Journal of South American Earth Sciences 89 (2019) 76–91

Contents lists available at ScienceDirect

Journal of South American Earth Sciences

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

Crustal deformation in the northern Andes – A new GPS velocity field T ∗ Héctor Mora-Páeza, , James N. Kelloggb, Jeffrey T. Freymuellerc, Dave Mencind, Rui M.S. Fernandese, Hans Diederixa, Peter LaFeminaf, Leonardo Cardona-Piedrahitaa, Sindy Lizarazoa,g, Juan-Ramón Peláez-Gaviriaa, Fredy Díaz-Milaa, Olga Bohórquez-Orozcoa, Leidy Giraldo-Londoñoa, Yuli Corchuelo-Cuervoa a Colombian Geological Survey, Space Geodesy Research Group, Colombia b University of South Carolina, USA c University of Alaska, Fairbanks, USA d University of Colorado and UNAVCO, USA e University of Beira Interior, Instituto D. Luiz, Portugal f The Pennsylvania State University, USA g Now at the University of Nagoya, Japan

ARTICLE INFO ABSTRACT

Keywords: We present a velocity field for northwestern South America and the southwest Caribbean based onGPS Space geodesy Continuously Operating Reference Stations in Colombia, Panama, Ecuador and Venezuela. This paper presents North Andean block the first comprehensive model of North Andean block (NAB) motion. We estimate that the NAB ismovingtothe Crustal deformation northeast (060°) at a rate of 8.6 mm/yr relative to the South America plate. The NAB vector can be resolved into a margin-parallel (035°) component of 8.1 mm/yr rigid block motion and a margin-normal (125°) component of 4.3 mm/yr. This present-day margin-normal shortening rate across the Eastern Cordillera (EC) of Colombia is surprising in view of paleobotanical, fission-track, and seismic reflection data that suggest rapid uplift (7km) and shortening (120 km) in the last 10 Ma. We propose a “broken indenter” model for the Panama-Choco arc, in which the Choco arc has been recently accreted to the NAB, resulting in a rapid decrease in shortening in the EC. The Panama arc is colliding eastward with the NAB at approximately 15–18 mm/yr, and the Panama-Choco collision may have been responsible for much of the uplift of the EC. The present on-going collision poses a major earthquake hazard in northwestern Colombia from the Panama border to Medellin area. Since the northeastward margin-parallel motion of the NAB is now greater than the rate of shortening in the EC, northeast trending right- lateral strike-slip faulting is the primary seismic hazard for the 8 million inhabitants of Bogota, the capital city of Colombia. There continues to be a high risk of a great megathrust earthquakes in southern Colombia along the Ecuador-Colombia trench. Trench earthquakes have only released a fraction of the energy accumulated in the Ecuador-Colombia trench since the 1906 Ecuador earthquake, and interseismic strain is accumulating rapidly at least as far north as Tumaco, the rupture area of the 1979 earthquake.

1. Introduction measurements performed in Costa Rica, Panama, Colombia, Venezuela and Ecuador, from the late 1980s and early 1990s demonstrated the Given the controversies over whether continental deformation is northeastward movement of the North Andes, Caribbean – North Andes best described as relative movements of rigid blocks or continuous convergence, and the ongoing rapid collision of the Panama arc with deformation, whether mountain building rates can be compared to the North Andes (e.g., Kellogg et al., 1990; Freymueller et al., 1993; convergent boundary processes, and how slip is partitioned at complex Trenkamp et al., 2002). convergent plate boundaries, northwestern South America offers a good Using GPS results from the first three CASA GPS campaigns field laboratory because it includes an active arc-continent collision, (1988–1991), Freymueller et al. (1993) showed evidence for northward margin-parallel slip, and active “flat-slab” and “normal” subduction. movement of the North Andes and convergence at the South Caribbean CASA (Central And South America) GPS Project campaign deformed belt. Kellogg and Vega (1995) used CASA GPS results to

∗ Corresponding author. E-mail addresses: [email protected], [email protected] (H. Mora-Páez). https://doi.org/10.1016/j.jsames.2018.11.002 Received 20 July 2018; Received in revised form 13 October 2018; Accepted 1 November 2018 Available online 03 November 2018 0895-9811/ © 2018 Elsevier Ltd. All rights reserved. H. Mora-Páez et al. Journal of South American Earth Sciences 89 (2019) 76–91 propose a rigid Panama block and rapid Panama-North Andes con- Panama; one station on Cocos Island, Costa Rica; seven stations in vergence. Trenkamp et al. (2002) presented CASA campaign data from Ecuador, including two IGS stations, GLPS and RIOP; and two stations 1991 to 1998 that showed wide plate boundary deformation and escape in Venezuela. Nine of the stations used are part of the COCONet Project tectonics from the subducting Carnegie Ridge along an approximately (Braun et al., 2012). This precise velocity field is based on permanent 2000 km long transform fault belt, known as the Eastern frontal Fault stations with a minimum of 2.5 years of observations (e.g., Blewitt and zone, or perhaps better named the North Andean Boundary Fault. This Lavallee, 2002). fault extends from the Gulf of Guayaquil in Ecuador to Venezuela, and Colombian data are from the GeoRED Project (Geodesia: Red de which could be a manifestation of the incipient dismemberment of the Estudios de Deformación), which is run by the Space Geodesy Research northwestern South American plate. That study also showed locking of Group of the Colombian Geological Survey (CGS, Servicio Geológico the subducting Nazca plate and strain accumulation in the Ecuador- Colombiano; formerly INGEOMINAS). Initiated in 2007 by the CGS, Colombia forearc at the latitude of the border between the two coun- GeoRED is a research and development project based on space geodesy tries, collision of the Panama arc with Colombia, and Caribbean-North technology to catalog and interpret the geodynamics and associated Andes convergence. Elastic modeling of horizontal displacements were hazards within the broad northwestern South America plate margin consistent with partial locking in the Ecuador subduction zone and a deformation zone, (Mora-Páez et al., 2018; Mora-Páez, 2006). The fully locked Panama-Colombia collision zone. GeoRED network currently has 108 operating sites, located on the Significant deformation in the region is driven by Cocos Ridge Nazca, South America and Caribbean plates, (Mora-Páez et al., 2018), subduction, Panama collision and subduction of the Caribbean plate although sites with less than 2.5 years are not used in this study. Data (e.g., van Benthem and Govers, 2010; Kobayashi et al., 2014). from Ecuadorean stations have been provided by the Geophysics In- Kobayashi et al. (2014) used GPS data from Panama, Costa Rica and stitute of the National Polytechnic University (Escuela Politécnica Na- Colombia and elastic block modeling to conclude that tectonic escape cional), (Mothes et al., 2013). Data from stations located in Panama, from Cocos Ridge collision drives northeast motion of the Panama re- Venezuela and Costa Rica have been obtained from the COCONet gion and subsequent collision with the North Andes and Choco blocks. Project (https://www.unavco.org/data/gps-gnss/gps-gnss.html) and Mora-Páez et al. (2016) measured velocities from nine GPS Con- two stations from the Panama Canal Authority. tinuously Operating Reference Stations (CORS) and twenty campaign sites in the northeast trending Eastern Cordillera of Colombia that 2.2. GPS data processing and analysis constrain the rate of this motion, showing oblique convergence, con- sisting of 8 mm/yr of right-lateral strike-slip and only 4 mm/yr of All GPS data have been processed with GIPSY-OASIS II software, v northwest-southeast shortening. Perez et al. (2018) used GPS data 6.3 developed by the Jet Propulsion Laboratory (JPL), California primarily from Venezuela, to estimate motion of the North Andean Institute of Technology (Bertiger et al., 2010; Zumberge et al., 1997). block (NAB) relative to South America, but their estimate only included Daily station coordinates are expressed in ITRF2008. The station velo- four sites in northern Colombia. Nocquet et al. (2014) used GPS data, cities (Supp. Data 1) are computed using the HECTOR software (Bos primarily from Ecuador and Peru, to quantify the margin-parallel et al., 2013), software developed at SEGAL (Space & Earth Geodetic northeastward motion of the North Andean sliver (NAS) or North An- Analysis Laboratory at the University of Beira Interior, Portugal) that is dean block (NAB) and the southeastward motion of the Peru sliver, but used to estimate the linear trend in time-series with temporal correlated their estimate for the NAB used only two sites in Colombia. noise. On the Pacific coast of Ecuador and southernmost Colombia, ob- A power-law plus white noise model was assumed. For each time served deformation includes a large contribution from elastic de- series a power-spectrum plot was generated from the residuals, and formation due to the locked part of the subduction interface (Trenkamp compared to the predicted power-spectrum of the noise model was et al., 2002; Kobayashi et al., 2014). However, that contribution varies compared with the observed power spectrum to verify that the proper considerably along strike, with high elastic strain observed in northern noise model has been used. Ecuador and nearly zero in southern Ecuador (Chlieh et al., 2014; Seasonal signals (an annual and semi-annual signal) have been in- Nocquet et al., 2014). Vallée at al. (2013) documented a long slow-slip cluded in the estimation of the secular velocities in order to reduce their event on a shallow locked patch along one part of the Ecuador sub- influence on the estimated velocities. We follow the current state-of-art duction interface (near Isla la Plata) using continuous GPS and broad- approach that assumes that the amplitude of such signals is constant at band seismic data. each considered period and described by a sinusoidal curve. Bos et al. In this paper, we present a new velocity field derived from the (2010) have demonstrated that when the time series are longer than GeoRED continuous GPS network, Colombia, and other CORS sites in 3–5 years, the remaining influence of the seasonal signals on the esti- northwestern South America and southwestern Caribbean that have at mated trend can be neglected. least 2.5 years of data. This precise and spatially extensive velocity field In order to use the same reference epoch, we select January 1, 2010 for 60 permanent stations allows us to make a more accurate estimate as reference epoch for all estimations instead of the midpoint of each of North Andean Block motion, and hence isolate motion of the North individual time-series. Table 2 indicates the data span for each station. Andes block from more localized deformation associated with the Panama arc – NAB collision, and Nazca “normal” and Caribbean “flat- 2.3. Velocity estimation and South American reference frame slab” subduction. The new velocity field will also be useful for thees- timation of earthquake hazard in the seismically active NAB, especially We estimated the angular velocity of South America with respect to near the large population centers of Bogotá, Medellín, Cali and Pasto, the ITRF2008 global reference frame (Altamimi et al., 2011). This es- Colombia. timate is from a previously unpublished model presented by Fernandes et al. (2016) for the motion of South America in ITRF, and is detailed 2. Data and methodology below because it had not been published before. We then removed this rotation to compute the relative motions of our stations with respect to 2.1. GPS data South America. This estimate is based on a GNSS velocity field esti- mated using exactly the same approach as for the GeoRED network, We have estimated a new geodetic velocity field in the region of the although the two solutions were done independently. Compared with North Andes using GPS data collected from 60 permanent stations in previously published present-day angular velocity models for the South Colombia, Panama, Costa Rica, Ecuador, and Venezuela: forty-six sta- American plate based on geodetic data, e.g., REVEL (Sella, 2002), tions in Colombia, including the IGS BOGT station; four stations in (Argus et al., 2010), ITRF2014 (Altamimi et al., 2016), this estimate

77 H. Mora-Páez et al. Journal of South American Earth Sciences 89 (2019) 76–91

Table 1 Estimated angular velocities for the SOAM plate with respect to ITRF2008 and NAB with respect to SOAM as computed in this study and by Nocquet et al. (2014). The North Andes relative to SOAM angular velocity is given in Cartesian coordinates in Supplemental Data 2.

° ° °/Myr °(a) °(a) °(b) °/Myr Plate Model Lat Lon ω σmax σmin ζmax σω

SOAM This Study −19.237 −133.603 0.1184 1.14 0.17 74.6 0.0071 NAB This Study 58.626 −174.848 0.0724 22.01 1.31 80.97 0.0017 Nocquet et al. (2014) 15.200 −83.400 0.2870 1.03 0.18 74.40 0.0180

(a) σmax, σmin: semi-major and semi-minor axes of the 1-σ error ellipse. (b) ζmax: azimuth of the semi-major axis reckoned clockwise from north.

Table 2 Site velocities (mm/yr) relative to stable South America, Fig. 2. Margin-normal (125°, NW-SE), and margin-parallel (035°, NE-SW) velocities, Fig. 5.

ID Long Lat East Vel (mm/ North Vel (mm/ Sig East Sig North Margin Normal Margin Parallel Margin Normal Margin Parallel yr) yr) Vel Vel Sigma Sigma

ACP1 280.050 9.371 22.2 1.7 0.3 0.3 17.2 14.1 0.3 0.3 ACP6 280.592 9.238 22.1 2.3 0.2 0.2 16.8 14.6 0.2 0.2 ALPA 287.082 11.528 13.9 3.0 0.9 0.6 9.6 10.4 0.7 0.8 AUCA 283.117 −0.641 1.9 −0.2 0.3 0.2 1.6 0.9 0.2 0.3 BAAP 286.446 4.072 −0.2 −0.6 0.5 0.2 0.2 −0.7 0.3 0.4 BACO 284.308 9.402 18.6 −2.4 0.9 1.0 16.6 8.6 1.0 1.0 BAEZ 282.113 −0.459 6.2 0.5 0.3 0.2 4.8 3.9 0.2 0.2 BAME 285.435 4.236 6.4 3.9 0.7 0.2 3.0 6.8 0.4 0.6 BAPA 285.342 5.466 7.5 4.2 0.3 0.2 3.8 7.8 0.3 0.3 BARU 284.410 10.258 17.0 −0.3 0.8 0.4 14.1 9.5 0.6 0.7 BASO 282.607 6.203 11.7 5.5 0.9 0.4 6.5 11.2 0.6 0.7 BOBG 286.642 8.312 11.9 4.6 0.5 0.3 7.1 10.6 0.4 0.5 BOGT 285.919 4.640 4.4 4.7 0.3 0.2 0.9 6.4 0.2 0.2 BUGT 283.004 3.826 9.7 4.9 0.3 0.2 5.1 9.6 0.2 0.3 CAPI 287.572 5.351 1.8 0.9 0.3 0.4 1.0 1.7 0.3 0.3 CIA1 283.643 3.505 7.6 4.2 1.3 0.6 3.9 7.8 0.9 1.1 CN19 289.952 12.612 18.2 3.3 0.7 0.6 13.1 13.1 0.6 0.6 CN28 280.966 8.625 22.7 5.3 0.7 0.5 15.5 17.3 0.6 0.6 CN33 279.673 8.487 25.3 −3.3 1.8 1.8 22.6 11.8 1.8 1.8 CN35 278.637 13.375 17.0 −2.9 0.6 0.4 15.6 7.4 0.5 0.5 CN37 284.737 10.793 14.0 3.6 1.1 1.6 9.4 11.0 1.5 1.3 CN38 288.012 12.222 16.9 3.1 0.5 0.3 12.0 12.2 0.3 0.4 CN40 291.042 12.180 17.6 1.9 0.7 0.7 13.3 11.7 0.7 0.7 COEC 282.213 0.716 7.3 2.6 1.0 0.4 4.5 6.3 0.7 0.9 CORO 284.712 9.328 17.0 1.1 0.3 0.3 13.3 10.7 0.3 0.3 CUC1 287.487 7.932 12.6 2.4 1.5 0.4 9.0 9.2 0.9 1.3 ESMR 280.276 0.935 22.2 6.8 0.5 0.3 14.3 18.3 0.4 0.5 GLPS 269.696 −0.743 54.8 2.1 0.2 0.2 43.7 33.2 0.2 0.2 GUAP 282.105 2.574 12.5 3.5 0.5 0.4 8.2 10.0 0.4 0.5 ISCO 272.944 5.544 55.1 65.8 0.7 0.5 7.4 85.5 0.6 0.6 MALO 278.394 4.003 53.1 4.2 0.6 0.3 41.1 33.9 0.4 0.5 MECE 286.288 7.107 9.5 4.5 0.3 0.4 5.2 9.1 0.4 0.3 MITU 289.768 1.261 −0.4 0.7 0.4 0.3 −0.7 0.3 0.3 0.4 MZAL 284.529 5.030 6.5 7.7 0.3 0.7 0.9 10.0 0.6 0.4 OCEL 288.384 4.271 −0.3 0.9 0.3 0.2 −0.8 0.6 0.2 0.3 PAL1 286.811 7.136 8.5 4.2 0.4 0.4 4.6 8.3 0.4 0.4 PASI 283.501 0.513 1.2 0.2 0.3 0.4 0.8 0.8 0.4 0.4 POVA 283.385 2.449 9.7 2.8 0.5 0.3 6.3 7.9 0.4 0.4 PUIN 292.097 3.851 0.0 −0.4 0.5 0.3 0.2 −0.4 0.4 0.5 QUIL 282.709 1.394 9.5 4.2 0.8 0.3 5.3 8.9 0.5 0.7 RIOP 281.349 −1.651 3.5 −4.1 0.6 0.4 5.2 −1.4 0.5 0.6 SAN0 278.284 12.580 18.2 −2.4 0.4 0.2 16.3 8.5 0.3 0.3 SEL1 284.471 6.191 9.0 4.6 0.3 0.2 4.7 8.9 0.3 0.3 SNLR 281.153 1.293 15.4 0.8 0.4 0.3 12.2 9.5 0.3 0.4 TICU 290.061 −4.187 −0.4 −0.1 0.3 0.3 −0.3 −0.3 0.3 0.3 TONE 283.861 6.324 9.3 5.4 0.5 0.2 4.5 9.7 0.3 0.4 TUCO 281.252 1.815 18.3 3.0 0.4 0.2 13.3 13.0 0.3 0.4 URRA 283.790 8.012 17.8 3.9 0.8 0.7 12.3 13.4 0.8 0.8 UWAS 287.609 6.451 5.1 2.4 0.3 0.3 2.8 4.8 0.3 0.3 VBUV 286.141 5.533 8.6 5.1 0.5 0.3 4.1 9.1 0.4 0.4 VDPR 286.752 10.436 13.8 4.7 0.2 0.2 8.7 11.8 0.2 0.2 VMAG 285.153 9.287 13.7 3.8 0.5 0.4 9.0 11.0 0.5 0.5 VNEI 284.745 3.062 3.9 3.6 0.3 0.3 1.1 5.1 0.3 0.3 VORA 283.278 7.818 19.4 1.5 1.8 1.5 15.0 12.3 1.6 1.7 VOTU 285.290 7.019 11.1 4.8 0.3 0.4 6.3 10.3 0.4 0.3 VPIJ 284.893 4.397 6.3 4.1 0.4 0.2 2.9 7.0 0.3 0.3 VPOL 285.139 10.794 13.1 6.3 0.5 0.2 7.1 12.6 0.3 0.4 VROS 285.677 4.847 4.9 4.1 0.5 0.3 1.7 6.2 0.4 0.4 VSJG 287.361 2.533 −0.4 0.2 0.6 0.3 −0.4 −0.1 0.4 0.5 VSJP 284.164 4.781 5.8 5.0 0.4 0.3 1.9 7.4 0.3 0.3

78 H. Mora-Páez et al. Journal of South American Earth Sciences 89 (2019) 76–91

Fig. 1. Distribution of the stations and the estimated residuals of the South American model used in this study. uses a much denser network of GNSS Continuously Operating Reference Consequently, most of the stations located in Argentina were excluded Station velocity solutions (Fernandes et al., 2017). The final estimate of because they contained co-seismic and post-seismic signals that could the angular velocity of South America used 91 stations that have more not be easily modeled, and they did not have enough data before the than 3.5 years of data span and that are within the limits of the stable Maule earthquake. part of the South America tectonic plate, and have residuals of < 1.5 Table 1 shows the estimated angular velocities for South America mm/yr relative to the final rigid South America model (61 stations and for the North Andes Block (NAB). Fig. 2 and Table 2 show the have residuals smaller than 1 mm/yr). Fig. 1 shows the distribution of relative motions of our North Andean network with respect to South the stations and the estimated residuals that demonstrate a good fit of America. the estimated angular velocities for most of the plate, in particular for our area of interest. The 91 stations used were a subset of stations from several addi- 2.4. Combined velocity field tional networks in addition to GeoRED, including RAMSAC from Argentina (81 stations processed); IBGE from Brasil (67); LISN (Low We combined our velocity field with the published results of Latitude Ionospheric Network) (43); and others, including IGS (20). All Nocquet et al. (2014), in order to provide a more complete picture of the time-series were analyzed and stations with abnormal features (e.g., velocities across the North Andes Block. As Nocquet et al. (2014) used a undocumented offsets, very noisy solutions, a problem at many LISN large number of sites in Ecuador, with only a few sites in Colombia, the stations due to unstable monumentation; no linear secular motions) two data sets complement one another. However, they defined the were removed. Many stations in these networks (particularly the ones stable South American plate using a different set of sites than used in installed for surveying and mapping purposes) have poor stability, but our model, so their realization of the South America-fixed frame is each network has some good sites that permit much more dense cov- slightly different than ours. We used the five overlapping sites between erage of the plate. the two data sets (AUCA, ESMR, SNLR, TUCO, RIOP), all located in One additional and larger problem when computing the angular northern Ecuador or southern Colombia, to compute the translation velocity of South America is the co-seismic or post-seismic effects in the between the two South America frames. The estimated translation dif- estimated secular motions due to recent large earthquakes. We have ference between the two sets of velocities is 0.05 mm/yr and 1.4 mm/yr investigated the effect of displacements associated with recent strong for the east and north components respectively, and this difference was earthquakes (in particular, Maule in 2010 and Iquique in 2014) at added to the Nocquet et al. (2014) velocities for the combined velocity stations several hundreds of kilometers from the epicenter. field, (Fig. 3). The data uncertainties are ∼0.5 and ∼0.3 mm/yr for the east and north components, so only the north component difference is

79 H. Mora-Páez et al. Journal of South American Earth Sciences 89 (2019) 76–91

Fig. 2. GPS vectors relative to stable South America, Table 2. One sigma error ellipses. North Andean Block (NAB). Eastern Cordillera (EC), East Andean fault system (EAFS), South Caribbean deformed belt (SCDB), Panama Block (PB), Bogota (BOGT), Medellin (SEL1), Pasto (QUIL), Tumaco (TUCO). significant. The residuals in this estimated translation are consistent (TUCO) moved rapidly east-northeastward at 22.2 ± 0.5 mm/yr and with the uncertainties. We estimated only a translation and not a 18.3 ± 0.4 mm/yr respectively prior to the April 16, 2016 M 7.8 translation plus rotation because the area spanned by the overlap sites Ecuador Pedernales subduction earthquake. These mean values are is small, which would make an unconstrained estimation of the rotation slightly greater than the mean vectors reported in Trenkamp et al. unstable in the presence of noise; the rotational difference is expected (2002), 20.8 ± 2.7 mm/yr (ESME) and 16.1 ± 3.76 mm/yr (TUMA) to be small in any case. The very small estimated difference between the respectively, but fall within the measurement uncertainties. frames justifies that assumption. We note, however, that the new vectors for several sites in south- west Colombia from 2°N to 5°N reported here differ from previous CASA results. For these sites, there is an apparent increase in eastward 3. Results motion in the new data set. For example, the Popayan vector (PPYN) was only 0.5 ± 4.7 mm/yr to the east (Trenkamp et al., 2002), and the 3.1. Comparison to previous velocity estimates Popayan vector reported here (POVA) is 9.7 ± 0.5 mm/yr to the east. The Cali vector (CALI) was only 0.1 ± 2.3 mm/yr to the east Our velocity field is based on continuous observations in permanent (Trenkamp et al., 2002), and the vector reported here (CIA1) at the sites, most of which are located near campaign sites used to estimate same location, is 7.6 ± 1.3 mm/yr. The Buenaventura vector (BUEN) velocities in previous studies (Trenkamp et al., 2002). Malpelo Island was reported by Trenkamp et al. (2002) as 4.5 ± 3.2 mm/yr, and the (MALO) and Galapagos Island (GLPS) on the Nazca oceanic plate con- nearby BUGT vector reported in this study is 9.7 ± 0.3 mm/yr. These verge east northeastward relative to stable South America at rates of differences are not related to flaws in the older GPS data analysis, as 53.1 ± 0.6 mm/yr and 54.8 ± 0.2 mm/yr, respectively (Table 2). The reprocessed solutions for the campaign GPS data confirm the older re- vectors estimated by Trenkamp et al. (2002) were 53.6 ± 2.1 mm/yr sults. It is more likely that the differences are due to the different data (MALS) and 58.2 ± 1.4 mm/yr (GALA), respectively. Near the Co- spans, the fact that solutions derived from continuous observations are lombia-Ecuador border, coastal sites Esmeraldas (ESMR) and Tumaco

80 H. Mora-Páez et al. Journal of South American Earth Sciences 89 (2019) 76–91

Fig. 3. Combined Velocity Field relative to stable South America. One sigma error ellipses. Yellow arrows (this study); white arrows, Nocquet et al. (2014) velocities; red circles: overlapping sites. North Andean Block (NAB), Eastern Cordillera (EC), Panama Block (PB). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) more robust than the ones derived from campaign measurements and (mainly in Ecuador, plus sites PSTO and BOGT in Colombia) and thus the use of different reference frames; Trenkamp et al. (2002) express may not be appropriate to extrapolate farther to the north. In addition, their results in ITRF96 while the velocities of this paper are expressed in the site BOGT shows anomalous horizontal motion relative to others in ITRF2008. A possible explanation for some of these differences is dis- the vicinity, perhaps due to groundwater extraction and compaction (it cussed in this paper in section 3.3.1. Nazca Subduction. subsides at a rate of ∼40 mm/yr), so it should not have been used and slightly biased the angular velocity estimate of Nocquet et al. (2014). 3.2. North Andes Block motion Replacement of BOGT with a more appropriate set of sites in Colombia, which were not available to Nocquet et al. (2014), is the major differ- Determining North Andes Block (NAB) motion relative to stable ence between our estimation of the North Andes motion and that of South America is important to accurately estimate slip partitioning into Nocquet et al. (2014). Perez et al. (2018) used nine velocity vectors to margin-parallel rigid body translation along the broad East Andean estimate the relative motion of the North Andes block relative to South Fault System (or Great North Andean Boundary Fault) (EAFS, Fig. 2) America, but only 4 vectors were from Colombia (including VDPR, and margin-normal elastic strain accumulation on the Ecuador-Co- ALPA, CN38 and MONT) and all the vectors used were north of 8.5° N. lombia trench, and permanent shortening and mountain building in the Thus, we re-estimated the North Andes Euler vector based on a set eastern Cordillera. Nocquet et al. (2014) and Chlieh et al. (2014) esti- of sites that span a larger part of the North Andes block (sites GPH1, mated the North Andes sliver (NAS) motion relative to stable South GYEC, CUER, TULC, QUIL, BAME, VPIJ, VSJP, BAPA, VBUV, VROS, America, but their estimates were based on data from a small area PAL1; (Figs. 2 and 4b). These sites are concentrated in southwestern

81 H. Mora-Páez et al. Journal of South American Earth Sciences 89 (2019) 76–91

Fig. 4. (a) Velocities relative to the North Andes block (NAB) (Table 3), One sigma error ellipses. Focal mechanisms: 1) October 17, 1992 (Wallace and Beck, 1993); 2) October 18, 1992 (Wallace and Beck, 1993); 3) September 14, 2016 (USGS National Earthquake Information Center, 2017). (b) Predicted motion of the NAB relative to stable South America from this study (yellow) and Nocquet et al. (2014) (white). Red circles: sites used to re-estimate the North Andes motion. No sites were selected north of 7.5° N, so that Panama-North Andes collision-related strain did not bias the angular velocity estimate. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Ecuador and central Colombia, with a few sites in northern Ecuador or Nocquet et al. (2014). After each trial inversion, residuals were checked southern Colombia. The selection of sites was optimized by trial and and the distribution of sites was reassessed. In the area of overlap, the error, starting from the sites chosen by Nocquet et al. (2014) with BOGT main differences from the site selection of Nocquet et al. (2014) were removed. Nocquet et al. (2014) demonstrated that elastic deformation that BOGT was not used in our estimate, and the sites we chose in from the subduction zone is negligible in southern Ecuador, but very southern Ecuador were slightly farther inland than those used by large in northern Ecuador. We thus began by selecting sites in southern Nocquet et al. (2014). Sites were restricted to sites west of the NAB- Ecuador and across central Colombia located away from the coast South American fault boundary, the EAFS. We selected no sites north of where subduction-related deformation should be minimal based on 7.5° N, so that Panama-North Andes and CA-North Andes convergence- previously published elastic models (Trenkamp et al., 2002; Kobayashi related strain did not bias the angular velocity estimate (e.g., Kobayashi et al., 2014; Nocquet et al., 2014), and where neighboring sites had et al., 2014). Since all of the sites used by Perez et al. (2018) were north similar velocities indicating minimal internal strain. Sites close to major of 8.5° N., our NAB velocity estimates differed significantly. crustal fault zones were also avoided to minimize the impact of elastic We estimate that the North Andes block rotates counter-clockwise at deformation from those faults. a rate of 0.072°/Ma about a pole located at 58.6°N, 174.8°W (Table 1). We then fit a rigid body rotation to these velocities relative toSouth Velocities relative to the North Andes block (Fig. 4a and Table 3), and America, assuming no elastic deformation, the same method used by the predicted block motion are shown in Fig. 4b, and clearly show

82 H. Mora-Páez et al. Journal of South American Earth Sciences 89 (2019) 76–91

Fig. 4. (continued) elastic strain accumulation from Nazca plate subduction, oblique con- Ecuador as being due to a small component of elastic strain from the vergence across the Eastern Cordillera, and the effects of Panama arc locked subduction zone, which we minimized by choosing sites slightly collision and Caribbean plate subduction (Fig. 4a). Because the pole lies more inland than Nocquet et al. (2014). In northern Colombia, the about 75–80° away from the block itself, the predicted block motions difference is larger, with our predicted North Andes block motion being are in nearly the same direction across the entire block, with the North 4.2 mm/yr more westward and 0.7 mm/yr more southward relative to Andes moving 8.6 mm/yr toward N60°E (Fig. 4b). Based on GPS mea- the prediction of Nocquet et al. (2014). This is mainly due to the fact surements in Ecuador and southern Colombia, Chlieh et al. (2014) that the BOGT velocity biased the Nocquet et al. (2014) pole. calculated a remarkably similar estimate of 8.5 ± 1.0 mm/yr for the northeastward long-term rigid motion of the “North Andean Sliver”. 3.3. Subduction earthquake cycle The motion of the North Andes relative to the Caribbean plate is similar to the estimate of Kobayashi et al. (2014). 3.3.1. Nazca Subduction Our estimated rotation pole is very different from that of Nocquet at Sites on the Pacific coast of Colombia and Ecuador move inland al. (2014), which was located near the Caribbean coast of Central relative to South America and the North Andes faster than those lying America (Table 1). In southern Ecuador our estimated block motion is further east (Figs. 2 and 4). Near the Colombia-Ecuador border, coastal about 1.5 mm/yr more westward and 1.8 mm/yr more northward re- sites Esmeraldas (ESMR) and Tumaco (TUCO) moved eastward relative lative to the result of Nocquet et al. (2014) (Fig. 4b); the difference in to the North Andes at 14.2 ± 0.5 mm/yr and 11.3 ± 0.4 mm/yr re- the north component is mostly due to the small difference in South spectively prior to the April 16, 2016 M 7.8 Ecuador subduction American frames. The east-west difference is because we have inter- earthquake. Trenkamp et al. (2002) modeled the eastward motion of preted the small amount of east-west contraction across southern these coastal sites as being due to pre-earthquake 50% elastic locking

83 H. Mora-Páez et al. Journal of South American Earth Sciences 89 (2019) 76–91

Table 3 Table 3 (continued) Site velocities (mm/yr) relative to NAB (Fig. 4a). ID Longitude Latitude Vel (E) Vel (N) Sig (E) Sig (N) ID Longitude Latitude Vel (E) Vel (N) Sig (E) Sig (N) PDNS −79.990 0.110 11.4 1.5 0.6 0.4 ACP1 −79.950 9.371 15.3 −2.4 0.3 0.3 POVA −76.615 2.449 2.7 −1.3 0.5 0.3 ACP6 −79.408 9.238 15.2 −1.8 0.2 0.2 PPRT −80.210 −0.120 12.6 3.1 0.5 0.5 AHUA −77.550 −1.060 −4.3 −3.8 0.7 0.4 PROG −80.360 −2.410 2.2 −0.3 0.5 0.3 ALPA −72.918 11.528 7.0 −1.1 0.9 0.6 PSTO −77.270 1.210 0.7 −0.3 0.5 0.4 AMAL −79.420 −4.580 −2.3 −5.5 0.5 0.3 PTGL −80.030 0.780 16.9 1.7 0.5 0.4 ARCA −70.750 7.080 −6.8 −3.3 0.5 0.5 PUEB −79.530 −1.550 2.9 −0.3 1.6 0.9 AUCA −76.880 −0.640 −5.6 −5.0 0.3 0.2 PUIN −67.903 3.851 −7.1 −4.5 0.5 0.3 AYAN −80.750 −1.980 5.5 0.7 1.0 0.8 PUYX −78.060 −1.500 −1.0 −4.5 1.7 0.7 BAAP −73.554 4.072 −7.3 −4.7 0.5 0.2 QUIL −77.291 1.394 2.5 0.1 0.8 0.3 BACO −75.692 9.402 11.6 −6.6 0.9 1.0 RIOP −78.650 −1.650 −2.7 −7.4 0.6 0.4 BAEZ −77.887 −0.459 −0.8 −3.6 0.3 0.2 RVRD −79.380 1.060 13.0 1.1 0.4 0.4 BALZ −79.900 −1.360 3.4 −0.4 1.2 0.9 SABA −80.220 −1.840 1.3 −0.6 1.2 0.8 BAME −74.565 4.236 −0.6 −0.3 0.7 0.2 SALN −80.990 −2.180 7.0 −0.5 0.6 0.4 BAPA −74.658 5.466 0.6 0.1 0.3 0.2 SAN0 −81.716 12.580 11.5 −6.5 0.4 0.2 BARU −75.590 10.258 10.1 −4.4 0.8 0.4 SEL1 −75.529 6.191 2.0 0.5 0.3 0.2 BASO −77.393 6.203 4.8 1.4 0.9 0.4 SNLR −78.840 1.290 8.8 −3.9 0.4 0.3 BOBG −73.358 8.312 5.0 0.5 0.5 0.3 SNTI −78.010 −3.040 −4.4 −6.0 1.2 1.3 BOGT −74.081 4.640 −2.6 0.6 0.3 0.2 SOZO −79.790 −4.330 −3.3 −5.4 0.8 0.4 BUEN −76.993 3.820 2.3 4.1 0.4 0.2 SRAM −79.560 −0.600 5.6 −1.2 2.0 1.5 BUGT −76.996 3.826 2.7 0.8 0.3 0.2 TICU −69.939 −4.187 −7.3 −4.2 0.3 0.3 CABP −80.420 −0.380 14.2 2.7 0.5 0.6 TONE −76.139 6.324 2.3 1.2 0.5 0.2 CALI −76.530 3.370 0.6 0.9 0.4 0.3 TOTO −78.670 −2.250 −2.1 −5.7 0.3 0.2 CAPI −72.428 5.351 −5.2 −3.3 0.3 0.4 TUCO −78.748 1.815 11.3 −1.1 0.4 0.2 CHIS −80.720 −1.050 8.8 2.5 0.4 0.3 TULC −77.700 0.810 −0.4 −0.4 0.4 0.3 CIA1 −76.357 3.505 0.7 0.0 1.3 0.6 URRA −76.210 8.012 10.8 −0.3 0.8 0.7 CN19 −70.049 12.612 11.3 −0.8 0.7 0.6 UWAS −72.391 6.451 −1.9 −1.8 0.3 0.3 CN28 −79.034 8.625 15.8 1.2 0.7 0.5 VBUV −73.859 5.533 1.6 1.0 0.5 0.3 CN33 −80.327 8.487 18.4 −7.4 1.8 1.8 VDPR −73.248 10.436 6.9 0.6 0.2 0.2 CN35 −81.363 13.375 10.3 −7.0 0.6 0.4 VMAG −74.847 9.287 6.8 −0.3 0.5 0.4 CN37 −75.263 10.793 7.1 −0.6 1.1 1.6 VNEI −75.255 3.062 −3.1 −0.6 0.3 0.3 CN38 −71.988 12.222 9.9 −1.0 0.5 0.3 VORA −76.722 7.818 12.4 −2.6 1.8 1.5 CN40 −68.958 12.180 10.6 −2.2 0.7 0.7 VOTU −74.710 7.019 4.1 0.7 0.3 0.4 CNJO −76.840 0.230 −4.8 −3.3 0.6 0.5 VPIJ −75.107 4.397 −0.7 0.0 0.4 0.2 COEC −77.787 0.716 0.4 −1.6 1.0 0.4 VPOL −74.861 10.794 6.1 2.2 0.5 0.2 CORO −75.288 9.328 10.1 −3.0 0.3 0.3 VROS −74.323 4.847 −2.1 0.0 0.5 0.3 CUC1 −72.513 7.932 5.6 −1.8 1.5 0.4 VSJG −72.639 2.533 −7.4 −3.9 0.6 0.3 CUEC −79.000 −2.880 −3.1 −5.8 0.4 0.3 VSJP −75.836 4.781 −1.2 0.8 0.4 0.3 CUER −79.530 −2.350 −2.3 −1.1 0.8 0.5 ZAMO −78.930 −4.050 −3.8 −4.8 0.3 0.3 CULA −78.690 0.140 4.8 0.6 1.3 1.1 ZHUD −79.000 −2.460 −1.7 −5.0 0.5 0.3 DAUL −79.990 −1.870 2.2 −0.2 0.7 0.4 DESV −79.920 −1.040 3.9 0.1 0.6 0.5 ELCH −77.800 −0.330 −3.2 −2.6 0.4 0.3 ESMR −79.720 0.930 14.2 2.3 0.5 0.3 above the subducting Nazca slab, down to a depth of 50 km. Such a FLFR −79.840 −0.350 7.9 −2.0 0.4 0.5 locking depth is probably too deep, as it would be substantially deeper FLOR −75.600 1.620 −5.7 −5.0 0.6 0.4 than the depth of slip in megathrust earthquakes observed in the GPH1 −79.910 −2.730 −1.0 −1.3 0.9 0.8 modern era. If the locking depth were reduced, the locking fraction on GUAP −77.895 2.574 5.5 −0.7 0.5 0.4 the shallower part of the interface would have to increase somewhat to GYEC −79.890 −2.140 0.1 0.3 0.5 0.4 HONA −79.150 −3.470 −3.2 −5.2 0.9 0.4 fit the data. HSPR −78.850 −0.350 5.3 −2.9 0.5 0.4 Just to the north, at Guapi (GUAP), the eastward component of HUAC −77.800 −0.700 −4.2 −3.4 0.6 0.6 velocity relative to the North Andes at the coast drops to ISPT −81.070 −1.260 28.4 1.0 0.4 0.3 5.5 ± 0.5 mm/yr. White et al. (2003) interpreted the reduction in JUJA −79.550 −1.890 3.4 0.1 1.1 0.6 LATA −78.620 −0.810 1.8 −1.7 0.2 0.2 apparent locking in southwest Colombia relative to northern Ecuador as LCOL −79.200 −0.240 5.2 −1.2 0.3 0.2 the result of viscoelastic relaxation in the lower crust following the LGCB −79.570 0.380 9.3 −0.7 0.5 0.3 1979 Mw 8.2 subduction earthquake. Inland and north of Guapi, our LIMO −76.620 −0.400 −6.9 −3.3 1.2 0.9 velocities have a substantially higher eastward component than the LITS −78.440 0.870 6.1 −0.8 0.3 0.3 1990s campaign velocities of Trenkamp et al. (2002). Although no data LJEC −79.190 −3.980 −3.1 −5.7 1.5 0.5 LORO −75.980 −1.610 −8.2 −1.6 0.6 0.5 from Guapi were available at that time, the model of White et al. (2003) MACH −79.960 −3.250 −1.9 −4.1 0.7 0.4 predicted a substantially smaller velocity than the velocity at Tumaco. MALO −81.606 4.003 46.1 0.1 0.6 0.3 Our current data set is mainly from the last few years while the velo- MECE −73.712 7.107 2.5 0.4 0.3 0.4 cities of Trenkamp et al. (2002) were from the mid-1990s; the time MITU −70.232 1.261 −7.4 −3.4 0.4 0.3 MOCA −79.500 −1.180 2.9 −0.6 0.5 0.4 difference is approaching the characteristic relaxation time ofthe White MONT −76.980 −2.060 −5.8 −3.4 0.6 0.5 et al. (2003) model. Thus we might expect the viscoelastic component MZAL −75.471 5.030 −0.5 3.5 0.3 0.7 to be reduced by a factor of 2–2.5 compared to the effect on the 1990s NARI −79.530 −3.140 −4.0 −5.2 0.5 0.3 data. This change is probably too small to explain the velocity differ- NEVA −75.290 2.930 0.5 0.9 0.7 0.8 ence at Popayan (sites POVA and PPYN), but might explain the change OCEL −71.616 4.271 −7.3 −3.2 0.3 0.2 PAJA −80.420 −1.550 3.6 −1.6 1.0 0.7 at Buenaventura (sites BUGT and BUEN) and Cali (sites CIA1 and CALI). PAL1 −73.189 7.136 1.5 0.1 0.4 0.4 If this explanation is true, then the slower eastward motion at Guapi PAPA −78.140 −0.380 −0.4 −0.5 0.5 0.4 probably represents a combination of decaying postseismic deformation PASI −76.499 0.513 −5.8 −3.9 0.3 0.4 along with an actual reduction in the extent of locking along the Co- lombian section of the margin.

84 H. Mora-Páez et al. Journal of South American Earth Sciences 89 (2019) 76–91

Fig. 5. (a) GPS margin parallel (035°) vectors relative to South America (Table 2). North Andes Block (NAB), Panama Block (PB); (b) GPS margin normal (125°) vectors relative to South America (Table 3). North Andean Block (NAB), Panama Block (PB).

3.3.2. Caribbean subduction relative to South America, Caribbean coastal sites still show large mo- San Andres and Providencia islands (SAN0 and CN35, two of the tions relative to the North Andes. CN38 and ALPA are moving eastward very few sites unequivocally located on the stable Caribbean plate) at 9.9 ± 0.5 mm/yr and 7.0 ± 0.9 mm/yr relative to the North Andes. obliquely converge east-southeastward with stable South America at The orientation of the vectors is very similar to the motion of sites in 18.2 ± 0.4 mm/yr and 17.0 ± 0.6 mm/yr respectively (Fig. 2) and Panamá relative to the North Andes. (Fig. 4a). southeastward with respect to NAB at 11.5 ± 0.4 mm/yr and 10.3 ± 0.5 mmyr respectively. Slow amagmatic Caribbean subduction 3.4. Arc-continent collision under the North Andes has been proposed by numerous authors based on a weak Wadati Benioff zone (e.g., Dewey, 1972; Kellogg and Bonini, The Panamá arc is rapidly colliding eastward with the North Andean 1982; Vargas and Mann, 2013; Bernal-Olaya et al., 2015), seismic to- block (NAB) at approximately 15–18 mm/yr. A result shown by these mographic evidence for a south-dipping, high velocity slab (van der new vector solutions is that the present deformation is probably asso- Hilst and Mann, 1994; van Benthem et al., 2013), seismic reflection ciated with the Panama arc collision and it is confined to the North profiles that show Caribbean acoustic basement underthrusting the Andes, north of latitude 7.5°N. VORA site (Atrato region, Colombia), for deformed belt (e.g., Silver et al., 1975; Ladd et al., 1984; Bernal-Olaya example, moves eastward at 12.4 ± 1.8 mm/yr relative to NAB. BASO et al., 2015), and plate motion models that require convergence site (Choco region, Colombia), however, just 200 km to the southwest, (Boschman et al., 2014; Kobayashi et al., 2014). moves eastward at only 4.8 ± 0.9 mm/yr also relative to NAB. The new GPS velocity vectors presented here suggest subduction- Geodetic evidence for active Panama-North Andes collision was re- related deformation in the overriding North Andean block. Even though ported by Mora-Páez (1995) and Kellogg and Vega (1995). Kobayashi we estimate a significant eastward motion of the North Andes block et al. (2014) interpreted the eastward motion of the Panama block as

85 H. Mora-Páez et al. Journal of South American Earth Sciences 89 (2019) 76–91

Fig. 5. (continued) tectonic escape from the subducting Cocos Ridge at the Middle America slip of the NAB relative to South America (15.0 ± 1.0 mm/yr) likely trench and modeled the Panama Block-NAB convergence as 12.2 mm/ reflects deformation related to the Panamá collision. yr to the southeast (124°). The area of this collision is also a zone of active seismicity. In 1992, 3.5. Margin-parallel “escape” two large shallow earthquakes occurred in the area (Fig. 4a, MS = 6.6 and 7.3) (Wallace and Beck, 1993), and the focal mechanisms are McCaffrey (1996) has shown that about half of all modern sub- consistent with compression normal to the Panamá-North Andes suture duction zones have mobile forearc blocks. Slip partitioning into margin- (Freymueller et al., 1993). A recent earthquake occurred on September parallel and margin-normal components within the overriding plate at 14, 2016, in the Mutatá region, Colombia, (Fig. 4a and 7.37°N oblique subduction zones frequently results in lithospheric blocks being 76.17°W), (Mw = 6.0, depth = 18 km) with a focal mechanism con- detached from the overriding plate. The forearc blocks are driven by sistent with northwest-southeast compression (USGS National plate coupling and are displaced relative to the overriding plate Earthquake Information Center, 2017). (McCaffrey, 2002). Since the Panamá–North Andes convergence zone involves the GPS vectors relative to stable South America (Fig. 2) can be parti- collision of two thick buoyant crustal blocks, the resulting deformation tioned into margin-parallel (035°, Fig. 5a and Table 2) and margin- is collision-like, unlike subduction zones where most of the convergence normal (125°) components (Fig. 5b and Table 2), assuming that the in the overriding plate is recoverable elastic strain associated with the average trend of the NAB-South America margin is 035° between 1° and earthquake cycle. Our new vectors are slightly smaller but generally 7°N latitude. At least ten sites in the North Andes (BAPA, BAME, CIA1, consistent with the Trenkamp et al. (2002) model for Panamá collision MECE,PAL1, POVA, QUIL, SEL1, VPIJ, VSJP) have margin-parallel related deformation in northern Colombia over a locked east-dipping vectors insignificantly different at the 95% confidence level from thrust fault zone. The Perez et al. (2018) estimate of relative eastward 8.6 ± 1.0 mm/yr, our estimate of the average motion of the North

86 H. Mora-Páez et al. Journal of South American Earth Sciences 89 (2019) 76–91

Andean block (8.6 mm/yr toward 060°, Fig. 4b). This NAB vector can be shortening and uplift of the Cordillera. Mora-Páez et al. (2016) inter- resolved into a margin-parallel (035°) component of 8.1 mm/yr and a preted the small GPS measured shortening as evidence for slow for- margin-normal component of 4.3 mm/yr. Mora-Páez et al. (2016) es- mation of the Eastern Cordillera over a period of up to 40 Myr. They timated right-lateral strike-slip shear along the northeast trending argue that the paleobotanical evidence for recent rapid uplift may re- Eastern Cordillera of 8.0 ± 1.7 mm/yr. Most of the northeastward present local uplift or may have been influenced by climate change or movement is accommodated along the broad East Andean Fault System invasive species from North America. In this paper, we present an al- (EAFS). Near Panamá and north of 7° N latitude the margin-parallel ternative explanation, a “broken indenter” model for the Panamá- rates increase up to 10.4 mm/yr (CORO, VMAG, VORA, VPOL), prob- Choco arc collision, that is consistent with the GPS data, paleobotanical, ably due to the Panamá arc–North Andes collision. radiometric, and geologic evidence. The “broken indenter model” is also compatible with GPS evidence for the accretion of the Choco arc to 3.6. Margin-normal mountain building the North Andes, as well as taking into account the collision of the Panama-Choco arc with the North Andes as a potential mechanism for The margin-normal (125°) components of the new GPS vectors re- crustal shortening and mountain building. lative to stable South America are shown in Fig. 5b and Table 2. The new velocity field for northwestern South America and the Margin-normal velocities within the North Andes vary from 0.9 to southwest Caribbean presented here records margin-normal shortening 16.6 mm/yr. Margin-normal components of velocity vary much more less than 4.1 mm/yr in the Eastern Cordillera of Colombia (based on 4 than do the margin-parallel components. Margin-parallel vectors are sites in this paper; 3.7 ± 0.3 mm/yr according to Mora-Páez et al., dominated by northeastward rigid NAB translation while margin- 2016). However, paleobotanical, fission-track, seismic reflection and normal vectors are dominated by the subduction earthquake cycle and well data from the range suggest rapid uplift (7 km) and shortening permanent deformation (mountain building). (120 km) in the last 10 Ma. This would imply an average rate of about Mountains are built by both viscous/plastic deformation and fault- 12 mm/yr for the last 10 Ma or about 3 times the present rate. related deformation. Compressive mountain belts, such as the Eastern The initial collision of the Panama-Choco arc and the North Andes Cordillera of Colombia, are primarily built by slip on reverse faults. occurred between 12 and 40 Ma (e.g., Coates et al., 2004; Barat et al., Earthquake focal mechanisms for the Eastern Cordillera are character- 2014; Montes et al., 2012). Montes et al. (2015) used cooling ages of ized by WNW-ESE compression on reverse faults (e.g., Taboada et al., magmas, U/Pb dating, paleomagnetic pole rotations, and Atlantic sea- 2000; Corredor, 2003; Cortés and Angelier, 2005). Both seismic and floor anomalies to propose that closure of the Isthmus of Panama oc- aseismic slip on the reverse faults produces horizontal shortening and curred at 15 Ma, with no space for trans-isthmian marine passages, vertical thickening (permanent deformation). The rate of mountain However, paleoceanographic studies show a decrease in the transport growth depends on the rate of shortening, the erosion rate, the dip of of deep and intermediate Pacific waters into the Caribbean by the thrust faults, and isostatic adjustments of the crust to the mountain 10–11 Ma, probably related to a closing Central American Seaway load. (Osborne et al., 2014; Sepulchre et al., 2014). Recorded changes in Using an elastic half-space model for a cross section across Ecuador, Caribbean water salinity (Haug et al., 2001) and the Great American Trenkamp et al. (2002) estimated 6 mm/yr permanent shortening on Biotic Interchange (Marshall et al., 1982) did not occur until 4.2 Ma and reverse faults in the Ecuadorian Andes along an east-west profile. Using 3.5 Ma. Coates and Stallard (2013) propose that the Indonesian Aus- a GPS permanent station in Ecuador (COEC) and two GPS stations in tralian Archipelago provides a model for the Panama Arc between 15 Colombia (QUIL and POVA), the margin-normal (125°) velocity com- and 3 Ma that accounts for the tectonic configuration, while also ac- ponents seem to confirm a shortening rate of 4.5–6 mm/yr, which re- counting for the marine fossil record, and the delayed Great American presents permanent deformation. Trenkamp et al. (2002) estimated the Biotic Interchange. far field component of eastward recoverable elastic displacement from About 12–15 Ma the Eastern Cordillera of Colombia became a major the trench as only about 2 mm/yr in the Ecuadorian Andes. To the sediment source (Gregory-Wodzicki, 2000; Mora et al., 2014), although north, in Colombia's Eastern Cordillera (BAME, UWAS, VBUV and some sediments were being supplied to the Eastern Foothills area as VROS), the shortening rates are less than 4.1 mm/yr. The Eastern early as the Oligocene (Mora et al., 2010a; Saylor et al., 2012). Pa- Cordillera of Colombia is 400–600 km from the trench, and the esti- leobotanical data (Fig. 6a; Wijninga, 1996; Gregory-Wodzicki, 2000) mated recoverable elastic trench related velocity is less than 1 mm/yr. and apatite fission-track ages (Mora et al., 2008; Mora et al., 2010b) Most of the observed shortening therefore is building the mountain by indicate that most of the uplift in the central and eastern flank of the either viscous/plastic or fault-related deformation. Explanations for the Eastern Cordillera occurred in the last 12 Ma. Paleo-precipitation data apparent discrepancy between this small margin-normal shortening and from the Upper Magdalena Valley indicate that a substantial orographic evidence for recent rapid uplift of the Eastern Cordillera are presented barrier was not fully established until 6–3 Ma, when > 1 km/m.y. of in the following Discussion section. material was exhumed (Anderson et al., 2016). Based on minimum Rapid margin-normal shortening of 7–17 mm/yr in northern exhumational ages provided by new apatite FTA data, as well as Colombia north of 7° N reflects some combination of elastic strain from thermal history modeling, Anderson et al. (2016) suggest that thrust- the subducting Caribbean plate, mountain building and faulting across induced rapid exhumation of the Garzón Massif and southern Eastern the broad plate boundary, including the Western and Central cordilleras Cordillera was focused between 6.4 Ma and 3 Ma. Tectonic uplift of Colombia and the Merida Andes of Venezuela (e.g., Kellogg and (Fig. 6a, after Egbue et al., 2014) is estimated from present-day total Bonini, 1982; Audemard and Audemard, 2002). We do not model the uplift and paleoelevation data. Elevation is shown in black with present deformation in this area in this paper. The permanent shortening in mean elevation just over 3 km and estimated paleoelevation based on northern Colombia and Venezuela is probably caused by the Panama paleobotanical and geomorphological data (black squares with error collision and Caribbean subduction. bars after Wijninga, 1996 and Gregory-Wodzicki, 2000). Tectonic uplift is estimated using present day total structural relief (10 km) and as- 4. Discussion suming erosion rates of 0.5 and 1.0 mm/a. Vertical velocities (mm/a) were measured at 4 of the CORS GPS sites in the Eastern Cordillera 4.1. Panamá Arc-North Andes collision (Fig. 2), BAME, 0.0 ± 0.9, UWAS, 1.3 ± 0.6, VBUV, 0.4 ± 0.9, and VROS, −1.1 ± 1.0. The mean GPS observed uplift rate (0.1 ± 0.9 There is an apparent discrepancy between the small GPS measured mm/a) is less than the maximum slopes of our predicted tectonic uplift margin-normal shortening in the Eastern Cordillera of Colombia and curves for the last 6 Ma (0.75–1.0 mm/a, Fig. 6a). However, the un- paleobotanical, radiometric, and geologic evidence for recent rapid certainties in the vertical measurements remain too large to test

87 H. Mora-Páez et al. Journal of South American Earth Sciences 89 (2019) 76–91

also shown for 60 km shortening, the minimum published estimates for the Eastern Cordillera (0.8 × 68 km; Cooper et al., 1995; Teixell et al., 2015), where k = 9. Northeastward escape rates for the North Andean block (Fig. 6b) are from the compilation by Egbue and Kellogg (2010). The present-day Panama-South America convergence rate is approxi- mately 23 mm/yr (Fig. 2 and Table 2), measured on three of the GPS stations (ACP1, ACP6, CN28), and 25 mm/yr at the CN36. Note that range-parallel shear “escape” has become a significant factor in the last 2 Ma, and range-normal shortening rates have apparently declined abruptly. The apparent dichotomy of rapid Late Miocene-Pliocene shortening in the Eastern Cordillera and very slow range normal shortening at present (new GPS results) can be explained by a “broken indenter” model (Fig. 7). The rigid Panama-Choco collision with the North Andes (15–18 mm/yr at present) produced rapid permanent deformation in the North Andes, especially after the closure of the Central American Seaway and formation of the land bridge in the last 10 Ma (Fig. 7b). The present-day GPS vectors at BASO and TONE (Figs. 2 and 4a and Table 2), however, suggest that the Choco block is no longer part of the rigid Panama indenter but has been accreted to the NAB (Fig. 7c). The small residual vectors at Colombian coastal sites BASO (4.8 ± 0.9 mm/ yr), BUGT (2.7 ± 0.4 mm/yr) and GUAP (5.5 ± 0.7 mm/yr), relative to the North Andean Block, could be interpreted as Nazca subduction- related elastic locking. The high levels of shallow seismicity near the Panama-Colombia border area (e.g., Ammon et al., 1994; Trenkamp et al., 2002; Cortés and Angelier, 2005; Vargas and Mann, 2013) are also consistent with a recent detachment zone between the Panama and Choco blocks. Examples of active arc-continent collisions include the Banda Arc-Australia continent collision, where GPS measurements re- veal that large sections of the SE Asian plate are progressively accreting to the edge of the Australian continent (Nugroho et al., 2009) and the Fig. 6. Uplift and shortening of the E Cordillera (after Egbue et al., 2014) a) arc-continent collision in Papua New Guinea where GPS results show Uplift history 12 Ma to Present estimated from total tectonic uplift (10 km) and that New Guinea has been fragmented into a complex array of micro- paleoelevation data (black squares with error bars after Wijninga, 1996; plates (Wallace et al., 2004). Gregory-Wodzicki, 2000) for erosion rates of 0.5 and 1.0 mm/yr. GPS uplift rate A 250 km offset in Wadati-Benioff zone seismicity at 5.5° Nor (0.0 ± 1.0 mm/a, red lines) can be compared to our estimated uplift rate “Caldas Tear” (Vargas and Mann, 2013) separates “normal” subduction (slope of tectonic uplift curve). b) Shortening rates for 60 and 120 km short- to the south from flat-slab subduction to the north. The slab offset has ening are estimated from uplift rates. The red square at time 0 represents the been interpreted as a tear in the subducting Nazca slab (e.g., Corredor, range-normal shortening GPS estimate from this study. The red line and black 2003; Syracuse et al., 2016) or the southern edge of the subducted squares with error bars are northeastward geologically measured “escape” rates Caribbean flat slab (e.g., Taboada et al., 2000; Yarce et al., 2014). compiled by Egbue and Kellogg (2010). (For interpretation of the references to However, the velocities relative to the NAB (Fig. 4a) presented with the colour in this figure legend, the reader is referred to the Web version ofthis new GPS data in this paper show no clear surficial velocity contrast article.) across the proposed “Caldas tear” at 5.5°N. The new velocity field presented in this paper highlights the in- unequivocally whether uplift rates are slowing at present. The Neogene crease in NAB deformation north of 7.5°N (Fig. 4a) related to present exhumation history from apatite fission track data and low-T thermo- day Panama-NAB convergence. If this interpretation is correct, the chronology (Mora et al., 2008; Parra et al., 2009) supports the rapid break in the Panama-Choco indenter and Choco accretion to the NAB tectonic uplift proposed for the last 6 Ma. Using apatite fission-track must have occurred very recently (in the last 1–2 Ma) to explain the data, Mora et al. (2010a) estimated exhumation rates of 1–1.5 mm/yr rapid Late Miocene-Pliocene shortening and uplift and the equally rapid for compressional structures in the eastern foothills of the Eastern slowdown in the margin-normal shortening rate. Cordillera over the last 3 Ma. Wittmann et al. (2011) used cosmogenic nuclide-based measurements to estimate denudation rates of 4.2. Early Pleistocene range-parallel shear 0.49–1.2 mm/yr for the Ecuadorian Andes. Note that the uplift curve for an erosion rate of 0.5 mm/yr shows tectonic uplift in the Eastern By the early Pleistocene (Fig. 6b), in addition to the compressional Cordillera distributed over at least 10 Ma. stress regime, a NE-SW strike-slip component was introduced as the The shortening rates for the Eastern Cordillera (Fig. 6b; Egbue and northern Andes began to “escape”. Egbue and Kellogg (2010) compiled Kellogg, 2012) since 12 Ma (the initiation of the Panama-South America field geologic estimates of northeastward displacement rates forthe collision) are estimated from the tectonic uplift rates (Fig. 6a) and North Andes with a mean estimated geologic slip rate for the last constrained to fit the maximum present margin-normal GPS measured 86,000 years of 7.6 mm/yr. The earliest measurements date back to the rates in the Eastern Cordillera relative to stable South America (BAME, opening of the Gulf of Guayaquil at 1.8 Ma, and the northeastward UWAS, VBUV, VROS), less than 4.1 mm/yr (Table 2). displacement of the North Andes has been interpreted as tectonic es- Shortening rates (v) are assumed to be proportional to tectonic cape from the Carnegie Ridge subducting at the Ecuador trench (Egbue uplift rates (u): v = ku, where k is a dimensionless proportionality and Kellogg, 2010; Nocquet et al., 2014; Chlieh et al., 2014). About constant. Shortening rates are shown for a total shortening of 120 km in 2 Ma, the aseismic Carnegie ridge, which was formed by the Galapagos the last 12 Ma (0.8 × 150 km; Dengo and Covey, 1993; Egbue and hotspot, arrived at the Colombia-Ecuador trench (Lonsdale and Kellogg, 2012); for 120 km shortening, k = 18. Shortening rates are Klitgord, 1978; Pedoja, 2003; Cantalamessa and Di Celma, 2004), and

88 H. Mora-Páez et al. Journal of South American Earth Sciences 89 (2019) 76–91

Fig. 7. Schematic evolution of Panama - North Andes. a) 12–15 Ma. initial collision of Panama-Choco arc and North Andes. b) 6 to 1 Ma. Panama-Choco N Andes collision. Permanent shortening and rapid uplift of E. Cordillera. c) Present. Choco arc is accreted to the North Andes. Upper Magdalena Valley (UMV), Eastern Cordillera (EC), Garzón Massif (GM). initiated the northeastward “escape” of the northern Andes (Egbue and The new velocity field for northwestern South America and south- Kellogg, 2010). Presently, in the Eastern Cordillera the escape rate west Caribbean presented here indicates margin-normal shortening of (8.1 mm/yr dextral slip on the EAFZ) is greater than the rate of range- less than 4.1 mm/yr in the Eastern Cordillera of Colombia. However, normal shortening (4.3 mm/yr). paleobotanical, fission-track, seismic reflection and well data fromthe range suggest rapid uplift (7 km) and shortening (120 km) in the last 5. Conclusions 10 Ma. This would imply an average rate of about 12 mm/yr for the last 10 Ma or about 3 times the present rate. This apparent discrepancy may This paper presents the first comprehensive model of North Andean be explained with the “broken indenter” model for the Panama-Choco block (NAB) motion. Previous estimates by Nocquet et al. (2014) and arc presented in this paper, in which the Choco arc has been recently Chlieh et al. (2014) were based on data from a small area (mainly in accreted to the NAB. The “broken indenter” model is an alternative to Ecuador, plus 2 sites in Colombia) and thus may not be appropriate to the Mora-Páez et al. (2016) interpretation of the small GPS measured extrapolate farther to the north. In northern Colombia, our predicted shortening as evidence for slow uplift of the Eastern Cordillera over a eastward NAB motion is 4.2 mm/yr faster relative to the prediction of period of up to 40 Myr. Additional GPS measurements are needed to Nocquet et al. (2014). The Perez et al. (2018) estimate of relative confirm the accretion of the Choco arc to the NAB. The Panama arcis eastward slip of the NAB relative to South America (15.0 ± 1.0 mm/ rapidly colliding eastward with the NAB at approximately 15–18 mm/ yr), based only on sites located north of 8.5°N, likely reflects de- yr. The Panama-Choco collision was likely responsible for much of the formation related to the Panamá collision, so our NAB velocity esti- uplift of the Eastern Cordillera. The present on-going collision poses a mates differed significantly. We estimate that the NAB is moving tothe major earthquake hazard from the Panama border to Medellin, Co- northeast (060°) at a rate of 8.6 mm/yr. lombia (SEL1, Fig. 2). Caribbean subduction under the NAB at about

89 H. Mora-Páez et al. Journal of South American Earth Sciences 89 (2019) 76–91

13 mm/yr poses an additional seismic hazard in northern Colombia. Petroleum Geology and Potential of the Colombian Caribbean Margin. AAPG Memoir The Colombia section of the Nazca-NAB trench continues to pose 108, pp. 247–270 ISBN 13: 978-0-89181-388-0. Bertiger, W., Desai, S., Haines, B., Harvey, N., Moore, A., Owen, S., Weiss, J., 2010. Single high risk of a great mega-subduction earthquake in southern Colombia. receiver phase Ambiguity resolution with GPS data. J. Geodes. https://doi.org/10. The 1942, 1958, 1979, and 2016 trench earthquakes have only released 1007/s00190-010-0371-9. a fraction of the energy accumulated in the Ecuador-Colombia trench Blewitt, G., Lavallée, D., 2002. Bias in geodetic site velocity due to annual signals: theory and assessment. In: Ádám, J., Schwarz, K.P. (Eds.), Vistas for Geodesy in the New since the great 1906 earthquake. Interseismic strain is accumulating Millennium. International Association of Geodesy Symposia. 125. Springer, Berlin, rapidly in the overriding plate at least as far north as Tumaco. Southern Heidelberg. Colombian coastal sites are moving more rapidly inland as the visco- Bos, M.S., Bastos, L., Fernandes, R.M.S., 2010. The influence of seasonal signals on the elastic mantle flow effects of the 1979 earthquake gradually decay. estimation of the tectonic motion in short continuous GPS time-series. J. Geodyn. 49 (3–4), 205–209. https://doi.org/10.1016/j.jog.2009.10.005. 2010. Residual motions within the NAB have important implications for Bos, M.S., Fernandes, R.M.S., Williams, S.D.P., Bastos, L., 2013. Fast error analysis of Nazca and Caribbean subduction earthquake cycles as well as for continuous GNSS observations with missing data. J. Geodyn. 87 (4), 351–360. mountain building in the Eastern Cordillera of Colombia. Boschman, L.M., van Hinsbergen, D.J.J., Torsvik, T.H., Spakman, W., Pindell, J.L., 2014. Kinematic reconstruction of the caribbean region since the early jurassic. Earth Sci. Presently, in the Eastern Cordillera the northeastward margin-par- Rev. 138, 102–136. https://doi.org/10.1016/j.earscirev.2014.08.007. allel “escape” rate (8.1 mm/yr) is greater than the rate of range-normal Braun, J., Mattioli, G., Calais, E., Carlson, D., Dixon, T.H., Jackson, M., Kursinski, R., shortening (4.3 mm/yr). Therefore, northeast trending right-lateral Mora-Páez, H., Miller, M., Pandya, R., Robertson, R., Wang, G., 2012. Focused study of interweaving hazards across the caribbean. Eos 93 (9), 89–90. strike-slip faulting is an increasing component of the seismic hazard for Cantalamessa, G., Di Celma, C., 2004. Origin and chronology of Pleistocene marine ter- the Eastern Cordillera and the 8 million inhabitants of the city of races of Isla de la Plata and of flat, gently dipping surfaces of the southern coastof Bogota. Cabo San Lorenzo (Manabí, Ecuador). J. S. Am. Earth Sci. 16, 633–648. Chlieh, M., Mothes, P., Nocquet, J.M., Jarrin, P., Charvis, P., Cisneros, D., Font, Y., Collot, Finally, we note that the current spatial density of coastal sites in J.Y., Villegas-Lanza, J.-C., Rolandone, F., Vallée, M., Regnier, M., Segovia, M., Colombia does not allow for a detailed image of elastic strain accu- Martin, X., Yepes, H., 2014. Distribution of discrete seismic asperities and aseismic mulation and strain partitioning; however, maturing time series of new slip along the Ecuadorian megathrust. Earth Planet Sci. Lett. 400, 292–301. https:// doi.org/10.1016/j.epsl.2014.05.027. sites will allow for a more detailed study in the future. Coates, A.G., Collins, L.S., Aubry, M.P., Berggren, W.A., 2004. The geology of the Darien, Panama, and the late Miocene-Pliocene collision of the Panama arc with north- Acknowledgements western South America. Geol. Soc. Am. Bull. 116, 1327–1344. https://doi.org/10. 1130/B25275.1. Coates, A., Stallard, R., 2013. How old is the Isthmus of Panama? Bull. Mar. Sci. 89 (4), The paper was improved by constructive comments from two 801–813. anonymous reviewers. We thank the Colombian Geological Survey for Cooper, M.A., Addison, F.T., Alvarez, R.M., Graham, R.H., Hayward, A.B., Howe, S., providing the GPS results. We also thank all of the GeoRED team for Martinez, J., Naar, J., Peñas, R., Pulham, A.J., Taborda, A., 1995. Basin development and tectonic history of the Llanos basin, eastern cordillera, and Middle Magdalena their support in installing sites and measuring positions. This research Valley, Colombia. AAPG (Am. Assoc. Pet. Geol.) Bull. 79 (10), 1421–1443. was supported by the government of the Republic of Colombia who Corredor, F., 2003. Seismic strain rates and distributed continental deformation in the provided the funds to the Colombian Geological Survey for the im- northern Andes and three-dimensional seismotectonics of the northwestern South America. Tectonophysics 372, 147–166. plementation of the GeoRED Project, identified with the BPIN code Cortés, M., Angelier, J., 2005. Current states of stress in the northern Andes as indicated 0043000220000 of the National Planning Department. We thank the by focal mechanisms of earthquakes. Tectonophysics 403, 29–58. Geophysics Institute of the National Polytechnic University (Escuela Dengo, C., Covey, M., 1993. Structure of the Eastern Cordillera of Colombia: implications for trap styles and regional tectonics. AAPG (Am. Assoc. Pet. Geol.) Bull. 77 (8), Politécnica Nacional) for providing data from some Ecuadorean stations 1315–1337. and also to UNAVCO for providing GPS data as part of the COCONet Dewey, J.W., 1972. Seismicity and tectonics of western Venezuela. Bull. Seismol. Soc. network. Am. 62, 1711–1751. Egbue, O., Kellogg, J., 2010. Pleistocene to present North Andean “escape”. Tectonophysics 489, 248–257. https://doi.org/10.1016/j.tecto.2010.04.021. Appendix A. Supplementary data Egbue, O., Kellogg, J., 2012. Three-dimensional structural evolution and kinematics of the PiedemonteLlanero, central Llanos foothills, eastern cordillera, Colombia. J. S. Supplementary data to this article can be found online at https:// Am. Earth Sci. https://doi.org/10.1016/j.jsames.2012.04.012. Egbue, O., Kellogg, J., Aguirre, H., Torres, C., 2014. Evolution of the stress and strain doi.org/10.1016/j.jsames.2018.11.002. fields in the Eastern Cordillera, Colombia. J. Struct. Geol. 58,8–21. Fernandes, R.M.S., Mora-Páez, H., LaFemina, P., Bos, M., 2016. Present-day GPS velocity References field of South America. In: Proceedings IASPEI Regional Assembly Latin-American Seismological Comission-LACSC, Seismology for Science and Science for Society, Abstract O190, p. 118, San José de Costa Rica, June 20-22, 2016. Altamimi, Z., Rebischung, P., Métivier, L., Collilieux, X., 2016. ITRF2014: a new release of Fernandes, R.M.S., Bos, M., Alothman, A., Mora-Páez, H., 2017. Comparing global and the International Terrestrial Reference Frame modeling nonlinear station motions. J. dedicated plate Angular velocity models: the cases of arabia And south America, Geophys. Res. Solid Earth 121, 6109–6131. https://doi.org/10.1002/2016JB013098. abstract G04-4-02, IAG symposia, G04. In: Earth Rotation and Geodynamics, IAG- Altamimi, Z., Collilieux, X., Métivier, L., 2011. ITRF2008: an improved solution of the IASPEI Joint Meeting. Kobe International Center, Kobe, Japan July30-August 4, international terrestrial reference frame. J. Geodyn. 85 457–73. https://doi.org/10. 2017. 1007/s00190-011-0444-4. Freymueller, J.T., Kellogg, J.N., Vega, V., 1993. Plate motions in the north Andean region. Ammon, C.J., Lay, T., Velasco, A.A., Vidale, J.E., 1994. Routine estimation of earthquake J. Geophys. Res. 98, 21853–21863. source complexity; the 18 October 1992 Colombian earthquake. Bull. Seismol. Soc. Gregory-Wodzicki, K.M., 2000. Andean paleoelevation estimates: a review and critique. Am. 84, 1266–1271. GSA Bulletin 112, 1091–1105. Anderson, V.J., Horton, B.K., Saylor, J.E., Mora, A., Tesón, E., Breecker, D.O., Ketcham, Haug, G., Tiedemann, R., Zahn, R., Ravelo, A.C., 2001. Role of the Panama uplift on R.A., 2016. Andean topographic growth and basement uplift in southern Colombia: oceanic freshwater balance. Geology 29, 207–210. https://doi.org/10.1130/0091- implications for the evolution of the Magdalena, Orinoco, and Amazon river systems. 7613(2001)029<0207:ROPUOO>2.0CO;2. Geosphere 12 (4), 1235–1256. https://doi.org/10.1130/GES01294.1. Kellogg, J.N., Bonini, W.E., 1982. Subduction of the Caribbean Plate and basement uplifts Argus, D.F., Gordon, R.G., Heflin, M.B., Ma, C., Eanes, R.J., Willis, P., Peltier, W.R., Owen, in the overriding South America plate. Tectonics 1 (3), 251–276. S.E., 2010. The angular velocities of the plates and the velocity of Earth's centre from Kellogg, J.N., Freymueller, J.T., Dixon, T.H., Neilan, R.E., Ropain, C.U., Camargo, S.M., space geodesy. Geophys. J. Int. 180, 913–960. https://doi.org/10.1111/j.1365-246X. Fernandez, B., Stowell, J.L., Salazar, A., Mora, J., Espin, L., Perdue, V., Leos, L., 1990. 2009.04463.x. First GPS baseline results from the north Andes, CASA UNO special issue. Geophys. Audemard, F.E., Audemard, F.A., 2002. Structure of the Mérida Andes, Venezuela: rela- Res. Lett. 17, 211–214. tions with the south America–caribbean geodynamic interaction. Tectonophysics Kellogg, J.N., Vega, V., 1995. Tectonic development of Panama, Costa Rica, and the co- 345, 299–327. lombian Andes: constraints from global positioning system geodesy studies and Barat, F., de Lépinay, B.M., Sosson, M., Müller, C., Baumgartner, P.O., Baumgartner-Mora, gravity. In: Mann, P. (Ed.), Geologic and Tectonics Development of the Caribbean C., 2014. Transition from the farallon plate subduction to the collision between south Plate Boundary in Southern Central America. Geological Society of America Special and Central America: geological evolution of the Panama Isthmus. Tectonophysics Paper 295, Boulder, Colorado. 622, 145–167. https://doi.org/10.1016/j.tecto.2014.03.008. Kobayashi, D., LaFemina, P., Geirsson, H., Chichaco, E., Abrego, A., Mora, H., Camacho, Bernal-Olaya, R., Mann, P., Vargas, C.A., 2015. Earthquake, tomographic, seismic re- E., 2014. Kinematics of the western caribbean collision of the Cocos Ridge and upper flection, and gravity evidence for a shallowly dipping subduction zone beneath the plate deformation. G-cubed 1671–1683. https://doi.org/10.1002/2014GC005234. Caribbean margin of northwestern Colombia. In: Bartolini, C., Mann, P. (Eds.), Ladd, J.W., Truchan, M., Talwani, M., Stoffa, P.L., Buhl, P., Houtz, R., Mauffret, A.,

90 H. Mora-Páez et al. Journal of South American Earth Sciences 89 (2019) 76–91

Westbrook, G.K., 1984. Seismic reflection profiles across the southern margin ofthe Pérez, O.J., Wesnousky, S.G., De La Rosa, R., Marquez, J., Uzcátegui, R., Quintero, C., Caribbean. In: Bonini, W.E., Hargraves, R.B., Shagam, R. (Eds.), The Caribbean-south Liberal, L., Mora-Páez, H., Szeliga, W., 2018. On the interaction of the North Andes American Plate Boundary and Regional Tectonics. Memoir - Geological Society of plate with the Caribbean and South American plates in northwestern South America America, pp. 153–159. from GPS geodesy and seismic data. Geophys. J. Int. 214 (3), 1986–2001. https:// Lonsdale, P., Klitgord, K.D., 1978. Structure and tectonic history of the eastern Panama academic.oup.com/gji/article/214/3/1986/5035819. Basin. GSA. Bulletin. 89, 981–999. Saylor, J., Horton, B., Stockli, D., Mora, A., Corredor, J., 2012. Structural and thermo- Marshall, L.G., Webb, S.D., Sepkoski, J.J., Raup, D.M., 1982. Mammalian evolution and chronological evidence for Paleogene basement-involved shortening in the axial the great american Interchange. Science 215 (4538), 1351–1357. Eastern Cordillera, Colombia. J. S. Am. Earth Sci. 39, 202–215. McCaffrey, R., 1996. Estimates of modern arc-parallel strain rates in fore arcs. Geology Sella, G.F., 2002. REVEL: a model for Recent plate velocities from space geodesy. J. 24, 27–30. Geophys. Res. 107. McCaffrey, R., 2002. Crustal block rotations and plate coupling. In: In: Stein,S., Sepulchre, P., Arsouze, T., Donnadieu, Y., Dutay, J.C., Jaramillo, C., Le Bras, J., Martin, Freymueller, J. (Eds.), Plate Boundary Zones 30. AGU Geodynamics Series, pp. E., Montes, C., Waite, A.J., 2014. Consequences of shoaling of the Central American 101–122. Seaway determined from modeling Nd isotopes. Paleoceanography 29, 176–189. Montes, C., Bayona, G., Cardona, A., Buchs, D.M., Silva, C.A., Morón, S., Hoyos, N., https://doi.org/10.1002/2013PA002501. Ramírez, D.A., Jaramillo, C.A., Valencia, V., 2012. Arc-continent collision and oro- Silver, E.A., Case, J.E., MacGillavry, H.J., 1975. Geophysical study of the Venezuelan cline formation: closing of the Central American seaway. J. Geophys. Res. 117, borderland. Geol. Soc. Am. Bull. 86, 213–226. B04105. https://doi.org/10.1029/2011JB008959. 2012. Syracuse, E.M., Maceira, M., Prieto, G.A., Zhang, H., Ammon, C.J., 2016. Multiple plates Montes, C., Cardona, A., Jaramillo, C., Pardo, A., Silva, J.C., Valencia, V., Ayala, C., Perez- subducting beneath Colombia, as illuminated by seismicity and velocity from the Angel, L.C., Rodríguez-Parra, L.A., Ramírez, V., Nino, H., 2015. Middle Miocene joint inversion of seismic and gravity data. Earth Planet Sci. Lett. 444, 139–149. closure of the central american seaway. Science 348 (6231), 226–229. Taboada, A., et al., 2000. Geodynamics of the northern Andes: subductions and in- Mora, A., Parra, M., Strecker, M.R., Sobel, E.R., Hooghiemstra, H., Torres, V., Vallejo tracontinental deformation (Colombia). Tectonics 19, 787–813. Jaramillo, J., 2008. Climatic forcing of asymmetric orogenic evolution in the Eastern Teixell, A., Teson, E., Ruiz, J.C., Mora, A., 2015. The structure of an inverted back-arc rift: Cordillera of Colombia. GSA Bulletin 120 (7/8), 930–949. insights from a transect across the Eastern Cordillera of Colombia near Bogota. In: In: Mora, A., Horton, B.K., Mesa, A., Rubiano, J., Ketcham, R.A., Parra, M., Blanco, V., Bartolini, C., Mann, P. (Eds.), Petroleum Geology and Potential of the Colombian Garcia, D., Stockli, D.F., 2010a. Migration of Cenozoic deformation in the Eastern Caribbean Margin 108. AAPG Memoir, pp. 499–516. Cordillera of Colombia interpreted from fission track results and structural re- Trenkamp, R., Kellogg, J.N., Freymuller, J.T., Mora, H., 2002. Wide plate margin de- lationships: implications for petroleum systems. AAPG (Am. Assoc. Pet. Geol.) Bull. formation, South Central America and Northwestern South America, CASA GPS ob- 94 (10), 1543–1580. https://doi.org/10.1306/01051009111. servations. J. S. Am. Earth Sci. 15, 157–171. Mora, A., Parra, M., Strecker, M.R., Sobel, E.R., Zeilinger, G., Jaramillo, C., Da Silva, S., USGS, September 14, 2016, Mutata Earthquake, Colombia, https://earthquake.usgs.gov/ Blanco, M., 2010b. The eastern foothills of the Eastern Cordillera of Colombia: an earthquakes/eventpage/us10006pdp#executive (accessed 07 November 2017). example of multiple factors controlling structural styles and active tectonics. Geol. Vallée, M., Nocquet, J.M., Battaglia, J., Font, Y., Segovia, M., Réhnier, M., Mothes, P., Soc. Am. Bull. 112, 1846–1864. https://doi.org/10.1130/B30033.1. Jarrin, P., Cisneros, D., Vaca, S., Yepes, H., Martin, X., Béthoux, N., Chlieh, M., 2013. Mora, A., Ketcham, R.A., Higuera-Diaz, I.C., Bookhagen, B., Jimenez, L., Rubiano, 2014. Intense interface seismicity triggered by a shallow slip event in the Central Ecuador formation of passive-roof duplexes in the colombian subandes and Peru. Lithosphere subduction zone. J. Geophys. Res. Solid Earth. 118, 2965–2981. https://doi.org/10. 6 (6), 456–472. 1002/jgrb.50216. Mora-Páez, H., 1995. Central and South America GPS Geodesy: Relative Plate Motions van Benthem, S.A.C., Govers, R., 2010. The Caribbean plate: pulled, pushed, or dragged? Determined from 1991 and 1994 Measurements in Colombia, Costa Rica, Ecuador, J. Geophys. Res. 115, B10409. https://doi.org/10.1029/2009JB006950. Panama and Venezuela. M.Sc. Thesis. University of South Carolina, Columbia, van Benthem, S.A.C., Govers, R., Spakman, W., WortelM, J.R., 2013. Tectonic evolution pp. 94. and the mantle structure under the Caribbean region. J. Geophys. Res. Solid Earth. Mora-Páez, H., 2006. National GPS Network for Geodynamics Research, BPIN Document 118, 3019–3036. https://doi.org/10.1002/jgrb.50235. and Supplementary Information, Project Proposal (in Spanish). Approved by van der Hilst, R., Mann, P., 1994. Tectonic implications of tomographic images of sub- National Planning Department. INGEOMINAS, Colombia, pp. 63 + 5 modules. ducted lithosphere beneath northwestern South America. Geology 22, 451–454. Mora-Páez, H., Mencin, D.J., Molnar, P., Diederix, H., Cardona-Piedrahita, L., Peláez- Vargas, C.A., Mann, P., 2013. Tearing and breaking off of subducted slabs as the result of Gaviria, J.R., Corchuelo-Cuervo, Y., 2016. GPS velocities and the construction of the collision of the Panama arc-indenter with northwestern South America. Bull. Seismol. eastern cordillera of the colombian Andes. Geophys. Res. Lett. 43, 8407–8416. Soc. Am. 103 (3), 2025–2046. https://doi.org/10.1785/0120120328. https://doi.org/10.1002/2016GL069795. Wallace, L.M., Stevens, C., Silver, E., McCaffrey, R., Loratung, W., Hasiata, S., Stanaway, Mora-Páez, H., Peláez-Gaviria, J.R., Diederix, H., Bohórquez-Orozco, O., Cardona- R., Curley, R., Rosa, R., Taugaloidi, J., 2004. GPS and seismological constraints on Piedrahita, L., Corchuelo-Cuervo, Y., Ramírez-Cadena, J., Díaz-Mila, F., 2018. Space active tectonics and arc-continent collision in Papua New Guinea: implications for geodesy infrastructure in Colombia. Seismol Res. Lett. 89 (2A), 446–452. https://doi. mechanics of microplate rotations in a plate boundary zone. J. Geophys. Res. 109, org/10.1785/0220170185. B05404. https://doi.org/10.1029/2003JB002481. Mothes, P.A., Nocquet, J.M., Jarrín, P., 2013. Continuous GPS network operating Wallace, T.C., Beck, S.L., 1993. The Oct. 17-18, 1992 Colombian earthquakes; slip par- throughout Ecuador. Eos Trans. AGU 94 (26), 229. titioning or faulting complexity. Seismol Res. Lett. 64, 29. Nocquet, J., Villegas-Lanza, J.C., Chlieh, M., Mothes, P.A., Rolandone, F., Jarrin, P., White, S.M., Trenkamp, R., Kellogg, J.N., 2003. Recent crustal deformation and the Cisneros, D., Alvarado, A., Audin, L., Bondoux, F., Martin, X., Font, Y., Régier, M., earthquake cycle along Ecuador-Colombia subduction zone. Earth Planet Sci. Lett. Vallée, M., Tran, T., Beauval, C., Maguiña-Mendoza, J.M., Martinez, W., Tavera, H., 216, 231–242. Yepes, H., 2014. Motion of continental slivers and creeping subduction in the Wijninga, V.M., 1996. Neogene ecology of the Salto de Tequendama site (2475 m alti- Northern Andes. Nat. Geosci. 7 (4), 287–291. https://doi.org/10.1038/NGEO2099. tude, Cordillera Oriental, Colombia): the paleobotanical record of montane and Nugroho, H., Harris, R., Lestariya, A.W., Maruf, B., 2009. Plate boundary reorganization lowland forests. Rev. Palaeobot. Palynol. 92, 97–156. in the active Banda Arc–continent collision: insights from new GPS measurements. Wittmann, H., von Blanckenburg, F., Guyot, J.L., Laraque, A., Bernal, C., Kubik, P.W., Tectonophysics 479, 52–65. https://doi.org/10.1016/j.tecto.2009.01.026. 2011. Sediment production and transport from in situ-produced cosmogenic 10Be Osborne, A.H., Newkirk, D.R., Groeneveld, J., Martin, E.E., Tiedemann, R., Frank, M., and river loads in the Napo River basin, an upper Amazon tributary of Ecuador and 2014. The seawater neodymium and lead isotope record of the final stages of Central Peru. J. S. Am. Earth Sci. 31 (1), 45–53. https://doi.org/10.1016/j.jsames.2010.09. American Seaway closure. Paleoceanography 29, 715–729. https://doi.org/10.1002/ 004. 2014PA002676. Yarce, Y., Monsalve, G., Becker, T.W., Cardona, A., Poveda, E., Alvira, D., Ordoñez- Parra, M., Mora, A., Sobel, E.R., Strecker, M.R., González, R., 2009. Episodic orogenic- Carmona, O., 2014. Seismological observations in Northwestern South America: front migration in the northern Andes: constraints from low-temperature thermo- evidence for two subduction segments, contrasting crustal thicknesses and upper chronology in the Eastern Cordillera, Colombia. Tectonics 28, TC4004. https://doi. mantle flow. Tectonophysics 637, 57–67. org/10.1029/2008TC002423. Zumberge, J.F., Heflin, M.B., Jefferson, D.C., Watkins, M.M., Webb, F.H., 1997. Precise Pedoja, K., 2003. Les terrasses marines de la marge Nord Andine (Equateur et Nord Péou): point positioning for the efficient and robust analysis of GPS data from large net- Relations avec le context géodynamique. These de Doctorat de l´Universite P. et M. works. J. Geophys. Res. 102. https://doi.org/10.1029/96JB03860. ISSN: 0148-0227. Curie, Paris, pp. 350.

91