Tectonophysics 633 (2014) 211–220

Contents lists available at ScienceDirect

Tectonophysics

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

Earthquake sequences in the southern block of the PernambucoLineament, NE Brazil: Stress field and seismotectonic implications

Heleno C. Lima Neto a, Joaquim M. Ferreira a,b,⁎, Francisco H.R. Bezerra a,c, Marcelo Assumpção d,Aderson F. do Nascimento a,b, Maria O.L. Sousa c, Eduardo A.S. Menezes b

a Programa de Pós-Graduação em Geodinâmica e Geofísica, Universidade Federal do Rio G. do Norte, Campus Universitário, Natal, RN 59078-970, Brazilb Departamento de Geofísica, Universidade Federal do Rio Grande do Norte, Campus Universitário, Natal, RN 59078-970, Brazilc Departamento de Geologia, Universidade Federal do Rio Grande do Norte, Campus Universitário, Natal, RN 59078-970, Brazild Instituto de Astronomia, Geofísica e ciências climáticas, Universidade de São Paulo, Cidade Universitária, São Paulo, SP 05508-900, Brazil

⁎ Corresponding author at: Departamento de GeofísicGrande do Norte, Campus Universitário, Natal, RN 59078-

http://dx.doi.org/10.1016/j.tecto.2014.07.0100040-1951/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 17 July 2013Received in revised form 9 July 2014Accepted 12 July 2014Available online 22 July 2014

Keywords:Earthquake seriesIntraplate South AmericaFocal mechanismStress fieldFault reactivation

The Pernambuco Lineament (PeL) is an E–W-striking, 700 km long shear belt located in the central part of theBorborema Province, Brazil. Seismic sequences have been monitored in the region since 1991, and they areclearly correlated with the main shear belt and its NE-trending branches. A new sequence of small earthquakesoccurred in April 2010 away from the main belt in the southern block of the lineament. We monitored thisseismicity using a ten station network for 154 days and used the observations to determine the nature of theseismicity and the stress patterns. In addition, we used data from five previous local seismograph networks todetermine the stress tensor. The results indicate that the seismicity is concentrated in three clusters, two ofwhich are associated with normal faults: the first strikes 96°, dips 51°, and has a rake of −65°, and the secondstrikes 253°, dips 64°, and has a rake of −120°. These faults are 2.2 km and 1.5 km long and 2.0–3.0 km and3.0–3.5 km deep, respectively, and are not correlated with the Pernambuco Lineament or its NE-trendingbranches. The inversion of focalmechanisms fromour investigation and previous studies indicates E–W-trendingcompression and N–S-trending extension that affects both the northern and southern blocks of the PeL. The newdata indicate that the fault behavior in the shear zone area is more complex than has previously been observedalong reactivated mylonitic belts.

a, Universidade Federal do Rio970, Brazil.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

Brazil is located in the eastern part of the South American plateand it has experienced earthquakes up to magnitude 6.2 and a max-imum intensity of VII MM (Barros et al., 2009; Berrocal et al., 1984;Ferreira et al., 1998). Among the main seismic areas of the countryis Borborema Province, in northeastern Brazil. Over the past fortyyears the seismic activity in this region has been characterized byevents with magnitudes of up to 5.2 mb and intensities up to VIIMMl (Assumpção, 1983, 1992; Berrocal et al., 1984; Ferreira et al.,1998), which generally occur as swarms of earthquakes that canlast more than 10 years, and sometimes include more than 1000events per day (Ferreira et al., 1998).

One of the most active areas in the Borborema Province is thePernambuco Lineament (PeL). The PeL is a ductile Brasiliano (PeL)(740–540 Ma) shear zone more than 700 km long (Fig. 1a) (Davisonet al., 1995). This shear zone and many others in the region were

reactivated in a brittle mode in the Cretaceous during the breakup ofSouth America and Africa (de Castro et al., 2008, 2012; Matos, 1992).Several earthquake series have been reported along this shear zoneand its NE-striking branches since the nineteenth century. In recentyears, seismic sequences have occurred in Caruaru (Ferreira et al.,1998, 2008), Belo Jardim (Lopes et al., 2010) and São Caetano counties(Lima Neto et al., 2013).

These studies showed that the seismic activity in each of the epicen-tral areas was related to the Pernambuco Lineament or its NE-trendingbranches. These seismic sequences have been reported as classic casesof the brittle reactivation of preexisting fabrics that behaved as zonesof weakness (Bezerra et al., 2011). However, in 2010, new sequencesof events occurred in the southern block of the PeL away from themain E–W-striking mylonitic belt and its NE-striking branches. Theseearthquake sequences cut across the preexisting ductile shear zonesand did not conform to the strict definition of swarms (Vidale andShearer, 2006) because local networks were not deployed when themain event occurred.

This study is the first to be performed with a local network of seis-mographic stations south of the Pernambuco Lineament. The first feltearthquake of the new sequences of events occurred on April, 18,

Fig. 1. (a) Location of the study area (red rectangle) and major geological features of the Borborema Province. The sedimentary basins in the figure: PaB— Parnaíba, PbB— Paraíba, PeB—

Pernambuco, and JaB— Jatobá (Almeida et al., 2000; Brito Neves et al., 2000). The red circles are the epicenters of earthquakes from Brazilian Seismic Bulletin catalog from 1720 to 2010;(b) distribution of focal mechanisms: A (Caruaru1991: Ferreira et al., 1998), B (Caruaru 2002: Ferreira et al., 2008), C (Belo Jardim2004: Lopes et al., 2010), D (Santa Luzia 2007: LimaNetoet al., 2013), E (São Caetano 2010: Lima Neto et al., 2013), F (this study), and G (this study).

212 H.C. Lima Neto et al. / Tectonophysics 633 (2014) 211–220

2010 and had amagnitude 2.8mR (Brazilian regional scale— equivalentto mb scale, Assumpção, 1983) and an intensity of VI MMI. This eventcaused panic in the town of Belém de Maria and in the village of Lajesde São José (Cupira County, Fig. 2). The 2.8mRmagnitude event that mo-tivated this study occurred along the Serra Verde seismogenic fault(Serra Verde cluster) and was the first event recorded in this area. Theepicentral determination of this event was done through analysis ofthe record of a regional station that operated 40 km away from this epi-central area. During the period of operation of a local network, otherevents were recorded in the Serra Verde cluster and in two otherseparate clusters — Barra do Chata and Lagoa dos Gatos, withmicrotremors ≤ 1.5 mR (Fig. 2).

This study presents the results of the campaign carried out fromApril, 21, 2010 to August, 10, 2010. The study confirmed the existenceof three seismic areas; the main area was located in Cupira County,and the two smaller seismic areas were located in the Agrestina andLagoa dos Gatos counties.

The objective of this study is to investigate the new seismic areas tothe south of the PeL, the stress patterns, and the causes of the seismicity.We show detailed hypocentral locations, focal mechanisms, and theregional stress pattern. The results indicate that the new seismicsequence cut across a preexisting ductile fabric, whereas the seismogenicfaults identified in previous studies reactivated these preexisting struc-tures.We also show the complexity associatedwith intraplate seismicityand the importance of the stress field determination.

2. Seismo-tectonic setting

The Borborema Province, which is located in the northeasternpart of the South American continent, is an intraplate area 900 km

long and 600 km wide. The province is limited to the west by theParnaíba Basin, to the east by coastal basins, and to the south bythe São Francisco Craton (Fig. 1a). The BP encompasses Archean,Paleoproterozoic, and Neoproterozoic terranes that were formedor reworked during the Brasiliano orogenic cycle at 750–540 Ma(Almeida et al., 2000; Brito Neves et al., 2000). The BorboremaProvince was deformed by a system of ductile shear zones thatform the boundaries between most of these terranes. The mainstructural features of the Borborema Province in the south are thePatos and Pernambuco (PeL) lineaments (Fig. 1a), which strike ap-proximate EW.

The Cretaceous sedimentary basins along the continental marginandwithin the BorboremaProvincewere formed by the brittle reactiva-tion of the major ductile lineaments during the breakup of Pangea intoSouth America and Africa (de Castro et al., 2008, 2012; Matos, 1992).The PeL forms thenorthern boundary of the Jatobá Basin and thebound-ary between the Paraíba and Pernambuco basins (Fig. 1a) (Lima Filhoet al., 2006; Matos, 1992).

The PeL and its branches are among the most seismic areas of theBorborema Province (Bezerra et al., 2011; Ferreira et al., 1998, 2008;Lima Neto et al., 2013; Lopes et al., 2010), whereas the Patos Linea-ment is nearly aseismic. The seismicity associated with the PeL andits vicinity has been recognized since the early nineteenth century(Ferreira and Assumpção, 1983), and the most active area is locatedbetween 100 and 180 km west of Recife (Fig. 1). In 1967,a sequenceof earthquakes, including a main event of magnitude 3.8 mR and in-tensity V MM, caused panic in the city Caruaru (Ferreira andAssumpção, 1983). A magnitude 4.0 mR event occurred in SãoCaetano in 2006 (Lima Neto et al., 2013) and reached an intensityof VI MM.

Fig. 2. The events of the Serra Verde, Barra do Chata, and Lagoa dos Gatos clusters (a). The Serra Verde (b) and Barra do Chata events (c) used in this study have a gap b180° and include, ten observations (P and S readings). The black star indicates theepicenter of 2.8mRmagnitude event thatmotivated this study. Themaximum errors for both datasets are rms≤ 0.03 s, erz≤ 0.3 km, and erh≤ 0.3 km. The cluster located south of Lagoa dos Gatos (d) was not used because the hypocenter errors arelarger due to the distribution of the stations, but is shown to visualize the earthquakes in the region.

213H.C.Lim

aNeto

etal./Tectonophysics633

(2014)211

–220

Fig. 3. Temporal distribution of the seismicity in the study area from 2010/04/21 to 2010/08/10 including the Serra Verde (a), Barra do Chata (b), and Lagoa dos Gatos (c) clusters.The vertical bars indicate the number of events recorded by the local network stations foreach cluster and the stars indicate the highest magnitudes of the events recorded duringthe operation of a local network.

214 H.C. Lima Neto et al. / Tectonophysics 633 (2014) 211–220

3. Methods

3.1. Data acquisition

The data acquisition was initiated on April, 21, 2010, when wedeployed a network of six seismographic stations. Four additional sta-tions were installed two weeks later. This local seismographic networkoperated until September 22, 2010 and was composed of three-component, short-period L4C3 seismometers. All of the stations usedReftek130 recorders with a sample rate of 250 samples per second.The epicenters recorded in the study area, the station locations, andthe locations of the Belém de Maria, Cupira, Agrestina, and Lagoa dosGatos counties are shown in Fig. 2. The daily activity of each cluster dur-ing the deployment is shown in Fig. 3.

3.2. Hypocentral determination

We used procedures that had already been tested by previousstudies in the region. We used the software SAC (Seismic AnalysisCode; Tapley and Tull, 1991) to determine the P and S arrival timesand thepolarity. As in other previous studies in theBorboremaProvince,we assumed a half-space model because the stations were deployed di-rectly over granitic–gneissic outcrops (Ferreira et al., 1987, 1995, 1998,2008; Takeya et al., 1989). In addition, the hypocenters were deter-mined using HYPO71 (Lee and Lahr, 1975).

The determination of the velocity model for the region used theWadati diagram (Kissslinger and Engdahl, 1973) to verify the consisten-cy of the readings and theVP/VS ratio for 31 events (a total of 151 P and Sreadings) (Fig. 4). These events were recorded well by the local net-work. The VP/VS ratio was 1.71. We performed a search betweenVP = 5.0 km/s and VP = 6.4 km/s to obtain the best velocity value,which was VP = 6.0 km/s. These values of VP/VS and VP had the lowestmean squared horizontal and vertical errors during the determinationof the hypocenters using HYPO71. We also applied the double-difference algorithm with the HYPOODD code (Waldhauser, 2001;Waldhauser and Ellsworth, 2000) to relocate the best events anddetermine the fault plane orientation. The epicenters determined dur-ing this campaign, and the Gravatá-Açu and Jurema shear zones areshown in Fig. 2.

3.3. Focal mechanism

Weused two software packages to determine the focalmechanisms.First, we used the FPFIT program (Reasenberg andOppenheimer, 1985),which performs a grid search to find the solution that minimizes aweighted sum of the differences in the polarities by considering boththe estimated variance of the data and the theoretical P-wave radiationamplitude. We used only high-quality P-wave polarities to obtain thesolution of the composite focalmechanisms. These focal mechanism so-lutions considered the hypocentral distribution (dip and strike) alongthe fault plane obtained with FPFIT. Second, we used the programFOCMEC (Snoke, 2003), which uses P- and S-waves, as well as theratio of the amplitudes between different phases (Kisslinger, 1980;Kisslinger et al., 1982). FOCMEC was used in areas with little data.

3.4. Stress tensor

We used the code gridfix (Michael, 1984, 1987) for the inversion ofthe focal mechanisms. This method assumes that the slip direction ofthe earthquake rupture is given by the ambient shear stress on thefault plane, which is usually considered a valid assumption.

The inversion was performed byminimizing the difference betweenthe observed slip direction and the calculated shear stress direction onthe fault plane. We used a grid search method to determine thedirections of the three principal stresses (S1, S2, and S3) and the shapefactor Φ = (S2 − S3) / (S1 − S3), where S1 and S3 are the greatest

and the least compressional stresses, respectively.Weused the code de-veloped by Michael (1987), which assumes that the slip directions ofthe earthquake ruptures are given by the ambient shear stress on thefault plane. The mismatch between the slip and shear directionson the fault plane can be due to either a non-uniform stress field or er-rors in the fault-plane solution. We used the nodal plane with thesmallest mismatch angle to determine the focal mechanismwith an un-known fault plane, at every step of the grid search. In addition, we useda grid search with steps of 2° in the stress directions and 0.1 in theshape factor.

Fig. 4. Compound Wadati diagram for the earthquakes determined using 31 events and151 data points (squares). The best fit (solid line) corresponds to the ratio VP/VS =1.710(±0.005).

215H.C. Lima Neto et al. / Tectonophysics 633 (2014) 211–220

We used the methodology described in previous studies (Ferreiraet al., 2008; Lopes et al., 2010) to determine the stress tensors.Our dataset includes seven focal mechanisms, which is the largest num-ber of focal mechanisms used to determine stress tensors in the area ofthe PeL to date. The data are classified as Quality A, which indicates thehigh quality of our results.

4. Results

4.1. Determination of hypocenters

The results of HYPO71 show that the seismic activity wasconcentrated in three clusters (Fig. 2): (1) the main cluster (SerraVerde— SV) is located in Cupira County, near station LJ; (2) the secondcluster (Barra do Chata — BC) is located in the Agrestina County,between stations CR and UB; and (3) the third cluster (Lagoa dosGatos— LG) is located south of the town of Lagoa dos Gatos near stationSC. It was not possible to obtain a good distribution of stations near theLG cluster due to heavy rains and a flood that destroyed roads in thearea. Therefore, the errors in the hypocentral determination of the LGcluster were higher than those in SV and BC clusters.

4.2. Focal mechanisms

We determined the focal mechanisms of two epicentral areas. Wecalculated the fault planes using the hypocenters selected for the SVand BC clusters before using the FPFIT program to determine the com-posite focal mechanism. We relocated the hypocenters selected for thefocal mechanism using the HypoDD code, which reduces the effects ofthe hypocenter errors due to the existent uncertainties in the velocitymodel used in Hypo71. Fig. 5 shows an example of waveforms recordedfrom the SV and BC clusters, and the relocated hypocenters and focalmechanisms for the SV and BC clusters are presented in Figs. 6(a) and7(a), respectively. We then selected fifteen events from the SV areaand eight events from the BC area to obtain the composite focal mecha-nisms. These events in both areas had arrival-time residuals(rms) ≤ 0.02 s, vertical errors (erz) ≤ 0.2 km, and horizontal errors(erh) ≤ 0.2 km. The largest azimuthal separation in degrees betweenstations (gap) was ≤180°, and the number of P and S readings (NO)was ≥08.

The free FPFIT solution for the SV dataset includes the followingparameters: a strike of 105° ± 14°, a dip of 60° ± 16°, and a rake

of −58° ± 13°. The BC dataset had the following results: a strikeof 230°±28°, a dip of 65°±15°, and a rake of−125°±17°. It is impor-tant to note that the FPFIT solutions are the best statistical results. How-ever, these solutions should be subjected to a critical analysis(Reasenberg and Oppenheimer, 1985). Strike and dip can be deter-mined by assuming that all of the earthquakes lie in the same plane.The projections of the hypocenters over vertical planes parallel and nor-mal to the strike direction are shown in Fig. 6 (SV cluster) and Fig. 7 (BCcluster). The strikes and dips of the best fitting planes to hypocenterswere s = 96° and d = 51° for the Serra Verde fault, and s = 253° andd = 64° for the Barra do Chata fault. The fault rake was determinedusing stations with alternating polarities in lower-hemisphere projec-tions (Figs. 6 and 7), which indicates proximity to one of the nodalplanes.

Figs. 6b and 7b show the focal mechanisms of the Serra Verde andBarra do Chata faults, respectively. The focal mechanism proposed forthe Serra Verde seismogenic fault strikes approximately EW, dips tothe south, and has normal slip. The focal mechanism proposed for theBarra do Chata seismogenic fault strikes approximate EW, dips to thenorth, and has normal slip. It was not possible to obtain a reliable focalmechanism for the LG cluster because of errors in the hypocenter deter-mination and the poor distribution of stations.

4.3. Regional stress

This study updated the stress tensor of the PeL area by adding newdata to existing dataset. Estimates of the orientations in the BorboremaProvince were presented earlier by Ferreira et al. (1998), Lopes et al.(2010), and Lima Neto et al. (2013). The new data in this study includethe new focal mechanisms of Serra Verde and Barra do Chata presentedin this study (Table 1, Figs. 6 and 7). Fault planes for all the focal mech-anisms were determined by the hypocentral distribution. We obtainedthe 3D inversion using the gridfix code (Michael, 1987) withoutrestricting the directions of the crustal stresses. The results indicatethat the main stresses are approximately horizontal and vertical(Fig. 8). These results are reliable because the focal mechanisms(Fig. 1b) were determined with local networks, and the fault planeswere previously defined.

Fig. 8 shows the best fit to the seven observed rakes and theazimuths/plunges of S1 and S3 and Table 2 shows their relatedmisfit pa-rameters. The focal mechanismswere fitted with an rms error of 2° andthe orientations of the maximum and minimum stresses in our cases,both of which were close to the horizontal. The stresses indicate astrike-slip regime with a nearly E–W deviatoric compression and anearly N–S deviatoric extension. The PeL area has a best-fitting shapefactor φ of 0.5, which implies that the E–W compression (S1–S2) is ap-proximately the same as the N–S extension (S2–S3).

To estimate the uncertainties in the stress tensor, we performed agrid search using a few tens of resampled of themechanism parameters(strike, dip, and rake) for the seven selected focalmechanisms. The sam-ples were randomly chosen from the original group of the focal mecha-nisms parameters (strike, dip, and rake) in 10° increments and also bychanging the nodal plane and rake of the mechanisms. The value of10° was chosen because the seven known arrangements have faultplanes; that is, the focal mechanisms are Quality A. The jackknife statis-tical method (Efron, 1982; Tukey, 1958), which is a resampling tech-nique especially useful for variance and bias estimation, was also usedto estimate the uncertainty. The jackknife estimator of a parameter isfound by systematically leaving out each observation from a dataset,calculating the estimate, and then finding the average of these calcula-tions. Given a sample of size N, the jackknife estimator is found by ag-gregating the estimates of each N − 1 estimate in the sample. Thiswas done using the seven focal mechanisms. In Table 2, the ‘range’ col-umn shows the maximum deviations in the S1 (and S3) orientations.The error in the form factor was obtained in a similar manner.

Fig. 5. Examples of P and S waveform of waves recorded by the local network. These data were used to determine the focal mechanisms of the Serra Verde and Barra do Chata clusters.

216 H.C. Lima Neto et al. / Tectonophysics 633 (2014) 211–220

5. Discussion

The results indicate that the two normal faults do not repre-sent the brittle reactivation of ductile shear zones. The seismic ac-tivity in the Cupira County indicated the existence of the SerraVerde seismogenic fault near the end of the Gravatá-Açú shearzone. Although the Serra Verde seismogenic fault strikes E–Wand has normal movement, it does not have the same strike asthe Gravatá-Açu shear zone and is not located along it (CPRMand DNPM, 2001). However, the Serra Verde fault coincides withan E–W-trending topographic negative anomaly (Fig. 6a). Like-wise, an E–W-trending negative topographic anomaly along a

nearby river coincides with the strike of the Barra do Chataseismogenic fault (Fig. 7a). We do not rule out the possibilitythat both seismogenic faults are the reactivation of preexistingCretaceous faults. Although it was not possible to obtain thefocal mechanism of the Lagoa dos Gatos cluster, it representsthe first seismicity described at this site; where there is no previ-ous earthquake record or mapped faults in the area (Fig. 2). Thesequences of Barra do Chata and Lagoa dos Gatos do not seem tobe directly associated with the April 18 event that occurred inthe Serra Verde Cluster. However, it is possible that the gradualrelease of stress in the region has triggered these other seismicareas.

Fig. 6. (a) Map view of 15 earthquakes (in yellow) selected to determine the focal mechanism in the Serra Verde region. Topographic lineaments are shown in red. (b) Projections in thevertical planes view perpendicular (AC) and parallel (AB) to the fault plane cluster in Serra Verde, (c) Composite focal mechanism for the Serra Verde area in the lower hemisphere, equalarea projection. Crosses and circles represent the first movements of compression and dilation, respectively. P and T are the axes of compression and expansion, respectively. FP indicatesthe fault plane. The maximum errors of the selected dataset are rms ≤ 0.02 s, erz ≤ 0.2 km, and erh ≤ 0.2 km.

217H.C. Lima Neto et al. / Tectonophysics 633 (2014) 211–220

The Pernambuco Lineament and its NE-striking branches involve theclear seismogenic reactivation of a major ductile shear zone (Bezerraet al., 2011). The Pernambuco Lineament and its NE-trending branchesto the north of the lineament have been considered to be reactivationof zones of weakness; earthquakes are clearly related to structuresobserved at the surface in epicentral areas such as Caruaru in 1991(Ferreira et al., 1998) and 2002 (Ferreira et al., 2008), Belo Jardim in2004 (Lopes et al., 2010), Santa Luzia in 2007 (Lima Neto et al., 2013),and São Caetano in 2010 (Lima Neto et al., 2013). However, the resultsof this study show that seismicity is not correlated with structuralfeatures in the southern block of the PeL, where both the Serra Verdeand Barra do Chata seismogenic faults cut across the preexisting ductilefabric (Fig. 2).

The shallow depths found in this study (2.0 to 3.5 km) have smalluncertainties and are reliable because each local net has at least one

station within about one focal depth of the epicenters. Shallow seis-micity is also reported in other cases in the nearby segments of thePernambuco Lineament: 4–6 km in Belo Jardim (Lopes et al., 2010),4–8 km near São Caetano (Lima Neto et al., 2013), and 3.5–5.0 inCaruaru (Ferreira et al., 2008). In the Borborema Province as awhole, seismicity occurs in the brittle upper crust between 1 and12 km (Assumpção, 1992; Bezerra et al., 2007, 2011; Ferreira et al.,1998, 2008). Seismicity occurring at depths as shallow as 1 km isnot uncommon in other parts of the Precambrian shield in Brazilsuch as in the São Francisco craton (Agurto et al., submitted;Chimpliganond et al., 2010) and in Neoproterozoic foldbelts in cen-tral Brazil (Assumpção and Sacek, 2013). The common occurrenceof intraplate seismicity in the topmost few km of the brittle crustmay be an indication of the relative importance of flexural stressesdue to crustal loads, as suggested by Ferreira et al. (1998) for

Fig. 7.Map view of eight earthquakes (in blue) selected to determine the focal mechanism in the Barra do Chata region. Topographic lineaments are shown in red. (b) Projections in thevertical planes perpendicular (AC) and parallel (AB) to the fault plane cluster in Serra Verde, (c) Composite focal mechanism for the Serra Verde area in the lower hemisphere, equal areaprojection. Crosses and circles represent the first movements of compression and dilation, respectively. P and T are the axes of compression and expansion, respectively. FP indicates thefault plane. The maximum errors of the selected dataset are rms ≤ 0.02 s, erz ≤ 0.2 km, and erh ≤ 0.2 km.

Table 1Focal mechanisms in the Pernambuco Lineament area.

Focal mechanism Locality Date Lat Long Depth Fault plane solution Azimuth Plunge Reference

(°) (°) (km) Strike Dip Rake P T

A Caruaru 1991 −8.28 −36.02 3.0–5.0 262 61 −81 194/72 346/16 Ferreira et al. (1998)B Caruaru 2002 −8.26 −35.95 3.0–5.0 232 70 −170 94/21 187/7 Ferreira et al. (2008)C Belo Jardim 2004 −8.32 −36.36 3.0–6.0 277 66 −80 206/68 360/20 Lopes et al. (2010)D Santa Luzia 2007 −8.26 −36.16 2.0–8.0 74 60 −145 289/45 19/0 Lima Neto et al. (2013)E São Caetano 2010 −8.30 −36.15 2.0–8.5 265 79 −91 174/56 356/34 Lima Neto et al. (2013)F Serra Verde 2010 −8.57 −35.91 2.0–3.0 96 51 −65 70/71 169/3 This studyG Barra do Chata 2010 −8.51 −35.93 3.0–3.5 253 64 −120 119/59 4/14 This study

218 H.C. Lima Neto et al. / Tectonophysics 633 (2014) 211–220

Fig. 8. Inversion of the stress tensor with seven fault planes (Table 1). The black arrowsindicate the direction of the greatest horizontal stress (S1 = SHmax), and the gray arrowsindicate the direction of the least horizontal stress (S3 = Shmin). The inversion wasperformed by minimizing the difference between the observed slip direction and theshear stress on each fault plane (this difference is shown by the thick segment on thefault plane).

219H.C. Lima Neto et al. / Tectonophysics 633 (2014) 211–220

Northeastern Brazil and Assumpção and Sacek (2013) for other partsof intraplate South America.

The results of inversion of seven focal mechanisms indicate nearlyE–Wcompression andnearlyN–S extension (Fig. 8). The principal stressdirections are consistent with the average estimates of Ferreira et al.(1998) and Lopes et al. (2010), and the results of this study provides ad-ditional details about the regional stress pattern. The SHmax direction ob-tained in this study agrees well with theoretical models of stresses inthe South American plate (e.g., Coblentz and Richardson, 1996).Additionally, all of the mechanisms in this area are normal or strike-slipwith a normal component,which suggests thatwe have a strong ex-tensional stress component. Assumpção (1992) and Ferreira et al.(1998) discussed the effect of the continental margin on the stressesin the Borborema Province. As proposed by Ferreira et al. (1998), sedi-ment loading at the continental shelf and continental rise coupledwith subsidence of the cooling oceanic lithosphere can generate largeextensional stresses normal to the coast in a peripheral bulge a fewhundred kilometers inland (e.g. Cloetingh et al., 1984; Stein et al.,1989). The Pernambuco and Paraíba basins are located at the easternend of the Pernambuco Lineament (Fig. 1), and extend offshore.The presence of the basin mayminimize the E–Wcompressional stress,which could trigger the seismicity in region. Future investigations witha larger number of focal mechanisms solutions in region will either

Table 2Results of stress inversion of focal mechanisms. The main stresses from the Pernambuco Lineamcertainties in the S1, S2 and S3 directions. φ = (S2 − S3) / (S1 − S3). N is the number of focastress in the fault plane.

S1 S2 S3

(Azimuth/plunge/range)

264.2/12.6/±1° 131.8/71/±11° 356.9/13.4/±2°

confirm or discard the stress field determined and its implicationsinferred in the present study.

Themajor ductile shear zones in the Borborema Province have threemain types of behaviors: (1) they are reactivated by the present-dayseismicity, such as the main fault and NE-striking branches of the PeL;(2) the area of the shear zone exhibits seismicity, but there is no corre-lation between the earthquakes andmapped structural features, such asin the southern part of the PeL presented in this study; and (3) the areaexhibits aseismic behavior, such as the Patos Lineament. Therefore, nosingle model can explain all of the intraplate seismic activity in generaland the intraplate seismicity related to these shear zones in particular.One example is the model in which intraplate seismicity is related tozones of weakness in the crust (Sykes, 1978). In this model, the correla-tion between seismogenic faults and preexisting ductile fabrics(Ferreira et al., 2008; Lopes et al., 2010) explains the behavior of thePernambuco Lineament and its NE-striking branches as zones of weak-ness. However, this model does not explain the aseismic behavior of thePatos Lineament, which has a similar geometry and orientation inrelation to the present-day stress field.

The relationship between seismogenic faults and preexisting fabricsis more complex than the behavior anticipated by Sykes (1978). In thePernambuco shear zone system, faults are frequently co-located withmajor shear zones and their branches (Ferreira et al., 2008; Lopeset al., 2010). Elsewhere in the region, however, faults do not use folia-tion planes as slip zones, but have a range of orientations to the local fo-liations (Bezerra et al., 2011; Ferreira et al., 1998). Studies of rift systemsin the area indicate that preexisting strained rocks control faultgeometry at the continental-scale. However, meso-scale brittle faults(10–100 s m) form structures that locally crosscut foliations. This mayexplain the range of seismogenic faults observed in the region, someof which are not related to the preexisting fabric (Kirkpatrick et al.,2013). As our study notes, this relationship between faults andpreexisting fabric is complex and requires further investigation.

6. Conclusions

Our study identified three new seismic clusters in the area of thePernambuco Lineament. This shear zone has been described as a uniquecase of the seismogenic reactivation of a ductile continental-scale struc-ture. The new seismogenic faults of Serra Verde (SV) and Barra do Chata(BC) seismogenic faults identified in this study are small and shallow.They represent active zones 2.2 km and 1.5 km long, respectively, andtheir depths range from 2.0 to 3.0 km for the SV and 3.0 to 3.5 km forthe BC. Both faults strike approximately E–W and have normal slip.Our 3D analysis of the stress tensor indicates that S1 and S3 are roughlyhorizontal and strike 264° and 356.9, respectively, which indicates thatthe faults are affected by approximately E–W-trending compressionand N–S-trending extension. These faults cut across the Precambrianductile shear zones and behave differently from previously mappedfaults that reactivated themain shear belt and its NE-trending branchesin the main and northern blocks of the PeL. No single model has beenproposed to explain the presence of seismogenic faults in intraplateareas. The seismicity of the Pernambuco shear zone is an indication ofthe complexity of intraplate seismicity.

ent area were constrained to be horizontal and vertical. ‘Range’ is an estimate of the un-l mechanisms used. Misfit angle is the difference between the observed slip and the shear

Range φ N Misfit angle

Ave. max.

±1° 0.5 ± 0.1 7 −1.2° 9.6°

220 H.C. Lima Neto et al. / Tectonophysics 633 (2014) 211–220

Acknowledgements

We thank Thomas Plenefischand and one anonymous reviewer fortheir numerous and constructive suggestions, which have greatly im-proved this work. This study was sponsored by the Brazilian ResearchCouncil (CNPq, Project Instituto Nacional de Ciência e Tecnologia emEstudos Tectônicos, INCT-ET, coordinated by R.A. Fuck) and the Pool ofgeophysical equipment (PegBr) coordinated by the NationalObservatory (ON, Brazil). The municipalities of Belém de Maria andCupira, and the Civil Defense (CODECIPE) of the State of Pernambucoprovided additional support for our study. AFN was sponsored by theInstituto Nacional em Ciência e Tecnologia em Geofísica do Petróleo(CNPq). JMF, FHRB,MA, andAFNhold CNPq PQgrants. HCLN also thanksCAPES for his PhD grant.

References

Agurto, H.,Assumpção, M.,Ciardelli, C.,Albuquerque, D.F.,Barros, L.V., França, G.S.L., 2014.The 2012–2013 Montes Claros earthquake series in the São Francisco Craton, Brazil:new evidence for non-uniform intraplate stresses in mid-plate South America.Geophys. J. Int. (submitted).

Almeida, F.F.M.,Brito Neves, B.B., Carneiro, C.D.R., 2000. The origin and evolution of theSouth American Platform. Earth Sci. Rev. 50, 77–111.

Assumpção, M., 1983. A regional magnitude scale for Brazil. Bull. Seismol. Soc. Am. 73,237–246.

Assumpção, M., 1992. The regional intraplate stress Field in South America. J. Geophys.Res. 97, 11889–11903.

Assumpção, M., Sacek, V., 2013. Intra-plate seismicity and flexural stresses in CentralBrazil. Geophys. Res. Lett. 40 (487–491), 2013. http://dx.doi.org/10.1002/grl.50142.

Barros, L.V., Assumpção, M., Quintero, R., Caixeta, D., 2009. The intraplate Porto dosGaúchos seismic zone in the Amazon Craton — Brazil. Tectonophysics. http://dx.doi.org/10.1016/j.tecto.2009.01.006.

Berrocal, J.,Assumpção, M.,Antezana, R.,Dias Neto, C.M.,Ortega, R., França, H.,Veloso, J.A.,1984. Sismicidade do Brasil. IAG-Universidade de São Paulo, CNEN.

Bezerra, F.H.R.,Takeya, M.K.,Sousa, M.O.L.,do Nascimento, A.F., 2007. Coseismic reactiva-tion of the Samambaia fault, Brazil. Tectonophysics 430, 27–39.

Bezerra, F.H.R.,do Nascimento, A.F.,Ferreira, J.M.,Nogueira, F.C.,Fuck, R.A.,Brito Neves, B.B.,Sousa, M.O.L., 2011. Review of active faults in the Borborema Province, intraplateSouth America — integration of seismological and paleoseismological data.Tectonophysics 510, 269–290.

Brito Neves, B.B., Santos, E.J.,Van Schmus, W.R., 2000. Tectonic history of the BorboremaProvince, northeastern Brazil. In: Cordani, U., Milani, E.J., ThomazFilho, A., Campos,D.A. (Eds.), Tectonic Evolution of South America: Proceedings of the 31stInternational Geological Congress, Rio de Janeiro, pp. 151–182.

Chimpliganond, C.,Assumpção, M., von Huelsen, M., França, G.S., 2010. The intracratonicCaraíbas-Itacarambi earthquake of December 09, 2007 (4.9 mb), Minas GeraisState, Brazil. Tectonophysics 480, 48–56.

Cloetingh, S.A.P.L.,Wortel, M.J.R.,Vlaar, N.J., 1984. Passive margin evolution, initiation ofsubduction and the Wilson cycle. Tectonophysics 109, 147–163.

Coblentz, D.D.,Richardson, R.M., 1996. Analysis of the South American intraplate stressfield. J. Geophys. Res. 100, 20245–20255.

CPRM (Brazilian Geological Survey), DNPM (National Department ofMineral Production),2001. Geological Map of Pernambuco State, Scale 1:100,000.

Davison, I.,McCarthy, M.,Powell, D.,Torres, H.H.F.,Santos, C.A., 1995. Laminar flow in shearzones: the Pernambuco shear zone, NE Brazil. J. Struct. Geol. 17, 149–161.

de Castro, D.L., Bezerra, F.H.R., Castelo Branco, R.M.G., 2008. Geophysical evidence ofcrustal-heterogeneity control of fault growth in the Neocomian Iguatu Basin, NEBrazil. J. S. Am. Earth Sci. 26, 271–285.

de Castro, D.L.,Bezerra, F.H.R.,Sousa, M.O.L.,Fuck, R.A., 2012. Influence of Neoproterozoictectonic fabric on the origin of the Potiguar Basin, northeastern Brazil and its linkswith West Africa based on gravity and magnetic data. J. Geodyn. 54, 29–42.

Efron, B., 1982. The Jackknife, the Bootstrap, and Other Resampling Plans. Society forIndustrial and Applied Mathematics, Philadelphia, Pa 9781611970319.

Ferreira, J.M.,Assumpção, M., 1983. Sismicidade do Nordeste do Brasil. Rev. Bras. Geofís. 1,67–88.

Ferreira, J.M.,Takeia, M.,Costa, J.M.,Moreira, J.A.,Assumpção, M.,Veloso, J.A.V.,Pearce, R.G.,1987. A continuing intraplate earthquake sequence near JoãoCâmara, NortheasternBrazil — preliminary results. Geophys. Res. Lett. 14, 1402–1405.

Ferreira, J.M.,Assumpção, M.,Costa, J.M.,Moreira, J.A.M.,Pearce, R.G., Takeya, M.K., 1995.Correlation of seismicity and water level in the Açureservoir — an example fromNortheast Brazil. Bull. Seismol. Soc. Am. 85, 1483–1489.

Ferreira, J.M.,Oliveira, R.T.,Takeya, M.K.,Assumpção, M., 1998. Superposition of local andregional stress in northeast Brazil: evidence from focal mechanism around thePotiguar marginal basin. Geophys. J. Int. 134, 341–355.

Ferreira, J.M.,Bezerra, F.H.R.,Sousa, M.O.L.,do Nascimento, A.F., Sá, J.M.,França, G.S., 2008.The role of Precambrian mylonitic belts and present-day stress field in the coseismicreactivation of the Pernambuco Lineament, Brazil. Tectonophysics 456, 111–126.

Kirkpatrick, J.D.,Bezerra, F.H.R.,Shipton, Z.K.,do Nascimento, A.F.,Pytharouli, S.I.,Lunn, R.J.,Soden, A.M., 2013. Scale-dependent influence of pre-existing basement shear zoneson rift faulting: a case study from NE Brazil. J. Geol. Soc. Lond. 170, 237–247.

Kisslinger, C., 1980. Evaluation of S to P amplitude ratios for determining focalmechanismfrom regional network observation. Bull. Seismol. Soc. Am. 70, 999–1014.

Kisslinger, C., Bowman, J.R.,Koch, K., 1982. Determination of focal mechanism from SV/Pamplitude ratios at small distances. Phys. Earth Planet. Inter. 30, 172–176.

Kissslinger, C.,Engdahl, E.R., 1973. The interpretation of the Wadatidiagram with relaxedassumptions. Bull. Seismol. Soc. Am. 63, 1723–1736.

Lee, W.H.K.,Lahr, J.C., 1975. HYPO71 (revised): a computer program for determining hy-pocenter, magnitude and first motion pattern of local earthquakes. U.S. GeologicalSurvey. Open File Rep. , pp. 75–311.

Lima Filho, M.F., Barbosa, J.A., Souza, E.M., 2006. Eventos tectônicos e sedimentaresnasBacias de Pernambuco e da Paraíba: implicações no quebramento do Gondwanae correlaçãocom a Bacia do Rio Muni. Geociências. , 25. UNESP, São Paulo, pp.117–126.

Lima Neto, H.C.L.,Ferreira, J.M.,Bezerra, F.H.R.,Assumpção, M.,do Nascimento, A.F.,Sousa,M.O.L.,Menezes, E.A.S., 2013. Upper crustal earthquake swarms in São Caetano: reac-tivation of the Pernambuco shear zone and trending branches in intraplate Brazil.Tectonophysics 608, 804–8011.

Lopes, A.E.V.,Assumpção, M.,do Nascimento, A.F., Ferreira, J.M.,Menezes, E.A.S.,Barbosa, J.R.,2010. Intraplate earthquake swarm in Belo Jardim, NE Brazil: reactivation of amajor Neoproterozoic shear zone (Pernambuco Lineament). Geophys. J. Int.180, 1303–1312.

Matos, R.M.D., 1992. The northeast Brazilian rift system. Tectonics 11, 766–791.Michael, A.J., 1984. Determination of stress from slip data: faults and folds. J. Geophys. Res.

89, 11517–11526.Michael, A.J., 1987. Use of focal mechanisms to determine stress: a control study. J.

Geophys. Res. 92, 357–368.Reasenberg, P.,Oppenheimer, D., 1985. FPFIT, FPPLOT and FPPAGE: Fortran computer pro-

grams for calculating and displaying earthquake fault-plane solutions. U.S. GeologicalSurvey. Open File Rep. 85–739 (109 pp.).

Snoke, J.A., 2003. FOCMEC: Focal Mechanism Determinations. A program for focal mech-anism determination by combined use of polarity and SV-P amplitude ratio data.http://www.geol.vt.edu/outreach/vtso/focmec/.

Stein, S.,Cloetingh, S.,Sleep, N.H.,Wortel, R., 1989. Passivemargin earthquake, stresses andrheology. In: Gregersen, S., Basham, P.W. (Eds.), Earthquakes at North-AtlanticPassive Margins: Neotectonics and Postglacial Rebound. Kluwer Academic, Boston,MA, pp. 231–259.

Sykes, L.R., 1978. Intraplate seismicity, reactivation of preexisting zones of weakness,alkaline magmatism, and other tectonism postdating continental fragmentation.Rev. Geophys. Space Phys. 16, 621–688.

Takeya, M.K.,Ferreira, J.M.,Pearce, R.P.,Assumpção, M.,Costa, J.M.,Sophia, C.M., 1989. The1986–1987 intraplate earthquakes sequence near JoãoCâmara, northeast Brazil —evolution of seismicity. Tectonophysics 167, 117–131.

Tapley,W.C.,Tull, J.E., 1991. SAC—Seismic Analysis Code, UserManual. Livermore NationalLaboratory (413 pp.).

Tukey, J.W., 1958. Bias and confidence in not-quite large samples (abstract). Ann. Math.Stat. 29 (2), 614.

Vidale, J.E., Shearer, P.M., 2006. A survey of 71 earthquake bursts across southernCalifornia: exploring the role of pore fluid pressure fluctuations and aseismic slip asdrivers. J. Geophys. Res. 111, B05312. http://dx.doi.org/10.1029/2005JB004034.

Waldhauser, F., 2001. HYPODD— a program to compute double-difference hypocenter lo-cations. U.S. Geological Survey Open-File Rep, pp. 01–113.

Waldhauser, F.,Ellsworth, W.L., 2000. A double-difference earthquake location algorithm:method and application to the Hayward fault. Bull. Seismol. Soc. Am. 90, 1353–1368.