Clumped isotope evidence for diachronous surface cooling of the Altiplano and pulsed surface uplift of the Central Andes

Earth and Planetary Science Letters 393 (2014) 173–181

Contents lists available at ScienceDirect

Earth and Planetary Science Letters www.elsevier.com/locate/epsl

Clumped isotope evidence for diachronous surface cooling of the Altiplano and pulsed surface uplift of the Central Andes Carmala N. Garzione a,∗ , David J. Auerbach a,1 , Johanna Jin-Sook Smith a,2 , Jose J. Rosario b , Benjamin H. Passey c , Teresa E. Jordan b , John M. Eiler d a

Department of Earth and Environmental Sciences, University of Rochester, Rochester, NY 14627, United States Department of Earth and Atmospheric Sciences, Cornell University, Ithaca, NY 14853, United States c Department of Earth and Planetary Sciences, Johns Hopkins University, Baltimore, MD 21218, United States d Division of Geological and Planetary Sciences, California Institute of Technology, Pasadena, CA 91125, United States b

a r t i c l e

i n f o

Article history: Received 30 September 2013 Received in revised form 11 February 2014 Accepted 12 February 2014 Available online 15 March 2014 Editor: T.M. Harrison Keywords: Andean plateau paleoelevation Altiplano clumped isotopes

a b s t r a c t Spatially extensive paleoelevation records of the Altiplano plateau are critical to determining the geodynamic mechanisms that formed and support high elevations over a broad area. Prior stable isotope data reveal a climate history for the northern Bolivian Altiplano that has been interpreted to show rapid surface uplift of 2.5 ± 1.0 km between ∼10 and 6 Ma. This study applies clumped isotope paleothermometry to paleosol carbonates formed at both a low-elevation site and temporally overlapping high-elevation sites in the southern Altiplano/Eastern Cordillera during the middle to late Miocene. Surface paleotemperature decreased by 14 ◦ C in the southern Altiplano/Eastern Cordillera relative to stable low-elevation paleotemperatures, implying surface elevation increase of 1.9 ± 0.7 km between 16 and 13 Ma and an additional 0.7 ± 0.6 km between 13 and 9 Ma. Both the large magnitude of surface temperature decrease and earlier onset (7 ± 4 Myr) in the south as compared to the north suggest rapid elevation increase by piecemeal removal of lower lithosphere beneath the plateau and possible northward lower crustal flow. © 2014 Elsevier B.V. All rights reserved.

1. Introduction The Altiplano plateau of western South America lies within the Central Andes above the subducting Nazca plate. The region stretches from ∼14 ◦ S to 22 ◦ S, reaches 300 km wide (measured approximately east to west), and has an average elevation of ∼4 km. It is enclosed by the Western Cordillera magmatic arc and Eastern Cordillera fold-thrust belt (Fig. 1), whose development has left the Altiplano internally drained since at least late Oligocene time (Horton et al., 2001). This resulted in the accumulation of an extensive sedimentary archive that reflects the Altiplano’s Neogene paleoclimate and surface uplift history. This record has the potential to inform us about the geodynamic processes that raise orogenic plateaus. The possible surface elevation histories can be described by two end-member models: rapid, substantial pulses of uplift vs. gradual, continuous uplift

*

Corresponding author. E-mail address: [email protected] (C.N. Garzione). 1 Now at Department of Geology and Geophysics, Yale University, New Haven, CT 06520, United States. 2 Now at ExxonMobil Corp., Houston, TX 77002, United States. http://dx.doi.org/10.1016/j.epsl.2014.02.029 0012-821X/© 2014 Elsevier B.V. All rights reserved.

(e.g., Barnes and Ehlers, 2009). Rapid surface uplift of portions of the Andes on the order of 1 km over no more than several million years would require specific geodynamic processes, such as the loss of the lower lithosphere (e.g., Houseman et al., 1981) or lower crustal flow (e.g., Husson and Sempere, 2003), whereas gradual surface uplift that reflects the rate of crustal shortening and thickening must be accompanied by continuous removal of the dense lower lithosphere via a process such as ablative subduction (Pope and Willett, 1998). 1.1. Existing paleoelevation constraints Previous to this study, few measurements of paleo-surface elevations constrain the processes responsible for the surface uplift of the Altiplano and the Eastern Cordillera, and those that do exist are solely from the north-central Altiplano and Eastern Cordillera (16 ◦ S to 19 ◦ S). The oldest information comes from 73–60 Ma coastal marine deposits, indicating that the region was at or near sea level at that time (Sempere et al., 1997). Structural and sedimentological data indicate that the Eastern Cordillera crust began to thicken by crustal shortening by at least Eocene time (DeCelles and Horton, 2003; Isacks, 1988; McQuarrie,

174

C.N. Garzione et al. / Earth and Planetary Science Letters 393 (2014) 173–181

Fig. 1. Topography of the central Andes between 15◦ and 25 ◦ S (SRTM30 dataset) showing physiographic divisions in italics, modified from McQuarrie (2002). Localities (red circles) are sections sampled in this study; white circles designate localities from northern Altiplano paleotemperature studies (Ghosh et al., 2006b; Gregory-Wodzicki, 2002; Gregory-Wodzicki et al., 1998). Diamonds identify major cities in Bolivia.

2002), likely resulting in some surface uplift. Thermochronometric data indicate that this shortening resulted in significant exhumation (∼4 to 7 km) of the Eastern Cordillera (Gillis et al., 2006; Barnes et al., 2008). Despite significant shortening and exhumation of the Eastern Cordillera, climate proxy data—paleotemperatures and the oxygen isotope composition of paleo-precipitation—suggest the Eastern Cordillera remained at relatively low elevations of 0–1.5 km in the late Oligocene to earliest Miocene time (Bershaw et al., 2010; Leier et al., 2013) and rose to ∼2.5 to 3 km in between 24 and 17 Ma (Leier et al., 2013). The mid-Miocene to earliest late Miocene record (∼11 to 10 Ma) of paleoelevation of the Altiplano, including leaf physiognomy, paleotemperatures, and oxygen isotope composition of paleo-precipitation, places the north-central Altiplano at low-elevations of 2 km. By ∼6 Ma, climate proxy estimates indicate the north-central Altiplano had reached nearmodern elevations (Bershaw et al., 2010; Garzione et al., 2006; Ghosh et al., 2006b). In combination, these studies suggest that the Eastern Cordillera experienced an earlier (early-middle Miocene) pulse of surface uplift while the adjacent Altiplano rose in late Miocene time. The record that emerges from data in the north-central Altiplano and Eastern Cordillera is one of significant regional climate change—cooling temperatures, changing oxygen isotope composition of precipitation, and shifting ecology in middle-late Miocene time—that is consistent with different pulses of surface uplift of the Eastern Cordillera and Altiplano. However, recent general circulation modeling (GCM) experiments have suggested that the response of regional climate to surface uplift is non-linear, due to significant thresholds in the response of climate to rising surface topography (Ehlers and Poulsen, 2009; Insel et al., 2012; Poulsen et al., 2010). The numerical results of these GCM experiments suggest that the surface elevation history of the Altiplano may be more gradual when threshold climate response is considered. Clarifying the nature and pace of the surface uplift history of the Altiplano requires better spatial resolution in the climate record; the existing record is limited almost entirely to the northern Altiplano. Here we expand the spatial extent of the paleoelevation record for the central Andes by determining middle to late Miocene surface paleotemperatures in the southern Altiplano and Eastern Cordillera (19 ◦ S to 22 ◦ S). The magnitude and temporal evolution of surface cooling across this region provides in-

sights into whether cooling was related to surface uplift or climate change. 2. Methods Conventional stable isotope approaches (e.g., O and H) provide a robust proxy of paleoelevation in regions that receive significant rainfall (>30 cm/yr) and show limited complexity in the source of vapor masses (e.g., Rowley and Garzione, 2007). In the arid southern Altiplano, however, climate simulations show high variability in O isotopic composition of rainfall and a less pronounced δ 18 O versus elevation gradient, with rainfall composition primarily influenced by precipitation amount (Insel et al., 2013). Under extremely arid conditions (<30 cm rainfall/yr), pedogenic carbonates, a proxy for paleo-rainfall composition, have been shown to strongly deviate from the isotopic composition of local meteoric water as a result of extreme evaporative enrichment of 18 O in soil water (Quade et al., 2007). In order to avoid the complexity of the lack of correspondence between δ 18 O values of rainfall and elevation, as well as soil–water evaporation associated with the aridity of the southern Altiplano (Garreaud et al., 2003; Minvielle and Garreaud, 2011) that would both tend to lead to a bias toward under-estimation of paleoelevation, this study is focused on the use of “clumped C–O isotope thermometry” that constrains the temperature of soil carbonate formation. The degree of “clumping” between 13 C and 18 O in carbonates (expressed as the parameter (47 )) is related to the temperature of carbonate formation (Eiler, 2007, 2011; Ghosh et al., 2006a). The carbonate clumped isotope thermometer can be applied to paleosol and lacustrine carbonates to retrieve information about surface paleotemperatures (Ghosh et al., 2006b; Huntington et al., 2010; Passey et al., 2010) that can be used to quantify climate change and elevation. We investigated paleosols in three high-elevation middle-late Miocene exposures of terrestrial sedimentary rocks (Cerdas, Quebrada Honda, and Quehua). Nodular paleosol carbonates were sampled at depths of ∼0.2 to 2 m below the tops of paleosols, where paleosol tops could be identified, in an effort to limit the magnitude of seasonal and diurnal variation in ground temperature (Fig. 2). Based on the Mack et al. (1993) classification scheme, the paleosols include calcisols and argillic calcisols. The Cerdas section includes argillic calcisols. The tops of paleosols are defined by

C.N. Garzione et al. / Earth and Planetary Science Letters 393 (2014) 173–181

175

Fig. 2. Miocene paleosols and Holocene soil from the Altiplano/Eastern Cordillera. Arrows point out the tops of paleosols where they are observable and vertical lines delineate the pedogenic carbonate (Bk ) horizon. A) Paleosol within the Cerdas section. Carbonates nodules ∼0.5 to 2 cm in diameter occur in the Bk horizon. B) Paleosol within the Cerdas section. The Bk horizon is ∼30 cm thick and contains carbonate nodules ∼1 to 2 cm in diameter. C) Paleosol within the Quehua section. Pedogenic carbonate nodules in the upper 5 cm of the Bk horizon have merged to form a carbonate bed. Below, nodules are 0.5 to 2 cm in diameter. D) Pleistocene–Holocene soil near the Quehua section. Bk horizon reflects diffuse carbonate and carbonate coatings on root traces. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

scour surfaces at the base of lenticular sand or conglomerate beds interpreted as fluvial channel deposits. At up to 2 m thick, these paleosols are the thickest observed within the three high altitude sections. The B horizon displays root traces and green-gray mottling within the upper reddish-brown argillic B horizon. Toward the top of the B horizon, there is decreasing carbonate content, decreasing grain size, increasing clay content, and increasing red coloration, with the upper ∼30 cm leached of carbonate. Paleosols from the Quebrada Honda section are much more calcareous (i.e., calcisols) and display nodular to bedded carbonate horizons. Because of sparse deposition of coarser channel deposits that typically define the tops of paleosols in fluvial/floodplain settings and

the lack of development of a well-defined argillic horizon, the tops of soils are difficult to discern and sampling depths are therefore defined based on the top of the carbonate horizon. Soil features include sparse mottling and abundant root traces. Toward the top of the section, the calcisols become increasingly bedded, consistent caliche formation at or near the surface. Quehua paleosols are thin (∼20 to 30 cm thick), nodular calcisols that contain sparse to abundant root traces and white to light grey mottling. Paleosols occur within very thick mudstone beds, and therefore their tops are difficult to distinguish. The extent of carbonate, root traces, and mottling are used to define the thickness and hence the top of the paleosol.

176

C.N. Garzione et al. / Earth and Planetary Science Letters 393 (2014) 173–181

Thin sections of paleosol carbonates were studied to ascertain the presence of diagenetic phases. We chose paleosol carbonates with no signs of sparry calcite. Several of the Subandes paleosols showed sparite. For those samples, we used a dental drill to sample primary micrite, avoiding spar. Samples were reacted in 100% H3 PO4 at 90 ◦ C in a common acid bath device (Passey et al., 2010). The evolved CO2 was purified by passage through multiple dry ice/ethanol traps (∼ −75 ◦ C) and a Porapak-Q gas chromatography column held at −10 ◦ C. Isotopologue ratios of purified CO2 were measured using a Thermo-Finnigan MAT 253 mass spectrometer, and all reported 47 values (a measure of enrichment in mass-47 CO2 , primarily 13 C18 O16 O) were normalized relative to high-temperature equilibrium gases (Huntington et al., 2009). These data were generated prior to the introduction of the ‘absolute reference frame’ of Dennis et al. (2011) and so are calibrated relative to the Caltech intralab reference frame of Huntington et al. (2009). Analytical errors are determined from the standard deviation of carbonate standards analyzed multiple times during the run sessions for the samples reported in Table 2. For two analytical sessions, the 2σ error of standards was 0.038h, slightly worse than the long-term laboratory average 2σ of 0.027h. Generic 95% confidence intervals for 47 were computed from the 1σ precision of standards and the number of samples N in each section (95% CI (1.96 × 1σ × ( N )−0.5 ). This uncertainty was propagated through the Ghosh et al. (2006a) temperature equation to arrive at the 2σ temperature uncertainty. 2.1. Temperature–elevation relationship

derived from paleosol temperatures (Quade et al., 2013). A step-bystep approach to calculating paleoelevation using this method, as well as the associated errors, is laid out in the discussion section. 3. Paleotemperature results

Analysis of paleosol carbonate gives the temperature of a soil during carbonate formation, the timing of which depends on local climate, vegetation, solar irradiance of the ground surface, and the seasonality of precipitation (Passey et al., 2010; Peters et al., 2013; Quade et al., 2011, 2013), factors discussed in Section 4. Using paleo-surface temperatures derived from paleosols to estimate paleoelevation requires an understanding of the temperature–elevation relationship. The modern air temperature– elevation lapse rate (−5 ◦ C/km) was derived from weather stations on the eastern flank of the central Andes (Gonfiantini et al., 2001). Regional GCMs of mean surface temperatures of the Altiplano plateau at various stages in its development (75%, 50%, 25%, 0% modern elevation) show that when the plateau was lower, climate (particularly atmospheric circulation) would have been different, resulting in changes to the air temperature–elevation gradient (Ehlers and Poulsen, 2009). These results indicate that lowering the Andes to 50% (2 km), 25% (1 km), and 0% (0 km) of their modern elevation yields mean temperatures that are warmer by +0.6 ◦ C, +3.4 ◦ C, and +6.6 ◦ C, respectively, than the modern lapse rate would indicate (Ehlers and Poulsen, 2009). We sum the weather station temperature data (Gonfiantini et al., 2001) with these model-based corrections to determine an “orographicallycorrected” temperature lapse rate (Fig. 3). The baseline temperature at 0 km is set by the MAAT of 28 ◦ C determined from middle to late Miocene paleosol carbonates from the Subandes at a location (Rio Iruya) that has remained near sea level during its depositional history (Tables 1 and 2) to yield the second degree polynomial equation:

y = (1.228 ± 0.898)(x − 14.62)2 − (146.8 ± 9.4)x

+ (3912 ± 211)

Fig. 3. Modern (uncorrected) temperature lapse rate of −5.0 ◦ C/km (Gonfiantini et al., 2001), and orographically-corrected (uplift-corrected) lapse rate (see Sections 2 and 4 for details). Open squares show mean annual air temperature (MAAT) data from weather stations on the eastern flank of the Andes (Gonfiantini et al., 2001). Gray diamonds show MAAT that have been adjusted 1) based on the orographic correction and 2) so that the 0 km elevation intersects 28 ◦ C, which is the calculated MAAT for the Subandes (Rio Iruya section) at close to 0 km. Solid squares show MAAT estimates determined from paleosol carbonate sites in the Altiplano and Eastern Cordillera. The positions of these temperatures on the orographicallycorrected lapse rate curve allow calculation of paleoelevations using the 2nd degree polynomial regression.

(1)

where y is paleoelevation and x is paleotemperature (Fig. 3). We then use this orographically-corrected relationship to estimate paleoelevations from calculated mean annual air temperature (MAAT)

Middle to late Miocene paleotemperatures (Tables 1 and 2) are derived from paleosol carbonates from three sedimentary sections in the southern part of the Altiplano and Eastern Cordillera, as well as the low-elevation section in the Subandes (Figs. 1 and 2). All sites have been previously dated through paleomagnetic correlation and through 40 Ar/39 Ar and K–Ar dating of tuffs (Hernández et al., 1999; MacFadden et al., 1990, 1995). High elevation sites in the Altiplano and Eastern Cordillera include middle Miocene Cerdas (3900 m) and Quebrada Honda (3500 m) sections (16.3–15.9 Ma and 13.4–12.9 Ma, respectively) and the late Miocene Quehua (3900 m) section (8.7–7.9 Ma). Although each of these locations are separated by ∼100 kilometers, they are all located within a similar physiographic and climate setting in the southern Altiplano/Eastern Cordillera. Importantly, the tops of these three sections are tied to a regional low relief surface (Gubbels et al., 1993; Kennan et al., 1997) that presumably formed when the region was much lower. Finally, samples from the middle to late Miocene Subandes section along the Rio Iruya (Hernández et al., 1999) provide the baseline temperature to which the orographicallycorrected temperature lapse rate is anchored (Eq. (1)). Paleosurface temperature estimates derived from clumped isotope analysis of paleosols from the low elevation Subandes show warm soil temperatures of 42 ± 4 ◦ C (2σ ) between 12.5 and 6 Ma over the time period of deposition of southern Altiplano paleosols, and lower temperatures of 30 ± 9 ◦ C at 5.9 Ma and 28 ± 9 ◦ C at 5.3 Ma (Table 1). Southern Altiplano/Eastern Cordillera paleosol carbonates from the Cerdas, Quebrada Honda, and Quehua sections show a ∼14 ◦ C decrease in soil temperatures from 35 ± 5 ◦ C to 21 ± 4 ◦ C between 16 and 9 Ma (Table 1). The majority of this temperature decrease, ∼9 ◦ C, occurred between 16 Ma and 13 Ma, and an additional ∼5 ◦ C temperature decrease occurred between 13 and 9 Ma. Soil carbonate from a Pleistocene–Holocene soil at Quehua (3900 m) yielded a temperature of 19 ± 8 ◦ C, similar to late Miocene paleotemperatures at the same locality. This decrease in soil temperature occurred earlier than similarly dramatic cooling

C.N. Garzione et al. / Earth and Planetary Science Letters 393 (2014) 173–181

177

Table 1 AgesA and stable isotope results from individual samples. 2σ errors are shown in parentheses unless otherwise specifiedB and are discussed in the text. Section/location/ sample # (sample depth)

AgeA (Ma)

δ 13 CPDB of

δ 18 OPDB of

47 of

carbonate

carbonate

extracted CO2

Temperature (◦ C)

MAAT (◦ C)C

Paleoelevation (km)D

Cerdas 20◦ 53 S, 66◦ 19 W 07CR01 (60 cm) 07CR02 (2 m) 07CR04 (unkn) 07CR09 (1.5 m)

16.29 16.28 16.25 16.15

−8.6 −9.5 −8.6 −7.6 −8.5

−11.5 −11.6 −10.3 −10.1 −13.8

0.604 (19) 0.602 (38) 0.603 (38) 0.610 (38) 0.603 (38)

35 (5) 35 (9) 35 (9) 33 (9) 35 (9)

19 (9) 21 21 18 20

1.1 (0.7) 1.0 1.0 1.3 1.0

Quebrada Honda 21◦ 58 S, 65◦ 08 W 07QH05 (20 cm) 07QH08 (60 cm) 07QH09 (60 cm) 07QH10 (1.2 m) 07QH11 (60 cm) 07QH13 (30 cm) 07QH15 (60 cm)

13.33 13.09 13.05 12.99 12.97 12.94 12.92

−9.0 −9.3 −8.8 −8.7 −8.7 −9.1 −9.1 −9.1

−8.6 −9.8 −8.7 −7.9 −8.3 −8.2 −9.4 −8.0

0.640 (14) 0.651 (38) 0.623 (38) 0.634 (38) 0.634 (38) 0.640 (38) 0.669 (38) 0.626 (38)

26 (3) 24 (8) 30 (9) 28 (9) 28 (9) 26 (9) 20 (8) 30 (9)

9 (5) 7 15 12 11 10 2 14

2.6 (0.6) 3.1 1.9 2.3 2.4 2.6 4.0 2.0

Quehua 19◦ 56 S, 66◦ 53 W 08QU17 (20 cm) 08QU18 (20 cm) 08QU26 (unkn) 08QU28 (20 cm) 08QU30 (20 cm)

8.98 8.90 7.60 7.22 7.14

−5.3 −5.8 −5.0 −5.7 −5.5 −4.5

−12.6 −12.8 −12.5 −12.4 −13.0 −12.5

0.665 (17) 0.652 (38) 0.671 (38) 0.673 (38) 0.655 (38) 0.673 (38)

21 (4) 24 (8) 20 (8) 19 (8) 23 (8) 19 (8)

2 (4) 7 2 1 6 1

3.7 (0.5) 3.1 4.1 4.1 3.3 4.1

Modern soil Quehua 08QU08 (20 cm)

0

−6.4 −6.4

−11.3 −11.3

0.676 (38) 0.676 (38)

19 (8) 19 (8)

0 (7) 1

4.3 (0.8) 4.3

12.5 8.9 8.9 7 .4 6 .8 6 .5 6

−11.3 −9.5 −11.6 −11.6 −13.3 −9.6 −12.1 −11.6

−8.7 −10.1 −9.5 −9.4 −6.7 −8.5 −8.2 −8.7

0.577 (14) 0.567 (38) 0.581 (38) 0.576 (38) 0.578 (38) 0.587 (38) 0.579 (38) 0.571 (38)

42 (4) 44 (10) 41 (10) 42 (10) 41 (10) 39 (10) 41 (10) 44 (10)

28 (5) 31 27 29 28 25 28 30

−9.5 −10.2 −8.9

−8.3 −8.1 −8.4

0.627 (27) 0.632 (38) 0.622 (38)

29 (6) 28 (9) 30 (9)

13 (6) 12 15

Subandes—Rio Iruya 22◦ 54 S, 64◦ 31 W RI-08 8 P.S. 4 (07) P.S. 4 (07) RI 3 P.S. 26 (07) RI 15 RI23 Late Miocene Subandes RI 27B RI29-2

5.85 5 .3

A Ages are determined by extrapolations from measured stratigraphic columns within previously dated stratigraphic sections (MacFadden et al. 1990, 1995; Hernández et al., 1999). B 1σ analytical uncertainties for δ 13 C and δ 18 O of carbonate average ±0.02 and ±0.01 respectively and are not shown in the table. C Mean annual air temperature is calculated using Eq. (3)1 of Quade et al. (2013). D Paleoelevations were determined from MAAT using Eq. (1).

Table 2 47 values reported in per mil(h), growth temperatures, and paleoelevations for the southern Altiplano/Eastern Cordillera, and north-central Altiplano. Errors for each section (in parenthesis) reflects 95% confidence intervals (2σ ) and are discussed in the text. Region (modern elev.)

Age (Ma)

Number of samples

47 of extracted CO2

Soil temperature (◦ C)

Subandes (0.4 km)

12.5 to 6.0

7

0.577 (0.014)

42 (4)

28 (5)

Southern Altiplano Cerdas (3.9 km) Quebrada Honda (3.5 km) Quehua (3.9 km) modern soil (3.9 km)

16.3 to 16.2 13.3 to 12.9 9.0 to 7.1 Holocene

4 7 5 1

0.604 0.640 0.665 0.676

35 26 21 19

19 9 2 0

North-central AltiplanoD Callapa (3.8 km) Callapa (3.8 km) Callapa (3.8 km)

11.4 to 10.3 8.6 to 6.8 6.3 to 5.8

8 2 4

0.631 (0.024) 0.680 (0.046) 0.706 (0.048)

A B C D

(0.019) (0.014) (0.017) (0.038)

(5) (3) (4) (8)

28 (2) 18 (4) 13 (6)

Calculated MAAT (◦ C)A

(9) (5) (4) (8)

12 (4)

−1 (4) −8 (5)

Paleoelevation estimate for Altiplano (km)B

1.1 (0.7) 3.0C (0.6) 3.7 (0.5) 4.3 (0.8) 2.4 (0.5) 4.5 (0.6) 5.7 (0.8)

Mean annual air temperature is calculated using Eq. (3)1 of Quade et al. (2013). Paleoelevations were determined from MAAT using Eq. (1). Adjusted by +0.4 km based on the difference between the modern and paleo-elevation of Quebrada Honda. Paleosol carbonate 47 temperature values and 2σ standard error of the mean are from Ghosh et al. (2006b).

observed in the northern Altiplano (∼16 ◦ C decrease between ∼10 and ∼6 Ma) (Ghosh et al., 2006b). The two 5.9 Ma samples from the Subandes section described above show relatively cool temperatures, which are unexpected in a region that has remained at low elevation throughout most of the Neogene. These anomalous results may reflect: (1) global cooling in the late Miocene (Zachos et al., 2001), (2) local paleoelevation changes as the Subandes section was incorporated into the deformation front of the Andes (Echavarria et al., 2003), (3) deeper

carbonate formation depths than older samples and/or (4) increasingly wet climate conditions in the Subandes (Mulch et al., 2010; Uba et al., 2007) that would tend to reduce ground warming (Passey et al., 2010) and might change the season of carbonate precipitation (Peters et al., 2013). We can rule out local paleoelevation changes since their depositional history indicates that they were deposited at near 0 km elevations throughout the time interval sampled in the section. Deeper carbonate formation and/or an increasingly wet climate could account for most of the observed

178

C.N. Garzione et al. / Earth and Planetary Science Letters 393 (2014) 173–181

13 ◦ C temperature shift. Instrumental records and clumped isotope temperatures in modern soils in Tibet show temperature differences of 7–11 ◦ C for the mean of the warmest monthly temperature associated with solar ground warming (Quade et al., 2013). The large magnitude of these temperature differences with depth in modern soils indicates that either changing the soil carbonate formation depth from shallow (near surface) to deeper (1 m) or increasing the density of vegetation from sparsely vegetated with exposed bare ground to densely vegetated could account for most of this temperature shift. Given that these samples are younger than any available high-elevation data from the southern Altiplano, they were not used to set the low-elevation baseline for our temperature reconstructions. 4. Discussion 4.1. Mechanisms of surface cooling A temperature change of the magnitude of 14 ◦ C in the southern Altiplano allows us to consider the various factors that might have brought about such cooling. The South American plate did not undergo significant latitudinal change during the middle to late Miocene (Pardo-Casas and Molnar, 1987). It is similarly unlikely that global or regional cooling can account for the significant cooling recorded in the southern Altiplano between 16 and 8 Ma, given that global cooling was no more than a few degrees (Pagani et al., 1999; Zachos et al., 2001; Poulsen and Jeffery, 2011). Poulsen and Jeffery (2011) show a that a doubling of pCO2 over pre-industrial levels, a level consistent with pCO2 estimates since early Miocene time, would result in an average global temperature increase of 2.8 ◦ C. Due to polar amplification of greenhouse warming, the tropics would show an even smaller amount of warming under 2x pCO2 conditions. The inference that this region did not experience significant global cooling is also supported up by the observation that the low elevation Rio Iruya temperature record shows relatively constant temperatures over the time interval captured in the high altitude records (Table 1; Fig. 4). Perhaps the strongest argument for a tectonic influence on surface temperature, rather than a regional or global cooling event is the observation that cooling of the southern Altiplano preceded cooling of the north-central Altiplano by 7 ± 4 Myr. That the observed cooling in the Altiplano is large and diachronous between the north-central (∼16 ◦ C between 10 and 6 Ma) and south (∼14 ◦ C between 16 and 9 Ma) implicates a large-scale regional phenomenon, such as changing surface elevation, as the cause of temperature changes at both locations. In addition to global and regional climate, the temperature of soil carbonate formation reflects two additional factors: (1) paleosol carbonates are primarily precipitated during the dry season (Breecker et al., 2009), and (2) they record ground, rather than air, temperature (Passey et al., 2010). The arid southern Altiplano (between ∼19 ◦ and 22 ◦ S) displays seasonal rainfall in the Austral summer (Dec.–Jan.–Feb.), when mid- and upper level easterly atmospheric flow supplies moisture that feeds convective storms over the plateau that result in mean precipitation = 5–10 cm/month (Minvielle and Garreaud, 2011; Garreaud, 1999). During the rest of the year, the Altiplano is dominated by mid-level westerly flow that brings very dry air from the Pacific, resulting in virtually no precipitation (Vuille and Ammann, 1997). Given the lack of fall–winter–spring precipitation and the very low rainfall amount associated with summer precipitation, we infer that soil carbonate forms in the summer season, associated with drying of soils between rainfall events. Differences in solar illumination, or ground warming, can increase deep soil temperatures by several degrees under conditions of less plant cover (Passey et al., 2010; Quade et al., 2013), but this

Fig. 4. Evolution of paleo-ground temperatures at high elevation (Altiplano–Eastern Cordillera) and neighboring low elevation (Subandes) locations through the middle to late Miocene (see Table 2 for data). The difference between high and low elevation paleotemperatures is used to estimate the elevations show in Fig. 5. See text for discussion about circled outliers.

mechanism cannot account for cooling in the Altiplano record because proxy records show increasing aridity and elevation during the Miocene (Garzione et al., 2006) that would both tend to reduce plant cover. 4.2. Paleoelevation estimates To account for higher middle to late Miocene temperatures in western South America associated with both climate change and the lower elevation of the central Andes, we estimated paleoelevation based on the temperature difference between low-elevation (Subandes) and potentially higher elevation (Altiplano) records (see Fig. 4). The following summarizes the steps that we took to calculated paleoelevations and the error associated with these estimates. Ground temperatures can be significantly elevated compared to air temperatures, especially in settings that have minimal vegetal cover (Passey et al., 2010; Quade et al., 2013). The study area in the southern Altiplano is associated with sparse ground cover, consisting of shrubs, grasses, and bare ground, characteristic of settings that should show significant ground warming. The 47 temperature estimate from the Pleistocene–Holocene soil near Quehua is 19 ◦ C, similar to observations at the closest Altiplano weather station at Vinto, Bolivia (elev. = 3710 m, 17 ◦ 57 S, 67 ◦ 06 W) (Weatherbase.com, 2012) that shows an average summer-time (Dec.–Jan.–Feb.) high of 19 ◦ C and summer-time mean of 10 ◦ C. The observation that recent soils reflect summertime high temperatures is consistent with strong solar ground warming in the southern Altiplano. This is observed in other arid mountain plateaus that are associated with minimal plant cover; for example, instrumental and clumped isotope thermometers show temperature differences of 7–11 ◦ C for the mean of the warmest monthly temperature in modern soils in Tibet, associated with solar ground warming (Quade et al., 2013). Calculation of the MAAT from the Pleistocene–Holocene soil at Quehua (see Step 1 below) gives MAAT = 0 ± 7 ◦ C, which is within error of the observed MAAT of 7 ◦ C at nearby Vinto, Bolivia. At Quehua, the similarity in temperature between the Pleistocene–Holocene soil (19 ± 8 ◦ C) and late Miocene paleosols (21 ± 4 ◦ C) (Table 2) suggests that this site had reached its approximate modern elevation (within ∼0.5 km) by ∼9 Ma.

C.N. Garzione et al. / Earth and Planetary Science Letters 393 (2014) 173–181

Step 1. To account for the effects of solar ground warming, we apply an empirical relationship that relates soil surface temperature to the MAAT (Eq. (3)3 from Quade et al., 2013). This equation is most relevant to the types of paleosols that we sampled because most samples were collected from shallow depths (20 to 60 cm) and show evidence for minimal to no erosion at their tops. Because we sampled at depth in paleosols, where soil temperature for the warm season may be cooler on average, this equation results in a minimum MAAT estimate that would tend to over-predict paleoelevations, thereby providing a systematic bias toward a smaller estimate of the magnitude of surface uplift between present day and the past. We note that this approach provides the most conservative estimate of surface uplift. 2σ errors in MAAT estimates were propagated using a bootstrap Monte Carlo approach and include analytical errors (described in the Methods section) and error in the regression3 from Quade et al. (2013) (Tables 1 and 2). Step 2. We determined the temperature–elevation lapse rates using the modern air temperature–elevation lapse rate (−5 ◦ C/km) on the eastern flank of the central Andes (Gonfiantini et al., 2001) in combination with regional GCMs of mean surface temperatures of the Altiplano plateau at various stages in its development (75%, 50%, 25%, 0% modern elevation) that show average warming of the central Andes by +0.6 ◦ C, +3.4 ◦ C, and +6.6 ◦ C for the 50%, 25%, 0% cases respectively (Ehlers and Poulsen, 2009) (Fig. 3). We sum the measured MAAT for individual weather stations (Gonfiantini et al., 2001) with these model-based corrections by linear extrapolation between elevations of 50%, 25%, and 0% of the modern Andes to determine a second degree polynomial curve that relates temperature to elevation (Fig. 3) (see Section 2.1 Temperature–elevation relationship for more details). No corrections are applied for stations above 2 km because the temperature change of the Central Andes region is minimal above the 50% Andes case (Ehlers and Poulsen, 2009). We call this second degree polynomial regression the orographically-corrected lapse rate because it takes into account the climate effects of lowering the central Andes. Step 3. The orographically-corrected lapse rate from Step 2 was adjusted along the x-axis to intersect the baseline temperature at 0 km (Fig. 3). This temperature (28◦ C) reflects the MAAT determined using Quade et al. (2013)3 (see Step 1) for middle to late Miocene paleosol carbonates from the Subandes at the Rio Iruya location that has remained near sea level throughout its depositional history (Tables 1 and 2). The orographically-corrected lapse rate equation is used to calculate paleoelevations from the mean of the MAAT calculated for each high altitude site (Tables 1 and 2). 2σ errors in these paleoelevation estimates are determined using a bootstrap Monte Carlo approach that includes the errors associated with variability in the T-elevation regression, as well as uncertainties in the MAAT estimates derived in Step 1. Additional 2σ errors in MAAT are added to sites that yielded paleoelevations of <2 km to take into account the uncertainty associated with the different climate conditions for a lower Andes (Ehlers and Poulsen, 2009). These additional errors are generously set equal to magnitude of the temperature difference for the lower Andes scenarios (i.e., additional 2σ errors of ±0.6 ◦ C for 2 km and ±3.4 ◦ C for 1 km). The 3 steps described above allow us to use the paleotemperature offset between high- and low-elevation samples and the orographically-corrected temperature lapse rate to calculate paleoelevations from Eq. (1) of 1.1 ± 0.7 km for Cerdas (∼16 Ma), 2.6 ± 0.6 km for Quebrada Honda (∼13 Ma), and 3.7 ± 0.5 km for Quehua (∼ 9–7 Ma) (Figs. 3 and 4). This calculation assumes 3 We used data provided by Quade (personal communication) represented in Fig. 4 of Quade et al. (2013). A bivariate linear fit to these data yielded the equation y = (1.22 ± 0.07)x − (23.18 ± 2.36), slightly different from Eq. (3) in Quade et al. (2013).

179

Fig. 5. Miocene–present elevation history of the Altiplano. Previously published estimates from the north-central Altiplano include oxygen isotopes of precipitation (Garzione et al., 2006), leaf physiognomy (Gregory-Wodzicki, 2002; Gregory-Wodzicki et al., 1998), and 47 paleothermometry (Ghosh et al., 2006b). The Ghosh et al. (2006b) paleoelevations were recalculated using the MAAT regression for soils (Eq. (3)1 in Quade et al., 2013) and the orographically-corrected temperature lapse rate (Eq. (1); Table 2). Note the general agreement between data from multiple proxies. Our data (black boxes) indicate that the southern Altiplano was at a similar elevation to the northern Altiplano at ∼16 Ma, but began to undergo rapid surface uplift between 16 and 13 Ma. The average rate of surface uplift is similar to the northern Altiplano records, but occurred 7 ± 4 Myr earlier.

that all observed temperature change is due to surface uplift and its attendant effects (e.g., changes in atmospheric circulation). We emphasize that this method is most accurate for paleosols that formed at higher elevations (e.g., Quehua section and Pleistocene– Holocene Quehua soil) because these soils formed under arid conditions and at shallow depths (Table 1), consistent with the conditions required for the conversion of soil surface temperature to MAAT carried out in Step 1. The older paleosols in the Cerdas and Quebrada Honda sections were sampled at deeper depths within the paleosol profiles (Table 1), where soil temperature for the warm season should be cooler on average (Quade et al., 2013). The equation used in Step 1 would produce a minimum MAAT estimate that would tend to predict higher paleoelevations. This systematic bias in older, lower elevation paleosols results in a smaller estimate of the magnitude of surface uplift between present day and the past. The three high elevation sections form the highest (youngest) stratigraphy beneath a regional low-relief paleo-surface that spans the Altiplano and Eastern Cordillera (Hoke and Garzione, 2008). Age constraints from these deposits bracket the age of this surface between middle and late Miocene. Because Quebrada Honda in the Eastern Cordillera resides at a lower modern elevation (3.5 km) than the sections within the Altiplano proper (3.9 km) within the same regional paleo-surface, the simplest scenario is that the Quebrada Honda site has always resided 0.4 km lower than the Altiplano. We therefore add 0.4 km to the Quebrada Honda elevation estimate to determine a paleoelevation of 3.0 ± 0.6 km for the Altiplano proper at ∼13 Ma. These estimates indicate surface uplift of 1.9 ± 0.7 km at an average surface uplift rate of 0.6 km/Myr between 16 and 13 Ma, and an additional 0.7 ± 0.6 km of surface uplift between 13 and 9 Ma (Fig. 5 and Table 2). Even if Quebrada Honda is excluded from this analysis because it is further removed from the southern Altiplano sites, the paleoelevation difference between the Cerdas and Quehua sites indicates surface

180

C.N. Garzione et al. / Earth and Planetary Science Letters 393 (2014) 173–181

uplift of 2.6 ± 0.7 km between 16 and 9 Ma at an average rate of 0.4 km/Ma. We recalculated paleoelevations from clumped isotope temperatures from paleosols from the north-central Altiplano (Ghosh et al., 2006b) using the same procedures described in Steps 1 though 3 above. Paleoelevations of 2.4 ± 0.5 km between 11.4 and 10.3 Ma, and 5.7 ± 0.8 km between 6.3 and 5.8 Ma indicate surface uplift of 3.3 ± 0.8 km between ∼10 and 6 Ma at an average rate of 0.8 km/Myr (Fig. 5). Although absolute elevations appear to be overestimated for the <8.6 Ma time period, we note that these new paleoelevation estimates result in a large magnitude of surface uplift, similar to previous studies (Ghosh et al., 2006b; Garzione et al., 2008; Gregory-Wodzicki, 2002). The most likely cause of the overestimate of paleoelevation is that paleosol carbonates in the north-central Altiplano formed under wetter climate conditions and at deeper depths compared to southern Altiplano paleosols. Based on the soil surface temperature versus MAAT regression3 from Quade et al. (2013), deeper soil carbonate formation would result in cooler MAAT estimates that would tend to overestimate paleoelevation. This problem could be addressed by analyzing low elevation paleosol carbonates from the Subandes in northern Bolivia to establish a baseline temperature record against which paleoelevations can be calculated for the north-central Altiplano. In addition, measuring clumped isotope temperature-depth profiles of modern high altitude soils in the north-central Altiplano would help to understand the relationship between soil carbonate temperature and MAAT (Quade et al., 2013).

also thickened (Garzione et al., 2006, 2007). If eclogitic lower crust and mantle lithosphere also experienced thickening of a similar magnitude, then surface uplift rates would be much lower (for example, 0 mm/yr if the rate of lower lithosphere thickening = rate of crustal thickening). In addition, the region of dominant crustal shortening during the middle to late Miocene time frame was focused within the Subandean zone (Echavarria et al., 2003), demonstrating that crustal thickening is spatially and temporally decoupled from the region of surface uplift within the Altiplano/Eastern Cordillera. Geophysical studies of the lithospheric structure of this region show that the eastern part of the southern Altiplano and the western part of the Eastern Cordillera have thickened crust, yet lack high density lower crust (eclogite) and mantle lithosphere (Beck and Zandt, 2002). There is also evidence for weak, ductile lower crust (Beck and Zandt, 2002; Chmielowski et al., 1999) beneath parts of the Altiplano and Eastern Cordillera. The sum of observations, including surface uplift rates, the diachroneity of surface uplift along strike of the Central Andes, and lithospheric structure based on geophysical observations support the removal of dense lower lithosphere and/or lower crustal flow as geodynamic mechanisms to account for rapid surface uplift. Based on the magnitude and timing of surface cooling along strike, we speculate that lithospheric removal occurred 7 ± 4 Myr earlier in the southern Altiplano. Surface uplift associated with lower lithospheric removal could have driven lower crustal flow toward the relatively low northern Altiplano, leading to additional crustal thickening, eclogite formation, and ultimately lower crustal/mantle removal.

5. Conclusions Acknowledgements The timing of surface uplift in the southern Altiplano preceded the north-central Altiplano by 7 ± 4 Myr. Dramatically thicker middle-late Miocene stratigraphic successions in the north-central Altiplano compared to the southern Altiplano (Allmendinger et al., 1997; Garzione et al., 2008) are compatible with the inferred middle Miocene surface uplift of the southern Altiplano, while the north-central Altiplano remained low. Between 14 and 9 Ma, stratigraphic thicknesses represent ∼5 km of basin fill in the northcentral Altiplano (Garzione et al., 2008), while the relatively high elevation southern Altiplano only experienced 200 to 300 m of deposition over this time period (MacFadden et al., 1990, 1995). These stratigraphic patterns underscore that the north-central Altiplano represented a deep basin in the middle Miocene relative to the surrounding southern Altiplano and Eastern and Western Cordilleras. Middle Miocene surface uplift of the southern Altiplano is also supported by observations from the Western Cordillera adjacent to the southern Altiplano, in both timing and magnitude, where middle Miocene rotation of the Western Slope of the Andes generated 2.0 ± 0.5 km of surface uplift, followed by 1.2 ± 0.6 km since the late Miocene (Jordan et al., 2010). Stable isotope evidence for the onset of hyperaridity during the middle to late Miocene in the Atacama Desert (Rech et al., 2010) is less specific about timing, but clearly indicates the development of a strong rain shadow on the Western Slope during that interval. The elevation records that we present here, combined with previous records, indicate rapid average rates of surface uplift between 0.4 km/Myr and 0.8 km/Myr in both the southern Altiplano and north-central Altiplano. Several lines of evidence support the removal of dense lower lithosphere and/or lower crustal flow as geodynamic mechanisms to account for rapid surface uplift. Surface uplift rates associated with crustal shortening are insufficient to account for the observed surface uplift rates. Crustal shortening rates for the middle Miocene to recent time frame (McQuarrie et al., 2005; Elger et al., 2005) distributed over the width of the central Andean plateau would predict surface uplift rates of 0.25 mm/yr, assuming that the lower dense lithosphere is not

This work was funded by NSF grant numbers 0635678 and 0908858 to Garzione. We would like to thank Bruce MacFadden and Roberto Hernández for helpful discussions and insight in the field, as well as Federico Anaya, Darin Croft, Juan Hernández and Greg Hoke for guidance in the field. Pennilyn Higgins is thanked for help with lab work. References Allmendinger, R.W., Jordan, T.E., Kay, S.M., Isacks, B.L., 1997. The evolution of the Altiplano–Puna plateau of the central Andes. Annu. Rev. Earth Planet. Sci. 25, 139–174. Barnes, J., Ehlers, T., 2009. End member models for Andean Plateau uplift. Earth-Sci. Rev. 97, 105–132. Barnes, J.B., Ehlers, T.A., McQuarrie, N., O’Sullivan, P.B., Tawackoli, S., 2008. Thermochronometer record of central Andean Plateau growth, Bolivia (19.5 ◦ S). Tectonics 27. http://dx.doi.org/10.1029/2007TC002174. Beck, S.L., Zandt, G., 2002. The nature of orogenic crust in the central Andes. J. Geophys. Res. 107. Bershaw, J., Garzione, C.N., Higgins, P., MacFadden, B.J., Anaya, F., Alvaregna, H., 2010. Spatial–temporal changes in Andean plateau climate and elevation from stable isotopes of mammal teeth. Earth Planet. Sci. Lett. 289, 530–538. Breecker, D.O., Sharp, Z.D., McFadden, L.D., 2009. Seasonal bias in the formation and stable isotopic composition of pedogenic carbonate in modern soils from central New Mexico, USA. GSA Bull. 121, 630–640. Chmielowski, J., Zandt, G., Haberland, C., 1999. The central Andean Altiplano–Puna magma body. Geophys. Res. Lett. 26, 783–786. DeCelles, P.G., Horton, B.K., 2003. Early to middle Tertiary foreland basin development and the history of Andean crustal shortening in Bolivia. GSA Bull. 115, 58–77. Dennis, K.J., Affek, H.P., Passey, B.H., Schrag, D.P., Eiler, J.M., 2011. Defining an absolute reference fram for ’clumped’ isotope studies of CO2 . Geochim. Cosmochim. Acta 75, 7117–7131. Echavarria, L.E., Hernandez, R.M., Allmendinger, R.W., Reynolds, J., 2003. Subandean thrust and fold belt of northwestern Argentina: Geometry and timing of the Andean evolution. Am. Assoc. Pet. Geol. Bull. 87, 965–985. Ehlers, T.A., Poulsen, C.J., 2009. Influence of Andean uplift on climate and paleoaltimetry estimates. Earth Planet. Sci. Lett. 281, 238–248. Eiler, J.M., 2007. “Clumped-isotope” geochemistry—The study of naturally-occurring, multiply-substituted isotopologues. Earth Planet. Sci. Lett. 262, 309–327.

C.N. Garzione et al. / Earth and Planetary Science Letters 393 (2014) 173–181

Eiler, J.M., 2011. Paleoclimate reconstructions using carbonate clumped isotope thermometry. Quat. Sci. Rev. 30, 3573–3588. Elger, K., Oncken, O., Glodny, J., 2005. Plateau-style accumulation of deformation Southern Altiplano. Tectonics 24. http://dx.doi.org/10.1029/2004TC001675. Garreaud, R.D., 1999. Multiscale analysis of the summertime precipitation over the central Andes. Mon. Weather Rev. 127, 901–921. Garreaud, R., Vuille, M., Clement, A.C., 2003. The climate of the Altiplano, observed current conditions and mechanisms of past changes. Palaeogeogr. Palaeoclimatol. Palaeoecol. 194, 5–22. Garzione, C.N., Hoke, G.D., Libarkin, J.C., Withers, S., MacFadden, B.J., Eiler, J.M., Ghosh, P., Mulch, A., 2008. Rise of the Andes. Science 320, 1304–1307. Garzione, C.N., Molnar, P., Libarkin, J.C., MacFadden, B.J., 2006. Rapid late Miocene rise of the Bolivian Altiplano: Evidence for removal of mantle lithosphere. Earth Planet. Sci. Lett. 241, 543–556. Garzione, C.N., Molnar, P., Libarkin, J.C., MacFadden, B.L., 2007. Reply to comment on “Late Miocene rise of the Bolivian Altiplano: Evidence for removal of mantle lithosphere”. Earth Planet. Sci. Lett. 259, 630–633. Ghosh, P., Adkins, J., Affek, H.P., Balta, B., Guo, W., Schauble, E., Schrag, D.P., Eiler, J.M., 2006a. 13 C–18 O bonds in carbonate minerals: A new kind of paleothermometer. Geochim. Cosmochim. Acta 70, 1439–1456. Ghosh, P., Garzione, C.N., Eiler, J.M., 2006b. Rapid uplift of the Altiplano revealed through 13 C–18 O bonds in paleosol carbonates. Science 311, 511–515. Gillis, R.J., Horton, B.K., Grove, M., 2006. Thermochronology, geochronology, and upper crustal structure of the Cordillera Real: Implications for Cenozoic exhumation of the central Andean plateau. Tectonics 25. http://dx.doi.org/10.1029/ 2005TC001887. Gonfiantini, R., Roche, M.-A., Olivry, J.-C., Fontes, J.-C., Zuppi, G.M., 2001. The altitude effect on the isotopic composition of tropical rains. Chem. Geol. 181, 147–167. Gregory-Wodzicki, K.M., 2002. A late Miocene subtropical-dry flora from the northern Altiplano, Bolivia. Palaeogeogr. Palaeoclimatol. Palaeoecol. 180, 331–348. Gregory-Wodzicki, K.M., McIntosh, W.C., Velasquez, K., 1998. Climatic and tectonic implications of the late Miocene Jakokkota flora, Bolivian Altiplano. J. South Am. Earth Sci. 11, 533–560. Gubbels, T.L., Isacks, B.L., Farrar, E., 1993. High-level surfaces, plateau uplift, and foreland development, Bolivian central Andes. Geology 21, 695–698. Hernández, R., Galli, C., Reynolds, J., 1999. Estratigrafía del Terciario en el noroeste Argentino. In: Gonzalez Bonorino, G., Omarini, R., Viramonte, J. (Eds.), Relatorio XIV Congreso Geológico Argentino. Salta, Argentina, pp. 316–328. Hoke, G.D., Garzione, C.N., 2008. Paleosurfaces, paleoelevation, and mechanisms for the late Miocene topographic development of the Altiplano plateau. Earth Planet. Sci. Lett. 271, 192–201. Horton, B.K., Hampton, B.A., Waanders, G.L., 2001. Paleogene synorogenic sedimentation in the Altiplano plateau and implications for initial mountain building in the central Andes. GSA Bull. 113, 1387–1400. Houseman, G., McKenzie, D.P., Molnar, P., 1981. Convective instability of a thickened boundary layer and its relevance for the thermal evolution of continental convergent belts. J. Geophys. Res., Solid Earth 86 (1978–2012), 6115–6132. Huntington, K., Eiler, J., Affek, H., Guo, W., Bonifacie, M., Yeung, L., Thiagarajan, N., Passey, B., Tripati, A., Daeron, M., 2009. Methods and limitations of ‘clumped’ CO2 isotope (47) analysis by gas-source isotope ratio mass spectrometry. J. Mass Spectrom. 44, 1318–1329. Huntington, K., Wernicke, B., Eiler, J., 2010. Influence of climate change and uplift on Colorado Plateau paleotemperatures from carbonate clumped isotope thermometry. Tectonics 29, TC3005. Husson, L., Sempere, T., 2003. Thickening the Altiplano crust by gravity-driven channel flow. Geophys. Res. Lett. 30. Insel, N., Poulsen, C.J., Ehlers, T.A., Sturm, C., 2012. Response of meteoric δ 18 O to surface uplift – Implications for Cenozoic Andean Plateau growth. Earth Planet. Sci. Lett. 317–318, 262–272. Insel, N., Poulsen, C.J., Sturm, C., Ehlers, T.A., 2013. Climate controls on Andean precipitation δ 18 O interannual variability. J. Geophys. Res., Atmos. 118, 9721–9742. http://dx.doi.org/10.1002/jgrd.50619. Isacks, B.L., 1988. Uplift of the central Andean plateau and bending of the Bolivian orocline. J. Geophys. Res. 93, 3211–3231. Jordan, T., Nester, P., Blanco, N., Hoke, G., Davila, F., Tomlinson, A., 2010. Uplift of the Altiplano–Puna plateau: A view from the west. Tectonics 29.

181

Kennan, L., Lamb, S., Hoke, L., 1997. High-altitude paleosurfaces in the Bolivian Andes: Evidence for late Cenozoic surface uplift. In: Widdowson, M. (Ed.), Paleosurfaces: Recognition, Reconstruction and Paleoenvironmental Interpretation. In: Geological Society (London), Special Publication, vol. 120, pp. 307–323. Leier, A., McQuarrie, N., Garzione, C., Eiler, J., 2013. Stable isotope evidence for multiple pulses of rapid surface uplift in the Central Andes, Bolivia. Earth Planet. Sci. Lett. 371, 49–58. MacFadden, B.J., Anaya, F., Perez, H., Naeser, C.W., Zeitler, P.K., Campbell Jr., K.E., 1990. Late Cenozoic paleomagnetism and chronology of Andean basins of Bolivia: Evidence for possible oroclinal bending. J. Geol. 98, 541–555. MacFadden, B.J., Anaya, F., Swisher III, C.C., 1995. Neogene paleomagnetism and oroclinal bending of the central Andes of Bolivia. J. Geophys. Res. 100, 8153–8169. Mack, G.H., James, W.C., Monger, H.C., 1993. Classification of plaeosols. Geol. Soc. Am. Bull. 105, 129–136. McQuarrie, N., 2002. The kinematic history of the central Andean fold-thrust belt, Bolivia: Implications for building a high plateau. GSA Bull. 114, 950–963. McQuarrie, N., Horton, B.K., Zandt, G., Beck, S., DeCelles, P.G., 2005. Lithospheric evolution in the Andea fold-thrust belt, Bolivia, and the origin of the central Andea plateau. Tectonophysics 399, 15–377. Minvielle, M., Garreaud, R.D., 2011. Projecting rainfall changes over the South American Altiplano. J. Climate 24, 4577–4583. Mulch, A., Uba, C.E., Strecker, M.R., Schoenberg, R., Chamberlain, C.P., 2010. Late Miocene climate variability and surface elevation in the central Andes. Earth Planet. Sci. Lett. 290, 173–182. Pagani, M., Arthur, M.A., Freeman, K.H., 1999. Miocene evolution of atmospheric carbon dioxide. Paleoceanography 14, 273–292. Pardo-Casas, F., Molnar, P., 1987. Relative motion of the Nazca (Farallon) and South American plates since late Cretaceous time. Tectonics 6, 233–248. Passey, B.H., Levin, N.E., Cerling, T.E., Brown, F.H., Eiler, J.M., 2010. High-temperature environments of human evolution in East Africa based on bond ordering in paleosol carbonates. Proc. Natl. Acad. Sci. 107, 11245–11249. Peters, N.A., Huntington, K.W., Hoke, G.D., 2013. Hot or not? Impact of seasonally variable soil carbonate formation on paleotemperature and O-isotope records from clumped isotope thermometry. Earth Planet. Sci. Lett. 361, 208–218. Pope, D.C., Willett, S.D., 1998. Thermal-mechanical model for crustal thickening in the central Andes driven by ablative subduction. Geology 26, 511–514. Poulsen, C.J., Jeffery, L., 2011. Climate change imprinting on stable isotopic compositions of high-elevation meteoric water cloaks past surface elevations of major orogens. Geology 39, 595–598. Poulsen, C.J., Ehlers, T.A., Insel, N., 2010. Onset of convective rainfall during gradual late Miocene rise of the central Andes. Science 328, 490–493. Quade, J., Garzione, C.N., Eiler, J., 2007. Paleoelevation reconstructions using paleosol carbonates. Rev. Mineral. Geochem. 66, 53–87. Quade, J., Breecker, D.O., Daëron, M., Eiler, J.M., 2011. The paleoaltimetry of Tibet: An isotopic perspective. Am. J. Sci. 311, 77–115. Quade, J., Eiler, J.M., Daëron, M., Achyuthan, H., 2013. The clumped isotope paleothermometer in soils and paleosol carbonate. Geochim. Cosmochim. Acta 105, 92–107. Rech, J.A., Currie, B.S., Shullenberger, E.D., Dunagan, S.P., Jordan, T.E., Blanco, N., Tomlinson, A.J., Rowe, H.D., Houston, J., 2010. Evidence for the development of the Andean rain shadow from a Neogene isotopic record in the Atacama Desert, Chile. Earth Planet. Sci. Lett. 292, 371–382. Rowley, D.B., Garzione, C.N., 2007. Stable isotope-based paleoaltimetry. Annu. Rev. Earth Planet. Sci. 35, 463–508. Sempere, T., Butler, R.F., Richards, D.R., Marshall, L.G., Sharp, W., Swisher, C.C. III, 1997. Stratigraphy and chronology of Upper Cretaceous–lower Paleogene strata in Bolivia and northwest Argentina. GSA Bull. 109, 709–727. Uba, C.E., Strecker, M.R., Schmitt, A.K., 2007. Increased sediment accumulation rates and climatic forcing in the central Andes during the late Miocene. Geology 35, 979–982. Vuille, M., Ammann, C., 1997. Regional snowfall patterns in the high, arid Andes. Clim. Change 36, 413–423. Weatherbase.com, 2012. Canty and Associates LLC. http://weatherbase.com/weather/ weather.php3?s=134258. Zachos, J., Pagani, M., Sloan, L., Thomas, E., Billups, K., 2001. Trends, rhythms, and aberrations in global climate 65 Ma to present. Science 292, 686–693.