
41
Expedition from the east to the west coast of the
interview island. This island has 45 elephants.
42
Another view of the elevated mangroves, on the
way to the interview island
43
(No Transcript)
44
The inflection line between uplift and subsidence
is a reasonable first order match for the downdip
extend of the rupture (excluding afterslip) The
offshore islands such as Simuelue, Nias, the
Batus, and the Mentawais are the expression of
the outer arc high/backthrust at the rear of the
accretionary wedge.
45
Landward vergence
Sound familiar?
46
Karig D.E, Moore G.F, Curray J.R, and Lawrence
M.B. Affiliation Dept of Geol Sci, C. U. I. N.
Y. U. S. A. E. H. D. E., 1980, Morphology and
shallow structure of the lower trench slope off
Nias Island, Sunda arc, American Geophysical
Union Geophysical Monograph 23, 179-208 In The
tectonic and geologic evolution of Southeast
Asian seas and islands. Part 1 (1980) p. 179-208
p.
47
(No Transcript)
48
Swath bathymetry from HMS Scott, March 2005. The
asymmetry of the folds is apparent, indicating
landward vergence.
49
(No Transcript)
50
New bathymetric data from the 26 December
mainshock region reveals local channels that most
likely contain the Holocene record of great
earthquakes. While the physiography differs
from Cascadia, localized channels should be good
depocenters. Images Courtesy L. McNeill.

51
Like Cascadia, vergence shifts southward to
landward. Probably due to the decrease in
sediment supply, this shift means higher
coupling, and therefore shorter earthquake
recurrence intervals in the south (100-200 years)
as opposed to the north (500-1000 years).
52
GPS data prior to the 26 December event suggested
poor coupling north of the Batu Islands,
nevertheless this poorly coupled region was the
source of the Mw 9.2 event and the 28 March 8.7
event.
Many people mistook poorly coupled for
unlikely to have a great earthquake
Nevertheless, poor coupling is probably linked to
overpressuring of the wedge due to rapid fan
accretion, as in Washington. Allternatively, and
equally good interpretation is that no further
strain was accumulating, signaling an imminent
earthquake. You decide
53
Chennai
What is the recurrence in India? This photo is
from Mahabalipuram, near Chennai on the east
cost of India. This site was hit by the 2004
tsunami. Below that deposit, this temple was
apparently hit by something, It is thought to be
700 years old, and may have been under
construction at the time. Last tsunami?
54
Lets look at some recent work dealing with
models and indicators of interplate locking.
This one, from Wells et al., 2003 JGR suggests
that forearc basins are asperities or locked
patches. This seems to work well on some
margins, notably Nankai where it was developed.
55
The theory is that locked patches result in
subduction erosion of the base of the upper
plate, which results in subsidence of the forearc
basin. The paper attempts to match slip zones
globally to forearc basins, using gravity as a
proxy.
In Nankai, the forearc basin seems to coincide
with the slip patch of the 1944 earthquake, a
little less well for the 1946 earthquake. Along
strike, the rupture boundaries are thought to
coincide with the basins.
56
(No Transcript)
57
(No Transcript)
58
(No Transcript)
59
(No Transcript)
60
In Alaska, the slip isnt so clearly centered in
a basin.
61
In Chile, the whole margin is collapsing due to
subduciton erosion.
62
(No Transcript)
63
(No Transcript)
64
(No Transcript)
65
Nehalem Bank
66
The model then, proposes that subsidence of
forearc basins is caused by basal subduciton
erosion of the upper plate at the locked patches.

Is this consistent with what we have seen of
forearc basins?
67
Houston we have a problem Sumatra.