Tearing and Breaking Off of Subducted Slabs as the Result of Collision of the Panama Arc-Indenter with Northwestern South America by Carlos A. Vargas and Paul Mann Abstract We present two regional, lithospheric cross sections that illustrate east- ward- and southeastward-dipping, subducted slabs to depths of 315 km beneath the surface of Colombia in northwestern South America. These cross-sectional interpre- tations are based on relocated earthquake hypocentral solutions, models supported on gravity and magnetic regional data, and coda-Q (Qc) tomography. The method of tomographic imaging based on spatial inversion of the coda wave has advantages of providing information on the lateral variations of the anelastic properties and ther- mal structure of the lithospheric system. Mapping of earthquake-defined Benioff zones combined with tomographic imaging reveals the presence of an ∼240 km long east–west-striking slab tear, named here the Caldas tear. The proposed Caldas tear separates a zone of shallow, 20°–30°-dipping, southeastward subduction in the area of Colombia adjacent to Panama and the Caribbean Sea, which is not associated with subduction-related volcanism, from an area of steeper, 30°–40°-dipping, slab adjacent to the eastern Pacific Ocean that is associated with an active north–south chain of active arc volcanoes. We propose that the Caldas slab tear separating these two distinct subducted slabs originally formed as the southern boundary of the Panama indenter, an extinct island arc that began subducting beneath northwestern South America about 12 Ma. The area south of the Panama indenter is Miocene oceanic crust of the Nazca plate, which subducts eastward beneath northwestern South America at normal angles and melts to form a north–south-trending active volcanic arc. In addition to the for- mation of the Caldas tear, we propose that impedance of the thicker crustal area of the Panama arc-indenter over the past 12 Ma may have led to down-dip break-off of previously subducted oceanic crust that is marked by an extremely concentrated and active earthquake swarm of intermediate-depth earthquakes beneath east-central Colombia. Introduction and Tectonic Setting Hypocentral solutions recorded by the Colombian Na- tional Seismological Network (CNSN) show an ∼240 km long, right-lateral offset of intermediate to deep events with azimuth of 102° (Fig. 1a,b). We infer this discontinuity in earthquakes to be a major slab tear which we have named the Caldas tear based on the location in the Caldas department of Colombia and the alignment of fault-related surface features (e.g., volcanism, faulting,mineral deposits, geothermal anoma- lies, etc.). Using the distribution of earthquakes >80 km, Ojeda and Havskov (2001) proposed that the discontinuity along the Caldas tear represented a boundary between two sub- ducted slabs with differing dips and strikes: the northern sub- duction zone, called the Bucaramanga subduction zone, has a shallower dip (27°) and more northeasterly strike, and the southern, called the Cauca subduction zone, has a steeper dip (35°–40°) and a more northerly strike (Fig. 1a). Regional compilations of Global Positioning Systems (GPS) data provide a quantitative tectonic framework for understanding the widespread crustal effects of the Panama arc collision on large areas of northwestern South America (Calais and Mann, 2009; Fig. 1a). GPS vectors in western Colombia show a marked decrease in velocities consistent with the ongoing collision of the Panama arc with north- western South America along a north–south-trending suture zone roughly parallel to the international boundary between Panama and Colombia (Adamek et al., 1988; Trenkamp et al., 2002; Corredor, 2003; Fig. 1a). The east–west direc- tion of GPS vectors shows that the effects of east–west shortening and indentation related to the collision of the Pan- ama arc remains relatively constant over a large, V-shaped, fault-bounded area of Colombia due east of the Panama arc- indenter (Fig. 1b). GPS vectors on the Maracaibo block of 2025 Bulletin of the Seismological Society of America, Vol. 103, No. 3, pp. 2025–2046, June 2013, doi: 10.1785/0120120328 Colombia and Venezuela show a more northerly direction of plate motions related to northward tectonic escape of the Maracaibo block into the southern Caribbean (Trenkamp et al., 2002). In contrast to this fairly uniform GPS velocity field of deformed crustal rocks produced by the Panama col- lision, underlying, eastward-dipping slabs change abruptly across the Caldas tear from dip angles of 30°–40° between latitudes ∼3:0°–5:6° N in southern Colombia, to dip angles of 20°–30° in the area north of ∼5:6° N (Ojeda and Havskov, 2001; Vargas et al., 2007). Two nests of concentrated intermediate-depth earth- quakes are present beneath Colombia (Fig. 1a). The Bucara- manga earthquake nest (BN) is found at a depth of ∼160 km on the down-dip extension of the southern (Bucaramanga) subduction zone and has an estimated volume dimension of ∼13 × 18 × 12 km (Schneider et al., 1987; Frohlich et al., Figure 1. (a) Tectonic map of northwestern South America and Panama showing plate boundaries, neotectonic fault systems, and se- lective distribution of hypocentral solutions of ∼30;000 earthquakes extracted from the entire catalog of the CNSN (∼102;000 events) during 1993–2012 with these criteria:mL ≥ 0:5; GAP ≤ 200; rms ≤ 0:5; error in latitude ≤10:0 km; error in longitude ≤10:0 km; and error in depth≤10:0 km. Color scale indicates depth of earthquakes. The north and south profiles symbolize the tomographic sections presented in this study. SMM, Santa Marta massif; CB, Choco block; WC, Western Cordillera; CC, Central Cordillera; EC, Eastern Cordillera; PR, Perija Range; GB, Guajira basin; LB, Llanos foreland basin; MMVB, Middle Magdalena Valley basin; RFZ, Romeral fault zone; SMBF, Santa Marta– Bucaramanga fault; PF, Palestina fault; CF, Cimitarra fault; MGF, Mulato–Getudo fault; HF, Honda fault; SFS, Salinas fault system; GF, Garrapatas fault; LFS, Llanos fault system; IF, Ibague fault; SR, Sandra ridge; BN, Bucaramanga nest; CN, Cauca nest; MN, Murindo nest; PIVC, Paipa–Iza volcanic complex; RSDV, Romeral and San Diego volcanoes. Yellow stars correspond to (1) the Tauramena earthquake (19 January 1995,Mw 6.5); (2) the Armenia earthquake (25 January 1999,Mw 6.2); and (3) the Quetame earthquake (24 May 2008, mL � 5:7). Sections AA0 and BB0 correspond to tomographic profiles presented in Figures 5 and 6. (b) Crustal isochron pattern of the Sandra ridge; pink- colored line, Caldas tear zone; arrows, station velocity GPS vectors relative to stable South America (after Calais and Mann, 2009). CHEP and BOGO are GPS stations used as reference to estimate the onset of the Panama-arc and South American plate collision. Other GPS stations in the Panama-arc collision area are MANZ, RION, BUCM,MONT, and CART. Faded blue arrow enclosing 102° azimuth of the approximately 240 km long, right-lateral offset of intermediate to deep events associated with the Caldas tear. 2026 C. A. Vargas and P. Mann 1995). Previous tectonic interpretations of the origin of the Bucaramanga nest vary from a zone of two slabs in contact (van der Hilst and Mann, 1994), two slabs overlapping (Taboada et al., 2000), or a single slab undergoing extreme bending (Cortés and Angelier, 2005) all occurring in the boundary area of the subducted northern (Bucaramanga) and southern (Cauca) subduction zones (Fig. 1a). The Cauca intermediate-depth earthquake nest (CN) is located ∼400 km southwest of the Bucaramanga nest on the trend of our pro- posed Caldas tear and has been previously interpreted by Cortés and Angelier (2005) as a bend in the slab in this area (Fig. 1a). There is no clear consensus among seismol- ogists for the tectonic interpretation of the two concentrated Colombian intermediate earthquake nests (Frohlich, 2006; Zarifi, 2006). The Caldas tear defines the northern limit of the active volcanic front of the northern Andes that has formed as a consequence of the steeper subduction of oceanic slab of normal thickness of the Nazca plate (Fig. 1a). Moreover, associated with active and inactive volcanoes, the east–west projected surface trace of the Caldas tear localizes an east– west alignment of some unusual volcanic rocks including adakites (Borrero et al., 2009; Fig. 1a). Other volcanic rocks in the vicinity of the east–west-trending Caldas tear include the Plio-Pleistocene Paipa-Iza volcanic complex in the Eastern Cordillera of Colombia and the Romeral and San Diego volcanoes (Pardo et al., 2005). The presence of these east–west aligned volcanic rocks along with locally elevated geothermal gradient values (Vargas et al., 2009) suggests that the Caldas tear may penetrate the upper crust as a fault zone and provide a conduit for the upward rise of magmas and hydrothermal fluids produced by melting of the slabs on ei- ther side of the Caldas tear (Fig. 2). Furthermore, recent, shallow-focus, strong motion events such as the Tauramena earthquake (19 January 1995; Mw 6.5, h � 25� 10 km), the Quindio earthquake (25 January 1999; Mw 6.2, h � 18:6 km), and the Quetame earthquake (24 May 2008; mL 5.7; h � superficial) are all in alignment with the surface trace of the Caldas tear. Previous tomographic studies using both local and regional earthquakes of varying resolution have produced differing tectonic interpretations for slabs in this area (van der Hilst and Mann, 1994; Taboada et al., 2000; Vargas et al., 2007). In this paper, we present the results of an in- tegrated geophysics and geologic study that improves the 3D imaging of the interactions between the eastward-moving Panama indenter and its collisional area in northwestern South America. Data and Methods The following sections describe data and procedures used to estimate hypocentral solutions, the attenuation and its spatial distribution, the simultaneous 2D inversion of gravity and magnetic data, and the correlation of these results with focal mechanisms, geothermal gradients, geological maps, and a high-resolution seismic profile, seeking to define the geometry of the Caldas tear and its geotectonic implications in the northwestern corner of South America. Hypocentral Solutions and Estimation of the Coda-Wave Attenuation A catalog has been compiled of ∼102;000 earthquake locations calculated by the CNSN during the period 1993– 2012 (mL ≤ 6:8). Hypocentral solutions were estimated by using a seismological array of 17 short-period instruments (T � 1 s) of the CNSN and complemented by 13 stations associated with local volcanic monitoring systems and also foreign networks (Panama, Ecuador, and Venezuela). Final solutions were calculated with the HYPOCENTER program and the velocity model proposed by Ojeda and Havskov (2001). Then, 9338 waveforms associated with 7645 regional earthquakes (3:0 ≤ mL ≤ 6:5; 1993–2012) were selected for estimating the decay rate of the coda amplitudes (Q−1c , coda attenuation). The selected events, on basis of a significant number of stations that recorded them (Table 1), have epicen- tral distances to stations ranging between 22.6 and 690.0 km and depths varying between 0 and 222.0 km. Figure 2 shows all events used for the Q−1c plotted on a map of northwestern South America along with tectonically significant earth- quake focal mechanisms including the Quindio, Quetama, and Tauramena events aligned along the Caldas tear and in- termediate-depth focal mechanisms from the Bucaramanga and Cauca nests. Estimations of the Q−1c were done using the Single Backscattering model proposed by Aki and Chouet (1975). This model assumes that the coda of a local earthquake is composed of the sum of secondary S waves produced by heterogeneities distributed randomly and uniformly within the lithosphere. The coda is the portion of a seismogram cor- responding to back-scattered S-waves. The estimation of Q−1c used the following equation: P�ω; t� � 2g�θ�jS�ω�j 2 βt2 e � −ωt Qc � ; (1) where P�ω; t� is the time-dependent coda power spectrum, ω is the angular frequency, β is the shear-wave velocity, jS�ω�j is the source spectrum, and g�θ� represents the directional scat- tering coefficient. The g�θ� term has been defined as 4π times the fractional loss of energy by scattering, per unit travel dis- tance of primary waves, and per unit solid angle of the radi- ation direction θmeasured from the direction of primary wave propagation. Using these assumptions, the geometrical spread- ing is assumed to be proportional to r−1, which only applies to body waves in a uniform medium. The source factor can be treated as a constant value for single frequency. According to equation (1), Q−1c values can be obtained as the slope of the least-squares fit ofLn�t2 × P�ω; t�� versus ωt, for t > tβ, where tβ represents the S-wave travel time (Haskov et al., 1989). The Tearing and Breaking Off of Subducted Slabs as the Result of Collision of the Panama Arc-Indenter 2027 time-dependent coda power spectrum was calculated using the mean squared amplitudes of the coda Aobs�ω; t� from band- pass-filtered seismograms around a center frequency. In order to take into account the deep structure using coda waves, β � 4:64 km=s was assumed and calculated as a weighted average of S-wave speeds in the whole earth vol- ume covered by the scattered waves (Badi et al., 2009). All records were filtered in a chosen frequency band and then used a coda-wave time window (W) of 20 s, starting from 2 × tβ s�tstart�. The average lapse time, defined as tc � tstart� W=2 ranges between 11.0 and 384.0 s. These large tc values ensure the sampling of regional structures. Attenuation esti- mates were performed with short-period records (T � 1 s) at several frequencies (Table 1). Then we chose estimates in the frequency band 1–3 (2� 1) Hz because of the high availabil- ity and geographical distribution of observations regarding other frequencies; and also best values of correlation coeffi- cients, the root mean square (rms), and signal-to-noise ratio. In addition, it has been reported that the study region presents Q−1c values in this frequency band with errors ≤5% (e.g., see Vargas et al. (2004)). In general, errors seen along Q−1c estimations are acceptable, for example, the rms of all esti- mations vary between 0.07 and 1.79 (μ � 0:24, σ � 0:07) and the coefficients of correlation are oscillating between Figure 2. Epicenter projection of events used during the coda-wave-attenuation (Q−1c ) estimation. Colored circles, earthquakes; blue squares, locations of all seismological stations used in this paper; gray stars are shown with large focal mechanisms, and the most recent and surficial strong-motion events occurring along the Caldas tear are shown by banded-gray polygon. The main focal mechanisms reported by the NEIC-USGS (mb ≥ 4:0) defining the Bucaramanga nest to the northeast and the Cauca nest to the southwest are shown; pink areas identified in the epicentral location of these nests are two main geothermal gradient anomalies reported by Vargas et al. (2009). 2028 C. A. Vargas and P. Mann −0:5 and −0:97 (μ � −0:67, σ � 0:11). Table 2 presents a statistical summary of the main parameters related with the estimation of Q−1c values for 30 seismological stations. Figure 3a shows an example of typical waveform used dur- ing this analysis, as well the corresponding record filtered for the chosen frequency band. The attenuation factor (Q−1c ) is suggested as a decay factor for the coda-wave amplitudes. Figure 3b presents histograms for Q−1c values and their cor- relation coefficients, as well as distributions for the epicentral distances, focal depths, and local magnitudes of the events analyzed. Figures 2 and 3b emphasize the presence of attenu- ation contrasts in the region and at least two sources of events, one of them surficial and dispersed, and the other located at an intermediate depth (linked to the nests of Buca- ramanga and Cauca). Tomographic Imaging Using Coda-Wave Attenuation Mukhopadhyay and Sharma (2010) have proposed that the variation of Q−1c with tc shows a direct relationship with depth. These authors interpreted that Q−1c values related to scattering processes that penetrate >200 km depth are con- trolled by a crust and a relatively more transparent mantle. These results support the idea thatQ−1c estimated with a large tc is representative of a large sampled volume and large sampled depths. A corollary of this hypothesis is that the Q−1c value must be near to the intrinsic absorption (Q−1i ) con- trolled mainly by the mantle. Following these ideas, Vargas et al. (2004) developed a regional tomographic study using stations of the CNSN with relative large tc, (up to 180 s) and found that the Q−1c values are near to the Q−1i values for almost all stations, meaning that a large portion of the upper mantle is being sampled. Other studies have suggested a direct relation between the thermal field and anelastic attenu- ation (Faul and Jackson, 2005; Priestley and McKenzie, 2006; Yang et al., 2007). The physical meaning of this rela- tionship is not been completely understood, but Karato and Jung (1998) proposed that the higher water content in the asthenosphere significantly reduces the seismic-wave veloc- ities through anelastic relaxation and increasing temperature. Convergent margins such as Colombia, which involve large amounts of sediments and water mobilized during the sub- duction processes, are a likely site of large contrasts in anelastic attenuation in the subducted lithospheric slabs. Given the ease to estimate Q−1c , we can use this obser- vation for highlighting regional structures related to contrasts in rigidity (e.g., crust or lithospheric plates). One way to regionalize Q−1c is based on the work of Malin (1978) who, expanding on the work of Aki (1969) and Aki and Chouet (1975), realized that the first-order scatterers responsible for the generation of coda waves at any given tc can be located on the surface of an ellipsoid with earthquake and station locations as foci (Singh and Herrmann, 1983). In the ellip- soidal volume sampled by coda waves at any time t, Pulli (1984) defined the large semi-axis as a1 � βt=2, and defined the small semi-axis as a2 � a3 � �a21 − r2=4�1=2, where r is the source–receiver distance of the ellipsoid. The horizontal projection of this volume is coincident with the elliptical en- velope proposed by several authors as the area occupied by the scattered energy of the coda-wave record (Mitchell et al., 1997; Mitchell and Cong, 1998; Xie, 2002; Vargas et al., 2004). Following these observations and knowing the values of tc, W, and β, it is possible to deduce the volumes of the ellipsoidal shells where the seismic energy is scattered. HenceQ−1c values estimated with large tc correspond to large sampled volumes, and vice versa. Based on these hypotheses we can perform a generalized inversion for regionalizing Q−1c . For the purpose of the inversion, we define a geo- graphic grid around the seismic station that also encloses the hypocenter. We recognize that each measuredQ−1c is an aver- age estimate Q−1av (or Q−1apparent) for the volume as sampled by the ellipsoidal shell given by VTOTAL Qav � X j VBlock-j Qj ; (2) where VBlock-j is the fraction of volume (block) sampled by the ellipsoidal shell with the true attenuation coefficient Q−1j (or Q−1true). Assuming a constant S-wave velocity of propaga- tion, the volume traveled by a ray that leaves the hypocenter moves outward to the ellipsoidal shell as defined by the observation time of the coda and is scattered to the receiver, can be determined. Equation (2) can be written as Table 1 Estimated Values of Coda-Wave Attenuation (Q−1c ) at Various Frequencies Q−1c × 103 Coefficient of Correlation rms Signal/Noise Frequency Waveforms Analyzed Min Mean±Std Max Min Mean±Std Max Min Mean±Std Max Min Mean±Std Max 2 9338 1.7 7.1±2.7 47.6 −0.97 −0.67±0.11 −0.5 0.07 0.24±0.07 1.79 2.0 16.2±47.5 985.5 8 5421 0.8 1.8±0.7 20.4 −0.96 −0.60±0.08 −0.5 0.18 0.32±0.06 2.67 2.0 11.0±25.6 842.6 12 4441 0.6 1.2±0.4 1.2 −0.94 −0.59±0.08 −0.5 0.15 0.35±0.06 1.89 2.0 9.7±21.1 600.9 16 3741 0.5 0.9±0.3 9.4 −0.95 −0.58±0.07 −0.5 0.17 0.35±0.09 4.13 2.0 9.0±18.1 315.7 Also presented are extreme values, averages, standard deviations, and quality parameters. Q−1c values at 2 Hz were used to estimate tomograms because of the high availability of observations regarding other frequencies, best values of correlation coefficients, rms, and signal-to-noise ratio. A power law equation for all Q−1c observations suggested a high-frequency dependence of the attenuation in this region: Q−1c �f� � �13:2� 0:6� × 103f−�0:97�0:06�. Tearing and Breaking Off of Subducted Slabs as the Result of Collision of the Panama Arc-Indenter 2029 1 Qav � 1 Q1 VBlock-1 VTOTAL � � � � � 1 Q1 VBlock-1 VTOTAL � � � � � 1 Qn VBlock-n VTOTAL ; (3) where the ratio VBlock-j=VTOTAL is the volume fraction associated with the total scattered-wave travel path spent in the jth block. If the process is repeated for each station– hypocentral pair, the entire region is sampled. Equation (3) is of the form a1x1 � � � � � aixi � � � � � anxn � y; (4) where y � � 1 Qav � xi � � 1 Qi � ai � � VBlock-i VTOTAL � : Then, a least-squares estimation of the xi is given by the compact matrix equation AX � Y where A is a (k × n) coefficient matrix, X is a (n × 1) vector, Y is a (k × 1) vector, and k is the number of station–hypocenter pairs. A linear inversion of the matrix equation was formulated as an iter- atively damped least-squares technique (Levenberg, 1944; Marquardt, 1963). The damping factor (σ), which adds to the diagonal parameters of the matrix, was computed -1 -0.9 -0.8 -0.7 -0.6 -0.50 200 400 600 800 1000 1200 1400 1600 Fr eq ue nc y µ=-0.67 σ=0.11 Coefficient of correlation 0 100 200 300 400 500 600 700 800 9000 500 1000 1500 2000 2500 3000 3500 4000 4500 Fr eq ue nc y µ=155.0 σ=96.0 Epicentral distance 0 50 100 150 200 250 0 500 1000 1500 2000 2500 3000 3500 4000 Fr eq ue nc y µ=87.7 σ=65.6 Depth (km) 0 1 2 3 4 5 6 7 0 500 1000 1500 2000 2500 3000 3500 Fr e qu en cy µ=3.1 σ=0.7 (a)(a) (b)(b) 0 10 20 30 40 50 0 500 1000 1500 2000 2500 3000 3500 4000 Fr e qu en cy σ Figure 3. (a) Example of waveform used for estimating the Qc values. Upper trace represents the original record of an earthquake recorded by a short-period seismological station of the CNSN. Middle trace represents the filtered record in frequency band 1–3 (2� 1) Hz. Lower trace represents the decay envelope of the coda wave in a window of 20 s, starting from 2 × tβ s. Qc value was obtained as the slope of the least-squares fit of Ln�t2 × P�ω; t�� versus ωt (dashed line with arrowheads), for t > tβ, where tβ represents the S-wave travel time (Haskov et al., 1989). (b) Histograms for Q−1c values and their correlation coefficients, as well as distributions for the epicentral distances, focal depths, and local magnitudes of all events analyzed. 2030 C. A. Vargas and P. Mann Ta bl e 2 Se is m ol og ic al St at io ns of th e C N SN U se d in th is St ud y St at io n L on gi tu de (° ) L at itu de (° ) A lti tu de (m as l) W av ef or m s A na ly ze d Q − 1 c × 1 0 3 t c C oe ff ic ie nt s of C or re la tio n E pi ce nt ra l D is ta nc es M in M ea n± St d M ax M in M ea n± St d M ax M in M ea n± St d M ax M in M ea n± St d M ax A N IL − 75 .4 0 4. 49 23 00 24 8 2. 9 6. 8± 1. 6 17 .2 12 .1 89 .3 ± 51 .5 23 1 − 0. 94 − 0. 67 ± 0. 11 − 0. 50 26 .7 16 6. 3± 90 .2 41 3. 6 B A R − 73 .1 8 6. 58 18 64 75 4 1. 7 6. 1± 2. 5 32 .3 11 .8 71 .2 ± 19 .2 22 7 − 0. 94 0. 67 ± 0. 11 − 0. 50 26 .6 14 2. 0± 33 .7 41 4. 6 B C IP − 79 .8 4 9. 17 61 5 5. 5 6. 9± 2. 0 12 .2 10 4. 7 12 6. 4± 15 .7 14 6 − 0. 86 − 0. 69 ± 0. 10 − 0. 59 20 0. 7 23 8. 7± 27 .5 27 2. 8 B E T − 75 .4 4 2. 68 54 0 66 2. 9 6. 6± 3. 0 14 .9 16 .7 68 .1 ± 56 .3 24 6 − 0. 91 − 0. 68 ± 0. 11 − 0. 50 46 .7 13 6. 8± 98 .5 44 8. 0 B R I − 72 .7 9 7. 72 14 27 46 2. 9 6. 0± 3. 5 47 .6 13 .1 81 .6 ± 36 .7 15 0 − 0. 97 − 0. 66 ± 0. 13 − 0. 50 24 .7 15 9. 6± 65 .8 28 0. 2 C H I − 73 .7 3 4. 63 31 40 72 4 2. 7 6. 8± 3. 0 20 .0 11 89 .4 ± 53 .3 26 3 − 0. 97 − 0. 69 ± 0. 11 − 0. 50 24 .5 17 3. 5± 93 .8 47 7. 6 C L IM − 77 .8 9 0. 94 42 32 17 2. 5 7. 0± 5. 0 12 .0 31 .7 78 .9 ± 35 .5 13 2 − 0. 91 − 0. 69 ± 0. 11 − 0. 53 73 .0 15 5. 5± 62 .1 24 7. 6 C O D − 73 .4 4 9. 94 10 8 10 4 3. 4 7. 3± 3. 1 18 .2 19 .3 47 .3 ± 48 .7 18 0 − 0. 95 − 0. 70 ± 0. 12 − 0. 50 33 .8 10 0. 2± 85 .3 33 2. 2 C PA S − 77 .2 5 1. 22 26 20 8 5. 1 8. 8± 4. 6 15 .9 16 .6 14 .1 ± 10 .2 38 .4 − 0. 91 − 0. 72 ± 0. 12 − 0. 56 29 .1 42 .2 ± 17 .8 84 .7 C R U − 76 .9 5 1. 57 27 61 33 0 2. 5 6. 7± 2. 9 16 .1 17 .6 79 .2 ± 60 .2 33 9 − 0. 95 − 0. 69 ± 0. 11 − 0. 50 30 .8 15 6. 0± 10 5. 3 61 0. 4 C TA B − 74 .2 5. 01 35 00 2 5. 6 8. 1± 8. 2 14 .5 13 2. 4 13 3± 0. 85 13 4 − 0. 86 − 0. 70 ± 0. 22 − 0. 55 24 9. 2 25 0. 2± 1. 48 25 1. 3 C TA U − 74 .0 4 5. 20 38 68 5 4. 3 8. 0± 6. 7 28 .6 52 .8 96 .2 ± 31 .2 11 9. 1 − 0. 9 − 0. 73 ± 0. 12 − 0. 6 10 9. 9 18 5. 9± 54 .5 22 5. 9 C U M − 77 .8 3 0. 94 34 20 23 4 3. 0 7. 0± 3. 2 22 .7 12 .9 82 .7 ± 57 .9 38 4 − 0. 97 − 0. 69 ± 0. 12 − 0. 50 22 .6 16 2. 3± 10 1. 3 69 0. 0 G C A L − 77 .4 2 1. 21 23 53 8 7. 9 10 .3 ± 1. 8 13 .5 14 .6 22 .8 ± 21 .1 2 69 .4 − 0. 91 − 0. 83 ± 0. 05 − 0. 74 25 .6 57 .4 ± 37 .0 13 9. 0 G C U F − 77 .3 5 1. 23 38 00 56 3. 7 7. 4± 2. 7 18 .2 22 .0 52 .3 ± 40 .1 15 2 − 0. 93 − 0. 70 ± 0. 11 − 0. 50 38 .5 10 8. 1± 70 .5 28 3. 5 G U A − 72 .6 3 2. 54 21 7 11 3. 6 6. 2± 2. 9 10 .2 20 .3 17 4. 4± 85 .5 22 7 − 0. 86 − 0. 67 ± 0. 11 − 0. 51 53 .0 32 2. 7± 14 9. 7 41 5. 3 H E L − 75 .5 3 6. 19 28 15 21 6 3. 0 6. 1± 2. 1 19 .6 21 .0 11 0. 2± 40 .9 19 0 − 0. 96 − 0. 65 ± 0. 10 − 0. 50 36 .8 21 0. 0± 71 .5 35 0. 2 M A L − 77 .3 4 4. 01 75 39 1 2. 4 6. 2± 0. 0 15 .9 14 .9 65 .1 ± 37 .5 32 3 − 0. 95 − 0. 67 ± 0. 11 − 0. 50 43 .6 13 1. 4± 65 .7 58 2. 4 M A R A − 75 .9 5 2. 84 22 07 78 3. 0 5. 9± 2. 4 18 .2 17 .9 87 .6 ± 65 .2 25 9 − 0. 92 − 0. 65 ± 0. 11 − 0. 51 31 .3 17 0. 8± 11 4. 1 46 9. 9 N O R − 74 .8 7 5. 57 53 6 59 6 2. 5 5. 9± 2. 1 20 .4 16 .3 11 3. 4± 35 .5 29 7 − 0. 94 − 0. 66 ± 0. 11 − 0. 50 28 .5 21 5. 8± 62 .1 53 7. 4 O C A − 73 .3 2 8. 24 12 64 33 04 2. 0 6. 0± 2. 3 21 .7 16 .4 10 2. 8± 23 .1 28 7 − 0. 96 − 0. 66 ± 0. 11 − 0. 50 28 .7 19 7. 4± 40 .4 52 0. 3 O TA V − 78 .4 5 0. 24 34 92 26 4. 3 7. 6± 2. 1 13 .9 23 .6 70 .2 ± 25 .7 14 7 − 0. 88 − 0. 73 ± 0. 12 − 0. 53 58 .8 14 0. 3± 44 .9 27 4. 1 PC O N − 76 .4 2. 33 42 94 12 0 2. 0 6. 6± 3. 2 16 .9 12 .5 72 .4 ± 52 .8 28 9 − 0. 96 − 0. 67 ± 0. 11 − 0. 50 39 .4 14 4. 1± 92 .4 52 2. 6 PR A − 74 .8 9 3. 71 46 8 31 3 2. 5 6. 1± 2. 7 19 .6 17 .6 10 1. 1± 61 .3 25 9 − 0. 95 − 0. 67 ± 0. 11 − 0. 50 30 .8 19 4. 5± 10 7. 4 47 0. 1 R O SC − 74 .3 3 4. 86 30 20 14 9 2. 9 5. 9± 2. 4 33 .3 17 .9 88 .2 ± 47 .6 20 4 − 0. 95 − 0. 66 ± 0. 12 − 0. 50 31 .3 17 1. 8± 83 .3 37 3. 6 R R E F − 75 .3 5 4. 9 47 43 20 5 2. 5 6. 1± 2. 3 12 .7 11 3. 7 11 8. 1± 48 .8 22 0 − 0. 92 − 0. 65 ± 0. 10 − 0. 50 41 .5 22 4. 1± 85 .4 40 2. 7 R U S − 73 .0 8 5. 89 36 97 15 0 2. 6 6. 0± 2. 7 20 .4 17 .5 72 .6 ± 39 .9 24 0 − 0. 94 − 0. 66 ± 0. 11 − 0. 50 30 .6 14 4. 5± 69 .7 43 7. 9 SD V − 70 .6 3 8. 88 16 20 64 2. 8 5. 8± 2. 0 12 .5 11 4. 5 17 6. 3± 23 .7 26 2 − 0. 90 − 0. 65 ± 0. 11 − 0. 50 21 7. 9 32 6. 0± 41 .4 47 6. 5 SO L − 77 .4 1 6. 23 38 85 0 3. 1 7. 3± 3. 1 25 .6 17 .2 64 .6 ± 39 .8 30 5 − 0. 97 − 0. 71 ± 0. 11 − 0. 50 30 .1 13 0. 6± 69 .6 55 1. 1 T O L − 75 .3 2 4. 59 25 77 25 8 2. 5 6. 3± 2. 6 20 .0 19 .8 10 0. 4± 56 .1 24 8 − 0. 95 − 0. 67 ± 0. 11 − 0. 50 34 .7 19 3. 1± 98 .2 45 1. 9 T he st at io ns de te ct ed 76 45 ea rt hq ua ke s (1 99 3– 20 12 ) th at w er e us ed fo r es tim at in g 93 38 Q − 1 c va lu es in fr eq ue nc y ba nd 1– 3 (2 � 1 ) H z an d co da -w av e tim e w in do w (W ) of 20 s. t c va lu es an d re la tiv el y la rg e ep ic en tr al di st an ce s al lo w ed us to es tim at e th e co da -w av e to m og ra ph y re gi on al ly . Tearing and Breaking Off of Subducted Slabs as the Result of Collision of the Panama Arc-Indenter 2031 automatically for each iteration (Hoerl and Kennard, 1970; Hoerl et al., 1975). According to this technique, the solution and resolution matrixes can be found for the following equations: X � �ATA� σ2I�−1ATY; (5) and R � �ATA� σ2I�−1ATA: (6) Similar procedures for the Q−1c imaging have been used in previous works in order to establish a deterministic char- acterization of the heterogeneity in the lithosphere as an alternative technique for traditional tomographic measure- ments (O’Doherty et al., 1997; Lacruz et al., 2009). Resolution and Reliability A spatial inversion of attenuation of 32 × 32 × 8 blocks with block dimensions of ∼60 km �latitude�× ∼50 km �longitude� × ∼40 km �thickness� was designed in order to detect relevant structures in the region. We qualified the tomographic inversion by means of three approaches: (a) hit count of ellipsoidal shells; (b) solving controlled tests; and (c) mapping the diagonal elements of the resolution ma- trix (RDE) by using equation (6). The hit count is a very rough quality estimation that only tells about summing up the number of ellipsoids that contribute to the solution of a block. Based on this discretized volume, we mapped the hit count with the available data in eight layers (0, 45, 90, 135, 180, 225, 270, and 315 km; Fig. 4a). Although a large part of northwestern South America (including Colombia, western Venezuela, eastern Panama, and northern Ecuador) is covered by ellipsoidal shells (over 500 crossings), it is in northern Colombia and northwestern Venezuela (71° W to 76° W; 5° N to 10° N; 0–180 km depth) where the largest number of shells run through the blocks (based on more than 5000 hits per block). This approach emphasizes the impor- tance of the Bucaramanga nest data in the estimation of to- mographic images. To incorporate the second approach, we evaluated the efficiency of the method described above solving the 3D di- rect problem. The ellipsoidal shell volume associated with all foci (pairs of earthquake and station) were used to relate the Q−1c values in two controlled test boards. The first one was for appraising large domains of attenuation, for example, a zone with flat subduction in the north, and the other zone with normal subduction in the south (Fig. 4b). The second test board evaluated the ability of the method to detect small anomalies by use of a typical chessboard (Fig. 4c). In tests, we assigned two values of Q−1true that represent attenuation contrasts (1=70 and 1=200) into the 3D grid. Then we esti- mated the Q−1av values (or theoretical Q−1c values) for all el- lipsoids (each one related to foci [earthquake–station]) by estimating the weighted average of Q−1true involved in the vol- ume of each ellipsoid. For the 3D inverse problem, we estimated the fraction of volumes associated with each Q−1av in order to establish equation (3). Using all foci related to the events selected in this study, we assembled the compact matrix of equation (4) and then we inverted the Q−1av values using equation (5). Finally, a spatial interpolation of the Q−1av values was done based on the Kriging method (Oliver and Webster, 1990) and presented on Mercator projection. Figure 4d shows the results of the inversion for the synthetic experiment based on two domains of contrasting attenuation (Fig. 4b). This experiment is comparable to a slab-tearing model for which two zones with different angles of subduction, are related to different attenuations. This hypothetical model linked a flat subduction zone in the north (lower attenuation) and a nor- mal subduction zone in the south (higher attenuation). In general, the available data may allow detection of large struc- tures with significant contrasts of attenuation as much as ∼270 km depth. On the other hand, Figure 4e presents the inversion for a chessboard based on two areas of contrast- ing attenuation (Fig. 4c). This experiment suggests that the available data may allow detection of smaller bodies (e.g., 100 km × 100 km × 60 km) with significant contrasts of attenuation, mainly in Colombia, and as much as ∼180 km depth. After several trials of accurate resolution and sta- bility, the spatial inversion of attenuation with real data was performed with the same grid (32 × 32 × 8 blocks). Figure 4f,g shows results of the tomographic estimation and their maps of the RDE at different depths. Because each RDE shows the amount of independence in the solution of one model parameter (RDE oscillates between 0 and 1), the larger value of the RDE for one model parameter suggests a more independent solution for this parameter. 3D inversion presents higher RDE values (e.g., >0:4) limited by the avail- ability and geographical concentration of Q−1c values, indi- cating that the method is useful for areas for which a large stacking of attenuation observations is present. From the available earthquake data, the tomographic solution of the attenuation efficiently images large areas of the crust and upper mantle of northwestern South America including Colombia, eastern Panama, and western Venezuela with sampling depths reaching >315 km (Fig. 4f). Because the ellipsoids related to deeper hypocentral solutions can sample profounder volumes, the 3D inversion may detect the thermal influx from the mantle adequately. In order to infer the geometry of the Caldas tear and its relationships with the adjacent Nazca and Caribbean plates, we made two regional cross sections: (1) a northern sec- tion (AA0, Fig. 1a) from the Caribbean plate to the Llanos foreland basin of eastern Colombia and crossing the inter- mediate-depth earthquakes of the Bucaramanga nest; (2) a southern section designed for imaging the corridor between the Nazca plate and the Llanos basin and crossing the inter- mediate-depth earthquakes of the Cauca nest (BB0, Fig. 1a). As discussed subsequently, it is essential to incorporate all 2032 C. A. Vargas and P. Mann available geophysical data for proper interpretation of the tomograms along these sections. Integrating Earthquake Data with Regional Seismic-Reflection Lines Hypocentral solutions of the CNSN (rms < 0:3 s; GAP < 200; stations ≥6; error in latitude ≤10:0 km; error in longitude≤10:0 km; error in depth≤5:0 km)were plotted on the tomographic profiles along two 60 km wide corridors (Figs. 5 and 6). Because seismicity in a corridor parallel to the northern section is sparse, we have included an interpreta- tion of theTrans-Andeanmegaregional seismic-reflection line that extends from the Caribbean coast to the Eastern Cordil- lera of Colombia (Vargas et al., 2010) and to the northern to- mographic section. This 383 km long reflection line is a 20 s Figure 4. Resolution, reliability, and results of the spatial inversion of attenuation based on a geometry of 32 × 32 × 8 blocks with dimensions of ∼60 km �latitude� × ∼50 km �longitude� × ∼40 km �thickness�. Coda-wave tomograms were estimated with 9338 Q−1c ob- servations associated with 7645 regional earthquakes (mL ≤ 6:7; 1993–2012) in the frequency band 1–3 (2� 1) Hz. (a) Hit count of ellip- soidal shells, suggesting that the 3D inversion of Q−1c may solve large areas of Colombia, eastern Panama, western Venezuela, and northern Ecuador. (b) Synthetic model that represents two large domains of attenuation (e.g., a zone with flat subduction in the north and normal subduction in the south, limited by a slab tear). The contrasts of attenuation incorporated into the model to evaluate the effectiveness of the method wereQ−1c � 1=200 andQ−1c � 1=70. (c) Chessboard with smaller and regular distribution of attenuation contrasts. As in the previous case, the contrasts of attenuation incorporated into the model were Q−1c � 1=200 and Qc−1 � 1=70. (d) 3D inversion of the synthetic model presented in (b) suggesting that the distribution of the available data may allow detection of large structures with significant contrasts of attenuation as much as ∼270 km depth. (e) 3D inversion of the chessboard model presented in (c) suggesting that the distribution of the available data may permit detection of smaller bodies (e.g., ∼100 km × 100 km × 60 km) with significant contrasts of attenuation, mainly in Colombia, and as much as ∼180 km depth. (f) Results of the tomographic inversion with the available data. (g) Maps of the RDE at different depths. Higher RDE values (e.g., ≥0:5) indicate zones efficiently solved. However, these higher RDE values were limited by the geographical concentration of Q−1c values, indicating that the method is useful for areas where a large stacking of attenuation observations is present. Tomographic solution of the attenuation efficiently images large areas of the crust and upper mantle of northwestern South America including Colombia, eastern Panama, and western Venezuela with sampling depths reaching>315 km. High-attenuation anomalies suggest that Buca- ramanga and Cauca seismic nests may be related to asthenospheric emplacements. (Continued) Tearing and Breaking Off of Subducted Slabs as the Result of Collision of the Panama Arc-Indenter 2033 Figure 4. Continued. 2034 C. A. Vargas and P. Mann Figure 4. Continued. Tearing and Breaking Off of Subducted Slabs as the Result of Collision of the Panama Arc-Indenter 2035 record and 200-fold, and shows the subduction geometry of northern (Bucaramanga) slab dipping at a shallow angle to the southeast beneath northwestern Colombia (Fig. 7). Although reflectors are difficult to distinguish in deeper areas of the line, the overall distribution of deep reflectors dips east- ward in same amount as the subducted slab on the gravity and magnetic model in Figure 5c. Deep reflections are concen- trated within what we interpret as the lower crust whereas the uppermantle appears more transparent (Tittgemeyer et al., 1999). Prominent reflections in the upper crust can be corre- lated with major sedimentary basins such as the Sinu–San Ja- cinto, Lower Magdalena, Middle Magdalena, and Eastern Cordillera, as well as major faults such as the Romeral fault zone (RFZ). This major fault separates oceanic crustal rocks in thewestern terranes of Colombia and continental basement in eastern Colombia (Cediel et al., 2003). In general, seismicity east of the Romeral is more concentrated in the older and more anisotropic continental crust of northwestern South America. Gravity and Magnetic Modeling Coincident gravity andmagnetic models were completed for this study based on regional information (Maus et al., 2007; Sandwell and Smith, 2009; National Hydrocarbons Agency of Colombia, 2010; Figs. 5b and 6b). The gravity and magnetic data was merged with the 90 m elevation topo- graphic information available from NASA (Jarvis et al., 2008), with corrections from the International Gravity Stand- ardization Net 1971 (IGSN71), the World Geodetic System 1984 (WGS-84), the International Geomagnetic Reference Field (IGRF) and the Observed Magnetic Intensity (Hinze et al., 2005; Maus et al., 2005). The final database allowed us to estimate free air and magnetic anomalies. We then cal- culated Bouguer anomalies using densities of 2:67 g=cm3 for Figure 5. Section crossing the northern Panama-arc indenter and its down-dip Bucaramanga nest (Fig. 1a, AA0). (a) Geologic and geo- thermal observations; (b) gravity and magnetic data; (c) interpreted tomographic section. Green dots with vertical bars that represent vertical errors, hypocentral solutions in a 60 km wide corridor. Plotted events have the following selection criteria: rms < 0:3 s; GAP ≤ 200; stations ≥6; error in latitude ≤10:0 km; error in longitude ≤10:0 km; error in depth ≤5:0 km. Some representative focal mechanisms (beach balls) have been also plotted. 2036 C. A. Vargas and P. Mann land and 1:03 g=cm3 for marine water. In order to estimate gravity and magnetic model responses (Talwani et al., 1959; Talwani and Heirtzler, 1964; Geosoft, 2010) and com- paringwith the observed data, we used values of density,mag- netization, and magnetic susceptibility shown on Table 3. In addition, the gravity and magnetic models were constrained with the seismological and seismic data, as well as geologic transects compiled with superficial cartography and represen- tative seismic lines (Section AA0, Figs. 1a and 5; and Section BB0, Figs. 1a and 6; Lopez, 2004). Because of restrictions on the data use of the Trans- Andean megaregional seismic-reflection line, it was not possible to estimate refraction travel-time tomography for correlating with the profiles presented in this paper. However, in order to interpret the thermal and tectonic structure in this region, we used the velocity anomalies reported by Vargas et al. (2007) and van der Hilst and Mann (1994) that show similar distribution anomalies. In general, trends of high velocities match with slabs suggested by the gravity and magnetic models. Other Geophysical Information for Constraining the Interpretation Seismic attenuation has been used for examining the acoustic contrast between the upper mantle and the litho- sphere because it is believed that this seismic factor is a physical parameter closely related to the thermal state of the volume sampled by the waves (Faul and Jackson, 2005; Priestley and McKenzie, 2006; Yang et al., 2007). Therefore, in order to evaluate the empirical superficial response of the lithospheric thermal field, we plotted geothermal anomalies reported from oil wells in Colombia (Vargas et al., 2009) onto the two sections (Figs. 5a and 6a). Furthermore, topo- graphic response and focal mechanisms compiled from NEIC are plotted on the two profiles. Figure 6. Coda-wave-attenuation section crossing the southern part of the Panama indenter and the Cauca nest (Fig. 1b, BB0). (a) Geo- logic and geothermal observations; (b) gravity and magnetic data; (c) interpreted section. Green dots with vertical bars that represent vertical errors, hypocentral solutions in a 60 km wide corridor. Plotted events have the following selection criteria: rms < 0:3 s; GAP ≤ 200; stations ≥6; error in latitude ≤10:0 km; error in longitude ≤10:0 km; error in depth ≤5:0 km. Tearing and Breaking Off of Subducted Slabs as the Result of Collision of the Panama Arc-Indenter 2037 To support our interpretation, we have presented a total of eight variables along the profiles: hypocenter solutions, focal mechanisms, coda-wave attenuation, gravity, magnetic and geothermal anomalies, and geologic and topographic data derived from seismic-reflection profiles. The two sec- tions flank the intersection of the Panama arc-indenter and the Caldas tear to the north (Section AA0, Figs. 1a and 5) and south (Section BB0, Figs. 1a and 6). The profiles are ex- tended to the west from the Caribbean Sea and Pacific Ocean to the Llanos foreland basin in the east. Both images cross Figure 7. (a) Deep seismic profile nearly parallels the tomographic section AA0 (see inset map). Seismic image was assembled with three segments of the Trans-Andean Seismic Line acquired by the National Hydrocarbons Agency of Colombia (ANH; Vargas et al., 2010). (b) An interpretation of the seismic line. Black lines, suggested sedimentary basins and deeper reflections; blue lines, faults and main tectonic features (e.g., the Romeral fault zone in bold line); yellow line, suggested detachment surface associated with the Caribbean plate subduction. Red dots with vertical bars that represent vertical errors, hypocenter solutions in a 60 km wide corridor. Plotted events have the following selection criteria: rms < 0:3 s; GAP ≤ 200; stations ≥ 6; error in latitude ≤10:0 km; error in longitude ≤10:0 km; error in depth ≤5:0 km. Depth–time relation has been estimated by using several oil wells in the area (Vargas et al., 2010). Table 3 Physical Properties Expressed in SI for the Materials Used in Gravity and Magnetic Models Unit Density (kg=m3) Susceptibility Magnetization (A=m) Caribbean and Nazca plates 3:30 × 103 1:26 × 10−1–1:38 × 100 1:00 × 10−3–1:45 × 100 Lower continental crust 2:78 × 103 1:26 × 10−5 1:26 × 10−5–1:90 × 100 Upper continental crust 2:31 × 103 3:21 × 10−2–4:90 × 10−2 5:51 × 10−1–1:00 × 100 Mantle 3:15 × 103 Accreted sediments and oceanic crust 2:07 × 103–2:35 × 103 1:26 × 10−1 1:00 × 10−3–8:01 × 10−1 Oceanic sediments 1:68 × 103–2:00 × 103 1:26 × 10−5 1:00 × 10−3–4:50 × 10−1 Water 1:03 × 103 2038 C. A. Vargas and P. Mann the northern Andes, the Bucaramanga and Cauca seismic nests of intermediate depth earthquakes, and major faults including the Romeral, the Llanos, the Santa Marta– Bucaramanga, the Garrapatas, and the Ibague faults. Results and Discussion Northern Transect Crossing the Caribbean Plate and the Bucaramanga Nest Our northern transect shows the presence of an oceanic crust with thicknesses greater than 20 km related to the Car- ibbean oceanic plateau with a shallow subduction angle of of events. Although lineal regression of the temporal series of events is of low confidence, this unidirectional, westward displacement of events may be caused by down-dip and southwestward propagation of tearing of the subducted Caribbean slab. Southern Tomographic Transect Crossing the Nazca Plate and the Cauca Nest The Nazca oceanic slab has been modeled with an ∼15–22 km deep crustal thickness and a constant dip angle of 30°–40° to a depth >150 km beneath the active volcanic line (Figs. 1a and 6). The volcanic line is underlain by high-attenuation anomalies indicative of a normal melting range for the subducted oceanic slab (Figs. 2, 4c, 6). High- attenuation anomalies and the presence of shallow to inter- mediate seismicity around the Romeral fault zone suggest this major strike-slip provides another major upward conduit for the release of upper mantle heat. A large low-attenuation anomaly corresponds to the low geothermal gradient ob- served between the Colombian trench along the Pacific mar- gin and the Central Cordillera. The low geothermal gradient coincides with thick volcanic and sedimentary material ac- creted to western Colombia, mainly in the Western Cordillera and the Baudo Range. The accretion of this area may have accompanied a proposed westward jump in subduction from the Romeral fault zone in the Central Cordillera to the present Colombia trench (Cediel et al., 2003). An additional attenuation anomaly observed on tomographic data indicates a prominent high thermal inflow from the mantle that is located beneath the forebulge of the Llanos basin and is ap- proximately coincident with the largest geothermal anomaly measured in this basin. Seismicity of the Cauca nest is highly dispersed in depth (70–150 km) and its time evolution seems more complex than the Bucaramanga nest (Fig. 9). Even with the low con- fidence of the lineal regression, seismicity from the CNSC catalog shows an eastward displacement of events. The Cauca nest exhibits earthquakes with focal mechanisms ranging from pure gravitational collapse to strike-slip in the north and reverse with strike-slip component in the south (Fig. 2). Nature of the Caldas Tear In addition to the hypocenter solutions that initially revealed the ∼240 km long offset of intermediate to deep seismicity of the Caldas tear, we found that differential displacements between the southern of the Caldas tear in the South American plate and the Panama arc-indenter, based on GPS observations, suggest a trend of decreasing displacement 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 -77 -76.5 -76 -75.5 Lo ng itu de (°) 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 3.5 4 4.5 5 (a) (b) (c) 5.5 La tit ud e(° ) 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 -200 -150 -100 -50 0 D ep th (km ) Time (year) y(x) = a (x - b) a = 0.021514 b = 5553.5 R = 0.40882 y(x) = a x + b a = 0.0085141 b = -12.525 R = 0.12753 y(x) = a x + b a = 1.0488 b = -2211.1 R = 0.19102 Figure 9. Temporal evolution of the hypocentral parameters of the Cauca nest. Earthquake information provided by the CNSN from 1993 to 2012. The Cauca nest events were selected around the point 76.3° W, 4.5° N. Parameters of selection were Radii � 0:5°; 0 ≤ rms ≤ 0:3 s; h ≥ 60 km; GAP ≤ 200; (a) error in longitude ≤10:0 km; (b) error in latitude ≤10:0 km; (c) error in depth ≤10:0 km. Error bars have been associated with each event. Dashed polygon, the linear trend of trend of occurrences estimated by least-squares means. 2040 C. A. Vargas and P. Mann toward the east (Trenkamp et al., 2002; e.g., BOGO versus MZAL, RION, BUCM,MONT, and CART GPS stations). As- suming the BOGO station as the reference point south of the Caldas tear, and the CHEP station as the reference point on the Panama indenter, we estimate ∼24 mm=year of active right- lateral displacement across the Caldas tear. The hypothesis of lateral homogeneity of the crust, constant displacement rate, and a seismic offset along Caldas tear of ∼240 km, would suggest an ∼10:0 Ma initiation of the Panama arc–Colombia collision (∼240 km=24 mm=year). Geologic field observa- tions in the Panama-arc (Coates et al., 2004) suggest that the age for the initiation of the Panama arc collision with northern South America occurred between 12.8 and 7.1 Ma, which is consistent with our estimated ∼10:0 Ma initiation of the tear propagation. In addition, because the origin of the east–west Sandra ridge occurs between 9 and 12Ma (Defant et al., 1992; Lonsdale, 2005), and this structure is collinear with the Caldas tear, we propose that the right-lateral lineament defined by the Caldas tear and the Sandra ridge, constitutes a major area of lithospheric weakness along the southern flank of the Panama arc-indenter. Although there is no evidence for recent activity of the Sandra ridge due to lack of near-bottom instrumentation in the Pacific offshore of Colombia and Panama, recent earth- quakes and the adakite magmatism along the Caldas tear may indicate that this lineament localizes upper crustal fault con- duits that allowed the upward migration of magmatic fluids and are associated with elevated geothermal anomalies. Offset of intermediate- to deep-seismicity that defines the Caldas tear is also coincident with inactive volcanoes of adakite compo- sition and geothermal anomalies (Fig. 2). The adakites of the Ruiz volcanic complex with ages ranging from 0.97–0.05 and 1.8–0.6 Ma (Borrero et al., 2009) and the Paipa–Iza volcanic complex dated 1.9–2.5 Ma (Pardo et al., 2005) are a likely consequence of magmatism related to this progressive slab tear. Low ratio 87Sr=87Sr (0.705) and the presence of xenoliths of metamorphic rocks in this last volcanic complex (J. M. Jaramillo, personal comm., 2012) support the proposed break-off interpretation south of the Bucaramanga nest and east of the Caldas tear. Surficial evidence of this lithospheric tear are restricted to presence of mineral deposits, hydrocarbon occurrences, and some geomorphological anomalies. High-grade mineral deposits of platinum, gold, and copper exploited in the min- ing areas called Condoto (Tistl, 1994), Marmato (Ordoñez, 2001), Quinchia (INGEOMINAS, 1999), La Colosa (Gil- Rodriguez, 2010), and Cerro de Cobre (McLaughlin and Arce, 1970) are near, or collinear with, the Caldas tear and exhibit ages ranging between 6 and 20 Ma (see blue hexa- gons on map in Fig. 10a). In addition, significant changes in distribution and trend of oil and gas seeps, as well as the hydrocarbon fields on both sides of the Caldas tear, suggest that this structure may also affect the geometry of several sedimentary basins (e.g., Llanos foreland, Eastern Cordillera, Middle Magdalena Valley). But likely the most prominent geomorphological evidence is coming from hydraulic behav- ior of the main rivers that cross the south-to-north-flowing Magdalena and Cauca rivers of northwestern South America. After flowing 200 km from their sources, these rivers occupy broad river valleys. Downstream, rivers passing the Caldas tear lineament change their morphology from broad valleys to steeper relief gorges near the surface projection of the Caldas tear. The Cauca River near Supia town (Fig. 10a,b) reduces to a narrow channel only 150 m wide whereas the Magdalena River near the town of Honda reduces to a 250 m wide channel (Fig. 10a,c). In contrast to the broader, 42 km wide valleys observed upstream (e.g., Bolivar and Guamo towns in Fig. 10a,d), in both areas the steeper gradient and more narrow rivers produce rapids that are an impediment to navigation. The narrowness of these rivers favored the down- stream economic development of urban settlements such as Honda and Supia from Spanish colonial times. However, the quake that occurred on 16 June 1805 that destroyed Honda and other nearby towns shows us that the Caldas tear could form a major source of earthquakes and seismic hazard in this region. The Caldas tear may localize large crustal earthquakes, including the recent strong-motion activity associated with the Tauramena earthquake (19 January 1995, Mw 6.5). Two types of focal mechanisms have been proposed for this event, one of which suggests an east–west-oriented right lateral movement (Dimate et al., 2003; Fig. 2). Similarly, seismic activity in the area of the Armenia earthquake (25 January 1999, Mw 6.2) shows east–west alignment of after- shocks with the Caldas tear. This long right Caldas tear does not explain the Quetame earthquake (24 May 2008, mL 5.7) with left-lateral slip, in which the Caldas tear could have played an important role for controlling the propagation of north–south-trending faults (such as this event or any re- lated to the Llanos fault system) in a similar way as has been suggested in the southwest Colombia margin where trans- verse faults reduce coupling between adjacent segments (Collot et al., 2004). We removed the crustal earthquakes from the database of events located by the CNSN between 1993 and 2012 in order to illustrate the upper surface of the subducted slabs beneath Colombia (Fig. 11a). The 3D model of this surface images the Caldas tear and flat-slab subduction geometry that we relate to the presence of the Panama arc-indenter (Ra- mos and Folguera, 2009). The northern border of the indenter becomes diffuse and does not appear to form a distinctive tear as seen south of the indenter (see dashed line in Fig. 11a). The Caribbean plate changes from flat to steep subduction toward the northeast. South of the Caldas tear (∼2:5° N), there is a shift to a new pattern of intermediate and deep seismicity associated with flat subduction along with the development of a broad area of active volcanoes in southern Colombia and northern Ecuador. Figure 11b presents a con- ceptual model that explains the kinematic role of the Panama arc-indenter whose southern boundary is defined by the Sandra ridge and the Caldas tear. In this model the coupling of the Panama arc with the Caribbean plate could generate a change in buoyancy of the lithospheric system and Tearing and Breaking Off of Subducted Slabs as the Result of Collision of the Panama Arc-Indenter 2041 consequently the northern region of indentation has condi- tions that favor flat subduction. In the east, the Caribbean plate suddenly changes its subduction angle and produces a break off of the slab around the location of the Bucara- manga nest. South of the Caldas tear, the Nazca plate is sub- ducting beneath the South American plate with a steeper angle and a faster rate. The Cauca nest is a combined product of eastward decoupling of plates along the Caldas tear and flexure during the subduction process. A corollary of our model for the Panama indenter and the formation of the Caldas tear is the eastward indentation of geologic features. Inspection of theMap of Quaternary faults and folds of Colombia (Paris et al., 2000) suggests that some branches of the Romeral fault zone south of the Caldas tear show right-lateral offset from parallel faults including the Palestina, Cimitarra, Mulato–Getudo, Honda, or Bituima faults to the north. Other faults with southwest–northeast trend, such as the Ibague and Garrapatas, could be part of a transverse strike-slip fault at the surface level overlying the Caldas tear (Fig. 1a). These results raise new questions about the regional evolution in northwestern South America. For example, the Great Arc of the Caribbean has been de- fined along the south Caribbean region but disappears once it enters the Guajira basin and the Santa Marta Massif north of Colombia (Ostios et al., 2005). As a consequence of the Panama-arc collision, it is possible that this regional feature Figure 10. Surficial evidences of the Caldas tear related to mineral deposits, hydrocarbon occurrences, and geomorphological anomalies. (a) Blue hexagons, map of distribution of high-grade mineral deposits of platinum, gold, and copper: (1) Condoto, (2) Marmato, (3) Quinchia, (4) La Colosa, and (5) Cerro de Cobre. Black dots, oil and gas seepages. Purple circles are giant hydrocarbon fields. Blue stars are other oil and gas fields. These hydrocarbon occurrences suggest that the Caldas tear also is affecting the geometrical configuration of several sedi- mentary basins. White stars, hydraulic anomalies of the Cauca and Magdalena rivers on the Caldas tear (rapids on the Supia and the Honda). White circle and square are places upstream the rivers where there are broad valleys (Bolivar and Guamo). (b) Rapids of the Cauca River near Supia town that overlies the Caldas tear. (c) Rapids of the Magdalena River near Honda town that overlies the Caldas tear. (d) Broad valley observed upstream of the Cauca River near Bolivar town. Similar landscape is observed in Guamo town where the valley width of the Magdalena River reaches >40 km wide. 2042 C. A. Vargas and P. Mann has been offset and displaced eastward by the Panama indenter. Conclusions • The eastward-directed collision of the buoyant Panama arc-indenter with northwestern South America produces a distinctive V-shaped pattern of crustal deformation and widens the northern Andes in Colombia and Venezuela (Fig. 1b). The Panama collision initiated ∼10 Ma and con- tinues to be active as shown by GPS data. • The southern edge of the Panama indenter is associated with the proposed Caldas slab tear at latitude ∼5:6° N. This tear extends for ∼240 km in an east–west direction and is collinear with the ∼9–12 Ma, now extinct, east–west- oriented Sandra oceanic spreading ridge on the unsubducted Figure 11. (a) Seismic surface estimated by interpolation and filtering of ∼68;000 local earthquakes (h ≥ 10:0 km). Blue lines, shore line of northwest South America. Bold black lines, limits of the convergent margins. Bold red lines, the southern border of the Panama indenter that includes the Sandra ridge and the Caldas tear. Bucaramanga nest is related to a break-off process that is propagating toward the south- west. Triangles, red (active) and green (inactive) volcanoes. Orange dashed lines, wireframe model suggested for indicating the subduction geometry of the Caribbean plate. (b) Schematic 3D model suggesting flat subduction on the northern side of the weakness zone formed by the Sandra ridge and the Caldas tear. Caribbean plate suddenly changes its subduction angle and promotes a break off of the slab around the location of the Bucaramanga nest. South of the weakness zone, the Nazca plate subducts beneath the South American plate with a steeper angle and faster displacement. Probably the Cauca nest is the combined product of eastward decoupling of plates along the Caldas tear as well as flexion and discrete movements of the plate during subduction. The Murindó earthquake nest located in proximity to the Panamanian and Colombian border, may be response to convergent accommodation between the Panama arc-indenter and the Caribbean plate. Tearing and Breaking Off of Subducted Slabs as the Result of Collision of the Panama Arc-Indenter 2043 oceanic Nazca plate to the west (Fig. 1b). We postulate that the Caldas tear may have formed as a zone of lithospheric weakness along the now subducted part of the inactive Sandra spreading ridge. • The ∼240 km long Caldas tear is a narrow, east–west- trending boundary between two subducted slabs of differ- ent dip. The northern zone is the down-dip extension of the Panama arc, has a shallower dip, and is not associated with active arc volcanism. The southern zone has a steeper dip and is associated with an active volcanic front (Fig. 11a,b). • The Caldas tear also localizes angular difference of the subduction geometry in both geophysical sections pre- sented in this work (Figs. 5 and 6). The lineament defined by the ∼240 km long offset of the deep seismicity along ∼5:6° N; the eruption of the north–south belt of active volcanism and presence of extinct magmatic bodies with adakite composition along the lineament; the occurrence of high-grade mineral deposits and geothermal gradient anomalies; the different patterns associated with oil and gas manifestations; the distribution of major oil and gas deposits north and south of the Caldas tear as well as the GPS measurements and strong-motion events with right- lateral movements, support the existence of the Caldas tear. Data and Resources Waveforms and preliminary hypocentral solutions of the Colombian territory were supplied by the Geological Sur- vey of Colombia (INGEOMINAS). Bouguer and magnetic data provided for the National Hydrocarbons Agency of Colombia (http://www.anh.gov.co/es/index.php?id=82, last accessed November 2012) were used for modeling geologic sections using the GM-SYS profile module of the Oasis Montaj software (Geosoft, 2010). This software calculates the gravity and magnetic model response based on the meth- ods of Talwani et al. (1959), and Talwani and Heirtzler (1964). GM-SYS uses a 2D, flat-earth model for the gravity and magnetic calculations. Each structural unit or block extends to plus and minus infinity in the direction perpen- dicular to the profile. The earth is assumed to have topogra- phy but no curvature. The model also extends plus and minus 30,000 km along the profile to eliminate edge effects. The 90 m elevation topographic information used for the gravity modeling is available from the CGIAR-CSI SRTM 90 m database (http://srtm.csi.cgiar.org, last accessed November 2012). Focal mechanisms reported by NEIC were used in this work (http://earthquake.usgs.gov/earthquakes/eqarchives/ sopar/, last accessed November 2012). Acknowledgments This work was partially funded by the industry sponsors of the CBTH project of the University of Houston and by fellowship support from the University of Texas at Austin. Earthquake, gravity, magnetic, seismic, and geothermal data were kindly provided by Agencia Nacional de Hidrocar- buros, Universidad Nacional de Colombia, INGEOMINAS, and the following research projects: 1233-333-18664, Contract 201-2006 (COLCIENCIAS); 1233-487-25728, Contract 589-2009 (COLCIENCIAS); CGL2005-04541- C03-02 and CGL2008-00869/BTE (UPC, MICCIN, FEDER). We also thank the Associate Editor Heather DeShon, and two anonymous reviewers for their helpful reviews of this paper. References Adamek, S., C. Frohlich, and W. Pennington (1988). Seismicity of the Caribbean-Nazca boundary: Constraints on microplate tectonics of the Panama region, J. Geophys. Res. 93, 2053–2075. Aki, K. (1969). Analysis of the seismic coda of local earthquakes as scattered waves, J. Geophys. Res. 74, 615–631. Aki, K., and B. Chouet (1975). Origin of coda waves: Source, attenuation and scattering effects, J. Geophys. Res. 80, 615–631. Badi, G., E. Del Pezzo, J. M. Ibanez, F. Bianco, N. Sabbione, and M. Araujo (2009). Depth dependent seismic scattering attenuation in the Nuevo Cuyo region (southern central Andes), Geophys. Res. Lett. 36, no. L24307, doi: 10.1029/2009GL041081. Borrero, C., L. M. Toro, M. Alvaran, and H. Castillo (2009). Geochemistry and tectonic controls of the effusive activity related with the ancestral Nevado del Ruiz volcano, Colombia, Geofísc. Int. 48, no. 1, 149–169. Calais, E., and P. Mann (2009). A combined GPS velocity field for the Caribbean plate and its margins (abstract #G33B-0657), American Geophysical Union, (Fall Meet.), G33B-0657. Cediel, F., R. P. Shaw, and C. Caceres (2003). Tectonic assembly of the northern Andean block, in The circum-Gulf of Mexico and the Caribbean: Hydrocarbon habitats, basin formation, and plate tectonics, AAPG Memoir 79, 1–34. Coates, A. G., L. S. Collins, M. P. Aubry, and W. A. Berggren (2004). The geology of the Darien, Panama, and the late Miocene-Pliocene colli- sion of the Panama arc with northwestern South America, Bull. Geol. Soc. Am. 116, nos. 11–12, 1327–1344. Collot, J. Y., B. Marcaillou, F. Sage, F. Michaud, W. Agudelo, P. Charvis, D. Graindorge, M. A. Gutscher, and G. D. Spence (2004). Are rupture zone limits of great subduction earthquakes controlled by upper plate structures? Evidence from multichannel seismic reflection data ac- quired across the northern Ecuador–southwest Colombia margin, J. Geophys. Res. 109, B11103, doi: 10.1029/2004JB003060. Corredor, F. (2003). Seismic strain rates and distributed continental defor- mation in the northern Andes and three-dimensional seismotectonics of northwestern South America, Tectonophysics 372, 147–166. Cortés, M., and J. Angelier (2005). Current states of stress in the northern Andes as indicated by focal mechanisms of earthquakes, Tectonophy- sics 403, 29–58. Defant, M. J., T. E. Jackson, M. S. Drummond, J. Z. De Boer, H. Bellon, M. D. Feigenson, R. C. Maury, and R. H. Stewart (1992). The geochem- istry of young volcanism throughout western Panama and southeastern Costa Rica: An overview, J. Geol. Soc. 149, 569–579. Dimate, C., L. A. Rivera, A. Taboada, B. Delouis, A. Osorio, E. Jimenez, A. Fuenzalida, A. Cisternas, and I. Gomez (2003). The 19 January 1995 Tauramena (Colombia) earthquake: Geometry and stress regime, Tectonophysics 363, 159–180. Faul, U. H., and I. Jackson (2005). The seismological signature of temper- ature and grain size variations in the upper mantle, Earth Planet. Sci. Lett. 234, 119–134. Frohlich, C. (2006). Deep Earthquakes, Cambridge Univ. Press, Cambridge, U.K., 574 pp. Frohlich, C., K. Kadinsky-Cade, and S. D. Davis (1995). A reexamination of the Bucaramanga, Colombia, earthquake nest, Bull. Seismol. Soc. Am. 85, 1622–1634. Geosoft (2010). GM-SYS profile module, Oasis montaj processing and mapping. Gil-Rodriguez, J. (2010). Igneous Petrology of the Colosa Gold-Rich Porphyry System (Tolima, Colombia),M. Sc. Thesis, University of Arizona, 51 pp. Haskov, J., S. Malone, D. McClury, and R. Crosson (1989). Coda-Q for the State of Washington, Bull. Seismol. Soc. Am. 79, 1024–1038. 2044 C. A. Vargas and P. Mann Hinze, W. J., C. Aiken, J. Brozena, B. Coakley, D. Dater, G. Flanagan, R. Forsberg, T. Hildenbrand, G. R. Keller, J. Kellogg, R. Kucks, X. Li, A. Mainville, R. Morin, M. Pilkington, D. Plouff, D. Ravat, D. Roman, J. Urrutia-Fucugauchi, M. Veronneau, M. Webring, and D. Winester (2005). New standards for reducing gravity data: The North American gravity database, Geophysics 70, no. 4, J25–J32. Hoerl, A. E., and R. W. Kennard (1970). Ridge regression: Biased estimation for nonorthogonal problems, Technometrics 12, 55–67. Hoerl, A. E., R. W. Kennard, and K. F. Baldwin (1975). Ridge regression: Some simulation, Comm. Stat. 4, 105–123. INGEOMINAS (1999). Risaralda, Department Mining Inventory, Internal report (in Spanish). Jarvis, A., H. I. Reuter, A. Nelson, and E. Guevara (2008). Hole-filled SRTM for the globe version 4, available from the CGIAR-CSI SRTM 90m Database (http://srtm.csi.cgiar.org; last accessed July 2012). Karato, S., and H. Jung (1998). Water, partial melting and the origin of the seismic low velocity and high attenuation zone in the upper mantle, Earth Planet. Sci. Lett. 157, 193–207. Lacruz, J., A. Ugalde, C. A. Vargas, and E. Carcolé (2009). Coda-wave attenuation imaging of Galeras Volcano, Colombia, Bull. Seismol. Soc. Am. 99, no. 6, 3510–3515. Levenberg, K. (1944). A method for the solution of certain nonlinear problems in least squares, Q. Appl. Math. 2, 164–168. Lonsdale, P. (2005). Creation of the Cocos and Nazca plates by fission of the Farallon plate, Tectonophysics 404, 237–264. Lopez, C. (2004). Upper crust models of Colombia, INGEOMINAS, Bogota, internal Rept., 64 pp. Malin, P. E. (1978). A first order scattering solution for modeling lunar and terretrial seismic coda, Ph.D. Dissertation, Princeton University, Princeton, New Jersey. Marquardt, D. W. (1963). An algorithm for least squares estimation of nonlinear parameters, J. Soc. Ind. Appl. Math. 11, 55–67. Maus,S., S.Macmillan,T.Chernova,S.Choi,D.Dater,V.Golovkov,V.Lesur, F. Lowes, H. Luhr,W.Mai, S.McLean, N. Olsen,M. Rother, T. Sabaka, A. Thomson, and T. Zvereva (2005). The 10th generation international geomagnetic reference field, Phys. Earth Planet. Int. 151, 320–322. Maus, S., T. Sazonova, K. Hemant, J. D. Fairhead, and D. Ravat (2007). National Geophysical Data Center candidate for the World Digital Magnetic Anomaly Map, Geochem. Geophys. Geosyst. 8, Q06017, doi: 10.1029/2007GC001643. McLaughlin, D. H., andM. Arce (1970). Economic geology of the Zipaquira quadrangle and adjoining area, Department of Cundinamarca, Colombia, U.S. Geol. Surv. Open-File Rept., 141 pp. Mitchell, B. J., and L. Cong (1998). Lg coda Q and its relation to the structure and evolution of continents: A global perspective, Pure Appl. Geophys. 153, 655–663. Mitchell, B. J., Y. Pan, J. Xie, and L. Cong (1997). Lg coda variation across Eurasia and its relation to crustal evolution, J. Geophys. Res. 102, 22,767–22,779. Mukhopadhyay, S., and J. Sharma (2010). Attenuation characteristics of Garwhal–Kumaun Himalayas from analysis of coda of local earth- quakes, J. Seismol. 14, doi: 10.1007/s10950-010-9192-9. National Hydrocarbons Agency of Colombia (2010). Total Bouguer Anoma- lies Map (MABT) of Colombia, National Hydrocarbons Agency of Colombia (ANH). O’Doherty, K. B., C. J. Bean, and J. McCloskey (1997). Coda wave imaging of the Long Valley caldera using a spatial stacking technique,Geophys. Res. Lett. 24, 1545–1550. Ojeda, A., and J. Havskov (2001). Crustal structure and local seismicity in Colombia, J. Seismol. 5, no. 4, 575–593. Oliver, M. A., and R. Webster (1990). Kriging: A method of interpolation for geographical information system, Int. J. Geogr. Inf. Syst. 4, no. 3, 313–332. Ordoñez, C. O. (2001). Caracterização isotópica Rb-Sr e Sm-Nd dos principais eventos magmáticos nos Andes Colombianos, Ph.D. Thesis, Instituto de Geociências, Universidade de Brasilia, Brasilia, 197 pp. (in Portuguese). Ostios, M., F. Yoris, and H. G. Avé Lallemant (2005). Overview of the south- east Caribbean-South American plate boundary zone, Spec. Pap. Geol. Soc. Am. 394, 53–89. Pardo, N., H. Cepeda, and J. M. Jaramillo (2005). The Paipa volcano, Eastern Cordillera of Colombia, South America: Volcanic stratigraphy, Earth Sci. Res. J. 9, no. 1, 3–18. Paris, G., R. L. Dart, and M. N. Manchette (2000). Map of Quaternary faults and folds of Colombia and its offshore regions, U.S. Geol. Surv. Open-File Rept., Scale 1:2,500,000. Priestley, K., and D. McKenzie (2006). The thermal structure of the lithosphere from shear wave velocities, Earth Planet. Sci. Lett. 244, 285–301. Pulli, J. J. (1984). Attenuation of coda waves in New England, Bull. Seismol. Soc. Am. 74, 1149–1166. Ramos, V. A., and A. Folguera (2009). Andean flat-slab subduction through time, Spec. Publ. Geol. Soc. Lond. 327, 31–54. Sandwell, D. T., and W. H. F. Smith (2009). Global marine gravity from retracked Geosat and ERS-1 altimetry: Ridge segmentation versus spreading rate, J. Geophys. Res. 114, B01411, doi: 10.1029/ 2008JB006008. Schneider, J. F., W. D. Pennington, and R. P. Meyer (1987). Microseismicity and focal mechanisms of the intermediate-depth Bucaramanga nest, Colombia, J. Geophys. Res. 92, no. B13, 13,913–13,926, doi: 10.1029/JB092iB13p13913. Singh, S., and R. B. Herrmann (1983). Regionalization of crustal coda Q in the continental United States, J. Geophys. Res. 88, 527–538. Taboada, A., L. A. Rivera, A. Fuenzalida, A. Cisternas, H. Philip, H. Bijwaard, J. Olaya, and C. Rivera (2000). Geodynamics of the Northern Andes; subductions and intracontinental deformation (Colombia), Tectonics 19, 787–813. Talwani, M., and J. R. Heirtzler (1964). Computation of magnetic anomalies caused by two-dimensional bodies of arbitrary shape, in Computers in the mineral industries, Part 1: Stanford Univ. Publ., Geological Sciences, G. A. Parks (Editor), Vol. 9, 464–480. Talwani, M., J. L. Worzel, and M. Landisman (1959). Rapid gravity computations for two-dimensional bodies with application to the Mendocino submarine fracture zone, J. Geophys. Res. 64, 49–59. Tistl, M. (1994). Geochemistry of platinum-group elements of the zoned ultramafic Alto Condoto complex, northwest Colombia, Econ. Geol. 89, no. 1, 158–167. Tittgemeyer, M., F. Wenzel, T. Ryberg, and K. Fuchs (1999). Scales of heterogeneities in the continental crust and upper mantle, Pure Appl. Geophys. 156, 29–52. Trenkamp, R., J. N. Kellogg, J. T. Freymueller, and H. Mora (2002). Wide plate margin deformation, southern Central America and northwestern South America, CASA GPS observations, J. S. Am. Earth Sci. 15, no. 2, 157–171. van der Hilst, R. D., and P. Mann (1994). Tectonic implications of tomo- graphic images of subducted lithosphere beneath northwestern South America, Geology 22, 451–454. Vargas, C. A., C. Alfaro, L. A. Briceño, I. Alvarado, andW. Quintero (2009). Mapa Geotérmico de Colombia, 2009, in Proceedings of X Simposio Bolivariano Exploración Petrolera en Cuencas Subandinas, Colombia (in Spanish). Vargas, C. A., P. Mann, C. Gomez, L. A. Briceño, and C. Rey (2010). Trans- Andean mega-regional seismic reflection line extending from the Caribbean coast to Cordillera Oriental of Colombia: Implications for hydrocarbon exploration, AAPG Annual Convention: Unmasking the Potential of Exploration & Production, New Orleans, Abstract 90104. Vargas, C. A., L. G. Pujades, and L. A. Montes (2007). Seismic structure of south-central Andes of Colombia by tomographic inversion, Geofísc. Int. 46, no. 2, 117–127. Vargas, C. A., A. Ugalde, L. G. Pujades, and J. A. Canas (2004). Spatial variation of coda wave attenuation in northwestern Colombia, Geophys. J. Int. 158, 609–624. Xie, J. (2002). Lg Q in the eastern Tibetan Plateau, Bull. Seismol. Soc. Am. 92, 871–876. Tearing and Breaking Off of Subducted Slabs as the Result of Collision of the Panama Arc-Indenter 2045 Yang, Y., D. W. Forsyth, and D. S. Weeraratne (2007). Seismic attenuation near the East Pacific Rise and the origin of the low-velocity zone, Earth Planet. Sci. Lett. 258, 260–268, doi: 10.1016/j.epsl.2007.03.040. Zarifi, Z. (2006). Unusual subduction zones: Case studies in Colombia and Iran, Ph.D. Thesis, University of Bergen, Norway, 78 pp. Department of Geosciences Universidad Nacional de Colombia–Sede Bogotá Carrera 45 No 26-85–Edificio Manuel Ancizar Bogotá D.C.–Colombia, 111321
[email protected] (C.A.V.) Department of Earth and Atmospheric Sciences University of Houston 4800 Calhoun Boulevard Houston, Texas 77004
[email protected] (P.M.) Manuscript received 3 November 2012 2046 C. A. Vargas and P. Mann