Tectonophysics 717 (2017) 214–225

Contents lists available at ScienceDirect

Tectonophysics

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

Moment tensor solutions along the central Lesser Antilles using regionalbroadband stations

O'Leary González a,b,⁎, Valerie Clouard a, Jiri Zahradnik c

a Observatoire Volcanologique et Sismologique de Martinique, Institut de Physique du Globe de Paris, Franceb Centro Nacional de Investigaciones Sismológicas (CENAIS), Ministerio de Ciencia, Tecnología y Medio Ambiente, Cubac Faculty of Mathematics and Physics, Charles University in Prague, Czech Republic

⁎ Corresponding author at: CENAIS, Cuba.E-mail address: oleary@cenais.cu (O. González).

http://dx.doi.org/10.1016/j.tecto.2017.06.0240040-1951/© 2017 Published by Elsevier B.V.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 16 November 2016Received in revised form 13 June 2017Accepted 20 June 2017Available online 22 June 2017

Using waveform data gathered from the seismological networks of the Lesser Antilles, we calculate 38 momenttensors for earthquakes with M ≥ 3, from 2013 to middle of 2015 by full waveform inversion. Nine of these mo-ment tensor solutions are in good agreement with those previously reported by other institutions, it providessome guarantee for 29 newmoment tensors for the central region of the Lesser Antilles. For earthquakes withinthe upper Caribbean lithosphere, our results evidence that extensional and strike-slip focal mechanisms are pre-dominant, resulting from the intra-plate deformation produced by the subduction of the North America andSouthAmerican Plates under the Caribbean Plate,whereas very few thrust events are observed. For deeper earth-quakes (N90 km), our results compare well with older focal mechanisms from previous studies, showing normaloblique or strike slip faulting within the subducted slab. However, the inversion for most of the deeper events isless reliable (as documented, for example, by their larger condition numbers). We use the newly obtained mo-ment magnitudes to estimate the scaling relationship with the local magnitude MLv computed by the regionalseismic network. The two magnitudes are consistent for earthquakes with magnitudes M N 4, with a slopeclose to unity. Further work is needed to precise the scaling relation for M b 4.

© 2017 Published by Elsevier B.V.

Keywords:Moment tensorFocal mechanismLesser AntillesISOLASubduction zone

1. Introduction

The complex seismicity in the Lesser Antilles region is mostly relatedto the subductionof theNorth and SouthAmericanplates beneath the Ca-ribbean plate (Fig. 1). Regional earthquake catalogs (Bengoubou-Valériuset al., 2008; Massin et al., in prep.) show that the subduction processcreates a 100–200 km wide deformation zone accompanied by shallowseismicity, from the trench to the Lesser Antilles Islands (where the activevolcanic arc is located) and an intermediate depth seismicity along theslab observed down to 200 km. The instrumental seismicity indicatesimportant lateral variations along the subduction zone, in particular inthe central Lesser Antilles (between 16°N and 15°N, a N150 kmwide re-gion, where shallow seismicity is low and thrust earthquakes are rare(Fig. 1). Nevertheless the eastern Caribbean boundary has the classicalcharacteristics of a subduction zone (e.g., Byrne et al., 1988). To thesouth of the Tiburon Ridge is located a thick accretionary prism, whereturbidites from the Orinoco sediments are accreted (Westbrook, 1975;Westbrook et al., 1982; Pichot et al., 2012). In this area, a seismic gaphas been identified and explained by a low coupling due to high sedimentpresence (Dorel, 1981;Wadge and Shepherd, 1984). The Tiburon Ridge is

the northern limit of these sediments on theAtlantic seafloor (Brown andWestbrook, 1987). The area between the Tiburon Ridge and BarracudaRidge represents the diffuse plate boundary zone between the Southand the North America plates (Patriat et al., 2011). The Caribbean plateconvergence with the American plates is stable since ca. 20 Myr (e.g.,Bouysse and Westercamp, 1990; Meschede and Frisch, 1998; Pindelland Kennan, 2009), with a slow frontal to transitional convergence rateof ~2 cm/year (e.g., DeMets et al., 2000; Lopez et al., 2006; Symithe etal., 2015). Recent GPS data analysis indicates no relative motion of theLesser Antilles islands with respect to the Caribbean plate, and the ob-servedCaribbeanplate active faultswithin the arc and the forearc only ac-cumulate strain at a rate of less of 2mm/yr (Symithe et al., 2015). Amajorissue in the region is the estimation of the seismic and tsunami potentialof the subduction zone. Only few large historical earthquakes have oc-curred, the 1843 and the 1839 events (Fig. 1) being the largest ones(Bernard and Lambert, 1988; Feuillet et al., 2011; Hough, 2013). Basedon regional reported intensities and damages, Bernard and Lambert(1988) proposed amagnitude between 7,5 and 8 and an interplate thrustevent for the 1843 earthquake, while adding felt reports from NorthAmerica leads to a magnitude N8.4 (Hough, 2013), with large uncer-tainties on its location. However, the lack of tsunami associated to the1843 event is puzzling, whereas in the case of the 1839 earthquake, asmall tsunami reported in Martinique (Clouard et al., 2017) is consistent

Fig. 1. Regional settings and GCMT focal mechanisms for earthquakes with M N 5, from 1950 to 2015 (Dziewonski et al., 1981). The convergent plate velocity comes from DeMets et al.(2000). The subsurface dimensions of 1839 and 1843 events are obtained from Wells and Coppersmith's (1994) rupture area regression versus magnitude, with epicenter andmagnitude of 8.3 for 1843 following Hough (2013) and epicenter and magnitude of 7.5 for 1839 following Bernard and Lambert (1988).The red dashed line is the deformation front ofthe accretionary prism. The black rectangle bounds the study area. The upper inset shows the location of the Eastern Caribbean. Bathymetry comes from ETOPO1 (Amante and Eakins,2009). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

215O. González et al. / Tectonophysics 717 (2017) 214–225

with a 7.5 thrust event as proposed by Bernard and Lambert (1988) andFeuillet et al. (2011).

For the 1950–1978 period, Stein et al. (1982) determined 17 focalmechanisms in the Lesser Antilles and found only one thrust eventnear Antigua. Thus, they proposed that the central Lesser Antilles islargely decoupled and aseismic. A recent GPS study of the area arrivesat a similar conclusion (Symithe et al., 2015). The instrumental seismic-ity since 1976 has registered only 6 earthquakes with magnitude N6(Fig. 1) in the central part of the Lesser Antilles. Except the Martinique2007 earthquake (M= 7.3) at intermediate depth, they are all shallowand too sparse to understand the strain partitioning of the area nor thestate of coupling of the subduction.

The French Volcanological and Seismological Observatories in Gua-deloupe and Martinique (OVS-IPGP), in addition with the Seismic Re-search Center (SRC-UWI) are the regional seismic network operatorsin charge of the Lesser Antilles seismicity monitoring. Since 1973, theyroutinely share their data and determine the hypocenters and localmagnitudes of the earthquakes. During the four ultimate years, inorder to improve the regional seismic monitoring and to contribute tothe Caribbean Early Warning System, the OVS-IPGP and SRC-UWI

have greatly increased their detection networks with the installationof several new high quality broadband stations (Anglade et al., 2015)along the Lesser Antilles arc. In addition, an effort has been done toproduce a common regional catalog of the seismicity, named the CDSAcatalog (Centre de Données Sismologiques des Antilles, http://www.seismes-antilles.fr), based on a compilation of all phases from theOVS-IPGP, the SRC-UWI and others seismic operators (Massin et al., inprep.). However, for a better understanding of the focal processes andtheir relationship with the tectonic environment, the style of faultingand moment tensors also need to be determined.

For earthquakes with M N 5, these parameters are routinelydetermined by international agencies by inversion of seismic wave-forms recorded in the far field (e.g., GCMT, Project http://www.globalcmt.org, IRIS catalog, http://ds.iris.edu, GEOSCOPE Observatory,http://geoscope.ipgp.fr, etc.). However, up to date, only 48moment ten-sor solutions have been determined in our studied region since 1979(Fig. 1), 12 of them in the last decade. Focal mechanisms can also be de-termined by other methods based on P polarities and S to P amplituderatios (e.g., Buforn, 1994), but these methods need a good azimuthalcoverage of the study area. In our study area, using P-wave polarities

Fig. 2.Map of the stations (triangles) used for determining the moment tensor solutions. The grey zone is the study area and detail of the Martinique Island is zoomed.

Fig. 3. Representation of a grid search for the 20140225 09:58:15.80 UTC earthquake. a) Vertical: in 10 trial source positions around the hypocenter (starting at the depth of 5 km, verticalstep 3 km), and around the origin time in the range of ±3 s (step of 0.45 s). b) Horizontal: at the previously determined depth of 14 km, the horizontal and vertical axis numbers are thesource position numbers (with a step of 3 Km). The filling colors of the beach balls correspond to the double-couple percentage (DC scale on the upper right), the best-fitting solution (thatof the largest correlation) is the red beach ball. This is a representative example, the real grid search differs on the “step” as mentioned in the text. (For interpretation of the references tocolour in this figure legend, the reader is referred to the web version of this article.)

216 O. González et al. / Tectonophysics 717 (2017) 214–225

Fig. 4. Example of thewaveform fit for the 20150514 04:16:50 UTC earthquake. Black lines represent the real seismogram, red lines the synthetic one generated for the calculatedmomenttensor solution. Frequency ranges from 0.06 to 0.09 Hz. Blue numbers are the variance reduction of the individual components. (For interpretation of the references to colour in this figurelegend, the reader is referred to the web version of this article.)

217O. González et al. / Tectonophysics 717 (2017) 214–225

recorded by a dense offshore temporary network, Ruiz et al. (2011) de-termined 22 focal mechanisms offshoreMartinique and Dominica. Theyfound a diffuse seismicity concentrated in the inner forearc, not related

Fig. 5. Fault plane solutions (brown beach balls) of the earthquakes used during thecalibration procedure (Set 1) and of some older events (green beach balls), before theinstallation of the new regional broadband stations, from GCMT and Russo et al., 1992.The black rectangle shows the study area.

to any specific structure, mostly located near the base of the Caribbeancrust and near the interface. However, due to insular configuration ofthe Lesser Antilles permanent seismic network and the fact that 95%of the seismicity is offshore, the seismic station distribution (Fig. 2)does not allows to use these methods.

The moment tensor determination by full waveform inversion hasbecome a routine practice at many observatories due to the availabilityof high quality data (Havskov and Ottemöller, 2010). In this paper, wepropose to use the newly up-graded Lesser Antilles seismic networkto provide a complete moment-tensor catalog of low-magnitudeearthquakes up to middle of 2015. Another innovation is that wecomplement the newly obtained moment tensor by their uncertaintyestimate. The moment magnitudes (Mw) that we obtain are also com-pared with the commonly used local magnitudes MLv (measured onthe vertical component). Preliminary tectonic implications of the resultsare discussed.

2. Data and method

2.1. Data selection

For this study, we use the CDSA catalog to identify earthquakes withM N 3.0, whose signal has been registered by at least 4 regional broadband 3-component seismic stations (Fig. 2 and the Electronic Supple-ment #1S summarize the most important features for each station).Within these data, we select two sets of earthquakes:

2.1.1. Set 1It is selected for calibration purpose, and represents 9 regional earth-

quakes for which moment tensors were determined by internationalagencies.

2.1.2. Set 2Is a set of 29 earthquakes, inside our study area andwith good signal

to noise ratio in the frequency band of the regional waveform inversion(typically below 0.1 Hz), for which the moment-tensor have not yetbeen calculated.

The velocity model for the waveform inversion is taken from Dorel(1978). This 1D model consists of three crustal layers with 3, 12, and15 km of thicknesses and P velocities (Vp) of 3.5, 6.0, and 7.0 km/s

Table 1Comparison of the inversion results of this study (Set 1) and from GCMT (Dziewonski et al., 1981 and Ekström et al., 2012).a

Event # Origin timeyyyymmddhh:mm:ss.ss

Lat Lon depth(centroid GCMT)(centroid this study)

Δ(km)

MLv MwGCMT

Mwthis study

MomenttensorGCMT

DC%GCMT

Momenttensorthis study

DC%this study

VR%this study

CNthis study

K°

1 20120704 21:29:27.00 18.32–62.998118.23–62.9478

167.9 5.6 4.9 4.8 87 94 32 5 14

2 20130430 06:56:48.46 17.69–61.975517.64–62.0846

104.6 5.5 5.3 5.3 89 83 68 2.4 4

3 20130708 09:14:18.00 18.24–61.583717.74–61.5334

174.0 5.2 5.1 4.9 55 79 73 4 12

4 20131012 02:10:32.00 10.86–62.157810.90–62.2060

78.4 6.3 6 5.8 70 98 57 4.7 13

5 20140218 09:27:13.92 14.76–58.891414.85–58.8912

182.3 6.5 6.4 6.4 96 58 45 2.2 21

6 20140516 11:01:42.75 17.18–60.241217.09–60.3926

26.0 5.9 5.9 5.9 81 83 50 2.4 14

7 20140915 21:09:08.83 14.34–60.141614.41–60.057

75.1 5.2 5.2 5.0 42 47 88 3.2 14

8 20141207 15:35:44.00 13.72–60.271513.89–60.194

69.3 4.7 4.9 4.5 b 92 75 5.5 28

9 20141219 19:49:30.19 16.37–61.8211616.19–61.78120

163.3 5.7 5.6 5.5 87 91 57 3.1 14

a Δ hypocentral distance of the nearest station used for the inversion; DC is are the percent of the double couple, VR the variance reduction and K is the Kagan-angle deviation of ourresult from GCMT.

b Unknown information.

218 O. González et al. / Tectonophysics 717 (2017) 214–225

respectively; the mantle model is represented by a half space of Vp8.0 km/s. The Vp/Vs ratio for all the layers is 1.76. Thismodel is routinelyused for earthquake location by the OVS-IPGP network in the Lesser An-tilles (Bengoubou-Valérius et al., 2008). The model values were derivedfrom seismic refraction profiles, averaged for land and oceanic regionsof the Lesser Antilles. Density values come from the relationship ofNafe and Drake (1960), while the attenuation quality factors Qp andQs are determined by the regression of Graves and Pitarka (2004). Theearthquake parameters are those of the CDSA catalog and theinstrument response parameters come from the IPGP Data Center(http://centrededonnees.ipgp.fr). We verify the correct sensor orienta-tion, using the 3-component P-polarities and azimuths followingHavskov et al. (2002). The stations with inconsistent results areremoved from the inversion.

In order to select the optimum frequency band for filtering duringthe inversion, each component is initially evaluated in several frequencybands using the most updated version of SAC software (Goldstein,1999). ISOLA codes (Sokos and Zahradník, 2008 and 2013) are used tocalculate the Signal to Noise Ratio (SNR) of the events and to identifyand remove the records spoiled by instrumental disturbances(Zahradník, and Plešinger, 2005 and 2010; Vackář et al., 2015). For

each earthquake, at least 10 usable components and a SNR ≥ 2 in thefrequency band used for the inversion are required.

The (deviatoric) moment tensors are calculated with ISOLA by fullwaveform inversion, using the least-squared method, while thecentroid position and time are grid-searched. The spatial grid search isperformed first vertically below epicenter in an interval around thehypocenter depth (with a step of 1 km), and in an interval of ~3 saround the origin time (step of 0.09 s) (Fig. 3a). The second grid searchis usually performed in a horizontal plane with typical steps rangingfrom 2 to 4 km (Fig. 3b), fixing the reference depth from the first search.The best-fitting solution (Fig. 4) is characterized by the VarianceReduction percentage (VR), or correlation (where correlation squaredequals VR), the Double-Couple percentage (DC) and the ConditionNumber (CN). CN serves for relative evaluation of the moment-tensorresolvability: the lower CN means the better resolvability (for exactdefinitions, see Krizova et al., 2013).

2.2. Calibrations results (Set 01)

In order to check the consistency of the inversion parameters tobe used (e.g. instrumental calibration of the seismic stations, velocity

Fig. 6. New focal mechanisms from this study (numbering from Table 2): a) Projection ofthe new focal mechanisms along profil P), in km. 0 is taken at the volcanic arc (redtriangle). The horizontal dashed line is the average depth of Caribbean crust. The thickgrey line represent the Wadati-Benioff plan: it is constrained by the results from Koppet al. (2011), Evain et al. (2013) and Laigle et al. (2013) and prolongated deeper inagreement with our data; b) The bathymetry is a compilation of ETOPO1 (Amante andEakins, 2009) and Aguadomar cruise (Deplus, 1998). The red dashed line is the limit ofthe backstop (Laigle et al., 2013). Red triangles are the active volcanoes and the volcanicarc is the orange dashed line. (For interpretation of the references to colour in this figurelegend, the reader is referred to the web version of this article.)

219O. González et al. / Tectonophysics 717 (2017) 214–225

model, attenuation values, etc.), the moment tensors of 9 earthquakes(Fig. 5) previously determined by other sources and archived in IRIS cat-alog in http://ds.iris.edu/spud/momenttensor (Dziewonski et al., 1981,and Ekström et al., 2012) are recalculated using the regional stationsof our study. Due to the magnitude range of the selected earthquakes(Mw N 4.5), the frequency band for the inversion is chosen in a range

between 0.01 and 0.06 Hz. This bandwidth is consistent with the filterssuggested by Dreger (2002) for these magnitude values. The results arepresented in Table 1.

Themean departure from theMw values of GCMT catalog is as smallas 0.12. The differences are found more frequently for events far fromour study area, where the scaling law for MLv as well as the crustalmodel (Dorel, 1978) and their corresponding attenuation values areless appropriate.

The approach proposed by Kagan (1991) describes the differencebetween two pure double-couple source models; the Kagan's angle(K-angle) is used to quantify the difference between each new focalmechanism and its corresponding GCMT solution. Our results show ingeneral, relatively low Kagan's angle values, formost of the earthquakes(≤21°) and their mean value of 15°, which indicates a good agreementwith GCMT. Such results with our first dataset of highest magnitudeevents, clearly indicates that the instrumental calibration and the veloc-ity model used are relevant, and enables the process of weaker events.

3. Newmoment tensors solutions (Set 02)

Our full waveform inversion provides a set of 29 new moment ten-sor solutions for the Central Lesser Antilles (Fig. 6). Results are reportedin Table 2 and in the following, each event is identified by the numberthat appears on this table. We now use again the Kagan's angle but fora completely different purpose, to check the uncertainty of themomenttensor solutions as the deviation of each acceptable solution from thebest-fitting one (Zahradník and Custodio, 2012). The K-angle values inTable 2 represents the mean of the deviation of all solutions (see exam-ple in the K-angle histogram, Fig. 7), and all of them are in acceptablerange (Zahradník and Custodio, 2012). Overall, the solutions are charac-terized by adequate values of the variance reduction percentage (VR),the double-couple percentage (DC) and the condition number (CN).For easternmost events (#11 to 14), in spite of the in general large NCand SNR, relatively large K-angle and CN values as well as low VR prob-ably reflect the negative influence of the insular configuration of ournetwork geometry. The 2 least resolved events are #17 and #28,which coincide with relatively large CN and small number of compo-nents (NC) used during the processing. Detailed results for all eventsare in the Electronic Supplement S2 as well as the waveform fit in theElectronic Supplement S3.

4. Discussion

4.1. Relationship between the focal mechanism and the source depth

The shallow seismicity accommodates the intra-Caribbean deforma-tion by normal, normal-oblique or strike slip faulting in the forearc(Stein et al., 1982; Gagnepain-Beyneix et al., 1995). Among the 29new moment tensors obtained, 18 occurred in the upper lithosphereat shallow depth (Fig. 6). To the north, events #5 and #24, at respective-ly 22 and 34 km-depth, are located on the southern wall of the NNEtrending Antigua Valley (Fig. 8). Because of the geometry of this andother regional valleys, it has been proposed that they were normalfault-bounded grabens (Feuillet et al., 2002). The normal mechanismswe found confirm that and indicate that this trough is an active grabenassociated to normal faults down to crustal depths. #9 earthquake, lo-cated on the northern wall of Marie-Galante Graben (Fig. 8) coincideswith previous normal earthquakes like 1992-08-03 and 2001-01-05events (Fig. 5, and more information about previous earthquakes usedfor comparison in Electronic Supplement S4) and also indicates an ac-tive graben. #21 normal faulting earthquake is close and similar to theMw = 6.2 2004 Les Saintes earthquake located on the major normalfault of the Roseau Graben extending from Les Saintes to Dominica(Bazin et al., 2010). It is located at the southern extremity of Roseaufault system, below the no longer active Roseau underwater volcano.#15 event is located at the top of the Arawak Valley, which, by analogy,

Table 2New focal mechanisms derived in this study (Set 2).a

Event # Origin timeyyyymmddhh:mm:ss.ss

Lat Lon depth(hypocentr)(centroid)

Min - maxΔ (km)

MLv Mw StrikeDiprakeNP1/NP2

Mon Ten DC%

VR%

NC SNR K° CN

1 20130508 03:45:22.58 14.06–60.571214.03–60.544

44–98 3.8 3.6 277 19 21167 83 108

57 67 15 7 7 2.4

2 20130628 07:53:23.00 14.05–60.581214.12–60.288

53–255 3.6 3.5 224 64–157124 69–27

94 65 21 4 4 2.2

3 20131018 04:01:56.58 16.17–60.684116.12–60.7352

68–209 4.6 4.1 102 76–17310 83–14

70 67 30 5 4 2.3

4 20131022 07:53:17.73 15.01–60.413415.05–60.4146

91–147 4.4 3.9 150 26 74348 66 98

75 50 15 3 4 2.1

5 20131024 23:34:49.47 17.17–61.03417.20–61.0022

93–184 3.9 3.7 326 63–96160 28–78

82 45 18 5 6 4.4

6 20131125 20:22:02.14 15.70–61.25 12015.70–61.31 119

126–178 4.5 4.3 281 34–7988 56–97

71 60 17 6 8 8.7

7 20140110 03:14:02.72 16.15–61.94 18116.15–61.91 185

181–284 4.6 4.4 81 58–7174 84–148

82 42 18 3 6 5.9

8 20140209 10:54:51.05 14.48–60.611714.56–60.509

28–209 4.0 3.6 111 66–78263 26–115

90 70 17 2 5 2.3

9 20140225 09:58:15.80 16.11–61.261416.14–61.2614

35–209 4.2 3.6 235 44–127102 57–60

68 50 20 4 4 2.2

10 20140514 23:44:10.89 14.76–60.452714.87–60.283

56–184 3.7 3.1 331 67–172238 83–23

68 50 19 2 5 4.3

11 20140516 12:30:21.00 17.22–60.173017.14–60.287

140–316 5.2 4.9 291 18–154176 82–73

100 67 23 10 6 3.6

12 20140516 12:34:54:00 17.19–60.153117.11–60.185

138–314 4.9 4.8 304 16–164199 86–75

92 48 15 3 12 7.5

13 20140516 16:52:25.70 17.03–60.311417.11–60.375

114–291 4.9 5.1 287 3–176193 90–85

69 43 30 22 7 11.2

14 20140517 14:47:37.47 17.09–60.321517.20–60.297

118–297 4.5 4.6 302 16–124157 76–81

66 35 16 17 7 3.6

15 20140601 06:06:33.00 15.38–60.721615.36–60.6919

87–147 4.0 3.5 280 41–6164 55–113

70 42 24 2 3 2.1

No. Origin timeyyyymmddhh:mm:ss.ss

Lat Lon depth(hypocentr)(centroid)

Min - maxΔ (km)

MLv Mw StrikeDiprakeNP1/NP2

Foc Mec DC%

VR%

NC SNR K° CN

16 20140617 12:25:44.13 15.73–61.21 10315.73–61.3295

111–154 3.6 3.7 64 47–175330 86–43

88 40 12 2 8 4.8

17 20140710 12:54:59.83 14.90–61.43 17814.98–61.32 159

181–209 4.1 4.2 48 81 173139 83 9

74 54 11 4 18 14.6

18 20140717 04:42:58.41 14.97–60.5258 64–356 4.4 4.3 211 23 110 88 52 51 6 4 2.0

220 O. González et al. / Tectonophysics 717 (2017) 214–225

Table 2 (continued)

No. Origin timeyyyymmddhh:mm:ss.ss

Lat Lon depth(hypocentr)(centroid)

Min - maxΔ (km)

MLv Mw StrikeDiprakeNP1/NP2

Foc Mec DC%

VR%

NC SNR K° CN

15.02–60.3846 10 69 82

19 20140805 20:10:41.21 15.54–60.801815.54–60.7515

69–170 4.3 3.7 355 80–161262 72–10

89 64 33 5 3 2.3

20 20140805 22:14:19.23 15.54–60.801815.54–60.7115

70–170 4.0 3.5 351 82–167259 77–8

90 48 28 2 3 3.0

21 20140809 13:15:50.38 15.70–61.491015.72–61.514

23–113 3.4 3.2 304 73–111176 27–41

91 69 21 3 5 2.7

22 20140818 20:53:14.13 15.35–61.05 11915.35–61.05 120

126–160 3.7 3.6 125 83–42221 48–171

80 46 17 2 8 6.4

23 20141012 09:28:19.88 16.78–60.513216.67–60.5137

87–239 3.9 4.1 5 88 22274 68 177

36 65 19 5 4 2.5

24 20141125 05:33:01.00 16.74–61.112116.74–61.0834

59–118 4.2 3.8 9 27–101201 63–84

96 56 15 2 5 2.6

25 20150214 23:00:07.41 14.70–61.05814.73–61.074

12–20.4 2.9 3.0 287 54–94113 37–85

60 64 13 3 12 5.3

26 20150216 21:52:21.42 15.41–60.133215.45–60.1326

110–307 4.4 4.1 268 43–146152 68–52

78 59 34 5 3 2.3

27 20150228 12:28:04.09 15.38–61.21 13615.41–61.21 137

131–160 4.7 4.6 183 67–15583 67–22

93 82 20 8 11 7.7

28 20150416 14:49:31.71 14.48–60.477914.51–60.3669

72–168 4.3 3.9 244 50 11147 81 139

94 60 12 2 19 7.4

29 20150514 14:16:50.00 13.97–60.465413.95–60.3845

72–120 5.0 4.3 109 45–17211 78–133

76 80 12 10 7 3.2

a Min - max Δ are hypocentral distance of the nearest and farthest station used for the inversion; NC is the number of components used; SNR the signal to noise ratio; K is the meanKagan angle of the family of the acceptable solutions and CN is the condition number. For the remaining symbols, see legend of Table 1.

221O. González et al. / Tectonophysics 717 (2017) 214–225

is another active graben. #26 earthquake is located at the northwesternlimit of the Arawak basin, a deep forearc basin, with the buried southernpart of theKarukera Spur (Evain et al., 2013). It can results from the sub-sidence of the Arawak basin or from the uplift of the Karukera Spur, as itis the case for its northern part. These normal faulting earthquakes mayalso be explained due to the back-arc extension and the correspondinglithospheric boudinage of the Caribbean upper plate. The back-arc ex-tension is a typical process of the West-directed subduction zones(Doglioni and Panza, 2015) due to the kinematics of the plates involvedas described Doglioni et al. (2007).

Events #24 and 29 are respectively normal and oblique normalearthquakes, located at 34 and 45 km-depth. Fig. 6 indicates that theyare potentially within the mantle wedge, above the subducting slab,and below the upper crust. This kind of event has been evidenced inour studied area by Laigle et al. (2013) and related to the “supraslabearthquake” (Uchida et al., 2010). They could indicate that, like inTohoku, the seismogenic portion of the interface could extend deeperthan in other subduction zones.

Events 11 to 14 are aftershocks of the M = 5.9, 16/05/2014 normalearthquake (in Table 1). They are located at the intersection ofsubducted Barracuda ridge with toe of the backstop (Fig. 8). The back-stop is the forearc region significantly stronger than the region justtrenchward of it (Byrne et al., 1993). In the Lesser Antilles case, thebackstop geometry exhibits a contact with the accretionary wedge dip-ping trenchward (e.g., Westbrook, 1975) and it is the island arc crustthat serves as backstop (Christeson et al., 2003). These 4 events are sim-ilar, at 5–7 km-depth, ie. very shallow, with homogeneous magnitudebetween 4.6 and 5.1. The vertical dip angle we found traduces the char-acteristic uplift and faulting of the upper plate over a subducting relief(Dominguez et al., 1998) and contribute to the uplift of the BarracudaRidge observed since the Early Pleistocene related to the bending ofthe lithosphere prior subduction and to the North and South Americaplate motion (Pichot et al., 2012).

Strike-slip mechanisms are sparse in the Caribbean upper litho-sphere and we find only 4 shallow strike-slip events: #19 earthquakeand its aftershock #20, 2 h later, #10 and #2. The lack of shallow

Fig. 7. The uncertainty estimates for the earthquake of 20140225 09:58:15.80UTC: a) the family of the acceptable solutions; b) Histogram of the Kagan's angle of the acceptable calculatedsolutions with respect the best-fitting solution.

222 O. González et al. / Tectonophysics 717 (2017) 214–225

strike-slip events in our studied area, in addition with geodetic analysisof Symithe et al. (2015) is another argument against the northern LesserAntilles forearc bloc proposed by López et al. (2006). Respectively,strike-slip mechanism is the most common mechanism for intraslabearthquakes as seen with #3, 7, 16, 17, 22, 27, 28 and 29 strike slipevents, at depth between 40 and 185 km (Fig. 6). Part of theseearthquakes around 15° may be related not only to the subduction pro-cess but also to the relative displacements between the subductedNorthern and Southern American plates. The 6 deeper earthquakes(depth N 90 km), in general, have relatively larger CNnumber and largeruncertainty (K-angle), but the resulting models have a comparablefaulting type solutions than the older events from GCMT located inthe same region and depth. For example: the pairs of events #16and 2005-02-08 (Fig. 5) and #27 and 1996-09-24 have a normaloblique faulting and events #17 and 1982-01-08 have an strike slipfaulting.

Finally, we find only one area where thrust faulting earthquakesoccur, located eastward Martinique (events # 4 and # 18). These 2events are located near the 1946-05-21 (Ms = 6.0) earthquake(Bernard and Lambert, 1988; Russo et al., 1992), the 1983-03-03 and2008-02-06 (Mw = 5.3) earthquakes, and near several others thrustfaulting earthquakes from Ruiz et al. (2011). This area also correspondsto the rupture area of 1839 earthquake. While our short time windowdatabase does not contain other thrust events, the rupture area of1843 earthquake, from Desirade Island up to latitude 18°N is a secondzone where thrust earthquakes have occured (Fig. 8). Commonly, themajor moment release at subduction zone occurs at the interface be-tween the subducting plate and the overriding plate (e.g., Scholz,1998). However, in the central Lesser Antilles, this coupled interfaceseems to be limited to the 2 areas that we identified, matching the sub-surface rupture areas of the 2 large historical subduction earthquakes.Conversely, in the rest of the studied area, our results are consistentwith the hypothesis that the central Lesser Antilles is decoupled as al-ready proposed by Stein et al. (1982) and more recently by Symithe etal. (2015).

4.2. The Mw versus MLv scaling

The localmagnitude (MLv) is the Richter (1935)magnitude, derivedfrom the maximum amplitude of the velocity signal on the verticalcomponent and an attenuation law. MLv is computed routinely by theregional network operators (e.g., OVSM-IPGP and OSVG-IPGP). Topotentially improve their attenuation parameters and thus homogenizethe catalog, we use our obtained moment magnitudes (Mw) to

investigate their relationship with MLv. It is expected that the Mw –MLv scaling differs for stronger and weaker events. For example, Mw≈ MLv for 4 b M b 6 (Kanamori, 1983) and Mw ~ 0.67MLv for M b 4(Ottemöller and Havskov, 2003). Excluding events 1 and 4 fromTable 1, which are too far from our study region, our data give (Fig. 9):

Mw ¼ 0:94 MLvþ 0:15 for 4bMb6

Mw ¼ 0:53 MLvþ 1:53 for MN4

For M b 4, the latter relationship is less well constrained due to theconsiderable data scattering. More data are needed to precise the Mw-MLv relation for M b 4 and to make a comparison like Bindi et al.(2005) for the North of Italy, Scherbaum and Stoll (1983) for Germany,Zuniga et al. (1988) for Hawaii, etc. Nevertheless, for M N 4, the presentdata shows that MLv is a good proxy to Mw.

5. Conclusions

Using the records of the new broad band seismic network operatingin the Lesser Antilles, a set of 38 moment tensor solutions for earth-quakes with M ≥ 3 has been determined by full waveform inversion.Where other sources exist (like GCMT), our solutions compare wellwith them. The analysis of the new moment tensors for earthquakesin the Caribbean upper lithosphere indicates that normal, oblique andstrike slip faulting are predominant, and reflects the intraplate deforma-tions of the Caribbean plate from the subduction to the volcanic arc, inparticular on the southern walls of most of the regional grabens. Inour study area, only two thrust earthquakes located eastward ofMartinique are present at depth between 40 and 90 km. This scarcityis consistent with GCMT catalog and may be explained, between otherscauses, by the hypothesis of the low coupling subduction in the centralLesser Antilles (e.g. Stein et al., 1982 and Symithe et al., 2015). However,two areasmatching the subsurface rupture areas of the 2 large historicalsubduction earthquakes could define two patches of higher coupling.Intermediate depth earthquakes (deeper than 100 km) are locatedwithin the subducting slab and correspond to a normal oblique or strikeslip faulting related to the subduction process. Nevertheless, in general,the largest uncertainties correspond to the inversion procedures ofthese deeper earthquakes. The comparison of our new Mw with theMLv is consistent with Kanamori (1983) results for M ≥ 4. However,for magnitudes b4, the considerable data scatter requires more infor-mation to establish the Mw-MLv relation. The computation of Mwfrom spectral analysis (Edwards et al., 2010) could provide constrainson this law. This important problem requires further investigation

62˚W 61˚W 60˚W

14˚N

15˚N

16˚N

17˚N

2227

3

4

28

29

18

7

16

22

6

17

1

2

5

8

10

19 20

21

2324

25

26

9

15

1213

1411

Martinique

Dominica

Saint Lucia

Guadeloupe

Les Saintes

Antigua

Marie Galante

Desirade

0

40

80

120

160

200Depth−km

1843

1839

BarracudaRidge

TiburonRidge

Ant

igua

Val

ley

Marie-Galante Graben

Arawak Basin

Desirade Graben

Arawak Valley

Karukera S

pur

Fig. 8.Distribution of existing focal mechanisms in the study area from GCMT, Ruiz et al. (2013), and this study (with numbering from Table 2). Ellipses are the subsurface dimensions of1839 and 1843 events and enclose the two areas where thrust focal mechanisms are found. Names of the submarine features are in brown. The red dashed line is the limit of the backstop(Laigle et al., 2013). Bathymetry is a compilation of ETOPO1 (Amante and Eakins, 2009) and Aguadomar cruise (Deplus, 1998). (For interpretation of the references to colour in this figurelegend, the reader is referred to the web version of this article.)

223O. González et al. / Tectonophysics 717 (2017) 214–225

before establishing a reliable magnitude scale for Lesser Antillesearthquakes.

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.tecto.2017.06.024.

Acknowledgements

We are grateful to the team of the Volcanological and SeismologicalObservatory ofMartinique (OVSM/IPGP) for all thework facilities and toIRIS and OVS-IPGP for the seismic records provided. We acknowledge

the constructive discussions with the IPGP Seismology team. We thankMario Ruiz for sharing its Lesser Antilles focal mechanism database andChristineDeplus andMartin Patriat for sharing regional bathymetric da-ta. Maps and Fig. 9 were done using the GMT software (Wessel et al.,2013). Comments from Editor-in-Chief Rob Govers helped to improvethe manuscript and we also wish to thank two anonymous reviewersfor their constructive remarks. This research was supported by theOVSM/IPGP, the European Interreg Caraïbe IV (31420) TSUAREG pro-ject, the CENAIS/CITMA (P211SCU-003) and the Czech Science Founda-tion grant GACR-14-04372S. This is an IPGP contribution #3849.

2.5

3.0

3.5

4.0

4.5

Mag

nitu

de (

Mw

)

2.5 3.0 3.5 4.0 4.5

Magnitude (MLv)

Mw = 0.52MLv + 1.53

3.5

4.0

4.5

5.0

5.5

6.0

6.5

Mag

nitu

de (

Mw

)

3.5 4.0 4.5 5.0 5.5 6.0 6.5

Magnitude (MLv)

Mw = 0.94MLv + 0.15

a) b)

Fig. 9. Scaling of the moment magnitude Mw (from this study) with the local magnitude MLv: a) for 4 ≤ M ≤ 6 and b) 2.9 ≤ M ≤ 4.

224 O. González et al. / Tectonophysics 717 (2017) 214–225

References

Amante, C., Eakins, B.W., 2009. ETOPO1 1 arc-minute global reliefmodel: procedures, datasources and analysis. NOAA Technical MemorandumNESDIS NGDC-24. National Geo-physical Data Center, NOAA http://dx.doi.org/10.7289/V5C8276M.

Anglade, A., Lemarchand, A., Saurel, J.-M., Clouard, V., Bouin, M.-P., De Chabalier, J.-B., Tait,S., Brunet, C., Nercessian, A., Beauducel, F., Robertson, R., Lynch, L., Higgins, M.,Latchman, J., 2015. Significant technical advances in broadband seismic stations inthe Lesser Antilles. Adv. Geosci. 40:43–50. http://dx.doi.org/10.5194/adgeo-40-43-2015.

Bazin, S., Feuillet, N., Duclos, C., Crawford, W., Nercessian, A., Bengoubou-Valérius, M.,Beauducel, F., Singh, S.C., 2010. The 2004–2005 Les Saintes (FrenchWest Indies) seis-mic aftershock sequence observed with ocean bottom seismometers. Tectonophysics489, 91–103.

Bengoubou-Valérius, M., Bazin, S., Bertil, D., Beauducel, F., Bosson, A., 2008. CDSA: anewseismological data center for the French Lesser Antilles. Seismol. Res. Lett. 79 (1),90–102.

Bernard, P., Lambert, J., 1988. Subduction and seismic hazard in the northern Lesser Antil-les: revision of the historical seismicity. Bull. Seismol. Soc. Am. 78 (6), 1965–1983.

Bindi, D., Spallarossa, D., Eva, C., Cattaneo, M., 2005. Local and duration magnitudes inNorthwestern Italy, and seismic moment versus magnitude relationships. Bull.Seismol. Soc. Am. 95, 592–604.

Bouysse, P., Westercamp, D., 1990. Subduction of Atlantic aseismic ridges and Late Ceno-zoic evolution of the Lesser Antilles island arc. Tectonophysics 175, 349–380.

Brown, K.M., Westbrook, G.K., 1987. The tectonic fabric of the Barbados Ridge accretion-ary complex. Mar. Pet. Geol. 9, 71–81.

Buforn, E., 1994. Methods for estimation of focal mechanism of earthquakes. Física de laTierra, Núm. 6. Editorial Computence, Madrid, pp. 113–139.

Byrne, D.E., Davis, D.M., Syke, L.R., 1988. Loci and maximum size of thrust earthquakesand the mechanics of the shallow region of subduction zones. Tectonics 7, 833–857.

Byrne, D.E., Wang, W., Davis, D.M., 1993. Mechanical role of backstops in the growth offorearcs. Tectonics 12, 123–144.

Christeson, G.L., Bangs, N.L., Shipley, T.H., 2003. Deep structure of an island arc backstop,Lesser Antilles subduction zone. J. Geophys. Res. 108 (B7):2327. http://dx.doi.org/10.1029/2002JB002243.

Clouard, V., Roger, J., Moizan, E., 2017. Tsunami deposits inMartinique related to the 1755Lisbon earthquake. Natural Hazards and Earth System Sciences, accepted in July 2017.

DeMets, C., Jansma, P.E., Mattioli, G.S., Dixon, T.H., Farina, F., Bilham, R., Calais, E., Mann, P.,2000. GPS geodetic constraints on Caribbean-North America plate motion. Geophys.Res. Lett. 27, 437–440.

Deplus, C., 1998. AGUADOMAR Cruise, RV L'Atalante. (http://dx.doi.org/10.17600/98010120).

Doglioni, C., Panza, G., 2015. Polarized plate tectonics. In R. Dmowska (Ed.). Adv. Geophys.56:1–167. http://dx.doi.org/10.1016/bs.agph.2014.12.001.

Doglioni, C., Carminati, E., Cuffaro, M., Scrocca, D., 2007. Subduction kinematics and dy-namic constraints. Earth Sci. Rev. 83 (3–4):125–175. http://dx.doi.org/10.1016/j.earscirev.2007.04.001.

Dominguez, S., Lallemand, S., Malavieille, J., von Huene, R., 1998. Upper plate deformationassociated with seamount subduction. Tectonophysics 293, 207–224.

Dorel, J., 1978. Sismicité et structure de l'arc des Petites Antilles et du bassin Atlantique.Université Pierre et Marie Curie, p. 326 (Ph.D. thesis).

Dorel, J., 1981. Seismicity and seismic gap in the Lesser Antilles arc and earthquake hazardin Guadalupe. Geophys. J. R. Astron. Soc. 67, 679–695.

Dreger, D.S., 2002. Time-Domain Moment Tensor INVerse Code (TDMT_INVC) Version1.1. Berkeley Seismological Laboratory (18p).

Dziewonski, A.M., Chou, T.-A., Woodhouse, J.H., 1981. Determination of earthquake sourceparameters from waveform data for studies of global and regional seismicity.J. Geophys. Res. 86:2825–2852. http://dx.doi.org/10.1029/JB086iB04p02825.

Edwards, B., Allmann, B., Fah, D., Clinton, J., 2010. Automatic computation of momentmagnitudes for small earthquakes and the scaling of local to moment magnitude.Geophys. J. Int. 183, 407–420.

Ekström, G., Nettles, M., Dziewonski, A.M., 2012. The global CMT project 2004–2010: cen-troid-moment tensors for 13,017 earthquakes. Phys. Earth Planet. Inter. 200–201, 1–92012. (doi:10.1016/j.pepi.2012.04.002).

Evain, M., Galve, A., Charvis, P., Laigle, M., Kopp, H., Bécel, A., Weinzierl, W., Hirn, A., Flueh,E.R., Gallart, J., the Lesser Antilles Thales scientific party, 2013. Structure of the LesserAntilles subduction forearc and backstop from 3D seismic refraction tomography.Tectonophysics 603, 55–67.

Feuillet, N., Manighetti, I., Tapponnier, P., Jacques, E., 2002. Arc parallel extension and lo-calization of volcanic complexes in Guadeloupe, Lesser Antilles. J. Geophys. Res. 107(B12):2331. http://dx.doi.org/10.1029/2001JB000308.

Feuillet, N., Beauducel, F., Tapponnier, P., 2011. Tectonic context of moderate to large his-torical earthquakes in the Lesser Antilles and mechanical coupling with volcanoes.J. Geophys. Res. 116, B10308. http://dx.doi.org/10.1029/2011JB008443.

Gagnepain-Beyneix, J., Lépine, J.C., Nercessian, A., Hirn, A., 1995. Experimental study ofsite effects in the Fort-the-France area (Martinique Island). Bull. Seismol. Soc. Am.85 (2), 478–495.

Goldstein, P., 1999. SAC User's Manual. University of California, Lawrence LivermoreLaboratory.

Graves, R.W., Pitarka, A., 2004. Broadband time history simulation using a hybridapproach. Proc. 13th World Conf. Earthquake Eng., Vancouver, Canada, PaperNo. 1098.

Havskov, J., Ottemöller, L., 2010. Routine Data Processing in Earthquake Seismology.Springer Science+Business Media B.V. http://dx.doi.org/10.1007/978-90-481-8697-6_5.

Havskov, J., Bormann, P., Schweitzer, J., 2002. Earthquake location. In: Bormann, P. (Ed.),New Manual of Seismological Observatory Practice. GeoForschungsZentrum,Potsdam.

Hough, S.E., 2013. Missing great earthquakes. J. Geophys. Res. 118:1098–1108. http://dx.doi.org/10.1002/jgrb.50083 (2013).

Kagan, Y.Y., 1991. 3-D rotation of double-couple earthquakes sources. Geophys. J. Int. 106:709–716. http://dx.doi.org/10.1111/j.1365-246X.1991.tb06343.x.

Kanamori, H., 1983. Magnitude scale and quantification of earthquakes. Tectonophysics93, 185–199.

Kopp, H., Weinzierl, W., Bécel, A., Charvis, P., Evain, M., Flueh, E.R., Gailler, A., Galvé, A.,Hirn, A., Kandilarov, A., Klaeschen, D., Laigle, M., Papenberg, C., Planert, L., Roux, E.,Trail and Thales teams, 2011. Deep structure of the central Lesser Antilles islandarc: relevance for the formation of continental crust. Earth Planet. Sci. Lett. 304,121–134.

Krizova, D., Zahradnik, J., Kiratzi, A., 2013. Resolvability of isotropic component in regionalseismic moment tensor inversion. Bull. Seismol. Soc. Am. 103, 2460–2473.

Laigle, M., Hirn, A., Sapin, M., Bécel, A., Charvis, P., Flueh, E., Diaz, J., Lebrun, J.-F., Gesret, A.,Raffaele, R., Galve, A., Evain, M., Ruiz, M., Kopp, H., Bayrakci, H., Weinzierl, W., Hello,Y., Lépine, J.-C., Viodé, J.-P., Sachpazi, M., Gallart, J., Kissling, E., Nicolich, R., 2013. Seis-mic structure and activity of the central Lesser Antilles subduction from an integratedapproach: similarities with the Tohoku forearc. Tectonophysics 603:1–20. http://dx.doi.org/10.1016/j.tecto.2013.05.043.

López, A.M., Stein, S., Dixon, T., Sella, G., Calais, E., Jansma, P., Weber, J., LaFemina, P., 2006.Is there a northern Lesser Antilles forearc block? Geophys. Res. Lett. 33, L07313.http://dx.doi.org/10.1029/2005GL025293.

225O. González et al. / Tectonophysics 717 (2017) 214–225

Massin, F., Beauducel, F., Clouard, V., Vorobieva, I., Bertil, D., Bengoubou-Valérius, M.,Bouin, M.-P., Amorese, D., Dessert, C., the Centre de Données Sismologiques desAntilles project team, 2017. Automatic Picking and Probabilistic Location for Earth-quake Assessment in the Lesser Antilles Subduction Zone (1972–2013) (in prep).

Meschede, M., Frisch, W., 1998. A plate-tectonic model for the Mesozoic and Early Ceno-zoic history of the Caribbean Plate. Tectonophysics 296, 269-29.

Nafe, J.E., Drake, C.L., 1960. In: Hill, M.N. (Ed.), Physical Properties of Marine Sediments.The Sea vol.3. Wiley-Interscience, NewYork, pp. 794–815.

Ottemöller, L., Havskov, J., 2003. Moment magnitude determination for local and regionalearthquakes based on source spectra. Bull. Seismol. Soc. Am. 93, 203–214.

Patriat, M., Pichot, T., Westbrook, G.K., Umber, M., Deville, E., Bénard, F., Roest, W.R.,Loubrieu, B., the ANTIPLAC Cruise Party, 2011. Evidence for quaternary convergenceacross the North America–South America plate boundary zone, east of the Lesser An-tilles. Geology 39:979–982. http://dx.doi.org/10.1130/G32474.1.

Pichot, T., Patriat, M., Westbrook, G.K., Nalpas, T., Gutscher, M.A., Roest, W.R., Deville, E.,Moulin, M., Aslanian, D., Rabineau, M., 2012. The Cenozoic tectonostratigraphic evo-lution of the Barracuda Ridge and Tiburon Rise, at the western end of the NorthAmerica-South America plate boundary zone. Mar. Geol. 303-306:154–171. http://dx.doi.org/10.1016/j.margeo.2012.02.001.

Pindell, J., Kennan, L., 2009. Tectonic evolution of the Gulf of Mexico, Caribbean and north-ern Sotuh America in the mantle reference frame: an update. The Geology and Evo-lution of the Region between North and South America, Geological Society ofLondon, Special Publication (In press).

Richter, C.F., 1935. An instrumental earthquake magnitude scale. Bull. Seismol. Soc. Am.25 (1–2), 1–32.

Ruiz, M., Galve, A., Monfret, T., Sapin, M., Charvis, P., Laigle, M., Evain, M., Hirn, A., Flueh, E.,Gallart, J., Diaz, J., Lebrun, J.F., The Lesser Antilles Thales scientific party, 2011. Seismicactivity offshoreMartinique and Dominica islands (Central Lesser Antilles subductionzone) from temporary onshore and offshore seismic networks. Tectonophysics 603:68–78. http://dx.doi.org/10.1016/j.tecto.2011.08.006.

Russo, R.M., Okal, E.A., Rowley, K.C., 1992. Historical seismicity of the southeastern Carib-bean and tectonic implications. Pure Appl. Geophys. 139, 87–120.

Scherbaum, F., Stoll, D., 1983. Source parameters and scaling laws of the 1978 SwabianJura (southwest Germany) aftershocks. Bull. Seismol. Soc. Am. 73, 1321–1343.

Scholz, C.H., 1998. Earthquakes and friction laws. Nature 391:37–42. http://dx.doi.org/10.1038/34097.

Sokos, E.N., Zahradník, J., 2008. ISOLA a Fortran code and a Matlab GUI to perform multi-ple point source inversion of seismic data. Comput. Geosci. 34, 967–977.

Sokos, E.N., Zahradník, J., 2013. Evaluating centroid-moment-tensor uncertainty in thenew version of ISOLA software. Seismol. Res. Lett. 84, 656–665.

Stein, S., Engeln, J.F., Wiens, D.A., Fujita, K., Speed, R.C., 1982. Subduction seismicity in theLesser Antilles Arc. J. Geophys. Res. 87, 8642–8664.

Symithe, S., Calais, E., de Chabalier, J.B., Robertson, R., Higgins, M., 2015. Current blockmo-tions and strain accumulation on active faults in the Caribbean. J. Geophys. Res.http://dx.doi.org/10.1002/2014JB011779.

Uchida, N., Kirby, S.H., Okada, T., Hino, R., Hasegawa, A., 2010. Supraslab earthquake clus-ters above the subduction plate boundary offshore Sanriku, northeastern Japan:seismogenesis in a graveyard of detached seamounts? J. Geophys. Res. 115, B09308.http://dx.doi.org/10.1029/2009JB006797.

Vackář, J., Burjánek, J., Zahradník, J., 2015. Automated detection of disturbances in seismicrecords; MouseTrap code. Seismol. Res. Lett. 86:442–450. http://dx.doi.org/10.1785/0220140168.

Wadge, G., Shepherd, J.B., 1984. Segmentation of the Lesser Antilles subduction zone,Earth planet. Sci. Lett. 71, 297–304.

Wells, D., Coppersmith, K., 1994. New empirical relationships among magnitude, rupturelength, rupture width, rupture area, and surface displacement. Bull. Seismol. Soc. Am.84, 974–1002.

Wessel, P., Smith, W.H.F., Scharroo, R., Luis, J.F., Wobbe, F., 2013. Generic mapping tools:improved version released. EOS Trans. AGU 94, 409–410.

Westbrook, G.K., 1975. Structure of crust and upper mantle in region of Barbados andLesser Antilles. Geophys. J. R. Astron. Soc. 43:201–242. http://dx.doi.org/10.1111/j.1365-246X.1975.tb00632.x.

Westbrook, G.K., Smith, M.J., Peacock, J.H., Poulter, M.J., 1982. Extensive underthrusting ofundeformed sediment beneath the accretionary complex of the Lesser Antilles sub-duction zone. Nature 300:625–628. http://dx.doi.org/10.1038/300625a0.

Zahradník, J., Custodio, S., 2012. Moment tensor resolvability: application to SouthwestIberia. Bull. Seismol. Soc. Am. 102 (3):1235–1254. http://dx.doi.org/10.1785/0120110216.

Zahradník, J., Plešinger, A., 2005. Long-period pulses in broadband records of near earth-quakes. Bull. Seismol. Soc. Am. 95 (5):1928–1939. http://dx.doi.org/10.1785/0120040210.

Zahradník, J., Plešinger, A., 2010. Toward understanding subtle instrumentation effects as-sociated with weak seismic events in the near field. Bull. Seismol. Soc. Am. 100 (1):59–73. http://dx.doi.org/10.1785/0120090087.

Zuniga, F.R., Wyss, M., Scherbaum, F., 1988. A moment magnitude relation for Hawaii.Bull. Seismol. Soc. Am. 78, 370–373.