EART 118 SeismotectonicsLecture 17
CONCEPTS:
Variable frictional properties seem ubiquitous
Subduction Zone Geometry…
Earthquakes!
No seismicsignal
Episodic SlipSteady
SlipTremor
36-40 mm/yr loading rate Coast moves ~2 cm/yr today (SNARF)Margin-wide recurrence: 550 years (var ~200 yrs); 313 years into the eq cycle
~10 meters of post-1700 accumulated slip deficitLast eq appears to have been margin-wide (M9)
LVZLOC
Sealed plate boundary Permeable plate
boundary
Forearc
Arc volcano
Serpentinized mantle wedge
Overpressuredoceanic crust
Layer 3
Layer 1Layer 2 Oceanic crust Oceanic mantle
Spreading ridge
Hydrothermal circulation
EcoglitizationLFE hypocentersW-B hypocenters
MODEL
1. upper oceanic crust intensely hydrated through hydrothermal circulation at ridge;2. sediments carpet seabottom, plate cools, free water incorporated in hydrous minerals;3. upon subduction, increased P/T cause prograde metamorphic reactions, freeing fluids at lithostatic pore pressures;4. pressure maintained by impermeable plate boundary above and massive gabbro below5. metamorphic reactions culminate near 45 km depth with eclogitization and ruptureof plate boundary seal;6. fluids escape into overlying mantle wedge causing serpentinization and pore pressure reduction
Base of Seismogenic Zone
• CATACLASITE (P-sensitive)
~350°C - ~450°C(quartz vs. feldspar)
• MYLONITE (T-sensitive)
Cascadiavelocity
field(SNARF)
2010
vertical
2010
+
Szeliga et al., 2008
Chapman and Melbourne., 2009
Many magnitude 6s
Subtract from convergence rate of 34 mm/yr
•~Half of convergence is accommodated by large ETS events
•All imaged slip occurs below 25 km depth, above 40 km
• But max slip ~ 1/smoothing
6 Gomberg and Peng, 2010
Tremor (and slow slip) vs LFE
vs earthquake
TimevsDistancealongstrike
• Along-strikepropagationvelocityvaries– 7to12km/day
• Generaltendencytomoveupdip– WechandCreager
• Originatein3places
Houston,Delbridge,Wech,CreagerNatureGeoscience,inrevision
Non-volcanic tremor always correlates with GPS-detectable ETS events
• Space and time correlation between slow slip and tremor detection along nearly the entire margin
Brudzinski etc 2009
GPS-inferred slip: Mw= 6.75 PNSN tremor
August 2010 event
Slip Tremor?
Melbourne
extendsupdipfrom
SimilarresultsfromSchmidtandKrogstad
ETS may delineate a 25 km lower limit to interseismic strain accumulation
After Hyndman, Dragert, Wang, etc, 1992-2003
A rough forecast of future slip after full recurrence interval
Tremor, VLFs, SSEs - All Slow, Interplate Earthquakes
TREMOR EPICENTRAL DISTRIBUTION NANKAI vs CASCADIA
Nonvolcanic Deep TremorAssociated with Subduction in
Southwest JapanKazushige Obara
Deep long-period tremors were recognized and located in a nonvolcanic regionin southwest Japan. Epicenters of the tremors were distributed along the strikeof the subducting Philippine Sea plate over a length of 600 kilometers. The depthof the tremors averaged about 30 kilometers, near the Mohorovic discontinuity.Each tremor lasted for at most a few weeks. The location of the tremors withinthe subduction zone indicates that the tremors may have been caused by fluidgenerated by dehydration processes from the slab.
Long-period events and tremors with typi-cal periods in the range of 0.2 to 2 s areoften observed at active volcanoes and re-flect the internal dynamics of the volcanicsystem (1). A possible tremor-generatingmechanism is flow-induced oscillation inchannels transporting magmatic fluid (2).We have identified and studied anomalouslong-period tremors from a nonvolcanic
area in southwest Japan by using the Na-tional Research Institute for Earth Scienceand Disaster Prevention’s (NIED) high-sensitivity seismograph network (Hi-net),which is composed of about 600 stationsinstalled throughout Japan to detect mi-croearthquakes (3). The densely distributedhigh-sensitivity seismic stations provide ahigh-level detection capability for micro-earthquakes and offer us an opportunity tofind and investigate very small amplitudetremors. Because the amplitudes of thesetremors are very small, it is difficult to
identify them with a single station or asparse network.
We observed small-amplitude tremors thatlasted from a few minutes to a few days, asshown in fig. S1A. The tremors were observedsimultaneously at several Hi-net stations, whichindicates that they are not related to artificialnoise. The predominant frequency of the trem-ors ranged from 1 to 10 Hz and was lower thanthat of ordinary earthquakes of similar size (10to 20 Hz). We transferred the raw seismogramto root-mean-square (rms) amplitude for thefiltered output (fig. S1B), and tremors wereclearly seen for time windows of 35 to 50 min.The envelope shapes of the tremors were verysimilar at different stations. The envelopes hadgradual rise times and differed from those of anormal earthquake, which has a spike-like en-velope shape. The similar envelope amplitudepattern seemed to have been propagated with avelocity of 4 km/s, which we roughly estimatedfrom paste-up traces plotted with the increasingepicentral distance. This means that the sourceof the tremors was located at a deep portion andthe envelopes were propagated not by P-wave,but by S-wave velocity. Because it was verydifficult to identify the initial P- and S-waveonset for the hypocentral determination, weapplied a cross-correlation technique to get thedistribution of the relative arrival time of theenvelope (4). The spatial distribution of the
National Research Institute for Earth Science andDisaster Prevention, Tenno-dai 3-1, Tsukuba, Ibaraki,305-0006, Japan. E-mail: [email protected]
Fig. 1. Epicentral distribution of the deep long-period tremors in the yearof 2001. The circle indicates the center of epicenters of tremors calcu-lated every hour. The crosses represent the Hi-net stations. The depth
contour line indicates the maximum frequent depth-distribution ofearthquakes inside the subducting Philippine Sea plate, and the gray linerepresents the leading edge of the subducting Philippine Sea plate (11).
R E P O R T S
www.sciencemag.org SCIENCE VOL 296 31 MAY 2002 1679
on N
ovem
ber
26, 2010
ww
w.s
cie
ncem
ag.o
rgD
ow
nlo
aded fro
m
Wech & Creager (2012) Obara (2002)
• tremor locations parallel 30-40 km slab contours• tremor width narrower in Japan (~30 km) than Cascadia (~80 km)
20, 30, 40 km
Tremor Mechanism
Slow shear slip, not fluid flow.
Does tremor in Japan, Cascadia, California, Costa Rica, Mexico, Alaska,… share the same mechanism?
Responds fast to stress(talk by Chastity Aiken)
Rubinstein, Vidale, et al.
…Turned on its side
Tremor under the San Andreas
Earthquakes
Tremorsand episodic slow slip?
Tremor under the San Andreas
Nadeau, 2005
Parkfield
Much of what we know about fault behavior at
depth comes from seismicity
Source Properties from SeismogramsEarthquake
???
P and S arrivals, polarity, waveform modeling, etc...
Source: unr.edu1 minute
Tremor
40 minutes
Matched filter technique:Template Waveforms
10 minutes 25 seconds
Shelly et al., Nature, 2007
Scan template through continuous data,sum correlations
13
Stacked LFE Templates
Single LFE template 100 LFE stack template
stacking
P-wave S-wave
17
Parkfield Tremor
Locations
• 88 stacked LFE templates• Located by P and S arrivals
on stacked waveforms, using a 3D velocity model.
• Sources extend 75 km both NE and SW of Parkfield
Shelly and Hardebeck, GRL, 2010
18
• 12+ years of continuous data
• ~50 trillion cross-correlation measurements
• ~850,000 events detected since mid-2001 (3000-30,000 per family)
• Detectible tremor activity in some area every day
Parkfield Tremor Catalog
Shelly and Hardebeck, GRL, 2010
Conclusion 1.1: The San Andreas fault does not end at the base of the “seismogenic zone.” Tremor sources are located on the deep extension of the fault, in the lower crust.
Conclusion 1.2: At least some portions of the deep fault deform brittlely. Tremor contains seismic waves of 30+ Hz even with temperatures ~500-600ºC
Conclusion 1.3: Tremor amplitude varies coherently along strike, possibly reflecting control from variations in geology and/or fluids.
What does it mean? (1)
Tremor Migration: characteristic velocity?
29
Recurrence: Shallower sources have larger, less frequent bursts
Cum
ulat
ive
even
ts
1 year
Cum
ulat
ive
even
ts
1 year
Cum
ulat
ive
even
ts
1 year
Cum
ulat
ive
even
ts
1 year
Shelly and Johnson, GRL, 2011