TECTONICS, VOL. 24, TC4020, doi:10.1029/2004TC001675, 2005

Plateau-style accumulation of deformation: Southern Altiplano

Kirsten Elger, Onno Oncken, and Johannes Glodny GeoForschungsZentrum Potsdam, Potsdam, Germany Received 5 May 2004; revised 17 December 2004; accepted 23 March 2005; published 31 August 2005.

[1] Employing surface mapping of syntectonic during the Paleogene, initially reactivating crustal sediments, interpretation of industry reflection- weak zones and by thermal weakening of the crust seismic profiles, gravity data, and isotopic age dating, with active magmatism mainly in the Neogene stage. we reconstruct the tectonic evolution of the southern Citation: Elger, K., O. Oncken, and J. Glodny (2005), Plateau- Altiplano (20–22S) between the cordilleras style accumulation of deformation: Southern Altiplano, Tectonics, defining its margins. The southern Altiplano crust 24, TC4020, doi:10.1029/2004TC001675. was deformed between the late Oligocene and the late Miocene with two main shortening stages in the Oligocene (33–27 Ma) and middle/late Miocene 1. Introduction (19–8 Ma) that succeeded Eocene onset of shortening at the protoplateau margins. Shortening [2] Although considerable advance has been made in recent years in understanding the processes involved in rates in the southern Altiplano ranged between 1 and the formation of orogenic plateaus, the precise temporal 4.7 mm/yr with maximum rates in the late Miocene. and spatial patterns of uplift and lateral progradation of Summing rates for the southern Altiplano and the deformation in plateaus beyond the last few million years is Eastern Cordillera, we observe an increase from still shrouded in darkness. This situation is largely due to a Eocene times to the late Oligocene to some 8 mm/yr, lack of systematic data with the potential to date the timing followed by fluctuation around this value during the of deformation and uplift, as well as of detailed structural Miocene prior to shutoff of deformation at 7–8 Ma and data that reveal the internal architecture of the world’s transfer of active shortening to the sub-Andean fold large orogenic plateaus, Tibet and Altiplano [e.g., Dewey and thrust belt. Shortening inverted early Tertiary et al., 1988; Isacks, 1988; England and Molnar, 1997; graben and half graben systems and was partitioned Allmendinger et al., 1997; Shen et al., 2001]. Young, in three fault systems in the western, central, and syntectonic and posttectonic sediments largely cover both plateaus and restrict access to the subsurface. Yet, models of eastern Altiplano, respectively. The east vergent fault large-scale orogeny and plateau formation heavily rely on systems of the western and central Altiplano were assumptions that relate to the early stages of plateau synchronously active with the west vergent Altiplano evolution and the conditions governing these. This is west flank fault system. From these data and from particularly true for the Andean Altiplano, which, due to section balancing, we infer a kinematically linked its setting in an ocean-continent convergence system, has western Altiplano thrust belt that accumulated a been called a plate tectonic paradox [e.g., Allmendinger et minimum of 65 km shortening. Evolution of this belt al., 1997]. Because of the unsatisfactory data situation, contrasts with the Eastern Cordillera, which reached published ideas on accumulation of deformation in the peak shortening rates (8 mm/yr) in between the above Altiplano range from various inferred stages in plateau two stages. Hence local shortening rates fluctuated evolution [e.g., James, 1971; Pardo-Casas and Molnar, across the plateau superimposed on a general trend of 1987; Isacks, 1988; Sempere et al., 1990; Allmendinger and Gubbels, 1996; Kley, 1996; Allmendinger et al., 1997; increasing bulk rate with no trend of lateral Lamb and Hoke, 1997; Lamb et al., 1997; Lamb, 2000; propagation. This observation is repeated at the Hindle and Kley, 2002] to suggestions of continuous east- shorter length and time scales of individual growth ward progradation of deformation, starting at the Western structures that show evidence for periods of enhanced Cordillera, with a propagation style much like that observed local rates at a timescale of 1–3 Myr. We interpret in classic foreland fold and thrust belts [e.g., Horton et al., this irregular pattern of deformation to reflect a 2001; McQuarrie, 2002]. To date, the available data in the plateau-style of shortening related to a self-organized Altiplano do not provide a complete quantitative record of state of a weak crust in the central South American back the spatial and temporal pattern of strain accumulation arc with a fault network that fluctuated around the during plateau formation that also sufficiently resolves the critical state of mechanical failure. Tuning of this state earlier stages. may have occurred by changes in plate kinematics, [3] In the present paper we attempt to provide a first nearly complete record of magnitude and timing of shortening across the southern Altiplano plateau (compare Copyright 2005 by the American Geophysical Union. Figure 1) by the mapping of syntectonic sediments, by 0278-7407/05/2004TC001675$12.00 identifying structures, their relative ages and magnitudes of

TC4020 1of19 TC4020 ELGER ET AL.: ALTIPLANO DEFORMATION TC4020

Figure 1

2of19 TC4020 ELGER ET AL.: ALTIPLANO DEFORMATION TC4020 deformation from the interpretation of industry reflection- shortening during the early to middle Miocene [e.g., Kley, seismic profiles, and by isotopic age dating of deformation, 1996; Allmendinger et al., 1997; Lamb and Hoke, 1997; including a reevaluation of published ages. Our analysis Mu¨ller et al., 2002]. The adjoining sub-Andean fold-thrust focuses on the southern Altiplano at 20–22S. This segment belt accumulated most shortening since the Miocene [Baby is characterized by the highest density of geophysical and et al., 1997; Kley and Monaldi, 1998] and still shows active geological data in the central Andes, all provided by a growth at its eastward deformation front as evidenced from variety of recent projects [e.g., Wigger et al., 1994; Beck active seismicity and GPS data [Lamb, 2000; Hindle and et al., 1996; Swenson et al., 2000; Yuan et al., 2000; Kley, 2002; Bevis et al., 1999]. Haberland and Rietbrock, 2001; Brasse et al., 2002; Go¨tze [6] In contrast to the plateau flanks, the southern Alti- and Krause, 2002; ANCORP Working Group, 2003]. In plano evolution is only recorded by few structural and age addition, high-resolution industry reflection seismic sections data [e.g., Baby et al., 1997; Rochat et al., 1999; Welsink et covering the entire plateau were made available to this al., 1995] (see Figure 1a). Most of the area is covered by project by the Bolivian national oil industry (courtesy of Quaternary fluvial and lacustrine sediments, including sev- YPFB). eral large saltpans (e.g., Salar de Uyuni, Figure 1), and by posttectonic volcanic rocks from the active magmatic arc of 2. Geological Framework the Western Cordillera. Most of the exposed pre-Cenozoic basement is made up of mainly unmetamorphic and mildly [4] In spite of ongoing subduction for more than 200 Myr, folded Ordovician to Devonian clastic sediments while the large-scale deformation of the upper plate and formation of western margin of the plateau is built of deformed Late the Andes high plateau only commenced during early to Paleozoic and Mesozoic igneous and sedimentary rocks. In mid-Tertiary times [Sempere et al., 1990; Lamb and Hoke, southern Bolivia, Cretaceous sediments unconformably 1997; Allmendinger et al., 1997; Horton et al., 2001]. The overlie this basement. These sediments were partly depos- resulting central Andean mountain belt is built from a series ited in a shallow marine environment [Fiedler et al., 2003] of morphotectonic units (Figure 1d): the Longitudinal and mark the last marine ingression prior to plateau forma- Valley in the forearc with an average elevation of some tion. This postrift sequence is related to the Potosi basin 1000 m and thin sedimentary infill; the 4500–5500 m high [Fiedler et al., 2003], a branch of a major rift system in the Chilean Precordillera with Paleozoic and Mesozoic base- south (‘‘Salta Rift’’), the site of which largely matches the ment, overlain by a thin Neogene cover; the Western extent of the later Eastern Cordillera. Cordillera built by the Neogene to recent magmatic arc [7] The establishment of internal drainage early in the with peak elevations of up to 6000 m; the flat Altiplano Andean evolution with two cordilleras confining the proto- Plateau at some 3800 m average elevation forming a plateau area was responsible for storing sediment that 200-km-broad intramontane basin; the Eastern Cordillera reflects the entire deformation history (compare Vilque well reaching 5000–6000 m altitude, a doubly vergent thick- [Welsink et al., 1995]; see Figure 1a). In the southern skinned thrust system; the inter-Andean and sub-Andean Altiplano, the late Paleocene Santa Lucia formation and zones representing an east vergent, ‘‘classical’’ thin-skinned the Eocene/Oligocene Potoco formation [Jordan and foreland fold-thrust belt that accumulated most of the back Alonso, 1987; Sempere et al., 1997; Horton et al., 2001], arc shortening; and finally the down flexed, but otherwise a series of coarse to fine-grained clastics and red beds undeformed Chaco at some 500 m elevation, a 200-km- conformably overlie the Cretaceous/Paleocene deposits wide foreland basin with a wedge-shaped Neogene clastic (Figure 2). Rochat et al. [1999] and Jordan et al. [2001] fill of up to 5 km thickness. suggested that these sediments were partly deposited in [5] Recent work has shown that deformation and uplift of local extensional basins. This sequence is in turn overlain the western flank of the Plateau and Precordillera occurred by the Miocene continental clastic San Vicente Formation between the late Oligocene to late Miocene at very low, [Martı´nez et al., 1994; Mertmann et al., 2003]. To the east mildly fluctuating strain rates [Victor et al., 2004] following this formation is delimited by the San Vicente Thrust zone an Eocene/?early Oligocene precursor shortening stage (Figure 1a) marking the western topographic limit of the [Haschke and Guenther, 2003]. In contrast, the Eastern Eastern Cordillera. The San Vicente Formation is charac- Cordillera and inter-Andean belt accumulated substantially terized by significant lateral variations in thickness, grain more shortening starting in the late Eocene, with most size, and detrital composition as well as by unconformable

Figure 1. (a) Geological sketch map of the southern Altiplano (section of map of central Andes by Reutter et al. [1994] including our own mapping results) superimposed on shaded relief of the topography showing the distribution of pre-late Neogene rocks, lower-hemisphere equal-area stereonet projections with bedding and regional fold axes, and the location of reflection seismic lines. (b) Bouguer anomaly map (provided by H. J. Go¨tze; see Go¨tze and Krause [2002] for description of gravity field) of area shown in Figure 1a depicting positive anomalies closely related to faults and folds mapped at the surface and observed in the seismic sections. Note the chain of anomalies in the nearly completely sediment-covered western Altiplano with similar style as obvious from the exposed central Altiplano unit. (c) Location of the working area. (d) Major morpho-tectonic units of the southern central Andes. See color version of this figure at back of this issue.

3of19 TC4020 ELGER ET AL.: ALTIPLANO DEFORMATION TC4020

and age with those identified by Lamb and Hoke [1997] farther north in the Altiplano.

3. Plateau Architecture 3.1. Structural Style

[8] In outcrop, major folds often associated with thrust faults characterize the southern Altiplano structural style. The folds have subvertical axial planes, a parallel geometry, and are nearly cylindrical with NNE-SSW oriented fold axes plunging less than 5 north or south (see Figure 1a). Bedding surfaces on the fold limbs have well developed slickensides usually oriented downdip (flexural-slip folds). Sediments at the limbs of folds often exhibit growth structures. Apart from folding, most of the shortening is recorded by thrusts rooting below the anticlines. Strike-slip faults are of very minor importance. Major extensional structures are not observed on the southern Altiplano, except for rare small-scale faults, which usually show displacements of no more than a few decimeters to meters. [9] At surface, the moderately eastward dipping thrust system of the San Vicente Fault Zone (SVFZ) separates the Eastern Cordillera in the east from the sediment-covered Altiplano in the west (Figure 1a). In the plateau area to the west, we subdivide the southern Altiplano into three tectono- stratigraphic domains based on its internal architecture: the eastern, the central, and the western southern Altiplano domain. These are also particularly well imaged by the high-resolution Bouguer map (Figure 1b). In the areas exposing deformed Tertiary rocks positive Bouguer anoma- lies perfectly trace anticlines and fault hanging walls. Hence the Bouguer map can be seen to highlight the structural trend of folds and faults beneath the younger cover. [10] The eastern domain, which also includes the Tertiary/ Quaternary Lı´pez Basin, is the buried tip of the Eastern Cordillera’s west vergent, thin-skinned western margin with Figure 2. Schematic stratigraphic columns of the widely spaced open anticlines. The central Altiplano do- (a) central and (b) eastern parts of the southern Altiplano main is formed by a thrust system separated from the with thickness as mapped in the field, in the Vilque well Eastern Cordillera by the undeformed Lı´pez basin: The [Welsink et al., 1995], and in seismic sections (Foldout 1). Khenayani-Uyuni Fault Zone (KUFZ) represents its key Age data collected in this study and by Silva-Gonza´lez part and is the most important structural element of the [2004] are shown at respective sample location in the southern Altiplano [Sempere et al., 1990; Martı´nez et al., stratigraphic column. Superscript 1 indicates that ages of 1994; Welsink et al., 1995; Rochat et al., 1999]. It represents base of Potoco and Molino Formations are from Sempere et an east vergent leading imbricate fan of up to four major, al. [1997, and references therein]; asterisks indicate that N-S to NNE-SSW trending, emergent thrusts (dips are 20– ages are from Silva-Gonza´lez [2004]. 30 at frontal thrust to subvertical at westernmost thrust). Along the thrusts, Paleozoic to Oligocene sediments were displaced eastward over the Cenozoic fill of the Lı´pez Basin (Figure 3). The hanging walls show folded and faulted onlaps on the Paleozoic basement. Isotopic ages from Paleozoic sedimentary basement with remnant Late Meso- intercalated volcanic rocks of the southern Altiplano range zoic, and a thinned to incomplete Cenozoic cover between some 27.4 and 5 Ma, interpreted to bracket the (Figure 3a). The footwalls of the individual thrusts are timing of deposition of the San Vicente Formation and the characterized by thrust front synclines containing growth supposedly contemporaneous deformation [Grant et al., structures with steep western limbs and up to 2 km long, 1979; Sempere et al., 1990; Soler et al., 1993; Lamb and gently dipping eastern limbs. Because thrust hanging wall Hoke, 1997; Horton et al., 2002]. Late Miocene to recent cutoffs are eroded, we assess displacement from down- volcano-sedimentary deposits unconformably overlie all plunge projection of cutoff geometries mapped in the preexisting units. The above sequences, summarized in southern parts of the area (Figure 1a). The western parts Figure 2, show significant similarities in depositional style of the central Altiplano system is built from west vergent

4of19 TC4020 ELGER ET AL.: ALTIPLANO DEFORMATION TC4020

the identical gravity signature as compared to the exposed central Altiplano, appears to have a similar internal archi- tecture. Its western part is also exposed farther north (18– 19S) in the Western Cordillera [Charrier et al., 2002; Garcı´a et al., 2002].

3.2. Reflection Seismic Data

[12] Surface observations are complemented by industry reflection data that image the crust down to some 4 s two- way time (TWT) depth. The definition of seismic facies and their correlation with mappable stratigraphic formations and well data mainly is shown in Figure 9 in the auxiliary material1. [13] Seismic lines 10007 and 10010 (Foldout 1) illustrate the lateral (north-south) variability of the Lı´pez Basin and deformation front of the Eastern Cordillera. Toward depth (line 10007 in Foldout 1), the Potoco Formation shows a westward thinning wedge with conformable top lap reflec- tor geometries at both the upper and lower sequence boundaries. The overlying San Vicente Formation is built from three subsequences, defined here as the OMsv1, OMsv2 (Oligo-Miocene San Vicente Formation), and Pilkhaua Subsequence. The OMsv1 and OMsv2 Subse- quences are characterized by relative uniform thickness across the entire section. The Pilkhaua Subsequence is absent in line 10007, most probably due to erosion (com- pare, however, its occurrence on the west flank of Vilque Anticline in line 10010). In the syncline east of the Vilque Well small normal faults are local features, related to the Figure 3. Structures from Khenayani-Uyuni Fault Zone collapse of the forelimb of the eastern Anticline above system. (a) Unconformity between San Vicente Formation evaporite layers within the OMsv2 Subsequence (600 m (Pilkhaua Subsequence?) and Silurian deposits in hanging of evaporites drilled in the Vilque Well). The most prom- wall of Khenayani-Uyuni Fault. (b) Steeply inclined San inent unconformity of line 10007 is the truncation at the Cristo´bal Fault with subvertical hanging wall structure base of the OMsv1 Subsequence near the deformation front (Potoco Formation) intruded by San Cristo´bal Volcano. (c) (DF, see Foldout 1e, 1f) above eroded top lap terminations Growth strata in footwall of San Cristo´bal Fault (Pilkhaua of the Potoco and El Molino Formations. A second struc- Subsequence). tural disconformity in the western part of the Vilque Anticline between the OMsv1 and OMsv2 Subsequences (Foldout 1e) is related to the blind tip of the hanging wall fault propagation folds with 6–10 km wavelength overlying ramp of the Vilque fault bend fold. blind thrusts. The folds have associated small thrust top [14] At its eastern border, the southern part of the Lı´pez basins filled with wedge-shaped, onlapping deposits. This basin (line 10010, Foldout 1f) exhibits thinning and pro- entire central Altiplano system is continued to the north in gressive onlaps of reflectors of the Potoco Formation and the exposed Tambo-Tambillo area (Figure 1a) [cf. Lamb and OMsv1 Subsequence against a basement high. As in line Hoke, 1997]. 10007, the unconformity at the base of the OMsv1 Subse- quence overlies well-developed top lap terminations from [11] The largely sediment-covered western Altiplano do- main only exposes deformed rocks at the Yazo´n Peninsula the Potoco Formation With the onset of deposition of the in the southern Salar de Uyuni (21S, 68 W), where OMsv2 Subsequence the downlapping reflectors in the folded strata of the San Vicente Formation crop out western part of the basin indicate a westward prograding (Figure 4). The structure of the Yazo´n Peninsula is a west system of foresets. Locally developed top lap terminations vergent, NNE-SSW trending fold with a long eastern back below the eastern OMsv2 Subsequence show that this limb, a steep west facing forelimb and an associated thrust sequence boundary was also partly erosive (Foldout 1e, 1f). front basin filled with growth strata. In spite of the isolated [15] The most obvious difference between the northern nature of this exposure, the Bouguer map (Figure 1d) shows section (10007) and the southern section (10010, compare that the Yazo´n structure is part of a major NNE-SSW Foldout 1) is the smaller basin width in the southern section. trending system of folds and/or faulted and uplifted rocks A second difference between lines 10007 and 10010 is a under the volcanic and Salar cover (Figure 4). This nearly unknown structural unit in the western Altiplano signifi- 1Auxiliary material is available at ftp://ftp.agu.org/apend/tc/ cantly contributes to its deformation architecture and, from 2004TC001675.

5of19 TC4020 ELGER ET AL.: ALTIPLANO DEFORMATION TC4020

Figure 4. Exposed part of western Altiplano thrust system (see Figure 1a for location, Yazo´n Peninsula). (a) Satellite image. (b) Geological map with lower-hemisphere equal-area stereonet projection showing regional fold axis and bedding. (c) Cross section with local age data. (d) SW-ward view of roof and frontal limb of Yazo´n Anticline with onlapping growth sediments of Pilkhaua Subsequence. southward decrease of thrust displacement related to for- of the seismic image correlate with steeply inclined strata or mation of the Vilque Anticline, a classical fault-bend with footwall units buried beneath deformed imbricates. anticline, in the north. The most significant observation of The strongly folded and imbricated sequences of the central Foldout 1, however, is the strong similarity between both Altiplano can be observed in nearly continuous outcrop 5– sections in terms of the characteristics of seismic facies, 10 km south of the seismic section (Figure 1a). Line 10023 similar positions of unconformities, and the structural style shows an east vergent imbricate fan building the eastern part allowing correlation between lines. of the system; west vergent fault bend folds dominate the [16] In the central Altiplano domain (lines 10023 and western part. The analysis of seismic facies and their 2593, see Foldout 1) surface outcrops at sites of poor quality correlation with surface geology exhibits the same, albeit

6of19 TC4020 LE TA. LILN DEFORMATION ALTIPLANO AL.: ET ELGER 7of19

Foldout 1. Cross section of the southern Altiplano based on seismic lines 2553 (stacked section), 10023, 2593, and 10007 (migrated sections), courtesy of YPFB (see Figure 1 for location) and their time-depth conversion using average interval velocities of the lines as shown in inset (bold line, average; thin lines, variation along line). Western Altiplano profile is projected into the common section. Structures are shown by name or numbers (eastern Altiplano) with nearby isotopic ages projected along strike. Restored section shows late Oligocene stage prior to onset of sedimentation of San Vicente Formation Details show (a) western deformation front of western Altiplano thrust system; (b) wedge-shaped westward thinning of Potoco Formation and youngest fold of central Altiplano thrust system folding the onlapping Pilkhaua Subsequence; (c) fault-propagation folds of San Vicente Formation with piggyback basins filled with Pilkhaua Subsequence that onlap neighboring folds; (d) thrust units bounded by Corregidores (right) and San Cristo´bal Faults (left) with Potoco beds truncated by San Vicente Formation and growth strata of the Pilkhaua Subsequence; (e) buried western deformation front of the Eastern Cordilleran Thrust System truncated by the base of San Vicente Formation (OMsv1); (f) western deformation front of the Eastern Cordilleran Thrust System along the southern line 10010 (see Figure 1 for location) with truncation of older structures by the basal San Vicente Formation. The latter shows internal truncations in the east (between OMsv1 and OMsv2), westward prograding sedimentary foresets in the west (OMsv2), and growth structures at the top of the sequence (Pilkhaua Subsequence). Numbers indicate sequence of fault activation. (See enlarged version of this figure in the HTML.) TC4020 TC4020 ELGER ET AL.: ALTIPLANO DEFORMATION TC4020

Figure 5. (e) Surface outcrop and (a and b) interpreted part of seismic line 10023 (see Foldout 1 for location of detail) with backlimb of Allka Orkho fold structure (in the west) and Ines fold (in the east) and intervening basin filled with Pilkhaua Subsequence. The basin fill shows stacked generations of onlapping reflector terminations onto both folds and decreasing intensity of folding upsection in discontinuous steps marked as (c and d) individual sequences (see text for details; PT, pretectonic; ST, syntectonic; PST, posttectonic after ST1 and erosion of Allka Orkho fold). Bars in San Vicente Formation in Figure 5a) indicate reflector sequences bounded by continuous high-amplitude reflections.

more deformed and dissected, seismic sequences as detected the lower sequence boundary of the overlying San Vicente in the eastern Altiplano (Foldout 1 and Figure 9 in the Formation evolves from conformable in the footwall of the auxiliary material). This includes the prominent double individual faults to erosive in their hanging wall positions reflector of the El Molino Formation, the seismically (top lap terminations of the Potoco reflectors, see arrow in transparent facies of the Potoco Formation, and the contin- Foldout 1d; see also Figure 3c). uous high-amplitude reflectors of the San Vicente Forma- [18] The main body of the San Vicente Formation has a tion In the central Altiplano, subdivision of the San Vicente more constant thickness, but also thins westward (line Formation into the OMsv1 and OMsv2 Subsequences is not 10023 in Foldout 1) and is very thin on the Khenayani- possible any more. The overlying Pilkhaua Subsequence Uyuni Fault Zone as compared to the immediately adjacent is most distinct as infill of local thrust top basins (e.g., areas. Contacts with underlying strata are conformable in Foldout 1c and Figure 5). most cases except for the above erosional unconformity at [17] The Potoco Formation forms a continuously east- in hanging wall positions (Corregidores Fault, Khenayani- ward thickening wedge that reaches its maximum thickness Uyuni Fault, and Allka Orkho Fault). In the uppermost San in the hanging wall of the San Cristo´bal Fault and again Vicente Formation, thrust top basins filled with onlapping exhibits conformable sequence boundaries. Its thickness is deposits show wedge-shaped sedimentary bodies defining significantly smaller east of the San Cristo´bal Fault. From the Pilkhaua Subsequence. These Pilkhaua-filled basins seismic analysis, map interpretation, and field observations, generally do not extend below 0.5–0.8 s TWT depth. West

8of19 TC4020 ELGER ET AL.: ALTIPLANO DEFORMATION TC4020 of the westernmost anticline, the seismic profile was ac- dipping thrusts (KUF, CF, SCF, IPF, see Foldout 1) with quired over undeformed lacustrine sediments south of the steep to moderately dipping hanging wall strata underlain Salar de Uyuni (see Foldout 1) that cover eroded top laps by a gently west dipping detachment (Foldout 1, line of the Pilkhaua Subsequence at flanks of major folds 10023). In comparison to earlier interpretations [e.g., Baby (Foldout 1b). et al., 1997; Lamb and Hoke, 1997; Rochat et al., 1999; [19] In the western Altiplano, the structural architecture McQuarrie and DeCelles, 2001], this geometry, as deducted was only accessible at the Yazo´n peninsula and in a seismic from surface as well as from seismic data (compare stack section of poor quality crossing the Salar de Uyuni Foldout 1), accommodates significantly more shortening. (example and tentative interpretation of the line drawing of Large parts of the hanging walls were eroded during line 2553 is shown in Foldout 1a). Interestingly, structural thrusting. Eroded hanging wall cutoffs were therefore analysis exhibits a doubly vergent imbricate fan and fault reconstructed from downplunge projection of the map. bend fold system quite similar to the central Altiplano The less deformed, west vergent part of the central Altipla- structure. In addition, this unit shows a similar style of no Fault System exhibits four fault propagation folds above Bouguer anomalies as the central Altiplano (Figure 1d). eastward dipping, mostly blind, thrusts (Foldout 1, line Judging from reflection styles, there is restricted evidence 10023). The folds have associated syntectonic basins filled for the presence of the El Molino and Potoco Formations with sediments of the Pilkhaua Subsequence. Initial short- (see Figure 9 in the auxiliary material) or they are indistin- ening prior to deposition of the San Vicente Formation, as guishable from the San Vicente Formation In contrast, the evidenced from its erosive basis, only affected the external Pilkhaua Subsequence is clearly identified by its reflector parts of the doubly vergent thrust system, namely the architecture. Both San Vicente Subsequences are also ob- Khenayani-Uyuni and the Corregidores imbricates in the served at surface (Yazo´n Peninsula) and show the same east and the Allka Orkho Fault in the west. The eroded onlapping relationship as in the central and eastern Altipla- hanging wall cutoffs were reconstructed from extrapolating no domains (Figure 4). All units are covered and onlapped stratigraphic thickness from the footwall to the hanging by younger sediments and volcanics of the Salar de Uyuni wall. Where thickness changes are evidenced, like the basin. reduced Potoco Formation thickness in the KUFZ area, we used erosion estimates derived from fission track mod- eling. Ege [2004] calculated some 0.9–2.4 km of erosion 3.3. Section Balancing for the Potoco Formation prior to the onset of San Vicente [20] Using surface data and the depth-converted inter- deposition in this area. Restoration of the central Altiplano preted line drawing (depth conversion based on averaged revealed a total of 40 km horizontal contraction (13 km of interval velocities of the seismic lines, see inset in Foldout 1), which were accumulated in the early stage) with a predom- we iteratively constructed and restored a balanced section inance of east directed thrusting (see Foldout 1). 1 with the software packages GEOSEC (Paradigm) and [23] The main body of the San Vicente Formation below 2D-MOVE1 (Midland Valley Exploration Ltd.) using the Pilkhaua Subsequence exhibits a uniform thickness flexural flow as the dominant deformation mechanism as unaffected by the west vergent structures. In contrast, in indicated from field observations. We determined fault the dominantly east verging thrusts, the same formation is geometry at depth from seismic interpretation and depth- very thin or absent, or represented only by Pilkhaua-style to-detachment calculations using the fold geometries of deposits (compare Foldout 1, line 10023). This fact requires various fault-related folds [cf. Epard and Groshong, 1993]. either continuous or repeated activity of the KUFZ, in order [21] The eastern Altiplano domain is shown to be part of to maintain local relief, or very high relief formed during the a west vergent thin-skinned fold-and-thrust belt with widely initial stage that was not downgraded before the late stage of spaced ramp anticlines and open fault-bend folds (Foldout 1, deposition of the San Vicente Formation In both cases, local line 10007). Here, most shortening was accumulated prior relief with respect to neighboring domains persisted until to deposition of the San Vicente Formation in the western the sedimentary surface in these reached the surface eleva- part of the section (4.5 km shortening). All thrusts join a tion of the eroded KUFZ area and then started to transgress gently eastward dipping detachment, which is characterized the latter. In summary, deformation in the central Altiplano by a double flat ramp geometry, descending eastward in two locally started prior to deposition of the San Vicente steps from six to nine kilometers depth within lower Formation and continued throughout San Vicente deposi- Paleozoic sediments and probably linking with the San tion in the KUFZ area, while all of the central Altiplano Vicente fault. Minor reactivation (1.5 km shortening) structures were active during deposition of the Pilkhaua during a later stage (post-OMsv2 deposition) was responsi- Subsequence. ble for the formation of two wide ramp anticlines east of the [24] For the western Altiplano thrust system, we only earlier deformation front. This interpretation is in contrast to aimed at a semiquantitative model interpreting mainly the the extensional model of Welsink et al. [1995] but is similar seismic section based on reflection styles and geometries. to that of Baby et al. [1997] and Mu¨ller et al. [2002] The doubly vergent style of deformation, again with a showing, however, significantly more detail. predominance of east directed thrusting, is similar to that [22] In the central Altiplano domain, the Khenayani- of the central Altiplano system. However, for lack of Uyuni Fault Zone (KUFZ) forming the eastern part of this resolution of structures, the measured shortening of 21 km doubly vergent thrust system is built from four westward (see Foldout 1) is a minimum estimate only. A separation of

9of19 TC4020 ELGER ET AL.: ALTIPLANO DEFORMATION TC4020 shortening into two stages is not detectable due to the poor KE06-99 and KE10-99 were taken from the eastern flank of quality of the seismic line and lack of exposure. the Vilque Anticline in the eastern Altiplano, sample KE19- [25] Balancing shows that all thrusts in the central and 99 is from the eastern flank of the Ines Anticline in the western Altiplano thrust systems merge into a basal detach- central Altiplano, and samples KE11-99 and KE12-99 come ment, which lies at depths between 9 and 12 km in the from the northern part of the Yazo´n Peninsula in the western Paleozoic sediments (compare eastern Altiplano) and dips Altiplano (Figure 1). 1–2 westward (Foldout 1). Apart from the dominant east [31] The history of three syntectonic basins of the Pil- directed thrusting, indications for this detachment geometry khaua Subsequence is constrained as follows: For the basin were derived from the following two observations: west of the Yazo´n Anticline (western Altiplano), a late [26] 1. The Late Cretaceous El Molino Formation forms a pretectonic tephra sample is from the youngest predeforma- subhorizontal regional near sea level in the east rising to tional strata of the San Vicente Formation, from the verti- 1500 m elevation in the western Altiplano. Exceptions to cally inclined forelimb of the Yazo´n Anticline (KE28-00). this are the strongly deformed external parts of the central KE24-00 was taken only 50 m farther west, in the basal part Altiplano doubly vergent thrust system, i.e., the Khenayani- of the steeply inclined (48) syntectonic Pilkhaua sediments, Uyuni Fault Zone and the Allka Orkho Fault. There, the El 5 m above the basal conglomerate (Figure 4). The samples Molino Formation was uplifted along faults and locally KE01-00 and KE30-00 were interbedded in the syntectonic crops out at an altitude of 4000 m (Foldout 1). basin west of the Ines Anticline, central Altiplano (Figure 5). [27] 2. The floor of the basins filled with Pilkhaua KE39-00, KE40-00, and KE41-00 were sampled from the Subsequence has a constant elevation of some 3200 m west syntectonic basin west of the central Anticline (Foldout 1). of the Khenayani-Uyuni Fault dropping to 2100 m east of The crystal tuff KE40-00 is only slightly inclined, while the fault. Because the Pilkhaua Subsequence has a nearly KE39-00 lies horizontally. KE41-00 is a volcanic boulder in identical age throughout the southern Altiplano (see below a volcaniclastic layer within the steeper parts of the basin. and Figure 2) and lacks lateral facies changes, this indicates The latter two tuffs were deposited during or after the latest (1) joint uplift of the latter basins in the west and (2) a stages of thrusting. Sample KE13-00 is a posttectonic planar decollement geometry below this part of the plateau ignimbrite of the southern central Altiplano (southwest of without major variation of dip angle that would have Julaca, Figure 1 and Foldout 1). 40 39 triggered formation of ramp anticlines and associated dif- [32]The Ar/ Ar analytical data are summarized in ferential uplift of the above units. Table 1. An important observation is the presence of [28] The restored section (see Foldout 1) reveals that variable amounts of excess Ar in almost all samples, both shortening was preceded by an extensional stage that led in hornblende and biotite. We therefore evaluated the data to the formation of wedge-shaped half graben fills of the using both the apparent age spectra and the information Eocene/early Oligocene Potoco Formation [cf. Jordan et al., from Ar-Ar isochrons [cf. Miller et al., 1998]. In some 2001]. The eastern border of the central half graben is samples, excess Ar is restricted to the first, low-temperature formed by a normal fault that was inverted as the San degassing steps, and high-temperature degassing steps con- Cristo´bal reverse fault. A similar half graben structure may stitute plateau ages, which we regard as valid extrusion ages build the Lı´pez Basin and also appears plausible for explain- for the respective tuffs (KE13-00 biotite, KE24-00 biotite, ing the seismic architecture of the western Altiplano. KE40-00 biotite, KE28-00 amphibole, KE39-00 amphibole, Estimated extension of the protoplateau domain was less and KE41-00 amphibole; Table 2). Samples with persis- than 10–15%. tence of excess Ar throughout the degassing spectra allowed determination of reasonably precise Ar-Ar isochron ages (samples KE01-00 biotite, KE39-00 biotite; Table 2), sim- 4. Geochronology: The 40Ar/39Ar and K-Ar ilarly interpreted as extrusion ages. Age Data [33] The K/Ar analytical data are presented in Table 2. All K/Ar apparent ages for pre-Pilkhaua tuffs of the San Vicente [29] To further decipher the Cenozoic deformation history Formation are higher than 18 Ma. The observation of of the central and western Altiplano, we sampled volcani- abundant presence of excess Ar in the Ar/Ar data set sets clastic intercalations at key locations along the cross section important limitations on the geological significance of the for 40Ar/39Ar and K/Ar isotopic age determinations. Linking K-Ar ages from similar rocks. Apparently, the melts that age data with structural constraints and stratigraphy aimed formed the felsic tuffs and ignimbrites were rich in argon at providing time brackets for specific deformation incre- with elevated 40Ar/36Ar ratios. Correspondingly, Ar solubil- ments. Samples have been selected to cover the entire ity is known to be particularly high in felsic, rhyolithic melts sedimentation and deformation history. Sample material [Kelley, 2002]. It is very likely that the presence of excess Ar mostly consisted of felsic, rhyolithic tuffs with abundant is a general phenomenon in phenocrysts from Altiplano tuffs indications for textural and physico-chemical disequilibria. and ignimbrites that are petrologically and genetically sim- An outline of data acquisition procedures and a detailed ilar to the rocks that were analyzed by the Ar/Ar method in evaluation of analytical data are given in the auxiliary the present study. The conventional K/Ar approach depends material. on the assumption that all Ar trapped during the magmatic [30] The sedimentation history of the San Vicente For- stage is of atmospheric composition. The K-Ar dates are thus mation is surveyed with five crystal tuff samples. Samples equivalent to the total gas ages calculated from Ar/Ar data

10 of 19 TC4020 ELGER ET AL.: ALTIPLANO DEFORMATION TC4020

Table 1. Sample Details and Summary of 40Ar/39Ar Data Collected by Incremental Heatinga

Sieve Fraction UTM-x, Strike/ Tectonic Plateau TGA, Isochron Initial Sample Rock Type Material (mm) UTM-y Dip Position Age, Ma Ma Age, Ma 40Ar/36Ar

KE01-00 tephra biotite >0.25 662592, 7677470 290/60 syntectonic — 29.7 ± 5.3 15.8 ± 1.2 366.1 ± 20.6 CF: 0.9878 KE13-00 ignimbrite biotite >0.355 644433, 7669689 315/15 posttectonic 9.4 ± 0.3 10.1 ± 0.4 9.8 ± 0.2 242.7 ± 47.2 CF: 0.9990 KE24-00 tephra biotite 0.160–0.355 605882, 7731505 288/48 syntectonic 17.1 ± 0.7 17.1 ± 1.1 17.0 ± 0.9 330.5 ± 47.8 CF: 0.9933 KE28-00 tephra hbl 0.2–0.5 605967, 7730893 323/88 pretectonic 20.9 ± 2.1 24.0 ± 2.9 20.6 ± 0.1 447.9 ± 31.8 CF: 0.9999 KE39-00 crystal tuff biotite 0.355–0.5 675418, 7678968 horizontal syntectonic — 15.8 ± 2.6 7.6 ± 0.9 400.6 ± 13.2 CF: 0.9710 KE39-00 crystal tuff hbl 0.355–0.5 675418, 7678968 horizontal syntectonic 8.3 ± 1.2 9.2 ± 1.4 8.5 ± 0.2 307.0 ± 45.5 CF: 0.9991 KE40-00 crystal tuff biotite 0.5–1 679815, 7684634 272/15 syntectonic 7.9 ± 0.2 8.1 ± 0.3 8.2 ± 0.1 279.1 ± 20.9 CF: 0.9997 KE41-00 crystal tuff hbl 0.125–0.25 680936, 7682507 257/60 syntectonic 29.3 ± 2.0 32.2 ± 2.2 29.2 ± 0.6 317.6 ± 75.2 CF: 0.9994

aIsochron age is age from 40Ar/36Ar versus 39Ar/36Ar isotope correlation; TGA is total gas age. Errors are given at the 1s level. Abbreviations are cl, clast; hbl, hornblende. Preferred ages are in bold type. CF is correlation coefficient of 40Ar/36Ar versus 39Ar/36Ar isochron. Detailed analytical data are presented in the auxiliary material. that are systematically too high if excess Ar is present (see [35] The formation of structures related to deposition of the Ar-Ar data of Table 2). Since presence of excess Ar must the Pilkhaua Subsequence occurred between 19 and also be suspected for the samples analyzed with the K-Ar 7 Ma. In the western Altiplano, the earliest onset of method, all our K-Ar data as well as previously published shortening was found at the Yazo´n Anticline (youngest K-Ar data from the same area and rocks have to be pretectonic tuff, 20.6 ± 0.1 Ma, Ar-Ar; sample ke28-00 interpreted as maximum ages for eruption. overlain by the oldest syntectonic tuff, 17.1 ± 0.7 Ma, Ar- Ar, sample ke24-00, Figure 4). However, interpretation of the seismic section farther north indicates that these samples 5. Age and Pattern of Deformation may be underlain by as much as several 100 m of onlapping Pilkhaua Subsequence that records an earlier onset of 5.1. Constraints on Deformation Stages deformation. The oldest undeformed andesitic lava covering [34] The youngest structures observed are small normal the western Altiplano thrust system (at northern margin of faults that dissect the syntectonic sediments of the Pilkhaua Salar de Uyuni) was dated at 11.1 ± 0.4 Ma [Baker and Subsequence in the eastern Altiplano (to the east of the Francis, 1978] marking a minimum age for the end of Vilque Anticline, Foldout 1) but are covered by undeformed shortening here (compare similar observations in Western Quaternary sediments. These faults are local features related Cordillera farther north at 18–19SbyCharrier et al. to underlying evaporates (see chapter on seismic interpre- [2002] and Garcı´a et al. [2002]). tation) and were only observed in the eastern part of the [36] Syntectonic sedimentation in the doubly vergent main cross section (line 10007). thrust system of the central Altiplano began at a similar

Table 2. Sample Details and Summary of K-Ar Dataa

Rock UTM-x, Strike/ Tectonic 40Ar0,b 40Ar0/ Average Average Age, 40 40 0 Sample Type Material UTM-y Dip Position ppm Artotal Ar , ppm K, % K, % Ma

KE06-99 crystal biotite 750589, 7664554 108/20 pretectonic 0.007841, 0.008093 0.302, 0.112 0.007967 6.242, 6.291 6.266 18.2 ± 0.5 tuff KE10-99 crystal biotite 751833, 7668609 069/03 pretectonic 0.005046, 0.005200 0.227, 0.222 0.005123 4.124, 4.137 4.130 17.8 ± 0.5 tuff KE11-99 crystal biotite 607111, 7757827 270/55 pretectonic 0.010510, 0.010850 0.274, 0.298 0.010680 6.298, 6.065 6.182 23.6 ± 0.6 tuff KE12-99 crystal fsp 607111, 7757827 275/55 pretectonic 0.000626, 0.000550 0.066, 0.077 0.000588 0.343, 0.353 0.348 24.2 ± 1.2 tuff KE19-99 crystal glass 667540, 7669542 120/55 pretectonic 0.005949, 0.005676 0.325, 0.281 0.005812 3.348, 3.380 3.367 24.7 ± 0.6 tuff

a 40 10 0 10 40 Abbreviation is fsp, feldspar; Ar refers to in situ radiogenic Ar. Constants used are lb = 4.962 10 /yr; l0 + l0 = 0.581 10 /yr; K/Ktotal = 1.193 104 g/g. Errors given at the 1s level. bReplicate analyses of the same sample (every sample was measured twice).

11 of 19 TC4020 ELGER ET AL.: ALTIPLANO DEFORMATION TC4020 time: The oldest syntectonic age (15.8 ± 1.2 Ma, Ar-Ar, sedimentary record in the San Vicente Formation with ke01-00) was obtained from a 60 WNW dipping tephra in material provenance from the east [Horton et al., 2002], the thrust front basin west of the west vergent Ines Anticline deformation in the adjoining parts of the Eastern Cordillera (see Figure 5, note, however, that the sample position is (i.e., San Vicente thrust and hanging wall to the east) must underlain by 200 m of Pilkhaua deposits at depth). In the have peaked between 25 and 17 Ma, during deposition of east vergent part, the oldest K-Ar maximum age was the westward prograding OMsv2. According to Mu¨ller et al. derived from a vertically inclined tuff in the footwall of [2002], the nearby San Vicente thrust was sealed by the San Cristo´bal Fault (13.7 ± 0.4 Ma, sample ps99-17 volcanics with an age of 17 Ma marking the end of the [Silva-Gonza´lez, 2004]; note again underlying Pilkhaua main stage of deformation and uplift of the margin of the deposits of 400–600 m thickness). The end of deformation Eastern Cordillera. This matches the end of deposition of is indicated by undeformed dacitic lava, which unconform- the OMsv2 Subsequence of the San Vicente Formation. ably overlies steeply inclined San Vicente strata in the [39] The locally developed truncation of tilted hanging hanging wall of the westernmost fault of the KUFZ wall sequences below the San Vicente/OMsv1 across major system, dated at 11.0 ± 0.5 Ma (K-Ar, sr99-10 [Silva- parts of the southern Altiplano is evidence for pre-San Gonza´lez, 2004]). Subhorizontal to horizontal tuffs in Vicente deformation. The almost complete lack of volcanic growth sequences yielded 7.9 ± 0.2 and 8.3 ± 1.2 Ma intercalations in the Potoco Formation, which was deposited (ke41-00, Ar-Ar, bedding 272/15; ke39-00, Ar-Ar, horizon- prior to this deformation increment, prevents deriving tal bedding) in the central parts of the doubly vergent precise age constraints from syntectonic deposits. A single system, and 11.0 ± 0.3 Ma (K-Ar maximum age, ps99-26 tuff age within intensely folded Potoco sediments in the [Silva-Gonza´lez, 2004]) in the footwall of the San Cristo´bal footwall of the Corregidores Fault yielded a K-Ar maximum Fault. Subhorizontal tuffs sealing the Khenayani-Uyuni age of 40.4 ± 1.1 Ma (sr99-03 [Silva-Gonza´lez, 2004]). The Fault (10 ± 0.3; sample SR99-15 [Silva-Gonza´lez, 2004]) oldest dated volcanic rock of the San Vicente Formation in indicate that deformation at the KUFZ was shut off between the central Altiplano gave 29.3 ± 1.9 Ma (volcanic boulder 10 and 11 Ma (compare similar observations on Tambo- in volcaniclastic deposit; sample ke41-00, amphibole Ar- Tambillo structure toward the north [Lamb and Hoke, Ar). Moreover, near sites of early shortening (eastern 1997]). The progressively tilted deposits of growth strata margin of KUFZ) the basal conglomerate overlying Eocene in the hanging wall of the KUFZ indicate continuous to Paleozoic sediments contains Silurian clasts from the thrusting since 14 Ma (oldest dated Pilkhaua deposit). adjacent eroded hanging walls and is overlain by a basaltic- However, the parallel and folded reflectors of several 100 m andesitic tuff with a maximum age of 27.4 ± 0.6 Ma (K-Ar, of San Vicente deposits underlying this sample may indicate sample sr99-12 [Silva-Gonza´lez, 2004]). Finally, Ege [2004] slow or absent deformation earlier. On the other hand, the reports apatite fission track ages from the uplifted Paleozoic absence of San Vicente deposits older than 15–18 Ma in the basement rocks of the KUFZ faults, ranging between 28 and KUFZ structure, their unconformable basis covering all 33 Ma. In conclusion, this early shortening stage in the deeper units in the hanging wall anticlines, and the thick- central Altiplano is bracketed between 35 and 27 Ma. ness of 3–3.5 km for the San Vicente deposits in the [40] In the eastern Altiplano domain, the anticlinal hinges footwall require continuous or repeated fault activation formed during this stage were eroded prior to the deposition during San Vicente deposition in order to maintain local of the lowermost OMsv1 sediments. Some 700 m above the relief and to provide accommodation space (see above). base of the latter, a K/Ar age of 25.3 ± 1 Ma was determined [37] In the eastern Altiplano domain, the laterally uniform (sample an14-2000 [Silva-Gonza´lez, 2004]). Considering thickness and overall conformable boundaries of the OMsv1 average sedimentation rates (compare Figure 2b) the base and OMsv2 Subsequences indicate a late Miocene age of of the San Vicente Formation, and hence the period of this deformation for the Vilque and the eastern Anticlines (see early deformation increment, is constrained to between 28 Foldout 1). The youngest OMsv2 age in the eastern Alti- and 32 Ma at the western margin of the Eastern Cordillera. plano (16–17 Ma, extrapolated from samples ke10/99, 17.8 ± 0.5 Ma, and ke 06/99, 18.2 ± 0.5, 200 m below top OMsv2) is a maximum age of formation of these 5.2. Evolution of Shortening structures. Onlaps and divergent reflectors to the west of [41] Using the data on horizontal shortening as derived the Vilque Anticline (line 10010, see Foldout 1) indicate from section construction and the newly reported as well as that its growth was coeval with the deposition of the published ages from synkinematic deposits, we analyze the Pilkhaua Subsequence. The youngest mildly tilted tuff in evolution of shortening in more detail with the strategy the Pilkhaua Subsequence north of the Vilque anticline has outlined by Victor et al. [2004]: Shortening related to each an age of 7.6 ± 0.2 Ma [Silva-Gonza´lez, 2004]. fault is distributed over the period of its activity as con- [38] Except for the observations in the KUFZ area, strained from the sedimentary record in the immediately undisturbed deposition of the San Vicente Formation in adjoining syntectonic deposits. The age estimate for the the entire southern Altiplano indicates a stage of local oldest beds recording the onset of deformation and of the tectonic quiescence. In contrast, its more complex internal youngest deformed are based on the above isotopic data or sequence architecture toward the margin of the Eastern were estimated from extrapolation using average sedimen- Cordillera (with westward prograding foresets) must have tation rates (the latter are 100 m/Myr ± 20 m/Myr as derived witnessed nearby uplift east of the study area. From the for the Pilkhaua Subsequence across the Altiplano with

12 of 19 TC4020 ELGER ET AL.: ALTIPLANO DEFORMATION TC4020 local maximum of 200 m/Myr in the center of the Lı´pez 10010 in Foldout 1f). Shortly before or during initial basin) and some 150–250 m/Myr or 150–200 m/Myr for deposition of OMsv1, shortening propagated westward into the deeper members of the San Vicente Formation in the the Lı´pez basin and, subsequently, retreated back east to the eastern and central Altiplano domain. respectively; compare San Vicente thrust. The latter accommodated subsequent Figure 2). To assess the maximum or minimum period of shortening at the Eastern Cordillera front between 27 and fault activity we included the respective errors in isotopic 17 Ma (i.e., during OMsv1 and OMsv2 [Mu¨ller et al., age dating and the sedimentation rate estimate used to 2002]). Following shutoff of the San Vicente Thrust, calculate the ages of the base and top of syntectonic deformation again propagated westward, forming the Vilque deposits. Potential error in determining heave is estimated Anticline, until its end at 7 Ma. at ±10% (+20% in less well constrained sections). Using [45] Across the entire southern Altiplano, the Pilkhaua- these estimates, we assessed the highest and lowest possible filled basins exhibit 3 to 5 internal subsequences based on shortening rates for each fault over its activity period. observation of reflector terminations and reflector packages Summing these numbers for all kinematically linked faults (compare Foldout 1 and Figures 5c and 5d). These packages provides minimum shortening rates (minimum shortening show a trend toward long high-amplitude reflections at their and maximum activity period) and maximum shortening respective tops. This trend in reflection signal reflects a rates (maximum shortening and minimum activity period) progressive change in grain size, porosity, and/or average for a thrust system. bed thickness. In the Vilque well, Welsink et al. [1995] [42] Figure 6 shows the summed results merged with describe the San Vicente Formation drilled to consist of two equivalent data from the literature for the western plateau major fining upward sequences matching the two seismic flank and the Eastern Cordillera at 21S. As mentioned sequences in line 10007 that crosses the well site (compare above, two main stages of shortening can be identified Foldout 1). Mertmann et al. [2003] also describe sequences across the Altiplano up to the western border of the of several 10 to 500 m thickness showing internal fining Eastern Cordillera. An earlier Oligocene stage (between upward cycles in the San Vicente Formation While these 35 and 27 Ma) has averaged local shortening rates sequences are very probably also influenced by climatic between 0.1 and some 3 mm/yr (not resolved in western fluctuations the nature of the sediments as syntectonic Altiplano for lack of exposure). This stage is followed by deposits next to tectonically created relief indicates that an early Miocene lull in deformation. Shortening resumed changes of the balance between relief formation and relief with a distinct local acceleration to some 1.7–3 mm/yr degradation have the most important impact on the local (west and center) in the middle to late Miocene (20– sedimentation. 10 Ma). In most of the Altiplano, rates were higher [46] From the age span recorded in the Pilkhaua Subse- during the later stage while the eastern Altiplano shows quence (10 Ma, Figure 2), the above number of internal the opposite relationship. At 7–8 Ma, deformation ceased sequences amounts to substages of locally enhanced rates of everywhere. From these figures, bulk Cenozoic strain formation of relief at recurrence intervals of 2 ± 1 Ma for an rates in the southern Altiplano were at an average of individual structure. The San Vicente Formation shows a 3.5 1016 s1 with a middle Miocene maximum of 2 laterally varying number of 4–7 reflector packages of 1014 s1 (Figure 7c). similar properties as mentioned above spanning the time between 29 and 19 Ma (Figure 5a and Foldouts 1c, 5.3. Deformation Partitioning 1e, and 1f). This time span yields the same recurrence interval as for the Pilkhaua Subsequence. Hence the major [43] Inspection of the data at higher temporal and deformation periods in the Altiplano identified in Figures 6c spatial resolution underlines the complex mode of strain and 6d that typically encompass some 5–15 Myr are broken accumulation discussed in the following with two exam- down locally into shortening rate peaks at a typical ples. In the symmetrical syntectonic basin (Pilkhaua timescale of 1 to 3 Myr for individual structures. Because Subsequence) separating the adjacent Ines and Allka of the variable number of these internal sequences between Orkho faults (Figure 5), growth sediments record alter- neighboring structures, the structures within a coupled fault nating and contemporaneous activity of both faults: system may at best exhibit a partial synchronization. More stage 1, activity of Allka Orkho Fault and minor uplift probably, they record random, mostly unsynchronized of Ines fold (ST1); stage 2, growth of Ines fold, quiescent distribution of these fluctuations of shortening rate. Allka Orkho Fault (ST2); stage 3, growth of Allka Orkho Fault dominates over Ines fold (ST3); and stage 4, the Ines Fold again takes over. At the scale of the entire 6. Discussion central Altiplano unit, we observe Miocene initiation of first the east vergent KUFZ structures, succeeded by 6.1. Deformation Age and Pattern formation of new thrusts (14–11 Ma) in all parts of this [47] The new data in this study together with published thrust system. Finally, active deformation between 11 and data complete a cross section of the central Andean Plateau 8 Ma prograded westward with minor shortening in the and its margins at 20–22S, except for a 40-km-long internal parts of the system and a quiescent eastern segment of the Western Cordillera (the Recent magmatic deformation front. arc), where Miocene to Quaternary volcanic deposits pre- [44] The eastern Altiplano shows a similar evolution for vent access to the subsurface and no high-resolution reflec- the deformation front of the Eastern Cordillera (section tion seismic data are available. However, observations from

13 of 19 TC4020 ELGER ET AL.: ALTIPLANO DEFORMATION TC4020

Figure 6. Distribution of deformation ages, shortening, and magmatism across the southern central Andes (21S) based on published and own data. (a) Compilation of deformation ages: Western Flank [Victor et al., 2004], Precordillera [Haschke and Guenther, 2003], Altiplano [this study; Ege, 2004; Silva- Gonza´lez, 2004], Eastern Cordillera [Gubbels et al., 1993; Mu¨ller et al., 2002], inter-Andean [Kley, 1996; Ege, 2004], and sub-Andean [Kley, 1996]. (b) Balanced cross section at 21S compiled from Victor et al. [2004] for Altiplano West Flank; own data (Altiplano), and Mu¨ller et al. [2002] for Eastern Cordillera and sub-Andean; Moho and Andean Low-Velocity Zone (ALVZ) from receiver function data [Yuan et al., 2000]. Line drawing in the middle crust indicates locations of strong reflectivity in ANCORP seismic line [cf. ANCORP Working Group, 2003]. The western Altiplano cross section was projected along strike of gravity anomalies (Figure 1) 50 km southward. (c) Estimates of maximum and minimum periods of active faulting and related folding in the units building the Altiplano (see Foldout 1 for names and location of structures). (d) Cumulative shortening rates of the western, central, and eastern Altiplano thrust systems, calculated from fault activity periods and heave (see text for explanation; result for western Altiplano is extrapolated from observation of ages at Ya´zon peninsula to all faults observed in the seismic line considering the maximum possible error). Bold line delineates average shortening rates based on sliding average of 3 Myr sampling width. Shortening rate of the Precordillera from Victor et al. [2004] (complemented by Haschke and Guenther [2003]). For the Eastern Cordillera we estimated average shortening rates for individual subunits based on sections and age data summarized by Mu¨ller et al. [2002]. Because of the lack of resolution, there are both very few preserved piggyback basins and incomplete age dating of individual structures; the uncertainty here is higher than in the plateau area. See color version of this figure at back of this issue. exposed Neogene deposits of the Western Cordillera area the entire southern Altiplano Plateau is here identified as a farther north (18–19S[Garcı´a et al., 2002; Charrier et al., series of up to four doubly vergent thrust systems, which 2002]) show similar age properties as the Precordillera and vary in size and in amount of shortening (Figure 6). Most of the western Altiplano at 21S and may be projected to the these resulted from inversion of earlier extensional struc- unexposed part of the section at 21S. The structural style of tures. Inversion commenced in the Eocene in the west

14 of 19 TC4020 ELGER ET AL.: ALTIPLANO DEFORMATION TC4020

(Precordillera domain [Haschke and Guenther, 2003]) and only affected the protoplateau area farther east in the Oligocene. [48] Accumulation of shortening in the past 45 Myr (see Figures 6 and 7) initiated at the protoplateau margins, i.e., the Eastern Cordillera and Chilean Precordillera at around 40–45 Ma ago [e.g., Sempere et al., 1997; Lamb and Hoke, 1997; Mu¨ller et al., 2002; Haschke and Guenther, 2003] followed by a break in the west. The subsequent evolution shows peak activity to repeatedly shift between the Eastern Cordillera and the more central and western parts of the evolving plateau area. In the final stage at 11–8 Ma, contraction affected the entire southern Altiplano plateau area dying out everywhere at around 7–8 Ma. Shortening then migrated to the eastern flank of the plateau where eastward migrating deformation led to the formation of the broad, thin-skinned thrust belt of the inter-Andean and sub- Andean Ranges [Kley, 1996]. [49] Synchronization of peak shortening periods in the western systems (Altiplano west flank, western and central Altiplano) and their anticorrelation with those from the Eastern Cordillera, exhibiting repeated advance and retreat of its western deformation front that is unrelated to the remaining Altiplano, indicates kinematic coupling of various thrust systems: One is rooted in the Chilean Precordillera/western Altiplano and the other in the Eastern Cordillera. The existence of the here newly defined western Altiplano Thrust Belt (see Figures 8 and 6b), coupling the three western thrust units, is also corroborated by the here identified gently westward dipping detachment underlying the systems. These findings make the more eastern parts of this belt (western and central Altiplano thrust systems) a broad east facing, low-tapered prowedge accumulating most shortening (60 km). This prowedge is paired by the steeply inclined, slowly moving and narrow retrowedge forming the Western Flank Monocline/Precordillera of the Altiplano (3–4 km shortening [cf. Victor et al., 2004]). [50] Although the data imply significant synchronicity of deformation stages across parts of the plateau at time frames >5 Myr, more detailed inspection of the seismic sequence architecture of the syntectonic basins reveals significant short-term complexity with repeated retreat of the deforma- tion front (Eastern Cordillera) or repeated alternate thrust activation in the individual doubly vergent systems, and last, but not least, variations in local fault slip rates at local timescales of 1–3 Myr. Hence tectonic deformation rates would be seen to peak at various timescales between the isotopically resolved scales of several million years down to at least the subisotopic geological record exhibiting discon-

Figure 7. Cumulative shortening and shortening rate evolution of the southern Altiplano (sub-Andean plateau flank not included). (a) Summed minimum and maximum shortening rates across the southern Altiplano. Bold line depicts average based on sliding 3 Myr average. (b) Stacked average shortening rates of individual thrust systems building the southern Altiplano (from Figure 6). (c) Cumulative shortening.

15 of 19 TC4020 ELGER ET AL.: ALTIPLANO DEFORMATION TC4020

here reported details along the section at 21S may well be representative for most of the Altiplano to the north (com- pare to observations by Lamb and Hoke [1997]). Toward the south, the Puna plateau currently exhibits seismic activity across its entire area, bulk GPS-based shortening at some 6 mm/yr distributed over the entire Puna [Klotz et al., 2001], an increase of shortening rate since the Pliocene [Ramos et al.,2002],andthegrowthofthreetofour regularly spaced thrust systems across strike that confine intervening basins. This situation is highly similar to the late Miocene stage of synchronous plateau-wide shortening in the southern Altiplano prior to eastward expansion with an associated acceleration of rates. [53] Last, but not least, it is interesting to note that this evolution of shortening rates has a complex relation with the evolution of plate convergence rates between the South American and the oceanic Nazca plates [cf. Pardo-Casas and Molnar, 1987; Somoza, 1998]. Only the early, Eocene to Oligocene, evolution of upper plate shortening may be seen to correlate with two stages of enhanced plate conver- gence rates at 50 to 44 Ma and 25 to 20 Ma. In contrast, the decrease of Neogene convergence rate from some 15 cm/yr Figure 8. Outline of western Altiplano thrust belt and of at 20 Ma to the present value of 6.5 cm/yr [Klotz et al., the western deformation front of the Eastern Cordillera 2001] is anticorrelated with a shortening rate increase [cf. thrust belt drawn on shaded relief map of Altiplano Plateau. also Hindle and Kley, 2002]. We therefore speculate that Note narrow intervening basin between both belts (Lı´pez plate convergence rate was only a relevant factor controlling Basin (LB) in the south and Toledo Basin (TB) in the north) upper plate shortening rate in the initial stages. as well as restricted outcrop of prowedge of western Altiplano Thrust Belt in Corque (C), Tambo-Tambillo (TT), 6.2. Shortening, Crustal Thickening, and Surface Uplift and the here analyzed San Cristo´bal (SC) and Yazo´n (Y) [54] The distribution of upper crustal shortening concluded areas. from the two thrust systems building the plateau (western Altiplano thrust system, 65 km; Eastern Cordillera, tinuous deformation of the southern Altiplano crust over a 85 km [cf. Mu¨ller et al., 2002]) does not match an range of scales. equivalent thickness distribution of the deeper crust beneath these (compare Figure 6b). While there is slightly thicker [51] The pattern of plateau-wide bulk pure shear contrac- tion followed at around 10 Ma by a switch to bulk over- crust underlying the plateau margins, various geophysical thrusting of the plateau toward the east as identified by data (see Figure 6) [Wigger et al., 1994; Beck et al., 1996; Allmendinger and Gubbels [1996] is here refined by resolv- Yuan et al., 2000; ANCORP Working Group, 2003] indicate ing the pure shear stage into several substages with alternate that the plateau crust everywhere exceeds 60 km thickness. or overlapping activity of two major thrust systems. The Moreover, plateau surface uplift as recorded by Gregory- observed lack of any large-scale eastward or westward Wodzicki [2002] with enhanced rates since the late Miocene migrating deformation front over the entire plateau, i.e., is succeeding the main stage of upper crustal shortening in between the Western Flank and the Eastern Cordillera, is in the plateau domain. Hence upper crustal shortening as contrast to recently published models of continuous east- discussed here is geometrically as well as temporally ward plateau expansion based on an interpretation of the decoupled from deeper crustal and mantle lithosphere sedimentological record [McQuarrie and DeCelles, 2001; deformation. This is in line with late Neogene to Recent Horton et al., 2002; McQuarrie, 2002]. underthrusting of the plateau domain from the east and deep crustal flow as suggested by various authors [i.e., [52] The summed rates of all subsystems building the Altiplano highlight and reflect the important individual Allmendinger and Gubbels, 1996, and references therein]. stages in Altiplano deformation that were previously repeat- From the lack of a significant near surface deformation edly identified across the plateau as key deformation response to this stage, we conclude a near complete me- phases [e.g., Sempere et al., 1990; Lamb and Hoke, 1997; chanical decoupling of the upper crust from the deeper crust Allmendinger et al., 1997]. Their reinterpretation in the light since at least the late Miocene. of the here presented data would suggest these to be 6.3. Magmatism and Deformation fluctuations of shortening rates superimposed on the general trend of increase of bulk shortening velocity of the upper [55] The evolution of the Altiplano was paralleled by arc plate identified by Hindle and Kley [2002]. Moreover, the and back arc volcanic activity since the late Oligocene, widespread occurrence of these periods of enhanced short- largely coeval therefore to the period of shortening and ening rates throughout the plateau would suggest that the largely matching the lateral extent of the plateau. This link

16 of 19 TC4020 ELGER ET AL.: ALTIPLANO DEFORMATION TC4020 has been suggested to reflect a general genetic relationship [59] This conclusion and the above observations throw between magmatic activity and plateau-forming processes additional light on the dynamics of plateau formation. On in terms of thermally controlled weakening of the upper the basis of numerical modeling, the mechanical evolution plate [e.g., Isacks, 1988; Allmendinger et al., 1997; Kay et underlying plateau formation (see Wdowinski and Bock al., 1999; Babeyko et al., 2002; ANCORP Working Group, [1994], England and Molnar [1997], Pope and Willett 2003]. [1998], Shen et al. [2001], and Vanderhaeghe et al. [56] In the southern Altiplano, the initiation of volcanic [2003] for details) involves heating and related weakening activity (27–29 Ma) slightly postdates the onset of the first of the lower crust as a result of continuous crustal thicken- shortening increment (35–27 Ma) and the early stage of ing and melting. However, most of these models invoke an deposition of the San Vicente Formation (compare initial domain of localized deformation, which expands Figures 6a and 7a) (compare similar observations in other laterally forming classical self-similar growing wedges until parts of the plateau [e.g., Kay et al., 1999; Riller et al., mechanical decoupling at the base entails surface flattening 2001; Ramos et al., 2002]). Volcanism continued until 10– and the formation of a high plateau with two topographi- 8 Ma in the back arc plateau with ongoing magmatism in cally distinct margins bounding a hot, viscous deeper the present arc [Evernden et al., 1977; Fornani et al., lithosphere. The resulting plateau lacks a lateral mass 1993; Soler et al., 1993; Kennan et al., 1995], largely gradient forcing lateral migration of deformation. Subse- overlapping therefore with the main stage of shortening of quent shortening within the plateau would be delocalized the entire southern Altiplano domain. Hence it would with an irregular pattern in time and space coincident with appear that near surface magmatism in the southern the here reported observations. In contrast to the model Altiplano locally followed the onset of deformation with predictions, however, our results show that plateau initiation a minor time lag, but that it paralleled the general trend of in the Altiplano area occurred by growth of two indepen- increased shortening rates of the southern Altiplano pla- dent thrust systems at the site of the later plateau margins teau (compare Figure 7a). during the early stages. These do not show much evidence [57] On the basis of these observations, it appears plau- for lateral expansion until the sub-Andean ranges developed sible that the thermal evolution of the lithosphere resulting with concordant shutdown of deformation of the plateau. In in transient presence of melts may have sufficiently me- the case of the central Andes, first-order anisotropies chanically weakened the upper crust of the Altiplano to probably controlled this initial localization of deformation enhance global deformation rates through time as observed and, hence the subsequent evolution: the hot magmatic arc for the plateau (compare classical model by Isacks [1988]). crust in the kinematic axis of the western Altiplano thrust Modeling results by Babeyko et al. [2002] showed that system, and large-scale mechanical heterogeneity resulting advective melt transfer through the crust with associated from repeated rifting of the crust at the site of the later crustal melting requires some previous crustal shortening Eastern Cordillera. and thickening, which is equivalent to the correlation observed here. 7. Conclusions 6.4. Plateau-Style Deformation [60] The southern Altiplano structure formed during a [58] Along with the above observation of fluctuating complex shortening history that peaked in the early Oligo- shortening rates, the highly irregular spatial and temporal cene (<35–27 Ma) and middle/late Miocene (19–7 Ma). pattern of strain accumulation at all scales hints at a self- These stages were preceded by an extensional stage that similar distribution of deformation during plateau growth resulted in formation of a series of grabens across the entire since the Eocene with a lack of characteristic length and Altiplano, which later localized the formation of several timescales in the processes shaping the plateau. Fault slip, thrust systems. The three western systems, the west facing or periods of enhanced local strain rates, on a fault network Precordillera/Western Cordillera and the east facing western that is in failure equilibrium and has an approximately and central Altiplano systems probably form a coupled fractal size distribution (compare observations by Marrett retroprowedge system that is separate from the eastern and Allmendinger [1991] in the Puna) would be expected to Altiplano/Eastern Cordillera system. Both cordilleras delim- generate the type of scale-independent response found in the iting the later plateau area started shortening earlier during syntectonic sedimentary record of the Altiplano. It is the Eocene (40–45 Ma). Subsequently, their deformation obvious that the number of small tectonic events, at least history was decoupled with joint activity only in the middle partly reflected in individual layers deposited in the syntec- Miocene. Horizontal contraction on the southern Altiplano tonic basins, is significantly larger than the number of gradually subsided between 11 and 7 Ma as indicated by the reflection seismic sequences within each basin. Their num- age of undeformed volcanic rocks sealing the contractional ber, in turn, is lower than the few larger ‘‘displacement structures. events’’ building the Andes, like the rapid strong shortening [61] Deformation of the Andean Plateau at 21Sis of the Eastern Cordillera in the Miocene. These observa- characterized by a spatially and temporally irregular fault tions suggest that plateau deformation in the Andes is activity pattern, at a generally increasing bulk rate of plateau essentially accommodated by a series of interconnected shortening during the Oligocene/Miocene. Largely synchro- faults, which are close to failure equilibrium helping to nized peaks at the 5–15 Myr timescale within a kinemat- distribute strain. ically coupled thrust system were complemented by local

17 of 19 TC4020 ELGER ET AL.: ALTIPLANO DEFORMATION TC4020 fluctuations of strain accumulation at individual faults at a turn, may have been responsible for discontinuous acceler- typical timescale of 1–3 Myr that show no evidence of ation of upper plate shortening rates during slowdown of lateral synchronization. These observations preclude sys- convergence. tematically propagating deformation. The spatial and tem- poral pattern of deformation indicates that the plateau probably was self-organized fluctuating about a critical state [62] Acknowledgments. Research for this project was funded by the German Science Foundation (DFG) under the auspices of the Collaboration of failure throughout its evolution. This style of deformation Research Center 267 ‘‘Deformation Processes in the Andes’’ and the is fundamentally different from smaller-scale orogenic Leibniz-Program (On7/10-1). This support is gratefully acknowledged. A wedges that exhibit more continuous lateral growth. The great number of people have helped through discussion of our results as well as in finalizing the manuscript. First and foremost, the former Bolivian Altiplano plateau has probably had low relief since its Oil Industry (YPFB) provided the seismic sections in the context of the initiation and evolved from partly independent activity of ANCORP cooperation: David Tufin˜o and Ramiro Sua´rez (YPFB) as well as two thrust belt systems rather than from a single initially Oscar Aranibar and Eloy Martı´nez (Chaco SA) provided valuable support in seismic data exchange and interpretation; K. Munier and P. Goˆni doubly vergent orogen. Previously suggested thermal weak- processed the satellite images; M. Dziggel has drawn most of the figures; ening of the plateau crust is reflected in a loose spatial and W. Frank (Vienna University) has provided the Ar/Ar age data; T. Vietor, temporal relationship between deformation and, slightly K.-J. Reutter, D. Mertmann, and Herna´n Castillo have supported field work; later onset of, volcanism. Eocene to Oligocene crustal J. Herwig has prepared samples for age dating; and H. Ege, E. Scheuber, and P. Silva-Gonza´lez provided important additional age data. Last, but not shortening with some relation to plate convergence rates least, G. Wo¨rner and an anonymous reviewer provided thoughtful reviews. preceded thermal weakening of the crust. The latter, in We thank all these colleagues for their collaboration and help.

References Allmendinger, R. W., and T. Gubbels (1996), Pure Ege, H. (2004), Exhumationsgeschichte des Altiplano Haschke, M., and A. Guenther (2003), Balancing crus- and simple shear plateau uplift, Altiplano-Puna, und der Ostkordillere Su¨d-Boliviens anhand von tal thickening in arcs by tectonic vs. magmatic Argentina and Bolivia, Tectonophysics, 259(1– Apatit-Spaltspur-Datierungen, Ph.D. thesis, 159 means, Geology, 31, 933 – 936. 3), 1 – 13. pp., Freie Univ., Berlin, Germany. Hindle, D., and J. Kley (2002), Displacements, strains Allmendinger, R. W., T. E. Jordan, S. M. Kay, and B. England, P., and P. Molnar (1997), Active deformation and rotations in the central Andean plate boundary, L. Isacks (1997), The evolution of the Altiplano- of Asia: From kinematics to dynamics, Science, in Plate Boundary Zones, Geodyn. Ser., vol. 30, Puna Plateau of the central Andes, Annu. Rev. 278, 647 – 650. edited by S. Stein and J. T. Freymueller, pp. Earth Planet. Sci., 25, 139 – 174. Epard, J.-L., and R. Groshong Jr. (1993), Excess area 135 – 144, AGU, Washington, D. C. ANCORP Working Group (2003), Seismic imaging of and depth to detachment, AAPG Bull., 77, 1291 – Horton, B. K., B. A. Hampton, and G. L. Waanders a convergent continental margin and plateau in the 1302. (2001), Paleogene synorogenic sedimentation in central Andes (Andean Continental Research Pro- Evernden, J. F., S. J. Kriz, and M. C. Cherroni (1977), the Altiplano Plateau and implications for initial ject 1996 (ANCORP’96)), J. Geophys. Res., Potassium-argon ages of some Bolivian rocks, mountain building in the central Andes, Geol. 108(B7), 2328, doi:10.1029/2002JB001771. Econ. Geol., 72, 1042 – 1061. Soc. Am. Bull., 113, 1387 – 1400. Babeyko, A. Y., S. V. Sobolev, R. B. Trumbull, O. Fiedler, K., D. Mertmann, and V. Jacobshagen (2003), Horton, B. K., B. A. Hampton, B. N. Lareau, and E. Oncken, and L. L. Lavier (2002), Numerical mod- Cretaceous marine ingressions in the southern Po- Baldellon (2002), Tertiary provenance history of els of crustal scale convection and partial melting tosı´ basin of southern Bolivia: Tectonic and eu- the northern and central Altiplano (central Andes, beneath the Altiplano-Puna Plateau, Earth Planet. static control, Rev. Tecn. Yacimientos Petrol. Bolivia): A detrital record of plateau-margin tec- Sci. Lett., 199(3 – 4), 373 – 388. Fiscales Bolivianos, 21, 157 – 164. tonics, J. Sediment. Res., 72(5), 711 – 726. Baby, P., P. Rochat, G. Mascle, and G. Herail (1997), Fornani, M., L. Pozzo, L. Bailly, J. Leroy, and M. G. Isacks, B. L. (1988), Uplift of the central Andean Pla- Neogene shortening contribution to crustal thicken- Bonhomme (1993), Miocene volcanic centers in teau and bending of the Bolivian Orocline, J. Geo- ing in the back arc of the central Andes, Geology, the southern Altiplano of Bolivia: The Cerro Mor- phys. Res., 93, 3211 – 3231. 25, 883 – 886. okho ad Cerro Bonete area (Sur Lı´pez), paper pre- James, D. E. (1971), Plate tectonic model for the evo- Baker, M. C. W., and P. W. Francis (1978), Upper sented at Second ISAG, Colloques et Seminaires, lution of the central Andes, Geol. Soc. Am. Bull., Cenozoic volcanism in the central Andes: Ages Inst. de Rec. Sci. pour le De´v. en Coop., Oxford, 82, 3325 – 3346. and volumes, Earth Planet. Sci. Lett., 41(2), U.K. Jordan, T. E., and R. N. Alonso (1987), Cenozoic stra- 175 – 187. Garcı´a, M., G. He´rail, R. Charrier, G. Mascle, M. tigraphy and basin tectonics of the Andes Moun- Beck, S. L., G. Zandt, S. C. Myers, T. C. Wallace, P. G. Fornari, and C. Perez de Arce (2002), Oligo- tains, 20 –28 south latitude, AAPG Bull., 71,49– Silver, and L. Drake (1996), Crustal-thickness var- cene-Neogene tectonic evolution of the Altiplano 64. iations in the central Andes, Geology, 24,407– of northern Chile (18 – 19S), paper presented at Jordan, T. E., M. Burns, R. Veiga, F. Pangaro, S. Kel- 410. Andean Geodynamics, 5th International Sympo- ley, and C. Mpodozis (2001), Extension and basin Bevis, M., E. C. Kendrick, R. Smalley Jr., T. Herring, sium, Inst. de Rech. et Dev., Toulouse, France, formation in the southern Andes caused by in- J. Godoy, and F. Galban (1999), Crustal motion 16 – 18 Sept. creased convergence rate: A mid-Cenozoic trigger north and south of the Arica deflection: Comparing Go¨tze, H.-J., and S. Krause (2002), The central Andean for the Andes, Tectonics, 20, 308 – 324. recent geodetic results from the central Andes, gravity high, a relic of an old subduction complex?, Kay, S., C. Mpodozis, and B. Coira (1999), Neogene Geochem. Geophys. Geosyst., 1(1), doi:10.1029/ J. S. Am. Earth Sci., 14(8), 799 – 811. magmatism, tectonism, and mineral deposits of the 1999GC000011. Grant, J. N., C. Halls, W. A. Salinas, and N. J. Snelling central Andes (22 to 33S latitude), in Geology Brasse, H., P. Lezaeta, V. Rath, K. Schwalenberg, W. (1979), K-Ar ages of igneous rocks in mineraliza- and Ore Deposits of the Central Andes, edited by Soyer, and V. Haak (2002), The Bolivian Altiplano tion in part of the Bolivian tin belt, Econ. Geol., 74, B. J. Skinner, Spec. Publ. Soc. Econ. Geol., 7,27– conductivity anomaly, J. Geophys. Res., 107(B5), 838 – 851. 59. 2096, doi:10.1029/2001JB000391. Gregory-Wodzicki, K. M. (2002), A late Miocene sub- Kelley, S. (2002), Excess argon in K-Ar and Ar-Ar Charrier, R., A. Chavez, S. Elgueta, G. He´rail, J. Flynn, tropical-dry flora from the northern Altiplano, Bo- geochronology, Chem. Geol., 188(1 – 2), 1 – 22. D. Croft, A. Wyss, and M. Garcı´a (2002), Rapid livia, Palaeogeogr. Palaeoclimatol. Palaeoecol.,, Kennan, L., S. Lamb, and C. Rundle (1995), K-Ar tectonic and paleogeographic evolution of the Chu- 180(4), doi:10.1016/S0031-0182(01)00434-5. dates from the Altiplano and Cordillera Oriental cal Anticline, Altiplano of Arica, northern Chile, Gubbels, T. L., B. L. Isacks, and E. Farrar (1993), of Bolivia: Implications for Cenozoic stratigraphy paper presented at Andean Geodynamics, 5th Inter- High-level surfaces, plateau uplift, and foreland and tectonics, J. S. Am. Earth Sci., 8(2), 163 – 186. national Symposium, Inst. de Rech. et Dev., Tou- development, Bolivian central Andes, Geology, Kley, J. (1996), Transition from basement-involved to louse, France, 16 – 18 Sept. 21, 695 – 698. thin-skinned thrusting in the Cordillera Oriental of Dewey, J. F., R. M. Shackleton, C. Chang, and Y. Sun Haberland, C., and A. Rietbrock (2001), Attenuation southern Bolivia, Tectonics, 15, 763 – 775. (1988), The tectonic evolution of the Tibetan Pla- tomography in the western central Andes: A de- Kley, J., and C. R. Monaldi (1998), Tectonic shorten- teau, Philos. Trans. R. Soc. London, Ser. A, 327, tailed insight into the structure of a magmatic arc, ing and crustal thickness in the central Andes: How 379 – 413. J. Geophys. Res., 106, 11,151 – 11,167. good is the correlation?, Geology, 26, 723 – 726.

18 of 19 TC4020 ELGER ET AL.: ALTIPLANO DEFORMATION TC4020

Klotz, J., G. Khazaradze, D. Angermann, C. Reigber, dillera, southern Bolivia (21S), Tectonics, 21(5), Seminaires, Inst. de Rech. Sci. pour le De´v. en R. Perdomo, and O. Cifuentes (2001), Earthquake 1037, doi:10.1029/2001TC001340. Coop., Oxford, U.K. cycle dominates contemporary crustal deformation Pardo-Casas, F., and P. Molnar (1987), Relative motion Somoza, R. (1998), Updated Nazca (Farallon)-South in central and southern Andes, Earth Planet. Sci. of the Nazca (Farallon) and South American plates America relative motions during the last 40 My: Lett., 193, 437 – 446. since Late Cretaceous time, Tectonics, 6, 233 – 248. Implications for mountain building in the central Lamb, S. (2000), Active deformation in the Bolivian Pope, D.-C., and S. D. Willett (1998), Thermal-me- Andean region, J. S. Am. Earth Sci., 11(3), 211 – Andes, South America, J. Geophys. Res., 105, chanical model for crustal thickening in the central 215. 25,627 – 25,653. Andes driven by ablative subduction, Geology, 26, Swenson, J. L., S. L. Beck, and G. Zandt (2000), Crus- Lamb, S., and L. Hoke (1997), Origin of the high 511 – 514. tal structure from broadband regional waveform plateau in the central Andes, Bolivia, South Amer- Ramos, V., E. O. Cristallini, and D. J. Pe´rez (2002), modeling: implications for the composition of thick ica, Tectonics, 16, 623 – 649. The Pampean flat slab segment of the central An- continental crust, J. Geophys. Res., 105, 607 – 621. Lamb, S., L. Hoke, L. Kennan, and J. Dewey (1997), des, J. S. Am. Earth Sci., 15, 59 – 78. Vanderhaeghe, O., S. Medvedev, P. Fullsack, C. Beau- Cenozoic evolution of the central Andes in Bolivia Reutter, K.-J., R. Do¨bel, T. Bogdanic, and J. Kley mont, and R. A. Jamieson (2003), Evolution of and northern Chile, in Orogeny Through Time,edi- (1994), Geological map of the central Andes be- orogenic wedges and continental plateaux: Insights ted by J. P. Burg and M. Ford, Geol. Soc. Spec. tween 20 and 26S, in Tectonics of the Southern from al thermal-mechanical models overlying sub- Publ., 121, 237 – 264. Central Andes, edited by K.-J. Reutter, E. Scheu- ducting mantle lithosphere, Geophys. J. Int., 153, Marrett, R., and R. Allmendinger (1991), Estimates of ber, and P. J. Wigger, Springer, New York. 27 – 51. strain due to brittle faulting; sampling of fault po- Riller, U., I. Petrinovic, J. Ramelow, M. Strecker, and Victor, P., O. Oncken, and J. Glodny (2004), Uplift of pulations, J. Struct. Geol., 13, 735 – 738. O. Oncken (2001), Late Cenozoic tectonism, col- the western Altiplano plateau: Evidence from the Martı´nez, C., E. Soria, H. Uribe, A. Escobar, and A. lapse caldera, and plateau formation in the central Precordillera between 20 and 21S (northern Hinojosa (1994), Estructura y evoluciu¨n del Alti- Andes, Earth Planet. Sci. Lett., 188, 299 – 311. Chile), Tectonics, 23, TC4004, doi:10.1029/ plano suroccidental: El sistema de cabalgamiento Rochat, P., G. Herail, P. Baby, and G. Mascle (1999), 2003TC001519. de Uyuni-Khenayani y su relaciu¨n con la sedimen- Bilan crustal et controˆle de la dynamique e´rosive et Wdowinski, S., and Y. Bock (1994), The evolution of taciu¨n Terziaria, Rev. Tec. Yacimientos Petrol. Fis- se´dimentaire sur les me´canismes de formation de deformation and topography of high elevated pla- cales Bolivianos, 15(3 – 4), 247 – 264. l’Altiplano, C. R. Acad. Sci., Ser.e II, 328(3), 189 – teaus: 1. Model, numerical analysis, and general McQuarrie, N. (2002), The kinematic history of the 195. results, J. Geophys. Res., 99, 7103 – 7119. central Andean fold-thrust belt, Bolivia: Implica- Sempere, T., G. Herail, J. Oller, and M. G. Bonhomme Welsink, H. J., E. Martinez, O. Aranibar, and J. Jaran- tions for building a high plateau, Geol. Soc. Am. (1990), Late Oligocene-early Miocene major tec- dilla (1995), Structural inversion of a Cretaceous Bull., 114, 950 – 963. tonic crisis and related basins in Bolivia, Geology, rift basin, southern Altiplano, Bolivia, in Petroleum McQuarrie, N., and P. DeCelles (2001), Geometry and 18, 946 – 949. Basins of South America, edited by A. J. Tankard, structural evolution of the central Andean back- Sempere, T., R. F. Butler, D. R. Richards, L. G. Mar- R. S. Sua´rez, and H. J. Welsink, AAPG Mem., 62, thrust belt, Bolivia, Tectonics, 20, 669 – 692. shall, W. Sharp, and C. C. Swisher III (1997), Stra- 305 – 324. Mertmann, D., E. Scheuber, P. Silva-Gonza´lez, and K.- tigraphy and chronology of Upper Cretaceous/ Wigger, P., et al. (1994), Variations of the crustal struc- J. Reutter (2003), Tectono-sedimentary evolution lower Paleogene strata in Bolivia and northwest ture of the southern central Andes deduced from of the southern Altiplano: Basin evolution, thermo- Argentina, Geol. Soc. Am. Bull., 109, 709 – 727. seismic refraction investigations, in Tectonics of the chronology, and structural geology, Rev. Tec. Yaci- Shen, F., L.-H. Royden, and B. C. Burchfiel (2001), Southern Central Andes, edited by K.-J. Reutter, E. mientos Petrol. Fiscales Bolivianos, 21, 17 – 22. Large-scale crustal deformation of the Tibetan Pla- Scheuber, and P. J. Wigger, pp. 23 – 48, Springer, Miller, J. S., M. T. Heizler, and C. F. Miller (1998), teau, J. Geophys. Res., 106, 6793 – 6816. New York. Timing of magmatism, basin formation, and tilting Silva-Gonza´lez, P. (2004), Der su¨dliche Altiplano im Yuan, X., et al. (2000), Subduction and collision pro- at the west edge of the Colorado River extensional Tertia¨r: Sedimenta¨re Entwicklung und tektonische cesses in the central Andes constrained by con- corridor: results from single crystal 40Ar/39Ar geo- Implikationen, Ph.D. thesis, Freie Univ., Berlin, verted seismic phases, Nature, 408, 958 – 961. chronology of Tertiary rocks in the Old Woman Germany. Mountains area, southeastern California, J. Geol., Soler, P., G. Carlier, M. G. Bonhomme, and M. Fornani 106, 195 – 209. (1993), Field observations and K-Ar dating of the K. Elger, J. Glodny, and O. Oncken, GeoForschungs- Mu¨ller, J. P., J. Kley, and Jacobshagen (2002), Struc- Cerro Chiar Kkollu (southern Altiplano, Bolivia), Zentrum Potsdam, Telegrafenberg, D-14473 Potsdam, ture and Cenozoic kinematics of the Eastern Cor- paper presented at Second ISAG, Colloques et Germany. ([email protected])

19 of 19 TC4020 ELGER ET AL.: ALTIPLANO DEFORMATION TC4020

Figure 1

2of19 TC4020 ELGER ET AL.: ALTIPLANO DEFORMATION TC4020

Figure 1. (a) Geological sketch map of the southern Altiplano (section of map of central Andes by Reutter et al. [1994] including our own mapping results) superimposed on shaded relief of the topography showing the distribution of pre-late Neogene rocks, lower-hemisphere equal-area stereonet projections with bedding and regional fold axes, and the location of reflection seismic lines. (b) Bouguer anomaly map (provided by H. J. Go¨tze; see Go¨tze and Krause [2002] for description of gravity field) of area shown in Figure 1a depicting positive anomalies closely related to faults and folds mapped at the surface and observed in the seismic sections. Note the chain of anomalies in the nearly completely sediment-covered western Altiplano with similar style as obvious from the exposed central Altiplano unit. (c) Location of the working area. (d) Major morpho-tectonic units of the southern central Andes.

3of19 TC4020 ELGER ET AL.: ALTIPLANO DEFORMATION TC4020

Figure 6. Distribution of deformation ages, shortening, and magmatism across the southern central Andes (21S) based on published and own data. (a) Compilation of deformation ages: Western Flank [Victor et al., 2004], Precordillera [Haschke and Guenther, 2003], Altiplano [this study; Ege, 2004; Silva- Gonza´lez, 2004], Eastern Cordillera [Gubbels et al., 1993; Mu¨ller et al., 2002], inter-Andean [Kley, 1996; Ege, 2004], and sub-Andean [Kley, 1996]. (b) Balanced cross section at 21S compiled from Victor et al. [2004] for Altiplano West Flank; own data (Altiplano), and Mu¨ller et al. [2002] for Eastern Cordillera and sub-Andean; Moho and Andean Low-Velocity Zone (ALVZ) from receiver function data [Yuan et al., 2000]. Line drawing in the middle crust indicates locations of strong reflectivity in ANCORP seismic line [cf. ANCORP Working Group, 2003]. The western Altiplano cross section was projected along strike of gravity anomalies (Figure 1) 50 km southward. (c) Estimates of maximum and minimum periods of active faulting and related folding in the units building the Altiplano (see Foldout 1 for names and location of structures). (d) Cumulative shortening rates of the western, central, and eastern Altiplano thrust systems, calculated from fault activity periods and heave (see text for explanation; result for western Altiplano is extrapolated from observation of ages at Ya´zon peninsula to all faults observed in the seismic line considering the maximum possible error). Bold line delineates average shortening rates based on sliding average of 3 Myr sampling width. Shortening rate of the Precordillera from Victor et al. [2004] (complemented by Haschke and Guenther [2003]). For the Eastern Cordillera we estimated average shortening rates for individual subunits based on sections and age data summarized by Mu¨ller et al. [2002]. Because of the lack of resolution, there are both very few preserved piggyback basins and incomplete age dating of individual structures; the uncertainty here is higher than in the plateau area.

14 of 19