lable at ScienceDirect

Journal of South American Earth Sciences 76 (2017) 11e24

Contents lists avai

Journal of South American Earth Sciences

journal homepage: www.elsevier .com/locate/ jsames

Crustal structure of north Peru from analysis of teleseismic receiverfunctions

Cristobal Condori a, b, George S. França a, *, Hernando J. Tavera b, Diogo F. Albuquerque a,Brandon T. Bishop c, Susan L. Beck c

a Instituto de Geociencias, Universidade de Brasília, Campus Universitario Darcy Ribeiro, Asa Norte, Brasília, 70910-900 Brazilb Instituto Geofísico del Perú IGP, Ciencias de la Tierra S�olida CTS, Calle Calatrava Mz. F Lt. 10 Urb., Camino Real Camacho-La Molina Lima, Peruc University of Arizona UA, Department of Geosciences, Gould-Simpson Bldg. 537, USA

a r t i c l e i n f o

Article history:Received 16 August 2016Received in revised form13 February 2017Accepted 15 February 2017Available online 22 February 2017

Keywords:Northern PeruReceiver functionCrustal thicknessVp/Vs ratio

* Corresponding author. Observat�orio Sismol�ogicUniversidade de Brasília, Campus Universitario Darcy70910-900 Brazil.

E-mail addresses: cristobal.condori@igp.gob.pe (C.(G.S. França), hernando.tavera@igp.gob.pe (H.J. Tave(D.F. Albuquerque), brandontbishop@email.arizona.email.arizona.edu (S.L. Beck).

http://dx.doi.org/10.1016/j.jsames.2017.02.0060895-9811/© 2017 Elsevier Ltd. All rights reserved.

a b s t r a c t

In this study, we present results from teleseismic receiver functions, in order to investigate the crustalthickness and Vp/Vs ratio beneath northern Peru. A total number of 981 receiver functions wereanalyzed, from data recorded by 28 broadband seismic stations from the Peruvian permanent seismicnetwork, the regional temporary SisNort network and one CTBTO station. The Moho depth and averagecrustal Vp/Vs ratio were determined at each station using the H-k stacking technique to identify thearrival times of primary P to S conversion and crustal reverberations (PpPms, PpSs þ PsPms). The resultsshow that the Moho depth correlates well with the surface topography and varies significantly fromwestto east, showing a shallow depth of around 25 km near the coast, a maximum depth of 55e60 kmbeneath the Andean Cordillera, and a depth of 35e40 km further to the east in the Amazonian Basin. Thebulk crustal Vp/Vs ratio ranges between 1.60 and 1.88 with the mean of 1.75. Higher values between 1.75and 1.88 are found beneath the Eastern and Western Cordilleras, consistent with a mafic composition inthe lower crust. In contrast values vary from 1.60 to 1.75 in the extreme flanks of the Eastern andWesternCordillera indicating a felsic composition. We find a positive relationship between crustal thickness, Vp/Vs ratio, the Bouguer anomaly, and topography. These results are consistent with previous studies inother parts of Peru (central and southern regions) and provide the first crustal thickness estimates for thehigh cordillera in northern Peru.

© 2017 Elsevier Ltd. All rights reserved.

1. Introduction

The complex structural and tectonic features of northern Peruare due to the rapid convergence (~60e70 mm/yr) between thesubducting oceanic Nazca plate and the continental South Amer-ican plate since the Mesozoic (DeMets et al., 1990; Kendrick et al.,2003; Norabuena et al., 1999; Villegas, 2009a). One of the mostdramatic consequences of this tectonic process is the presence ofthe Andes (Dewey and Bird, 1970), which are characterized by thesecond-largest plateau in the world, and complexities in crustal

o, Instituto de Geociencias,Ribeiro, Asa Norte, Brasília,

Condori), georgesand@unb.brra), diogofarrapo@gmail.comedu (B.T. Bishop), slbeck@

thickness, major deflections, variations in the angle of subduction(Gutscher et al., 1999), and along-strike variations in the width ofhigh topography (Dalmayrac et al., 1980).

The boundary between the Earth's crust and the mantle, calledthe Moho, comprises a major change in chemical, rheological, andseismic properties. The Moho has an important relationship tomany geological processes of great importance, such as earth-quakes, volcanism, and orogeny. Detailed information on the vari-ation of crustal properties and thickness are crucial forunderstanding the mechanism of crustal thickening, crustal evo-lution and the degree of isostatic compensation of a region (Yuanet al., 2002; Assumpç~ao et al., 2013).

Crustal thickness in Peru, especially in the southern and centralregions, has been the subject of many geophysical studies, usinggravity modeling (Fukao et al., 1989, 1999; Tassara and Echaurren,2012) and seismic data (James, 1971; Ocola et al., 1971; Tavera,1990; Phillips et al., 2012; Phillips and Clayton, 2014; Bishop

C. Condori et al. / Journal of South American Earth Sciences 76 (2017) 11e2412

et al., 2014; Ryan et al., 2016) to understand tectonic and dynamicprocesses, as well as the relationship between surface topography(isostasy and mountain building) and crustal thickness. Addition-ally crustal structure studies have been carried out along thePeruvian continental margin using wide-angle seismic techniques(Hussong et al., 1976; Kulm et al., 1981; Krabbenh€oft et al., 2004).

The local crustal structure beneath northern Peru is compara-tively poorly known. The first of crustal thickness estimates werecomputed using refraction and reflection data by Hussong et al.(1976), giving estimates of Moho depth from 8 to 12 km beneaththe continental margin and 30 km beneath the Pacific coast ofnorthern Peru (Fig. 1). Aranda and Assumpç~ao (2013) estimated acrustal thickness between 35 and 40 km under the AmazonianBasin at the Brazil-Peru border (Fig. 1). Additional estimates of

Fig. 1. Tectonic map of northern Peru, showing the location of seismic stations used in thisseismic studies (Hussong et al., 1976, circles; Aranda and Assumpç~ao, 2013, squares), blacklines represent the outline of the Andean Cordillera (600 m elevation contour). BrownDalmayrac et al., 1987; Tavera and Buforn, 1998). Abbreviations: MFZ: Menda~na Fracture ZAmazonian Basin; and HD: Huancabamba Deflection, represented by the thick red dashedpretation of the references to colour in this figure legend, the reader is referred to the web

crustal thickness on the continental-scale have been developed forSouth American, based on a data compilation from seismic refrac-tion experiments, receiver function analysis, and surface-wavedispersion (Chulick et al., 2013; Pav~ao et al., 2012; andAssumpç~ao et al., 2013, 2015).

Receiver functions are one of the most commonly used tech-niques for the study of the Earth's crustal structure beneath abroadband seismic station. Teleseismic P-waveforms contain thecombined effects of the earthquake source, the propagation me-dium and the instrument response. To determine Earth structurebeneath a broadband seismic station, the effects of local structure,can be isolated from others factors using signal deconvolution(Langston, 1979; Owens et al., 1984; Ammon et al., 1990; Ammon,1991). The resulting waveform of the isolated site response is

study and the main tectonic features and local crustal thickness result from previousdotted lines represent contours from the Slab1.0 model (Hayes et al., 2012). Dark graydashed lines indicate morphostructural units (modified from Audebaud et al., 1973;one; FA: Forearc; WC: Western Cordillera; EC: Eastern Cordillera; SA: Sub-Andes; AB:line. Line AA0 corresponds to the schematic cross-section show in Fig. 10. (For inter-version of this article.)

C. Condori et al. / Journal of South American Earth Sciences 76 (2017) 11e24 13

known as a receiver function (RF).To better understand the heterogeneous structure of northern

Peru, we have used teleseismic data recorded at broadband seismicstations and analyzed RFs from each station in order to estimate forfirst time the spatial variations in crustal thickness and the Vp/Vsratio beneath all of these stations.We usedmainly data collected bythe Peruvian National Seismic Network, and also data from aregional seismic network, and the recent compilation of informa-tion on crustal thickness by Assumpç~ao et al. (2013, 2015). Thecrustal thickness results from this study will contribute toimproving the regional crustal model for the Andean Chain.

2. Tectonic framework

Northern Peru forms part of the Central Andes (Gansser, 1973)where the tectonics (Andean Orogeny) have been driven by sub-duction processes since the early Mesozoic (M�egard, 1987). Thecomplex structural evolution of the region has given rise to theformation of different geomorphological structures causingcompressive and extensional stresses and resulting in uplift, crustalshortening and thickening, and formation of foreland basins (Solerand S�ebrier, 1990).

According to Martinod et al. (2010), the Andean orogenesis hasinvolved three major shortening episodes that correspond to pe-riods of rapid trench-perpendicular plate convergence along theAndean margin: 1) the Late Cretaceous Peruvian Orogeny, 2) thePalaeocene-Eocene Incaic Orogeny, and 3) the Neogene QuechuanOrogeny. The latest episode was apparently associated with ridgeimpingement and the onset of flat subduction between ~15 and 11Ma. (Hampel, 2002; Rosenbaum et al., 2005), which increased platecoupling and crustal shortening, reduced arc magmatism and vol-canic activity, and enhanced hydrothermal activity (Gutscher et al.,2000; Rosenbaum et al., 2005; Martinod et al., 2010).

Our study focuses on the segment bounded between latitude0�S and 11�S, where the Andean Cordillera reaches elevations of~4000 m and undergoes a major change in trend from NW to NNEat 6�S called the Huancabamba Deflection (HD) which changes thetopography of the Andes between central Peru and Ecuador,(Cobbing, 1981; M�egard, 1987; Mourier et al., 1988; Mitouard et al.,1990). Furthermore, in this region, the subducted Nazca plate has a“flat-slab” geometry (Gutscher et al., 2000; Ramos and Folguera,2009).

The Peruvian shallow or flat-slab subduction segment (3�S to15�S) is by far the most extensive region of flat-slab subduction inthe world (Gutscher et al., 2000), inwhich the Nazca Plate subductsat a normal dip (~40�) to a depth of around 100 km (Fig. 1), mostlikely lying directly beneath the overriding continental lithosphere,and then extends horizontally at this depth for several hundredkilometers before steepening again to the east (Cahill and Isacks,1992; Phillips and Clayton, 2014). The region is characterized bythe absence of arc volcanism and low heat flow (~40e80 mW/m2)(Henry and Pollack, 1988; Cardoso et al., 2010; Haraldsson, 2011), incontrast to the normally dipping regions to the south.

According to Audebaud et al. (1973), Megard (1978) andDalmayrac et al. (1980), our study region can be subdivided into thefollowing main morphostructural units which strike parallel to thePacific coast from west to east. These are (1) the Forearc, (2) theWestern Cordillera, (3) the Eastern Cordillera, (4) Sub-Andes, and(5) the Amazonian Basin.

◦ The Forearc (FA) is composed mainly of Precambrian andPaleozoic marine sedimentary and metamorphic rocks, andextends from the trench up to the western boundary of theAndean Cordillera following a NW-SE orientation, withmaximum elevations between 900 and 1200 m.

◦ The Western Cordilleras (WC) represents a morphological andstructural asymmetry about 150 km wide and reaches eleva-tions ranging from 3500 to 5000 m. It is dominated by theCoastal Batholith, which consist of multiple intrusions. Theolder units show the effects of compressive deformation thatpresumably occurred during the late Eocene to Miocene(Megard, 1978; Wipf, 2006).

◦ The Eastern Cordillera (EC) began to form during the Hercynian(Devonian) period and has a Precambrian basement. It reacheselevations of ~4000 m and width increases from ~100 km in thenorth to 150 km in the south. It is composed of crystalline andplutonic rocks overlain by Paleozoic shallow marine and conti-nental deposits (Dalmayrac et al., 1980). The main structures inthe region include open folds and steep thrust faults.

◦ The Sub-Andes (SA) belt is a region of eastward convergingcompressive structures that have decreasing magnitudes ofdeformation toward the Brazilian Shield (Dumont et al., 1991). Itcorresponds to a fold and thrust belt which has developed sinceMiocene times (Tankard et al., 1995) on a heterogeneous sub-stratum inherited from pre-Andean Paleozoic, Triassic andJurassic basins.

◦ The Amazonian Basin (AB) is comprised of Paleozoic andMesozoic marine sediments which are overlain by massivecontinental deposits of Tertiary age (Megard, 1978; Su�arez et al.,1983). The deposits have been faulted and folded most exten-sively at the Andean margin, but deformation decreases towardthe east where the sedimentary layer thins onto a foreland basinon the Brazilian Shield.

3. Data

We have used data from regional and international networksdistributed in northern Peru. The first data set was collected from15 permanent broadband seismic stations and recorded between01/2012 and 02/2015 by the National Seismological Network (NSN)(Fig. 1), operated by the Instituto Geofísico del Perú (IGP). Thesecond data set was obtained from the past deployment in NWPeru of a network (SisNort) of 12 temporary local broadbandseismic stations operating from 11/2008 to 12/2009. Finally, datafrom one auxiliary seismic station (AS077-ATH) from the CTBTO(Comprehensive Nuclear-Tes-Ban Treaty Organization) networkwere also examined for the period from 01/2012 to 02/2015. Theseismic stations were equipped with Trillium and Guralp CMG-40Tbroadband seismometers, with Nanometrics (Trident, Taurus)Digitalizers and Reftek 130s.

We obtained records of teleseismic direct P-phase arrivals fromeach seismometer for the calculation of teleseismic RFs. The in-formation for the teleseismic earthquakes was extracted from thePDE catalogue of US National Earthquake Information Center(NEIC). We selected events with magnitudes greater than 5.5 Mbthat occurred at epicentral distances ranging between 30� and 95�,which in general provided reasonably good coverage in terms ofback-azimuth, helping to reduce the influence of lateral crustalstructure variations on the analysis of our result (Fig. 2).

4. Methodology

The RF is a well-established seismological method, and has beenextensively used in recent decades to estimate the crustal thicknessbeneath three-component seismic stations, through the identifi-cation of major impedance contrasts using the P to S convertedphase obtained through the deconvolution of the vertical from theradial components of a seismogram in either the frequency domainor the time domain (Langston, 1979; Owens et al., 1984; Zandt and

Fig. 2. The map in azimuthal projection shows the distribution of teleseismic events with magnitude larger than 5.5 Mb used in this study. Red circles represent earthquakeepicenter locations used for the computation of receiver functions, large circles represent the distances of 30� and 95� from the center of the seismic network (blue triangle) insidethe red box, blue lines represent tectonic plate boundaries. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of thisarticle.)

C. Condori et al. / Journal of South American Earth Sciences 76 (2017) 11e2414

Ammon, 1995). The P to S converted phase from the Moho in RFwaveforms and their relative arrival times can be used to constrainthe crustal thickness and the Vp/Vs ratio beneath a seismic station(Zandt and Ammon, 1995; Zhu and Kanamori, 2000), and then thecrustal composition can sometimes be inferred from the averageVp/Vs (or Poisson's ratio).

Prior to RF calculation, we automatically identified the P arrivaltime for each event using the global velocity model (IASP91)included in the TauP software (http://www.seis.sc.edu/taup/)(Crotwell et al., 1999). Then we detrended and tapered the signal.We visually inspected each and manually selected those with highsignal-to-noise ratios. We rotated the two horizontal componentsto obtain radial and transverse components. Stations with orien-tation errors in the north component azimuth, were rotated usingthe analysis of the particle motions of teleseismic P waves in eachstation. To eliminate the effects of the ocean or storm waves, sur-face waves and high frequency noise we applied a band-pass filterto each waveform with corner frequencies of 0.05 and 5 Hz.

In order to remove the source and instrument response, RFswere obtained by deconvolving vertical components from radialcomponents using the time domain iterative deconvolution tech-nique (Ligorría and Ammon, 1999) with a maximum of 500 itera-tions. We tested a range of different Gaussian (“a”) width factors ateach station, the best resulting RFs were calculated using “a” valuesof 1.0, 1.5, and 2.5. Each “a” value is sensitive to a different fre-quency range, which makes the RFs sensitives to different features.

The greater the “a” factor value the more sensitive the resulting RFis to thin features (Frassetto et al., 2011; Poveda et al., 2015). Weused the lowest “a” values for places where a thick crust was ex-pected (beneath the Andean Cordillera), and the largest values forlocations where a thin crust was expected, especially for stationslocated at the flanks of the Andean Cordillera. Finally, the radial andtransverse RFs obtained through the deconvolution, were exam-ined in order to eliminated low quality RFs. Only RFs with variancereduction of greater than 70% were selected for use in subsequentanalysis.

We used the H-k stacking method of Zhu and Kanamori (2000),to estimate crustal thickness (H) and Vp/Vs (k) ratio (Poisson'sRatio) beneath all stations. We used an a priori assumption of6.1 km/s for the average crustal P-wave velocity, following the localcrustal velocity model proposed by Lindo (1993) and Villegas(2009b) for this study area.

RFs were stacked using the method of Zhu and Kanamori(2000), which uses the times of the converted phase and multi-ples to obtain estimates of the depth of an interface and average Vp/Vs ratio above the interface by summing the observed RF ampli-tudes at the expected times of the P-to-S conversions (Zhu andKanamori, 2000). A search is done over a range of depths and Vp/Vs ratios based using different many events.

The H-k stacking technique sums the amplitudes at the pre-dicted arrival times of Moho Ps and multiples (PpPs andPpSs þ PsPs) of the radial RF for different crustal thicknesses (H)

C. Condori et al. / Journal of South American Earth Sciences 76 (2017) 11e24 15

and Vp/Vs ratios (k). The optimumvalues are determined by findingthe maximum amplitudes of the H-k stack. Given the greaterrobustness of the P to S conversion at theMoho comparedwith thatof the multiples, we used weighting factors for the Ps, PpPs, andPpSs þ PsPs of 0.7, 0.2, and 0.1, respectively, as recommended byZhu and Kanamori (2000). However, there is a limitation to thistechnique and cases where it may fail, due to the interference ofcrustal reverberated phases (Cassidy and Ellis, 1993; Zhu andKanamori, 2000).

When using the H-k method, the best solution has an associatedsurrounding region enclosed by an ellipse which represents thearea where the stack is within one standard error of the maximumamplitude value (Eaton et al., 2006); we use the ellipse to define theuncertainty in H and Vp/Vs using the bootstrap method of Efronand Tibshirani (1991), which involved repeating the stacking pro-cedure 200 times, each time using a resampled data set selected atrandom from the original data set. A time window chosen tocalculate the RFs from�5 s before the P-wave to (35e65) s after theP-wave to retain the Moho conversion phases as well as reverber-ation phases. Erratic traces or outliers were visually removed, andafter this selection we obtained 981 RFs from 28 stations.

Fig. 3 shows the results of RFs for FIC, AS077-ATH and IQT sta-tions that are located in three different geomorphological units,(the FA, WC, and the AB, respectively). We stacked the individualtraces after applyingmove-out correction for the Ps-phases at threestations (the traces at the top of each figure). The Ps-phases at theMoho are correlated with the individual receiver functions. The RFsare arranged by epicentral distance, stations FIC and IQT show clearMoho converted Ps phases with positive peaks approximately at 3.5and 3.8 s, while the station AS077-ATH shows the Ps phase at 7.5 s.The multiple phases (PpPs and PpSs þ PsPs) do not show up clearlyin the single station RFs and this may be caused by the interferenceof crustal reverberations associated with the complex mid crustalfeatures in the study area. The use for a normal move-out

Fig. 3. Plot of individual radial RFs with different Gaussian filters (a) and time window as atectonic regimes. a) Station FIC of the SisNort Network located close to the coastal Peru-EcuadIQT station is located in the Amazonian Basin (a ¼ 1.5). The top panel in each figure showcorresponding to the Ps phase from the Moho and calculated arrival times from H-k stacki

correction to align the Ps phases at each station can also misalignsthe multiple phases, further obscuring these phases in the stackedtraces.

To identify potential, approximates times of multiple phase ar-rivals, we calculated theoretical travel-time curves of phases Ps andmultiples, using the optimal H and k values. For station FIC, the Ps,phase can be identified at ~3.5 s, while potential PpPs andPpSs þ PsPs phases may correspond to the arrivals near ~12.1 s and~15.6 s after the first P arrival (Fig. 3a). For station AS077-ATH, weidentify the Ps, arrival at approximately 7.5 s, and potential PpPsand PpSs þ PsPs phases at ~21.3 s and ~28.8 s (Fig. 3b). Finally forthe station IQT the large amplitude Ps, arrival can be identified at3.8 s, and potential PpPs and PpSs þ PsPs phases arrivals near~15.1 s and ~18.9 s respectively (Fig. 3c). The optimal estimate of Hand k values (Fig. 4) are H ¼ 30.6 ± 2.7 km, k ¼ 1.75 ± 0.07 for FIC,H ¼ 54.2 ± 3.2 km, k ¼ 1.88 ± 0.06 for AS077-ATH, andH ¼ 36.0 ± 2.7 km, k ¼ 1.74 ± 0.06 for IQT. These values show anobvious change in crustal thickness values consistent with thevariations in topography.

In order to enhance the Moho converted Ps-phase, we stackedall individual RFs for each seismic station (28 stations), afterapplying the Ps move-out correction. The traces have been groupedfor each geomorphological units and sorted with consideration ofthe topography variation in the region. Ps conversions from theMoho beneath northern Peru are typically visible and wellobserved on the stacked RFs (Fig. 5), due to the large contrast invelocity between the crust and the mantle, and show the arrivaltimes of the Ps conversions that reflect the geometry of crustal andmantle interfaces. In contrast the reverberation phases are harderto observe, mainly due to the complexity of crustal structure in theAndes and to misalignment of multiple phases resulting from thenormal move-out correction applied to better identify the Psarrival.

function of epicentral distance and back azimuth at three stations located in differentor border (a ¼ 2.5). b) AS077-ATH station located in the Western Cordillera (a ¼ 1.5). c)s the stack for all the traces after the move-out correction has been applied, pulses

ng for multiples (PpPs and PsPs þ PpSs) are indicated with arrows respectively.

Fig. 4. Results of H-k stacking for three different seismic stations showing the best crustal thickness and Vp/Vs ratios at each station. a) Station FIC of the SisNort Network locatedclose to the coastal Peru-Ecuador border. b) AS077-ATH station (CTBTO) located in the Andean Cordillera. c) IQT station (NSN) is located in the Amazonian Basin. The grid search wascarried out for an assumed average crustal P-wave velocity of 6.1 km/s. Colors are stacking surface with contours showing percentage values of the objective function and the redline gives the percent confidence ellipse obtained from bootstrapping the dataset. (For interpretation of the references to colour in this figure legend, the reader is referred to theweb version of this article.)

C. Condori et al. / Journal of South American Earth Sciences 76 (2017) 11e2416

5. Results

Crustal thickness and Vp/Vs ratio are two important factors incharacterizing the crustal structure and physical properties of thecrust (Zandt and Ammon, 1995). We obtained maps for crustalthickness (Figs. 6 and 7) and Vp/Vs ratio (Fig. 8) variationsmeasured at 28 seismic stations in the study area. The maps indi-cate significant lateral variations from west to east in the crustalstructure of northern Peru. Detailed results from each station arealso listed in Table 1.

5.1. Crustal thickness

Fig. 6 shows the independent values of crustal thickness at eachstation estimated in this study of northern Peru. In general, thevalues in the Andean Cordillera with the crustal roots progressivelythickening from NNW (55 km) to SSE (62 km), while on the flanksthe values range between 25 and 30 km in FA and 35 and 40 in theSA and AB. This result is in good agreement with topography, andthis crustal thickness tendency is consistent with the continental-scale crustal thickness models of South America derived fromdifferent data set.

A smoothed contourmap of crustal thickness in northern Peru isgiving in Fig. 7, showing the distribution of crustal thickness innorthern Peru and surrounding areas based on data from this study(Fig. 6). Adjacent areas outside of station coveragewere constrainedbased on the regional scale crustal thickness map from Assumpç~aoet al. (2013, 2015). We note that measured crustal thickness rangefrom 22 km to 30 km in the FA zone, which reflect RF results atstations FIC (30.6 km), CLB (28.2 km), PCM (29.5 km), CBT(22.5 km), BYV (27.9 km), and CHO (27.1 km). Thicker crust appearsbeneath the coast near to the Peru-Ecuador border, with an averagevalue of 35 km beneath stations CHL (33.8 km), LCN (36.9 km), andMTP (35.7 km).

Crustal thickness varies from 50 to 55 km beneath the WC andEC, showing a clear increase in depth to Moho from north to south,with maximum values appearing in central Peru beneath stations

HCO (63.3 km), YANA (62.7 km), and OXA (57.2 km). These valuesare consistent with previous result for central Peru proposed byBishop et al. (2014), who estimated a crustal thickness of up to62 km for the area from RFs. In the SA zone, between the EC and theAB (Fig. 6), the crustal thickness decreases to an approximate rangeof 37 kme44 km (stations MOY, TAR, BLV, and NIEV).

Overall, thinner crust is observed in the AB with values between35 km and 40 km (station PUC, YRM, and IQT). This is consistentwith results from Aranda and Assumpç~ao (2013) who estimated acrustal thickness of 35e40 km in this region.

The similar pattern in crustal thickness have also been found inprevious studies derived from different techniques such as CCPstacks, gravity and surface waves (Fukao et al., 1989; Aranda andAssumpç~ao, 2013; Assumpç~ao et al., 2013, 2015; Chulick et al.,2013; Bishop et al., 2014). Overall, our results agree well with thecontinental-scale models that were mentioned previously, whileimproving the resolution in many areas beneath the study area.

5.2. Vp/Vs ratio

Fig. 8 shows the crustal Vp/Vs ratios under each station. We donot present an interpolated map, given the great uncertainty in thelateral variation of this variable. Resulting Vp/Vs ratios show thebulk or average value and range between 1.62 and 1.88 with anaverage of 1.75 for the entire area. Vp/Vs values between 1.77 and1.88 are observed at stations CHA(1.77), YLS(1.79), BAG(1.88),AS077-ATH(1.88), YANA(1.80), OXA (1.83), HCO(1.85) beneath theWC and EC and are associated with regions of high crustal thick-ness. This may indicate the presence of relatively weak materialswith mafic compositions (Musacchio et al., 1997), small amounts ofpartial melt or anisotropy resulting from subduction processes. Thecomparatively low Vp/Vs values (<1.75) are mostly located in theFA, SA and AB at stations CHO(1.70), BYV(1.63) NIEV(1.70), andIQT(1.74) (Fig. 8).

Fig. 9 summarizes the crustal thickness and Vp/Vs results foreach station and shows a clear correlation between these twovariables. Uncertainties in crustal thickness estimates are less than

Fig. 5. Cross-section showing the stacking radial receiver functions for all stations (28). Move-out corrections have been applied corresponding to each station prior to stacking.Only data showing clear Ps multiples and the P direct wave are displayed, others multiples are not displayed. Phase arrivals are marked with black arrows.

C. Condori et al. / Journal of South American Earth Sciences 76 (2017) 11e24 17

3.8 km while the crustal thickness varies between 22 and 65 km.Vp/Vs values can have uncertainties as high as 0.07. Given the rangeof maximum and minimum uncertainties, the Vp/Vs ratio valuesvary between 1.65 and 1.85 for the crust in northern Peru.

6. Discussions

The distribution of the crustal thicknesses and Vp/Vs ratios inour study area reflect the characteristics of different geomorpho-logical units in northern Peru. Although previous models have usedless information for their studies, our observed values of crustalthickness in WC an EC, compared to FA and AB, are consistent withexpectations from regional gravity modeling (Fukao et al., 1989)and tomography studies (Lloyd et al., 2010), and are also compa-rable with those estimated by Aranda and Assumpç~ao (2013),

Chulick et al. (2013), Bishop et al. (2014), and Assumpç~ao et al.(2013, 2015). In general, all previous crustal models for northernPeru suggest that the greatest crustal thickness occurs in the EC andWC. The RFs that we present in this study confirm these observa-tions and inferences.

6.1. Crustal thickness and Vp/Vs ratio

The crustal thickness varies between 25 and 55 km in northernPeru and the surrounding area (Fig. 7). We have found a sharp in-crease in crustal thickness from north to south along the AndeanCordillera that matches variations in the elevation. We find averagecrustal thicknesses of 50 km near the Peru-Ecuador border in theHuancabamba Deflection region (6�S) where elevations reach atmost 1800 mwhile values to the south vary from 55 to 60 km near

Fig. 6. Crustal thickness values at each station analyzed in this study.

C. Condori et al. / Journal of South American Earth Sciences 76 (2017) 11e2418

central Peru (consistent with the results of Bishop et al. (2014))where elevation reach maximum values of 4300 m.

For the purposes of comparison, topography has been labeled onthe top of each profile in Fig. 10. The section A-A0 extends from theFA, crossing the Andean Cordillera and into the AB. Lateral varia-tions of the elevation (ETOPO1), the Bouguer gravity anomaly(Fukao et al., 1999; F€orste et al., 2014), crustal thickness (this study),average Vp/Vs ratios (this study) and heat flow (Cardoso et al.,2010; Haraldsson, 2011) are shown along a W-E profile (Fig. 10).The cross-section shows high Bouguer gravity anomalies thatsuggest the presence of high density material at depth, which isconsistent with the crustal thickness; on the other hand, it alsoshows a quite asymmetric gravity anomaly associated with the WCand the EC, in marked contrast with the symmetric pattern in thetopographic profile. According to Fukao et al. (1989) gravity andcrustal models indicate the WC is approximately in isostatic equi-librium, while the EC is not, and this contrast in mechanical stateand the difference in recent tectonics suggest that the WC and theEC have not been uplifted entirely through the same tectonic pro-cess. We see high values under the WC, that supports this hy-pothesis. Our model of crustal thickness inferred from RFs shows

values between 50 and 55 km under the WC (Fig. 7), while theestimate by Fukao et al. (1989) were between 45 and 50 km.

To better analyze the crustal isostatic compensation under theWestern and Eastern Cordilleras. We used a topography model(GTOPO30, Lindquist et al., 2004) to derive the predicted Airyisostatic crustal thickness (Molnar and England, 1990).

Along the cross-section A-A0 the Western and Eastern Cordil-leras merge with the average elevation for the WC being ~2.8 kmand the EC elevation being ~2.4 km as measured off spatiallyfiltered topography (with features larger than 50 km � 50 kmpassed) (Fig. 10a). Assuming isostatic compensation at the Mohoand using standard values for a crustal density of 2.8 g/cm3, amantle density of 3.3 g/cm3, and a reference crustal column of35 km at sea level we calculate crustal thickness values of 53.5 kmand 49.5 km for the Western and Eastern Cordilleras respectively.This is very similar to the observed values of crustal thickness alongthe A-A' cross-section of 54 km at station AS077-ATH in theWC and47 km at station CHA in the EC. These estimated crustal thicknessvalues based on crustal isostasy are within the error bars calculatedfrom the receiver function analysis in this study (Fig. 10c).

We conclude that along the A-A' cross-section the crust is

Fig. 7. Crustal thickness map in kilometers beneath northern Peru and adjacent areas. The values include topographic elevation. In addition to our results, we have also includeddata from previous studies compiled by Assumpç~ao et al. (2013, 2015). Lines AA0 correspond to schematic cross-section showing in Fig. 10. Contour interval is 5 km.

C. Condori et al. / Journal of South American Earth Sciences 76 (2017) 11e24 19

isostatically compensated at theMoho. The crustal thickness profileis very similar to the pattern predicted by Fukao et al. (1999)'sgravity measurements with the largest negative mGal value cor-responding to the highest elevation and thickest crust (Fig. 10b).Further to the south the elevation increases for both the Westernand Eastern Cordilleras as does the calculated crustal thicknesses(stations HCO, YANA and OXA) consistent with the assumption ofisostatic compensation at the Moho. To the north stations BAG andSIG have crustal thicknesses of ~54.5 km and ~53 km respectivelybut at a lower elevation, hence this region is likely not locallyisostatically compensated.

The different geological and tectonic characteristic between WCand EC have been reveled by different studies. According to Konoet al. (1989), two dominant process occur in the WC and the EC.The first, in the WC, the dominant process is the emplacement ofmagma into the crust. In the EC, the dominant process is crustalshortening due to compression from the Brazilian shield. RecentlyVillegas-Lanza et al. (2016) found evidence for different dynamicpatterns and crustal deformation in the WC and EC using GPS data,revealing that the rigid motion of the Peruvian FA extends from the

oceanic trench axis to the WC and EC boundary. This region movessoutheastward at 4e5 mm/yr relative to a stable South Americareference frame. GPS data also indicate that the SA shortening in-creases southward by 2e4 mm/yr.

The Vp/Vs ratio (or Poisson's ratio) offers critical constraints ondetermining crustal thickness and is sensitive to rock composition,generally increasing with higher mafic content (Zandt and Ammon,1995). Laboratory experiments have shown that many physical andchemical factors can induce variations of the average crustal Vp/Vsratio (Christensen, 1996). Vp/Vs ratios show variations withchanges in pressure and temperature. However, the mineral con-tent of rocks plays a significant role in Vp/Vs ratio variation. Theamount of quartz and plagioclase feldspar present in commonigneous rocks has the dominant effect on the Vp/Vs ratio (Chevrotand van der Hilst, 2000). In general, mafic rocks have a high Vp/Vsratio while felsic rocks have a low Vp/Vs ratio, due to their differentmineral contents.

The Vp/Vs ratio values are not distributed in a uniform pattern,but in general the Vp/Vs ratios correlate overall with geomorpho-logical units. We see that high average values (Fig. 10) can be

Fig. 8. Vp/Vs ratios for the station analyzed in this study.

C. Condori et al. / Journal of South American Earth Sciences 76 (2017) 11e2420

associated with regions of greater crustal thickness (WC and EC)and low values are mostly located in the coastal FA, the SA, and theAB where the crustal thickness is thinnest.

The variation in Vp/Vs ratios indicate significant lateral het-erogeneity beneath northern Peru. The FA, SA, and AB have thelowest Vp/Vs ratios (1.62e1.76), which imply that the crust may bemore felsic in composition. In contrast, high Vp/Vs ratios(1.78e1.88) are observed around the WC and EC. These higher Vp/Vs ratios indicate that the crust beneath these areas is probablymoremafic in composition particularly associated to an existence ofold volcanicarc.

In addition, the region presents values of heat flow that rangebetween 40 and 80 mW/m2 (Cardoso et al., 2010; Haraldsson,2011), which corresponds to the variation of Vp/Vs ratio andcrustal thickness (Fig. 10). This could be reflecting zones withremnants of magma partially molten in the crust which in turncause the hydrothermal activity and elevated temperatures seen innumerous surface manifestations including hot springs in this re-gion that have been linked to the geothermal gradient (Vargas andCruz, 2010).

We suggest the high Vp/Vs ratio beneath the high topographycould correspond to the Cenozoic volcanic deposits and any mafic

lower crustal material that was emplaced during the Cenozoicbefore the onset of completely flat slab subduction and the ongoingmigration of the Nazca Ridge (Antonijevic et al., 2015). To investi-gate this hypothesis we suggest the future application of tech-niques like tomography and seismic arrays scale CCP stacking of RFsin this region.

Our Hk-stacking analysis shows that the average Vp/Vs ratio ofthe crust for the entire area is 1.75. This value is consistent with theVp/Vs ratio of 1.75 used in previous RF studies in central Peru.Bishop et al. (2014) usedWadati diagrams (Wadati, 1933) in centralPeru for the PULSE array and determined a Vp/Vs ratio of1.75 ± 0.01 and this value is consistent with values determined byHk-stacking immediately south of our study area (Phillips et al.,2012) and more generally in the central Andes (Cunninghamet al., 1986; Dorbath and Granet, 1996). Expressed in terms ofPoisson's ratio, our average value is comparable to the global crustalaverage value (0.26e0.27) reported by Christensen (1996) andZandt and Ammon (1995).

The study region is characterized by crustal complexities andintense tectonic activity (active orogenesis, extensional andcompressional processes), and these complexities are reflected inthe RFs. However, the spatial density of seismometers and the good

Table 1Summary of Crustal Thickness (H) and Vp/Vs Ratio and their associated uncertainties fromH-k stacking for each seismic stations, using P-wave velocity of 6.1 km/s andweightsof W1 ¼ 0.7, W2 ¼ 0.2, and W3 ¼ 0.1 respectively.

N Station Net. Longitude Latitude Elevation H (km) Vp/Vs RFs a

01 PCM SNN �79.656 �7.407 37 29.5 ± 0.4 1.75 ± 0.02 11 2.502 CLB SNN �81.236 �4.253 4 28.2 ± 1.9 1.72 ± 0.06 19 2.503 MTP SNN �80.194 �3.683 80 35.7 ± 1.1 1.87 ± 0.06 16 2.504 LCN SNN �80.549 �4.641 145 36.9 ± 0.6 178 ± 0.03 19 2.505 CHL SNN �80.158 �5.095 165 33.8 ± 2.3 1.71 ± 0.05 29 2.506 FIC SNN �80.097 �5.918 78 30.6 ± 2.7 1.75 ± 0.07 20 2.507 CLL SNN �78.910 �6.184 1578 53.8 ± 1.9 1.77 ± 0.04 25 1.508 SIG SNN �79.012 �5.125 1818 52.6 ± 1.5 1.76 ± 0.04 32 1.509 BAG SNN �78.508 �5.636 747 54.5 ± 2.5 1.88 ± 0.06 27 1.510 MOY SNN �76.970 �6.075 959 39.8 ± 0.6 1.74 ± 0.05 13 2.511 BLV SNN �76.592 �7.006 330 39.4 ± 1.9 1.81 ± 0.06 25 2.512 BYV SNN �81.065 �5.7989 170 27.9 ± 1.1 1.63 ± 0.06 15 2.513 AS077-ATH CTBTO �78.395 �7.135 3151 54.2 ± 3.2 1.88 ± 0.06 49 1.514 TBM NSN �80.322 �3.523 24 31.9 ± 2.2 1.72 ± 0.06 15 2.515 YRM NSN �76.129 �5.897 149 37.9 ± 0.5 1.76 ± 0.03 58 2.516 CHO NSN �80.962 �5.167 221 27.1 ± 0.5 1.70 ± 0.02 43 2.517 NIEV NSN �77.866 �4.497 230 44.3 ± 0.4 1.68 ± 0.02 18 1.518 PCH NSN �79.682 �6.005 662 31.8 ± 1.1 1.74 ± 0.04 80 2.519 CHA NSN �77.877 �6.227 2370 47.2 ± 1.6 1.77 ± 0.07 53 1.520 TAR NSN �76.357 �6.496 358 37.1 ± 2.8 1.62 ± 0.05 61 2.521 TICA NSN �77.901 �7.917 2819 52.8 ± 1.8 1.78 ± 0.03 35 1.522 CBT NSN �78.521 �9.128 53 22.5 ± 0.8 1.68 ± 0.06 34 2.523 YANA NSN �76.112 �10.638 3835 62.7 ± 1.0 1.80 ± 0.02 27 1.024 OXA NSN �75.398 �10.578 1840 57.2 ± 2.4 1.83 ± 0.04 34 1.025 PUC NSN �74.668 �8.397 142 37.6 ± 1.1 1.78 ± 0.04 32 2.526 IQT NSN �73.320 �3.816 105 36.0 ± 2.7 1.74 ± 0.06 26 2.527 HCO NSN �76.249 �9.952 1966 63.3 ± 3.8 1.85 ± 0.04 20 1.028 YLS NSN �77.889 �8.847 3208 45.7 ± 1.0 1.79 ± 0.03 75 2.5

SNN-SisNetwork, CTBTO Network, NSN National Seismological Network, RFs e Number of seismic traces in each station, Gaussian filter with (a).

Fig. 9. Vp/Vs ratio versus crustal thickness for the stations analyzed in this study. Bars associated with each circles represent the estimated uncertainty. See Table 1 for numericalvalues.

C. Condori et al. / Journal of South American Earth Sciences 76 (2017) 11e24 21

azimuthal coverage of data have helped us in the first order esti-mate of the crustal structure andMoho geometry beneath northernPeru. These results should be useful for future studies of crustalstructure and evolution in northern Peru, however further inves-tigation is required for a better understanding of the seismologicalfeatures we have observed.

7. Conclusions

We have investigated the crustal structure beneath northernPeru, a region that has not been studied in detail previously, byapplying the RFmethod to estimate the crustal thickness and Vp/Vsratio at 28 broadband seismic stations. The distribution of the

Fig. 10. Cross section showing profiles along the section A-A0 from Fig. 1 and 7: a) Topography showing change in elevation associated with the different geomorphological units innorthern Peru, the black line represent the topography (GTOPO30, arc second) spatially filtered to remove features smaller than 50 km � 50 km to better match the expected spatialwavelength of isostatic compensation. b) Bouguer anomaly. c) Average Crustal Thickness interpolated from previous studies compiled by Assumpç~ao et al. (2013, 2015), averagevalues for each station along the section A-A0 and based on calculations assuming crustal isostasy under the Western and Eastern Cordilleras. d) Average Vp/Vs ratio and GeothermalHeat Flow (Cardoso et al., 2010).

C. Condori et al. / Journal of South American Earth Sciences 76 (2017) 11e2422

crustal thickness and Vp/Vs ratio show a good correlation with theregion's known geological and tectonic features.

Imaging of the P-to-s conversion from theMoho suggest that thecrustal thickness is around 25 km near the coast, have a maximumthickness of 55e60 km beneath the Andean Cordillera, and35e40 km further to the east in the Amazonian basin. The het-erogeneous and complex crustal composition is revealed by Vp/Vsratios that have high values (1.78e1.88) in the WC and EC and low

values (1.62e1.77) in the FA, SA, and AB.

� The maximum crustal thickness was found beneath the WC,consistent with isostatic hypothesis. The EC crust is not as thickas the WC, which implies an isostatic imbalance. The crust thinsto the west beneath the FA and to the east beneath the SA andAB.

C. Condori et al. / Journal of South American Earth Sciences 76 (2017) 11e24 23

� The Vp/Vs ratios correlate with the variations in heat flow,crustal thickness, and topography with the high elevations ofthe WC and EC having high Vp/Vs ratios and the low elevationsof the FA, SA and AB having lower Vp/Vs ratios. This suggests adominantly felsic composition in the middle crust and an in-termediate to more mafic composition at the base of the crustunder high elevations.

� A rapid transition is seen in crustal thickness between the FAand WC, compared to a more gradual transition between EC, SAand AB.

� Our results provide some constraints on understanding themechanism of uplift and crustal thickening in the region andhelp to increase the database of crustal thicknesses for betterunderstanding the geodynamic of Northern Peru.

Acknowledgements

The data used in this study were provided by the Peruvian Na-tional Seismic Network and the Instituto Geofísico del Perú. Thisresearch was supported by CAPES. We would like to thank theObservatorio Sismol�ogico-OBSIS and Universidade de Brasília-UnB.We also thank Marcelo Assumpç~ao (IAG-USP) for providingcompiled data of crustal thickness for adjacent regions to our studyarea and for valuable contributions to the manuscript. We alsothank to Marcelo Peres for his contributions. GSF thanks CNPq forhis PQ grants (307255/2013-1).

References

Ammon, C.J., 1991. Isolation of receiver effects from teleseismic P waveforms. Bull.Seismol. Soc. Am. 81, 2504e2510.

Ammon, C.J., Randall, G.E., Zandt, G., 1990. On the nonuniqueness of receiverfunction inversions. J. Geophys. Res. 95, 15303. http://dx.doi.org/10.1029/JB095iB10p15303.

Antonijevic, S.K., Wagner, L.S., Kumar, A., Beck, S.L., Long, M.D., Zandt, G., Tavera, H.,Condori, C., 2015. The role of ridges in the formation and longevity of flat slabs.Nature 524, 212e215. http://dx.doi.org/10.1038/nature14648.

Aranda, N., Assumpç~ao, M., 2013. Crustal thickness in the northern Andes fromteleseismic pP and sS precursors. In: 13th International Congress of the Bra-zilian Geophysical Society. Society of Exploration Geophysicists and BrazilianGeophysical Society, EXPOGEF, Rio de Janeiro, Brazil, pp. 1781e1785. http://dx.doi.org/10.1190/sbgf2013-366, 26e29 August 2013.

Assumpç~ao, M., Bianchi, M., Albuquerque, D.F., França, G.S.L.A., Barros, L., 2015.Crustal Thickness Map in South America: an updated version. Extend Abstract1st Brazilian Symposium on Seismology 1e4.

Assumpç~ao, M., Feng, M., Tassara, A., Juli�a, J., 2013. Models of crustal thickness forSouth America from seismic refraction, receiver functions and surface wavetomography. Tectonophysics 609, 82e96. http://dx.doi.org/10.1016/j.tecto.2012.11.014.

Audebaud, E., Capdevila, R., Dalmayrac, B., Laubaucher, G., Marocco, R., Mattauer, M.,Megard, F., Paredes, J., 1973. Les traits geologiques essentials des Andes Cen-trales (Perou-Bolivie). Revue Geographie Physique, Geologique et Dinamique 15(1e2), 73e114.

Bishop, B., Beck, S., Zandt, G., Scire, A., Wagner, L., Long, M., Tavera, H., 2014. Trenchto andean Thrust Front: Evidence for Coupling of the Peruvian Flat Slab to theOver-riding South American Plate. American Geophysical Union. Fall Meeting2014, abstract #T23A-4629 2014AGUFM.

Cahill, T., Isacks, B.L., 1992. Seismicity and shape of the subducted Nazca plate.J. Geophys. Res. 97, 17503. http://dx.doi.org/10.1029/92JB00493.

Cardoso, R.R., Hamza, V.M., Alfaro, C., 2010. Geothermal resource base for SouthAmerica: A continental perspective. Proceedings World Geothermal Congress2010, Bali, Indonesia, 25-29 April 2010, 6 pp.

Cassidy, J.F., Ellis, R.M., 1993. S wave velocity structure of the northern Cascadiasubduction zone. J. Geophys. Res. 4407e4421.

Chevrot, S., van der Hilst, R.D., 2000. The Poisson ratio of the Australian crust:geological and geophysical implications. Earth Planet. Sci. Lett. 183, 121e132.http://dx.doi.org/10.1016/S0012-821X(00)00264-8.

Christensen, N.I., 1996. Poisson's ratio and crustal seismology. J. Geophys. Res. 101,3139. http://dx.doi.org/10.1029/95JB03446.

Chulick, G.S., Detweiler, S., Mooney, W.D., 2013. Seismic structure of the crust anduppermost mantle of South America and surrounding oceanic basins. J. S. Am.Earth Sci. 42, 260e276. http://dx.doi.org/10.1016/j.jsames.2012.06.002.

Cobbing, E.J., 1981. Tillites at the Base of the Possible Early Palaeozoic MarconaFormation, Southwest Coastal Peru. In: Hambrey, M.J., Harland, W.B. (Eds.),Earth's Pre-Pleistocene Glacial Record. Cambridge University Press, Cambridge,pp. 899e901.

Crotwell, H.P., Owens, T.J., Ritsema, J., 1999. The TauP Toolkit: flexible seismic travel-time and ray-path utilities. Seismol. Res. Lett. 70, 154e160. http://dx.doi.org/10.1785/gssrl.70.2.154.

Cunningham, P.S., Roecker, S.W., Hatzfeld, D., 1986. Three-dimensional P and S wavevelocity structures of southern Peru and their tectonic implications. J. Geophys.Res. 91, 9517e9532.

Dalmayrac, B., Marocco, R., Laubacher, G., 1987. Caracteres generaux de l’evolutiongeologique des Andes peruviannes. ORDTOM, pp. 122e501.

Dalmayrac, B., Marocco, R., Laubacher, G., 1980. Geologie des Andes peruviannes,Caracteres generaux de l’evolution geologique des Andes peruviennes. ORD-TOM, pp. 96e217.

DeMets, C., Gordon, R.G., Argus, D.F., Stein, S., 1990. Current plate motions. Geophys.J. Int. 101, 425e478. http://dx.doi.org/10.1111/j.1365-246X.1990.tb06579.x.

Dewey, J.F., Bird, J.M., 1970. Mountain belts and the new global tectonics. J. Geophys.Res. 75, 2625e2647.

Dorbath, C., Granet, M., 1996. Local earthquake tomography of the Altiplano and theEastern Cordillera of northern Bolivia. Tectonophysics 259, 117e136.

Dumont, J.F., Deza, E., Garcia, F., 1991. Morphostructural provinces and neotectonicsin the Amazonian lowlands of Peru. J. S. Am. Earth Sci. 4, 373e381. http://dx.doi.org/10.1016/0895-9811(91)90008-9.

Eaton, D.W., Dineva, S., Mereu, R., 2006. Crustal thickness and Vp/Vs variations inthe Grenville orogen (Ontario, Canada) from analysis of teleseismic receiverfunctions. Tectonophysics 420, 223e238.

Efron, B., Tibshirani, R., 1991. Statistical data analysis in the computer age. Science253, 390e395.

F€orste, C., Bruinsma, S., Marty, J.-C., Flechtner, F., Abrykosov, O., Dahle, C.,Lemoine, J.M., Neumayer, K.-H., Biancale, R., Barthelmes, F., K€onig, R., 2014.EIGEN-6C4 The Latest Combined Global Gravity Field Model Including GOCEData up to Degree and Order 2190 of GFZ Potsdam and GRGS Toulouse. GFZData Services.

Frassetto, A.M., Zandt, G., Gilbert, H., Owens, T.J., Jones, C.H., 2011. Structure of theSierra Nevada from receiver functions and implications for lithospheric foun-dering. Geosphere 7 (4), 898e921. http://dx.doi.org/10.1130/GES00570.1.

Fukao, Y., Kono, M., Yamamoto, A., Saito, M., Nawa, K., Giesecke, A., Perales, P., 1999.Gravity measurements and data reduction for Bouguer Anomaly Map of Peru.Bull. Earthq. Res. Inst. Univ. Tokyo 74, 161e266.

Fukao, Y., Yamamoto, A., Kono, M., 1989. Gravity anomaly across the PeruvianAndes. J. Geophys. Res. 94, 3867. http://dx.doi.org/10.1029/JB094iB04p03867.

Gansser, A., 1973. Facts and theories on the Andes. J. Geol. Soc. 129, 93e131.Gutscher, M.A., Olivet, J.-L., Aslanian, D., Eissen, J.-P., Maury, R., 1999. The “lost Inca

Plateau”: cause of flat subduction beneath Peru? Eart planet. Sci. Lett. 171,335e341.

Gutscher, M.A., Spakman, W., Bijwaard, H., Engdahl, E.R., 2000. Geodynamics of flatsubduction: seismicity and tomographic constraints from the Andean margin.Tectonics 19, 814e833.

Hampel, A., 2002. The migration history of the Nazca Ridge along the Peruvianactive margin: a re-evaluation. Earth Planet. Sci. Lett. 203, 665e679. http://dx.doi.org/10.1016/S0012-821X(02)00859-2.

Haraldsson, I.G., 2011. Geothermal activity and development in South America:Short overview of the status in Bolivia, Chile, Ecuador and Peru. Presented at“Short Course on Geothermal Drilling, Resource Development and PowerPlants”, Santa Tecla, El Salvador, January 16e22 18.

Hayes, G.P., Wald, D.J., Johnson, R.L., 2012. Slab1.0: a three-dimensional model ofglobal subduction zone geometries. J. Geophys. Res. 117, B01302. http://dx.doi.org/10.1029/2011JB008524.

Henry, S.G., Pollack, H.N., 1988. Terrestrial heat flow above the Andean subductionzone in Bolivia and Peru. J. Geophys. Res. 93, 15153. http://dx.doi.org/10.1029/JB093iB12p15153.

Hussong, D.M., Edwards, P.B., Johnson, S.H., Campbell, J.F., Sutton, G.H., 1976. Crustalstructure of the PerueChile trench: 8�e12�S latitude. In: The Geophysics of thePacific Ocean Basin and its Margins. AGU Geophysical Monograph, vol. 19,pp. 71e81.

James, D.E., 1971. Andean crustal and upper mantle structure. J. Geophys. Res. 76,3246e3271. http://dx.doi.org/10.1029/JB076i014p03246.

Kendrick, E., Bevis, M., Smalley, R., Brooks, B., Vargas, R.B., Lauría, E., Fortes, L.P.S.,2003. The NazcaeSouth America Euler vector and its rate of change. J. S. Am.Earth Sci. 16, 125e131. http://dx.doi.org/10.1016/S0895-9811(03)00028-2.

Kono, M., Fukao, Y., Yamamoto, A., 1989. Mountain building in the central Andes.J. Geophys. Res. Solid Earth 94, 3891e3905. http://dx.doi.org/10.1029/JB094iB04p03891.

Krabbenh€oft, A., Bialas, J., Kopp, H., Kukowski, N., Hübscher, C., 2004. Crustalstructure of the Peruvian continental margin from wide-angle seismic studies.Geophys. J. Int. 159, 749e764. http://dx.doi.org/10.1111/j.1365-246X.2004.02425.x.

Kulm, L.D., Prince, R.A., French, W., Johnson, S., Masias, A., 1981. Crustal structureand tectonics of the central Peru continental margin and trench. Geol. Soc. Am.Mem. 154, 445e468.

Langston, C.A., 1979. Structure under Mount Rainier, Washington, inferred fromteleseismic body waves. J. Geophys. Res. 84, 4749. http://dx.doi.org/10.1029/JB084iB09p04749.

Ligorría, J.P., Ammon, C.J., 1999. Iterative deconvolution and receiver-functionestimation. Bull. Seism. Soc. Am. 89.

Lindo, R., 1993. Seismotectonique des andes du Perou central: Apport des donnessismologiques de haute precision. Ph.D. Thesis. Universite Louis-Pasteur,Strasbourg, France, p. 74.

C. Condori et al. / Journal of South American Earth Sciences 76 (2017) 11e2424

Lindquist, K.G., Engle, K., Stahlke, D., Price, E., 2004. Global topography and ba-thymetry grid improves research efforts. Eos, Trans. Am. Geophys. Union 85,186. http://dx.doi.org/10.1029/2004EO190003.

Lloyd, S., van der Lee, S., França, G.S., Assumpç~ao, M., Feng, M., 2010. Moho map ofSouth America from receiver functions and surface waves. J. Geophys. Res. 115,B11315. http://dx.doi.org/10.1029/2009JB006829.

Martinod, J., Husson, L., Roperch, P., Guillaume, B., Espurt, N., 2010. Horizontalsubduction zones, convergence velocity and the building of the Andes. EarthPlanet. Sci. Lett. 299, 299e309. http://dx.doi.org/10.1016/j.epsl.2010.09.010.

Megard, F., 1978. Etude geologique des Andes du P�erou Central. Memoires ORSTOM86, 310.

M�egard, F., 1987. Cordilleran Andes and marginal Andes: a review of Andean ge-ology north of the Arica elbow (18�s). Geodyn. Ser. 18, 71e95. http://dx.doi.org/10.1029/GD018p0071.

Mitouard, P., Kissel, C., Laj, C., 1990. Post-Oligocene rotations in southern Ecuadorand northern Peru and the formation of the Huancabamba deflection in theAndean Cordillera. Earth Planet. Sci. Lett. 98, 329e339. http://dx.doi.org/10.1016/0012-821X(90)90035-V.

Molnar, P., England, P., 1990. Late Cenozoic uplift of mountain ranges and globalclimate change: chicken or egg? Nature 346, 29e34. http://dx.doi.org/10.1038/346029a0.

Mourier, T., M�egard, F., Rivera, L., Arguedas, A.P., 1988. L’�evolution m�esozoique desAndes de Huancabamba (nord P�erou-sud Ecuateur) et l’hypoth�ese de l’accr�etiondu bloc Amotape-Tahuin. Bull. Soc. G�eol. Fr. 4, 69e79.

Musacchio, G., Mooney, W.D., Luetgert, J.H., Christensen, N.I., 1997. Composition ofthe crust in the Grenville and Appalachian provinces of North America inferredfrom V P/V S ratios. J. Geophys. Res. 102, 15225. http://dx.doi.org/10.1029/96JB03737.

Norabuena, E.O., Dixon, T.H., Stein, S., Harrison, C.G.A., 1999. Decelerating Nazca-South America and Nazca-Pacific Plate motions. Geophys. Res. Lett. 26,3405e3408. http://dx.doi.org/10.1029/1999GL005394.

Ocola, L., Meyer, R.L., Aldrich, L.T., 1971. Cross crustal structure under Peru-Boliviaaltiplano. In: Earthquakes Notes, vol. 42. Seismological Society of America,pp. 3e4.

Owens, T.J., Zandt, G., Taylor, S.R., 1984. Seismic evidence for an ancient rift beneaththe Cumberland Plateau, Tennessee: a detailed analysis of broadband tele-seismic P waveforms. J. Geophys. Res. 89, 7783. http://dx.doi.org/10.1029/JB089iB09p07783.

Pav~ao, C.G., França, G.S., Marotta, G.S., Menezes, P.H.B.J., Neto, G.B.S., Roig, H.L., 2012.Spatial interpolation applied to crustal thickness in Brazil. J. Geogr. Inf. Syst. 4(2), 142e152.

Phillips, K., Clayton, R.W., 2014. Structure of the subduction transition region fromseismic array data in southern Peru. Geophys. J. Int. 196, 1889e1905. http://dx.doi.org/10.1093/gji/ggt504.

Phillips, K., Clayton, R.W., Davis, P., Tavera, H., Guy, R., Skinner, S., Stubailo, I.,Audin, L., Aguilar, V., 2012. Structure of the subduction system in southern Perufrom seismic array data. J. Geophys. Res. 117, B11306. http://dx.doi.org/10.1029/2012JB009540.

Poveda, E., Monsalve, G., Vargas, C.A., 2015. Receiver functions and crustal structureof the northwestern Andean region, Colombia. J. Geophys. Res. Solid Earth 120,2408e2425. http://dx.doi.org/10.1002/2014JB011304.

Ramos, V.A., Folguera, A., 2009. Andean flat-slab subduction through time. Geol.Soc. Lond., Spec. Publ. 327, 31e54.

Rosenbaum, G., Giles, D., Saxon, M., Betts, P.G., Weinberg, R.F., Duboz, C., 2005.Subduction of the Nazca Ridge and the Inca Plateau: insights into the formationof ore deposits in Peru. Earth Planet. Sci. Lett. 239, 18e32. http://dx.doi.org/10.1016/j.epsl.2005.08.003.

Ryan, J., Beck, S., Zandt, G., Wagner, L., Minaya, E., Tavera, H., 2016. Central Andeancrustal structure from receiver function analysis. Tectonophysics. http://dx.doi.org/10.1016/j.tecto.2016.04.048.

Soler, P., S�ebrier, M., 1990. Nazca slab retreat versus compressional deformation inthe central Andes since late Oligocene times. Symposium internationalg�eodynamique andine: R�esum�es des communications 187e190.

Su�arez, G., Molnar, P., Burchfiel, B.C., 1983. Seismicity, fault plane solutions, depth offaulting, and active tectonics of the Andes of Peru, Ecuador, and southernColombia. J. Geophys. Res. 88, 10403. http://dx.doi.org/10.1029/JB088iB12p10403.

Tankard, A.J., Su�arez, S.R., Welsink, H.J., 1995. Petroleum geology of the sub-Andeanbasins of Peru. Petroleum basins S. Am. AAPG Menoir 62, 423e444.

Tassara, A., Echaurren, A., 2012. Anatomy of the Andean subduction zone: threedimensional density model upgraded and compared against global-scalemodels. Geophys. J. Int. 189, 161e168.

Tavera, H., 1990. Interpretaci�on de las Anomalías de Estaci�on a partir de fases P yPKIKP para la regi�on del Perú Central. Boletín Sociedad Geol�ogica del Perú 81,47e54.

Tavera, H., Buforn, E., 1998. Sismicidad y sismotect�onica de Perú. Fisica de la Tierra10, 187e219.

Vargas, V., Cruz, V., 2010. Geothermal Map of Perú, in: Proceedings WorldGeothermal Congress 2010. Bali, Indonesia, p. 7.

Villegas-Lanza, J.C., Chlieh, M., Cavali�e, O., Tavera, H., Baby, P., Chire-Chira, J.,Nocquet, J.-M., 2016. Active tectonics of Peru: heterogeneous interseismiccoupling along the Nazca megathrust, rigid motion of the Peruvian Sliver, andSubandean shortening accommodation. J. Geophys. Res. Solid Earth 121,7371e7394. http://dx.doi.org/10.1002/2016JB013080.

Villegas, J.C., 2009a. Processing and Analysis of the LISN Permanent GPS Network:Preliminary Tectonic Results. Nice-France.

Villegas, J.C., 2009b. Modelos de Velocidad Unidimensionales para las RegionesNorte, Centro y Sur del Perú, a partir de la Inversi�on de los Tiempos de arribo delas ondas P y S de sismos locales. Tesis de Ingeniería U.N.S.A, pp. 56e75.

Wadati, K., 1933. On the travel time of earthquake waves. Part II. Geophys. Mag. 7,101e111.

Wipf, M.A., 2006. Evolution of the Western Cordillera and Coastal Margin of Peru:evidence from low-temperature Thermochronology and Geomorphology. SwissFed. Inst. Technol. Zürich 163.

Yuan, X., Sobolev, S., Kind, R., 2002. Moho topography in the central Andes and itsgeodynamic implications. Earth Planet. Sci. Lett. 199, 389e402. http://dx.doi.org/10.1016/S0012-821X(02)00589-7.

Zandt, G., Ammon, C., 1995. Continental crust composition constrained by mea-surements of crustal Poisson's ratio. Nature 374, 152e154.

Zhu, L., Kanamori, H., 2000. Moho depth variation in southern California fromteleseismic receiver functions. J. Geophy. Res. 105, 2969e2980.