Earth and Planetary Science Letters 305 (2011) 215–225

Contents lists available at ScienceDirect

Earth and Planetary Science Letters

j ourna l homepage: www.e lsev ie r.com/ locate /eps l

Earthquake swarms in circum-Pacific subduction zones

S.G. Holtkamp ⁎, M.R. BrudzinskiMiami University Geology Department, 114 Shideler Hall, Oxford OH, 45056, USA

⁎ Corresponding author. Tel.: +1 513 235 8915.E-mail addresses: holtkasg@muohio.edu (S.G. Holtka

(M.R. Brudzinski).

0012-821X/$ – see front matter © 2011 Elsevier B.V. Adoi:10.1016/j.epsl.2011.03.004

a b s t r a c t

a r t i c l e i n f o

Article history:Received 4 October 2010Received in revised form 2 March 2011Accepted 4 March 2011Available online 30 March 2011

Editor: P. Shearer

Keywords:Earthquake swarmsSubduction zones

We systematically and manually search through clusters of earthquakes along circum-Pacific subductionzones to identify potential earthquake swarms. In total, we find 266 potential earthquake swarms: 180 weclassify as megathrust and 68 we classify as volcanic due to their proximity to the megathrust or to volcanoes.We focus on the megathrust swarms and demonstrate that: (1) the number of events in a swarm is not afunction of the largest earthquake in the swarm, (2) swarms exhibit an approximately constant rate ofseismicity that lasts until after the mean timing of events in the swarm, (3) the timing of the largestearthquake in the sequence is no different than the timing of any other earthquake in the sequence, (4) ourcatalogs of earthquakes comprising swarms (~9000 events) have high b-values (1.5 to 2), and (5) whenearthquake swarms are considered as single events using total duration and cumulative moment, they appearto be consistent with the slow earthquake magnitude-duration scaling law presented by Ide et al. (2007). Thefirst three observations, along with the observation that swarms can span very large areas compared to theircumulative seismic moment, argue against static stress triggering as a driving mechanism for earthquakeswarms. Along strike propagation velocities are observed for several swarms, showing epicentral propagationof ~10 km/day, similar to other documented slow slip events. Together, this evidence implies that aseismicslip along the megathrust is likely an important mechanism for the generation of megathrust earthquakeswarms in circum-Pacific subduction zones. We then conduct a comparison of swarms and large megathrustearthquakes, finding evidence that the two are broadly anti-correlated: megathrust segments with largeearthquake swarm gaps are more likely to experience large (MwN8) megathrust events. We characterize theubiquity of megathrust swarms at different margins, and suggest that fault properties along Marianas-typemargins may allow for earthquake swarms to occur regularly, but other margins may rely on other variables,such as the subduction of a ridge or seamount, to facilitate the generation of megathrust earthquake swarms.

mp), brudzimr@muohio.edu

ll rights reserved.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

Relationships between earthquakes are observed by the cluster-ing of earthquakes in space and time. This clustering commonlyoccurs as mainshock–aftershock (MS–AS) sequences, which aregenerally interpreted to contain the initial rupture of a fault (themainshock) and a decaying cascade of smaller ruptures on or verynear to the initial rupture plane (aftershocks) (Lay and Wallace,1995). In fact, aftershock sequences are often used to define therupture plane of the associated mainshock (e.g., Sykes, 1971; Utsuand Seki, 1954).

Clustering of earthquakes in space and time can also occur asearthquake swarms, which are empirically defined as an increase inseismicity rate above the background rate without a clear triggeringmainshock earthquake (Hill, 1977; Mogi, 1963; Sykes, 1970).Earthquake swarms are often associated with volcanic regions andare studied because of their relationship to eruptions or intrusions of

magmatic material (Benoit and McNutt, 1996). Earthquake swarmshave been documented in areas not associated with active volcanism,such as transform faults (Lohman and McGuire, 2007; Shibutani et al.,2002) and hydrothermal systems (Fischer and Horalek, 2003; Heinikeet al., 2009). Triggering mechanisms for these non-volcanic swarmsrange from associated aseismic slip on associated faults (Lohman andMcGuire, 2007) to movement of volatiles in hydrothermal systems(Heinike et al., 2009).

Earthquake swarms at subduction margins not associated withvolcanism have been documented in New Zealand (Evison andRhoades, 1993), Japan (Fujinawa et al., 1983; Matsuzawa et al., 2004),Kamchatka (Slavina et al., 2007; Zobin, 1996), Mexico (Zobin, 1996),and South America (Holtkamp et al., 2011; Lemoine et al., 2001).Studies of earthquake swarms at these convergent margins have beenmotivated by their potential relation to large megathrust events,although the mechanisms behind swarm nucleation and potentialinteraction with large megathrust events remains debated (Evisonand Rhoades, 1993; Llenos et al., 2009).

Most swarms documented in literature were located with local orregional scale seismic networks, often including offshore networks,and utilize local earthquake catalogs with lower magnitude

216 S.G. Holtkamp, M.R. Brudzinski / Earth and Planetary Science Letters 305 (2011) 215–225

thresholds (e.g., Evison and Rhoades, 1993; Flueh et al., 1998; Vidaleand Shearer, 2006). While the heterogeneity of seismic networksprevents a global study of this type, the goal of this paper is to initiatea catalog of earthquake swarms along Circum-Pacific subductionzones using the global scale Preliminary Determination of Epicenters(PDE) data set. The core of this work is an expansion of the manualearthquake swarm search conducted by Holtkamp et al. (2011) overthe South American continent.

2. Methods

Wedownload and examine the complete PDE catalog from 1973 to2010 over the following regions: South America, Mexico/CentralAmerica, Alaska, Kurile-Kamchatka, Japan, Taiwan/Manila/Philippines,Sumatra, Vanuatu, and Tonga/New Zealand. Since earthquake swarmshave been defined empirically in the past (e.g., Hill, 1977), we beginwith our definition of an earthquake swarm that agrees withpreviously defined swarm properties (detailed below). We define anearthquake swarm to be a noticeable increase in seismicity rate abovea visually established background seismicity rate without a cleartriggering mainshock. Swarms typically have many earthquakes nearthe magnitude of the largest earthquake in the cluster so they do notfollow Baths Law, which states that the largest aftershock is typicallyone moment magnitude smaller than the triggering mainshock. Wefind thatmany earthquake swarms have abrupt onset and terminationof seismicitywhen compared to background seismicity (e.g., without adecay in seismicity rate as in a decaying aftershock sequence). We usethis to help determine if a cluster is a swarm, but it is not a require-ment, as it is likely that relatively abrupt termination is a necessaryoutcome of the visual swarm determination. Fig. 1 outlines theseobservations with a representative swarm example. In contrast, Fig. 2shows a typical mainshock–aftershock (MSAS) sequence, in which themainshock is first in the sequence and is typically one moment mag-nitude larger than the second largest earthquake (Baths Law), and thesequence typically fades into the background seismicity rate withoutan abrupt termination.

We use these criteria to search through all major circum-Pacificsubduction zones for clusters of earthquakes that appear swarm-like.For each region, we systematically examine each apparent cluster ofseismicity (apparent as a vertical line of dots in the bottom panel ofFigs. 1 and 2). Clusters that appear to have a triggering mainshock orare dominated by a single event are discarded, while the remainingclusters are marked as having swarm-like characteristics. Back-ground seismicity rates in the PDE catalog are highly variable in twoways: (1) reported seismicity rates from 1973 to 2010 vary by about2 orders of magnitude, likely due to increased instrumentation, and(2) background seismicity varies within each region studies,sometimes drastically (e.g., central Chile, from 30 to 35°S, accountsfor half of the seismicity in the South and Central American PDEcatalog).

In regions with high background seismicity rate (e.g. Alaska,central Chile), visual characterization of swarms becomes moredifficult. In these cases, larger earthquake magnitudes (~1 Mw larger)or larger increases in seismicity rate (e.g., several tens of earthquakesin a period of days to weeks) are necessary to distinguish the cluster,but not both. For example, Holtkamp et al. (2011) find a swarm at thePapudo seamount, South America, without an increase in earthquakemagnitudes because there were several tens of earthquakes in a fewdays. In areas with low background seismicity rate (e.g. southern Chileand Bonin-Marianas Trench) seismicity rate increases can be detectedeven if only a few earthquakes are large enough to be recorded byregional networks. In Puerto Aysen, southern Chile, for example, weidentified two earthquake swarms (1991 and 2007) despite findingless than 15 regionally recorded earthquakes in the PDE catalog. In thecase of the 2007 swarm, a local seismic network recorded over 6000earthquakes without a mainshock (Mora et al., 2008), supporting

the use of our approach in cases of limited earthquake numbers in thePDE catalog. For a more detailed examination of the visual detectionmethodology, see Supplementary Figs. S1 and S2.

In considering ways to pursue an automated swarm detectionapproach instead, we found that previous studies successfully imple-menting an automated detection have often relied on a uniformbackground seismicity level and magnitude threshold, which areconditions that cannot be met in our global study. For example, themethod of Vidale and Shearer (2006) constructed an unbiased auto-mated burst detection algorithm that exploited a uniform backgroundseismicity rate, but with limited spatial and temporal scale. Yet evenwithin that dataset, visual classification of swarms was still required.Since we aim to produce a swarm catalogwhich is not limited in spaceand time and is produced from a global catalog with widely varyingbackground seismicity rate and magnitude threshold (both vary byseveral orders of magnitude), it does not allow us to assume aconstant background seismicity rate or magnitude threshold. As aresult, we rely on a visual swarm detection algorithm. While ourvisual search is likely incomplete, we are encouraged that the swarmcharacteristics we present in the next section closely resemble thoseof Vidale and Shearer (2006).

Since magnitude plays a role in defining earthquake swarms, weseek to establish a consistent magnitude measurement in our catalogsearch. First, with regards to catalog completeness, we find that inrecent years completeness is ~Mw=4 along major convergentmargins. However, in the earlier decades of the catalog, complete-ness was ~Mw=5. Secondly, magnitudes given in the PDE catalogare either locally constrained (ML) or regionally/globally con-strained (waveform-constrained moment magnitudes forMwN~4.5 in the past 20 yrs and body wave magnitudes for∼4bMwb∼4.5). In this analysis, locally constrained magnitudes areignored as there is no clear conversion to moment magnitudes.When only body wave magnitudes are given, a conversion tomoment magnitudes is performed by adding 0.31 to the body wavemagnitude (based on an empirical law given by Stein andWysession,(2003)). Prior to 20 yrs ago, only MwN~6 had waveform-con-strained moment magnitudes reported and so earthquakes smallerthan this are converted from body wave magnitudes. Consideringthat magnitude differences in MS–AS sequences are ~1 (Bath's Law),these minor adjustments we make to try to establish a consistentmagnitude measurement are not likely to influence swarmdetection.

3. Characteristics of earthquake swarms

In total, we find 266 potential earthquake swarms (Fig. 3). Wenext attempt to classify them according to the tectonic regimewherethey occurred. There exists a bimodal distribution of swarms insubduction zones: those near the seismogenic megathrust and thosenear the volcanic arc (perhaps best seen in Supplementary Figs. S3and S4). 180 swarms lie within the 0 and 50 km depth to slabinterface contours, and we classify these as megathrust earthquakeswarms. The PDE catalog does not have the epicentral or depthresolution to determine whether these earthquakes represent actualmegathrust faulting, but these swarms show thrusting focalmechanisms for every case where magnitudes were large enoughto have CentroidMoment Tensor (CMT) solutions (about one quarterof swarms, 47 of the 182). In any case, the proximity of these swarmsto the plate interface indicates that the megathrust is playing aprominent role in their formation.

We classify 68 swarms as volcanic, which we define as occurringwithin ~50 km of an active volcano in the Smithsonians Global Vol-canism Program (GVP) database. These swarms are typically shallow(in the crust) and many are associated with volcanic eruptions ordocumented volcanic activity. We list 18 swarms as other becausethey don't fit the megathrust or volcanic swarm definitions. These

Fig. 1. Example of an earthquake swarm. (Top Panel) Map view of the seismicity displayed in the middle panel and associated with the 1980 Vanuatu swarm. Red triangles areHolocene volcanoes from the Global Volcanism Program. Dashed lines give depth to slab contours in 50 km increments. Colors of circles are relative time as defined by the color bar atthe top of the middle panel. Small map shows the regional context. (Middle Panel) Earthquake magnitude vs. time for ~3 weeks around the swarm. (Bottom Panel) Earthquakemagnitude versus time over the region defined in the top panel for 15 yrs surrounding the swarm with vertical bars representing the time shown in the top two panels.

217S.G. Holtkamp, M.R. Brudzinski / Earth and Planetary Science Letters 305 (2011) 215–225

include swarms that occurred in the outer rise (e.g. Izu-Bonin trench)and backarc spreading centers (e.g. the Andaman Sea backarc sprea-ding center). A list of all earthquake swarms shown in Fig. 3 areincluded in Supplementary Tables S1 (megathrust), S2 (volcanic), andS3 (other).

Next, we quantitatively classify and characterize our megathrustswarm catalog by following the methodology of Vidale and Shearer(2006). Fig. 4 shows the number of events in each swarm against thelargest earthquake in that swarm. For this figure, MSAS sequencesare simply a random sampling of mainshocks from each subductionzone studied. This figure has two important characteristics: (1) thereis no tendency among swarms to have a larger number of events

with a bigger largest event, that is, the number of events is notcontrolled by the largest event (which is expected for and seen inMSAS sequences), and (2) swarms and MSAS sequences plot in twodistinct, separated regions in this plot, effectively separating MSASsequences from swarms and providing a quantitative measurementof swarminess.

To further examine the swarminess of our megathrust swarmcatalog, we investigate the relative timing of events within the swarm.We do this with the time normalization method of Vidale and Shearer(2006), where the timing of each event in a swarm is normalized suchthat the mean event timing is 1. The relative timing of events withinthe swarms can then be averaged for all 180 swarms. Fig. 5 shows this

Fig. 2. Example of a mainshock–aftershock (MSAS) sequence. Layout is similar to Fig. 1. (Top Panel) Map view of the seismicity displayed in the middle panel and associated with a1996 Aleutian MSAS sequence. (Middle Panel) Earthquake magnitude vs. time for ~1 week around the sequence showing the typical Omoris Law trend where the rate of aftershocksis proportional to the inverse of time since the mainshock. (Bottom Panel) Stars mark mainshocks of productive MSAS sequences.

218 S.G. Holtkamp, M.R. Brudzinski / Earth and Planetary Science Letters 305 (2011) 215–225

result for the timing of all events and the timing of the largest event inthe sequence. We compare our normalized swarm to that of Vidaleand Shearer (2006) and find remarkably consistent results betweenthe two studies despite large differences in region and scale betweenthese studies 5. In both cases (global: circum-Pacific and local: South-ern California), we find three consistent points. First, the initial peakcontains ~15% of the earthquakes. We agree with the interpretation inVidale and Shearer (2006) that this peak is likely composed of theaftershocks of the initiating earthquake, which is sometimes one ofthe larger in the sequence, or that the largest earthquake and itsaftershocks disproportionately occurs early in the sequence. Second,there is an approximately constant rate of earthquakes at about 5% per

time period after the initial peak that lasts until after the mean time.Third, the seismicity rate diminishes rapidly after ~1.4 normalizedtime, perhaps in accordance with an Omori-type decay law. Thesecond point is particularly convincing evidence that our catalog is infact swarms.

Fig. 5b is identical to 5a, but for only the largest earthquake ineach sequence. The similarity between Fig. 5a and b is furtherevidence of the swarminess because the largest earthquake is nodifferent than any other earthquake. Also, point (3) suggests that theswarm-like behavior may on average stop after 1.4 normalized timein the sequence. If the seismicity rate then decays with an Omori-type law, the end of the sequences would be best explained as

Fig. 3.Map of earthquake swarm seismicity. Blue circles represent earthquakes associated with megathrust earthquake swarms, green circles represent volcanic earthquake swarms,and black circles are other earthquake swarms as defined in the main text. Earthquake swarms were found in every subduction zone examined, but appear to be more common inMariana-type margins. Red triangles are Holocene volcanoes from the Global Volcanism Program, plate boundary model is from Bird (2003), and colored background is seafloorbathymetry.

1

10

100

1000

Num

ber

of E

Qs

M

w=

5

5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0

Largest EQ (Mw)

MS-

AS

“Swar

my”

MS-

AS

Legend

Swarm-like

Mainshock-Aftershock (MS-AS)

in Between Swarm and MS-AS

“Sw

arm

y”

MS-

AS

>

Fig. 4. Number of earthquakes in a sequence relative to the magnitude of largest event for swarms and mainshock–aftershock sequences (MSAS). Only earthquakes greater thanMw=5, which we consider to be a global catalog threshold, are included. The number of earthquakes in swarms, unlike MSAS, is not a function of the largest earthquake in thesequence. We compare our results with those for a local catalog in Southern California (Vidale and Shearer, 2006). In both cases, “swarmy” sequences and MSAS sequences can beeffectively separated by a line.

219S.G. Holtkamp, M.R. Brudzinski / Earth and Planetary Science Letters 305 (2011) 215–225

Normalized 180 Swarm AverageTiming of all events

t-1

t-1/2

0

5

10

15

20

Per

cent

of E

arth

quak

es

0.0 0.5 1.0 1.5 2.0 2.5 3.0

Normalized time

Normalized time

9a

Timing of largest event in each sequence

9b

0

5

10

15

20

Per

cent

of E

arth

quak

es

0.0 0.5 1.0 1.5 2.0 2.5 3.0

Timing of largest earthquake in each sequencet-1

t-1/2

a

b

Fig. 5. Temporal distribution of earthquakes within all swarms by normalizing time tothe total duration of each swarm and then stacking all 180 swarms together. Themethod of time normalization is taken from Vidale and Shearer (2006), to which werefer the interested reader for a thorough description of the method. (a) Timing of allearthquakes in each swarm. We plot two Omori-style rate functions, showing that theydo not fit for most of the normalized swarm. Common results for our analysis and thatfor a local catalog in Southern California (Vidale and Shearer, 2006) are: (1) an initialburst with ~15% of earthquakes, (2) roughly steady rate that lasts until normalized time1.4, which is after the mean time of 1, and (3) after time 1.4, the seismicity rate dropsoff, perhaps following the Omori-style curve (implying the end of seismicity iscomprised of aftershocks of the swarm events). (b) Relative timing of the largest eventin the sequence. The similarity between (a) and (b) suggests that the largest earth-quake is not special, as it does not have a higher probability to occur at any particulartime as any other earthquake in the sequence.

0

1

2

3

4

5

Log(

Cou

nt >

Mag

)

4 6 8

Magnitude (Mw)

Megathrust Swarm EQs: -1.50 +- 0.03Volcanic Swarm EQs: -2.01 +- 0.07Other Swarm EQs: -2.09 +- 0.062006 PDE Catalog: -1.04 +- 0.02

Slope (b-value)

Fig. 6.Magnitude-frequency relations for earthquakes within swarms. Each earthquakethat was part of a swarm is included in this panel, separated by their spatial catego-rization. B-value is calculated by least squares fit over the region which shows a linearrelationship.

220 S.G. Holtkamp, M.R. Brudzinski / Earth and Planetary Science Letters 305 (2011) 215–225

aftershocks of the earthquakes in the sequence. This may lead toslightly overestimated durations, which will be important inSection 4.1.

Fig. 6 shows magnitude-frequency relations for the individualearthquakes within the swarms. Globally, b-values (the log-linearslope of the magnitude-frequency relation) are around 1. The entire2006 ISC catalog has a b-value of around 1.04 (Fig. 6). Our catalogs ofearthquakes within swarms (megathrust, volcanic, and other, ~9000events total) have high b-values (1.5 to 2), indicating that swarmsare deficient in larger magnitude events compared to other earth-quakes. This is not unusual, as high b-values are often documentedwithin individual earthquake swarms, for example in volcanicregions (Lay and Wallace, 1995). We do not calculate b-values for

the overall magnitude of swarms because there are too few toproduce a large enough magnitude range over which a magnitudefrequency relationship is linear.

While megathrust swarms exist in every subduction zone, theyappear to be more common in Vanuatu, Tonga, Kamchatka (Kurile),and Alaska (Aleutian) and less common in Japan, Central America,most of South America, and Sumatra (Table S1). However, the regionswe report to more commonly have swarms all have low backgroundseismicity levels, so there is likely some bias in this observation. InSouth America, we find a strong correlation between the location ofmegathrust swarms and the subduction of oceanic ridges or sea-mounts (Fig. 3) (Holtkamp et al., 2011). This correlation has also beenpreviously recognized in Japan (Fujinawa et al., 1983), but does notappear to be the case in other areas (i.e., Vanuatu, Tonga, or Izu-Bonin-Mariana; Supplementary Figs. S3–S5). Therefore, we suggest thatcertain subduction zones, perhaps Marianas-type, are more proneto experiencing earthquake swarms over broader portions of theinterface while other subduction zones require an external factor,such as the subduction of an oceanic ridge, for earthquake swarms tooccur.

In some subduction zones, Vanuatu for example, earthquakeswarms can cover large areas. Fig. 7 shows a remarkable example ofthis. This sequence in 1980 began with a Mw=7.1 MSAS sequencenear the southernmost edge of the margin. Over the next two years, aseries of 5 earthquake swarms occurred to the northwest occupyingan along-strike distance of ~200 km of the megathrust. Despite fillingthis 200 km wide region, the largest earthquake was Mw=6.2. Thetotal seismic moment release for these swarms was~Mw=7.0.

Several of the megathrust swarms (e.g. Figs. 1 and 8, and the 2006Copiapo, Chile swarm (Holtkamp et al. (2011))) show apparentalong-strike migration of epicenters over time. Fig. 8 shows this alongstrike propagation for the 2008 Tonga swarm and give epicentralpropagation velocities of 8.5±1.9 km/day. However, the PDE catalogonly has enough spatial resolution to show propagation or expan-sion of epicenters for the largest and broadest swarms, so potentialmigration of swarms at smaller scales cannot be determined from thisstudy.

4. Discussion

We document ~5500 individual earthquakes within the 180megathrust swarms which have a total magnitude of Mw=8.0. Of

Fig. 7. An intriguing succession of MSAS and swarm sequences over a 2 yr period in Vanuatu. Layout is similar to Fig. 1. MSAS sequences marked by stars are labeled 1–3 and swarmsare labeled a–f. The largest earthquake in the section of arc south of−21°S and west of 171°E is Mw=6.2, remarkable considering that nearly the entire ~20,000 km2 area producedseismicity.

221S.G. Holtkamp, M.R. Brudzinski / Earth and Planetary Science Letters 305 (2011) 215–225

~286,000 earthquakes in these regions, ~9000 (3%) are associatedwith earthquake swarms, but the swarm earthquakes account for only0.1% of the moment release. Despite the small percentage of earth-quakes and moment release, we believe swarms on the megathrustcan give valuable insight into the physical properties of the megath-rust. In particular, we find that aseismic slip may be an importantfactor in the generation of earthquake swarms, and that earthquakeswarms may indicate variations in fault properties (i.e., coupling) andthe limits of large megathrust earthquakes.

4.1. Magnitude-duration relations of earthquake swarms

Since earthquake swarms generally have abrupt onset andtermination of seismicity with respect to the background seismicityrate, we can use them to quantify the duration of each earthquake

swarm. We find that megathrust swarms last as short as less than aday and as long as several months. Combining this duration withthe total cumulative seismic moment for each swarm, Fig. 9 showsthat the magnitude-duration relation of circum-Pacific earthquakeswarms matches fairly well with the proposed scaling law for slowearthquakes by Ide et al. (2007). However, our selection criteria andthe magnitude completeness of the PDE catalog restrict the domainin Fig. 9 which we are able to sample. While we are not able to quan-tify this restriction absolutely, we have highlighted the portion ofFig. 9 sampled by our search and warn that, to some extent, the ab-solute position of the swarm points in this figure follows from ourmethodology.

We note that there is not a clear linear trend in our magnitude vsduration observations. There are potentially large and unquantifi-able errors of duration and magnitude estimates. For duration, we

Fig. 8. Example of swarmwith amigration of epicenters. Layout is similar to Fig. 1. (Top Panel) Map view of the seismicity displayed in themiddle panel and associated with the 2008Tonga/New Zealand swarm. (Middle Panel) Earthquake magnitude vs. time for ~7 weeks around the swarm. (Bottom Panel) Along strike migration of epicenters.

222 S.G. Holtkamp, M.R. Brudzinski / Earth and Planetary Science Letters 305 (2011) 215–225

have tried to address this by only showing megathrust swarms thatoccur in areas of low background seismicity (event magnitudes~1 Mw above background seismicity or rates several orders ofmagnitude higher than background seismicity rates) in order toreduce the impact of incorrectly estimating the duration of an event.We are left with a little over half (94) of the originally detectedmegathrust swarms that are then plotted in Fig. 9. For furtherjustification of the culling process for swarm durations, seeSupplementary Figs. S1 and S2.

For the estimate of magnitude, it is not clear how the aseismicmoment release, which is plotted in the Ide et al. (2007) figure,should correspond with seismic moment release, which we plot ontop of the Ide et al. (2007) figure. The ratio of seismic to aseismicmoment releasemay vary by region, and thus perhaps be governed bycoupling, stressing rate, or some other factor. For example, lineartrends consistent with the slow slip scaling law can be shown for the

Izu-Bonin-Mariana (IBM) and Kurile-Kamchatka regions (Supple-mentary Fig. S6). If swarms in these regions are controlled by slowslip, this would imply that the ratio of seismic to aseismic momentrelease is greater than 1 for IBM and less than 1 for Kurile-Kamchatka.There has also been recent discussion on whether individual slow slipphenomena show the same linear scaling relation between momentand duration as the overall trend (e.g., Japan slow earthquakes, shownon the original Ide et al. (2007) plot, and Houston (2008) on lowfrequency earthquakes).

So, while individual earthquakes within the swarms follow thetraditional scaling law for earthquakes, the correlation of swarmswith other slow earthquake processes implies that slow slip mayplay a causative role in the occurrence of earthquake swarms. In fact,correlation between earthquake swarms and aseismic slip hasalready been observed by Lohman and McGuire (2007), Ozawa etal. (2007), andWolfe et al. (2007) in other settings. Aseismic slip has

Slow e

arth

quak

es

Regular earthquakes

Gap1 hour

1 day

1 month

1 year

LFE

VLF

SSE

ETS

Silent EQ

a

b

cd

e

Mw

1 2 3 4 5 6 7

-2

-1

0

1

2

3

4

5

6

7

8

9

log

[Cha

ract

eris

tic d

urat

ion

(s)]

10 11 12 13 14 15 16 17 18 19 20 21

log [Seismic moment (N m)]

SlowSlowea

rthqu

ake

ear

thqu

ak

thq

ReRegguulalar er eararttrrr hqhquakesuakes

1 hour

1 day

1 month

1 year

LLFE

VLF

a

b

c

Magnitude Threshold

TemporalProximityThreshold

Megathrust earthquake swarms

Fig. 9. Magnitude–duration relationships for earthquake swarms and MSAS sequencesplotted on top of Fig. 2 from Ide et al. (2007). Gray shaded region indicates areas whereglobal seismicity catalogs cannot sample. Duration is time between the first and lastevents in the swarm or sequence, and magnitude is the total seismic moment ofearthquakes that comprise the swarm or sequence.

223S.G. Holtkamp, M.R. Brudzinski / Earth and Planetary Science Letters 305 (2011) 215–225

been documented in every major subduction zone with sufficientgeodetic observation capabilities, either in the form of slow slipevents (e.g. Japan (Hirose et al., 1999), Cascadia (Rogers andDragert, 2003), New Zealand (Douglas et al., 2005), Alaska (Ohtaet al., 2006), and Central America (Kostoglodov et al., 2003)) oraseismic afterslip (e.g. South America (Pritchard et al., 2007),Kamchatka (Bürgmann et al., 2001), and Sumatra (Hsu et al.,2006)). Aseismic slip events commonly show along strike propaga-tion velocities of ~10 km/day (Boyarko and Brudzinski, 2010;Lohman and McGuire, 2007; Shelly et al., 2007; Wech and Creager,2008), so the observation that some swarms appear to bepropagating with this approximate velocity (Fig. 8) also suggeststhat aseismic slip may be an important factor.

Based on the magnitude-duration scaling law proposed in Fig. 9, aMw=8+ swarm could last on the order of years to decades (Meadeand Loveless, 2009), such that one would currently be indistinguish-able from the background seismicity rate. This is noteworthy as itcautions against over-interpreting that our catalog of megathrustswarms saturates at magnitude ~7.0, which could imply there is alimiting factor on source size for the swarms. We do not appear tohave enough time or catalog resolution yet to determine if the sizeof megathrust swarms saturates. We see that volcanic swarms startto saturate between Mw=6 and Mw=6.5. We expect the size ofvolcanic swarms to saturate as there is simply not enough fault areaaround a volcano to produce large slip areas.

4.2. Relationships between earthquake swarms and the megathrustseismogenic zone

Holtkamp et al. (2011) noted that in South America, swarms thatconcentrated along the Carnegie Ridge in Ecuador and the NazcaRidge in Peru occurred in areas of long standing seismic gaps along themegathrust seismogenic zone. If these earthquake swarms are causedby aseismic moment release, this would suggest that these areas are

not accumulating significant long term strain and may never ruptureas part of a large megathrust event. If the extensive and pervasiveswarms we find in Marianas-type subduction zones such as Vanuatuand Tonga (Supplementary Figs. S4 and S5) are releasing significantmoment aseismically, this may help explain why earthquakes in theseregions do not exceedMw~7, as a significant area of themegathrust isfrequently releasing moment aseismically and preventing megathrustrupture growth through these regions.

To test the hypothesis that swarms are affecting the megathrustrupture cycle by presenting barriers to rupture propagation, eitherby releasing strain aseismically or signifying an area of the plateinterface that is not accumulating significant strain, we haveattempted to quantify the pervasiveness of swarms and comparethis figure to key subduction parameters. To quantify the pervasive-ness of swarms, we measure the largest along strike distance(separation, or gap) that has not had an earthquake associatedwith an earthquake swarm. We measure this for each 500 kmsegment along strike, defined by Syracuse and Abers (2006). Foreach along strike segment, we also characterize the largest mega-thrust earthquake in the past century (for earthquakes since 1964,we filter the CMT for focal mechanisms with plunges greater than45° and for earthquakes prior to that we perform a literature search).This allows us to compare our measurement of swarm pervasiveness(inversely swarm separation), largest characteristic earthquake, andkey subduction parameters listed in Syracuse and Abers (2006)(Fig. 10).

If swarms represent barriers to megathrust earthquake propaga-tion, we could expect that large gaps between earthquake swarms aremore likely to rupture with a single earthquake. In Fig. 10 (a), we seethat for regions with swarm separations greater than 600 km, 100% ofthe regions had earthquakesMwN8. It appears that fault sectionswithlarge swarm gaps are more likely to rupture with larger magnitudeevents. This conclusion is reinforced by the observation that thenorthern 10° of Tonga is nearly devoid of megathrust swarms andcontains three times as many Mw7 earthquakes as Kermadec to thesouth, which is almost saturated with earthquake swarms (back-ground seismicity does not vary much between these two regions, sothere is little chance for bias).

If swarms are a result of decreased coupling, we might expect atrend with subducting plate dip angle (higher dip angles are oftenassociated with slab roll back and trench retreat which results inlittle to no forearc subduction erosion, indicating that there isdecreased coupling on the interface). Fig. 10 (b) shows that higherdip angles (dip anglesN50°) are associated with smaller swarmseparations, indicating that decreased coupling may aid in swarmgeneration. Some potential correlations with plate age/thermalparameter also exist for larger values of age/thermal parameterand swarm separation: large age/thermal parameter is alwaysassociated with small swarm separation and large swarm separationis always associated with small age/thermal parameter. While wefind broad positive correlations between some parameters, we findno evidence for linear relationships at this point. Nevertheless, eachpiece of evidence suggests that swarms thrive on a weakly coupledinterface.

5. Conclusions

We present results from a search of the PDE catalog forearthquake swarms along major circum-Pacific subduction zones,finding 182megathrust and 68 volcanic earthquake swarms. Many ofthese are documented in literature and most are likely felt by localpopulations, but our study presents the first comprehensive catalogof these events in the shallow subduction environment.

Preliminary analysis of the individual swarms in the catalogreveals 3 main discussion points. (1) For several swarms which arelarge enough, we notice an apparent along strike migration of

0 20 40 60 80 100 120 140Incoming Plate Age (Ma)

0 20 40 60 80 100Plate Convergence Velocity (mm/yr)

10 20 30 40 50 60 70 80 90Dip angle (degrees)

0

200

400

600

800

1000

1200

1400

Sw

arm

Sep

arat

ion

(km

)

0

200

400

600

800

1000

1200

1400

Sw

arm

Sep

arat

ion

(km

)

0

200

400

600

800

1000

1200

1400

Sw

arm

Sep

arat

ion

(km

)

0

200

400

600

800

1000

1200

1400

Sw

arm

Sep

arat

ion

(km

)

0

200

400

600

800

1000

1200

1400

Sw

arm

Sep

arat

ion

(km

)

6 7 8 9 10Largest EQ in Region

0 10 20 30 40 50 60 70 80Thermal Parameter

Fig. 10. Investigation of controls on swarm pervasiveness represented by swarm separation, which is the largest gap between swarms measured in each region. Key subductionparameters are from (Syracuse and Abers, 2006) for 500 km along strike sections of circum-Pacific margins used in this study.

224 S.G. Holtkamp, M.R. Brudzinski / Earth and Planetary Science Letters 305 (2011) 215–225

epicenters at a rate of ~10 km/day, a rate typically associated withaseismic slip events such as ETS events. (2) We show that swarms aremore common in some subduction zones than others, perhaps morepervasive in Marianas-type margins. (3) Swarms can cover large (e.g.,N~10000 km2) areas without the occurrence of a large earthquake, aswas the case in Vanuatu, 1980–1982, in which a series of swarms filledin 200 km along strike of the megathrust without an earthquakegreater than Mw=6.2.

Further analysis shows that earthquakes associated with ourdetected swarms have b-values between 1.5 and 2. We quantify theswarminess of our catalog by showing that the number of events inthe swarm is not a function of the largest event, that there is anapproximately constant rate of seismicity which lasts until after themean event timing in the swarm, and that the timing of the largestearthquake is identical to the timing of any other earthquake (thelargest earthquake is just another earthquake). These characteristicsof swarminess, along with the observation that swarms can cover amuch larger area than their cumulative moment release wouldsuggest, rule out a simple static stress triggering driving mechanismfor this catalog of earthquake swarms. Since earthquake swarmsgenerally have abrupt onset and termination, we use them toquantify the swarm duration. We found that the moment-durationrelationship for swarms agrees remarkably well with the momentproportional to duration relationship presented by Ide et al. (2007)for slow earthquakes, as opposed to the moment proportional to thecube of duration for traditional earthquakes. These pieces ofevidence lead us to suggest that aseismic slip is an importantmechanism for the generation of megathrust earthquake swarms incircum-Pacific subduction zones. Additionally, specific fault prop-erties along Marianas-type margins may allow for earthquakeswarms to occur regularly but other margins may rely on other

features, such as the subduction of a ridge or seamount, to occur. Thepervasiveness of earthquake swarms along margins such as Tonga-Kermadec and Vanuatu may indicate the release of larger momentaseismically, which would help explain the lack of great (MwN8)megathrust earthquakes along their margins: pervasive strainrelease along the margin prevents the growth of large contiguousruptures.

Supplementarymaterials related to this article can be found onlineat doi:10.1016/j.epsl.2011.03.004.

Acknowledgments

This project was supported by the NSF-EAR/EarthScope CAREERAward 847688 (MB) and a NASA NESSF Fellowship (SH). The researchbuilds on the initial swarm investigation in South America of SH, M.Pritchard and R. Lohman. We thank J. Vidale for providing data andcomments.

References

Benoit, J.P., McNutt, S.R., 1996. Global volcanic earthquake swarm database and preli-minary analysis of volcanic earthquake swarm duration. Ann. Geofisica 39, 221–229.

Bird, P., 2003. An updated digital model of plate boundaries. Geochem. Geophys.Geosyst. 4, 1027.

Boyarko, D.C., Brudzinski, M.R., 2010. Spatial and temporal patterns of nonvolcanictremor along the southern Cascadia subduction zone. J. Geophys. Res. 115.

Bürgmann, R., Kogan, M.G., Levin, V.E., Scholz, C.H., King, R.W., Steblov, G.M., 2001.Rapid aseismic moment release following the 5 December, 1997 Kronotsky,Kamchatka, earthquake. Geophys. Res. Lett. 28, 1331–1334.

Douglas, A., Beavan, J., Wallace, L., Townend, J., 2005. Slow slip on the northernHikurangi subduction interface, New Zealand. Geophys. Res. Lett. 32.

Evison, F.F., Rhoades, D.A., 1993. The precursory earthquake swarm in New Zealand—hypothesis tests. N Z J. Geol. Geophys. 36, 51–60.

225S.G. Holtkamp, M.R. Brudzinski / Earth and Planetary Science Letters 305 (2011) 215–225

Fischer, T., Horalek, J., 2003. Space-time distribution of earthquake swarms in theprincipal focal zone of the NW Bohemia/Vogtland seismoactive region: period1985–2001. J. Geodyn. 35, 125–144.

Flueh, E., Vidal, N., Ranero, C.R., Hojka, A., von Huene, R., Bialas, J., Hinz, K., Cordoba, D.,Danobeitia, J.J., Zelt, C., 1998. Seismic investigation of the continental margin off-and onshore Valparaiso, Chile. Tectonophysics 288, 251–263.

Fujinawa, Y., Eguchi, T., Ukawa, M., Matsumoto, H., Yokota, T., Kishio, M., 1983. The 1981earthquake swarm off the Kii peninsula observed by the ocean bottomseismometer array. J. Phys. Earth 31, 407–428.

Heinike, J., Fischer, T., Gaupp, R., Gotze, J., Koch, U., Konietzky, H., Stanek, K., 2009.Hydrothermal alteration as a trigger mechanism for earthquake swarms: theVogtland/NW Bohemia region as a case study. Geophys. J. Int. 178, 1–13.

Hill, D.P., 1977. Model for earthquake swarms. J. Geophys. Res. 82, 1347–1352.Hirose, H., Hirahara, K., Kimata, F., Fukii, N., Miyazaki, S., 1999. A slow thrust slip event

following the two 1996 Hyugada earthquakes beneath the Bungo Channel,Southwest Japan. Geophys. Res. Lett. 26, 3237–3240.

Holtkamp, S., Pritchard, M., Lohman, R., 2011. Earthquake swarms in South America.Geophys. J. Int. In revision GJI–S–09–0667.R1.

Houston, H., 2008. Toward a spectral source model of the ETS process. Eos Trans. AGU89 (53).

Hsu, Y.J., Simons, M., Avouac, J.P., Galetzka, J., Sieh, K., Chlieh, M., Natawidjaja, D.,Prawirodirdjo, L., Bock, Y., 2006. Frictional afterslip following the 2005 Nias-Simeulue earthquake, Sumatra. Science 312, 1921–1926.

Ide, S., Beroza, G.C., Shelly, D.R., Uchide, T., 2007. A scaling law for slow earthquakes.Nature 447, 76–79.

Kostoglodov, V., Singh, S.K., Santiago, J.A., Franco, S.I., Larson, K.M., Lowry, A.R., Bilham,R., 2003. A large silent earthquake in the Guerrero seismic gap, Mexico. Geophys.Res. Lett. 30, 1807.

Lay, T., Wallace, T.C., 1995. Modern Global Seismology. Academic Press.Lemoine, A., Madariaga, R., Campos, J., 2001. Evidence for earthquake interaction in central

Chile: the July 1997-September 1998 sequence. Geophys. Res. Lett. 28, 2743–2746.Llenos, A.L., McGuire, J.J., Ogata, Y., 2009. Modeling seismic swarms triggered by

aseismic transients. Earth Planet. Sci. Lett. 281, 59–69.Lohman, R.B., McGuire, J.J., 2007. Earthquake swarms driven by aseismic creep in the

Salton Trough, California. J. Geophys. Res. Solid Earth 112, B04405.Matsuzawa, T., Uchida, N., Igarashi, T., Okada, T., Hasegawa, A., 2004. Repeating earth-

quakes and quasi-static slip on the plate boundary east off northern Honshu, Japan.Earth Planet. Space 56, 803–811.

Meade, B.J., Loveless, J.P., 2009. Predicting the geodetic signature of mw 8 slow slipevents. Geophys. Res. Lett. 36.

Mogi, K., 1963. Some discussions on aftershocks, foreshocks, and earthquake swarms—the fracture of a semi finite body caused by an inner stress origin and its relation tothe earthquake phenomena. Bull. Earthquake Res. Inst. 41, 615–658.

Mora, C., Comte, D., Russo, R., Gallego, A., Mocanu, V., 2008. The Liquiñe-Ofqui faultsystem and its relationship to the Aysen (Southern Chile) 2007 seismic swarm. EosTrans. AGU 89.

Ohta, Y., Freymueller, J., Hreinsdottir, S., Suito, H., 2006. A large slow slip event and thedepth of the seismogenic zone in the south central Alaska subduction zone. EarthPlanet. Sci. Lett. 247, 108–116.

Ozawa, S., Suito, H., Tobita, M., 2007. Occurrence of quasi-periodic slow-slip off the eastcoast of the Boso peninsula, Central Japan. Earth Planet. Space 59, 1241–1245.

Pritchard, M.E., Norabuena, E.O., Ji, C., Boroschek, R., Comte, D., Simons, M., Dixon, T.H.,Rosen, P.A., 2007. Geodetic, teleseismic, and strong motion constraints on slip fromrecent southern Peru subduction zone earthquakes. J. Geophys. Res. Solid Earth 112.

Rogers, G., Dragert, H., 2003. Episodic tremor and slip on the Cascadia subduction zone:the chatter of silent slip. Science 300, 1942–1943.

Shelly, D.R., Beroza, G.C., Ide, S., 2007. Non-volcanic tremor and low-frequencyearthquake swarms. Nature 446, 305–307.

Shibutani, T., Nakao, S., Nishida, R., Takeuchi, F., Watanabe, K., Umeda, Y., 2002. Swarm-like seismic activity in 1989, 1990 and 1997 preceding the 2000 Western Tottoriearthquake. Earth Planet. Space 54, 831–845.

Slavina, L.B., Pivovarova, N.B., Levina, V.I., 2007. a study in the velocity structure ofDecember 5, 1997, M-w=7.8 Kronotskii rupture zone, Kamchatka. J. Volcanol.Seismology 1, 254–262.

Stein, S., Wysession, M., 2003. An Introduction to Seismology, Earthquakes, and EarthStructure. Blackwell Publishing.

Sykes, L.R., 1970. Earthquake swarms and sea-floor spreading. J. Geophys. Res. 75, 6598.Sykes, L.R., 1971. Aftershock zones of great earthquakes, siesmicity gaps, and

earthquake prediction for Alaska and Aleutians. J. Geophys. Res. 76, 8021.Syracuse, E.M., Abers, G.A., 2006. Global compilation of variations in slab depth beneath

arc volcanoes and implications. Geochem. Geophys. Geosyst. 7.Utsu, T., Seki, A., 1954. Relation between the area of aftershock region and the energy of

the main shock. J. Seism. Soc. Jpn 7, 233–240.Vidale, J.E., Shearer, P.M., 2006. A survey of 71 earthquake bursts across southern

California: exploring the role of pore fluid pressure fluctuations and aseismic slip asdrivers. J. Geophys. Res. 111.

Wech, A.G., Creager, K.C., 2008. Automated detection and location of Cascadia tremor.Geophys. Res. Lett. 35, L20302.

Wolfe, C.J., Brooks, B.A., Foster, J.H., Okubo, P.G., 2007. Microearthquake streaks andseismicity triggered by slow earthquakes on the mobile south flank of KilaueaVolcano, Hawaii. Geophys. Res. Lett. 34, L23306.

Zobin, V.M., 1996. Earthquake clustering in shallow subduction zones: Kamchatka andMexico. Phys. Earth Planet. Inter. 97, 205–218.