Glacier expansion in southern Patagonia throughout the Antarctic cold reversal

Juan L. García1*, Michael R. Kaplan2, Brenda L. Hall3, Joerg M. Schaefer2, Rodrigo M. Vega4, Roseanne Schwartz2, and Robert Finkel5 1Instituto de Geografía, Facultad de Historia, Geografía y Ciencia Política, Pontifi cia Universidad Católica de Chile, Campus San Joaquín, Avenida Vicuña Mackenna 4860, comuna Macul, Santiago 782-0436, Chile 2Geochemistry, Lamont-Doherty Earth Observatory, Palisades, New York 10964, USA 3Earth Sciences Department and Change Institute, University of Maine, Orono, Maine 04469, USA 4Instituto de Ciencias de la Tierra y Evolución, Universidad Austral de Chile, Campus Isla Teja, Valdivia, Chile 5Earth and Planetary Science Department, University of California–Berkeley, Berkeley, California 94720, USA

ABSTRACT essential for understanding its cause, as well Resolving debated climate changes in the southern middle latitudes and potential telecon- as the cryosphere-atmosphere-ocean links that nections between southern temperate and polar latitudes during the last glacial-interglacial operated during the late glacial to transition is required to help understand the cause of the termination of ice ages. Outlet gla- transition (Ackert et al., 2008). ciers of the Patagonian Ice Fields are primarily sensitive to atmospheric temperature and also We use 10Be and 14C techniques to establish precipitation, thus former ice margins record the extent and timing of past climate changes. a detailed reconstruction of ice fl uctuations 38 10Be exposure ages from moraines show that outlet glaciers in Torres del Paine (51°S, south during the entire ACR in the Torres del Paine Patagonia, Chile) advanced during the time of the Antarctic cold reversal (ACR; ca. 14.6– National Park (51°S, 73°W; Fig. 1), southern 12.8 ka), reaching a maximum extent by ~14,200 ± 560 yr ago. The evidence here indicates Chile. Torres del Paine has one of the prime late that the South Patagonian Ice Field was responding to late glacial climate change distinctly glacial moraine records in the southern middle earlier than the onset of the European stadial (ca. 12.9 ka). Major glacier latitudes. The excellent preservation and conti- recession and in the Torres del Paine region occurred by 12.5 ka and thus early in nuity of moraines, as well as their geographical the Younger Dryas. We provide direct evidence for extensive ice in Patagonia at the very start location, make them ideal to test hypotheses of the ACR that agrees with atmospheric and marine records from the Southern Ocean and of late glacial climate change at the middle . Atmospheric conditions responsible for the early late glacial expansion at Torres latitudes, ~51°S (Figs. 1 and 2). In Torres del del Paine resulted from a climate reorganization that prompted a northern migration of the Paine, previous work (Marden and Clapper- south westerly wind belt to the latitude of Torres del Paine at the onset of the ACR chronozone. ton, 1995) defi ned four distinct moraines belts

INTRODUCTION Figure 1. Location of Tor- Understanding millennial-scale climate vari- ODP1233 200 km res del Paine (white box CLD LGM STF N ability that interrupted the last deglaciation in main image) in south- AC (18.0–11.5 ka) affords insight into the nature ern South America. Solid and cause of the termination of ice ages. One arrows depict inferred prominent event, the Antarctic cold reversal approximate location of Lago General Carrera/Buenos Aires south westerly wind belt HPN (ACR, ca. 14.6–12.8 ka; Lemieux-Dudon et at present (Miller, 1976) 46°40'S al., 2010) in the high southern polar latitudes, HPS and different key peri- PB moraines was contemporaneous with the Bølling-Allerød ods during last glacial- ACR TDP warm period in the north and ended at the onset interglacial transition as PRESENT inferred from sedimento- SM of the Younger Dryas stadial (ca. 12.9–11.7 ka; SAF logical and paleoecologi- TF Blunier and Brook, 2001), but its cause remains BC cal records of Lamy et al. HS1 - YD CDD obscure. Recent studies (Strelin et al., 2011; (2004), Heusser (2003), Putnam et al., 2010a; Kaplan et al., 2010) Moreno et al. (1999), and show evidence for a late glacial ice expansion Anderson et al. (2009). PF

White dashed lines indi- 58°20'S in southern middle latitudes near the end of the cate approximate present ACR ca. 13.0 ka, followed by substantial gla- positions of Polar Front cier recession in the subsequent millennium. (PF), Sub-Antarctic Front However, marine and ice-core evidence indi- (SAF), and Subtropical Southern Ocean Front (STF). TDP—Torres cates environmental changes associated with del Paine; CLD—Chil- the onset of the ACR much earlier than 13.0 ka ean Lake District; AC— (EPICA Community Members, 2004 [EPICA— Archipiélago de Chiloé; European Project for Ice Coring in Antarctica]; HPN—Hielo Patagónico Norte; HPS—Hielo Pa- Blunier and Brook, 2001; Barker et al., 2009], 70°00'S tagónico Sur; PB—Puerto and the nature of climate dynamics through- Bandera; SM—Strait of out the ACR around Patagonia remains unclear Magellan; TF—Tierra (Kaplan et al., 2008; Sugden et al., 2005). Thus, del Fuego; BC—Beagle resolving the timing and structure of climate Channel; CD—Cordillera 81°40'W 70°00'W 58°20'W Darwin ice cap; LGM— changes throughout this time period on land is ; ACR—Antarctic cold reversal; HS1—Heinrich stadial event; YD— Younger Dryas; ODP—Ocean Drilling Program. Base map modifi ed from University of Maine *E-mail: [email protected]. Environmental Change Model (http://ecm.um.maine.edu).

GEOLOGY, September 2012; v. 40; no. 9; p. 859–862; Data Repository item 2012241 | doi:10.1130/G33164.1 | Published online 23 July 2012 ©GEOLOGY 2012 Geological | September Society 2012of America. | www.gsapubs.org For permission to copy, contact Copyright Permissions, GSA, or [email protected]. 859 E E E faces (Schaefer et al., 2009) allowing precision Laguna Geomorphology σ 13.8 ± 0.3 LA0704 River scarp River bed in our ages that averages 3.9% (1 ; Fig. DR3 11.2 ± 0.5 LA0512 Azul TDP IV 14.6 ± 0.4 LA0703 Alluvial fan Talweg in the Data Repository). Our exposure ages are 15.0 ± 1.3 LA0522 LA0707 N 14.5 ± 0.3 Talus slope Mire 13.8 ± 0.5 LA0728 10 14.0 ± 0.3 LA0901 Lake terrace Bedrock calculated using a Be production rate based on 14.6 ± 0.6 LA0727 12.312.3 ± 00.1.1 * Lake shoreline a New Zealand site (Putnam et al., 2010b; see Glacial geomorphology 10 12.512.5 ± 00.07.07 N the Data Repository). This Be production rate TDP I main outwash plain Vega Baguales TDP I moraine recently has been confi rmed in the Lago Argen- TDP I moraine ridge tino area (Kaplan et al., 2011), <100 km north TDP I ACR main outwash plain 13.8 ± 0.4 RP0701 TDP III TDP II ACR outwash plain of Torres del Paine. We detected and rejected 14.4 ± 0.4 RP0705 9.3 ± 0.7 LA0714 ACR moraine 13.9 ± 0.5 RP0703 13.9 ± 0.5 LA0715 the outliers by applying the Grubbs (1969) test ACR moraine ridge 16.9 ± 0.6 LA0716 and a 2σ criteria (see the Data Repository). Post-ACR outwash plain 13.8 ± 0.6 LA0720 21.3 ± 0.8 LA0732 Meltwater channel Scoured bedrock RESULTS Physiography * For each respective moraine belt, the 10Be N 12.512.5 ± 00.04.04 100 m interval topographic countour Lake boulder ages show a normal distribution and 12.512.5 ± 00.1.1* River Road exhibit high internal consistency after exclud- 10Be sample 14C sample

14.6 ± 0.5 RP0815 N ing outliers (Fig. DR5; see the Data Reposi- 15.0 ± 0.5 RP0817 14.1 ± 0.5 VN0525 1 0.5 0 km 1 tory). Boulders from the TDP II moraines 13.9 ± 0.4 VN0526 13.4 ± 0.5 RP0820 yielded ages ranging from 13.4 to 15.0 ka, with 13.9 ± 0.4 RP0903 an arithmetic mean of 14.2 ± 0.5 ka (n = 14) 14.1 ± 0.4 RP0906 13.7 ± 0.6 RP0905 (Fig. DR5A). The TDP III moraine boulders 14.0 ± 0.6 RP0904 range from 13.7 to 15.0 ka, with a mean of 14.1 ± 0.5 ka (n = 10) (Fig. DR5B), and those of the TDP IV moraines (including 2 ages recalcu-

N Laguna lated from Moreno et al., 2009) yielded ages of Amarga Rio de las Chinas io 13.8–15.3 ka, with a mean of 14.1 ± 0.7 ka (n = Rio Paine 13.7 ± 0.4 SAR0705 6) (Fig. DR5C). The resulting 10Be mean ages 12.6 ± 0.7 SAR0719 N 14.3 ± 0.4 SAR0907 13.9 ± 0.5 SAR0718 indicate that deposition of all three moraine 13.9 ± 0.4 SAR0725 14.0 ± 0.6 SAR0908 systems occurred, within error, during the 15.3 ± 0.6 SAR0721 same time interval, and thus rapidly. The num- 18.7 ± 1.0 SAR0724 8.7 ± 0.3 SAR0701 13.8 ± 0.9 SAR0723 13.0 ± 0.4 SAR0702 ber, size, and continuity of the moraine ridges 16.4 ± 0.9 SAR0722 7.4 ± 0.5 SAR0906 suggest that the ice was active and capable of 14.3 ± 0.4 SAR0703 14.3 ± 0.4 SAR0713 eroding, transporting, and depositing a large Lago Sarmiento de Gamboa volume of sediment during their formation. In E EEthe Lago Sarmiento, the TDP II moraine cross- Figure 2. Glacial geomorphic map of Laguna Azul–Lago Sarmiento area in Torres del Paine cuts the older TDP I moraine, suggesting that at National Park (inset with red box delineating extent of main map). White boxes in main map least the most extensive ACR moraine in Torres show 10Be cosmogenic-exposure ages (black) and calibrated radiocarbon ages (pink; close del Paine represents a glacial expansion, rather minimum deglacial ages; asterisks represent ages from Moreno et al., 2009) in thousands than just a stillstand during retreat. In addition, σ 10 of years before today and before present, respectively (±1 ). Two Be ages (VN0525 and near Río Paine, deformed lake beds occur in VN0526) are recalculated from Moreno et al. (2009) using New Zealand 10Be production rate. 10Be ages in italics and gray color are statistically determined outliers. TDP—Torres del the TDP IV moraines, indicating that ice read- Paine; ACR—Antarctic cold reversal. vanced over proglacial lake deposits (Marden and Clapperton, 1995). Maximum glacier expansion occurred by (from outer to inner, A–D) thought to have short gaps between both lake basins, suggesting 14.2 ± 0.5 ka (TDP II). Local ice retreat (typi- been deposited during Last Glacial Maximum that Laguna Azul and Lago Sarmiento ice lobes cally 1–3 km) occurred after TDP II moraine (LGM) conditions. This study shows that, at merged and formed a single continuous ice mass deposition. Then, ice readvanced and depos- present, the timing, extent, and structure of the with a >20 km terminus during their formation ited the TDP III and TDP IV moraines. Despite LGM remain unknown in the Torres del Paine (Fig. 2; see the GSA Data Repository1). these local glacier fl uctuations, for the entire region. To avoid confusion with moraines of dif- ACR time ice was still extensive (95% of full ferent ages having similar labels in other sites METHODS late glacial extent) relative to the present icefi eld in southern South America (e.g., Sugden et al., We collected 38 boulders from the TDP II, divide and outlet glacier margins. A sediment 2005), we rename these moraines (i.e., A–D of TDP III, and TDP IV moraines adjacent to core obtained at the Vega Baguales meltwater Marden and Clapperton, 1995) here: A = Torres Laguna Azul and Lago Sarmiento. We mea- conduit (Fig. 2) yielded a close 14C minimum del Paine (TDP) I, B = TDP II, C = TDP III, sured 10Be concentrations of the boulder sur- calibrated (cal) age of 12,460 ± 70 cal yr B.P. and D = TDP IV. These moraine sets normally (see the Data Repository) for glacier retreat occur within 2–3 km of each other, including at 1GSA Data Repository item 2012241, supporting from the TDP IV moraine position in Torres del Laguna Azul and Lago Sarmiento (Fig. 2), and text, tables, and fi gures describing the physiographi- Paine. This in agreement with previous 14C data are ≥45 km from present-day ice margins. The cal features of the study area, geochronological data from the region (Moreno et al., 2009) indicating sharp morphology of the TDP II–IV moraines and methods, and southern glacier dynamic during that ice was receding at several sites in the area the Antarctic cold reversal, is available online at 10 contrasts with that of the TDP I moraines, www.geosociety.org/pubs/ft2012.htm, or on request by 12,500 cal yr B.P. Taken together, the Be 14 which are wide, prominent landforms. TDP II, from [email protected] or Documents Secre- and C ages indicate that ice remained at the IV III, and IV moraines can be traced with only tary, GSA, P.O. Box 9140, Boulder, CO 80301, USA. position until before 12.5 ka.

860 www.gsapubs.org | September 2012 | GEOLOGY DISCUSSION AND CONCLUSIONS Figure 3. Paleoclimate re- Late Early This study expands and refi nes earlier pio- cords discussed in text. A: 10Be moraine ages IS A neering work (Strelin et al., 2011; Sagredo et al., from Torres del Paine PB 2011; Moreno et al., 2009; Sugden et al., 2005; (TDP) II, TDP III, and TDP Fogwill and Kubik, 2005) that proposed the IV moraines (this study, BH and two 10Be ages from existence of expanded glaciers during the ACR TDP IV Moraines period in southern South America, but lacked Moreno et al., 2009). See Tables DR1 and DR2 TDP III the extensive directly 10Be-dated chronological (see footnote 1). Birch TDP II data presented here. That is, we show with direct Hill (BH) (Putnam et al., B 10Be dating that glaciers in Torres del Paine were 2010a) and Irishman Ba- far (≥45 km) from the present-day ice at the very sin (IS) (Kaplan et al., 10 0 20

2010) outer moraines are ) start of the ACR, ~1300 yr before the onset of 1

in New Zealand; radiocar- 6 10 14 C k.y.

the Younger Dryas (YD). Be and C chronolo- 40 30 bon-dated Puerto Ban- 2 Polar species (%) gies in Torres del Paine together afford the fi rst dera moraines (PB) are 45 gcm 3 terrestrial evidence for both the onset and the in Argentina (Strelin et ( D al., 2011). B: Ocean-water duration of the ACR in southern Patagonia. Col- 12 temperatures as inferred Opal flux lectively, with the records from glacial basins from polar foraminiferal O north (Strelin et al., 2011) and south (Sagredo et 18 species in South Atlantic E al., 2011) of Torres del Paine, we conclude that (core TNO57–21) (Barker an unknown amount of regional glacier retreat et al., 2009). C: South-

ern Ocean upwelling 2 had occurred after the LGM, before the onset of G. bulloides from biogenic opal fl ux 2.5 2 1.5 1 0.5 the ACR. This includes what happened immedi- (Anderson et al., 2009). ately prior to the late glacial period. D: Sea-surface tempera- F 3 We hypothesize that the prominent expres- tures in southeast Pacifi c Dome C CO sion of the ACR at Torres de Paine could have Ocean (Ocean Drilling 200 220 240 260 Program Site 1233) in- been due to the shift of the westerly belt close ferred from Globigerina O 18 to 51°S (Fig. 1), which would have brought not bulloides δ18O (Lamy et al., G -38 -36 -34 only cold conditions, but also peak precipita- 2004). E: Dome C (Ant- Byrd arctica) atmospheric CO tion to the glacial catchment. Late glacial pollen 2 concentrations (Monnin -40 records south of 53°S in the southernmost tip O et al., 2001) placed on 18 of South America (e.g., Heusser, 2003) support Greenland Ice Sheet Proj- -42 such a northward shift of the westerly belt and ect 2 (GISP2) time scale GISP2 also may imply a concurrent northward shift of (Marchitto et al., 2007). F: the Antarctic Polar Frontal Zone, probably as far Polar atmospheric mean -42 -40 -38 -36 YD ACR HS1 LGM annual temperatures de- as the latitude of the Strait of Magellan (Sugden rived from Byrd (Byrd et al., 2005). 10 12 14 16 18 20 Station, Antarctica) (ka) Our moraine chronology suggests that gla- core δ18O isotopic record ciers in the southern middle latitudes responded (Lemieux-Dudon et al., 2010). G: Polar atmospheric mean annual temperatures derived from North Atlantic GISP2 ice core δ18O isotopic record (Greenland) (Stuiver and Grootes, 2000). to the onset of the ACR, as recorded in Antarc- LGM—Last Glacial Maximum; HS1—Heinrich stadial 1; ACR—Antarctic cold reversal; YD— tic ice cores (EPICA Community Members, Younger Dryas. 2004; Blunier and Brook, 2001) (Fig. 3F). Previous work (Strelin et al., 2011; Putnam Figure 4. Late glacial Antarctic Cold Reversal et al., 2010a; Kaplan et al., 2010) showed readvances at southern PB (n = 11) glaciers in the southern middle latitudes cul- middle latitudes. Glacial IS (n = 9) minating at the end of the ACR chronozone. records suggest late gla- BH I__Outer (n = 2) Moreover, New Zealand glaciers could have cial expansions at begin- readvanced much earlier, at the start of the ning and end of Antarctic cold reversal (see the Data TDP (n = 30) ACR, but data are sparse (Putnam et al., 2010a; Repository [see footnote Figs. 3A and 4). The climate signal from south- 1] for more details). TDP— ern glaciers (Fig. 4; see the Data Repository) Torres del Paine II, III, and Relative Probability BH I (n = 27) is consistent with several recently obtained IV moraines (this study); PB—Puerto Bandera mo- ocean-atmosphere records shown in Figure 3, raines at Lago Argentino, suggesting a coupling between oceanic, atmo- Argentina (Strelin et al., 11.5 12.0 12.5 13.0 13.5 14.0 14.5 15.0 15.5 16.0 17.0 spheric, and cryospheric systems in the middle 2011); BH—Birch Hill mo- Age (ka) and high latitudes of the Southern Hemisphere. raines at Lake Pukaki, The onset of glacier readvance in south Pata- New Zealand (Putnam et al., 2010a); IS—Irishman Basin (New Zealand) outer moraines (Kaplan et al., 2010). Solid line represents Patagonian sites; dashed line represents New Zealand sites. gonia and perhaps in New Zealand (Putnam et al., 2010a) during the early phase of the

ACR (Fig. 3A) was contemporaneous with an (Fig. 3E) at 14.6 ka. Following ~1600 yr of and atmospheric CO2 rates after 13.0 ka coin- inferred decline in the Southern Ocean upwell- ACR conditions (Lemieux-Dudon et al., 2010; cided with rapid glacier retreat in Patagonia ing rate (Anderson et al., 2009) (Fig. 3C), and EPICA Community Members, 2004; Blunier and New Zealand (e.g., Strelin et al., 2011; likely reduced CO2 outgassing of the Southern and Brook, 2001), the resumed glacial to inter- Putnam et al., 2010a; Kaplan et al., 2010; Ocean to the atmosphere (Monnin et al., 2001) glacial rise in the Southern Ocean upwelling Moreno et al., 2009).

GEOLOGY | September 2012 | www.gsapubs.org 861 Our study adds another important element Blunier, T., and Brook, E.J., 2001, Timing of mil- concentrations over the Last Glacial termina- to the scenario that signifi cant latitudinal shifts lennial scale climate change in Antarctica and tion: Science, v. 291, p. 112–114, doi:10.1126 of the westerly wind belt through termination I Greenland during the last glacial period: Sci- /science.291.5501.112. ence, v. 291, p. 109–112, doi:10.1126/science Moreno, P.I., Lowell, T.V., Jacobson, G.L., Jr., and (Fig. 1) (Anderson et al., 2009; Denton et al., .291.5501.109. Denton, G.H., 1999, Abrupt vegetation and 2010) and resulting changes in Southern Ocean Denton, G.H., Anderson, R.F., Toggweiler, J.R., climate changes during the last glacial maxi- upwelling (Barker et al., 2009; Anderson et al., Edwards, R.L., Schaefer, J.M., and Putnam, mum and last termination in the Chilean Lake 2009; Bianchi and Gersonde, 2004; Toggweiler A.E., 2010, The Last Glacial termination: District: A case study from Canal de la Puntilla Science, v. 328, p. 1652–1656, doi:10.1126 (41°S): Geografi ska Annaler, v. 81, p. 285–311, et al., 2006) seem to have played a prime role /science.1184119. doi:10.1111/j.0435-3676.1999.00059.x. in the late glacial climate regime of the south- EPICA Community Members, 2004, Eight glacial Moreno, P.I., Kaplan, M.R., Francois, J.P., Villa-Mar- ern middle latitudes. Our fi ndings support late cycles from an Antarctic ice core: Nature, tinez, R., Moy, C.M., Stern, C.R., and Kubik, glacial climatic links in the southern middle and v. 429, p. 623–628, doi:10.1038/nature02599. P.W., 2009, Renewed glacial activity during the high latitudes (Fig. 3) due to migrations in the Fogwill, C.J., and Kubik, P.W., 2005, A glacial stage Antarctic cold reversal and persistence of cold spanning the Antarctic Cold Reversal in Torres conditions until 11.5 ka in southwestern Pata- Antarctic Polar Frontal Zone and westerly belt del Paine (51°S), Chile, based on preliminary gonia: Geology, v. 37, p. 375–378, doi:10.1130 linked to Southern Ocean temperature (Bian- cosmogenic exposure dating: Geografi ska An- /G25399A.1. chi and Gersonde, 2004). In addition, our data naler, v. 87, p. 403–408, doi:10.1111/j.0435-3676 Putnam, A.E., Denton, G.H., Schaefer, J.M., Bar- support the atmospheric bipolar seesaw model .2005.00266.x. rell, D.J.A., Andersen, B.G., Finkel, R.C., of climate change at the end of the last gla- Grubbs, F.E., 1969, Procedures for detecting outlying Schwartz, R., Doughty, A.M., Kaplan, M.R., observations in samples: Technometrics, v. 11, and Schlüchter, C., 2010a, Glacier advance ciation (Denton et al., 2010), thus suggesting p. 1–21, doi:10.2307/1266761. in southern middle-latitudes during the Ant- that the latitudinal migration of the westerly Heusser, C.J., 2003, Ice age southern Andes: A arctic Cold Reversal: Nature Geoscience, v. 3, belt throughout the termination could be tied chronicle of paleoecological events: Amster- p. 700–704, doi:10.1038/ngeo962. to major oceanographic changes in the North dam, Netherlands, Elsevier, 240 p. Putnam, A.E., Schaefer, J.M., Vandergoes, M., Bar- Kaplan, M.R., Moreno, P.I., and Rojas, M., 2008, Gla- rell, D., Kaplan, M., Goehring, B., Schwartz, Atlantic basin. In this scenario, throughout cial dynamics in southernmost South America R., Finkel, R., Kelley, S., and Denton, G.H., the ACR, the westerly belt was shifted north, during Marine Isotope Stage 5e to the Younger 2010b, In situ cosmogenic 10Be production-rate bringing cold temperatures and increased pre- Dryas chron: A brief review with a focus on cos- calibration from the Southern Alps, New Zea- cipitation to Torres del Paine, in response to a mogenic nuclide measurements: Journal of Qua- land: Quaternary Geochronology, v. 5, p. 392– reinvigorated Atlantic meridional overturning ternary Science, v. 23, p. 649–658, doi:10.1002 409, doi:10.1016/j.quageo.2009.12.001. /jqs.1209. Sagredo, E.A., Moreno, P.I., Villa-Martínez, R., circulation during the Northern Hemisphere Kaplan, M.R., Schaefer, J.M., Denton, G.H., Barrell, Kaplan, M.R., Kubik, P.W., and Stern, C.R., Bølling-Allerød warm period. D.J.A., Chinn, T.J.H., Putnam, A.E., Andersen, 2011, Fluctuations of the Última Esperanza B.G., Finkel, R.C., Schwartz, R., and Doughty, ice lobe (52°S), Chilean Patagonia, during the ACKNOWLEDGMENTS A.M., 2010, Glacier retreat in New Zealand last glacial maximum and termination 1: Geo- The National Geographic Society, the Churchill during the Younger Dryas stadial: Nature, morphology, v. 125, p. 92–108, doi:10.1016/j Exploration Fund, the Graduate Student Government v. 467, p. 194–197, doi:10.1038/nature09313. .geomorph.2010.09.007. at the University of Maine, and the Comer Science and Kaplan, M.R., Strelin, J.A., Schaefer, J.M., Denton, Schaefer, J.M., Denton, G.H., Kaplan, M., Putnam, Education Foundation supported this research. We are G.H., Finkel, R.C., Schwartz, R., Putnam, A.E., A., Finkel, R.C., Barrell, D.J.A., Andersen, grateful to CONAF (Corporación Nacional Forestal) Vandergoes, M.J., Goehring, B.M., and Travis, B.G., Schwartz, R., Mackintosh, A., Chinn, T., Región de Magallanes and Torres del Paine National S.G., 2011, In-situ cosmogenic 10Be production and Schluchter, C., 2009, High-frequency Holo- Park (Chile), Víctor García, Marcelo Arévalo, Stefan rate at Lago Argentino, Patagonia: Implications cene glacier fl uctuations in New Zealand differ Krauss, Mario Pino, and Patricio Moreno for sup- for late-glacial climate chronology: Earth and from the Northern Signature: Science, v. 324, port and assistance during fi eld campaigns. We thank Planetary Science Letters, v. 309, p. 21–32, p. 622–625, doi:10.1126/science.1169312. Marcelo Solari, Esteban Sagredo, Aaron Putnam, and doi:10.1016/j.epsl.2011.06.018. Strelin, J.A., Denton, G.H., Vandergoes, M.J., Ninne- Christopher Marden for discussions on this research. Lamy, F., Kaiser, J., Ninnemann, U., Hebbeln, D., mann, U.S., and Putnam, A.E., 2011, Radio carbon FONDECYT (Fondo Nacional de Desarrollo Cientí- Arz, H.W., and Stoner, J., 2004, Antarctic tim- chronologies of the late glacial Puerto Bandera fi co y Tecnológico) grant 11110381 supported this ing of surface water changes off Chile and Pa- moraines, southern Patagonian icefi eld, Argen- publication. This is Lamont-Doherty Earth Observa- tagonian Ice Sheet response: Science, v. 304, tina: Quaternary Science Reviews, v. 30, p. 2551– tory contribution 7543. p. 1959–1962, doi:10.1126/science.1097863. 2569, doi:10.1016/j.quascirev.2011.05.004. Lemieux-Dudon, B., Layo, E., Petit, J.R., Wael- Stuiver, M., and Grootes, P.M., 2000, GISP2 oxy- REFERENCES CITED broeck, C., Svensson, A., Ritz, C., Barnola, gen isotope ratios: Quaternary Research, v. 53, Ackert, R.P., Jr., Becker, R.A., Singer, B.S., Kurz, J.M., Narcisi, B.M., and Parrenin, F., 2010, p. 277–284, doi:10.1006/qres.2000.2127. M.D., Caffee, M.W., and Mickelson, D.M., 2008, Consistent dating for Antarctic and Greenland Sugden, D.E., Bentley, M.J., Fogwill, C.J., Hulton, Patagonian glacier response during the Late ice cores: Quaternary Science Reviews, v. 29, N.R.J., McCulloch, R.D., and Purves, R.S., Glacial–Holocene transition: Science, v. 321, p. 8–20, doi:10.1016/j.quascirev.2009.11.010. 2005, Late-glacial glacier events in southern- p. 392–395, doi:10.1126/science.1157215. Marchitto, T.M., Lehman, S.J., Ortiz, J.D., Flückiger, most South America: A blend of ‘northern’ and Anderson, R.F., Ali, S., Bradtmiller, L.I., Nielsen, J., and van Geen, A., 2007, Marine radiocarbon ‘southern’ hemispheric climatic signals?: Geo- S.H.H., Fleisher, M.Q., Anderson, B.E., and evidence for the mechanism of deglacial at- grafi ska Annaler, v. 87, p. 273–288, doi:10.1111

Burckle, L.H., 2009, Wind-driven upwelling in mospheric CO2 rise: Science, v. 316, p. 1456– /j.0435-3676.2005.00259.x. the Southern Ocean and the deglacial rise in at- 1459, doi:10.1126/science.1138679. Toggweiler, J.R., Russell, J.L., and Carson, S.R., Marden, C.J., and Clapperton, C.M., 1995, Fluctua- mospheric CO2: Science, v. 323, p. 1443–1448, 2006, Midlatitude westerlies, atmospheric doi:10.1126/science.1167441. tions of the South Patagonian Icefi eld during the CO2, and climate change during the ice ages: Barker, S., Diz, P., Vautravers, M., Pike, J., Knorr, last glaciation and the Holocene: Journal of Qua- Paleoceanography, v. 21, PA2005, doi:10.1029 G., Hall, I.R., and Broecker, W.S., 2009, Inter- ternary Science, v. 10, p. 197–209, doi:10.1002 /2005PA001154. hemispheric Atlantic seesaw response during /jqs.3390100302. the last deglaciation: Nature, v. 457, p. 1097– Miller, A., 1976, The climate of Chile, in Schwerdt- 1102, doi:10.1038/nature07770. feger, W., ed., World survey of climatology 12: Bianchi, C., and Gersonde, R., 2004, Climate evo- of Central and South America: Am- Manuscript received 29 December 2011 lution at the last deglaciation: The role of the sterdam, Netherlands, Elsevier, p. 113–145. Revised manuscript received 8 April 2012 Southern Ocean: Earth and Planetary Science Monnin, E., Indermühle, A., Dällenbach, A., Flück- Manuscript accepted 10 April 2012 Letters, v. 228, p. 407–424, doi:10.1016/j.epsl iger, J., Stauffer, B., Stocker, T.F., Raynaud,

.2004.10.003. D., and Barnola, J.M., 2001, Atmospheric CO2 Printed in USA

862 www.gsapubs.org | September 2012 | GEOLOGY