University of New Hampshire University of New Hampshire Scholars' Repository

Master's Theses and Capstones Student Scholarship

Fall 2009

Developing a 'Little Ice Age' glacial chronology in the southern Peruvian Andes using lichenometry and cosmogenic beryllium-10 surface exposure dating

Jean R. Taggart University of New Hampshire, Durham

Follow this and additional works at: https://scholars.unh.edu/thesis

Recommended Citation Taggart, Jean R., "Developing a 'Little Ice Age' glacial chronology in the southern Peruvian Andes using lichenometry and cosmogenic beryllium-10 surface exposure dating" (2009). Master's Theses and Capstones. 498. https://scholars.unh.edu/thesis/498

This Thesis is brought to you for free and open access by the Student Scholarship at University of New Hampshire Scholars' Repository. It has been accepted for inclusion in Master's Theses and Capstones by an authorized administrator of University of New Hampshire Scholars' Repository. For more information, please contact [email protected]. DEVELOPING A 'LITTLE ICE AGE' GLACIAL CHRONOLOGY IN THE

SOUTHERN PERUVIAN ANDES USING LICHENOMETRY AND COSMOGENIC

10BE SURFACE EXPOSURE DATING

BY

JEAN R. TAGGART

B.S., Beloit College 2006

THESIS

Submitted to the University of New Hampshire

in Partial Fulfillment of

the Requirements for the Degree of

Master of Science

in

Earth Sciences: Geology

September, 2009 UMI Number: 1472084

INFORMATION TO USERS

The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleed-through, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

UMI

UMI Microform 1472084 Copyright 2009 by ProQuest LLC All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code.

ProQuest LLC 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106-1346 This thesis has been examined and approved.

0 Thesis Director, Joseph M. Licciardi Associate Professor of Earth Sciences

V

JulJaG.Bryce V Associate Professor of Geochemistry

leredith A. Kell'Ay ^(J Assistant Professor of Earth Sciences Dartmouth College

IpTUS" Date DEDICATION

For M.J.

iii ACKNOWLEDGEMENTS

I would like to thank, first and foremost, my advisor Joe Licciardi who has enabled me to complete in this project. Thanks to Joe for allowing me to be involved in his endeavors in Peru, for help in the field with my component of the project, for teaching me the ropes in the lab, and for endless time, encouragement, advice, input, funding, and editing my grant proposals and thesis. I have learned a great deal from working with Joe on this project. I would also like to thank my committee members, Julie Bryce and

Meredith Kelly, for their time invested in, patience with, and involvement in this project.

Thanks for their input and collaboration.

Thanks to everyone who provided assistance in the field, lab, and with various areas of this work: Anton Seimon, Barbara Mathe, Tom Lowell, Yves and Elena Chemin,

Pablo, Dylan Rood, Bob Finkel, Vincent Jomelli, Delphine Grancher, Olga Solomina, and Antoine Rabatel.

Thanks for the encouragement, advice, and support from Patrick, my parents, and my friends. Thanks also to my peers at UNH - including Florencia Prado, Mimi Szeto and Deb Goodwin - for their helpful language skills and technological expertise.

This project would not have been possible without generous funding in the form of a Teaching Assistantship, research grant, and travel assistance from the UNH

Department of Earth Sciences; a student research grant from the Geological Society of

America, a Grant-in-Aid of Research from Sigma Xi, a Graduate TA Achievement award from UNH CEPS, and travel assistance from the UNH Graduate School.

iv TABLE OF CONTENTS

DEDICATION iii

ACKNOWLEDGEMENTS iv

LIST OF FIGURES viii

LIST OF TABLES ix

ABSTRACT x

CHAPTER PAGE

I INTRODUCTION 1

1.1 Project Overview 1

1.2 Study Location 4

1.3 Geologic Setting 6

1.4 Climate Setting 8

1.5 'Little Ice Age' Glacial History of the Peruvian Andes 11

II METHODS 14

2.1 Geomorphic Mapping 14

2.2 Moraine Dating Methods 14

2.2.1 Lichenometric Dating 15

2.2.2 Cosmogenic 10Be Surface Exposure Dating 20

2.3 ELA Reconstruction 25

2.3.1 Modern Mass Balance and ELA 27

v 2.3.2 Paleo-Glacier Reconstructions and ELA 29

2.3.3 ELA Depression and Paleoclimate Implications 31

III RESULTS 32

3.1 Geomorphic Relationships 32

3.2 Moraine Chronology 36

3.2.1 Lichenometry Results 3 6

3.2.2 10Be Surface Exposure Ages 39

3.3 ELA and Paleoclimate Results 43

IV DISCUSSION 48

4.1 Geomorphic Relationships 48

4.2 LIA Moraine Chronology 49

4.2.1 Lichenometric Age Relationships 49

4.2.2 10Be Age Considerations 52

4.2.3 Comparison of Lichenometric and 10Be Ages 54

4.2.4 Possible Differences in Glacier Response Time 54

4.3 Comparison with Tropical Andean Glacier Fluctuations 56

4.4 Comparison to Global LIA Chronologies 58

4.5 ELA Determinations and Paleoclimate Inferences 59

4.6 Comparison with Other Proxies 61

4.7 Possible Climate Drivers 66

V CONCLUSIONS 71

REFERENCES 73

APPENDICES 83

VI APPENDIX A: LICHEN MEASUREMENTS 84

APPENDIX B: SAMPLE SITES AND DESCRIPTIONS 92

APPENDIX C: GOOGLE EARTH ELEVATION CALIBRATION 98

APPENDIX D: GEOMORPHIC PHOTOGRAPHS 101

APPENDIX E: 10BE SAMPLE PREPARATION AND PRODUCTION RATE

CALIBRATION 109

VII LIST OF FIGURES

FIGURE PAGE

1. Peru location map 5

2. Geologic map of the Cuzco region 7

3. Meteorological Data from Cuzco, Peru 9

4. Published Cordillera Blanca Rhizocarpon sp. growth curves 18

5. Landsat TM images of the Cordillera Vilcabamba 28

6. Geomorphic map of field area 33

7. Histogram of lichen diameters 36

8. Reconstructed Rhizocarpon sp. growth curve 37

9. Lichenometric calendar and relative ages 38

10. 10r> 42 Be ages 11. ELAs and AELAs 45

12. Modern and paleo-glacier reconstructions 46

13. Percent of glacier surface area loss 47

14. Comparison of global LIA glacial records 57

15. Comparison with tropical ice core records 62

16. Comparison with climate change proxies 65

Al. Tucarhuay geomorphic photographs 101

A2. Rio Blanco geomorphic photographs 104

A3. Sisaypampa geomorphic photographs 106 LIST OF TABLES

TABLE PAGE

1. Dates of the LIA culminations in Peru and northern Bolivia 12

2. Calibration points used in Cordillera Blanca lichen growth curves 21

3. Lichen data from the Cordillera Vilcabamba 38

4. 10Be results 40

5. Comparison of 10Be ages (alternative scaling schemes) 41

6. ELA estimates 44

7. Percent of glacier surface area loss 47

8. AAR and THAR values for tropical Andean glaciers 60

9. Modern ELAs and LIA ELA depressions in the tropical Andes 61

Al. Lichen measurements 84

A2. Calibration of Google Earth elevations 98

IX ABSTRACT

DEVELOPING A 'LITTLE ICE AGE' GLACIAL CHRONOLOGY IN THE

SOUTHERN PERUVIAN ANDES USING LICHENOMETRY AND COSMOGENIC

,0BE SURFACE EXPOSURE DATING

by

Jean R. Taggart

University of New Hampshire, September, 2009

The timing and causes of tropical climate changes during the Holocene are important and unresolved issues in paleoclimatology. Glacier chronologies are crucial for discerning the role of the tropics in global climate change, but past glacier activity in this region remains poorly documented. In this study, mapping has identified two prominent glacier advances in three valleys in the Cordillera Vilcabamba (13°20'S). 10Be dating reveals that the most recent glacier culminations occurred during the late AD

1700's to early 1800's, which corresponds to the late 'Little Ice Age' period (LIA; AD

1350-1860). Lichenometric dating suggests near-coeval LIA moraine stabilization in all mapped valleys. The late LIA culmination in the Cordillera Vilcabamba is broadly correlative with glacier chronologies in Europe, North America, and northern Patagonia.

However, the timing of events in southern Peru differs from culminations in Alaska and southern Patagonia and from patterns of glaciation in New Zealand. Reconstructed equilibrium line altitudes (ELAs) of glaciers in the Vilcabamba reveal an ELA rise of

x -165-200 m since the LIA, suggesting that temperatures ~1.1-1.3°C cooler could have sustained glaciers at their LIA position. The difference between L;LAs of early Holocene and LIA glaciers is small relative to the ELA rise since the LIA, which highlights the magnitude of the LIA climate oscillation in the tropics. The favored climate hypothesis responsible for sustaining more advanced Vilcabamba glaciers includes a southward displacement of the Intertropical Convergence Zone during the LIA which may have enhanced moisture delivery. The new glacier chronologies developed here augment other high-resolution Holocene tropical records, thereby allowing a fuller understanding of inter-hemispheric climate processes and linkages.

XI CHAPTER I

INTRODUCTION

1.1 Project Overview

Over 99% of all present-day tropical glaciers are located in the Andes, and more than 70% are in the Cordilleras of Peru (Kaser and Osmaston, 2002). Tropical glaciers are extremely sensitive to temperature, precipitation, humidity, and insolation variations and are among the most robust indicators of small oscillations in climate (Wagnon et al.,

1999; Kaser and Osmaston, 2002). The Peruvian Andes contain extensive and well- preserved evidence of past glacier fluctuations, but the timing of recent glacial advances in the tropics and subtropics remains poorly documented. Chronologies of latest

Holocene glacial culminations in the tropical Andes are limited by the scarcity of organic material required for radiocarbon dating of glacial deposits, and are constrained only to within broad limits (between ca. AD 1350 and 1720) from 14C ages and lichenometric dating (e.g., Mercer and Palacios, 1977; Mercer, 1984; Rodbell, 1992a; Seltzer, 1992;

Rabatel et al, 2005; Jomelli et al., 2008; Rodbell et al., 2009). A better understanding of recent Peruvian glacier fluctuations is imperative for understanding past climate changes and the role of the tropics in the origin, transmission, and magnification of climate signals across the globe (e.g., Cane and Clement, 1999). My research focuses on the 'Little Ice

Age', which represents the most recent climate oscillation of the quasi-stable Holocene,

1 and as such has important implications for recognizing anthropogenic change that underlies natural climate variability.

The 'Little Ice Age' (hereafter LIA) is best known in the North Atlantic region as an interval between approximately AD 1350-1860 when temperatures were cooler and glaciers reached their most advanced positions since the early Holocene (Grove, 1998;

2004). Available records show that the cooling varied in space and time (e.g., Nesje and

Dahl, 2003) and there is no consensus on the exact timeframe of the LIA. In the North

Atlantic region, onset of cooling began around AD 1100-1400 and glaciers attained maximum positions in the AD 1600-1800's (e.g., Bickerton and Matthews, 1993; Grove,

1998; Grove, 2004; Holzhauser et al., 2005; Luoto et al., 2008; Axford et al., 2009;

Thomas and Briner, 2009). Whether the LIA was truly a global phenomenon remains a contentious and unresolved issue as historical records are very scarce outside of Europe and the timing and magnitude of climatic responses across the globe varied markedly during this period (e.g., Schaefer et al., 2009).

Beyond terrestrial North Atlantic records, the LIA was characterized by arid conditions in Central and South America (Haug et al., 2001; Hodell et al., 2005), diminished heat transport by the Gulf Stream (Lund et al., 2006), and anomalously high tropical sea surface temperatures and salinities north of the Equator (Hendy et al., 2005;

Lund and Curry, 2006). However, the driving mechanisms that initiated the LIA and transmitted this climate signal across the globe are not fully understood. Suggested controls on glacial advances associated with the LIA include temperature, precipitation

(Thompson et al., 1985, 1995), solar activity (Crowley et al., 2000), oceanic and atmospheric circulation (Lund and Curry, 2006; Lund et al., 2006), and volcanic activity

2 (Free and Robock, 1999), but the relative importance of these various controls remains uncertain.

This study tests the hypothesis that the last major glacial culmination in the southern Peruvian Andes was a regional event which occurred during the northern hemisphere LIA. The research strategy focuses on developing glacial-geologic records in three adjacent valleys in the Cordillera Vilcabamba of southern Peru. Geomorphic maps demarcate former glacial extents, and the timing of glacier culminations is resolved with lichenometry and ' Be surface exposure dating. Results suggest that the most recent glacier advance culminated during the late LIA period. The geomorphic mapping and

10Be exposure dating programs expand on previous work initiated in the Rio Blanco valley (Licciardi et al., 2007), whereas the lichen measurements presented here constitute the first lichenometric investigation in the Cordillera Vilcabamba. The new glacial- geologic records developed here place limits on past climate conditions through equilibrium line altitude (ELA) reconstructions of glaciers at dated moraine positions.

Climate reconstructions constrain the maximum temperature rise since the LIA culmination, while the inferred climate forcings include a precipitation decrease since the

LIA. This new record of LIA glacial culminations in the Cordillera Vilcabamba fills an important geographic gap in existing regional ice core and terrestrial paleoclimate records, thus increasing spatial and temporal coverage for identifying and assessing patterns of climate change during the late Holocene in the tropical Andes.

3 1.2 Study Location

The Cordillera Vilcabamba (13°20'S, 72°33'W) is one of three major mountain ranges in the central Peruvian Andes (Figure 1). The Vilcabamba is the fourth most glacierized mountain range in Peru, with an estimated 173 km ice-covered area, after the

Cordilleras Blanca (723 km2), Vilcanota (539 km2), and Central (174 km2) (Morales-

Arno, 1999). Despite the extensive present-day glacier coverage, the glacial history of the Vilcabamba is poorly known. The only previous geologic investigations in the

Cordillera Vilcabamba include bedrock geology mapping (Egeler and de Booy, 1961;

Bowman, 1968; Marocco, 1978), landslide investigations near Machu Picchu (e.g.,

Vilimek et al., 2006), and geomorphic mapping and sample collection for cosmogenic

10Be surface exposure dating in 2006 (Licciardi et al., 2007; 2008).

The tallest peak, Nevado Salcantay (6271 m asl; 13°20'S, 72°32'W), is flanked by valley glaciers on all sides. Other high mountains, including Nevados Tucarhuay to the southwest (5910 m asl; 13°21'S, 72°35'W), and Sacsarayoc (5991 m asl; 13°15'S,

72°48'W) and Choquetacarpo (5512 m asl; 13°13'S, 72°51'W) to the northwest, are similarly glacier-clad. Extensive glacial deposits occur in the valleys radiating from these high peaks. Moraines deposited by glaciers emanating from Salcantay and

Tucarhuay were chosen as the focus of this study due to their prominence, preservation, accessibility, and proximity to granodiorite bedrock (see section 1.3).

Field investigations were initiated by J.M. Licciardi in 2006 in the Rio Blanco valley on the south side of Nevado Salcantay. Two dominant moraine positions in the

Rio Blanco valley were informally designated the "inner" and "outer" moraines. This previous work, which used 10Be dating to determine the ages of early Holocene (8.6 ± 0.3

4 ho' 1 J* "^P A y A :f

IPS :ma?§

I ^ t 1 :^-

Cordillera -' Blanca T * Cordillera hfO°S Vilcabamba

Quelccaya *% ice cap EJ: ••'"/*" VM L.3KC . 4* I'i f. 75° W mm- _l

Figure 1. Map showing the location of the Cordillera Vilcabamba in southern Peru, along with nearby paleoclimate study sites mentioned in the text. The Cordillera Vilcabamba is situated in the central Peruvian Andes (pink), which are flanked by the eastern (oriental) and western (occidental) Peruvian Andes (blue and yellow, respectively) (after Morales-Arno, 1999).

5 ka) and late LIA (AD 1810 ± 20) glacial culminations in the Rio Blanco drainage, is further detailed in Licciardi et al. (2007, 2008, 2009). New field work in support of my thesis was conducted in summer 2008, and expanded on the previous work to include investigations of glacial deposits beyond the Rio Blanco valley, encompassing the

Tucarhuay valley south of Nevado Tucarhuay and the Sisaypampa valley east of Nevado

Salcantay. The objectives of the 2008 field expedition were to build on the cosmogenic l0Be glacial chronology in the Cordillera Vilcabamba and to determine the age of the inner moraines independently using lichenometry. New mapping (Tucarhuay and

Sisaypampa valleys), along with lichenometric dating (Tucarhuay, Rio Blanco, and

Sisaypampa valleys) and 10Be surface exposure dating (Tucarhuay and Sisaypampa valleys) of moraines are discussed here, and are also included in an in-press publication

(Licciardi et al., 2009).

1.3 Geologic Setting

The tallest mountains in the Cordillera Vilcabamba are composed of an igneous granodiorite to quartz monzonite intrusion referred to as the 'Vilcabamba batholith' or the 'Machu Picchu batholith' (Bowman, 1968; Marocco, 1978; Vilimek et al., 2006)

(Figure 2). The Vilcabamba batholith intruded into an envelope of thermally metamorphosed Pre-Cambrian quartzites and phillites during a strong orogenic event associated with extensive folding during the Permo-Triassic (Upper Paleozoic). The bedrock geology of the Cordillera Vilcabamba is pertinent to this study because the boulders within moraines are sourced primarily from the Vilcabamba batholith, and to a lesser extent from the metamorphosed sedimentary rock into which the granodiorite 6 ifc •S 11*30' % V PES* ,0°

U'00' .ism

•>•

1T3G' -12 ¥)

13 00' - cusco BOO'

.'&

v -•>,

25 50 km T

Figure 2. Geologic map of the Cusco region, modified from Marocco (1978). Locator map of Peru depicts location of the Cusco province. Black box indicates location of field sites in the Cordillera Vilcabamba. Pink unit corresponds to the Vilcabamba batholith and dark brown unit corresponds to metasedimentary lithologies. intrudes. These boulders provide a stable substrate for lichen colonies and contain ample quartz for 10Be surface exposure dating (-30% quartz in granodiorite; quartz veins in metasedimentary rock).

1.4 Climate Setting

The Cordillera Vilcabamba is located in the outer tropics at 13°S latitude, where insolation receipt is nearly constant throughout the year. The seasonal range in mean daily temperature is relatively small compared with more pronounced diurnal temperature variations (Kaser and Osmaston, 2002). Meteorological records (1954-1970) from nearby Cuzco (3312 m asl; 13°33'S, 71°59'W) document mean temperatures (Johnson et al., 1976; climate data available from www.senamhi.gob.pe/) (Figure 3a). According to these records, mean maximum monthly temperatures fluctuate between 19° (January-

June) and 21°C (October-December), with a yearly average of 20°C. Mean minimum monthly temperatures fluctuate between ~1°C (from June-July) and ~7°C (from

December-February), with a yearly average of 5°C.

Precipitation in southern Peru is dominantly controlled by the position of the

Inter-Tropical Convergence Zone (ITCZ) and trends in the El Nino/Southern Oscillation

(ENSO) (Johnson, 1976; Vuille and Keimig, 2004). In the ITCZ the trade winds sweep in from the Atlantic over 1800 km away and converge in a low-pressure zone, causing warm moisture-laden air to rise and release heavy precipitation. The ITCZ maintains a more southerly position during the wet season (December-April) in southern Peru and shifts north during the drier months, occupying its most northerly position at the height of the dry season (June-August). During the dry season, available moisture originates in the

8 Figure 3. Meteorological data (1954-1970) from Cuzco, Peru (3312 masl; 13°33'S, 71°59'W). (A) Mean monthly precipitation, (B) mean monthly maximum and minimum temperatures, and (C) mean monthly relative humidity (%). Data sourced from Johnson (1976).

MAMJ J ASON D

Month Pacific when the height of the coastal inversion layer is relatively low. A pronounced east-west precipitation gradient exists in the central Andes, such that on average the eastern Cordilleras receive 600-1000 mm of precipitation annually, whereas the western cordilleras receive 50-400 mm of precipitation annually (Vuille and Keimig, 2004).

Records from the nearby Machu Picchu meteorological station (25 km north of field site;

13°10'S, 72°32'W, 2563 m asl) recorded 1946 mm year"1 mean annual rainfall (1965-

1976 and 1999-2005), 70% of which fell between December and April (Vilimek et al.,

2006). Meteorological data from Cuzco register 750 mm year"1 mean annual rainfall

(1954-1970), with December-March monthly precipitation exceeding 100 mm (Johnson et al., 1976) (Figure 3b). Precipitation varies on interannual timescales in association with ENSO, during which the central Andes receive lower than average precipitation (El

Nino events) or higher than average precipitation (La Nina events) (Vuille and Keimig,

2004).

Seasonal variations in modern glacier accumulation and ablation in the outer tropics are, dominated by moisture availability (Kaser and Osmaston, 2002).

Accumulation is limited to periods of high precipitation and to the high-elevation areas of glaciers. Ablation is controlled mainly by albedo, which is tightly linked to solid precipitation (Favier et al., 2004). Intense melting occurs during periods of low accumulation when low-albedo bare ice is exposed and short-wave energy receipt is high.

Records of monthly mean relative humidity (1954-1970) indicate highest average humidity values in the wet season (December-May) when monthly mean relative humidity values range from 55-66% (Figure 3c) (Johnson et al., 1976).

10 1.5 'Little Ice Age' Glacial History of the Peruvian Andes

LIA glacier advances in subtropical Peru and Bolivia are poorly dated but most records place the culmination between the AD 1400s and 1800s (Table 1). The earliest published account of historical glacial events in Peru dates back to 1943 when Broggi

(1943) described observed retreat of glaciers from recently held positions. Pioneering investigations of presumed LIA moraines in the tropical Andes were conducted by

Clapperton (1972, 1983), who noted evidence of recent glaciations in the Cordillera

Blanca (9°S, 78°W). Peruvian moraines thought to represent the LIA remained undated until Mercer and Palacios (1977) and Mercer (1984) obtained radiocarbon ages from bog peat incorporated in moraines marking the most recent glacial advances near the

Quelccaya Ice Cap in the Cordillera Vilcanota (14°S, 70°W) (Table 1). These studies led by Mercer established maximum limiting ages of 630 ± 65 14C years (AD 1310-1410) and 270 ± 80 14C years (AD 1510-1810) for recent moraines in the Vilcanota and near

Quelccaya, respectively. When calibrated, these radiocarbon ages correspond to multiple calendar age intercepts of comparable probabilities, and thus place relatively broad maximum limits on the LIA culmination age (Table 1). Wright (1984) found no evidence for any glacial advances within the past four centuries based on radiocarbon dating peat from deltas in lake sediments in west-central Peru (10°S, 76°W). Rodbell (1992a) used radiocarbon and lichenometric dating to document glacial advances within the last several hundred years in the Cordillera Blanca. Goodman et al. (2001) analyzed soil development and 14C in the Cordillera Vilcanota near the Quelccaya Ice Cap to constrain the maximum LIA advance ages to 330 ± 50 and 270 ± 80 14C years (AD 1550 ± 40 and

1770 ± 40). Similarly, when calibrated, radiocarbon ages correspond to multiple calendar

11 Table 1. Dates of the LIA culmination in Peru and northern Bolivia from previous investigations. Ages obtained with lichenometry and 10Be are direct. Ages obtained with radiocarbon and soil development methods are either maximum or minimum limiting ages. Ice core proxies identify the duration of the LIA in the region but not the culmination. LIA culmination Location Dating method Source (moraine age, years AD) after 1680 ±80+ C. Vilcanota radiocarbon ages Mercer and Palacios, after 1320 ±65+ 1977; Mercer, 1984 after 1620 ±50+ soil development and Goodman et al., + C. Vilcanota after 1680 ±80 radiocarbon 2001 radiocarbon and Kelly et al., 2007; Between 1600 and 1700 C. Vilcanota cosmogenic 10Be 2008 last several hundred years C. Blanca radiocarbon and Rodbell, 1992 lichenometry late 1600's C. Blanca lichenometry Solomina et al., 2007 1590-1720 C. Blanca lichenometry Jomelli et al., 2008 after 570 ± 70+ C. Real radiocarbon ages Seltzer, 1992 before 170 ±90+ AD 1630 ± 30 C. Real lichenometry Rabatel et al., 2005 C. Vilcabamba, AD 1810 ±20 cosmogenic 10Be Licciardi et al., 2009 Rio Blanco valley Thompson et al., duration of LIA: 1500-1900 Quelccaya Ice Cap ice core proxies 1985; 1986

"""Radiocarbon ages reported by Mercer and Palacios (1977), Mercer (1984), Seltzer (1992), and Goodman et al. (2001) have been calibrated to calendar age equivalents using CALIB 5.0.2 (Stuiver et al., 2005) and the values reported here are those with the greatest age probabilities. Multiple intercepts of comparable probabilities exist for most of the above radiocarbon ages. Radiocarbon age 630 ± 65 yBP corresponds to calendar ages AD 1310-1360 (p=0.597) and AD 1380-1410 (p=0A03) (Mercer and Palacios, 1977; Mercer, 1984). Radiocarbon age 270 ± 80 corresponds to calendar ages AD 1510-1600 (p=0.28), AD 1620-1700 (p=0.35), and AD 1730- 1810 (p=0.37) (Mercer and Palacios, 1977; Mercer, 1984). Radiocarbon age 170 ± 90 corresponds to calendar ages AD 1670-1740 (p=0.332), AD 1800-1820 (p=0.203), AD 1830- 1900 07=0.328), and AD 1910-1950 (p=215) (Seltzer, 1992). Radiocarbon age 270 ± 80 corresponds to calendar ages AD 1620-1700 (p=035) and AD 1730-1810 (p=0.37) (Goodman et al., 2001). Radiocarbon ages reported in the text and table are rounded to the nearest decade.

age intercepts of comparable probabilities (Table 1). In the Cordillera Real, Bolivia

(16°S, 69°W), Seltzer (1992) used radiocarbon ages from peat bogs to bracket the most recent LIA advance to between 570 ± 70 to 170 ± 90 14C years (AD 1420 ± 30 and 1710

12 ± 40), which also correspond to multiple calendar age intercepts of comparable probabilities (Table 1).

Recent lichenometric studies in the Cordillera Blanca, Peru place the LIA maximum in the early AD 1600's (Solomina et al., 2007; Jomelli et al., 2008). Jomelli et al. (2008; 2009) also report that some glaciers in the Cordillera Blanca advanced during the early LIA, around AD 1200-1350. Lichenometric dating in the Cordillera Real,

Bolivia places the LIA culmination in the late AD 1600's (Rabatel et al., 2005; 2008).

The lichenometric studies in the Cordilleras Blanca and Real also identified several glacier pauses or minor readvances between AD 1730 and 1870. Licciardi et al. (2007;

2008) and Kelly et al. (2007; 2008) reported the first 10Be moraine exposure ages for late

Holocene moraines in the region. Using a combination of 10Be and radiocarbon dating,

Kelly et al. (2007; 2008) dated the LIA culmination in the Cordillera Vilcanota to AD

1600-1700. Licciardi et al. (2009) dated the LIA culmination in the Rio Blanco valley in the Cordillera Vilcabamba (13°S, 72°W) to AD 1810 ± 20. A complementary constraint on the timing of the LIA in the southern tropics comes from Thompson et al. (1985;

1986), who provided the first tropical ice core record of the LIA and interpreted snow accumulation rates, 6180 fluctuations, and other ice core proxies as evidence for the influence of LIA climate between AD 1500-1900.

13 CHAPTER II

METHODS

2.1 Geomorphic Mapping

Lateral and end moraines along with other glacio-geomorphic features deposited by two valley glaciers were mapped on the south side of Nevado Tucarhuay and on the east side of Nevado Salcantay. Surficial mapping was conducted with the aid of a

1:100,000 scale topographic map (Machupicchu, Peru; Instituto Geografico Nacional, edition 1-IGN, series J631, sheet 2344; 1998) and augmented with detailed descriptions of moraine location, preservation, morphology, stratigraphic relationships, and occurrence of glaciofluvial outwash surfaces. Moraine crest coordinates and elevations were measured to within ± 8.5-20 m with a GPS (Appendix B). Moraine portions not directly surveyed were triangulated using a clinometer and compass (see section 3.2).

GPS coordinates of moraine crests were uploaded and plotted on Google Earth and

Google Terrain base maps, overlain with the IGN topographic map.

2.2 Moraine Dating Methods

This study is the first in the Andes to combine lichenometric and 10Be age control on the same geomorphic features. The strategy of employing multiple dating methods enables the development of robust glacial chronologies.

14 2.2.1 Lichenometric Dating

Lichenometry is a traditional surface exposure dating technique which has been used to date moraines in Peru since the pioneering work of Rodbell (1992a) (e.g.,

Solomina et al., 2007; Jomelli et al., 2008). Lichens are symbiotic associations of algae and fungi which grow in small bush shapes (foliose lichen) or flat disc shapes (crustose lichen) on rocks and other hard substrates. Rhizocarpon, a yellow-green crustose lichen which grows radially from its center, is the genus on which most lichenometry studies rely due to its ubiquity and ease of identification. Rhizocarpon lichens are typically identifiable only to the subgenus level {Rhizocarpon subgenus Rhizocarpon; hereafter

Rhizocarpon sp.) as different species are indistinguishable from each other in the field

(Innes, 1985; Rodbell, 1992a). Lichenometric dating relies on the colonization of stable rock surfaces and an independently calibrated lichen growth rate in order to obtain an exposure age of a surface (Beschel, 1950; Innes, 1985).

Long and short axes of Rhizocarpon sp. were measured on flat boulder surfaces on stable portions of lateral and end inner moraine crests in the Tucarhuay, Rio Blanco, and Sisaypampa valleys. Nearly circular thalli were measured to reduce the risk of including coalesced lichen. The single largest lichen per boulder was measured with a flexible, transparent plastic ruler to the nearest mm, with an estimated measurement uncertainty of ± 0.5 mm. In the Tucarhuay valley, the lichen data set size (n=97) was limited by the number of stable surface boulders present on the largest inner moraine. In the Rio Blanco valley («=300) diameters were measured on the terminal portion of the moraine and on the east moraine crest (Appendix A). In the Sisaypampa valley (n=100)

15 diameters were measured on the south crest of the inner moraine. Outer moraines in the three valleys were not considered for lichenometric study because the early Holocene ages of these moraines (Licciardi et al., 2009) lie well beyond the oldest age limit for reliable lichen dating.

Factors affecting lichen growth rate and colonization time include substrate lithology, moisture availability and precipitation, radiation, altitude, aspect, wind exposure, snow cover, and temperature (Beschel, 1950; Benedict, 1967; Innes, 1985;

Benedict, 1990; Rodbell, 1992a). Lichen aspect was not recorded in this study, but thalli from all aspects were measured and such effects are assumed to average out. Remaining factors, with the exception of substrate lithology, are assumed to be comparable among moraines and should not affect relative age comparisons. Regarding substrate lithology,

Rodbell (1992a) noted a variation in lichen size between granodiorite and fine-grained metasedimentary rocks from several cirque moraines in the Cordillera Blanca. On three young moraine groups, Rhizocarpon sp. thalli were larger on granodiorite boulders than on metasedimentary boulders by a ratio 1.18:1. Rodbell (1992a) attributed the offset to a slower colonization time on fine-grained metasedimentary boulders relative to the more quickly colonized coarse-grained granodiorite rocks. In this study, the Rio Blanco and

Sisaypampa moraines boulders are composed predominantly of granodiorite, whereas the

Tucarhuay moraine boulders are primarily metasedimentary (Figure 2; Appendix D).

Three lichen dating methods were employed to determine moraine ages in the field area. The first method uses an averaging maxima approach, which is a traditional method whereby five, ten, or more thalli diameters are typically considered (e.g., Innes,

1984; McCarroll, 1993). The averaging maxima approach was developed in the 1980's

16 in order to circumvent problems associated with an earlier method developed by Beschel

(1961) that relied on the use of the single largest lichen as an age predictor. It has been argued that Beschel's 'single largest' approach may hinge on anomalously large lichen, and that the average of several largest lichen should be used to reduce this possibility and reduce standard deviations associated with ages (Innes, 1984). Innes (1984) determined that five is the optimal number of diameters when averaging maxima (hereafter 'Five

Largest'), and his approach is implemented in this study.

The Five Largest method was first applied in the Cordillera Blanca by Rodbell

(1992a) and subsequently refined by Solomina et al. (2007) (Figure 4). The Five Largest growth curve used in this study was constructed using control points presented in

Solomina et al. (2007). Solomina's calibration points include 11 sites on the Pacific side of the Cordillera Blanca and one reportedly reliable radiocarbon age published by

Rodbell (1992a). This radiocarbon age is a maximum limiting radiocarbon date from a peat sample contained in a moraine dated to 1570 ± 170 14C years BP (AD 1470 ±160 years ago; ± la) obtained by Rodbell (1992a), and recalibrated here with CALIB 5.0.2

(Stuiver et al., 2005) (Table 2). Anomalously large lichen diameters were excluded from analysis following Solomina et al. (2007), after Calkin and Ellis (1980), whereby the largest lichen is considered to be anomalous if its size exceeds the next largest lichen by

20% or more. Equations for Solomina's curve were not published, therefore equations for the Cordillera Blanca Rhizocarpon sp. growth curves were determined here following methods modified from Solomina et al. (2007), according to the formula: log (y) = a + bx, where y is moraine age (years), x is lichen size (mm), and a and b are constants

17 6000 8000

1000 1500 2000 Age (years ago)

Figure 4. Previously published lichen growth curves for the Cordillera Blanca based on the average of the Five Largest lichen extreme value criteria. A. Preliminary curve redrawn from Rodbell (1992a). Error bars are for radiocarbon dates within la. B. Preliminary curves redrawn from Solomina et al. (2007). Gray dots are control points used to construct the growth curve. The two black curves depict limits constrained by the oldest (1630 cal years ago) and youngest (1340 cal years ago) bracketing ages on the radiocarbon date obtained by Rodbell (1992a). Dashed lines display 20% error intervals for both curves.

18 determined by a least squares regression of log-age versus thallus diameter. To avoid a systematic bias in age estimates, two growth curves were constructed based on the upper and lower limits of Rodbell's radiocarbon date and an error envelope encompasses an additional 20% error associated with these ages. To account for variations in substrate lithology, lichen diameters from metamorphic boulders on the Tucarhuay inner moraine were multiplied by 1.18 to normalize values to granitic lithologies in order to be directly comparable with Rio Blanco and Sisaypampa diameters (Rodbell, 1992a).

The second lichen dating method used here is the Generalized Extreme Value method (GEV), devised by Cooley et al. (2006) and Naveau et al. (2007) in order to resolve two distinct problems associated with traditional lichen dating methods (e.g., Five

Largest), namely that (1) separating lichens into two groups (diameter and age) is statistically arbitrary; and (2) traditional growth curve methods are unable to propagate age uncertainties (Jomelli et al., 2008). The entire distribution of lichen diameters is utilized for calibration points and the 50 largest lichen diameters are utilized from the sample sites to calculate ages, hence ages obtained with this method comparable to those obtained with an 'averaging maxima' approach. Diameters are modeled with the GEV distribution upon which a Bayesian hierarchical model is built (Cooley et al., 2006;

Naveau et al., 2007). In brief, the GEV distribution depends on three parameters (u, a, £) which describe the location, scale, and shape of the distribution and can be identified with a growth curve. A Monte Carlo Marcov Chain (MjCMC) resampling strategy is run for over 100,000 iterations in order to determine the best combination of ft, a, and £ parameters. Associated error is calculated using the standard mean and variance of the

19 age distribution. Full details of the method are given in Cooley et al. (2006), Naveau et

al. (2007) and Jomelli et al. (2008).

Calibration points used for the GEV method include 19 sites from Pacific and

Atlantic facing slopes, (Solomina et al., 2007) and 10 new calibration points developed by Jomelli et al. (2008) in the Cordillera Blanca from archaeological monuments mainly

on the Atlantic side of the mountain range. Age uncertainties associated with the

calibration points are presented in Table 3. GEV ages were computed at the Centre

National de la Recherche Scientifique (CNRS) in France by Vincent Jomelli and

Delphine Grancher.

The third lichen dating method used in this study is the 98% quantile method

(Lowell et al., 2005). This method has been used to construct growth curves in New

Zealand, but is employed here as a measure of relative (rather than calendar) moraine ages. The 98% quantile is defined as the size of a single thallus that is larger than 98% of the sample population regardless of sample size (Lowell et al., 2005). Similar to the

GEV, the 98% quantile method considers the entire distribution of thallus diameters and therefore improves upon traditionally used 'averaging maxima' methods which critically depend on the total number of lichen diameters measured. Lowell and others (2005) determined that the 98% quantile was the best metric suited for comparing lichen populations of different sizes, as is the case in this study.

2.2.2 Cosmogenic 10Be Surface Exposure Dating

A growing number of studies have demonstrated the utility of terrestrial in-situ cosmogenic nuclide surface exposure dating of Andean moraines as a means to

20 Table 2. Calibration points used in Solomina et al. (2007) and Jomelli et al. (2008) growth curves. Site name Age Method Source (years AD) Honcopampa 775 ±125 five radiocarbon ages from ceramic fragments r archaeologic and lichen diameter consistency Quilcayhuanca valley 775 ±125 with Honcopampa r Huacramarca 1050 ±150 historical period attributed to Huacramarca r pottery and architecture attributed to Late Llanganuco valley 1 1300 ±100 Intermediate period r pottery and architecture attributed to Late Llanganuco valley 2 1300 ± 100 Intermediate period r Inca period; Ibarra, personal communication Macarjirca 1485 ±47 with Jomelli r not reported; using lichen species Lecanora Pueblo Viello (Viejo) not reported rupicola and Orphinospora r Huaritambo 1485 ±47 Inca period r not reported; using rare Rhizocarpon and more Incarraga not reported abundant Lecanora rupicola and Orphinospora r Huanuco Pampa not reported not reported r+ 1970s rockfall deposits due to road construction s Cancaraca-2 + 1932-1948 second terrace of the lake s Cancaraca-2 + 1923-1924 terminal moraine dam of the lake s Cancaraca-2 + 1970/1971 end moraine, partly in the lake s Allicocha + 1948-1962 moraines of hanging glacier s Allicocha + 1923-1924 glacier advance 1923-1924 s Atlante + 1923-1924 glacier advance 1923-1925 s Atlante + 1970s youngest moraine ridge s Atlante + 1951 ±7 proglacial till s* Broggi + Broggi 1940 ±8 proglacial till s* s*+ Broggi 1975 ±5 rockfall deposits + 1943/1944 ±9 s* Yanamarey proglacial till + Yanamarey 1939 moraine s* s*+ Yanamarey 1948 proglacial till surface + Uruachraju 1968 moraine s* s*+ Uruachraju 1985 moraine + Artesonraju 1932 moraine s* s*+ Artesonraju 1961 moraine + Huascaran 1970 debris flow s* Quilloc moraine 1470 ±160 radiocarbon dating R*

J - New points added to the growth curve for GEV analysis by Jomelli et al. (2008). Lichen Rhizocarpon sp. were used unless otherwise specified. S - Calibration points from Solomina et al. (2007). R - Calibration point from Rodbell (1992a). * Demarcates calibration points used to create the Solomina et al. (2007) growth curve using the average of the Five Largest lichen from sites on the Pacific side of the Cordillera Blanca. + Demarcates calibration points used to create the Jomelli et al. (2008) GEV growth curve. Jomelli et al. (2008) assigned error equivalent at ± 10 years to all S calibration points.

21 reconstruct past climate (e.g., Kaplan et al., 2005; Douglass et al., 2005; Smith et al.,

2005, 2008; Zech et al., 2007; Kelly et al., 2007, 2008; Licciardi et al., 2009). Surface

boulders on moraine crests are deposited during the culmination of glacier activity, hence

their exposure ages may be used to determine the onset of glacier retreat from an

advanced position. This study relies on the production of cosmogenic 10Be in crystalline

quartz to determine exposure ages of moraine boulders and, in turn, the timing of ice

retreat and moraine stabilization.

Rock samples were collected with a hammer and chisel from six large, well-

preserved boulders on stable inner moraine crests in the Tucarhuay and Sisaypampa

valleys for 10Be analysis following established sampling protocols (e.g., Licciardi and

Pierce, 2008). Boulders were sampled from nearly horizontal surfaces where possible to

minimize shielding effects. Most samples from inner moraines retained glacial polish

(Licciardi et al., 2009). Boulder edges and corners were avoided in order to minimize

edge effects of 10Be production. Boulders higher than the ground surface by ~1 m or

more were targeted for sampling to avoid potential snow shielding and exhumation

effects (Appendix C). Shielding from the horizon was measured with a clinometer, along with surface strike and dip measurements, in order to determine corrections for shielding

effects on 10Be production. Samples PE08-2, PE08-3, and PE08-14 are from quartz veins

in metasedimentary boulders from the inner Tucarhuay moraine, but sample PE08-14 did not yield sufficient quartz for a 10Be measurement and is not discussed further (Figure 6).

Samples PE08-4, PE08-5, and PE08-6 are from granitic boulders on the inner

Sisaypampa moraine (Figure 6).

22 Sample preparation for Be analysis was conducted at the University of New

Hampshire according to well-established procedures (Licciardi, 2000). Two duplicate samples (PE08-4D and PE08-6D) were prepared at the Lamont-Doherty Earth

Observatory (LDEO) for the purpose of inter-lab comparison. Standard techniques of rock crushing, grinding, isolation of quartz by repeated leaching in HF / HNO3, and preparation of target material (BeO) by ion-exchange chromatography and selective precipitation are presented in Appendix E. 10Be/9Be ratios were measured at the

Lawrence Livermore National Laboratory Center for Accelerator Mass Spectrometry

(LLNL-CAMS) under the direction of Robert C. Finkel and Dylan H. Rood. All age calculations and shielding corrections were made using the CRONUS-Earth (Cosmic-Ray produced NUclide Systematics on Earth) 10Be exposure age calculator Version 2.2, available online at http://hess.ess.washington.edu/math (Balco et al., 2008).

The ages presented here are calculated using a 10Be production rate determined from a high altitude (4045 m) calibration site in central Peru (9.65°S) (Farber et al., 2005) whose altitude and latitude are comparable to the Vilcabamba sites (Appendix E). The reference sea-level high-latitude (SLHL) 10Be production rate derived from the Peruvian calibration site of Farber et al. (2005) is 4.23 ± 0.09 atoms g"1 yr"1 (la, in quartz; Lifton et al. (2005) scaling), which is -9% lower than the SLHL 10Be production rate based on the current combined CRONUS calibration site data set following Lifton scaling (4.6 atoms g"1 yr"1) (Lifton, personal communication, May 2009). The SLHL 10Be production rate must be scaled to the location of each sample because nuclide production varies with altitude, latitude, and atmospheric pressure. 10Be production rates at the Cordillera

Vilcabamba field sites require only slight latitudinal and altitudinal scaling from the high-

23 altitude low-latitude Farber et al. (2005) calibration site, which minimizes potential

scaling uncertainties, compared with use of the global calibration dataset which relies on

production rates biased toward low-altitude sites at northern mid-latitudes.

Atmospheric conditions in the Andes deviate from the standard atmosphere, most

markedly at high altitudes. The non-standard atmosphere for the central Andes (Farber et

al., 2005) is used to determine the atmospheric pressure at boulder site altitudes. These

non-standard atmospheric parameters are equivalent to those applied by the calculator's

default height-pressure relationship (Balco et al., 2008). With a best fit to Andean meteorological data from eight stations, the atmospheric conditions comprise a sea-level pressure of 1012.8 mbar, an adiabatic lapse rate of 6.5 mK m"1, and a sea-level

temperature of 301.73 K (Farber et al., 2005).

Factors including snow cover, topographic shielding, and surface erosion can

affect sample exposure ages (Gosse and Phillips, 2001). If snow cover is present, it will

shield geomorphic surfaces from cosmic rays and yield anomalously young surface

exposure ages, in which case shorter boulders would be expected to have greater snow

cover and yield comparatively younger ages than taller boulders. Snowpack data are not

available for the field site, but no inverse age-height relationship is seen in this study, hence snow cover is assumed to be negligible. Surrounding topography can also block

cosmic rays from the horizon and thereby result in anomalously young ages. Age calculations account for topographic shielding effects (Appendix B). Surface erosion of

sampled boulders is assumed to be negligible because glacial polish was noted on most inner moraine boulders.

24 2.3 ELA Reconstruction

The equilibrium line altitude (ELA) corresponds to the altitude on a glacier

surface where net annual inputs equal losses. Changes in ELA over time are often used

in paleoclimate studies to provide insight on glacier sensitivity and lend clues about

climate driving mechanisms (Benn and Lehmkuhl, 2000). The location of the ELA is a

function of local climate, mainly temperature and precipitation, but to a lesser degree is

also controlled by humidity and cloud cover/insolation (Vuille et al., 2008). The most

common and practical approach to understanding glacier-climate interactions relies on

changes in temperature and precipitation (Kaser and Osmaston, 2002). Glaciers recede

and ELAs rise if temperatures increase and/or precipitation decreases, and conversely,

glaciers advance and ELAs fall if temperatures decrease and/or precipitation increases.

In this study, ELAs have been determined for modern glaciers and reconstructed for paleo-glaciers following three commonly used methods (Porter, 2001). The strategy of

employing multiple ELA methods enables the development of robust ELA and paleoclimate estimates.

The first method is the Maximum Elevation of Lateral Moraines (MELM), which provides a reliable and robust means of determining the minimum altitude of former glacier ELAs and relies on the principle that lateral moraines are formed ohly in the ablation zone, i.e., below the ELA (Benn et al., 2005). Ice in the ablation zone flows upward and out toward the glacier margins, bringing with it debris which is deposited along the margins of the ablation zone; the uppermost limit of this debris is the ELA

(Meierding, 1982). Where moraines are poorly preserved, MELM is less effective at determining ELA location because the elevation of lateral moraines falls below the actual

25 ELA. Where moraines are well preserved, MELM provides a secure method of

determining paleo-ELAs and can be particularly reliable where AAR values are poorly

constrained (Richards et al., 2000; Benn et al., 2005). In this study, MELMs of paleo-

glaciers are considered to be the most direct and reliable indicator of ELA in the

Cordillera Vilcabamba, and thus serve as a reference upon which the other ELA methods

are based. The MELM method is used here to determine ratios for the following two

ELA methods, akin to the approach followed in the Cordillera Real by Seltzer (1992).

The second ELA reconstruction method uses the ratio of the accumulation area to the ablation area, or the Accumulation Area Ratio (AAR), which assumes that, under steady-state conditions, the accumulation area of a glacier occupies some fixed proportion of the glacier area (Meierding, 1982; Benn et al., 2005). Application of AAR takes into account glacier hypsometry, but does not consider annual mass balance measurements. AAR values that have been typically applied to glaciers in tropics and sub-tropics of Peru and northern Bolivia range from 0.5-0.82 (e.g., Klein et al., 1999;

Porter, 2001; Kaser and Osmaston, 2002; Smith et al., 2005).

The third ELA method employed here uses the Toe-to-Headwall Altitude Ratio

(THAR), which assumes that the ELA can be approximated by a fixed ratio between the altitudes of the terminus and head of the glacier (Meierding, 1982; Benn et al., 2005).

THAR does not account for glacier hypsometry or glacier mass balance, but can be a valuable tool for determining ELAs where detailed hypsometries are lacking and AAR ratios are not well established. Typical THAR values employed in the Andean subtropics range from 0.2-0.5 (e.g., Osmaston, 1975; Rodbell, 1992b; Seltzer, 1992; Porter, 2001;

Mark et al., 2002).

26 2.3.1 Modern Mass Balance and ELA

The modern ELA was calculated for five glaciers in the Moyoc, Yanama,

Sisaypampa, Otiyoc, and Huascacocha valleys using AAR and THAR. The optimal

AAR and THAR coefficients were first determined on paleo-glaciers, according to the best fit with MELM elevations (see section 2.3.2) and those coefficients were then applied to find modern ELAs.

The modern glaciers were selected for analysis based on their relatively simple glacier hypsometries and the potential to determine ELA changes over time (i.e., also the presence of corresponding inner moraines that were easily identified in the field and on topographic maps, aerial photographs, Landsat Thermal Mapper (TM) imagery, and

Google Earth images). Glaciers in the Rio Blanco and Tucarhuay valleys were excluded from modern ELA analysis because they are partly reconstituted, such that portions of the glacier are separated by steep, unglaciated rock slopes and ice transfer occurs via avalanches. The ELA of a reconstituted glacier is generally impossible to determine reliably because the theoretical ELA may be located anywhere on the glacier or avalanche track (Benn and Lehmkuhl, 2000). In the Tucarhuay valley, identifying the modern ELA is further complicated because the existing glaciers cannot confidently be discerned from each other on Landsat TM images.

Modern glacier extents were inferred from Landsat TM images from the dry season (21 July 2001) acquired from the United States Geological Survey (USGS) Global

Visualization Viewer (GloVis) (Figure 5). Only high-quality satellite images were chosen for this study based on quality of data, recent acquisition date, absence of cloud

27 Figure 5. Landsat TM images of the Cordillera Vilcabamba during the dry season (acquisition date 21 July 2001). A. True-color, B. False-color, C. NDSI, and D. TM band 4/TM band 5 images were used for mapping modern glacier extents and identifying glacial deposits for mapping paleo-glacier extents. cover, absence of recent snow cover, and a capture date toward the end of the dry season.

The USGS GloVis images used in this study are radiometrically and geometrically corrected with ground control and elevation control points. Thresholded ratio images were used for accurately distinguishing clean ice and snow from clouds, rock, vegetation, and other materials. The ratio TM band 4/TM band 5 was applied as it enhances contrast in snow zones (Williams et al., 1991), especially in areas obscured by shadow (Hall et al.,

1987; Paul et al., 2002). The Normalized Difference Snow Index (NDSI = [TM band 2 -

TM band 5] / [TM band 2 + TM band 5]) was employed because it is used to effectively distinguish snow from soil, vegetation, rock, and clouds, and is effective for mapping snow cover over rugged terrain (Figure 5) (Hall et al., 1995; Sidjak and Wheate, 1999).

28 2.3.2 Paleo-Glacier Reconstruction and ELA

The glacier margins corresponding to inner moraines in the Cordillera

Vilcabamba were reconstructed by following cirque morphology and crests of lateral and

end moraines for six glaciers according to established methods (Benn and Luhmkuhl,

2000; Benn et al., 2005; Krusic et al., 2009). Reconstructed glaciers include those in the

Rio Blanco, Moyoc, Yanama, Sisaypampa, Otiyoc, and Huascacocha valleys. Above the

elevation of the highest preserved ice-marginal deposits, former glacier boundaries were

identified by breaks in slope along valley walls and above the margins of existing glacier heads where slopes higher on the valley wall were deemed too steep to have held a

glacier. The glacier margin corresponding to the Rio Blanco outer moraine was reconstructed by following the crests of lateral and end moraines.

Glacier-margin elevations were estimated from the base of moraine crests using

Google Earth images. Elevations obtained from contoured Google Earth terrain maps were found to be on average 30 ± 21 m lower than ~100 waypoint elevations measured in the field by GPS (Appendix C). The GPS waypoint elevations are accurate to ± 20 m or less, implying that Google Earth and GPS elevations agree within respective error. To ensure consistency, Google Earth elevations were used in all ELA reconstructions.

Contour lines with 20 m intervals were interpolated over the paleo-glacier surfaces, drawn in curved lines mimicking the shape of modern glaciers where glaciers are not currently present, and drawn along topographic lines where modern glaciers exist.

MELMs were estimated from lateral moraine elevations obtained from Google

Earth images for a total of 8 former glaciers (the six aforementioned reconstructed

29 glaciers plus two other glaciers located at 13°12'34"S/70°50'17"W and

13oll'18"S/70o46'53"W). Lateral moraine elevations were verified with clinometer

sightings in the Rio Blanco and Sisaypampa valleys. It was noted where maximum

elevations of preserved lateral moraines may be limited by very steep terrain or obliterated by landslides and/or outwash, as is common in high-elevation glacial environments. The most appropriate AAR and THAR values were determined by iteratively identifying best-fit coefficients which yielded ELAs in best accordance with the mean of MELM ELAs for all paleo-glaciers at the LIA culmination (4720 m) (see section 3.3). All reconstructed paleo- and modern glacier ELAs are calculated with these best-fit values (AAR = 0.60 ±0.10; THAR = 0.40 ± 0.05).

Some error is likely associated with paleo-glacier reconstruction. The largest potential sources of error include ambiguities in identifying headwall elevations, distinguishing dynamically connected versus independent glacier lobes, interpolation of moraine elevations from Google Earth images, and glacier margin estimations. It is assumed that paleo-glaciers selected for ELA determinations were not reconstituted.

Moreover, the Moyoc, Yanama, Otiyoc, and Huascacocha paleo-glacier margins

(moraines) were not directly dated, but instead assumed to be of comparable age to the inner Tucarhuay, Rio Blanco, and Sisaypampa moraines dated with 10Be and lichenometry in this study and by Licciardi (2009) based on similarities in distance from headwalls and morphological appearance on satellite images.

30 2.3.3 ELA Depression and Paleoclimate Implications

ELA depressions are used here to infer the magnitude of climate forcing necessary to sustain glaciers at more advanced positions. Differences between modern

and LIA ELAs have been calculated (AELAM.LIA) in the Sisaypampa, Moyoc, Yanama,

Otiyoc, and Huascacocha valleys. The difference between the LIA and early Holocene

ELA has been calculated (AELALIA-EH) in the Rio Blanco valley only. Paleotemperatures based on ELA depressions have been calculated for many tropical glaciers (e.g., Seltzer,

1994; Porter, 2001; Ramage et al., 2005). Here, annual mean temperature reductions required to sustain glaciers at their formerly advanced positions are calculated using a lapse rate of 6.5°C km"1 (Johnson, 1976). In this calculation, the precipitation component is ignored because snow accumulation data were not accessible. This simple temperature lapse-rate approach is commonly employed where precipitation data are insufficient for rigorous paleo-precipitation estimates (e.g., Benn et al., 2005; Ramage et al., 2005) and, although it assumes that temperature is the only control on glaciers, this approach provides a useful first-order estimate of climate change in southern Peru during the LIA.

31 CHAPTER III

RESULTS

3.1 Geomorphic Relationships

Field mapping in the Tucarhuay and Sisaypampa valleys identified two prominent moraine complexes in each valley, referred to as "inner" and "outer" moraines, which largely mimic the geomorphic expression of moraines in the Rio Blanco valley (Figure

6). The inner moraines are extremely sharp-crested and slumping is common on their ice-proximal sides. Outer moraine crests form subrounded ridges and mounds. Surface boulder frequency is higher on the inner moraines than for the corresponding outer moraines in all valleys. Detailed geomorphic photographs can be found in Appendix D.

In the Tucarhuay valley a large outer moraine occurs as two ridges with subrounded crests (photo Ala) separated by a breach. The outer moraine is downvalley from three adjacent inner moraine loops, defined here as "western", "central", and

"eastern" inner moraines (photos Alb,c,d). The predominant boulder lithology on moraines at Tucarhuay is metasedimentary, but some granitic boulders also occur. The

Tucarhuay central inner moraine (Figure 6; photos Ale) is the largest of the three inner

Tucarhuay moraines and is the only one of the three directly surveyed in this study. The central inner moraine was the focus of lichen measurements and sample collection for

10Be measurements. The inner moraines are dissected by breaches (photos Alc,d), presumably cut during glacial meltwater outbursts or perhaps later when the moraine

32 NEVADOSALCANTAY 6271m

PEO^ J^^tJ^ mm r Sisaypampa ; y / /- • valley /

NEVADOTUCARHUAY 5910 m o

./';/ Rio /ff Blanco V valley H 3 km

•^futwashrj,. PE08-14 ,-.N „„^ ,x cH - inner moraine 10 v outer moraine > le sample location Tucarhuay PE08-3 \X^/r kame terraces - bog valley triangulated crest < "•n Jake . w -•-'Stream - inferred crest Intermittent stream

Figure 6. Geomorphic map of field area in the Cordillera Vilcabamba, Salcantay vicinity, created on a Google Earth topographic map base imaee. Moraine crests, outwash surfaces, and 10Be sample locations are indicated. failed behind a dammed lake. Two blue-green glacial meltwater lakes are dammed upvalley by the central moraine (photo Ale). An alluvial fan emanates from the breach in the western inner moraine (photo Aid). The area between the central inner moraine and the outer moraines is punctuated by three amorphous mounds which may represent poorly preserved moraines, slumping, or moraine breach material. These mounds are unsuitable for 10Be surface exposure dating as they lack large boulders. The gap between the inner and outer moraines is covered by an extensive outwash surface interpreted to be contemporaneous with the inner moraines. Another outwash surface associated with the outer moraines extends downvalley beyond the outer moraine ridges (photo Ala). A stream carrying glacial meltwater, the Quebrada Tucarhuay, cuts into the upper outwash plain between the central and western inner moraines, flows downslope, and cuts through the breach area of the outer moraine ridges.

In the Rio Blanco valley (described in more detail in Licciardi et al., 2009), a single prominent inner moraine loop (photo A2c) occurs 2 km upvalley from two ridges with subrounded crests that comprise the outer moraine (photo A2a). Large granitic boulders occur on the inner and outer moraines. Lichen measurements on the Rio Blanco valley inner moraine come from the eastern lateral moraine ridge and the terminus. The outer moraine encloses a bog, which was probed to a depth of >2 m (photo A2b). Several small but discernible ridges are adjacent to the most prominent outer moraine ridges, suggesting modest glacier fluctuations during moraine deposition.

Two major moraine sets in the Sisaypampa valley include an outermost moraine complex of lateral ridges (photo A3a) and one large inner moraine (photos A3b,c).

Boulders on moraines in the Sisaypampa valley are predominantly granitic. Boulders on

34 the southern loop of the inner moraine were the focus for lichen measurements in this valley. The inner moraine is characterized by a large ridge breached at its terminus by a braided meltwater channel (photo A3b). Several smaller ridges are discernible, suggesting some degree of glacier fluctuation (photo A3c). On the south side of the valley the outer moraine complex consists of seven to ten small nested ridges (photo

A3a). The distal outer moraine ridges hug the slope of a topographic rise and extend uphill to form a loop extending to the west, indicating a lobe of the glacier pushed up into a tributary valley as it advanced. Other outer ridges mimic this curvilinear trace with the exception of the most proximal nested ridge which instead parallels the inner moraine

(photo A3d). A kame terrace hugs the slope of the topographic rise above the outermost moraine and lacks large boulders, which makes it unsuitable for 10Be surface exposure dating. The kame terraces and a parabolic cross-valley profile downvalley from the outer moraines (photo A3b) indicate more extensive earlier glacial advances, perhaps corresponding to late glacial events recorded in other Peruvian cordilleras (e.g., Smith et al., 2008). Older glacial deposits have likely been wiped out by active mass wasting and fluvial scouring. Outer moraine ridges occur on the north side of the valley but are more muted. A small stream carrying glacial meltwater from the tributary valley to the southwest cuts through the succession of outer moraines before flowing downvalley into the Quebrada Sisaypampa. A small upper outwash infilling occurs between the outer moraine and the nested ridges, and was most likely was deposited concurrently with the nested ridges. A more extensive lower outwash plain emanates from the breach in the inner moraine and occurs downvalley.

35 3.2 Moraine Chronology

3.2.1 Li chenometry Results

Lichen diameters from moraine boulders in the Tucarhuay, Rio Blanco, and

Sisaypampa valley are represented in histograms (Figure 7). Diameters were used to compute the calendar age of the moraines with the Five Largest and GEV approaches, as well as relative moraine ages with the 98% quantile technique. Five Largest and 98% quantile results discussed below are adjusted for metasedimentary boulder lithology at

Tucarhuay, following Rodbell (1992a).

Tucarhuay Rio Blanco Sisaypampa

n=WO

0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35 Lichen diameter (mm) Lichen diameter (mm) Lichen diameter (mm) Figure 7. Histogram of lichen diameters measured on the Tucarhay, Rio Blanco, and Sisaypampa inner moraines.

Averages of the Five Largest lichen diameters are 23.8, 28.6, and 24.2 mm for the

Tucarhuay, Rio Blanco, and Sisaypampa inner moraines, respectively. The growth curve after Solomina et al. (2007) is shown in Figure 4 and reconstructed here (Figure 8).

Corresponding ages estimated from the reconstructed growth curve (Figure 8) are 140 ±

20, 200 ± 30, and 150 ± 20 years (AD 1870 ± 20, 1810 ± 30, and 1860 ± 20) for the

Tucarhuay, Rio Blanco, and Sisaypampa moraines, respectively (Table 4; Figure 9).

Ages agree within 2a between the three inner moraine ages, and correspond to the late

LIA period.

36 0 500 1000 1500 2000 Age (years ago)

Figure 8. Reconstructed growth curve and control points for Rhizocarpon sp. Rhizocarpon lichen, created after Solomina et al. (2007) for lichen in the Cordillera Blanca. Black dots represent control points. Black lines represent maximum and minimum age estimates which limit the oldest and youngest age estimates of the control points. Dashed lines display 20% error intervals for each curve. Growth curves are constructed according to the formula log (y) = a + bx, where y is moraine age (years ago), x is lichen diameter (mm), and a and b are constants determined by a least squares regression of log-age versus thallus diameter. Equations for the curves are as follows: upper: log (y) = 1.3144 + 0.0335x; lower: log (y) = 1.3658 + 0.0342x; upper with 20% error: log (y) = 1.3296 + 0.0315x; and lower with 20% error: log (y) = 1.3534 + 0.0358x. Growth curves are best fit lines to the control points, most of which are at the lower end of the curve. As such, the curves are weighted toward younger samples and smaller lichen diameters, and consequentially the curves do not pass directly through the oldest control point.

37 Table 3. Lichen data from the Cordillera Vilcabamba.

Five Largest GEV 98% quantile Moraine n (mm) Age (AD) Age (AD) (mm) Tucarhuay 93 20.2 1900 ±10 1890-20/+10 22.8 Tucarhuay* 93 23.8 1870 ±20 27.0 Rio Blanco 300 28.6 1800 ±30 1840-10/+20 26.0 Sisaypampa 100 24.2 1860 ±20 1890-20/+10 27.0 Tucarhuay* represents metasedimentary adjusted boulders to granitic lithology. Lichen diameters were measured to the nearest mm. Error for lichen diameters is ± 0.5 mm. Five Largest ages were calculated from the growth curves which encompass 20% error associated with calibration points.

Lichen diameter (mm) 15 20 25 30 j_ 110±10/140±20 Tucarhuay o 22.8 / 27.0 •0- 120-20/+10

•—% i 200 ± 30 Rio Blanco 26.0 170-10/+20

150 ±20 Sisaypampa 27.0 120-20/+10

T T —r T- 0 100 200 300 400 500 Age (years ago)

NUMERICAL AGE INDICATORS RELATIVE AGE INDICATORS Five Largest t0i GEV B 98% quantile

unadjusted .. unadjusted ,-, unadjusted K> u 98% quantile Five Largest "V GEV

Figure 9. Lichenometric calendar and comparative ages, including moraine ages according to the growth curve and GEV methods, and comparative moraine ages according to the 98% quantile method. Error bars represent la age uncertainty; error bars on 98% values are smaller than symbol. 'Unadjusted' samples (open symbols, gray text) have not been multiplied by 1.18, according to Rodbell (1992a)

38 Results from GEV analysis are presented in Table 4 and Figure 9. Using the GEV method, the Tucarhuay, Rio Blanco, and Sisaypampa moraines are 120 -20/+10 years,

170 -10/+20, and 120 -20/+10 years, respectively (AD 1890 -20/+10, 1840 -10/+20, and

1890 -20/+10), corresponding to the late LIA period. Lithologic adjustments were not conducted by colleagues at CNRS and therefore the Tucarhuay age cannot be directly compared to the ages in the other two valleys. Rio Blanco and Sisaypampa GEV ages overlap within 50 years.

The 98% quantile lichen thallus diameters are 27.0, 26.0, and 27.0 mm for the

Tucarhuay, Rio Blanco, and Sisaypampa moraines, respectively (Table 4; Figure 9). 98% quantile thallus sizes agree within the ±0.5 mm ruler measurement error for all three moraines.

3.2.2 10Be Surface Exposure Ages

10Be concentrations and corresponding ages are presented in Tables 5 and 6.

Tucarhuay, Rio Blanco, and Sisaypampa 10Be ages are also presented in Licciardi et al.

(2009). All ages discussed below use the Lifton (Li) scaling scheme (Lifton et al., 2005;

Balco et al., 2008). The mean of boulder ages is interpreted here to reflect the timing of moraine occupation and abandonment.

Closely agreeing ! Be exposure ages for two boulders sampled on the Tucarhuay inner moraine are 270 ± 50 (PE08-2) and 270 ± 40 (PE08-3) years, yielding a mean age

270 ± 30 years (AD 1740 ± 30). The mean is assigned error equivalent to the propagated analytical uncertainty, which is larger than the standard deviation of the two ages (3 years).

39 Table 4. Be results fromCordiller a Vilcabamba samples

Sample Qtz Ht Lat Lon Elev Press Thick Shield ,0Be X 4 (g) (m) (deg S) (degW) (m) (mbar) (cm) corr (10 atg-') blk Tucarhuay moraine PE08-2 23.377 2.30 13.3811 72.5859 4307 607.1 1.75 0.9521 0.980 ±0.178 5 PE08-3 33.474 1.35 38.3809 72.5861 4306 607.2 0.75 0.9521 1.003 ±0.132 7 Sisaypampa moraine PE08-4 40.767 1.95 13.3417 72.5191 4506 592.2 1.50 0.9618 1.125 ±0.107 9 PE08-4* 54.888 1.95 13.3417 72.5191 4506 592.2 1.50 0.9618 1.303 ±0.038 28 PE08-5 41.521 1.15 13.3427 72.5151 4399 600.2 2.00 0.9734 0.579 ±0.103 5 PE08-6 40.457 0.8 13.3428 72.5146 4379 601.7 1.50 0.9731 0.911 ±0.109 7 PE08-6* 48.608 0.8 13.3428 72.5146 4379 601.7 1.50 0.9731 0.976 ±0.032 19 Samples were spiked with the LDEO low-level 9Be carrier prepared from shielded beryl by Roseanne Schwartz at LDEO and normalized to "'Be/'Be standard 07KNSTD3110. Measurement uncertainties reflect Iff analytical error only. "Ht" is boulder height to sample. "Thick" is rock sample thickness. "X blk" represents level of 10Be sample activity above the blank. Analyses of two blanks processed in parallel with the samples yielded a straight mean 10Be/Tie of 4.2 ± 3.0xl(T15 and one blank processed in parallel with LDEO-prepared duplicate samples PE08-4D and PE08-6D yielded a '"Be/Tie ratio of 2.6 ± 0.56x 10"'5. 10Be/9Be ratios are 5-28 times above the 10Be/9Be of corresponding blanks (Appendix E). AMS measurement precision on blank-corrected sample '"Be/^Be values ranges from 2.7-18.2%.

10Be measurements from the three boulders on the Sisaypampa moraine include two duplicates prepared by colleagues at LDEO. The duplicates, representing splits of quartz from the same sample and processed independently at UNH and LDEO, yield ages of 280 ± 30 (PE08-4; UNH) and 320 ± 10 (PE08-4D; LDEO) years, and 240 ± 30 (PE08-

6; UNH) and 250 ± 10 (PE08-6D; LDEO) years. The Iff inter-lab reproducibility of 10Be measurements from these very young sample duplicates increases confidence in the results. The mean 10Be surface exposure age of the inner Sisaypampa moraine is 240 ±

80 years (AD 1770 ± 80). The Rio Blanco inner moraine age determined by Licciardi et al. (2009) is 200 ± 20 years (AD 1810 ± 20), not including two old outliers attributed to

40 isotopes inherited from prior exposure. All three moraine ages agree within 2a and correspond to the late LI A period (Figure 10).

Table 5. Comparison of l0Be exposure ages calculated under alternative scaling schemes (in years, ± 1 a).

Sample St De Du Li Lm Tucarhuay moraine PE08-2 251 ±46 272 ± 50 262 ± 48 269 ±49 290 ± 53 PE08-3 254 ± 34 276 ± 37 266 ± 36 273 ± 37 295 ± 39 mean ± std dev — > 253 ± 29 274 ±31 264 ±30 271 ±31 293 ±33

Sisaypampa moraine PE08-4 260 ± 25 278 ± 27 268 ± 26 275 ± 26 302 ± 29 PE08-4D 302 ± 09 322 ±10 311 ±09 319±10 349 ±10 PE08-5 139 ±25 150±10 145 ± 26 149 ±27 162 ±29 PE08-6 220 ± 26 238 ± 28 229 ± 27 235 ± 28 225 ± 30 PE08-6D 236 ± 08 254 ± 09 245 ± 27 252 ± 09 274 ± 09 mean ± std dev --> 224 ± 80 240 ± 84 232 ±81 237 ± 83 259 ± 92 "St" - Lai (1991), Stone (2000); "De" - Desilets et al. (2006); "Du" - Dunai (2001); "Li" - Lifton et al. (2005); and "Lm" - Lai (1991), Stone (2000), Nishiizumi et al. (1989). See Balco et al. (2008) for details on all scaling schemes. Moraine age means were calculated as straight means, with the exception of duplicate pairs (PE08-4 and PE08-4D, PE08-6 and PE08-6D) which were first calculated as weighted means (i.e., biased toward more precise ages), and then treated as individual boulder ages, la analytical uncertainty represents the larger of either the standard deviation or the propagated analytical uncertainties on the l0Be measurements. 10Be ages reported in this study are expressed in years before collection (2008 for Tucarhuay and Sisaypampa; 2006 for Rio Blanco) and rounded to the nearest 10 years in the main text.

41 -I I I L_J I L -I I I I I I l_ j i i i Tucarhuay 270 ± 30

*o-

Rio Blanco 200 ±20

Sisaypampa 240 ± 80

T—>—'—•—r 0 200 400 600 800 1000 1200 Age (years ago)

10 10 ^ Beage •#• Be age duplicate 1 std.dev ^ outlying 10Be age ^ duplicate pair or 1 o from >•« 10Be age (Licciardi et al., 2008) mean

Figure 10. 10Be ages according to Lifton scaling scheme for the Tucarhuay, Sisaypampa (this study, and Rio Blanco moraines). Boulders are ordered from youngest to oldest. Error bars represent la uncertainty associated with each sample. Gray area represents 1 standard deviation or la uncertainty from internal mean. Gray symbols represent 10Be age duplicates processed at LDEO; red and yellow symbols represent samples processed at UNH (red, this study; maroon, Licciardi et al., 2009). Open symbols represent outliers not included in the mean.

42 3.3 ELA and Paleoclimate Results

Individual ELAs determined for a total of five modern glaciers, eight LIA

glaciers, and one early Holocene glacier are summarized in Table 7 and Figure 11. The

extents of modern and reconstructed paleo-glaciers are presented in Figure 12. The mean modern ELA is 4920 ± 60 following the AAR method (coefficient 0.60 ±0.10) and 4930

± 40 following the THAR method (coefficient 0.40 ± 0.05). The modern ELA for the north-facing glacier (Huascacocha) is -100 m higher than the modern mean; ELAs of southeast- and southwest-facing glaciers (Sisaypampa and Otiyoc) are -20-140 m lower than average; modern ELAs of south-facing glaciers are at or up to 100 m above the modern mean (Yanama and Moyoc).

At the LIA culmination, the mean AAR-derived ELA is 4730 ± 70 m and the mean THAR-derived ELA is 4710 ± 60 m (Figure 11). The LIA ELA for the north- facing Huascacocha glacier is higher than the LIA mean (-120 m); ELAs of southeast- and southwest-facing glaciers are within -50 m of the average, and ELAs of south-facing glaciers (including Rio Blanco) vary within -100 m of the average. Results suggest that

ELAs have risen by 160 ± 80 m (AAR) to 200 ± 60 m (THAR) since the LIA maximum

(Table 7). In comparison, glacier surface area decreased since the LIA maximum by

-30% to -45% (Figure 13; Table 8). North-, east- and west facing glaciers decreased the most drastically (43-46%) while south-facing glaciers shrank the least (32-34%) over this time interval. If the ELA rise since the LIA was forced solely by a change in temperature, LIA temperatures would have been 1.1 ± 0.5°C (AAR) to 1.3 ± 0.4°C

(THAR) cooler than present, assuming a 6.5°C/km adiabatic lapse rate.

43 Table 6. ELA estimates according to THAR, AAR, and MELM, AELA, and Atemperature. Rio Blanco is omitted from modern ELA because it is reconstituted; two glaciers are omitted from LIA ELA due to complex glacier hypsometries and limited Google Earth coverage. Modern ELA THAR 0.40 ±0.05 AAR 0.60 + 0.1 -0.1 MELM Rio Blanco — — — — — — Yanama 4954 43 5025 65 55 — Moyoc 4945 40 4935 45 60 — Sisaypampa 4878 48 4780 80 90 — Huascacocha 5056 57 5010 70 70 — Otiyoc 4808 26 4865 35 35 — Modern average 4928 43 4923 59 62 — LIA ELA THAR 0.40 ±0.05 AAR 0.60 + 0.1 -0.1 MELM Rio Blanco 4632 51 4610 30 20 4800 Yanama 4690 65 4855 105 105 4734 Moyoc 4758 56 4750 100 90 4767 Sisaypampa 4692 64 4710 80 80 4770 Huascacocha 4834 75 4770 40 70 — Otiyoc 4664 38 4710 55 50 4621 13°12'34770°50' — — — — — 4589 13°11'18772°46' — — — — — 4737 LIA average 4712 58 4734 68 69 4717 Early Holocene THAR 0.40 ±0.05 AAR 0.60 + 0.1 -0.1 MELM Rio Blanco 4520 67 4575 0 25 —

A ELAM-LIA THAR 0.40 ±0.05 AAR 0.60 + 0.1 -0.1 MELM Rio Blanco — — — — — — Yanama 264 65 170 105 105 — Moyoc 187 56 185 100 90 — Sisaypampa 186 64 70 80 80 — Huascacocha 222 75 240 70 70 — Otiyoc 144 38 155 55 50 — Average A 201 60 164 82 79 —

A ELALiA.EH THAR 0.40 ±0.05 AAR 0.60 + 0.1 -0.1 MELM Rio Blanco 112 67 35 30 20 —

A Temp.M.LIA THAR 0.40 ±0.05 AAR 0.60 + 0.1 -0.1 MELM Rio Blanco — — — — — — Yanama 1.7 0.4 1.1 0.7 0.7 — Moyoc 1.2 0.4 1.2 0.7 0.6 — Sisaypampa 1.2 0.4 0.5 0.5 0.5 — Huascacocha 1.4 0.5 1.6 0.5 0.5 — Otiyoc 0.9 0.2 1.0 0.4 0.3 — Average 1.3 0.4 1.1 0.5 0.5 —

A Temp.M.UA THAR 0.40 ±0.05 AAR 0.60 + 0.1 -0.1 MELM Rio Blanco 0.7 0.4 0.2 0.2 0.1 290'

280°

270°

260°

250'

Modem ELAs 'Little Ice Age'ELAs Early Holocene ELAs AAR THAR MELM AAR THAR MELM AAR THAR O D Yanama o o ° Rio Blanco • a Rio Blanco O D Moyoc o • a Yanama • "_' Sisaypampa • • Moyoc O u Huascacocha ::• Sisaypampa O • otiyoc o 0 • Huascacocha o o a Otiyoc Modern THAR ELA Other — — Modern AAR ELA o LIATHAR ELA LIAAARELA LIA MELM ELA

Figure 11. ELAs and AELAs in the Cordillera Vilcabamba. Rose plot of ELAs of modern, LIA, and early Holocene glaciers in m a.s.l. are according to glacier aspect. Open symbols represent modern ELAs, closed symbols represent LIA ELAs, and closed symbols with black outline represent early Holocene ELA. An AAR coefficient of 0.6 ± 0.1 and a THAR coefficient of 0.40 ± 0.05 were used (± represented by error bars). Dashed lines show the average ELA determined by various methods. Dashed lines show the average ELA for all glaciers examined in this study by various methods.

45 Figure 12. Modern and paleo-glacier reconstructions for Otiyoc, Moyoc, Huascacocha, Yanama, Sisaypampa, and Rio Blanco glaciers used for ELA estimations. "Coordinate" glaciers, defined in Table 7, are not shown here. 'Dynamically distinct ice lobes' represent adjacent glaciers or paleo-glaciers interpreted to have been separate from the reconstructed glaciers and as such have not been considered in ELA calculations.

The ELA of the Rio Blanco glacier during the early Holocene glacier culmination

(Figure 11) was determined to be 4575 ± 25 m (AAR) to 4520 ± 70 m (THAR). For the

Rio Blanco glacier only, the early Holocene ELA was thus lower by 10 ± 50 m (AAR) and 100 ± 70 m (THAR) relative to the ELA during the LIA (AELALIA-EH)- Assuming

ELA height is controlled solely by temperature, then early Holocene temperatures would have been between 0.2 ± 0.2°C (AAR) and 0.7 ± 0.4°C cooler than LIA temperatures in the Cordillera Vilcabamba.

46 • Otiyoc 190° 180° 170°

Figure 13. Percent of glacier surface area loss (LIA-modern) according to Landsat TM, 21 July 2001 glacier extents. Values are plotted according to glacier aspect and vary between -30% to -45% shrinkage.

Table 7. Percent of glacier surface area loss (LIA-modern) from 2001 glacier extents.

Shrinkage (LIA to Modern) Glacier Aspect % Yanama 185° /S 32 Moyoc 175° / S 34 Sisaypampa 115°/SE 43 Huascacocha 350° / N 46 Otiyoc 245° / SW 46 Average 40 Std Dev 7

47 CHAPTER IV

DISCUSSION

4.1 Geomorphic Relationships

The most recent glacial advance in the Cordillera Vilcabamba is recorded by a single prominent inner moraine crest. In contrast, the LIA is recorded by multiple moraines at most other sites in Peru and northern Bolivia (Rabatel et al., 2005; Jomelli et al., 2008; Kelly et al., 2008). A succession often closely-spaced LIA moraines occur in the Cordillera Real (Rabatel et al., 2005, Fig. 5), ten to twenty at the edges of the

Quelccaya Ice Cap, Cordillera Vilcanota (Kelly et al., 2008), and ten to fourteen in the

Cordillera Blanca (Jomelli et al., 2008). The moraine pattern differences could be due to variations in glacier hypsometry and valley geometry. For example, in the Cordillera

Vilcabamba where valleys are steep and glaciers are confined by valley walls, single, and likely composite, moraine crests are present. If moraine crests are composite, older till will be buried in the culminating ridges, which implies that glacier buildup may have been occurring throughout the LIA before glaciers reached the most recent maximum extents dated here. In contrast, in other cordilleras where glacier termini were less confined by topography, glacier geometry may have encouraged preservation of multiple moraine crests that record small glacier margin fluctuations.

In the Tucarhuay and Rio Blanco valleys the outer moraine occurs as a single prominent moraine crest whereas in the Sisaypampa valley the outer moraine occurs as a

48 sequence of seven to ten smaller lateral moraines. Analogous to the LIA moraines, the differences in outer moraine patterns are also likely due to the degree of glacier confinement between valleys. Confinement by valley walls may have limited small degrees of glacier fluctuation. For example, in the Rio Blanco valley the walls are steep and confining and the outer moraine is delineated by a single prominent moraine crest.

Conversely, the glacier in the Sisaypampa valley was able to push into a tributary valley and thereby record minor fluctuations of the glacier margin.

4.2 LIA Moraine Chronology

Both lichenometric and l0Be dating places the most recent glacier culmination in the Cordillera Vilcabamba late in the LIA period, building on pilot results from a single valley (Licciardi et al. 2007; 2008) to more thoroughly define the regional extent and timing of LIA glacier culminations. This interpretation is robust despite apparent age disparities between lichenometric and 10Be dating methods and between valleys. This study provides some of the first 10Be surface exposure ages for the most recent glacier culmination in the tropical Andes (Kelly et al., 2007, 2008; Licciardi et al., 2007, 2008), and is one of few studies to use both lichenometry and 10Be surface exposure dating on the same geomorphic surfaces (Howley, 2008).

4.2.1 Lichenometric Age Relationships

Moraine culmination calendar ages obtained with both the Five Largest and GEV lichenometry methods are based on growth curves established in the Cordillera Blanca.

Possible differences in lichen growth rates between mountain ranges may result in

49 inaccuracies in calendar lichen ages in the Vilcabamba. Disparities in growth rates between the two mountain ranges is unknown, however lichen growth depends on factors which include moisture availability and precipitation, radiation, altitude, aspect, wind

exposure, snow cover, or temperature (see section 2.2.1), which do differ somewhat between cordilleras. Climatic conditions recorded close to the Cordillera Blanca (~9°S,

77°W) in Cajamarca (7°08'S, 78°28'W, elevation 2621 m) and close to the Cordillera

Vilcabamba (13°20'S, 72°33'W) in Cuzco, Peru (13°33'S, 71°59'W, elevation 3312 m) from 1954-1970 may be indicative of inter-cordillera climate differences (Johnson,

1967). Compared to the Cordillera Blanca, Vilcabamba mean annual temperatures are similar, annual mean relative humidity is around 10% lower, mean annual precipitation is around 35 mm or 5% higher, and mean annual wind speeds are 2 knots or 30% higher.

However, these values may not be representative of regional differences because they are based on data from two meteorological stations with a 600 m difference in altitude. The lichen measurement site altitudes are similar between mountain ranges (4200-4400m)

(Jomelli et al., 2008). If the Vilcabamba lichen growth rate is different due to differences in moisture availability and precipitation, radiation, altitude, aspect, wind exposure, snow cover, or temperature, then boulders with exposure durations equivalent to those in the

Cordillera Blanca could host lichen with differing diameters in the Vilcabamba, thereby yielding anomalous exposure ages there.

Age relationships between valleys according to the Five Largest and 98% quantile are based on lichen diameters adjusted to a granitic lithology, namely by multiplying lichen diameters by 1.18 on the Tucarhuay inner moraine. The factor 1.18 was observed by Rodbell (1992a) in the Cordillera Blanca for the average of the Five Largest lichen

50 diameters on four late LI A moraines (Gueshque 2). No single Gueshque 2 moraine contained both granodiorite and metamorphic boulders. However, Rodbell (1992a) asserts that this trend is not statistically significant at the 0.05 confidence level, and as such some uncertainty is inherently associated with both the Five Largest and 98% quantile Tucarhuay calendar and relative ages.

There is an apparent age disparity in Five Largest results at the la uncertainty level between valleys that may indicate real age differences. The apparent age differences may be explained by the wide range of sample sizes (Tucarhuay «=93,

Sisaypampa n=100, Rio Blanco n=300). Tucarhuay and Sisaypampa Five Largest lichen moraine ages are derived from similar sample sizes, and overlap within Iff uncertainty.

In comparison, the Rio Blanco moraine, with a sample size 3 times larger than the

Tucarhuay and Sisaypampa populations, yields an apparently older age. These results are consistent with a theoretical increase in age bias toward larger samples (Innes, 1984). If all sample sizes were equal, it is possible that the age disparity between the moraines would decrease, a trend observed in the 98% quantile results (below). There is an apparent age disparity in GEV results as ages differ between valleys by 50 years. The explanation for differences in GEV ages is the same 'bias toward larger sample size' explanation used to explain Five Largest differences, above, as the 50 largest lichen diameters are used to obtain moraine culmination ages.

The numerical ages obtained with the Five Largest and GEV methods are similar for individual valleys and overlap within lcr. For example, the Rio Blanco inner moraine ages are 1800 ± 30 (Five Largest) and 1840 -10/+20 (GEV). Relatively small age discrepancies are likely due to differences inherent in the Five Largest and GEV methods.

51 For instance, different calibration points, calibration curves, iterative MCMC procedures, modeling strategies, and number of lichen used to calculate ages exist between the two lichen dating methods.

The 98% quantile results suggest the culmination of LIA moraine deposition in all three valleys was nearly contemporaneous. While no calendar ages are obtained with the

98% quantile method in the Cordillera Vilcabamba, I regard it as the most robust lichen index developed in this study for several reasons. First, it accommodates the different sample sizes collected in this study as it considers variable sample sizes equally. Second, unlike the Five Largest and GEV growth curves, the 98% quantile does not rely on control points that have considerable associated age uncertainty (Tables 2 and 3). Third, the Five Largest and GEV growth curves are based on control points that do not span a wide interval of time (Tables 2 and 3). Fourth, the 98% quantile is independent of probable differences in growth rates in different Peruvian cordilleras.

4.2.2. 10Be Age Considerations

The 10Be ages generated in this work are among the youngest ever obtained with this commonly-used surface-exposure dating method. Previously, the application of 10Be surface exposure dating has been difficult on young and historical surfaces because over short exposure times only very small amounts of terrestrial cosmogenic nuclides are produced, and these low nuclide concentrations are difficult to measure precisely (Finkel et al., 2008). The young ages obtained on these moraines demonstrate the recently- developed feasibility of dating historical surfaces with 10Be surface exposure dating

(Schaefer et al., 2009; Licciardi et al., 2009).

52 The Be age differences for each mean moraine age delivered by the five scaling schemes are within 30 years, or -15% percent, because minimal scaling is necessary from the 10Be production rate calibration site in Peru developed by Farber et al. (2005).

Calculation of exposure ages under any of the five published scaling schemes (Table 6) does not alter the assignment of moraines to the late LIA. Also, the agreement between age populations are not affected by systematic 10Be production rate and scaling uncertainties. However, 10Be age comparisons with records dated by other methods must take these uncertainties into consideration. 10Be ages may not be directly comparable to true calendar ages, but I am confident that they are fairly close given the realistic bounds of uncertainty in the 10Be production rate. While it is difficult to quantify the actual calendar age uncertainty, a reflection of its potential value in this case comes from the offset between the reference SLHL 10Be production rate calibrated in Peru (Farber et al.,

2005) and the production rate from current combined CRONUS calibration site data set

(Lifton, personal communication, May 2009). The Peru-derived production rate is 9% lower (see section 2.2.2) and could lead to anomalously old ages.

For low-level samples such as these, the uncertainty on the 10Be measurement is critically dependent on accurate knowledge of the blanks and how high the sample activity is above the blank. Because the blanks are well-known and reproducible (Table

5), coupled with the consistency of 10Be measurements in duplicate samples prepared independently between the UNH and LDEO labs, contamination with background meteoric 10Be is considered minimal.

Geologic and environmental factors, including boulder exhumation, boulder rotation, and snow cover, could lead to anomalously young ages. Due to coherency

53 between ages and the sampling strategy which only targeted tall boulders (~1 m) exhumation is considered improbable. Likewise, age agreement between boulders suggests that boulder rotation is unlikely. However, boulder PE08-5 is younger than the other two Sisaypampa boulders, may have be influenced by post-depositional boulder rotation, or snow or sediment cover. No relationship between boulder height and apparent age exist, therefore snow cover issues are considered negligible.

4.2.3 Comparison of Lichenometric and 10Be Ages

Lichenometric ages (Five Largest and GEV) are within 10-30 years of the 10Be age in the Rio Blanco valley, but are 90-130 years younger than the I0Be ages in the

Tucarhuay and Sisaypampa valleys. The age offset and age patterns between the two dating methods may be explained in several ways. First, any inaccuracies in the site- specific 10Be production rate used in this study could lead to an over- or underestimation of the timing of glacier culminations (see section 4.3). Secondly, differences between lichenometric and calendar ages could be due to the inaccuracy or inapplicability of the lichen growth rate from the Cordillera Blanca (see section 4.2.1). Finally, discordances between lichen and 10Be ages could be due to an older lichen age bias toward larger sample populations (see section 4.2.1).

4.2.4 Possible Differences in Glacier Response Time

Variations in moraine ages between valleys may reflect differences in glacier response time to changing climate conditions. Smaller glaciers typically respond more quickly to changes in climate than do larger glaciers, assuming comparable glacier

54 velocities and activity indices (Johanneson, 1989; Bahr et al., 1998; Oerlemans, 2005).

At the height of the LIA advance, the Rio Blanco glacier had the largest area of any glacier at the field site. The Sisaypampa glacier occupied about 85% as much area as the

Rio Blanco glacier, and the Tucarhuay central glacier occupied about 60% of the Rio

Blanco glacier area. Assuming that glacier response time is inversely correlated to glacier size, the Tucarhuay glacier would be expected to respond more quickly to a climate perturbation, the Sisaypampa glacier more slowly, and the Rio Blanco glacier response would be the slowest. The result of variable response times is such that moraine boulders would be exposed earlier for glaciers that retreated more quickly in response to ameliorating climate conditions, and vice versa. Assuming this relationship holds true, moraine boulders are expected to yield the oldest ages on the Tucarhuay inner moraine, intermediate ages on the Sisaypampa moraine, and youngest ages on the Rio Blanco inner moraine. These hypothesized variations in glacier response time are consistent with 10Be ages which indicate that the Tucarhuay moraine is the youngest, the Sisaypampa moraine is an intermediate age, and Rio Blanco moraine is the oldest. However, uncertainty on the mean Sisaypampa moraine 10Be age is too large (± 80 years) to exclude the possibilities that the Sisaypampa glacier receded either earlier (before Tucarhuay) or later

(after Rio Blanco). The glacier response time hypothesis is not supported by the most robust lichen index, the 98% quantile, which suggests near-coincident glacier retreat.

Considering the large uncertainty associated with the Sisaypampa moraine 10Be age and the lack of support by the lichen data, this hypothesis requires further exploration with additional chronological information from LIA moraines in the Cordillera Vilcabamba.

55 4.3 Comparison with Tropical Andean Glacier Fluctuations

Previously published LIA records from the Peruvian and Bolivian sub-tropics

(Table 1; see section 1.5) are not entirely consistent with the record from the Cordillera

Vilcabamba record. The 10Be dated LIA culminations in the Cordillera Vilcabamba were one to two centuries later than those dated the Cordillera Blanca, (Jomelli et al., 2008) and up to one century later than those in the Cordillera Real (Rabatel et al., 2005; 2008), but are consistent with radiocarbon ages in the Cordilleras Vilcanota (Mercer and

Palacios, 1977; Mercer, 1984) and Real (Seltzer, 1992) (Figure 14).

It is possible that the site-specific 10Be production rate used in this study, or a combination of factors associated with 10Be dating, may have led to an underestimation of calendar ages in the Cordillera Vilcabamba (see section 4.2.2), though this is not considered likely due to the large age difference (e.g., 30%) between ice advances in the

Cordilleras Blanca and Real. Another possibility is that uncertainties associated with lichen growth curves may yield ages too old in the nearby Cordilleras (see section 4.2.1).

Also, the radiocarbon ages obtained in nearby Cordilleras suggest the LIA culmination occurred after AD 1310-1810 (after 630 ± 65 and 270 ± 80 14C yr BP) in the Vilcanota

(Mercer and Palacios, 1977; Mercer, 1984) and before AD 1670-1950 (170 ± 90 14C yr

BP) in the Real (Seltzer, 1992). However, calendar ages associated with these 14C ages have multiple intercepts with similar probabilities (Table 1), such that age precision is not high enough to draw robust conclusions about whether the culminations occurred before those in the Vilcabamba. A final possibility is that the age differences reflect real regional differences in glacier culminations, though the various age uncertainties do not conclusively demonstrate this at the present time.

56 J 1 I I I l__l I I I I I I I I I I I I 1 I

Canada

Coastal

Brooks Range Iceland £ d

Norway

u, Norway I

$ Aletch I

Gorner -t Colorado a a c'3 -* Front Range New Zealand Q h. •—j—i H>-<)—i i (y- X> 0 0 L ] N.Patagonia S. Patagonia t- ] CVilcan.,Peru ] C Real, Bolivia ] CBIan,Peru TucC RioC O* Sis C > r x x x x x T—'—r 1 2000 1900 1800 1700 1600 1500 1400 1300 1200 1100 1000

Age (years AD)

Figure 14. Comparison of global LIA glacial records. Timing of the late LIA glacier culminations in the Cordillera Vilcabamba, Peru (bottom) depicted by red diamonds as 10Be ages compared with selected global glacier fluctuations. Error bars represent 1 a uncertainty. Advances in Canada (a. Luckman, 2000) and Coastal Alaska (b. Wiles et al., 2008) are derived from dendrochronology. Each bar shows the length of a tree ring record. Culminations in the Brooks Range, Alaska from lichenometric dating are shown in c. (Sikorski et al., 2009). Fluctuations in Iceland are derived from lichenometry and tephrochronology (d. Bradwell et al., 2005); the largest orange diamond represents the LIA culmination of Lambatungnajokull at the Vatnajokull ice cap, followed by smaller orange diamonds which represent minor readvances. Records in Norway (e. Nesje et al., 2008) are derived from lichenometry and historical sources; culmination represented by a purple diamond and glacier length variations are lines. Fluctuations in the Norway, Aletch, and Gorner glaciers in the Swiss Alps are reconstructed from historical accounts, tree rings, and radiocarbon data from fossil wood (f. Holzhauser et al., 2005). Graph (g. Benedict, 1973) shows the LIA culmination in the Colorado Front Range dated by lichenometry. Graph (h. Schaefer et al., 2009) shows culminations in New Zealand dated with 10Be, the magnitude of which is depicted by the size of gray diamonds. Fluctuations in northern Patagonia (i. Garibotti and Villalba, 2009) are depicted by brown diamonds. Radiocarbon ages of culminations in southern Patagonia are represented by black bars (j. Villalba, 1994). LIA culmination in the tropical Andes is represented in the Cordillera Vilcanota, Peru (k. Mercer and Palacios, 1977) as a dashed line, dated by radiocarbon dating, and the Cordilleras Real, Bolivia (1. Rabatel et al., 2005; 2008) and Blanca, Peru (m. Jomelli et al., 2008) as dated by lichenometry. The size of the diamonds represents the magnitude of the glacier advance. 5-7 4.4 Comparison to Global LIA Chronologies

The new results presented here provide an opportunity to compare chronologies on a global scale. The most salient and best-known LIA chronologies against which I draw a comparison include selected records from the North Atlantic, North America,

Patagonia, and New Zealand because these studies represent precisely dated glacial chronologies spanning a wide geographic representation (Figure 14). The most robust correlations are with glacier culminations in the Canadian Rockies (late AD 1700's-

1800's) (Luckman, 2000), southeastern Iceland (AD 1780's) (Bradwell et al., 2006),

Norway (early AD 1700's to the late 1800's) (Nesje et al., 2008), the Swiss Alps (early- mid AD 1800's) (Holzhauser et al., 2005), the northern Patagonian Andes (early-mid AD

1700's) (Garibotti and Villalba, 2009), and the Colorado Front Range (AD 1700-1900)

(Benedict, 1973). LIA glaciers in Alaska (McKay and Kaufman, 2009; Wiles et al.,

2008; Sikorski et al., 2009) and southern Patagonia (Villalba, 1994) culminated less than a century prior to glaciers in the Cordillera Vilcabamba. In contrast to the Cordillera

Vilcabamba record, over the past millennium glaciers in New Zealand reached a maximum extent in the mid AD 1400's followed by less extensive glacial episodes around AD 1600 and AD 1730-1890 (Schaefer et al., 2009). In other tropical locations, such as Africa and New Guinea, glaciers had a general maximum extent during the LIA and have receded since the second half of the 19th century, though many of these records are not well-constrained (Kaser, 1999). While the Vilcabamba glacier chronology is correlative with chronologies from several locations and the data presented here help identify regions of coherency in the global record, available records suggest that LIA glacial advances were not globally synchronous.

58 4.5 ELA Determinations and Paleoclimate Inferences

The determination here of optimal THAR and AAR coefficients for glaciers in the

Vilcabamba represents a valuable contribution for future ELA estimations in the subtropical Andes. The THAR and AAR coefficients (0.40 ± 0.05 and 0.60 ± 0.10, respectively) that best align with the paleo- and modern ELAs in the Cordillera

Vilcabamba are similar to values typically applied in the region. Previous studies have determined AAR values between 0.65 and 0.77 and THAR values between 0.35 and 0.'45

(Table 9). Congruence with other independently determined AAR and THAR coefficients in the subtropical Andes lends credence to my findings.

An ELA lowering of 160 ± 80 m (AAR) to 200 ± 60 m (THAR) during the LIA culmination in the Cordillera Vilcabamba is remarkably consistent with other estimates in the sub-tropical Andes (Table 10), providing confidence in these results. Following the temperature lapse rate of 6.5°C km"1 for the central Andes (Farber et al., 2005) and assuming that temperature forcing alone was responsible for glacier expansion, these

ELA lowerings correspond to a depression of 1.3 ± 0.4°C (AAR) and 1.1 ± 0.5°C

(THAR) in the Cordillera Vilcabamba during the LIA. These estimates essentially match predictions by Rabatel and others (2008) that temperatures would have had to be 1.1-

1.3°C cooler in the Cordillera Real, Bolivia at the LIA culmination. The climate reconstructions offered here represent upper bounds on temperature depressions (i.e., those required if temperature was the sole control). However it is unlikely that temperature was singularly responsible for LIA glacier fluctuations, but rather changes in precipitation probably also played a critical role, given that the Quelccaya ice core data

59 clearly show large variability in accumulation rates during the LIA (Thompson et al.,

1985; 1986) coupled with the observation that modern mass balance of sub-tropical glaciers is largely controlled by precipitation and albedo (Favier et al., 2004; Vuille et al.,

2008).

At the height of the early Holocene culmination, the ELA in the Rio Blanco valley was 35 ± 30 m (AAR) to 110 ± 70 m (THAR) lower than at the height of the LIA.

This ELA depression is not directly comparable to other studies in the region because no other early Holocene ELAs in the region outside the Cordillera Vilcabamba have been reconstructed. The estimation that early Holocene temperatures were 0.2 ± 0.2°C (AAR) to 0.7 ± 0.4 (THAR) cooler than the LIA demonstrates the relatively large magnitude and abruptness of the climate changes in the short interval of time since the LIA maximum, relative to the rest of the Holocene. The relationship between glacier aspect and ELA is ambiguous as no notable relationship has been observed.

Table 8. AAR and THAR values for tropical Andean glaciers, as determined independently in various studies.

Location AAR THAR Reference C. Vilcabamba 0.60 ±0.10 0.4 ±0.05 this study C. Blanca 0.65 Jomelli et al. (2008) C.Real 0.77 0.37 Seltzer (1992) Central Andes 0.67 0.35 (modern); 0.45 (paleo) Klein et al. (1999) tropical glaciers 0.65-0.70 Kaser and Osmaston (2002)

60 Table 9. Modern ELAs and LIA ELA depressions in the tropical Andes, as determined independently in various studies. LIA ELA Location Modern ELA (m asl) depressions (m) Reference C. Vilcabamba 4920 (AAR); 160 ±80 (AAR); this study 4930 (THAR) 200 ± 60 (THAR) C. Blanca 4900-5000 160 Clapperton(1990) C. Blanca -5100 Seltzer (1992) C. Blanca -5100 175 Jomelli et al. (2008) 5098 ± 98 (mass C. Real balance); 158 ±30 Rabatel et al. (2005; 2008) and 285 ± 50 5250 (AAR) NE tropical Andes Klein etal. (1999) 4500-4700 NW tropical Andes Klein etal. (1999) up to 5100 northern Patagonia Klein etal. (1999) up to 5700

4.6 Comparison with Other Proxies

The new Cordillera Vilcabamba glacial record augments other available high- resolution climate proxy records in the region and provides valuable insight for identifying plausible climate drivers. One of the most instructive comparators is the high-resolution Quelccaya ice core record 250 km southeast of the Cordillera Vilcabamba

(Thompson et al., 1985; 1986) (Figure 15). In the Quelccaya ice cores, an increase in particulates and associated conductivities from AD 1500-1900 has been attributed to increased wind velocities across the high altiplano of southern Peru due to changes in atmospheric circulation (Thompson et al., 1985; 1986). The period from AD 1500 to

1720 stands out as an interval of exceptionally high snow accumulation, presumably when Vilcabamba glaciers were at advanced positions, followed by a period of low

61 accumulation (drought) from 1720 to 1860, when Vilcabamba glaciers were receding from LIA maximum positions according to the ioTB e results of this study.

ei

£2

l—•—i—•—i—'—i—'—r T—•—l—"—l—'—l—'—l—'—r 2000 1900 1800 1700 1600 1500 1400 1300 1200 1100 1000 Age (years AD)

Figure 15. Comparison with tropical ice core records. The timing of late LIA glacier culminations in the Cordillera Vilcabamba, Peru (bottom) depicted by red diamonds as 10Be ages compared with tropical ice core proxy data. Error bars represent Iff uncertainty. The top three graphs (a., b., c.) show ice core proxies from the Quelccaya ice cap, Peru (Thompson et al., 1985; 1986). Darkened areas represent deviations from modern values. The Quelccaya 5180 is compared to 5180 records from other tropical cores Huascaran (d. Thompson et al., 1995), Illimani (e. Thompson et al., 1998), and Sajama (f. Ramirez et al., 2003).

62 Negative 6I80 excursions at Quelccaya are interpreted by Thompson et al. (1986)

to correspond to periods of lower temperatures from AD 1500 to 1900, with a notable

1 R

low from AD 1800-1820. While 5'°0 in ice cores is likely controlled in part by

atmospheric temperature, stable isotopes from tropical ice cores are difficult to interpret because the isotopic signature is also affected by amount of precipitation, water vapor recycling, and water circulation (Hoffman, 2003). Low-resolution tropical Andean ice core records from Huascaran, Peru (6048 m, 9°06'S,77°36'W), Illimani, Bolivia (6350 m, 16°37'S, 67°46'W), and Sajama, Bolivia (6542 m, 18°06'S, 68°53'W), similarly document negative 5180 and 5D excursions (Thompson et al., 1995; Thompson et al.,

1998; Ramirez et al., 2003). These low values may also correspond to lower temperatures over the period AD 1600-1800, AD 1700-1900, and AD 1500-present, in the Huascaran, Illimani, and Sajama cores, respectively. Ramirez et al. (2003) postulate that the stable isotope fluctuations are controlled by rainout intensity over the Amazon and the Altiplano whereby negative excursions reflect wetter conditions, which suggests that these periods were wetter, and not necessarily cooler. While it is not entirely certain that the 5180 values signify cooler temperatures, our results suggest that the Vilcabamba glaciers retreated toward the end of an inferred period of prolonged coldness. However, the favored scenario here is that Vilcabamba glacier retreat was forced by the drought and not the end of the cold period.

Taken together, proxy ice core data showing cooler temperatures and higher precipitation are interpreted to suggest that precipitation led to a buildup of Vilcabamba glaciers and caused them to advance, but the drought ensuing after AD 1720 eventually led to glacier retreat. The onset of retreat some time after drought conditions took hold

63 may have been a consequence of reduced ablation during a period of prolonged cold or

due to a lag in glacier response time to the lower accumulation rates.

Other tropical paleoclimate proxies are consistent with tropical Andean ice core

precipitation records. Marine and terrestrial tropical climate proxies during the LIA

support enhanced precipitation which may have sustained glaciers at more advanced

positions in Peru (Figure 16). At Lake Titicaca (16°-17.50°S, 68.5°-70°W), a large basin

representative of precipitation patterns across a wide area of tropical South America,

sediment core data indicate rising to overflowing water levels between AD 1500 and

present (Abbott, 1997; Baker et al., 2001), when Vilcabamba glaciers are assumed to

have been advancing. Hillyer et al. (2009) documented a pronounced dry event at small

Lake Pacucha, Peru (13°36'26"S, 73°29'42"W) around 750 calendar years before present

(BP), or AD 1250, likely predating Vilcabamba glacial advances. The dry event at

Pacucha somewhat predates the onset of LIA conditions at Quelccaya by -250 years, but has been correlated to a period of decreased accumulation at Quelccaya (Thompson et al.,

1985; 1986), and suggests some degree of uniformity in regional climate. In the nearby

Maracocha Lake basin (13° 13 S, 72° 12 W), Chepstow-Lusty et al. (2003) recorded a long, dry episode between AD 900-1800, which does not support the evidence for wetter conditions during the LIA. However, the resolution of the Chepstow-Lusty et al. (2003) lake record is likely not high enough to fully capture such abrupt climate events. Unkel et al. (2007) recorded a transition from arid to semi-arid conditions from mid-1300's to early 1700's in the coastal desert of Peru (14°30'S), indicating that climate became wetter, and supporting the possibly that a common climate forcing influenced both the

64 I • » • • T • . I • | , | , | , | , | , l|-

T—«—i—'—i—•—i—•—i—•—i—«—i—•—i—•—i—•—i—•—r 2000 1900 1800 1700 1600 1500 1400 1300 1200 1100 1000 Age (years AD)

Figure 16. Comparison with climate change proxies. The timing of late LIA glacier culmination in the Cordillera Vilcabamba, Peru (bottom) depicted by red diamonds as 10Be ages compared with climate records across the globe. Error bars represent Iff uncertainty. Graph (a.) represents solar irradiance derived from Antarctica (Bard et al., 2000), yellow, and from sunspot records (Lean et al.,1995), orange. Dalton, Maunder, Sporer, and Wolf minima are denoted with letters. Graph (b.) shows Northern Hemisphere temperature reconstructions from Esperer et al., (2002), light gray, and from Mann et al. (1999), dark gray. Graph (c) shows surface water 5180 in cores from the Dry Tortugas, with higher values reflecting higher Florida Current surface salinity (Lund and Curry, 2006). Graph (d.) shows volume transport of the Florida Current (Lund et al., 2006). Graph (e.) shows 5180 from coral in Panama (Linsley et al., 1994), black. Graph (f.) shows variations in 6180 from lake carbonates from the Yucatan (Hodell et al., 2005). Graph (g.) shows titanium % in Cariaco Basin sediments, with lower values representing greater aridity in Venezuela and a more southerly position of the ITCZ (Haugetal.,2001).

65 coastal desert and Vilcabamba glaciers at this time. Gutierrez et al. (2009) used sediments and biological productivity to record a large, rapid, and persistent change in climate, ocean circulation, and biogeochemical cycling off the coast of Peru (12°00'S,

72°42'W andl4°07'S, 76°30'W) toward the end of the LIA, which coincides with the retreat of Vilcabamba glaciers. Most of these climate records show that conditions were wetter in the tropical Andes during much of the LIA and that a rapid changes in climate and ocean circulation occurred at the end of the LIA. The records introduce the possibility that common climate driving mechanisms influenced moisture availability in terrestrial, marine, and Vilcabamba glacial records both leading up to the and following the LIA culmination.

4.7 Possible Climate Drivers

Suggested drivers of LIA climate in the tropics include changes in temperature, precipitation, and insolation driven by volcanic forcing (Solomina et al., 2007), solar variability (Bard et al., 1999; Polissar et al., 2006), ENSO activity (Vuille and Keimig,

2004), and changes in the ITCZ (Lund and Curry, 2006; Lund et al., 2006; Kelly et al.,

2008; Licciardi et al., 2009). The eruption of Huaynaputna in AD 1600 in southern Peru may have caused some degree of decreased insolation, but this single event was not likely the primary cause of glacier advance and retreat southern Peru (Silva and Zielinsky,

1998). Periods of decreased solar activity occurred at four distinct times during the LIA; the Wolf solar minimum ~ AD 1250-1350, the Sporer minimum around 1450-1550, the

Maunder minimum -1650-1750, and the Dalton minimum in the mid to late 1800's

(Figure 16) (Bard et al., 1999), which may have altered atmospheric temperatures and

66 influenced Vilcabamba LIA glacier activity, yet precisely how insolation may have

translated into near-surface temperature changes in this region remains undetermined.

More frequent El Nino events have been linked to periods of lower than average

precipitation and negative mass balances in the sub-tropical Andes (e.g., Vuille and

Keimig, 2004), and have been speculated to drive glacier retreat in the tropical Andes

(Jomelli et al., 2009). If modern ENSO and mass balance relationships were the same in the past, persistent or intense El Nino events should be linked to periods of glacier

ablation/retreat whereas persistent or intense La Nina events should be linked to periods of glacier advance/accumulation. Cobb et al. (2003) used 5180 records from tropical

Pacific corals to determine that the most intense ENSO activity in the central tropical

Pacific during the past millennium occurred during the mid 1600's. However, Moy et al.

(2002) reconstructed steadily decreasing ENSO frequency and intensity in southern

Ecuador from a high 1200 years ago, which signifies low ENSO activity during the LIA, and suggests no LIA ENSO-driven glacier response. Conversely, a study by Gergis and

Fowler (2006) revealed enhanced ENSO activity from 15 sites throughout the Pacific, in

Asia, and Africa in the AD 1800-1900's centuries and peaks in La Nina activity in the

AD 1500-1600's. This latter study suggests La Nina events may be responsible for sustaining glaciers through the LIA. These conflicting proxy records make it difficult to confidently associate ENSO forcing with Vilcabamba glacier retreat at the end of the

LIA, but links between ENSO and glacier activity are plausible.

Another plausible climate driver of glacier advances is a southward displacement of the ITCZ, which is postulated to supply enhanced moisture to the southern tropics

(Lund and Curry, 2006; Lund et al., 2006; Kelly, et al., 2008; Licciardi et al., 2009). On

67 inter-annual to inter-decadal timescales, small temperature shifts have been modeled to correspond to significant shifts in the latitudinal position of the ITCZ (Chiang et al.,

2002). This hypothesis postulates that cooler North Atlantic atmospheric and sea surface temperatures forced a prolonged southward latitudinal shift in the mean position of the

ITCZ. A southward displacement of the ITCZ adequately explains the changes observed in paleoclimate archives in both the Northern Hemisphere and in the tropics during the

LIA (Figure 16) (e.g., Haug et al., 2001; Hodell et al., 2005; Unkel et al., 2007; Sachs et al., 2009). Proxy data from the Florida Straits indicate a reduction in the strength of the

Gulf Stream during the LIA, which may have led to decreased northward transport of heat and a subsequent cooling of the North Atlantic (Lund et al., 2006). North Atlantic cooling is supported by historical accounts and high-resolution tree-ring records which document pronounced cold periods during the 1600's and 1800's (Mann et al., 1999;

Esperer et al., 2002). Cooling in the North Atlantic may have altered cross-equatorial sea-surface temperature (SST) gradients and caused southward migration of the ITCZ

(Koutavas and Lynch-Stieglitz, 2004). Coupled general circulation model (CGCM) studies support the notion that weakened phases of the thermohaline circulation (THC) cause the ITCZ to move to a more southerly position (Chiang, 2004).

Records of arid conditions north of the equator concurrent with wet conditions south of the equator support a southward displacement of the ITCZ during the LIA.

Haug et al. (2001) inferred decreased precipitation and runoff during the LIA off the northern coast of Venezuela (10°42.73'N). Records from the Yucatan Peninsula (20°N) suggest climate became more arid at the onset of the LIA (Hodell et al., 2005). Records from the Florida Straits (24°-25°N) indicate salinity increased during the LIA due in part

68 to aridity caused by southward migration of the ITCZ (Lund and Curry, 2006). From lake records in the Northern Line Islands, Galapagos and Palau, Sachs et al. (2009) document that the Pacific ITCZ was south of its modern position during the LIA. Linsley et al. (1994) document the return to a northern mean position of the ITCZ in Panama

(7°S) following the early AD 1800's.

For the LIA, I speculate that climate forcings involving southward displacement of the ITCZ may explain correlative glaciations in tropical South America and the circum-North Atlantic region, as proposed by Licciardi et al. (2009). I envision a cooling event affecting North Atlantic atmospheric and SSTs either caused by or amplified by a reduction in Gulf Stream transport. These cold conditions may have promoted glaciation in the circum-North Atlantic. A decrease in the strength of the Atlantic meridional overturning circulation would have triggered a southward shift of the ITCZ and supplied enhanced precipitation to drive the advance of tropical Andean glaciers. Subsequent glacier retreat would have occurred when the ocean-atmosphere system returned to pre-

LIA conditions and the ITCZ resumed a more northerly mean position at the end of the

LIA. Alternatively, La Nina activity may have allowed glaciers to persist at least through the 1600's, and/or enhanced ENSO activity at the end of the LIA could have triggered glacier ablation and caused glaciers to retreat from more advanced LIA positions. The postulated ITCZ and ENSO forcing mechanisms imply a dominance of precipitation, rather than temperature, controls on LIA glacier activity. Solar minima and volcanic activity may have acted as second-order controls on glaciation. However, without more detailed atmosphere-ocean climate modeling and without an unambiguous temperature proxy in the tropical Andes, these hypotheses remain conjectural. The tropical glacier

69 chronologies and LIA ELA estimates offered in this study provide critical paleoclimate information that will ultimately enable more accurate modeling of LIA climate drivers. CHAPTER V

CONCLUSIONS

Detailed mapping of glacial deposits in three valleys in the vicinity of Nevado

Salcantay in southern Peru defines large, prominent moraine crests corresponding to the most recent glacier advance. The inner moraines provide evidence for the largest glacier advance since the early Holocene in the Cordillera Vilcabamba (Licciardi et al., 2009).

The most recent advance in the Cordillera Vilcabamba culminated in the late AD 1700's to early AD 1800's according to 10Be surface exposure ages, corresponding to the late

LIA period as defined in northern high latitudes. The 98% quantile lichen results, which are considered the most robust lichenometric index, support 10Be ages as they indicate near-coeval moraine stabilization. Minor differences in Be ages between valleys may suggest a lag in response time to climate perturbations according to glacier size.

The LIA culmination in the Vilcabamba occurred one to two centuries after the

LIA culmination in the nearby Cordilleras Blanca and Real. Offsets in timing may reflect regional climate variations, or uncertainties in lichen and 10Be ages. The pattern of LIA culminations in the Cordillera Vilcabamba is broadly synchronous with North Atlantic and northern Patagonian glacier behavior, but differs slightly from Alaskan and southern

Patagonian records and most drastically from New Zealand records.

The best-fit ELA coefficients in the Cordillera Vilcabamba, 0.40 ± 0.05 (THAR) and 0.60 ±0.10 (AAR), agree well with previous estimates in the tropical Andes. ELA reconstructions indicate the ELA was depressed during the LIA by 160 ± 80 m (AAR) to

71 200 ± 60 m (THAR) relative to modern ELAs at 4920-4930 m asl. In the absence of precipitation changes, temperature depressions around 1.3 ± 0.4°C (AAR) to 1.1 ± 0.5°C

(THAR) would be required to sustain more advanced LIA glacier positions. The early

Holocene ELA was depressed by 35 ± 30 m (AAR) to 110 ± 70 m below the LIA EL A and temperatures would have been 0.2 ± 0.2 (AAR) to 0.7 ± 0.4°C lower, which is relatively modest in comparison to the large and abrupt climate change since the LIA.

This simple-lapse rate approach is a reasonable first-order estimation of climate change during the LIA, but more detailed monitoring of climate in the Cordillera Vilcabamba and more rigorous glacier-climate modeling is necessary to test possible scenarios.

While climate estimates inferred from ELA changes place limits on maximum temperature estimates and do not make precipitation predictions, the preferred scenario here favors precipitation as the dominant climate driver of glaciation during the LIA.

The favored hypothesis envisioned here is a southward displacement of the ITCZ during the LIA, which may have supplied heavy precipitation to southern Peru and sustained glaciers at more advanced positions. Marine and terrestrial records from Central America and the Caribbean support a southerly shift of the ITCZ during the LIA. An alternative climate driver involves potential links with ENSO variability, but conflicting records of

ENSO activity in the tropics during the LIA are difficult to reconcile with the timing of glacier retreat in the Vilcabamba. The results of this study advance knowledge of the

LIA in South America and contribute to an understanding of the global imprint of the

LIA. These findings can be used as valuable boundary conditions in future LIA climate modeling.

72 REFERENCES

Abbott., M.B., A 3500 C yr high-resolution record of water-level changes in Lake Titicaca, Bolivia/Peru. Quaternary Research, v. 47, p. 169-180.

Axford, Y., Geirsdottir, A., Miller, G.H., Langdon, P.G., 2009, Climate of the Little Ice Age and the past 2000 years in northeast Iceland inferred from chironomids and other lake sediment proxies. Journal of Paleolimnology, v. 41, p. 7-24.

Bahr, D.B., Pfeffer, W.T., Sassolas, C, Meier, M.F., 1998, Response time of glaciers as a function of size and mass balance: 1. Theory. Journal of Geophysical Research, v. 103, p. 9777-9782.

Baker, P.A., Seltzer, G.O., Fritz, S.C., Dunbar, R.B., Grove, M.J., Tapia, P.M., Cross, S.L., Rowe, H.D., Broda, J.P., 2001, The history of South American tropical precipitation for the past 25,000 years. Science, v. 291, p. 640-643.

Balco, G., Stone, J.O., Lifton, N.A., Dunai, T.J., 2008, A complete and easily accessible means of calculating surface exposure ages or erosion rates from 10Be and 26A1 measurements. Quaternary Geochronology, v. 3, p. 174-195.

Balco, G., Briner, J., Finkel, R.C., Rayburn, J., Ridge, J.C., Schaefer, J.M., 2009, Regional beryllium-10 production rate calibration for late-glacial northeastern North America. Quaternary Geochronology, v. 4, p. 93-107.

Bard, E., Raisbeck, G., Yiou, F., Jouzel, J., 2000, Solar irradiance during the last 1200 years based on cosmogenic nuclides. Tellus, v. 52B, p. 985-992.

Benedict, J.B., 1967, Recent glacial history of an alpine area in the Colorado Front Range, U.S.A. I. Establishing a lichen-growth curve. Journal of Glaciology, v. 6, p. 817-832.

Benedict, J.B., 1973, Chronology of cirque glaciations, Colorado Front Range. Quaternary Research, v. 3, p. 584-599.

Benedict, J.B., 1990, Experiments on lichen growth. I. Seasonal patterns and environmental controls. Antarctic and Alpine Research, v. 22, p. 244-254.

Benn, D.I., Lehmkuhl, F., 2000, Mass balance and equilibrium-line altitudes of glaciers in high- mountain environments. Quaternary International, v. 65/66, p. 15-29.

Benn, D.I., Owen, LA., Osmaston, H.A., Seltzer, G.O., Porter, S.C., and Mark, B.G., 2005, Reconstruction of equilibrium-line altitudes for tropical and sub-tropical glaciers. Quaternary International, v. 138-139, p. 8-21.

Beschel, R.E., 1950, Lichens as a measure of the age of recent moraines. Arctic and Alpine Research, v. 5, p. 303-309.

73 Beschel, R.E., 1961, Dating rock surfaces by lichen growth and its application to glaciology and physiography (lichenometry). In Raasch, CO., ed., Geology of the Arctic, v. 2, Toronto, University of Toronto Press, p. 1045-1062.

Bowman, L, 1968, The Andes of southern Peru: Greenwood Press, New York, 336 p.

Bradwell, T., Dugmore, A.J., Sugden, D.E., 2006, The Little Ice Age glacier maximum in Iceland and the North Atlantic Oscillation: evidence from Lambatungnajokull, southeast Iceland. Boreas, v. 35, p. 61-80.

Broggi, J.A., 1943, La desglaciacion actual de los Andes del Peru. Boletin de la Sociedad Geologica del Peru, v. XTV/XV, p. 59-90.

Calkin, P.E., Ellis, J.M., 1980, A lichenometric dating curve and its application to Holocene glacier studies in the Central Brooks Range, Alaska. Arctic and Alpine Research, v. 12, p. 245-264.

Cane, M., Clement, A.C., 1999, A role for the tropical Pacific coupled ocean-atmosphere system Milankovitch and millennial timescales. Part II: Global Impacts. In Clark, P.U., Webb, R.S., and Keigwin, L.D., eds., Mechanisms of Global Climate Change at Millenial Timescales, Washington D.C., American Geophysical Union, p. 373-383.

Chepstow-Lusty, A., Frogley, M.R., Bauer, B.S., Bush, B.M., Herrera, A.T., 2003, A late Holocene record of arid events from the Cuzco region, Peru. Journal of Quaternary Science, v. 18,491-502.

Chiang, J.C.H., Kushnir, Y., Giannini, A., 2002, Deconstructing Atlantic Intertropical Convergence Zone variability: Influence of the local cross-equatorial sea surface temperature gradient and remote forcing from the eastern equatorial Pacific. Journal of Geophysical Research, v. 107, no. Dl 4004, p. 1-19.

Chiang, J.C.H., 2004, Present-day climate variability in the tropical Atlantic: A model for paleoclimate changes? In Diaz, H.F., and Bradley, R.S., eds., The Hadley circulation: Present past and future, Dordrecht, The Netherlands, Kluwer Academic Press, p. 465-488. Clapperton, CM., 1972, The Pleistocene moraine stages of west-central Peru. Journal of Glaciology, v. 11, p. 255-263.

Clapperton, CM., 1983, The glaciations of the Andes. Quaternary Science Reviews, v.2, p. 83- 155.

Clapperton, CM., 1986, Glacial geomorphology, Quaternary sequence and paleoclimatic inferences in the Ecuadorian Andes. In Gardiner, V., ed., International Geomorphology 1986, Part II, Chichester, Wiley, p. 843-870.

Clapperton, CM., 1990, Quaternary glaciations in the Southern Hemisphere: An overview. Quaternary Science Reviews, v. 9, p. 299-304.

Cooley, D., Naveau, P., Jomelli, V., Rabatel, A., Grancher, D., 2006, A Bayesian hierarchical extreme value model for lichenometry. Environmetrics, v. 17, p. 555-574.

74 Crowley, T.J., 2000, Causes of climate change over the past 1000 years. Science, v. 289, p. 270- 277.

Desilets, D., Zreda, M., Prabu, T., 2006, Extended scaling factors for in situ cosmogenic nuclides. Earth and Planetary Science Letters, v. 246, p. 265-276.

Douglass, D.C., Singer, B.S., Kaplan, M.R., Ackert, R.P., Mickelson, D.M., Caffee, M.W., 2005, Evidence of early Holocene glacial advances in southern South America from cosmogenic surface-exposure dating. Geology, v. 33, p. 237-240.

Dunai, T.J., 2001, Influence of secular variation of the geomagnetic field on production rates of in situ produced cosmogenic nuclides. Earth and Planetary Science Letters, v. 193, p. 197-212.

Egeler, C.G., de Booy, T., 1961, Preliminary note on the geology of the Cordillera Vilcabamba (SE Peru), with emphasis on the essential pre-Andean origin of the structure. Geologie en Mijnbouw, v. 40e, p. 319-325.

Esperer, J., Cook, E.R., Schweingruber, F.H., 2002, Low-frequency signals in long tree-ring chronologies for reconstructing past temperature variability. Science, v. 295, p. 2250-2253.

Farber, D.L., Hancock, G.S., Finkel, R.C., Rodbell, D.T., 2005, The age and extent of tropical alpine glaciations in the Cordillera Blanca, Peru. Journal of Quaternary Science, v. 20, p. 759- 776.

Favier, V., Wagnon, P., Ribstein, P., 2004, Glaciers of the outer and inner tropics: A different behaviour but a common response to climatic forcing. Geophysical Research Letters, v. 31, L16403,p. 1-5.

Finkel, R.C., Schaefer, J.M., Schwartz, R., 2008, Exposure dating meets history: Precise I0Be dating of very young surfaces. Goldschmidt Conference Abstracts 2008, p. A269.

Fischer H, Werner, M., Wagenbach, D., Schwager, M., Thornsteinnson, T., Wilhelms, F., Kipfstuhl, J., Sommer, S., 1998, Little Ice Age clearly recorded in northern Greenland ice cores. Geophysical Research Lettters, v. 25, p. 1749-1752.

Free, M., Robock, A., 1999, Global warming in the context of the Little Ice Age. Journal of Geophysical Research, v. 104, p. 19057-19070.

Garibotti, I.A., Villalba, R., 2009, Lichenometric dating using Rhizocarpon subgenus Rhizocarpon in the Patagonian Andes, Argentina. Quaternary Research, v. 71, p. 271-283.

Gergis, J.L., Fowler, A.M., 2006, How unusual was the late 20th century El Nino-Southern Oscillation (ENSO)? Assessing evidence from tree-ring, coral, ice-core and documentary palaeoarchives, A.D. 1525-2002. Advances in Geosciences, v. 6, p. 173-179.

Goodman, A.Y., Rodbell, D.T., Seltzer, G.O., Mark, B.G., 2001, Subdivision of glacial deposits in southeastern Peru based on pedogenic development and radiometric ages. Quaternary Research, v. 56, p. 31-50.

Gosse, J.C., Phillips, F.M., 2001, Terrestrial in situ cosmogenic nuclides: theory and application. Quaternary Science Reviews, v. 20, p. 1475-1560.

75 Grove, J., 1988, The Little Ice Age: Routledge, London, 326 p.

Grove, J., 2004, Little Ice Ages Ancient and Modern: Routledge, London, 718 p.

Gutierrez, D., Sifeddine, A., Field, D.B., Ortlieb, L., Vargas, G., Chavez, F., Velazco, F., Ferriera, V., Tapia, P., Salvateci, R., Boucher, H., Morales, M.C., Valdez, J., Reyss, J.-L., Campusano, A., Boussafir, M., Mandeng-Yogo, M., Garcia, M., Baumgartner, T., 2009, Rapid reorganization in ocean biogeochemistry off Peru towards the end of the Little Ice Age. Biogeosciences, v. 6, p. 835-848.

Hall, D.K., Ormsby, J.P., Bindschadler, R.A., Siddalingaiah, H., 1987, Characterization of snow and ice reflectance zones on glaciers using Landsat TM data. Annals of Glaciology, v. 9, p. 104-108.

Hall, D.K., Foster, J.L., Chien, J.Y.L., Riggs, G.A., 1995, Determination of actual snow-covered area using Landsat TM and digital elevation model data in Glacier National Park, Montana. Polar Record, v. 31, p. 191-198.

Haug, G.H., Hughen, K.A., Sigman, D.M., Peterson, L.C., Rohl, U., 2001, Southward migration of the Intertropical Convergence Zone through the Holocene. Science, v. 293, p. 1304-1308.

Hillyer, R., Valencia, B.G., Bush, M.B., Silman, M.R., Steinitz-Kannan, M., 2009, A 24,700-yr paleolimnological history from the Peruvian Andes. Quaternary Research, v. 71, p. 71-82.

Hodell, D.A., Brenner, M., Curtis, J.H., Medina-Gonzalez, R., Can, E.I-C, Albornaz-Pat, A., Guilderson, T.P., 2005, Climate change on the Yucatan Peninsula during the Little Ice Age. Quaternary Research, v. 63, p. 109-121.

Hoffman, G., Ramirez, E., Taupin, J.D., Francou, B., Ribstein, P., Delmas, R., Diirr, H., Gallaire, R., Simoes, J., Schotterer, U., Stievenard, M., Werner, M., 2003, Coherent isotope history of Andean ice cores over the last century. Geophysical Research Letters, v. 30, 1179, p. 1 -4.

Holzhauser, H., Magny, M., Zumbuhl, H.J., 2005, Glacier and lake-level variations in west- central Europe over the last 3500 years. The Holocene, v. 15, p. 789-801.

Howley, M.W., 2008, A Late Glacial and Holocene chronology of the Castner glacier, Delta River Valley, Alaska. M.S. Thesis, University of New Hampshire, 78 p.

Innes, J.L., 1984, The optimal sample size in lichenometric studies. Arctic and Alpine Research, v. 16, p. 233-244.

Innes, J.L., 1985, Lichenometry. Progress in Physical Geography, v. 9, p. 187-254.

Johanneson, T., Raymond, C, Waddington, E., 1989, Time-scale for adjustment of glaciers to changes in mass balance. Journal of Glaciology, v. 35, p. 355-369.

Johnson, A.M., 1976, Climate of Peru, Bolivia, and Ecuador. In Schwerdtfeger, W., ed., World Survey of Climatology, v. 12, New York, El Sevier, p. 147-218.

76 Jomelli, V., Grancher, D., Brunstein, D., Solomina, O., 2008, Recalibration of the yellow Rhizocarpon growth curve in the Cordillera Blanca (Peru) and implications for LIA chronology. Geomorphology, v. 93, p. 202-212.

Jomelli, V., Favier, V., Rabatel, A., Brunstein, D., Hoffman, G., Francou, B., 2009, Fluctuations of glaciers in the tropical Andes over the last millennium and palaeoclimatic implications: A review. Palaeogeography, Palaeoclimatology, Palaeoecology, in press.

Kaplan, M.R., Douglass, D.C., Singer, B.S., Ackert, R.P., Caffee, M.W., 2005, Cosmogenic nuclide chronology of the pre-last glacial maximum moraines at Lago Buenos Aires, 46°S, Argentina. Quaternary Research, v. 63, p. 301-315.

Kaser G., Osmaston, H., 2002, Tropical Glaciers: Cambridge University Press, 207 p.

Kelly, M.A., Lowell, T.V., Schaefer, J.M., 2007, A Chronology of Late-Glacial and Holocene advances of Quelccaya Ice Cap, Peru, based on l0Be and radiocarbon dating. Eos Transactions AGU, v. 88, Fall Meeting Supplementary Abstract PP33B-1277.

Kelly, M.A., Lowell, T.V., Schaefer, J.M., Finkel, R.C., 2008, Late-Glacial and Holocene history of changes in Quelccaya Ice Cap, Peru. Eos Transactions AGU, v. 89, Fall Meeting Supplementary Abstract GC12A-03.

Klein, A.G., Seltzer, G.O., Isacks, B.L., 1999, Modern and last local glacial maximum snowlines in the Central Andes of Peru, Bolivia, and Northern Chile. Quaternary Science Reviews, v. 18, p. 63-84.

Kohl, C.P., Nishiizumi, K., 1992, Chemical isolation of quartz for measurement of in situ produced cosmogenic nuclides. Geochimica et Cosmochimica Acta, v. 52, p. 3583-3587.

Koutavas, A., Lynch-Stieglitz, J., 2004, Variability of the marine ITCZ over the eastern Pacific during the past 30,000 years: Regional perspective and global context. In Diaz, H.F., and Bradley, R.S., eds., The Hadley circulation: Present past and future, Dordrecht, The Netherlands, Kluwer Academic Press, p. 347-369.

Krusic, A.G., Prentice, M.L., Licciardi, J.M., 2009, Climatic implications of reconstructed early- mid Pliocene equilibrium-line altitudes in the McMurco Dry Valleys, Antarctica. Annals of Glaciology, v. 50, p. 31-36.

Lai, D., 1991, Cosmic ray labeling of erosion surfaces: in situ nuclide production rates and erosion models. Earth and Planetary Science Letters, v. 104, p. 424-439.

Lean, J., Beer, J., Bradley, R., 1995, Reconstruction of solar irradiance since 1610: Implications for climate change. Geophysical Research Letters, v. 22, p. 3195-3198.

Li, Y., Cole-Dai, J., Zhou, L., 2009, Glaciochemical evidence in an East Antarctica ice core of a recent (AD 1450-1850) neoglacial episode. Journal of Geophysical Research, v. 114, D08117, p. 1-11.

Licciardi, J.M., Schaefer, J.M., Lund, D.C., 2007, Cosmogenic 10Be dating of early and latest Holocene moraines on Nevado Salcantay in the southern Peruvian Andes. Abstract, AGU 2007 Fall Meeting.

77 Licciardi, J.M., Schaefer, J.M., Lund, D.C., Taggart, J.R., 2008, Precise surface exposure dating of early Holocene and Little Ice Age moraines in the Cordillera Vilcabamba of southern Peru. Eos Transactions AGU, v. 89, Fall Meeting Supplementary Abstract GC21A-0720.

Licciardi, J.M., Schaefer, J.M., Taggart, J.R., Lund, D.C., 2009, Holocene glacier fluctuations in the Peruvian Andes indicate northern climate linkages. Science, in press.

Licciardi, J.M., Pierce, K.L., 2008, Cosmogenic exposure-age chronologies of Pinedale and Bull Lake glaciations in greater Yellowstone and the Teton Range, USA. Quaternary Science Reviews, v. 27, p. 814-831.

Lifton, N.A., Bieber, J.W., Clem, J.M., Duldig, MX., Evenson, P., Humble, J.E., Pyle, R., 2005, Addressing solar modulation and long-term uncertainties in scaling secondary cosmic rays for in situ cosmogenic nuclide applications. Earth and Planetary Science Letters, v. 239, p. 140- 161.

Lifton, N.A., 2009, personal communication, May 2009.

Linsley, B.K., Dunbar, R.B., 1994, A coral-based reconstruction of the Intertropical Convergence Zone variability over Central America since 1707. Journal of Geophysical Research, v. 99, p. 9977-9994.

Lowell, T.V., Schoenenberger, K., Deddens, J.A., Denton, G.H., Smith, C, Black, J., Hendy, C.H., 2005, Rhizocarpon calibration curve for the Aoriaki/Mount Cook area of New Zealand. Journal of Quaternary Science, v. 20, p. 313-325.

Luckman, B.H., 2000, The Little Ice Age in the Canadian Rockies. Geomorphology, v. 32, p. 357-384.

Lund, D.C., Lynch-Stieglitz, J., Curry, W.B., 2006, Gulf Stream density structure and transport during the past millennium. Nature, v. 444, p. 601-604.

Lund, D.C., Curry, W., 2006, Florida Current surface temperature and salinity variability during the last millennium. Paleoceanography, v. 21, PA2009, p. 1-15.

Luoto, T.P., Nevalainen, L., Sarmaja-Korjonen, K., 2008, Multiproxy evidence for the 'Little Ice Age' from Lake Hamptrask, Southern Finland. Journal of Paleolimnology, v. 40, p. 1097- 1113.

Mann, M.E., Bradley, R.S., Hughes, M.K., 1999, Northern Hemisphere temperatures during the past millennium: Inferences, uncertainties, and limitations. Geophysical Research Letters, v. 26, p. 759-762.

Mark, B.G., Seltzer, G.O., Rodbell, D.T., Goodman, A.Y., 2002, Rates of deglaciation during the Last Glaciation and Holocene in the Cordillera Vilcanota-Quelccaya Ice Cap region, southeaster Peru. Quaternary Research, v. 57, p. 287-298.

Marocco, R., 1978, Geological study of Vilcabamba range. INGEOMIN, Bulletin 4, Series D, Lima, Peru, 157 p.

78 Matthes, F.E., 1939, Report of committee on glaciers, April 1939. Transactions, American Geophysical Union, p. 518-522.

Matthews, J.A., 1974, Families of lichenometric dating curves from the Storbreen gletschervorfeld, Jotunheimen, Norway. Norsk Geografisk Tidsskrift, v. 28, p. 215-235.

McCarroll, D., 1993, Modeling late-Holocene snow-avalanche activity: incorporating a new approach to lichenometry. Earth Surface Processes and Landforms, v. 18, p. 527-539.

McKay, N.P., Kaufman, D.S., 2009, Holocene climate and glacier variability at Hallet and Greyling Lakes, Chugach Mountains, south-central Alaska. Journal of Paleolimnology, v. 41, p. 143-159.

Meierding, T.C., 1982, Late Pleistocene glacial equilibrium-line altitudes in the Colorado Front Range: A comparison of methods. Quaternary Research v. 18, p. 289-310.

Mercer, J.H., Palacios, O.M., 1977, Radiocarbon dating of the last glaciations in Peru. Geology, v. 5, p. 600-604.

Mercer, J.H., 1984, Late Cainozoic glacial variation in South America south of the Equator. In Vogel, J.C., ed., Late Cainozoic Paleoclimates of the Southern Hemisphere, A.A. Balkema, Rotterdam, p. 45-58.

Morales-Arno, B., 1999, Glaciers of South America - Glaciers of Peru. In Williams, R.S., Jr., and Ferrigno, J.G., eds., Satellite Image Atlas of Glaciers of the World, U.S. Geological Survey Professional Paper 1386-1, online version 1.02.

Moy, CM., Seltzer, G.O., Rodbell, D.T., Anderson, D.M., 2002, Variability of El Nino/Southern Oscillation activity at millennial timescales during the Holocene epoch. Nature, v. 420, p. 162-165.

Naveau, P., Jomelli, V., Cooley, D., Grancher, D., Rabatel., A., 2007, Modeling uncertainties in lichenometry studies. Arctic, Antarctic, and Alpine Research, v. 39, p. 277-285.

Nesje, A., Dahl, S.O., 2003, The 'Little Ice Age' - only temperature? The Holocene, v. 13, p. 139-145.

Nesje, A., Bakke, J., Dahl, S.O., Lie, 0, Matthews, J.A., 2008, Norwegian mountain glaciers in the past, present and future. Global and Planetary Change, v. 60, p. 10-27.

Oerlemans, J., 2005, Extracting a climate signal from 169 glacier records. Science, v. 308, p. 675-677.

Osmaston, H.A., 1975, Models for the estimation of firnlines of present and Pleistocene glaciers. In Peel, R.F., Chisholm, M.D.I., Hagget, P., eds., Processing Physical and Human Geography. Bristol Essays, p. 218-245.

Paul, F., Kaab, A., Maisch, M., Kellenberger, T., Haeberli, W., 2002, The new remote-sensing- derived Swiss glacier inventory: I. Methods. Annals of Glaciology, v. 34., p. 355-361.

79 Polissar, P.J., Abbott, M.B., Wolfe, A.P., Bezada, M., Bradley, R.S., 2006, Solar modulation of Little Ice Age climate in the tropical Andes. PNAS, v. 103, p. 8937-8942.

Porter, S.C., 2001, Snowline depressions in the tropics during the Last Glaciation. Quaternary Science Reviews, v. 20, p. 1067-1091.

Rabatel, A., Jomelli, V., Naveau, P., Francou, B., Grancher D., 2005, Dating of Little Ice Age glacier fluctuations in the tropical Andes: Charquini glaciers, Bolivia, 16°S. C.R. Geoscience, v. 337, p. 1311-1322.

Rabatel, A., Macacha, A., Francou, B., Jomelli, V., 2006, Glacier recession on Cerro Charquini (16°S), Bolivia, since the maximum of the Little Ice Age (17th century). Journal of Glaciology, v. 52, p. 110-118.

Rabatel, A., Francou, B., Jomelli, V., Naveau, P., Grancher, D., 2008, A chronology of the Little Ice Age in the tropical Andes of Bolivia (16°S) and its implications for climate reconstruction. Quaternary Research, v. 70, p. 198-212.

Ramage, J.M., Smith, J.A., Rodbell, D.T., Seltzer, G.O., 2005, Comparing reconstructed Pleistocene equilibrium-line altitudes in the tropical Andes of central Peru. Journal of Quaternary Science, v. 20, p. 777-788.

Ramirez, E., Hoffman, G., Taupin, J.D., Francou, B., Ribstein, P., Caillon, N., Ferron, F.A., Landais, A., Petit, J.R., Pouyaud, B., Schotterer, U., Simeos, J.C., Stievenard, M., 2003, A new Andean deep ice core from Nevado Illimani (6350 m), Bolivia. Earth and Planetary Science Letters, v. 212, p. 337-350.

Richards, B.W.M., Benn, D.I., Owen, L.A., Rhodes, E.J., Spencer, J.Q., 2000, Timing of late Quaternary glaciations south of Mount Everest in the Khumbu Himal, Nepal. GSA Bulletin, v. 112, p. 1621-1632.

Rodbell, D.T., 1992b, Late Pleistocene equilibrium-line reconstructions in the northern Peruvian Andes. Boreas, v. 21, p. 43-52.

Rodbell, D.T., 1992a, Lichenometric and radiocarbon dating of Holocene glaciations, Cordillera Blanca, Peru. The Holocene, v. 2, p. 19-29.

Rodbell, D.T., Smith, J.A., Mark, B.G., 2009, Glaciation in the Andes during the Lateglacial and Holocene. Quaternary Science Reviews, doi:10.1016/j.quascirev.2009.03.012.

Sachs, J.P., Sachse, D., Smittenberg, R.H., Zhang, Z., Battisti, D.S., Golubic, S., 2009, Southward movement of the Pacific intertropical convergence zone AD 1400-1850. Nature Geoscience, v. 2, p. 519-525.

Schaefer, J.M., Denton, G.H., Kaplan, M., Putnam, A., Finkel, R.C., Barrell, D.J.A., Andersen, B.G., Schwartz, R., Mackintosh, A., Chinn, T., Schluchter, C, 2009, High-frequency Holocene glacier fluctuations in New Zealand differ from the Northern signature. Science, v. 324, p. 622-625.

Seltzer, G.O., 1992, Late Quaternary glaciations of the Cordillera Real, Bolivia. Journal of Quaternary Science, v. 7, p. 87-98.

80 Seltzer, CO., 1994, Climatic interpretation of alpine snowline variations on millennial time scales. Quaternary Research, v. 41, p. 154-159.

Servicio Nacional de Meteorologia e hidrologia del Peru, 2008, Clima Data Historicos Direcciones regionales, accessed 3 August 2009, www.senamhi.gob.pe/?p=200.

Sikorski, J.J., Kaufman, D.S., Manley, W.F., Nolan, M., 2009, Glacial-geologic evidence for decreased precipitation during the Little Ice Age in the Brooks Range, Alaska. Arctic, Antarctic, and Alpine Research, v. 41, p. 138-150.

Sidjak, R.W., Wheate, R.D., 1999, Glacier mapping of the Illecillewaet icefield, British Columbia, Canada, using Landsat TM and digital elevation data. International Journal of Remote Sensing, v. 20, p. 273-284.

Silva, S.L., Zielinsky, G.A., 1998, Global influence of the AD 1600 eruption of Huaynaputina, Peru. Nature, v. 393, p. 455-458.

Smith, J.A., Seltzer, G.O., Farber, D.L., Rodbell, D.T., Finkel, R.C., 2005, Early local Last Glacial Maximum in the tropical Andes. Science, v. 308, p. 678-681.

Smith, J.A., Mark, B.G., Rodbell, D.T., 2008, The timing and magnitude of mountain glaciations in the tropical Andes. Journal of Quaternary Science, v. 23, p. 609-634.

Solomina, O., Jomelli, V., Kaser, G., Ames, A., Berger, B., Pouyaud, P., 2007, Lichenometry in the Cordillera Blanca, Peru: "Little Ice Age" moraine chronology. Global and Planetary Change, v. 59, p. 225-235.

Stone, J., 2000, Air pressure and cosmogenic isotope production. Journal of Geophysical Research, v. 105, p. 23753-23759.

Stuiver, M., Reimer, P.J., Reimer, R.W., 2005, CALIB 5.0: WWW program and documentation.

Thomas, E.K., Briner, J.P., 2009, Climate of the past millennium inferred from varved proglacial lake sediments on northeast Baffin Island, Arctic Canada. Journal of Paleolimnology, v. 41, p. 209-224.

Thompson, L.G., Mosley-Thompson, E., Bolzan, J.F., Koci, B.R., 1985, A 1500-year record of tropical precipitation in ice cores from the Quelccaya ice cap, Peru. Science, v. 229, p. 971- 973.

Thompson, L.G., Mosley-Thompson, E., Dansgaard, W., Grootes, P.M., 1986, The Little Ice Age as recorded in the stratigraphy of the tropical Quelccaya ice cap. Science, v. 234, p. 361-364.

Thompson, L.G., Mosley-Thompson, E., Davis, M.E., Lin, P.-N., Henderson, K.A., Cole-Dai, J., Bolzan, J.F., Liu, K.-b., 1995, Late Glacial stage and Holocene tropical ice core records from Huascaran, Peru. Science, v. 269, p. 46-50.

Thompson, L.G., Davis, M.E., Mosley-Thompson, E., Sowers, T.A., Henderson, K.A., Zagorodnov, V.S., Lin, P.-N., Mikhalenko, V.N., Campen, R.K., Bolzan, J.F., Cole-Dai, J.,

81 Francou, B., 1998, A 25,000-year tropical climate history from Bolivian Ice Cores. Science, v. 282, p. 1858-1864.

Unkel, I., Kadereit, A., Machtle, B., Eitel, B., Kromer, B., Wagner, G., Wacker, L., 2007, Dating methods and geomorphic evidence of palaeoenvironmental changes at the eastern margin of the Southern Peruvian coastal desert (14°30'S) before and during the Little Ice Age. Quaternary International., v. 175, p. 3-28.

Vilimek, V., Klimes, J., Vlcko, J., Carreno, R., 2006, Catastrophic debris flows near Machu Picchu village (Aguas Calientes), Peru. Environmental Geology, v. 50, p. 1041-1052.

Villalba, R., 1994, Tree-ring and glacial evidence for the Medieval Warm Epoch and the Little Ice Age in southern South America. Climatic Change, v. 26, p. 183-197.

Vuille, M., Keimig, F., 2004, Interannual variability of summertime convective cloudiness and precipitation in the Central Andes derived from ISCCP-B3 data. Journal of Climate, v. 17, p. 3334-3348.

Vuille, M., Francou, B., Wagnon, P., Juen, I., Kaser, G., Mark, B.G., Bradley, R.S., 2008, Climate change and tropical Andean glaciers: Past, present, and future. Earth-Science Reviews, v. 89, p. 79-96.

Vuille, M., Kaser, G., Juen, I., 2008, Glacier mass balance variability in the Cordillera Blanca, Peru and its relationship with climate and the large-scale circulation. Global and Planetary Change, v. 62, p. 14-28.

Wagnon P., Ribstein, P., Kaser, G., Berton, P., 1999, Energy balance and runoff seasonality of a Bolivian Glacier. Global and Planetary Change, v. 22, p. 49-58.

Wiles, G.C., Barclay, D.J., Calkin, P.E., Lowell, T.V., 2008, Century to millennial-scale temperature variations for the last two thousand years indicated from glacial geological records of Southern Alaska. Global and Planetary Change, v. 60, p. 115-125.

Williams, R.S., Hall, D.K., Benson, C.S., 1991, Analysis of glacier facies using satellite techniques. Journal of Glaciology, v. 37, p. 120-128.

Wright, Jr., H.E., 1984, Late Glacial and late Holocene moraines in the Cerros Cuchpanga, Central Peru. Quaternary Research, v. 21, p. 275-285.

Zech, R., Kull, Ch., Kubik, P.W., Veit, H., 2007, Exposure dating of the Late Glacial and pre- LGM moraines in the Cordon de Dona Rosa, Northern/Central Chile (~31°S). Climate of the Past, v. 3, p. 1-14.

Zhang, P., Cheng, H., Edwards, R.L., Chen, F., Wang, Y., Yang, X., Liu., J., Tan, M., Wang, X., Liu., J., An, C, Dai, Z., Zhou, J., Zhang, D., Jia, J., Jin, L., Johnson, K., 2008, A test of climate, sun, and culture relationships from an 1810-year Chinese cave record. Science, v. 322, p. 940-942.

82 APPENDICES

83 APPENDIX A

LICHEN MEASUREMENTS

Table Al. Lichen measurement data from Tucarhuay, Rio Blanco, and Sisaypampa inner moraines in the Cordillera Vilcabamba.

Tucarhuay Rio Blanco Sisaypampa

Long Short Long Short Long Short No. axis axis No. axis axis No. axis axis (mm) (mm) (mm) (mm) (mm) (mm)

1 28 15 1 32 30 1 30 25 2 22 14 2 30 28 2 27 24 3 20 13 3 29 27 3 22 20 4 20 19 4 26 18 4 22 21 5 20 18 5 26 25 5 20 19 6 19 18 6 26 25 6 20 20 7 19 19 7 25 21 7 20 19 8 19 18 8 24 23 8 20 18 9 18 16 9 22 19 9 19 15 10 18 15 10 21 15 10 19 19 11 18 17 11 21 19 11 18 16 12 18 16 12 21 20 12 18 17 13 18 15 13 21 15 13 18 16 14 17 15 14 21 20 14 18 15 15 16 16 15 20 20 15 17 14 16 16 16 16 20 20 16 17 15 17 16 15 17 20 20 17 16 15 18 16 16 18 20 17 18 16 15 19 15 10 19 20 18 19 16 14 20 15 13 20 19 16 20 16 15 21 15 13 21 19 18 21 16 13 22 15 13 22 19 18 22 16 14 23 15 15 23 19 19 23 15 12 24 15 14 24 19 18 24 15 11 25 15 12 25 19 16 25 15 15

84 26 15 13 26 19 13 26 15 12 27 15 13 27 19 16 27 15 8 28 15 15 28 19 16 28 15 11 29 15 15 29 19 17 29 15 14 30 15 11 30 19 17 30 15 14 31 15 15 31 19 17 31 15 15 32 14 8 32 18 15 32 14 13 33 14 13 33 18 14 33 14 10 34 14 10 34 18 18 34 14 13 35 14 12 35 18 16 35 14 14 36 14 14 36 18 17 36 14 13 37 14 12 37 18 18 37 14 14 38 14 12 38 18 16 38 14 13 39 14 11 39 18 17 39 14 12 40 13 11 40 18 15 40 14 12 41 13 9 41 18 14 41 14 11 42 13 9 42 18 18 42 14 13 43 13 10 43 17 16 43 14 12 44 13 11 44 17 16 44 13 8 45 13 11 45 17 16 45 13 12 46 13 12 46 17 15 46 13 11 47 13 13 47 17 16 47 13 10 48 13 9 48 17 16 48 13 12 49 13 12 49 17 16 49 13 12 50 13 10 50 17 16 50 13 13 51 13 12' 51 17 15 51 13 12 52 12 10 52 17 16 52 13 11 53 12 11 53 17 15 53 13 12 54 12 10 54 17 17 54 13 12 55 12 11 55 17 16 55 13 13 56 12 11 56 16 14 56 13 11 57 12 11 57 16 14 57 13 12 58 12 12 58 16 14 58 13 12 59 12 10 59 16 15 59 13 13 60 11 8 60 16 15 60 13 12 61 11 8 61 16 15 61 12 8 62 11 9 62 16 15 62 12 9 63 11 11 63 16 14 63 12 11 64 11 11 64 16 15 64 12 10 65 11 11 65 16 16 65 12 11 66 11 10 66 16 15 66 12 12 oo

NN(SN(SNNN(N(N(SN O O O O aao\O\uo>0\Mhioin vo^^r^t^t^t^r~t^t^r^t--~r--ooooooooooooooooooooCT\ONO\0\0\C\ONONO\0\^-<

|fl«lflm^m1•^lfl^m'tnnvlK^'t•1•1•n1•H^tnnt1•t1•vl01•'^(^lmNMtflm1•

r^oo<^o^r^r^^>ovot-~ooo\o^cNroTj-u-ivor--ooa\O^H(Nr«-)-^-w-)vot---ooa\oooooooo vovo^ot^t^r~t^t^r^c^t^t^r^c»ooc»ooooooooooc»oo^^^o\ONONO\ONONO\^'^^^-H-H^H^-i^H

,—i O ^H O OOOO

^^H^OOOOOOOOOOOOO CTSOONOCTNOOOOOOOOOOO

vovDvor-~c~-r^r-r-c^c~~r-r^r^ooooooooooooooooooooCT\CT\CTNa\ 108 14 12 109 14 11 110 14 14 111 14 14 112 14 12 113 14 14 114 14 13 115 14 14 116 14 12 117 14 12 118 14 12 119 14 14 120 14 13 121 14 13 122 14 13 123 14 12 124 14 12 125 14 13 126 14 13 127 14 13 128 14 12 129 14 13 130 14 12 131 14 14 132 14 12 133 14 14 134 14 13 135 14 13 136 14 13 137 14 13 138 14 11 139 14 13 140 14 13 141 14 14 142 13 11 143 13 11 144 13 12 145 13 12 146 13 12 147 13 12 148 13 11 149 13 12 150 13 12 151 13 9 152 13 12 153 13 12 154 13 11 155 13 12 156 13 12 157 13 11 158 13 12 159 13 12 160 13 11 161 13 13 162 13 11 163 13 12 164 13 12 165 13 12 166 13 12 167 13 12 168 13 13 169 13 11 170 13 12 171 13 12 172 13 12 173 13 12 174 13 12 175 12 12 176 12 12 177 12 11 178 12 10 179 12 11 180 12 11 181 12 11 182 12 10 183 1? 10 184 12 12 185 12 10 186 12 12 187 12 10 188 12 12 189 12 11 190 12 10 191 12 11 192 12 11 193 12 11 194 12 12 195 12 12 196 12 11 197 12 11 198 12 10 199 12 10 200 12 11 201 12 10 202 12 12 203 12 10 204 12 12 205 12 9 206 12 11 207 12 12 208 12 12 209 12 10 210 11 10 211 11 11 212 11 9 213 11 10 214 11 10 215 11 9 216 11 10 217 11 11 218 11 10 219 11 10 220 11 10 221 11 10 222 11 8 223 11 11 224 11 11 225 11 10 226 11 10 227 11 10 228 11 10 229 11 11 230 11 11 231 11 10 232 11 10 233 10 9 234 10 10 235 10 10 236 10 8 237 10 10 238 10 10 239 10 9 240 10 10 241 10 9 242 10 9 243 10 9 244 10 9 245 10 10 246 10 10 247 10 9 248 10 10 249 10 10 250 10 9 251 10 10 252 10 9 253 10 9 254 10 10 255 10 10 256 10 9 257 10 9 258 10 10 259 10 10 260 10 10 261 10 9 262 10 10 263 10 12 264 10 9 265 10 10 266 10 9 267 10 10 268 10 10 269 10 10 270 10 9 271 9 9 272 9 9 273 9 8 274 9 8 275 9 9 276 9 9 277 9 9 278 9 9 279 9 7 280 9 8 281 8 7 282 8 8 283 8 7 284 8 7 285 8 7 286 8 8 287 8 7 288 8 7 289 8 7 290 8 7 291 8 8 292 8 8 293 7 6 294 7 7 295 7 7 296 7 6 297 7 6 298 6 5 299 6 5 300 4 4

91 APPENDIX B

SAMPLE SITES AND DESCRIPTIONS

92 Sample name: PE08-2 Location: S 13.38113 W 72.58591 Elevation: 4307 m 3-D Diff: ±10m Boulder height: 300 cm max; 160 cm min Lithology: Metasedimentary with quartz veins Strike/dip: 0°/0° Description: Large angular boulder with shelf-like surface due to foliation of the rock. Appears intact. Shielding data: (azimuths) 24, 80, 114, 135, 153, 185, 226, 252, 296, 340 (elevations) 31,11,15,2,2,17,32,18,20,34

Shielding results: 90

180 360 Azimuth

Photos:

93 Sample name: PE08-3 Location: S 13.38086 W 72.58605 Elevation: 4306 m 3-D Diff: ±20m Boulder height: 240 cm max; 30 cm min; 200 cm to sample Lithology: Metasedimentary with quartz veins Strike/dip: 0°/0° Description: Large boulder. Quartz vein on highest end of boulder sampled. Shielding data: (azimuths) 24, 80, 114, 135, 153, 185, 226, 252, 296, 340 (elevations) 31,11,15,2,2,17,32,18,20,34

Shielding results:

'•r---""""""-iL 80 120 180 240 300 360 Azimuth

Photos:

- **• «N4:M

;**-> % -

94 Sample name: PE08-4 Location: S 13.34172 W 72.51912 Elevation: 4506 m 3-D Diff: ± 16m Boulder height: 180-210 cm Lithology: Granodiorite Strike/dip: 287345° Description: Large, angular boulder 3 m off moraine crest. Some mineral etching, no polish; sample surface is from highest structural level and probably original. 180 cm to sample spot. Shielding data: (azimuths) 13, 66, 122, 155, 166, 213, 254, 219, 319, 358 (elevations) 20,4,5,12,9,25,19,32,26,38

Shielding results: 90

'i >••' ' " • i" 120 180 240 360 Atfmuth

Photos:

K>

:$•. '- 'A } . \ f

^m?\i •f\ '.ft* *«WH* %

95 Sample name: PE08-5 Latitude: S 13.34268 W 72.52505 Elevation: 4399 m 3-D Diff: ±8.5m Boulder height: 130 cm max; 100 cm to sample Lithology: Granodiorite Strike/dip: 0°/0° Description: Large, subrounded boulder right on moraine crest. Appears stable. Surface is reasonably smooth below lichen cover. Some mineral etching. Shielding data: (azimuths) 40, 72, 84, 102, 165, 185, 207, 240, 262, 294, 313 (elevations) 16,5,4,3,17,9,10,26,20,33,35

Shielding results: 90

120 180 360 Photos: Azimuth

96 Sample name: PE08-6 Location: S 13.34278 W 72.51457 Elevation: 4379 m 3-D Diff: ± 12.6 m Boulder height: 95 cm max; 63 cm to sample Lithology: Granodiorite Strike/dip: 90°/20° Description: Large, flat-topped boulder on a secondary ridge just inside moraine crest; some possibility of boulder rotation. Surface is very smooth below and comes off in flakes. Shielding data: (azimuths) 40, 72, 84, 102, 165, 185, 207, 240, 262, 294, 313 (elevations) 16,5,4,3,17,9,10,26,20,33,35

Shielding results: 90

180 360 Photos: Azimuth

97 APPENDIX C

GOOGLE EARTH ELEVATION CALIBRATION

Table A2. Calibration of Google Earth elevations as compared with GPS elevations in the Cordillera Vilcabamba. °

Waypoint GPS elev. Google Earth A Elevation Difference (m) (m) (m) %

PE06-1 4110 4080 30 0.7 PE06-2 4087 4057 30 0.7 PE06-3 4082 4059 23 0.6 PE06-4 4081 4050 31 0.8 PE06-6 4064 4034 30 0.7 PE06-7 4047 4006 41 1.0 PE06-8 4386 4357 29 0.7 PE06-9 4371 4322 49 1.1 PE06-10 4534 4472 62 1.4 PE06-11 4516 4456 60 1.3 PE06-12 4332 4284 48 1.1 PE06-13 4369 4320 49 1.1 PE06-14 4400 4386 14 0.3 TEA 4400 4368 32 0.7 TI2B 4394 4361 33 0.8 TI2C 4382 4351 31 0.7 TOD 4373 4333 40 0.9 THE 4361 4304 57 1.3 TI2F 4342 4283 59 1.4 TI2G 4327 4270 57 1.3 TI2H 4322 4262 60 1.4 TI2I 4318 4261 57 1.3 TI2J 4295 4258 37 0.9 TI2K 4287 4350 -63 -1.5 TI2L 4269 4344 -75 -1.8 TI2M 4289 4240 49 1.1 TON 4301 4257 44 1.0 TOO 4307 4263 44 1.0 TOP 4305 4271 34 0.8 TOQ 4313 4271 42 1.0 TOR 4321 4272 49 1.1 TM2A 4156 4151 5 0.1 TM2B 4164 4141 23 0.6 TM2C 4142 4121 21 0.5 TM3A 4171 4143 28 0.7 TM3B 4156 4132 24 0.6 TM3C 4145 4118 27 0.7 TUCO-1 4033 4020 13 0.3 TUCO-2 4084 4044 40 1.0 TUCO-3 4094 4060 34 0.8 TUCO-4 4114 4078 36 0.9 TUCO-5 4158 4124 34 0.8 TUCO-6 4144 4106 38 0.9 TUCO-7 4143 4094 49 1.2 TUCO-8 4143 4084 59 1.4 TUCO-9 4106 4065 41 1.0 TUCO-10 4082 4046 36 0.9 TUCO-11 4020 4010 10 0.2 SEI1 4550 4494 56 1.2 SEI2 4516 4470 46 1.0 SEI3 4499 4441 58 1.3 SEI4 4461 4411 50 1.1 SEI5 4422 4377 45 1.0 SEI6 4407 4351 56 1.3 SEI7 4383 4335 48 1.1 SEA1 4295 4301 -6 -0.1 SEA2 4271 4279 -8 -0.2 SEA3 4290 4263 27 0.6 SEA4 4287 4250 37 0.9 SEA5 4257 4234 23 0.5 SEA6 4242 4222 20 0.5 SEA7 4239 4207 32 0.8 SEA8 4222 4197 25 0.6 SEB1 4245 4228 17 0.4 SEB2 4263 4229 34 0.8 SEB3 4260 4242 18 0.4 SEB4 4266 4244 22 0.5 SEB5 4263 4246 17 0.4 SEB6 4275 4250 25 0.6 SEB7 4276 4254 22 0.5 SEB8 4284 4260 24 0.6 SEB9 4286 4271 15 0.3 SEC1 4299 4285 14 0.3 SEC2 4295 4275 20 0.5 SEC3 4290 4271 19 0.4 (/2C/3aiC/5ai!/3&O(/30OC/3C/5!Z)C/3C«l/3t/30OC/2C/3 dddddddod d d „ n.. o o o o K> K> K> K) K> 00 -J Q\ ivi 4^ m o o a > W5 >•

4^ *. 4^ *. 4^ 4^4^*>.4^4i.4^4^4^4i.4^4^4^4^4^

HHWWWU^WMWMMW WO\ 0\ ^ ^J 00 000\00<>>(JtO>00U>

4^4^4i.4^4i.4^4^4^4^4i.4^4i.4^4^4i.4^4^4^4^ OOOOVOOOOOO^O^OOOVOVOOOilJiLrtUiON •--UJV04i.^J00ON>>—'0000~-JtyiO\000000>—•

>-' K> K> ^JU>V0V000ts>000J00O'O~JU»UlO vo Ui O K>

pooooppopppppopppoo O i-» bbo«^b\l«!obo!oviNib\l>ib\'sj(o!til/iUi o o APPENDIX D

GEOMORPHIC PHOTOGRAPHS

Tucarhuay Valley

Figure Ala. View of the western segment of the Tucarhuay outer moraine and the lower outwash plain (photo looking south from the outer moraine crest). 4^"'1HfflH t \, :M • _ — ~ LT ,.-»-^^* ^ T?'5~ mH

Figure Alb. Boulder on eastern segment of Tucarhuay outer moraine. Central and eastern loops of inner moraines also visible (photo looking north from outer moraine). ^yfcm

••*** * •'; \ ••f-':. S*ciS i.m?

'*#T:*

Figure Ale. View of the central inner Tucarhuay moraine and the two dammed glacial lakes outer (photo looking southwest from the inner moraine crest). »#"-i( /Vx,/ ^ * 1 .•* , r "s "J» 4. m*. «fE

«&6 i? »!*, ,\

Figunir© AldL View of the western inner Tucarhuay moraine and alluvial fan emanating from the breach in the moraine (photo looking northwest from the inner moraine crest).

103 Rio Blanco Valley

Flgw© A2a. View of the outer and inner Rio Blanco moraines (photo looking north toward Salcantay). * •&•* .'.'$

:•*•**.

"'^-"•'•*^»k^.

Figure A2b. Outwash emanating from inner moraine which culminates at bog and outer moraine crests. Lateral outer moraine ridges visible on the left (photo looking south). •"$tS£ M ™

IEP^ y - « •<*? } j^ji •*»,,JB > 'i*^r| **L -v • ^ :% AMFS u W§!i Ifs fe*.

•as,*8*.*;

-'i^-fcsr-.' •^«P-

Figure A2c. View of the upper portions of the Rio Blanco inner moraine (photo from a ridge, looking west).

105 Sisaypampa Valley

V 5 \

Figure A3a. View of Sisaypampa outer moraine ridges, and kame terrace (photo from the inner moraine crest, looking southwest).

106 "^

St _ I *>^ rr •< <"< _* -* *s«^ r J^*>. "31 *.**£' " -'• vi fife**. 'wSSr^C" **" ~^ *"•

Fngmre A3Ibo Nested ridges and kame terrace (right), and northern outer ridges (left). Qutwash, inner crest, breach also visible (photo from inner moraine crest, looking west). 'Tl

•^

#• f > «f *-

•:°- i?J:-y-*' , •% --.s-.^. °x-

*-.*

&Hs&

.%»-, Mgunre A3e. View of the inner moraine and the small ridges on the inner moraine The innermost outer ridges are also visible (photo from the outer moraine ing northwest).

107 Figwe A3dL View of outer moraine, nested ridges, and inner moraine. Meltwater stream from the western tributary valley cuts through outer moraine and outermost nested ridges (photo from the outer moraine crest, looking northwest).

108 APPENDIX E

10BE SAMPLE PREPARATION AND PRODUCTION RATE CALIBRATION

Sample preparation

Rock samples were cut to appropriate thicknesses (1.5-2.0 cm) with a wet saw, crushed with a large jaw crusher and disk pulverizer, and dry sieved to 600-250 /xm.

Strongly magnetic minerals were removed with a hand magnet prior to acid leaching.

Crushed samples were sonicated ~70°C in a 5% HC1 / 0.03% H2O2 solution, then in 5%

HNO3 for pre-cleaning, to dissolve any lichen, organics, iron oxides, and carbonates.

Quartz was isolated by leaching samples 4—6 times in a 2% HF /1 % HNO3 solution at

~70°C (Kohl and Nishiizumi, 1992). After the final HF / HNO3 leach, remaining magnetic minerals were removed by one last pass with a hand magnet. 9Be carrier (~0.2 mg) was added to purified quartz samples before dissolution in concentrated HF and conversion to chloride form. Ion-exchange chromatography was conducted with anion and cation exchange columns to isolate Be from other ions. Ti was removed and

Be(OH)2 was selectively precipitated before oxidation to BeO at 1100°C. Samples were packed into cathodes with ~5-6 mg niobium powder in order to maximize ion current.

Reference production rate

For age calculations a reference 10Be production rate was established from a high- altitude calibration site in central Peru (9.65°S) at elevations (4045 m) comparable to the sites in this study, reported by Farber et al. (2005). Farber et al. (2005) reported I0Be concentrations from seven boulders, five of which have scattered Be concentrations.

The oldest two boulders, with 10Be concentrations which agree within \a uncertainty, were used for the calibration because they are least likely to have been affected by exhumation and/or erosion. These two boulders (duplicate pair PE98RU-43a, 43b; and

PE98RU-C44) were independently dated by tightly bracketing radiocarbon ages between

11.3 and 11.0 14C ka by Rodbell et al. (2000) and recalibrated here with CALIB 5.0.2 to

13.0 ± 0.1 cal ka (Stuiver et al., 2005).

110