Slow Earthquakes and Nonvolcanic Tremor€¦ · tremor and LFEs, were considered to be distinct when they were discovered. Earthquakes happen impulsively and complete quickly—usually

EA39CH10-Beroza ARI 29 March 2011 20:17

Slow Earthquakes andNonvolcanic TremorGregory C. Beroza1 and Satoshi Ide2

1Department of Geophysics, Stanford University, Stanford, California 94305;email: [email protected] of Earth and Planetary Science, University of Tokyo, Bunkyo, Tokyo 113-0033,Japan; email: [email protected]

Annu. Rev. Earth Planet. Sci. 2011. 39:271–96

First published online as a Review in Advance onFebruary 14, 2011

The Annual Review of Earth and Planetary Sciences isonline at earth.annualreviews.org

This article’s doi:10.1146/annurev-earth-040809-152531

Copyright c© 2011 by Annual Reviews.All rights reserved

0084-6597/11/0530-0271$20.00

Keywords

subduction, hazards, tectonics, faults

Abstract

Nonvolcanic tremor is observed in close association with geodetically ob-served slow-slip events in subduction zones. Accumulating evidence pointsto these events as members of a family of slow earthquakes that occur asshear slip on the downdip extensions of fault zones in a regime that is transi-tional between a frictionally locked region above and a freely slipping regionbelow. By virtue of their locations and their properties, slow earthquakes arecertain to provide new insights into the behavior of earthquakes and faultingand into the hazard they embody.

271

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.


INTRODUCTION

Intermittently, over the past 50 years, earthquake scientists have uncovered evidence for a classof earthquakes that occur unusually slowly (Benioff & Press 1958, Kanamori 1972, Kanamori& Stewart 1979, Yamashita 1980, Sacks et al. 1981; Silver & Jordan 1982). Slow earthquakesgenerate waves inefficiently relative to ordinary earthquakes, particularly at short periods, suchthat their signature is relatively strong excitation of long-period relative to short-period seismicwaves. Silent earthquakes are on the extreme end of the spectrum in that they are aseismic anddo not generate detectable seismic radiation (Beroza & Jordan 1990). These characteristics havebeen used in studies that search systematically through continuous seismic data for anomalouslyslow events (Beroza & Jordan 1990, Shearer 1994, Ekstrom 2006).

Transient, slow postseismic slip is known to occur after large earthquakes (Thatcher 1975), andin some cases it is so extensive that it may exceed the seismic moment of the earthquake precedingit (e.g., Heki et al. 1997). There also is evidence for slow seismic slip before some earthquakes(Kanamori & Cipar 1974, Cifuentes & Silver 1989, Ihmle & Jordan 1994, McGuire & Jordan2000). These observations suggest that isolated slow earthquakes might occur as well.

Much of the early evidence for slow earthquakes came from sparse data sets—in many cases, onlya single recording. Within the past decade, a new generation of dense and sensitive earthquake-monitoring networks has led to a series of discoveries that have clearly revealed an entire class ofslow earthquakes ranging in size from M ∼1 to at least M 7.6. Many of the discoveries concerningslow earthquakes have been possible only because of the deployment of a new generation ofrelatively densely spaced and highly sensitive earthquake-monitoring networks. These includeboth high-sensitivity borehole seismometer arrays and continuously recording global positioningsystem (GPS) networks. Although there was a great deal of conjecture about what these networksmight reveal when they were being planned, it is fair to say that none of the discoveries reportedhere were accurately anticipated. Indeed, some of the discoveries, such as repeating slow-slipevents that happen with clock-like regularity, are more bizarre than anything we could haveimagined.

Several reviews have already been written on slow earthquakes and tremor (Schwartz &Rokovsky 2007, Gomberg et al. 2010, Rubinstein et al. 2010); however, we are still in the earlystages of studying and trying to understand these phenomena. Progress is rapid and accelerating.Our review covers how some of the important discoveries were made, offers a snapshot of ourcurrent state of knowledge, and outlines some of the outstanding research issues. The rate ofdiscovery is such that what follows will soon be outdated.

LOW-FREQUENCY EARTHQUAKES

In the aftermath of the destructive M 6.9 earthquake in Kobe in 1995, the Japanese governmentinstituted a policy in which seismic data supplied by many universities and institutes in Japanwere gathered into a unified nationwide database maintained by the Japan Meteorological Agency( JMA). At about the same time, and as a consequence of the Kobe earthquake, a nationwidehigh-sensitivity borehole seismic station network, Hi-net (Obara et al. 2005), was constructed andoperated by the National Research Institute for Earth Science and Disaster Prevention (NIED).The combination of the unified database and Hi-net coverage enhanced detectability of seismicevents by a factor of ∼5 (Figure 1). These developments, coupled with open data access and thehigh rate of tectonic activity in Japan, sparked much of the recent progress we report in this review.Among the first discoveries to come out of these enhanced monitoring efforts were low-frequencyearthquakes.

272 Beroza · Ide

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.


0

1995 Kobe

Hi-netopen

Number of LFEs

Number of earthquakes

a

b

c

0

105

1042×105

1990 1995 2000 2005 2010

2002−2004

1992−1994

130˚ 132˚ 134˚ 136˚ 138˚ 140˚32˚

34˚

36˚

130˚ 132˚ 134˚ 136˚ 138˚ 140˚32˚

34˚

36˚

Figure 1(a) The number ofordinary earthquakes(blue bars, left axis) andlow-frequencyearthquakes (LFEs)(red circles, right axis)recorded in the JapanMeteorologicalAgency catalogbetween 1990 and2009. (b) Hypocenterdistribution (bluecircles) in western Japanfor 1992–1994. Circleradius scales withmagnitude. Shallowevents (depth <50 km)are shown. (c) Samedata as panel b for2002–2004; LFEs areshown in red.

www.annualreviews.org • Slow Earthquakes, Nonvolcanic Tremor 273

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.


LFE: low-frequencyearthquake

Network analysts, who pick arrival times from seismograms for routine hypocenter determi-nation by the JMA, accomplished the first discovery of unusual events from the next generationof earthquake-monitoring networks. A rule in place for routine work of the JMA prohibitedhypocenter determination for events without clear P waves; however, seismogram readers foundevents that could be located through their discernable S waves, even though they lacked clear Pwaves. The waveforms of these events have low amplitude and look “noisy” compared with ordi-nary earthquakes, but they show clear arrivals across many stations, and their dominant frequencyis clearly low compared with that of ordinary earthquakes. Some of these events occurred nearvolcanoes, but many were also found in the Tokai and Kii regions of the Nankai Trough, wheremegathrust earthquakes of M 8 have occurred in the past and are anticipated in the future. Despitethe fact that the nature of these events was unclear, the JMA decided not to ignore the many eventsthat were actually visible in the records. Instead, it identified them as a new class of events thatcame to be known as low-frequency earthquakes (LFEs), started to determine their hypocenters,and catalogued them (Nishide et al. 2000, Katsumata & Kamaya 2003). Although these unusualevents were discovered independently, LFEs would come to play a critical role in deciphering themechanism of deep, nonvolcanic tremor.

NONVOLCANIC TREMOR

Deep, nonvolcanic tremor is a weak but persistent shaking of Earth that was discovered in theNankai Trough subduction zone of Japan (Obara 2002). Dr. Kazushige Obara at the NIED dis-covered these new signals in Hi-net seismograms when checking the performance of the network.He was monitoring summary records, which are sets of thin waveforms of many stations alignedin time, and found continuous noise-like waveforms common to many stations. The signals werenot impulsive like those of regular earthquakes, and routine earthquake-location methods couldnot be applied to locate their origins. He developed a hypocenter determination code using thetiming of waveform envelopes and determined that the signals originated from deep within Earth,following a belt-like zone along the 30–35-km depth contours of the Philippine Sea plate (Obara2002). The spatial distribution was quite similar to that of the LFEs, but the two phenomena,tremor and LFEs, were considered to be distinct when they were discovered.

Earthquakes happen impulsively and complete quickly—usually in a matter of seconds. Deeptremor, conversely, shakes Earth for hours, days, or even weeks at a time. It is most prominent inthe 1–10-Hz frequency band. Deep tremor was termed nonvolcanic because there is a differentkind of tremor, volcanic tremor, that has been recognized and monitored for decades. As thename implies, volcanic tremor occurs in magmatic systems and is thought to be generated bythe coupling of fluids moving through a volcano’s magmatic plumbing system to the solid Earth(Chouet 1985, Julian 1994). Nonvolcanic tremor also is distinguished from volcanic tremor in thatit lacks the distinct peaks in frequency that often characterize the latter. The moniker nonvolcanicwas meant to help distinguish the two phenomena, but it is a bit awkward, and deep, nonvolcanictremor is increasingly referred to simply as tremor or tectonic tremor.

Tremor has numerous interesting properties. Obara (2002) observed that tremor is episodic:Periods of concentrated activity of days to weeks interrupt periods of months with little or notremor. He also found that the locus of tremor slowly migrates along-strike at speeds of approx-imately 10 km day−1. It occurs frequently in southwestern Japan, where the Philippine Sea plateis subducting, but is absent in northeastern Japan, where the Pacific plate is subducting. Thislatter observation suggested a role for fluids because the tremor belt corresponds to the depths atwhich metamorphic dehydration reactions would release water in the subducting lithosphere inthe Philippine Sea plate, as was pointed out by Julian (2002), Obara (2002), and Seno & Yamasaki

274 Beroza · Ide

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.


SSE: slow-slip event

(2003). In northeastern Japan, the older and colder Pacific plate subducts at roughly twice therate, and the corresponding reactions should occur at greater depths, where they are far removedfrom the plate interface. A key role for fluids was also suggested by some similarities betweendeep tremor and volcanic tremor. Magmatic fluids play a central role in models of volcanic tremor(Chouet 1985, Julian 1994).

SLOW-SLIP EVENTS

Slow earthquakes have the same mechanism as do ordinary earthquakes—that is, they are causedby slip on faults. However, slow earthquakes share a common and distinguishing characteristic:They take a long time to happen relative to ordinary earthquakes. This is why we describe themas slow. Because they are slow, the waves they generate, if they generate any detectable wavesat all, are unusually weak relative to an ordinary earthquake of the same size. The Introductionrecounts a short history of studies that point to the existence of unusually slow earthquakes, butbecause the seismic signature of these earthquakes is weak, most of the convincing evidence forthem has emerged only recently—following the deployment of new highly sensitive earthquake-monitoring networks. An important exception to this is the series of small, slow earthquakesobserved on the San Andreas Fault near San Juan Bautista, California (Linde et al. 1996). Theseslow events were small, with a cumulative Mw of only 4.8, and very shallow, extending to 8-kmdepth at most; however, they were clearly slow, there were five distinct events, and they wereobserved on multiple instruments.

A new era in the study of slow earthquakes opened with the discovery of protracted slow-slipevents (SSEs) in subduction zones. Hirose et al. (1999) found anomalies in GPS data near theBungo Channel, which separates Shikoku and Kyushu, Japan. They modeled the GPS anomaliesas a slow earthquake on the subduction interface with a duration of 300 days and Mw of 6.6.Although there were nearby earthquakes during this episode, the slow earthquake occurred at adifferent location, and its evolution did not follow the typical profile of initially rapid, then rapidlydecaying, postseismic slip.

On the other side of the Pacific, Herb Dragert and his colleagues at the Geological Surveyof Canada independently discovered slow earthquakes in Cascadia. Here again, greatly improvedearthquake-monitoring networks were essential to progress. They deployed their first continuousGPS site at Albert Head (station ALBH) on Vancouver Island in Victoria, British Columbia, in May1992. In October 1994, data from this site showed an inexplicable 5-mm westward displacementwith respect to a reference site at Penticton (station DRAO) that accumulated over a period ofabout a week. At the time, three other continuous GPS sites that constituted their fledgling GPSnetwork showed no similar offsets, indicating that the anomalous motion occurred at ALBH only.Unfortunately, there were no other GPS stations close to ALBH to corroborate this as a regional(tectonic) signal. These results prompted Dragert’s team to question the stability of the GPSmonument, even though the concrete pillar was solidly anchored in highly competent bedrock.

Because they had so few GPS sites, they made it a practice to establish four or five survey markersin the immediate vicinity of each GPS monument to monitor monument stability. Shortly afterthe initial installation of the ALBH antenna pier in 1992, a “footprint” survey tying the GPSmonument to five local survey markers was carried out with a geodimeter. Because of the ob-served shift in 1994, this survey was repeated in 1996, and the results indicated that the GPSmonument had moved less than 0.5 mm with respect to the five reference markers within a 95%confidence range of 1.3 mm. The researchers concluded that the pier was stable, and the 5-mmdisplacement remained unexplained until 2000, when an identical displacement was resolved forALBH in the August 1999 GPS data. This time, improved GPS coverage, improved GPS satellite


Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.


orbits, and improved analysis techniques allowed them to identify clearly the transient surface dis-placements at seven contiguous GPS sites on Vancouver Island and in northwestern WashingtonState. The observed displacements were readily explained by slow slip on the deep plate interface.Dragert et al. (2001) modeled the signals as an Mw 6.7 event with 2 cm of slip over an extensive(50 × 300 km) area that unfolded over a period of 1 to 2 weeks. The event differed notably fromthe Bungo Channel event in that there were no significant earthquakes nearby that might havesomehow triggered it. A large SSE also was reported in Guerrero, Mexico (Lowry et al. 2001).Although this event was recorded only on a single GPS station, subsequent events with similarsignatures in the same area were recorded by many more (Kostoglodov et al. 2003).

Meghan Miller and her colleagues at Central Washington University had noted periodic stepsin the GPS record at ALBH and had tried to understand them. The task was difficult because thetime series were short, there was a big gap in the GPS data around the time of the Gulf War,and there were not many nearby stations with which to compare ALBH, but once Dragert’s teammade the case for the SSE in August 1999, they were convinced that there were many of them.Meghan Miller recalls saying to her colleague Tim Melbourne about the 1999 SSE, “if there isone, there are eight.” Miller et al. (2002) reported that slow earthquakes in the region occurredrepeatedly—a total of eight times from 1992 to 2001—with a remarkably steady recurrence intervalof approximately 14.5 months. Szeliga et al. (2004) and Brudzinski & Allen (2007) have shownthat the recurrence interval varies along the Cascadia margin, and recurrence in Japan is morevariable still (Hirose & Obara 2005, 2006). Nevertheless, the extraordinary periodic behavior ofmany SSEs contrasts strongly with the behavior of ordinary earthquakes, which are characterizedby recurrence that appears to be far more irregular (e.g., Ellsworth et al. 1999).

EPISODIC TREMOR AND SLIP

Because both slow earthquakes in Cascadia and deep tremor in Japan occurred periodically andat similar depths, and because both subduction zones involved the subduction of young, warm,oceanic crust, Julian (2002) speculated that the two phenomena might occur together. This pos-sibility was confirmed by Rogers & Dragert (2003), who discovered that the repeated slow earth-quakes observed by Miller et al. (2002) were accompanied by deep tremor, that the two coincidedboth temporally and spatially, and that when slow slip did not occur, tremor was absent. Whatfollows is an account of the process of their discovery as related to us by Herb Dragert and GarryRogers.

At the time of the discovery of the Cascadia slow slip, Garry Rogers and his colleagues atthe Geological Survey of Canada examined the earthquake catalog as well as analog local seismicrecords to see if there was any seismic activity correlated with the slip. They failed to find anyconvincing correlation; however, the report of deep tremor (Obara 2002) and discussions withObara at the 2002 Western Pacific Geophysics Meeting led them to suspect that they may nothave been looking at the correct seismic signals in their correlation attempts. The fact that Obara’snonvolcanic deep tremors in southwestern Japan were characterized by (a) tremor activity lastingfrom a few days to a few weeks, (b) tremor locations that were aligned along-strike of the subductionzone, (c) tremor depths averaging 30 km, and (d ) tremor migrating at speeds of 10 to 13 km day−1

along-strike of the subduction zone led Herb Dragert to conclude that tremors and slip must berelated because these listed attributes applied to Cascadia slow slip.

This led them to consider two immediate questions: First, does Cascadia have tremor activ-ity similar to Japan’s? Second, can a relationship to the observed repeated slips be established?They quickly established that the answer to the first question was a resounding “yes.” For years,seismologists at Canada’s Pacific Geoscience Centre had observed puzzling increases in levels

276 Beroza · Ide

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.


ETS: episodic tremorand slip

of background seismic noise on analog records from seismometers located on Vancouver Island.These amplified noise levels were similar to those generated by Pacific storms but occurred outsidestorm periods.

Garry Rogers and colleagues established the exact timing of slip events observed between 1996and 2003 from the GPS data. With this history of slip episodes, they examined daily helicorderrecords to look for increases in background noise. It was an exciting afternoon, as each date, in turn,confirmed elevated noise levels throughout the period of each and every slip event. Subsequentdetailed hourly plots of the seismic digital data revealed the coherent pulses of seismic “noise”propagating across the network identical to the pulses identified by Obara in Japan. It took longerto establish that these prolonged, multiday tremor episodes occurred only during times of slip.Once the temporal and spatial correlations were clearly established, they decided to term thesepaired phenomena episodic tremor and slip (ETS). Soon after, Obara et al. (2004) established asimilar relationship in southwestern Japan. Because the slip episodes in Japan are somewhat smallerthan those in Cascadia, the SSE signature was detectable in crustal tilt measurements rather thanas a reversal of GPS-measured surface deformation.

Although deep, nonvolcanic tremor and SSEs were clearly recognized as being closely associ-ated, the nature of their relationship was unclear for several years, and some lingering controversyremains. SSEs were modeled as shear slip on the deep extension of the plate boundary. As men-tioned previously, initial ideas about the origin of deep tremor focused on the notion that themovement of fluids somehow generated the tremor signal. A hypothesis put forward by Kao et al.(2005) and modeled by Liu & Rice (2007) demonstrated that an unclamping effect arising fromstatic stress changes induced by slow slip was consistent with Kao et al.’s tremor locations.

TREMOR AS A SWARM OF LOW-FREQUENCY EARTHQUAKES

We return now to LFEs. Shelly et al. (2006) found that LFEs in Japan occurred almost exclusivelyduring tremor episodes and used waveform cross correlation—both among LFEs and betweenLFEs and Wadati-Benioff zone seismicity—to locate them precisely. The combination of preciseP- and S-wave arrival times allowed them to develop precise locations that localized the LFEsto the plate interface. In a follow-up paper, Shelly et al. (2007b) used a match-filter approachto demonstrate that most, if not all, tremor could be matched with known LFEs from the JMAcatalog. This result, combined with the similarity in the spectral behavior, led to the obviousinterpretation that tremor is a swarm of LFEs and that these LFEs occur as shear slip. Thisinterpretation was confirmed when Ide et al. (2007b) demonstrated that LFE first motions andwaveform-based moment tensor inversions indicated plate-boundary slip. Further support for thishypothesis comes from Cascadia, where particle motion analysis of tremor as it passed under aseismic array indicated plate-boundary slip (Wech & Creager 2007).

At this time, the location of tremor in Cascadia remains somewhat controversial. Locationsbased on the source-scanning algorithm (Kao & Shan 2004), which is a form of Kirchhoff mi-gration, indicate a wide range of depths, with a somewhat higher concentration near the plateinterface (Kao et al. 2005). La Rocca et al. (2009) developed a new approach that extracted theS-P time using cross correlation of array measurements or tremor ground motion. Their mea-surements placed the source of tremor at, or very near, the plate interface. They also found that atleast some of the tremor they observed was composed of LFEs with discernable P and S arrivals.Brown et al. (2008, 2009) used a network autocorrelation technique to detect LFEs within tremorand found that they localized to the plate interface in subduction zones in Shikoku, Cascadia,and Costa Rica. Array analysis of tremor in Cascadia indicates that the tremor source is locatedat or near the plate interface (Ghosh et al. 2009, 2010b). Preliminary results from a multiarray


Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.


VLF: verylow-frequencyearthquake

experiment for the 2010 ETS event also points to a source along the plate interface (Ghosh et al.2010a).

VERY LOW-FREQUENCY EARTHQUAKES

Ito et al. (2007) reported another category of slow earthquake discovered in Japan: very low-frequency earthquakes (VLFs). These VLFs have magnitudes of approximately Mw 3–3.5 andcharacteristic durations of approximately 20 s. Because of this longer duration, VLFs are mostapparent as long-period waveforms recorded on sensitive broadband seismographs. There aretwo populations of VLFs: one that occurs in the shallow accretionary prism offshore (Ito & Obara2006), and another that occurs at the same time and location as do tremor and larger-scale slowearthquakes (Ito et al. 2007). The shallow VLFs are consistent with deformation of the accretionaryprism, and the deep VLFs are consistent with plate-boundary slip.

LFEs and VLFs occur together. If LFEs occur in close succession, sometimes (although notalways) the waves have low-frequency components that, in aggregate, amount to VLFs. Whereasthe occurrence of LFEs appears to be a necessary condition for VLFs, the opposite is not true.Tremor, LFEs, and VLFs are sometimes visible in one seismic record. Figure 2 shows typicalETS activity in western Shikoku. For the SSE, the signal is visible only in tiltmeter records.The tremor, LFEs, and VLFs are recorded by seismometers. If we magnify some part of thebroadband seismogram, these events are clearly visible at different timescales and over differentfrequency ranges. Thus, when we calculate the spectrum of broadband record of ground velocity,we observe an almost flat spectral amplitude from low frequencies of approximately 0.01 Hz (inVLF band) to 10 Hz (in LFE band). This flat spectrum is interrupted by the prominent backgroundmicroseism peak (Figure 3). This is different from ordinary earthquakes, for which the velocityspectrum falls above the corner frequency as f−1. Spectra in both Cascadia and Nankai show high-frequency fall-off above 1 Hz, but a significant amount of this fall-off can be explained by seismicattenuation. Within the observable frequency range, tremor spectra are flat, like a sequence of deltafunctions.

Ide et al. (2008) reported somewhat larger (Mw ∼ 4) and longer (up to 200 s) VLFs. They chosenot to add another name, such as ultra low-frequency earthquakes, to describe them because itwas becoming apparent that these events were part of a family of slow earthquakes that spanneda wide range in both seismic moment and characteristic duration.

THE SLOW EARTHQUAKE FAMILY

Once tremor was established as a swarm of LFEs, it became clear that these diversely namedphenomena occurred at the same time and the same place. Furthermore, all of them occurredby the same mechanism: shear slip. Thus, it is natural to consider them different manifestationsof a single process. Moreover, slow earthquakes share common scaling characteristics that aredistinct from those of ordinary earthquakes. Figure 4 shows that ordinary earthquakes grow inseismic moment with the cube of their duration but that slow earthquakes grow in seismic momentlinearly with their duration (Ide et al. 2007a). A similar relation between SSEs and duration was alsopointed out by Schwartz & Rokosky (2007) and Aguiar et al. (2009). This constant moment rate alsosuggests that slow earthquakes share a common underlying mechanism. The gap in observationsfor events with durations from ∼200 s to 1 day may be attributable to the observational limitsof seismology (due to increasing noise at periods beyond 200 s) and geodesy (due to the weakdeformation signal from events smaller than Mw ∼ 6.0) (Ide et al. 2008).

278 Beroza · Ide

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.


April–May 2004

14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 1

April–May 2004

April–May 2004

Time (s) from April 25 21:55

Time (s) from April 25 21:56:30

14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 1

14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 1

0

20

40

60

80

100

120

km

0

20

40

N

60

80

100

120

km

0.0

0

1

–1

0

0.2

0

0

0.2

0.02

0.2

0.4

0.6

UWAH.E

UWAH.N

TSA.BHZ

IKTH.E

IKTH.NSSE

Tremor

LFE

VLF

SSE

Tiltmeter

Seismometer (2–8 Hz)



Seismometer (20–50 s)

μm s–1

μm s–1

μm s–1

μrad

0 60 120 180

0 10 20

P?S

a

cd

fe

b

Figure 2Typical episodic tremor and slip (ETS) activity in western Shikoku. (a) Space-time plot and (b) map view of tremor sources color codedby date (Ide et al. 2010). The vertical axes are common for panels a and b. The modeled slow-slip event (SSE) fault plane (Hirose &Obara 2005) is shown by a rectangle in panel b. (c) Tiltmeter records at two Hi-net stations (Hirose & Obara 2005). (d ) Seismometer inhigh-frequency (2–8 Hz) and low-frequency (20–50 s, 0.02–0.05 Hz) ranges for three minutes. The time is shown in panels a and e by ared line. (e) Seismometer records at Nishitosa (station TSA) maintained by the National Research Institute for Earth Science andDisaster Prevention. ( f ) Seismometer in high-frequency (2–8 Hz) range for 20 s. The time is shown in panel d by a red rectangle.Abbreviations: LFE, low-frequency earthquake; P, arrival time of P wave; S, arrival time of S wave; VLF, very low-frequencyearthquake.


Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.


−200

−180

−160

−140

−120

10–5

103

102

101

100

100

100 101

101

10–1

10–2

10–6

10–7

0.01 0.1 1 10Frequency (Hz)

Frequency (Hz)

Frequency (Hz)

Earthquakes M > 1Earthquakes M ≤ 1LFETremorBackground noise

a b

c

Noise

Tremorand VLF

MicroseismPS

D (

dB

, m2 s

–2 H

z–1)

Am

pli

tud

e (

m s

–1 H

z–1)

No

rma

lize

d a

mp

litu

de

Earthquakes, ML = 2.5–3.5Earthquakes, ML = 1.0–1.5Earthquakes, ML = 0.0–0.2

Largest tremorsBest-located tremorsBackground noise

Figure 3Examples of velocity spectra. (a) Broadband tremor record in the Kii Peninsula, Japan (after Ide 2008).(b) Low-frequency earthquake (LFE) and tremor in western Shikoku, Japan (Shelly et al. 2006). (c) Tremorin Cascadia (Kao et al. 2005). In each figure, the yellow area shows attenuation curves using exp(−π ft/Q), atravel time of t = 15 s, and Q in the range of 200–500. Abbreviations: M, Japan Meteorological Agencymagnitude; ML, local Richter magnitude; PSD, power spectral density; VLF, very low-frequency earthquake.

Further evidence that slow earthquakes form a separate population comes from the observationthat stress drops (Brodsky & Mori 2007) and scaled energy (seismically radiated energy divided byseismic moment) of slow earthquakes are orders of magnitude lower than those for ordinary earth-quakes (Ide et al. 2008). Despite observed small values of the scaled energy for slow earthquakes,it seems to be constant for slow earthquakes, at least regionally. Hirose et al. (2010) utilized aconstant related to scaled energy to measure long-term slip rate on the plate interface in Shikoku.

In an interesting twist on the slow earthquake scaling relation, Meade & Loveless (2009) arguethat if the relationship is scaled to great (Mw > 8) slow earthquakes, the slip velocity may neverexceed the plate convergence rate and thus may never register geodetically as a reversal of surface

280 Beroza · Ide

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.


1010

100 105 100 105

1015

1020

1010

1015

1020

1025

Mo

me

nt

(N-m

)

Duration (s) Duration (s)

Mo

me

nt

rate

(N

-m s

–1)

Figure 4Regular versus slow earthquake scaling. (Left) Moment (M0) versus duration τ ; (right) moment rate (Mo)versus duration τ . Both graphs plot ordinary earthquakes ranging in size from Mw 1.0 to 9.5 (blue triangles)and slow earthquakes ranging in size from Mw 1.5 to 7.5 (red circles). Plot on left shows moment for ordinaryearthquakes, M0 ∼ τ 3 (slope of 3). Plot on right shows moment rate for slow earthquakes, Mo ∼ τ (slope of1, or constant moment rate). Sources: Thatcher & Hanks 1973; Fletcher et al. 1984; Cocco & Rovelli 1989;Abercrombie 1995; Houston 2001; Prieto et al. 2004; Ide et al. 2007a, 2008.

deformation. Instead, such events might last for many decades and manifest themselves as apparentpartial elastic coupling, rather than as deformation transients.

THE GEOGRAPHY OF SLOW EARTHQUAKES

ETS has been found in many places. In the Nankai subduction zone of Japan, tremor usually occurstogether with SSEs, and their locations are similar. The exceptions to this are SSEs of extendedduration, such as the 2000–2003 Tokai SSE (Miyazaki et al. 2006) and the Bungo Channel SSEs in1997 and 2003 (Hirose & Obara 2005). The shallow parts of these long-term SSE slip zones do notshow tremor. In Cascadia, tremor occurs for more than 1,000 km following the iso-depth contoursof 30–40 km of the subducting Juan de Fuca plate. The extent of tremor in the dip direction iswider than that for Nankai but still less than 80 km (e.g., Gomberg et al. 2010), whereas theestimated SSE areas (Szeliga et al. 2008) are widely distributed both upward and downward ofthe tremor band. In Alaska, the station coverage is not as dense, but slow slip occurs next to theshallower locked area beneath the accretionary prism (Ohta et al. 2006), whereas tremor is locatednear the lower edge of the SSE area (Peterson & Christensen 2009). The above three areas havewide accretionary prisms and moderate relative plate velocity of 3–7 cm s−1. The subducting slabhas been strongly deformed in the Nankai Trough and to some extent in Alaska and Cascadia.There may be a correlation between SSE locations and places that feature higher slab curvature.

Whereas SSEs in Japan, Cascadia, and Alaska occur far (∼200 km) from the trench axis, SSEsin Mexico (e.g., Song et al. 2009) and Costa Rica (Outerbridge et al. 2010) are much closer tothe trench axis and at a similar depth to that of megathrust earthquakes (Figure 5). In theseareas, tremor is separated from the SSE areas and slightly deeper than in Nankai and Cascadia. InNew Zealand (McCaffrey et al. 2008, Delahaye et al. 2009) and Boso (Ozawa et al. 2007), SSEshave been discovered, but with an apparent lack of detectable tremor. Since these studies werepublished, however, there have been preliminary reports of tremor in New Zealand (A. Wech,personal communication). SSEs in New Zealand are also distributed over a range of depths.

Because our observational ability is strongly limited in most subduction zones, we do not knowthe systematics of where tremor does and does not occur. Nevertheless, in northeastern Japan,


Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.


10

20

30

40

50

60

70

km

020

0

40

0 km

020

0

40

0 km

020

0

40

0 km

020

0

40

0 km

020

0

40

0 km

020

0

40

0 km

02

00

40

06

00

km

−156°

−154°

−152°

–150°

−148°

−146°

−144°

−142°

58°

58°60°

02

00

40

06

00

80

01,

00

01,

20

0k

m

–12

8°

–12

6°

–12

4°

40°

42°

44°

46°

48°

50°

Ala

ska

Ca

sca

dia

Na

nk

ai

Ne

w Z

ea

lan

dM

ex

ico

Co

sta

Ric

a

Tre

mo

rS

SE

Co

nto

urs

o

f p

late

in

terf

ace

sM

eg

ath

rust

ear

thq

uak

e5

cm

ye

ar–

1

02

00

40

06

00

80

01,

00

0k

m

130° 132°

134°

136°

138°

140°

32°

32°

34°

34°

Bo

so

02

00

40

06

00

km

−102°

−100°

−98°

18

°

20

°

Gu

err

ero

Oa

xa

ca

02

00

40

06

00

km

174°

176°

178°

180°

–42°

−42°

−40°

02

00

40

06

00

km

−88°

−88°

−86°

−84°

8°10

°

12°

19

64

05

01

00

15

0O

cea

n-fl

oo

r a

ge

(M

Pa

)

Co

sta

Ric

a

Me

xic

o

Ne

wZ

ea

lan

d

Na

nk

ai

Ca

sca

dia

Ala

ska

0°

30

°6

0°

90

°1

20

°1

50

°1

80

°−

15

0°

−1

20

°−

90

°−

60

°−

30

°0

°−

60

°

−3

0°

0°

30

°

60

°

Sh

iko

ku

Kii

To

ka

i

19

23

Ka

nto

19

46

Na

nk

ai

19

44

T

on

an

ka

i

282 Beroza · Ide

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.


where the Pacific plate is subducting and observation conditions are the same as those in westernJapan, there is no tremor observed, although shallow VLFs (Asano et al. 2008) and afterslip (e.g.,Heki et al. 1997) clearly occur there. This difference has been attributed to the different thermalstructure of the subducting plate and the presence/absence of metamorphic dehydration reactionsat the depth of tremor ( Julian 2002, Obara 2002). As shown in Figure 5, the five subduction zonesthat are known to host tremor are located where young (<50 Ma) plate is subducting. Plates innortheastern Japan and New Zealand are much older. Plate boundaries involving subduction ofrelatively young lithosphere are not particularly common. Places similar to those mentioned aboveare Ecuador and southern Chile, and these should be considered likely spots for tremor to occur.

Tremor is not limited to subduction zones. Nadeau & Dolenc (2005) detected tremor in centralCalifornia under the San Andreas Fault near Parkfield. The frequency content is similar; it alsooccurs spontaneously without being triggered by waves from distant earthquakes, and it appears tooccur on the deep extension of the plate boundary—in this case, the San Andreas Fault (Shelly et al.2009). Possible sources of fluids exist, although they may be less obvious (Kirby et al. 2002). Shelly(2010a) identified thousands of LFEs in the lower crust over a ∼100-km stretch of the San AndreasFault, which strongly suggests that the fault is localized throughout the crust. He found evidencefor tremor migration along-strike at 15–80 km h−1 and showed that increased tremor activitynear Parkfield in the 3 months before the 2004 M 6 earthquake could have heralded acceleratedslip on the fault. Shelly & Hardebeck (2010) found that the depth of the LFEs constituting thetremor ranges from 18–28 km and spans both the creeping and locked segments of the centralSan Andreas Fault (Figure 6).

The true geographic extent of tremor and slow earthquakes is incompletely known. LFEs havebeen reported in the Apennines of Italy (Piccinini & Saccorotti 2008). Tremor-like swarms of tinyearthquakes were accidentally recorded during an active-source experiment in the Reelfoot Riftof the central United States (Langston et al. 2010). Triggered tremor has been observed over awide range of faults in California (Gomberg et al. 2008, Peng et al. 2009) and under the CentralRange of Taiwan (Peng & Chao 2008). Tremor and LFEs also have been observed deep beneaththe aftershock zone of the 2000 Mw 6.7 Western Tottori earthquake (Ohmi & Obara 2002, Ohmiet al. 2004). There have been sporadic reports of possible tremor in other places, as well. For thevast majority of Earth, we lack the locally recorded data required to discern tremor, so what weknow of its distribution must be considered incomplete.

TRIGGERED TREMOR

Miyazawa & Mori (2005) reported that the 2003 Mw 8.1 Tokachi-oki earthquake in northern Japantriggered LFEs in the Nankai subduction zone. The signal was modulated by long-period surface

←−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−Figure 5Tremor and slow-slip events (SSEs) in the world’s subduction zones. The world map shows colored ocean-floor age (Muller et al.2008), plate boundaries (dark blue lines, Bird 2003), and six subregions, for which individual detailed maps are shown. The references areas follows. Alaska: tremor (Peterson & Christensen 2009); SSE, the 1964 rupture area, and plate interface (Ohta et al. 2006). Nankai:tremor shown by low-frequency earthquakes determined by the Japan Meteorological Agency, SSEs (Hirose & Obara 2005, Miyazakiet al. 2006, Ozawa et al. 2007, Sekine et al. 2010); 1944 Tonankai earthquake (Kikuchi et al. 2003); 1946 Nankai earthquake (Sagiya &Thatcher 1999); and plate interface (Ishida 1992, Ide et al. 2010). Cascadia: tremor (Kao et al. 2009, Boyarko & Brudzinski 2010,Gomberg et al. 2010); SSE (Szeliga et al. 2008, Gomberg et al. 2010); and plate interface (McCrory et al. 2006). New Zealand: SSE andplate interface (McCaffrey et al. 2008). Mexico: tremor (Payero et al. 2008), SSE (Yoshioka et al. 2004), and megathrust events,analyzed by Song et al. (2009). Costa Rica: tremor (Brown et al. 2009); SSEs and megathrust events compiled by Outerbridge et al.(2010). Plate interfaces of Mexico and Costa Rica are from Slab1.0 (Hayes et al. 2009).


Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.


Parkfield

V04C

PPO

0 10 20 30 40 50 km10

120.5° W 120.0° W 121.0° W

35.5° N

36.0° N

36.5° N

PCA

PHOB

PCM

PSR

PSNPMR

PMCPAG

V05C

PBS

PSC

B079

PPG

B901

B900

PIRPBP

PHL PBM

YEG

SMMCAR

BPI

PLO

PHR

PCC

PMPB

PDR

PJU

PSA

PAR

PSMPJC

FROB

PWK

BVL BBGB

BBN

BRV

PBW

PPT

A

D

C

B

BBSan Benito

Bitterwater

Monarch Peak

Parkfield

Cholame

Simmler

–80 –60 –40 –20 0 20 40 60

30

20

10

0

De

pth

(k

m)

Along-fault distance (km)

Bitterwater Parkfield CholameNW SE

0

20

40

Slip

(cm)

Creeping Locked

A

D

C

B

50

40

30

20

10

0

Filte

red

am

plitu

de

(nm

s–

1)

SMR

Figure 6Map and cross section of tremor family locations and amplitude. (Top) Map view of low-frequencyearthquake/tremor locations (colored circles) and seismic stations used for location (triangles). Filled trianglesindicate borehole stations used for event detection. Open circles, plusses, and diamonds show interpretedtremor family locations from Shelly (2009), for reference. Gray dots show relocated microseismicity from1984 to 2003. Red star indicates hypocenter of 2004 M 6 earthquake. Black lines indicate faults. (Bottom)Along-fault cross section. Microseismicity within 5 km of the fault trace is in blue. Other symbols are thesame as in map view. Gray shading shows the coseismic slip and first 230 days of postseismic slip from the2004 earthquake. After Shelly & Hardebeck (2010).

284 Beroza · Ide

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.


waves and was strongest in the Bungo Channel, which was the site of an ongoing SSE. Miyazawa& Mori (2006) and Miyazawa et al. (2008) found that tremor in Japan was triggered by passingRayleigh waves, and they invoked induced pore pressure changes from dilatational deformationto explain them. Conversely, Rubinstein et al. (2007) found that tremor in Cascadia was triggeredby the passage of long-period Love waves from the 2002 Mw 7.8 Denali, Alaska, earthquake. Thetremor was triggered in synchrony with times when the Love waves were increasing shear stressacross the plate interface in a direction that would promote plate-boundary slip. Peng & Chao(2008) found that tremor under the Central Range of Taiwan was also triggered by Love waves,in that case from the 2001 Mw 7.8 Kunlun earthquake.

In addition to being triggered by earthquakes, tremor is strongly affected by tides. Shelly et al.(2007a) found that LFEs had a strong tidal signature at approximately 12.4 h for two tremorepisodes in Shikoku. They also observed that tremor amplitudes followed this same pattern.Rubinstein et al. (2008) found that three tremor episodes in Cascadia showed a pulsing of ac-tivity with tidal periods of 12.4 and 24–25 h, which correspond to the lunar and lunisolar tides,respectively, and they concluded that tremor must occur on very weak faults. Nakata et al. (2008)modeled changes in the Coulomb failure stress that resulted from tides and found that tremor wasadvanced by a few hours relative to the change in Coulomb stress. This argues against a simpleCoulomb failure model, and they interpreted tidal modulation of tremor in the context of a rate-and state-variable friction model of seismicity (Dieterich 1994). They found that tremor laggedpeak stress-rate changes by several hours and that the observations were consistent with low valuesof effective normal stress. They also reported preliminary results of tremor that did not appear torespond strongly to tidal forcing. This difference of sensitivity to tidal tremor is correlated withthe typical duration of tremor, suggesting a spatial difference of frictional properties on the plateinterface (Ide 2010a). Lambert et al. (2009) modeled stress on the plate interface and found thattremor activity peaked when the tidally induced shear stress favored plate-boundary slip. Theyalso discerned a daily, nontidal variation of unknown origin. Although the slow-slip signal is lesssignificant compared with tremor, Hawthorne & Rubin (2010) found evidence that it also is mod-ulated by tidal stress. Thomas et al. (2009) modeled tidal stresses for tremor on the deep extensionof the San Andreas Fault and used rate- and state-variable friction modeling to conclude that thefault responded to exceptionally low values of tidally induced shear stress. This, too, is consistentwith a weak and nearly critically stressed fault.

CHARACTERIZATION OF TREMOR

Regular earthquakes obey Gutenberg-Richter statistics, in which the log cumulative number ofevents is proportional to magnitude. What are the statistics of tremor? Tremor differs in behaviorfrom ordinary earthquakes, and the answer depends on how we define the size of tremor. On theone hand, if we measure the duration of tremor exceeding some threshold amplitude, the durationis proportional to an exponential function of the amplitude measured as the absolute value of eitherground displacement or velocity (Watanabe et al. 2007, Ide 2010b). On the other hand, if we defineevents as a cluster of tremor sources occurring close to one another and assume that the seismicmoment is proportional to the duration of the cluster (Ide et al. 2007a, Aguiar et al. 2009), thecumulative number seems to obey a Gutenberg-Richter relation with b = 1.0 (Creager et al. 2009).These two measures appear to contradict each other, and we need to understand how they can bereconciled. Spatial inhomogeneity of the tremor source likely plays a key role in this explanation.

Precise location of tremor reveals some interesting characteristics in its spatial distribution.Tremor does not occur homogeneously. In Cascadia, tremor during major ETS events tendsto occur in the shallower part of the tremor belt. Many smaller tremor swarms occur in the


Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.


deeper portions of the tremor belt during the inter-ETS period (Wech & Creager 2008). InNankai, Obara (2010) pointed out two lineations of tremor sources separated in the dip directionof subducting plate. These along-dip heterogeneities may be attributable to changes in pressureand temperature conditions. Tremor distribution is also highly heterogeneous in the along-strikedirection. Striations in the plate-subduction direction have been identified in Cascadia (Ghoshet al. 2010b) and Nankai (Ide 2010a). Subduction of heterogeneous structures, such as oceanicseamounts, provides a viable origin for these striations, which in turn suggests that the fine structureof the tremor distribution records and preserves subduction history over several million years (Ide2010a).

Shelly (2010b) has identified families of LFEs that exhibit remarkable behavior, including oneLFE that occurs every 3 days—except when its period doubles to 6 days. He identified ∼540,000LFEs for the 8.5-year period for which continuous seismic data are available (mid-2001 through2009). This is ∼2.5 times as many earthquakes (216,000) as those listed in the entire NorthernCalifornia Seismic Network catalog during that same time period. Shelly (2010b) documentedevents that exhibited tightly clustered recurrence intervals of ∼3 or ∼6 days, with abrupt transitionsbetween the two behaviors (Figure 7). It is unclear whether deep tremor in subduction zones—or,for that matter, ordinary earthquakes—might exhibit similar behavior, but getting a clearer pictureof the long-term recurrence behavior of the entire spectrum of slow earthquakes is an importantresearch goal.

CONTROLS ON TREMOR AND SLOW EARTHQUAKE OCCURRENCE

Fluids have been implicated in nonvolcanic tremor occurrence since the original discovery paper(Obara 2002). This is likely due in part to the similarity with volcanic tremor, which is thought tobe generated by fluid motion (Chouet 1985, Julian 1994), and in part from the observation thattremor occurs in southwestern Japan but not northeastern Japan. The latter finding is consistentwith the pressure and temperature conditions in those two subduction zones, and their controlof the depth at which water is liberated due to phase changes in the subducting slab (Peacock &Wang 1999, Julian 2002, Yoshioka et al. 2008, Peacock 2009). Direct evidence for high pore-fluidpressure in the region of tremor activity in Shikoku is apparent in tomographic images of P-and S-wave velocities (Shelly et al. 2006, Matsubara et al. 2009). Kao et al. (2005) found a widedistribution of tremor depths that correlate with independently imaged, highly reflective zones.Their interpretation is that tremor may be generated by fluid motions at these locations. Audetet al. (2009) used receiver functions to show that in Cascadia, tremor is associated with a narrowlow-velocity zone with Poisson’s ratio as high as 0.4 in the subducted oceanic crust. This stronglysuggests high, nearly lithostatic, pore-fluid pressure. Yoshioka et al. (2008) found that LFE activityin the Nankai Trough correlates well with slab dehydration rates found from detailed modelingof the subduction zone. This could explain why LFEs are relatively rare in Kyushu.

Although fluids appear to play a critically important role in deep tremor, other evidence pointsto the importance of additional factors. A recent structural survey in the Tokai region indicatesthat the upper bound of tremor is close to the Moho discontinuity (Kato et al. 2010). The Mohoplane appears to divide the slow-slip area in the Tokai area into a deeper tremor zone and ashallower, longer-term SSE region. This suggests that the material in the overriding plate playsan important role in the distribution of tremor and slow slip. The material in the overriding plateabove the tremor zone is wedge mantle, which is stagnant for a long time and suffers ongoingmetamorphism as suggested by Wada et al. (2008). Ide (2010a) observed sets of parallel striations intremor hypocenters with two orientations: one reflecting the current convergence direction of thePhilippine Sea plate, and the other reflecting the convergence direction that was operative several

286 Beroza · Ide

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.


San

Sim

eo

n E

Q

2002 2003 2004 2005 2006 2007 2008 2009 20100

2

4

6

8

10

Year

Tim

e s

ince

pre

vio

us

ev

en

t (d

ay

s)

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

Ma

xim

um

am

plitu

de

(no

rma

lized

)

2003.4 2003.6 2003.8 2004 2004.2 2004.4 2004.6 2004.8 20051

2

3

4

5

6

7

Year

Tim

e s

ince

pre

vio

us

ev

en

t (d

ay

s)a

b

Par

kfie

ld E

Q

San

Sim

eo

n E

Q

Ne

two

rk c

han

ge

s

Ne

two

rk

chan

ge

s

Ne

two

rk

chan

ge

s

Par

kfie

ld E

Q

Figure 7Recurrence interval versus time for a low-frequency earthquake (LFE) family south of Parkfield, California.(a) Recurrence history, mid-2001 to 2010. Colors indicate peak ground velocity of each burst, normalized tothe first percentile amplitude of all events in the family. Gray lines connect consecutive events. (b) Magnifiedview of part of the record in panel a. Note the distinct change in recurrence behavior after the 2004 Mw6 Parkfield earthquake. From Shelly (2010b). Abbreviation: EQ, earthquake.

million years ago. The persistence of the signature of past plate motion strongly suggests thatdifferences in material properties—due, for example, to metamorphism of subducting seamounts—play an important role in controlling tremor occurrence.

PHYSICAL MODELS FOR SLOW EARTHQUAKES

Numerous studies have modeled slow earthquakes using elastic material, and friction laws as theboundary condition on the fault interface. This traditional method for modeling cycles of ordinaryearthquakes was introduced into numerical modeling using a rate- and state-variable friction law


Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.


(Dieterich 1979, Ruina 1983) by Tse & Rice (1986). This pioneering study pointed out theexistence of “fast episodic creep events” for relatively large slip-weakening distance. Isolated SSEscan be modeled in this manner. After the discovery of a long-term SSE in Tokai (Ozawa et al.2002, Miyazaki et al. 2006), Kuroki et al. (2004) simulated the recurrence of slow slip using largeslip-weakening distance, with 3D complex plate interface structure.

Modeling short-term SSEs on the deep extension of megathrust faults requires additional com-plexity. Kato (2003), Shibazaki & Iio (2003), and Shibazaki & Shimamoto (2007) used friction lawswith a velocity cutoff, where friction changes from velocity weakening to velocity strengthening.Liu & Rice (2005) questioned the appropriateness of such a velocity cutoff and demonstrated thatspontaneous, aseismic slip events repeat near the downdip end of the seismogenic zone in their3D model. Liu & Rice (2007) and Rubin (2008) investigated conditions for SSEs in numericalmodels with rate- and state-variable friction laws. A large slip-weakening distance and low valuesof effective stress are common features of almost all models. Such models emulate an SSE as,essentially, a large-scale nucleation event without an ensuing instability.

In these models, the size of an SSE is determined mostly by geometry of the fault and the slip-weakening distance. The size is characteristic if the slip-weakening distance is homogeneous. Effec-tive stress also affects the size, but changing the stress by orders of magnitude is more difficult thanchanging the length scale. By changing the shape of the slow-slip zone, Shibazaki et al. (2010) pro-duced complex patterns of slow slip. The spatiotemporal slip pattern resembles the observed pat-tern in regions of the Nankai subduction zone (Figure 5). Unlike SSEs, smaller slow earthquakeshave not been well explained by fully dynamic models. Shibazaki et al. (2010) introduced a strongpatch in their slow-slip model to simulate VLFs. Ando et al. (2010) modeled LFEs and tremorusing a distribution of small asperities triggered by background slow slip. Developing models thatcan account for the swarm-like nature of tremor is likely to be an area of active future research.

Another approach to modeling SSEs posits that they arise from dilatant stabilization. Segall et al.(2010) have explored this possibility. In their model, dilatancy-induced pore-space leads to a dropin fluid pressure, which in turn increases the effective normal stress and inhibits the instability. Theassociation of SSEs with areas of high pore-fluid pressure supports this possibility. In their model,the competing effects of dilatancy and pore-fluid flow lead to a limit on slip and rupture speed.

Statistical features of tremor can be modeled kinematically by a simple stochastic model, theBrownian slow earthquake model (Ide 2008, 2010b), in which the tremor zone is approximatedby a circular area with a radius governed by the Langevin equation. This model can explain(a) the linear relationship between seismic moment and duration (Figure 3; Ide et al. 2007a),(b) nearly constant scaled energy for VLFs (Ide et al. 2008), (c) the observed flat velocity spectrum(Figure 2), and (d ) amplitude-duration statistics approximated by an exponential function (Watan-abe et al. 2007). Because the Langevin equation is a stochastic representation of the diffusionequation, it is also suitable for explaining the scale-dependent migration velocity of tremor (Shellyet al. 2007a,b; Ito et al. 2007) and diffusional migration (Ide 2010a). So far, the model is toosimple to describe the entire spectrum of ETS activity, including both long-range migration andsensitivity to tidal stress. Nevertheless, future comprehensive models of ETS should include someof the stochastic aspects of the Brownian slow earthquake model.

RELATION OF SLOW EARTHQUAKES TO REGULAR EARTHQUAKES

Because many of the slow earthquakes discussed above occur on the deep extensions of majorfaults, they will inevitably increase the stress on the shallower parts of the faults that are capable ofrupturing in large, damaging earthquakes (Dragert et al. 2004). How likely are tremor and slow slipto trigger nearby earthquakes? Of the SSEs known to have occurred, none have been accompanied

288 Beroza · Ide

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.


by a large earthquake; however, some have triggered small, nearby earthquakes (Segall et al. 2006,Liu et al. 2007). Mazzotti & Adams (2004) have calculated that SSEs on the deep extension ofthe plate interface in Cascadia may increase the conditional probability of a great earthquake by afactor of 30 to 100 over the background probability during the 2-week period in which they occur.Despite this dramatic increase, the probability during an ETS episode remains low because of thelong recurrence interval. The fact that the entire Cascadia subduction boundary could be the siteof nucleation of the next great earthquake, combined with the variable ETS recurrence over thatsame area, reduces the predictive value further.

The strong sensitivity of tremor to triggering by tides and surface waves of distant earthquakesindicates that it may be quite sensitive to changes in the state of stress at depth. There is thepotential, then, that there will be measurable changes in the behavior of tremor during the seismiccycle. There is a hint of such changes to tremor in the migration direction of LFEs under the SanAndreas Fault before the 2004 Parkfield earthquake (Shelly 2009), and certainly the Parkfield andnearby San Simeon earthquakes had a strong effect on the behavior of tremor afterward (Nadeau& Guilhem 2009, Shelly 2009). Although a possible relationship between tremor and/or slowearthquake behavior and megathrust rupture is intriguing, at this point it remains speculative.

Other important aspects of megathrust earthquake behavior have the potential to be informedby tremor and slow earthquakes. The Nankai Trough subduction zone presents a unique sit-uation with a combination of precisely located tremor over multiple tremor episodes and reli-able information on the slip distribution of previous megathrust earthquakes: the 1944 Mw 8.1Tonankai and 1946 Mw 8.1 Nankai events. As shown in Figure 5, the downdip edge of the slipdistribution in these megathrust earthquakes is outlined by the tremor and slow earthquake belt.This complementary distribution of tremor and large earthquake slip is sensible in that these slowearthquakes occur in the transitional region, but it also means that by monitoring and locatingslow earthquakes, we can identify the likely landward extent of future megathrust earthquakes.

In Cascadia, which is now known to host extremely large earthquakes (Atwater 1987) andwhere the last megathrust earthquake is thought to have occurred in 1700 (Satake et al. 1996), thedowndip extent of the next large earthquake is uncertain. The downdip extent of rupture has beenestimated through a combination of geodetic data and thermal modeling (Hyndman & Wang1995). If the same relationship between the slow earthquake belt and megathrust rupture holds inCascadia as it does in Japan, however, the rupture in the next megathrust earthquake would reach60 km inland of the Pacific coast. This is much closer to the major urban areas of the region thanthe results of Hyndman & Wang (1995) imply. Also, because the rupture would no longer stopoffshore, the potential size of future earthquakes is substantially larger than previously anticipated(Chapman & Melbourne 2009). This interpretation of the slipping-to-locked transition implied bythe ETS zone is consistent with GPS data (Chapman & Melbourne 2009). Furthermore, becausethe tremor zone responds to stresses as small as Earth tides and waves from distant earthquakes,it may be driven to slip relatively rapidly during large megathrust earthquakes. The tremor bandof 1–10 Hz is also the band of primary earthquake-engineering interest for most structures.

In other areas shown in Figure 5, including Mexico and Alaska, the dichotomy between theETS/SSE zone and the locked zone is less clear, and there may be 3D (along-strike) variationsthat are important. Finally, under the San Andreas Fault, tremor occurs at depths ranging from18 to 28 km (Shelly & Hardebeck 2010). If the same relationship between tremor and the lockedzone holds there, then large earthquake slip on the San Andreas Fault may extend to depthsbeyond the 15 km that is usually assumed. The assumption that large earthquake slip and ETSare complementary remains conjectural; however, because this topic is important for quantifyingseismic hazard, it will be receiving the attention of many scientists.


Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.


CONCLUSIONS

The discovery and rapidly growing understanding of slow earthquake behaviors has opened a newera in earthquake science. Slow earthquakes provide new and fundamental information on howfaults operate. By virtue of their locations and their properties, slow earthquakes are certain toprovide new insights into the behavior of regular earthquakes. In the process, they will help uscharacterize the hazard that earthquakes pose, and ultimately their risk to society.

DISCLOSURE STATEMENT

The authors are not aware of any affiliations, memberships, funding, or financial holdings thatmight be perceived as affecting the objectivity of this review.

ACKNOWLEDGMENTS

We thank K. Obara, H. Dragert, G. Rogers, M. Miller, and S. Tsukada for their helpful commentsand accounts. This research was partially supported by National Science Foundation grant EAR-0710835, JSPS KAKENHI (20340115), and MEXT KAKENHI (21107007).

LITERATURE CITED

Abercrombie RE. 1995. Earthquake source scaling relationships from –1 to 5 ML using seismograms recordedat 2.5-km depth. J. Geophys. Res. 100:24015–36

Aguiar AC, Melbourne TI, Scrivner CW. 2009. Moment release rate of Cascadia tremor constrained by GPS.J. Geophys. Res. 114:B00A05

Ando R, Nakata R, Hori T. 2010. A slip pulse model with fault heterogeneity for low-frequency earthquakesand tremor along plate interfaces. Geophys. Res. Lett. 37:L10310

Asano Y, Obara K, Ito Y. 2008. Spatiotemporal distribution of very-low frequency earthquakes in Tokachi-okinear the junction of the Kuril and Japan trenches revealed by using array signal processing. Earth PlanetsSpace 60:871–75

Atwater BF. 1987. Evidence for great Holocene earthquakes along the outer coast of Washington State. Science236:942–44

Audet P, Bostock MG, Christensen NI, Peacock SM. 2009. Seismic evidence for overpressured subductedoceanic crust and megathrust fault sealing. Nature 457:76–78

Benioff H, Press F. 1958. Progress report on long period seismographs. Geophys. J. R. Astron. Soc. 1:208–15Beroza GC, Jordan TH. 1990. Searching for slow and silent earthquakes using free oscillations. J. Geophys.

Res. 95:2485–510Bird P. 2003. An updated digital model of plate boundaries. Geochem. Geophys. Geosyst. 4(3):1027Boyarko DC, Brudzinski MR. 2010. Spatial and temporal patterns of nonvolcanic tremor along the southern

Cascadia subduction zone. J. Geophys. Res. 115:B00A22Brodsky EE, Mori J. 2007. Creep events slip less than ordinary earthquakes. Geophys. Res. Lett. 34:L16309Brown JR, Beroza GC, Ide S, Ohta K, Shelly DR, et al. 2009. Deep low-frequency earthquakes in tremor

localize to the plate interface in multiple subduction zones. Geophys. Res. Lett. 36:L19306Brown JR, Shelly DR, Beroza GC. 2008. An autocorrelation method to detect low frequency earthquakes

within tremor. Geophys. Res. Lett. 35:L16305Brudzinski MR, Allen RM. 2007. Segmentation in episodic tremor and slip all along Cascadia. Geology 35:907–

10Chapman JS, Melbourne TI. 2009. Future Cascadia megathrust rupture delineated by episodic tremor and

slip. Geophys. Res. Lett. 36:L22301Chouet B. 1985. Excitation of a buried magmatic pipe: a seismic source model for volcanic tremor. J. Geophys.

Res. 90:1881–93

290 Beroza · Ide

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.


Cifuentes IL, Silver PG. 1989. Low-frequency source characteristics of the great 1960 Chilean earthquake. J.Geophys. Res. 94:643–63

Cocco M, Rovelli A. 1989. Evidence for variation of stress drop between normal and thrust faulting earthquakesin Italy. J. Geophys Res. 94:9399–416

Creager KC, Wech A, Vidale JE. 2009. Episodic tremor and slip: cycles within cycles. Eos Trans. AGU 90(52),Fall Meet. Suppl., Abstract T21F-06

Delahaye EJ, Townend J, Reyners ME, Rogers G. 2009. Microseismicity but no tremor accompanying slowslip in the Hikurangi subduction zone, New Zealand. Earth Planet. Sci. Lett. 277:21–28

Dieterich JH. 1979. Modeling of rock friction: 1. Experimental results and constitutive equations. J. Geophys.Res. 84:2161–68

Dieterich JH. 1994. A constitutive law for rate of earthquake production and its application to earthquakeclustering. J. Geophys. Res. 99:2601–18

Dragert H, Wang K, James TS. 2001. A silent slip event on the deeper Cascadia subduction interface. Science292:1525–28

Dragert H, Wang K, Rogers G. 2004. Geodetic and seismic signatures of episodic tremor and slip in thenorthern Cascadia subduction zone. Earth Planets Space 56:1143–50

Ekstrom G. 2006. Global detection and location of seismic sources using surface waves. Bull. Seismol. Soc. Am.96:1201–12

Ellsworth WL, Matthews MV, Nadeau RM, Nishenko SP, Reasenberg PA, Simpson RW. 1999. A physicallybased earthquake recurrence model for estimation of long-term earthquake probabilities. U.S. Geol. Surv.Tech. Rep. OFR 99-522, U.S. Geol. Surv., Menlo Park, Calif.

Fletcher J, Boatwright J, Haar L, Hanks T, McGarr A. 1984. Source parameters for aftershocks of the Oroville,California, earthquake. Bull. Seismol. Soc. Am. 74:1101–23

Ghosh A, Vidale JE, Creager KC. 2010a. Tremor as observed by the Array of Arrays in Cascadia, Abstr. S11C-03.Presented at 2010 AGU Fall Meet., San Francisco, Dec. 13–17

Ghosh A, Vidale JE, Sweet JR, Creager KC, Wech AG. 2009. Tremor patches in Cascadia revealed by seismicarray analysis. Geophys. Res. Lett. 36:L17316

Ghosh A, Vidale JE, Sweet JR, Creager K, Wech A, Houston H. 2010b. Tremor bands sweep Cascadia.Geophys. Res. Lett. 37:L08301

Gomberg J, Rubinstein JL, Peng Z, Creager KC, Vidale JE, Bodin P. 2008. Widespread triggering of non-volcanic tremor in California. Science 319:173

Gomberg J, the Cascadia 2007 and Beyond Working Group. 2010. Slow-slip phenomena in Cascadia from2007 and beyond: a review. Geol. Soc. Am. Bull. 122:963–78

Hawthorne JC, Rubin AM. 2010. Tidal modulation of slow slip in Cascadia. J. Geophys. Res. 115:B09406Hayes GP, Wald DJ, Keranen K. 2009. Advancing techniques to constrain the geometry of the seismic rupture

plane on subduction interfaces a priori: higher order functional fits. Geochem. Geophys. Geosyst. 10:Q09006Heki K, Miyazaki S, Tsuji H. 1997. Silent fault slip following an interplate thrust earthquake at the Japan

Trench. Nature 386:595–98Hirose H, Hirahara K, Kimata F, Fujii N, Miyazaki S. 1999. A slow thrust slip event following the two 1996

Hyuganada earthquakes beneath the Bungo Channel, southwest Japan. Geophys. Res. Lett. 26:3237–40Hirose H, Obara K. 2005. Repeating short- and long-term slow slip events with deep tremor activity around

the Bungo channel region, southwest Japan. Earth Planets Space 57:961–72Hirose H, Obara K. 2006. Short-term slow-slip and correlated tremor episodes in the Tokai region, central

Japan. Geophys. Res. Lett. 33:L17311Hirose T, Hiramatsu Y, Obara K. 2010. Characteristics of short-term slow slip events estimated from deep

low-frequency tremors in Shikoku, Japan. J. Geophys. Res. 115:B10304Houston H. 2001. Influence of depth, focal mechanism, and tectonic setting on the shape and duration of

earthquake source time functions. J. Geophys. Res. 106:11137–50Hyndman R, Wang K. 1995. The rupture zone of Cascadia great earthquakes from current deformation and

the thermal regime. J. Geophys. Res. 100:22133–54Ide S. 2008. A Brownian walk model for slow earthquakes. Geophys. Res. Lett. 35:L17301Ide S. 2010a. Striations, duration, migration and tidal response in deep tremor. Nature 466:356–59


Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.


Ide S. 2010b. Quantifying the time function of nonvolcanic tremor based on a stochastic model. J. Geophys.Res. 115:B08313

Ide S, Beroza GC, Shelly DR, Uchide T. 2007a. A scaling law for slow earthquakes. Nature 447:76–79Ide S, Imanishi K, Yoshida Y, Beroza GC, Shelly DR. 2008. Bridging the gap between seismically and geode-

tically detected slow earthquakes. Geophys. Res. Lett. 35:L10305Ide S, Shelly DR, Beroza GC. 2007b. Mechanism of deep low frequency earthquakes: further evidence that

deep non-volcanic tremor is generated by shear slip on the plate interface. Geophys. Res. Lett. 34:L03308Ide S, Shiomi K, Mochizuki K, Tonegawa T, Kimura G. 2010. Split Philippine Sea plate beneath Japan.

Geophys. Res. Lett. 37:L21304Ihmle PF, Jordan TH. 1994. Teleseismic search for slow precursors to large earthquakes. Science 266:1547–51Ishida M. 1992. Geometry and relative motion of the Philippine Sea plate and Pacific plate beneath the

Kanto-Tokai district, Japan. J. Geophys. Res. 97(B1):489–513Ito Y, Obara K. 2006. Dynamic deformation of the accretionary prism excites very low frequency earthquakes.

Geophys. Res. Lett. 33:L02311Ito Y, Obara K, Shiomi K, Sekine S, Hirose H. 2007. Slow earthquakes coincident with episodic tremors and

slow slip events. Science 315:503–6Julian BR. 1994. Volcanic tremor: nonlinear excitation by fluid flow. J. Geophys. Res. 99:11859–77Julian BR. 2002. Seismological detection of slab metamorphism. Science 296;1625–26Kanamori H. 1972. Mechanism of tsunami earthquakes. Phys. Earth Planet. Inter. 6:346–59Kanamori H, Cipar JJ. 1974. Focal process of the great Chilean earthquake of May 22, 1960. Phys. Earth

Planet. Inter. 9:128–36Kanamori H, Stewart GS. 1976. Mode of strain release along the Gibbs Fracture Zone, Mid-Atlantic Ridge.

Phys. Earth Planet. Inter. 11:312–32Kao H, Shan SJ. 2004. The source-scanning algorithm: mapping the distribution of sources in time and space.

Geophys. J. Int. 157:589–94Kao H, Shan SJ, Dragert H, Rogers G. 2009. Northern Cascadia episodic tremor and slip: a decade of tremor

observations from 1997 to 2007. J. Geophys. Res. 114:B00A12Kao H, Shan SJ, Dragert H, Rogers G, Cassidy JF, Ramachandran K. 2005. A wide depth distribution of

seismic tremors along the northern Cascadia margin. Nature 436:841–44Kato A, Sakai S, Iidaka T, Iwasaki T, Hirata N. 2010. Non-volcanic seismic swarms triggered by circulating

fluids and pressure fluctuations above a solidified diorite intrusion. Geophys. Res. Lett. 37:L15302Kato N. 2003. A possible model for large preseismic slip on a deeper extension of a seismic rupture plane.

Earth Planet. Sci. Lett. 216:17–25Katsumata A, Kamaya N. 2003. Low-frequency continuous tremor around the Moho discontinuity away from

volcanoes in the southwest Japan. Geophys. Res. Lett. 30:1020–23Kikuchi M, Nakamura M, Yoshikawa K. 2003. Source rupture processes of the 1944 Tonankai earthquake

and the 1945 Mikawa earthquake derived from low-gain seismograms. Earth Planets Space 55:159–72Kirby S, Wang K, Brocher T. 2002. A possible deep, long-term source for water in the northern San Andreas

Fault system: a ghost of Cascadia subduction past? Eos Trans. AGU 83(47), Fall Meet. Suppl., Abstr.S22B–1038

Kostoglodov V, Singh SK, Santiago JA, Franco SI, Larson KM, et al. 2003. A large silent earthquake in theGuerrero seismic gap, Mexico. Geophys. Res. Lett. 30:1807

Kuroki H, Ito HM, Takayama H, Yoshida A. 2004. 3-D simulation of the occurrence of slow slip events inthe Tokai region with a rate- and state-dependent friction law. Bull. Seismol. Soc. Am. 94:2037–50

Lambert A, Kao H, Rogers G, Courtier N. 2009. Correlation of tremor activity with tidal stress in the northernCascadia subduction zone. J. Geophys. Res. 114:B00A08

Langston CA, Rieger DM, Magnani MB. 2010. Unusual microseisms seen in the Reelfoot Fault Zone, northernTennessee, from a reflection experiment. Bull. Seismol. Soc. Am. 100:377–83

La Rocca N, Creager KC, Galluzzo A, Malone S, Vidale JE, et al. 2009. Cascadia tremor located near theplate interface constrained by S minus P wave times. Science 323:620–23

Linde AT, Gladwin MT, Johnston MJS, Gwyther RL, Bilham RG. 1996. A slow earthquake sequence nearSan Juan Bautista, California in December 1992. Nature 383:65–69

292 Beroza · Ide

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.


Liu Y, Rice JR. 2005. Aseismic slip transients emerge spontaneously in three-dimensional rate and statemodeling of subduction earthquake sequences. J. Geophys. Res. 110:B08307

Liu Y, Rice JR. 2007. Spontaneous and triggered aseismic deformation transients in a subduction fault model.J. Geophys. Res. 112:B09404

Liu Y, Rice JR, Larson KM. 2007. Seismicity variations associated with aseismic transients in Guerrero,Mexico, 1995–2006. Earth Planet. Sci. Lett. 262:493–504

Lowry AR, Larson KM, Kostoglodov V, Bilham R. 2001. Transient fault slip in Guerrero, southern Mexico.Geophys. Res. Lett. 28:7909–22

Matsubara M, Obara K, Kasahara K. 2009. High-VP/VS zone accompanying non-volcanic tremors and slowslip events beneath southwestern Japan. Tectonophysics 472:6–17

Mazzotti A, Adams J. 2004. Variability of near-term probability for the next great earthquake on the Cascadiasubduction zone. Bull. Seismol. Soc. Am. 94:1954–59

McCaffrey R, Wallace LM, Beavan J. 2008. Slow slip and frictional transition at low temperature at theHikurangi subduction zone. Nat. Geosci. 1:316–20

McCrory PA, Blair JL, Oppenheimer DH, Walter SR. 2006. Depth to the Juan de Fuca slab beneath theCascadia subduction margin: a 3-D model for sorting earthquakes. U.S. Geol. Surv. Data Ser. DS-91,Version 1.2 (CD-ROM)

McGuire JJ, Jordan T. 2000. Further evidence for the compound nature of slow earthquakes: the PrinceEdward Island earthquake of April 28, 1997. J. Geophys. Res. 105:7819–27

Meade BJ, Loveless JP. 2009. Predicting the geodetic signature of Mw ≥ 8 slow slip events. Geophys. Res. Lett.36:L01306

Miller MM, Melbourne TI, Johnson DJ, Sumner WQ. 2002. Periodic slow earthquakes from the Cascadiasubduction zone. Science 295:2423

Miyazaki S, Segall P, McGuire JJ, Kato T, Hatanaka Y. 2006. Spatial and temporal evolution of stress and sliprate during the 2000 Tokai slow earthquake. J. Geophys. Res. 111:B03409

Miyazawa M, Brodsky EE, Mori J. 2008. Learning from dynamic triggering of low-frequency tremor insubduction zones. Earth Planets Space 60:e17–20

Miyazawa M, Mori J. 2005. Detection of triggered deep low-frequency events from the 2003 Tokachi-okiearthquake. Geophys. Res. Lett. 32:L10307

Miyazawa M, Mori J. 2006. Evidence suggesting fluid flow beneath Japan due to periodic seismic triggeringfrom the 2004 Sumatra-Andaman earthquake. Geophys. Res. Lett. 33:L05303

Muller RD, Sdrolias M, Gaina C, Roest WR. 2008. Age, spreading rates, and spreading asymmetry of theworld’s ocean crust. Geochem. Geophys. Geosyst. 9:Q04006

Nadeau RM, Dolenc D. 2005. Nonvolcanic tremors deep beneath the San Andreas Fault. Science 307:389Nadeau RM, Guilhem A. 2009. Nonvolcanic tremor evolution and the San Simeon and Parkfield, California,

earthquakes. Science 325:191–93Nakata R, Suda N, Tsuruoka H. 2008. Non-volcanic tremor resulting from the combined effect of Earth tides

and slow-slip events. Nat. Geosci. 1:676–78Nishide N, Hashimoto T, Funasaki J, Nakazawa H, Oka M, et al. 2000. Nationwide activity of low-frequency

earthquakes in the lower crust in Japan. Proc. Jpn. Earth Planet. Sci. Jt. Meet., Abstr. Sk-P002. Tokyo:Geod. Soc. Jpn.

Obara K. 2002. Nonvolcanic deep tremor associated with subduction in southwest Japan. Science 296;1679–81Obara K. 2010. Phenomenology of deep slow earthquake family in southwest Japan: spatiotemporal charac-

teristics and segmentation. J. Geophys. Res. 115:B00A25Obara K, Hirose H, Yamamizu F, Kasahara K. 2004. Episodic slow slip events accompanied by non-volcanic

tremors in southwest Japan subduction zone. Geophys. Res. Lett. 31:L23602Obara K, Kasahara K, Hori S, Okada Y. 2005. A densely distributed high-sensitivity seismograph network in

Japan: Hi-net by National Research Institute for Earth Science and Disaster Prevention. Rev. Sci. Instrum.76:021301

Ohmi S, Hirose I, Mori J. 2004. Deep low-frequency earthquakes near the downward extension of the seis-mogenic fault of the 2002 Western Tottori earthquake. Earth Planets Space 60:877–82

Ohmi S, Obara K. 2002. Deep low-frequency earthquakes beneath the focal region of the Mw 6.7 WesternTottori earthquake. Geophys. Res. Lett. 29:1807


Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.


Ohta Y, Freymueller JT, Hreinsdottir S, Suito H. 2006. A large slow slip event and the depth of the seismogeniczone in the south central Alaska subduction zone. Earth Planet. Sci. Lett. 247:108–16

Outerbridge KC, Dixon TH, Schwartz SY, Walter JI, Protti M, et al. 2010. A tremor and slip event onthe Cocos-Caribbean Subduction zone as measured by a global positioning system (GPS) and seismicnetwork on the Nicoya Peninsula, Costa Rica. J. Geophys. Res. 115:B10408

Ozawa S, Murakami M, Kaidzu M, Tada T, Sagiya T, et al. 2002. Detection and monitoring of ongoingaseismic slip in the Tokai region, central Japan. Science 298:1009–12

Ozawa S, Suito H, Tobita M. 2007. Occurrence of quasi-periodic slow-slip off the east coast of the Bosopeninsula, Central Japan. Earth Planets Space 59:1241–45

Payero JS, Kostoglodov V, Shapiro N, Mikumo T, Iglesias A, et al. 2008. Nonvolcanic tremor observed inthe Mexican subduction zone. Geophys. Res. Lett. 35:L07305

Peacock SM. 2009. Thermal and metamorphic environment of subduction zone episodic tremor and slip. J.Geophys. Res. 114:B00A07

Peacock SM, Wang K. 1999. Seismic consequences of warm versus cool subduction metamorphism: examplesfrom southwest and northeast Japan. Science 286:937–39

Peng Z, Chao K. 2008. Non-volcanic tremor beneath the Central Range in Taiwan triggered by the 2001 Mw

7.8 Kunlun earthquake. Geophys. J. Int. 175:825–29Peng Z, Vidale JE, Wech AG, Nadeau RM, Creager KC. 2009. Remote triggering of tremor along the San

Andreas Fault in central California. J. Geophys. Res. 114:B00A06Peterson CL, Christensen DH. 2009. Possible relationship between nonvolcanic tremor and the 1998–2001

slow slip event, south central Alaska. J. Geophys. Res. 114:B06302Piccinini D, Saccorotti G. 2008. First observations of nonvolcanic, long-period seismicity in the central

Apennines, Italy. Geophys. Res. Lett. 35:L12303Prieto GA, Shearer PM, Vernon FL, Kilb D. 2004. Earthquake source scaling and self-similarity estimation

from stacking P and S spectra. J. Geophys. Res. 109:B08310Rogers G, Dragert H. 2003. Episodic tremor and slip on the Cascadia subduction zone: the chatter of silent

slip. Science 300:1942–43Rubin AM. 2008. Episodic slow slip events and rate-and-state friction. J. Geophys. Res. 113:B11414Rubinstein JL, La Rocca M, Vidale JE, Creager KC, Wech AG. 2008. Tidal modulation of non-volcanic

tremor. Science 319:186–89Rubinstein JL, Shelly DR, Ellsworth WL. 2010. Non-volcanic tremor: a window into the roots of fault zones.

In New Frontiers in Integrated Solid Earth Sciences, ed. S Cloetingh, J Negendank, pp. 287–314. Dordrecht:Springer

Rubinstein JL, Vidale JE, Gomberg J, Bodin P, Creager KC, Malone SD. 2007. Non-volcanic tremor drivenby large transient shear stresses. Nature 448:579–82

Ruina A. 1983. Slip instability and state variable friction laws. J. Geophys. Res. 88:10359–70Sacks IS, Linde AT, Snoke JA, Suyehiro A. 1981. A slow earthquake sequence following the Izu-Oshima

earthquake of 1978. In Earthquake Prediction: an International Review, ed. DW Simpson, PG Richards.Maurice Ewing Ser. 4:617–28. Washington, DC: AGU

Sagiya T, Thatcher W. 1999. Coseismic slip resolution along a plate boundary megathrust: the Nankai Trough,southwest Japan. J. Geophys. Res. 104:1111–29

Satake K, Shimazaki K, Tsuji Y, Ueda K. 1996. Time and size of a giant earthquake in Cascadia inferred fromJapanese tsunami records of January 1700. Nature 379:246–49

Schwartz SY, Rokosky JM. 2007. Slow slip events and seismic tremor at circum-Pacific subduction zones. Rev.Geophys. 45:RG3004

Segall P, Desmarais EK, Shelly DR, Miklius A, Cervelli P. 2006. Earthquakes triggered by silent slip eventson Kilauea volcano, Hawaii. Nature 442:71–74

Segall P, Rubin AM, Bradley AM, Rice JR. 2010. Dilatant strengthening as a mechanism for slow slip events.J. Geophys Res. 115:B12305

Sekine S, Hirose H, Obara K. 2010. Along-strike variations in short-term slow slip events in the southwestJapan subduction zone. J. Geophys. Res. 115:B00A27

Seno T, Yamasaki T. 2003. Low-frequency tremors, intraslab and interplate earthquakes in Southwest Japan—from a viewpoint of slab dehydration. Geophys. Res. Lett. 30:2171

294 Beroza · Ide

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.


Shearer PM. 1994. Global seismic event detection using a matched filter on long-period seismograms. J.Geophys. Res. 99:13713–25

Shelly DR. 2009. Possible deep fault slip preceding the 2004 Parkfield earthquake, inferred from detailedobservations of tectonic tremor. Geophys. Res. Lett. 36:L17318

Shelly DR. 2010a. Migrating tremors illuminate deformation beneath the seismogenic San Andreas Fault.Nature 463:648–52

Shelly DR. 2010b. Periodic, chaotic, and doubled earthquake recurrence intervals on the deep San AndreasFault. Science 328:1385–88

Shelly DR, Beroza GC, Ide S. 2007a. Complex evolution of transient slip derived from precise tremor locationsin western Shikoku, Japan. Geochem. Geophys. Geosyst. 8:Q10014

Shelly DR, Beroza GC, Ide S. 2007b. Non-volcanic tremor and low-frequency earthquake swarms. Nature446:305–7

Shelly DR, Beroza GC, Ide S, Nakamula S. 2006. Low-frequency earthquakes in Shikoku, Japan, and theirrelationship to episodic tremor and slip. Nature 442:188–91

Shelly DR, Ellsworth WL, Ryberg T, Haberland C, Fuis GS, et al. 2009. Precise location of San AndreasFault tremors near Cholame, California using seismometer clusters: slip on the deep extension of thefault? Geophys. Res Lett. 36:L01303

Shelly DR, Hardebeck JL. 2010. Precise tremor source locations and amplitude variations along the lower-crustal central San Andreas Fault. Geophys. Res. Lett. 37:L14301

Shibazaki B, Bu S, Matsuzawa T, Hirose H. 2010. Modeling the activity of short-term slow slip events alongdeep subduction interfaces beneath Shikoku, southwest Japan. J. Geophys. Res. 115:B00A19

Shibazaki B, Iio Y. 2003. On the physical mechanism of silent slip events along the deeper part of the seismo-genic zone. Geophys. Res. Lett. 30:1489

Shibazaki B, Shimamoto T. 2007. Modeling of short-interval silent slip events in deeper subduction interfacesconsidering the frictional properties at the unstable–stable transition regime. Geophys. J. Int. 171:191–205

Silver PG, Jordan TH. 1982. Optimal estimation of scalar seismic moment. Geophys. J. R. Astron. Soc. 70:755–87

Song ATR, Helmberger DV, Brudzinski MR, Clayton RW, Davis P, et al. 2009. Subducting slab ultra-slowvelocity layer coincident with silent earthquakes in southern Mexico. Science 324:502–6

Szeliga W, Melbourne TI, Miller MM, Santillan VM. 2004. Southern Cascadia episodic slow earthquakes.Geophys. Res. Lett. 31:L16602

Szeliga W, Melbourne T, Santillan M, Miller M. 2008. GPS constraints on 34 slow slip events within theCascadia subduction zone, 1997–2005. J. Geophys. Res. 113:B04404

Thatcher W. 1975. Strain accumulation on the northern San Andreas Fault zone since 1906. J. Geophys. Res.80:4862–72

Thatcher W, Hanks TC. 1973. Source parameters of southern California earthquakes. J. Geophys. Res. 78:8547–76

Thomas AM, Nadeau RM, Burgmann R. 2009. Tremor-tide correlations and near-lithostatic pore pressureon the deep San Andreas fault. Nature 462:1048–51

Tse ST, Rice JR. 1986. Crustal earthquake instability in relation to the depth variation of frictional slipproperties. J. Geophys. Res. 91:9452–72

Wada I, Wang K, He J, Hyndman RD. 2008. Weakening of the subduction interface and its effects on surfaceheat flow, slab dehydration, and mantle wedge serpentinization. J. Geophys. Res. 113:B04402

Watanabe T, Hiramatsu Y, Obara K. 2007. Scaling relationship between the duration and the amplitude ofnon-volcanic deep low-frequency tremors. Geophys. Res. Lett. 34:L07305

Wech AG, Creager KC. 2007. Cascadia tremor polarization evidence for plate interface slip. Geophys Res. Lett.34:L22306

Wech AG, Creager KC. 2008. Automated detection and location of Cascadia tremor. Geophys. Res. Lett.35:L20302

Yamashita T. 1980. Quasi-static crack extension in an inhomogeneous viscoelastic medium—a possible mech-anism for the occurrence of aseismic faulting. J. Phys. Earth 28:309–26


Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.


Yoshioka S, Mikumo T, Kostoglodov V, Larson KM, Lowry AR, Singh SK. 2004. Interplate coupling and arecent aseismic slow slip event in the Guerrero seismic gap of the Mexican subduction zone, as deducedfrom GPS data inversion using a Bayesian information criterion. Phys. Earth Planet. Inter. 146:513–30

Yoshioka S, Toda M, Nakajima J. 2008. Regionality of deep low-frequency earthquakes associated with sub-duction of the Philippine Sea plate along the Nankai Trough, southwest Japan. Earth Planet. Sci. Lett.272:189–98

296 Beroza · Ide

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.

EA39-FrontMatter ARI 24 March 2011 6:51

Annual Reviewof Earth andPlanetary Sciences

Volume 39, 2011 Contents

Plate Tectonics, the Wilson Cycle, and Mantle Plumes: Geodynamicsfrom the TopKevin Burke � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 1

Early Silicate Earth DifferentiationGuillaume Caro � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �31

Building and Destroying Continental MantleCin-Ty A. Lee, Peter Luffi, and Emily J. Chin � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �59

Deep Mantle Seismic Modeling and ImagingThorne Lay and Edward J. Garnero � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �91

Using Time-of-Flight Secondary Ion Mass Spectrometryto Study BiomarkersVolker Thiel and Peter Sjovall � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 125

Hydrogeology and Mechanics of Subduction Zone Forearcs:Fluid Flow and Pore PressureDemian M. Saffer and Harold J. Tobin � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 157

Soft Tissue Preservation in Terrestrial Mesozoic VertebratesMary Higby Schweitzer � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 187

The Multiple Origins of Complex MulticellularityAndrew H. Knoll � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 217

Paleoecologic Megatrends in Marine MetazoaAndrew M. Bush and Richard K. Bambach � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 241

Slow Earthquakes and Nonvolcanic TremorGregory C. Beroza and Satoshi Ide � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 271

Archean Microbial Mat CommunitiesMichael M. Tice, Daniel C.O. Thornton, Michael C. Pope,

Thomas D. Olszewski, and Jian Gong � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 297

Uranium Series Accessory Crystal Dating of Magmatic ProcessesAxel K. Schmitt � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 321

viii

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.

EA39-FrontMatter ARI 24 March 2011 6:51

A Perspective from Extinct Radionuclides on a Young Stellar Object:The Sun and Its Accretion DiskNicolas Dauphas and Marc Chaussidon � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 351

Learning to Read the Chemistry of Regolith to Understand theCritical ZoneSusan L. Brantley and Marina Lebedeva � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 387

Climate of the NeoproterozoicR.T. Pierrehumbert, D.S. Abbot, A. Voigt, and D. Koll � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 417

Optically Stimulated Luminescence Dating of Sediments over the Past200,000 YearsEdward J. Rhodes � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 461

The Paleocene-Eocene Thermal Maximum: A Perturbation of CarbonCycle, Climate, and Biosphere with Implications for the FutureFrancesca A. McInerney and Scott L. Wing � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 489

Evolution of Grasses and Grassland EcosystemsCaroline A.E. Stromberg � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 517

Rates and Mechanisms of Mineral Carbonation in Peridotite:Natural Processes and Recipes for Enhanced, in situ CO2 Captureand StoragePeter B. Kelemen, Juerg Matter, Elisabeth E. Streit, John F. Rudge,

William B. Curry, and Jerzy Blusztajn � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 545

Ice Age Earth RotationJerry X. Mitrovica and John Wahr � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 577

Biogeochemistry of Microbial Coal-Bed MethaneDariusz Strapoc, Maria Mastalerz, Katherine Dawson, Jennifer Macalady,

Amy V. Callaghan, Boris Wawrik, Courtney Turich, and Matthew Ashby � � � � � � � � � � 617

Indexes

Cumulative Index of Contributing Authors, Volumes 29–39 � � � � � � � � � � � � � � � � � � � � � � � � � � � 657

Cumulative Index of Chapter Titles, Volumes 29–39 � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 661

Errata

An online log of corrections to Annual Review of Earth and Planetary Sciences articlesmay be found at http://earth.annualreviews.org

Contents ix

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2011

.39:

271-

296.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Seo

ul N

atio

nal U

nive

rsity

on

07/1

2/11

. For

per

sona

l use

onl

y.

Documents

Slow Earthquakes and Nonvolcanic Tremor€¦ · tremor and LFEs, were considered to be distinct when they were discovered. Earthquakes happen impulsively and complete quickly—usually