A SURFACE WATER δ18O BASELINE FOR ARCHAEOLOGICAL STUDIES OF SEASONALITY AND 

MOBILITY IN THE MAJES VALLEY AND PUCUNCHO BASIN OF SOUTHERN PERU 

 
By 
 

Emily Beatrice Peterson Milton 

 
 
 
 
 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

 

 

2020

A THESIS  

Submitted to 

Michigan State University 

in partial fulfillment of the requirements 

for the degree of 

Anthropology–Master of Arts 

 

 

A SURFACE WATER δ18O BASELINE FOR ARCHAEOLOGICAL STUDIES OF SEASONALITY AND 

MOBILITY IN THE MAJES VALLEY AND PUCUNCHO BASIN OF SOUTHERN PERU 

ABSTRACT 

 

 

 
By 
 

Emily Beatrice Peterson Milton 

 

Oxygen isotopes in archaeological human and animal teeth provide a potential means to 

study past migration. In this thesis I evaluate oxygen isotopes as a way to measure highland site 

occupation seasonality and human movement between elevation zones in the Central Andes. I 

focus on two questions: 1) Do surface water δ18O values vary predictably with altitude? and 2) 

Do surface water δ18O values reliably record the wet and dry season? Using water samples 

collected from the Majes Valley and Pucuncho Basin of southern Peru, I establish the seasonal 

and spatial distribution of δ18O. The final dataset represents 98 water samples collected during 

wet and dry seasons between 2018-2019, from elevations 36 to 4938 meters above sea level.  

 

Surface water δ18O and δD values in the study area are consistent with regional and 

global meteoric waters. Individual water bodies demonstrate relatively higher δ18O values 

during the dry season and lower δ18O values in the wet season, matching predictions. Lowland 

and highland surface waters demonstrate overlapping ranges of δ18O, making it impossible to 

predict the values of surface waters based on elevation alone. Rather, local δ18O appears to 

reflect evaporative processes influenced by stream order, catchment size, and the geographic 

position of a waterbody within the watershed. These data suggest oxygen isotopes are not 

independently suited to resolve questions of human mobility between elevation zones in the 

western Andes. Further testing is needed to understand the local seasonal and inter-zonal 

isotopic ecology. 

 

 

Copyright by 
EMILY BEATRICE PETERSON MILTON 
2020 
 

 

 

 

 

 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

For Marley. 

  iv 

 

ACKNOWLEDGEMENTS 

 

Many minds and hands contributed to the completion of this thesis. First, I thank my 

 
 

committee. To Dr. Katherine Moore, a most valued new mentor, for sharing her incredible 

reservoir of knowledge on South American anthropology and ecology; and for her unrelenting 

tenacity for responding to extensive, late-night inquiries. To Dr. Nathan Stansell for his steady 

hand in teaching South American climate and isotopic theory, the word “fractionation,” and his 

unsolicited interest in archaeological problems. To Doctor Hervé Bocherens for entertaining 

endless questions, instructing me in the careful art of biochemical analysis and interpretation, 

and encouraging creativity when faced with null results. To Dr. Gabe Wrobel, for his eager 

interest in an external project and for offering an open hand in a new department. Finally, to 

Dr. Jodie O’Gorman for her passionate enthusiasm to, “get it done!” 

 

To Dr. Kurt Rademaker, an exceptional advisor and mentor. I owe this research and the 

studies that will follow it to his initial and ongoing support, as well as his enthusiastic 

encouragement of novel and bizarre questions. His belief that students can make important 

contributions to science is perhaps unparalleled. I have such gratitude to him and the approving 

committee for NSF grant 1659015 for funding part of my fieldwork, education, and research 

along the way.  

 

To my fine comrades in research and adventure, Daniela Osorio and Sarah Meinekat, for 

helping to take notes while stuffing me full of laughter. To Brett Furlotte, an admirable field 

companion, union suit exemplar, and hilarious botanist.  

 

I thank the Zuniga family, without whom this research would not have been possible. I 

thank Efraim for his careful work with Marley, and Ermitaño for his adventurous car rides, 

  v 

 

fascinating narratives, and trustworthy dedication to helping a Gringa in need of water and 

tissue samples.  

 

To the lab personnel at the University of Tübingen and Northern Illinois University, 

especially Dr. Dorothée Drucker and Josh Schwartz, for lending critical insights to my method-

development and research design. To Dober Chala-Aldana for sharing and discussing his pilot 

data and samples. To Ruthann Yeaton for sticky notes containing key information.  

 

I thank Dr. Bill Lovis for taking time from his retirement to review and support my 

research and Dr. Joe Hefner for his critiques of my R plots and lax stats.  

 

To my early, but long-term mentors: Dr. Matt G. Hill for the intermittent reality checks; 

Dr. Larry Todd for commentary on glowing-teeth; Stephen Anderson for lessons in work-fun 

balance; and Matthew Neff for acting as a sounding board and editor. To all four for teaching 

me to think, question, and do archaeology as science and as anthropology.  

 

 

I am so grateful for my cohorts at Northern Illinois University, the University of 

Tübingen, and Michigan State University. I thank Jordi for “wife chats”; Mike, Steph, and Taylor 

for keeping my downtime spirited and the karaoke mic hot; Autumn, Jeff, Mari, and Susan, for 

Friday tarot readings and welcoming me to MSU; Malcolm, Alex, Ashley, and Ernesto – for the 

good times and late nights in ESPP. To my 2019 cohort and fast friends, for spouting anarchy 

and activism in theory and praxis. To my people of the past, especially Kinsey, Jacquelyn, and 

Sophia for forcing extended breaks called ‘vacations’; to Rachel, a talented archaeologist and 

editor; and to the best house mates anyone could get quarantined with: Jack, Kevin, Peanut, TC, 

and Vic - I adore four of you.  

  vi 

 

 

Finally, I state my enormous gratitude for my family. Bradley, for the consistent 

reminder that laurels are not cushions; Peg, for biscuit-filled care packages, and unyielding 

writing support throughout these many years; and Evan, for letting me sleep on your couch, 

floor, and spare bed and supplying lab materials when I needed them most. To all three 

nuclears and to my extended family: thank you for patiently listening to rants about alpacas, 

mountains, and teeth. I love you all very, very much.  

 

 

 

 vii 

 

 

TABLE OF CONTENTS 

LIST OF TABLES .............................................................................................................................. xi 

LIST OF FIGURES ........................................................................................................................... xii 

CHAPTER 1: INTRODUCTION .......................................................................................................... 1 
1.1 A Coast-Highland Connection ............................................................................................... 4 
1.2 Summary of Research Problems ........................................................................................ 11 
1.2.1 Baseline Data for a New Seasonality Proxy ................................................................. 12 
1.2.2 An Elevation Transect for Provenance Research ......................................................... 12 
1.3 Research Questions ............................................................................................................ 13 
1.4 Overview of Document Structure ...................................................................................... 14 

CHAPTER 2: BACKGROUND .......................................................................................................... 15 
2.1 Archaeological Context ...................................................................................................... 15 
2.1.1 Coastal Sites ................................................................................................................ 16 
2.1.1.1 Quebrada Jaguay .................................................................................................. 16 
2.1.1.2 Pampa Colorada 343 ............................................................................................ 19 
2.1.2 A High-Elevation Comparison ...................................................................................... 20 
2.1.2.1 Cuncaicha ............................................................................................................. 21 
2.1.3 Outstanding Research at Cuncaicha ............................................................................ 23 
2.1.3.1 Site Seasonality .................................................................................................... 23 
2.1.3.2 Human Provenance .............................................................................................. 25 
2.1.4 Section Summary ........................................................................................................ 26 
2.2 Study Area .......................................................................................................................... 27 
2.2.1 Local Geology .............................................................................................................. 29 
2.2.2 Hydrology .................................................................................................................... 30 
2.2.3 Highland Ecology ......................................................................................................... 32 
2.2.3.1 Fauna .................................................................................................................... 33 
2.2.3.2 Flora ..................................................................................................................... 34 
2.2.4 Section Summary ........................................................................................................ 35 
2.3 Stable Isotopes ................................................................................................................... 35 
2.3.1 Stable Oxygen .............................................................................................................. 36 
2.3.2 Oxygen Isotopes for Archaeological Research ............................................................ 36 
2.3.2.3 Analytical Interpretations ..................................................................................... 38 
2.3.3 Section Summary ........................................................................................................ 38 
2.4 Stable Isotopic Inputs to the Study Area ............................................................................ 39 
2.4.1 Input Sources for the Pucuncho Basin and Western Slope ......................................... 39 
2.4.1.1 Pacific Inputs ........................................................................................................ 41 
2.4.1.2 The Atlantic Source .............................................................................................. 42 
2.4.2 Stable Isotopic Controls .............................................................................................. 42 
2.4.2.1 Fractionation ........................................................................................................ 42 

 viii 

 

2.5 Climate Variability .............................................................................................................. 43 
2.5.1 Annual Variability, Seasonality .................................................................................... 43 
2.5.2 Inter-Annual Variability ............................................................................................... 45 
2.6 Proposed Study and Predictions ........................................................................................ 46 
2.6.1. Predictions .................................................................................................................. 46 
2.6.1.1 Andean Surface Water Seasonality ...................................................................... 46 
2.6.1.2 The Elevation Transect ......................................................................................... 49 

CHAPTER 3: METHODS ................................................................................................................. 50 
3.1 Field Methods and Collection Procedure ........................................................................... 50 
3.1.1 Elevation Sampling ...................................................................................................... 50 
3.1.2 Temporal Sampling ..................................................................................................... 52 
3.1.2.1 Daily Tests ............................................................................................................ 53 
3.1.2.2 Seasonal Tests ...................................................................................................... 53 
3.1.2.3 Yearly Tests .......................................................................................................... 53 
3.1.3 Sampling Procedure .................................................................................................... 54 
3.1.3.1 Labeling System .................................................................................................... 55 
3.2 Laboratory Analysis of Oxygen and Hydrogen ................................................................... 55 
3.3 Calculated Change for Temporal Tests ............................................................................... 56 
3.4 Software Analyses .............................................................................................................. 57 
3.4.1 Watershed and Stream Order Classifications .............................................................. 57 

CHAPTER 4: RESULTS .................................................................................................................... 59 
4.1 General Results .................................................................................................................. 59 
4.1.1 Results of Sample Collection ....................................................................................... 59 
4.1.2 Results of δ18O and δD Analysis .................................................................................. 63 
4.1.3 Stable Isotopic Variation In Andean Surface Waters ................................................... 71 
4.2 Results of Seasonal Sampling ............................................................................................. 72 
4.2.1 Local Variation in Andean Surface Waters Between Seasons ..................................... 73 
4.2.2 Results of Temporal Testing ........................................................................................ 76 
4.2.2.1 Boxplot Assessment for All Temporal Tests ......................................................... 77 
4.2.3 Daily Tests ................................................................................................................... 78 
4.2.3.1 Simultaneous Tests .............................................................................................. 78 
4.2.4 Seasonal Tests ............................................................................................................. 85 
4.2.5 Yearly Tests ................................................................................................................. 91 
4.2.5.1 Annual Tests ......................................................................................................... 94 
4.2.5.2 Interannual Tests .................................................................................................. 94 
4.2.6 Results of Temporal Comparisons ............................................................................... 97 
4.2.6.1 Change Over Multiple Time Scales, Controlled for Location ................................ 99 
4.3 Spatial Variation in Stable Oxygen in Southern Peru ....................................................... 102 
4.3.1 δ18O Variation Across an Elevation Transect ............................................................. 103 
4.3.2 Seasonal Variation in δ18O by Elevation .................................................................... 106 
4.3.2.1 Dry Season .......................................................................................................... 106 
4.3.2.2 Wet Season ........................................................................................................ 107 

  ix 

 

4.3.3 δ18O Variation Across Watersheds ............................................................................ 109 
4.3.3.1 Jaguay Drainage Basin ........................................................................................ 112 
4.3.3.2 Majes Drainage Basin ......................................................................................... 114 
4.3.3.3 Ocoña Drainage Basin ........................................................................................ 114 
4.3.4 Variation in δ18O by Stream Order ............................................................................ 114 
4.3.4.1 Jaguay Surface Waters ....................................................................................... 117 
4.3.4.2 Majes Surface Waters ........................................................................................ 117 
4.3.4.3 Ocoña Surface Waters ........................................................................................ 118 
4.3.5 Range of δ18O Across Elevations Throughout the Central Andes .............................. 120 
4.4 δ18O for Archaeological Human Provenance Research .................................................... 124 
4.4.1 Waters On and Off the Plateau ................................................................................. 125 
4.4.2 Abbreviated Vidal Ecozones ...................................................................................... 126 
4.4.3 The Humans of Cuncaicha ......................................................................................... 129 

CHAPTER 5: DISCUSSION AND CONCLUSIONS ........................................................................... 132 
5.1 General Comments .......................................................................................................... 132 
5.2 Seasonal δ18O Variation in Regional Surface Waters ....................................................... 132 
5.2.1 Prospectus for a Seasonal Proxy ............................................................................... 134 
5.3 Spatial δ18O Variation in the Central Andes ..................................................................... 135 
5.3.1 Reconciling the Watershed Effect and Archaeological Aims ..................................... 137 
5.4 Isotopic Methods ............................................................................................................. 139 
5.4.1 The Baseline: Isotopes Through Time, Across Space ................................................. 139 
5.4.2 Consider the Dentist: Microdrills and Teeth ............................................................. 141 
5.5 Next Steps ........................................................................................................................ 141 
5.5.1 Seasonality Research ................................................................................................. 141 
5.5.2 Mobility Research ...................................................................................................... 143 
5.6 For Posterity ..................................................................................................................... 144 
5.7 Summary of Conclusions .................................................................................................. 145 
5.8 A Case for Interdisciplinary Science ................................................................................. 146 

REFERENCES ............................................................................................................................... 148 
 
 

 

  x 

 

 

LIST OF TABLES 

Table 1: Summary of Results. In order of increasing elevation. Water Sample (WS) number 
indicates the year the sample was collected. .............................................................................. 65 

Table 2: Summary of Sampling for Each Field Season. ................................................................. 73 

Table 3: Results of Daily Tests. ..................................................................................................... 80 

Table 4: Statistical Results for Simultaneous Tests. ..................................................................... 84 

Table 5: Results of Seasonal Sampling. ........................................................................................ 86 

Table 6: Statistical Results of Seasonal Sampling. ........................................................................ 88 

Table 7: Results of Annual Sampling. ........................................................................................... 91 

Table 8: Results of Annual Statistical Tests. ................................................................................. 94 

Table 9: Results of Interannual Tests. .......................................................................................... 95 

Table 10: Results of Interannual Statistical Tests. ........................................................................ 95 

Table 11: Results of Elevation Subsets. ...................................................................................... 103 

Table 12: Results Divided by Watershed. ................................................................................... 109 

Table 13: Summary of δ18O Variation Among Stream Orders. ................................................... 115 

Table 14: Summary of Data from Bershaw et al. 2016. ............................................................. 120 

Table 15: δ18ODW, 87Sr/86Sr, and δ13CVPDB values from the humans of Cuncaicha. ...................... 129 

 

 

  xi 

 

 

LIST OF FIGURES 

Figure 1: Map of the Central Andean study region. Light blue circle shows study area. ............... 3 

Figure 2: Map of the study area. Sites include Pampa Colorada 343 and Quebrada Jaguay 280 at 
the coast, Cuncaicha in the highlands, and three Alca obsidian sources. ...................................... 5 

Figure 3: Aerial view of the three sites and Alca obsidian sources. Imagery from Google Earth 
Pro. ............................................................................................................................................... 16 

Figure 4: Overview of Quebrada Jaguay 280. View to the north during the dry season. ............ 19 

Figure 5: View of the shell middens at Pampa-Colorada 343. View to the northeast during the 
dry season. Photo curtesy of Kurt Rademaker. ............................................................................ 20 

Figure 6: Cuncaicha rockshelter archaeological site on the Andean Plateau. View to the south 
during the dry season. Nevado Coropuna is visible to the east behind the shelter. ................... 22 

Figure 7: Cross-section of an adult vicuña incisor showing parallel enamel growth lines 
characteristic of an ever-growing tooth. ...................................................................................... 25 

Figure 8: Map of Cuncaicha rock shelter, Pucuncho Basin, and Alca-1, 4, and 5 sources. ........... 28 

Figure 9: Map of Ocoña, Jaguay, and Majes watersheds. ............................................................ 31 

Figure 10: The prominent water bodies of the study area. These include high-altitude springs 
(upper right, dry season) and wetlands (upper left, wet season); as well as canals (lower left, dry 
season) and rivers and streams (lower right, dry season) found at all elevations. ...................... 32 

Figure 11: A herd of vicuña on the Andean Plateau, posed in front of Nevado Coropuna. Image 
shows llareta cushion plants and bunch grasses during the dry season. ..................................... 34 

Figure 12: Predicted seasonal variation in !18O in precipitation for the Pucuncho Basin based on 

OIPC modeled values for water sampling locations of Chala-Aldana et al., 2018. ....................... 44 

Figure 13: SENAMHI stream gauge data for 2018-2019 in the Majes (top) and Ocoña (bottom) 
watersheds. Each graph shows volume of stream discharge (left axis) over time (bottom axis). 
Together these images show the complete onset, peak, and decline of the 2018-2019 wet and 
dry seasons in the study area. ...................................................................................................... 48 

Figure 14: Map of the Survey Area. The survey route (light pink line) extends from Pampa 
Colorada 343 and Quebrada Jaguay 280 to Cuncaicha. Sampling crossed three watersheds, the 
Ocoña, Jaguay, and Majes Drainages. .......................................................................................... 51 

 xii 

 

Figure 15: Map of key sampling locations below the plateau. The confluence of Río Capiza 
(dashed yellow line) and the Río Colca (dashed blue line) mark the beginning of the Río Majes 
(blue area). The Río Blanco (dashed pink line) discharges over the plateau rim (dotted white 
line) around ~3500 masl. ............................................................................................................. 52 

Figure 16: Map provides an example of stream orders in the study area. .................................. 58 

Figure 17: Map of all sample locations within the survey area. Yellow triangles show dry season 
testing and blue circles indicate wet season sampling. ............................................................... 60 

Figure 18: Density of samples collected for wet (blue line) and dry (yellow line) seasons. The dry 
season shows a bimodal distribution of samples; most were collected below 1000 masl or 
above 3500 masl. The wet season shows a unimodal distribution with most samples collected 
from above 3500 masl. ................................................................................................................ 63 

Figure 19: Plot of δ18O and δD for all samples. The black line is the Global Meteoric Water Line 
(GMWL) δD = 8(δ18O) +10‰. The dashed black line reflects the local surface water line (LSWL) 
δD= 7.86 (δ18O) + 7.08‰. Class 1-5 indicates stream order, where first order streams are the 
smallest volume. The enriched samples below the line are discussed in Figure 21. ................... 72 

Figure 20: Seasonal datasets plotted with the respective wet and dry season Local Surface 
Water Lines (LSWL). Dry season samples (yellow dots) and LMWL (δD = 7.75 * δ18O + 5.52‰) 
are presented on the left and wet season samples (blue dots) and LMWL (δD = 8.15 * δ18O + 
11.22‰) presented on the right. Values that plot below the GMWL and LSWLs reflect enriched 
waters. ......................................................................................................................................... 74 

Figure 21: δ18O plotted against δD for sample locations replicated as part of the seasonality 
study. Dashed black line indicates the global meteoric waterline δD = 8.0 * δ18O + 10.00‰. Blue 
solid line δD = 8.10 * δ18O + 10.83‰. indicates the local meteoric water line during the wet 
season. Yellow dashed line represents the local meteoric water line during the dry season δD = 
7.36 * δ18O + 3.14‰. More samples appear to plot below the line during the dry season. ....... 76 

Figure 22: Box plot assessment for the per mil δ18O difference between paired temporal 
samples. Mean instrumental precision (0.03‰) shown by the dashed black line. Maximum 
instrumental precision (0.22‰) indicated by the dotted grey line. ............................................ 78 

Figure 23: Map of daily sampling locations. Replicate tests are dark blue: the Río Majes, SIM-01 
(383 masl); The Río Colca, SIM-02 (896 masl); the Río Grande, SIM-03 (2980 masl), SIM-04 (3396 
masl), and SIM-05 (3727 masl); Quebrada Pariaviri, SIM-06 (3967 masl); and the Río Arma, SIM-
07 (4085 masl). The diurnal sample (lighter blue), DIUR-01 (3936 masl) derives from the 
Quebrada de Rata. ....................................................................................................................... 81 

Figure 24: Observed δ18O for simultaneous (dark blue triangles) and diurnal (light blue circles) 
pairs. Samples include the Río Majes, SIM-01 (383 masl); The Río Colca, SIM-02 (896 masl); the 

 xiii 

 

Río Grande, SIM-03 (2980 masl), SIM-04 (3396 masl), and SIM-05 (3727 masl); Quebrada 
Pariaviri, SIM-06 (3967 masl); and the Río Arma, SIM-07 (4085 masl). The diurnal sample, DIUR-
01 (3936 masl) derives from the Quebrada de Rata. ................................................................... 82 

Figure 25: Change in δ18O between pairs collected for daily tests. Replicate tests (dark blue 
triangles) include the Río Majes, SIM-01 (383 masl); The Río Colca, SIM-02 (896 masl); the Río 
Grande, SIM-03 (2980 masl), SIM-04 (3396 masl), and SIM-05 (3727 masl); Quebrada Pariaviri, 
SIM-06 (3967 masl); and the Río Arma, SIM-07 (4085 masl). The diurnal sample (light blue 
circle), DIUR-01 (3936 masl) derives from the Quebrada de Rata. All demonstrate less change 
than the maximum machine error (0.22‰). ................................................................................ 83 

Figure 26: Map of the seasonal sample locations: the Río Majes, SEAS-01 (383 masl); The Río 
Grande, SEAS-02 (1379 masl); the Canal de Pacaychara, SEAS-03 (1811 masl); Huacllay Canal, 
SEAS-04 (1960 masl); Iray Canal, SEAS-05 (2521 masl); Quebrada de Rata, SEAS-06 (3936 masl); 
the Río Arma, SEAS-07 (4310 masl); Río Colunga, SEAS-08 (4333 masl); Río Cuncaicha, SEAS-09 
(4347 masl); Quebrada Japuchaca, SEAS-10 (4435 masl); Quebrada Pucunchitoyocc, SEAS-11 
(4717 masl); and Bofedal Pucunchitoyocc, SEAS-12 (4720 masl). ............................................... 87 

Figure 27: Plot showing δ18O values for each sample in the seasonal pairs. Data are coded by 
wet (blue triangles) and dry (yellow circles) seasons. The Majes (purple), Jaguay (blue), and 
Ocoña (coral) drainages are indicated. Location codes reflect: the Río Majes, SEAS-01; The Río 
Grande, SEAS-02; the Canal de Pacaychara, SEAS-03; Huacllay Canal, SEAS-04; Iray Canal, SEAS-
05; Quebrada de Rata, SEAS-06; the Río Arma, SEAS-07; Río Colunga, SEAS-08; Río Cuncaicha, 
SEAS-09; Quebrada Japuchaca, SEAS-10; Quebrada Pucunchitoyocc, SEAS-11; and Bofedal 
Pucunchitoyocc, SEAS-12. ............................................................................................................ 89 

Figure 28: Plot showing the difference in δ18O (‰) between seasonal sample pairs. Location 
codes in order of increasing elevation: the Río Majes, SEAS-01; The Río Grande, SEAS-02; the 
Canal de Pacaychara, SEAS-03; Huacllay Canal, SEAS-04; Iray Canal, SEAS-05; Quebrada de Rata, 
SEAS-06; the Río Arma, SEAS-07; Río Colunga, SEAS-08; Río Cuncaicha, SEAS-09; Quebrada 
Japuchaca, SEAS-10; Quebrada Pucunchitoyocc, SEAS-11; and Bofedal Pucunchitoyocc, SEAS-12.   
All samples demonstrate more change than maximum machine error (0.22‰) and maximum 
daily change (0.16‰). .................................................................................................................. 90 

Figure 29: The δ18O of surface waters sampled during the dry season in multiple years. Annual 
samples (<1 year; yellow circles) include: Quebrada Pariaviri, YR1-01 (3966 masl); Río Arma, 
YR1-02 (4300 masl); and Quebrada Cuncaicha, YR1-03 (4457 masl). Multiyear samples (>3 years; 
orange triangles) include the Río Majes, MYR-01 (383 masl); The Río Grande, MYR-02 (1379 
masl); Río Grande, MYR-03 (2980 masl); Río León Huactana MYR-04 (4384 masl); Quebrada 
Cuncaicha (3-year) MYR-05 (4457 masl); and Quebrada Cuncaicha (4-year) MYR-06 (4474 masl).
 ..................................................................................................................................................... 92 

Figure 30: Map of the yearly sample locations. Annual samples (<1 year) include: Quebrada 
Pariaviri YR1-01 (3966 masl); Río Arma, YR1-02 (4300 masl); and Quebrada Cuncaicha, YR1-03 

 xiv 

 

(4457 masl). Multiyear samples (>3 years) include the Río Majes, MYR-01 (383 masl); The Río 
Grande, MYR-02 (1379 masl); Río Grande, MYR-03 (2980 masl); Río León Huactana MYR-04 
(4384 masl); Quebrada Cuncaicha (3-year) MYR-05 (4457 masl); and Quebrada Cuncaicha (4-
year) MYR-06 (4474 masl). ........................................................................................................... 93 

Figure 31: Per mil difference in δ18O between surface waters sampled during the dry season in 
multiple years. Annual samples (<1 year; yellow circles) include: Quebrada Pariaviri, YR1-01 
(3966 masl); Río Arma, YR1-02 (4300 masl); and Quebrada Cuncaicha, YR1-03 (4457 masl). 
Multiyear samples (>3 years; orange triangles) include the Río Majes, MYR-01 (383 masl); The 
Río Grande, MYR-02 (1379 masl); Río Grande, MYR-03 (2980 masl); Río León Huactana MYR-04 
(4384 masl); Quebrada Cuncaicha (3-year) MYR-05 (4457 masl); and Quebrada Cuncaicha (4-
year) MYR-06 (4474 masl). ........................................................................................................... 96 

Figure 32: Per mil δ18O change for simultaneous (dark blue), diurnal (light blue), seasonal 
(green), one-year (yellow) and multi-year (orange) tests. Numbers correspond to sampling 
locations for each test in order of elevation. Maximum observed machine error indicated by the 
black dashed line (0.22‰).   Maximum daily change (0.16‰) indicated by the blue dashed line. 
Maximum yearly change (0.58‰) indicated by the orange dotted line. ..................................... 98 

Figure 33: Measured per mil difference in δ18O at the daily (yellow) and seasonal (red) scales for 
the Quebrada de Rata, Río Arma, and Río Majes. ...................................................................... 100 

Figure 34: Per mil change in the δ18O of surface waters at the daily (yellow), seasonal (red), and 
annual (blue) scales for the Río Arma, Río Grande, and Río Majes. Pairings from the Quebrada 
Cuncaicha reflect yearly change at a one- and four-year scale. ................................................. 102 

Figure 35: Map of all sampling locations illustrating the δ18O values of the collected surface 
waters. ....................................................................................................................................... 104 

Figure 36:Distribution of δ18O across elevation. The plot incorporates all samples. Four outliers 
include WS1962 (pink) in the Ocoña drainage basin, WS1815 (blue) in the Majes drainage basin, 
and the paired samples WS1945 and WS1965 (orange) from the Jaguay drainage basin. ........ 105 

Figure 37: Distribution of δ18O of Andean surface waters by elevation. Data are divided by 
season. Yellow dots (left) reflect the 2018 and 2019 dry season samples and blue triangles 
(right) the 2019 wet season samples. ........................................................................................ 108 

Figure 38: Dot plot of δ18O range for the three sampled drainage basins. The Ocoña expresses 
the most depleted surface water values overall, and the Jaguay the most enriched values. ... 110 

Figure 39: δ18O variation by elevation divided by individual watersheds. Jaguay drainage is on 
the left, Majes center, and Ocoña right. Data are coded by wet (blue triangles) and dry (yellow 
circles) seasons. .......................................................................................................................... 111 

 xv 

 

Figure 40: Map of the Jaguay drainage basin and the Quebrada de Rata samples. .................. 113 

Figure 41: Map showing stream order for sampling locations. ................................................. 116 

Figure 42: Stream order δ18O values for the Jaguay, Majes and Ocoña watersheds. ................ 119 

Figure 43: Map of Bershaw et al. 2016 (circles) and Milton 2018-2019 (triangles) water sample 
locations. δ18O values scaled to be consistent across both datasets. ........................................ 121 

Figure 44: Bershaw et al. 2016 (dark grey circles) and Milton’s 2018-2019 (light grey triangles) 
surface water samples plotted in a regression of δ18O and δD. The solid black line is the Global 
Meteoric Water Line (GMWL) δD = 8(δ18O) +10‰. ................................................................. 122 

Figure 45: The elevation distribution of δ18O in surface waters for both Bershaw et al. 2016 
(dark grey circles) and Milton 2018-2019 (light grey triangles) data. ........................................ 124 

Figure 46: Plot comparing the spread of possible δ18O values off (light grey circles) and on (dark 
grey triangles) the Andean Plateau. ........................................................................................... 126 

Figure 47: Dot plot of the spread of δ18O values observed in the surface waters across three 
Central Andean ecozones. Coastal (<500 masl), Yungas (500-2500 masl), and Highland (>2500 
masl). .......................................................................................................................................... 128 

Figure 48: δ18ODW values for the five humans of Cuncaicha plotted against surface water values 
from the study area. .................................................................................................................. 130 

Figure 49: The δ18ODW for the humans of Cuncaicha plotted against the δ18O values in surface 
waters from the Central Andes. ................................................................................................. 131 

 

 xvi 

 

CHAPTER 1: INTRODUCTION 

 

Our species experiences physiological stress at high-altitude because of environmental 

factors such as hypoxia, increased ultra-violet (UV) radiation, cold temperatures, and low 

humidity (Moore 1979, 2017). The Andes mountains of South America, which average 4000 

meters above sea level, constitute the second highest elevation landscape in the world. 

Evidence from multiple archaeological sites indicates humans began using the Andes at the end 

of the Last Ice Age (Capriles et al. 2016; Latorre et al. 2013; Rademaker et al. 2014; Yacobaccio 

2017). The broader behavioral and mobility strategies of the first Andeans, and the eventual 

processes of permanent occupation and animal domestication in the high Andes remain 

uncertain.  

 

Decades of archaeological excavations in South America have led some archaeologists 

to postulate that high-altitude adaptations developed over multiple millennia (Aldenderfer 

2006, 2011; Aldenderfer 1998). Research following this model proposes that phenotypical 

adaptations to altitude emerged through restricted short-term forays into the highlands from 

lower elevations (Capriles et al. 2018; Capriles et al. 2016; Haas et al. 2017). Others have 

proposed that humans may have entered the Andean system at the northern margins, 

developing adaptations as they migrated south, gradually travelling upward. In the latter 

model, the Andes would have functioned as a migration corridor for initial groups colonizing 

the continent (Osorio et al. 2017). A third potential suggests no adaptations are necessary for 

humans to stay at altitude long-term; rather, resource structure underlies the tenability of 

montane landscapes.  A detailed study of regional settlement systems could provide a viable 

 

1 

 

 

approach to unraveling the processes of high-altitude adaptation, initial land use practices, and 

the eventual domestication of llamas and alpacas in the Central Andes.  

 

In this thesis I present three linked archaeological sites that support a detailed study of 

settlement systems and adaptation in the Central Andes (Figure 1). I evaluate the potential of 

using oxygen isotopes to elucidate the connection between these sites. Oxygen isotopes 

preserved in the enamel apatite of human and animal teeth may be applied as tracers of both 

human mobility and climate seasonality. As the first phase of study, I have developed a 

localized oxygen isotope baseline to assess whether stable oxygen would provide a viable 

means to resolve anthropological questions in the region. In this introduction I present a brief 

history of the archaeological sites and review previous research to identify outstanding 

problems and questions requiring further investigation. I discuss my selection of oxygen 

isotopes to contribute to these issues and conclude with the proposed study and an outline for 

the document.   

 

2 

 

 

Figure 1: Map of the Central Andean study region. Light blue circle shows study area. 

 

3 

 

 

 

1.1 A Coast-Highland Connection 

 

Near the end of the Terminal Pleistocene (TP), after 14,000 years ago, the first people to 

colonize South America encountered a complicated backdrop of diverse ecosystems. The 

mountainous spine that bisects the continent separates the arid Pacific Coast from the lush 

tropics of the Amazon. On the western border, the steep orography of the Andes Mountains 

results in vertically stacked ecological zones, each characterized by unique resource structures 

(Vidal 1981). Studying the first peoples of South America requires the examination of these 

varied landscapes and associated adaptive strategies. 

 

 

In southern Peru, material evidence substantiates a connection among various 

ecological zones that began in the Terminal Pleistocene (TP) and continued well into the 

Holocene. Some of the earliest archaeological indicators of highland use are found along the 

Pacific Coast (Figure 2). In 1996, excavations at the maritime site, Quebrada Jaguay 280 (QJ-

280), revealed Alca obsidian flakes and a single tool in TP deposits (Sandweiss et al. 1998). The 

volcanic glass, Alca, outcrops at multiple locations to the north in the high Andes (Rademaker 

2006; Rademaker, Glascock, et al. 2013). As no natural processes can transport the tool stone 

to the coast, the presence of obsidian at QJ-280 points to an early use of the highlands by the 

first peoples of South America. The material connection extends through both time and space. 

Northwest of QJ-280, at the coastal site Pampa Colorada 343 (PC-343), Alca obsidian is also 

found in association with Early Holocene shell middens and human burials (McInnis 2006).  

 

4 

 

 

Figure 2: Map of the study area. Sites include Pampa Colorada 343 and Quebrada Jaguay 280 at the 

coast, Cuncaicha in the highlands, and three Alca obsidian sources. 

 

 

5 

 

 

 

In 2007, a key site for understanding the coastal obsidian was discovered approximately 

150 kilometers to the north. At 4480 meters above sea level, the Cuncaicha rockshelter in the 

Pucuncho Basin is the highest elevation Ice Age archaeological site in the western hemisphere 

(Rademaker et al. 2014; Rademaker 2012). The oldest levels date to the Terminal Pleistocene, 

between 12,600-11,200 years ago, a date range that overlaps the TP sequence at QJ-280 

(Rademaker and Hodgins 2018). Cuncaicha is contemporaneous with many of the oldest known 

sites in South America, and older than most in the Central Andean region (Rademaker, Bromley, 

et al. 2013). Artifacts from the TP deposits include a rich faunal assemblage of local ungulates 

and small mammals, and evidence for bone tool and bead production. The dense and diverse 

lithic assemblage consists of stone tools and debitage manufactured from highland raw 

materials, such as Alca obsidian. Charred botanicals and ash demonstrate that plants were 

incorporated into the diet and used for fuel, while unique crafting items such as quartz crystals 

and ochre suggest symbolic or creative activities were performed on site. In the Early Holocene, 

~ 11,800 to 8,200 years ago, non-local low-elevation raw materials and shell appear indicating 

an increased reliance on inter-zonal movement or trade; and three EH burials post-dating the 

occupation sequences suggest a change in site use and significance over time. 

 

Alca obsidian at the coast hints at an intriguing connection between low and highland 

ecosystems; a link manifested through either trade or direct transport. Today, the Pacific coast 

of southern Peru is extremely arid, and in some areas fresh-water, edible plants, and fauna are 

only seasonally available. In the local highlands, water and game are perennial resources, but 

the extreme elevation and cold inflicts an exacting toll on the human body. The physiological 

challenges at high elevation may have necessitated specific biocultural responses, including 

 

6 

 

 

cold-adapted technology or mobility strategies that used lower elevation zones. The obsidian 

link provides an archaeological mechanism to assess regional settlement systems and 

potentially, high-altitude occupation and adaptation. If obsidian arrived at the coast via a trade 

network it may reflect groups exercising residential movement within interior zones. 

Alternatively, direct transport would require logistical travel to the highlands but not 

necessitate an extended stay in the Andes.  

 

The obsidian data from the tree linked sites provide convincing evidence of high-altitude 

use from the time of human arrival in the area.  Resolving the connection between the regional 

low and high elevation landscapes, and nature of Cuncaicha’s changing deposits, requires 

secure evidence to reveal patterns of annual landscape use, migration strategies, and trade 

networks between the coast and high Andes. Suitable archaeological proxies for mobility may 

include (A) direct provenance data from human remains or, in the absence of biological tissues; 

or (B) more indirect measures, such as seasonality data from multiple sites across a region.  

 

Provenance research seeks to identify the origin of an individual or item by tying their 

chemical signatures to source locations in the environment.  Such analyses are commonly 

applied to raw materials or human remains to trace how far humans have transported items, or 

themselves, across a landscape. In the Central Andean region, the application of provenance 

analyses to the human remains at Cuncaicha and PC-343 could reveal whether groups resided 

predominantly at low, middle, or high elevations.  

 

Data revealing season of site use will help determine the time of year high-elevation 

living took place, and whether that was restricted to a single season (wet or dry). Further, 

comparing seasonality data from the coast and the highlands could reveal whether groups 

 

7 

 

 

occupied the landscapes at the same or different times of the year. While the former might 

imply a trade connection, the latter would more likely reflect inter-zonal movement by a single 

group.  

 

 Macro botanical and tooth eruption data from Cuncaicha’s juvenile fauna suggest a 

wet-season occupation between March-April (Moore 2016b). However, neither analytical 

technique is equipped to identify a dry season occupation. For this reason, Moore (2016) 

suggested a stable isotopic study of the preserved adult camelid (vicuña or guanaco) teeth, as a 

potential method of resolving the full season of site use at Cuncaicha.  

 

Stable isotopes could provide a viable method of analysis for both seasonality and 

provenance research. Isotopes occur in semi-predictable ratios throughout the environment 

and do not decay over time, thus they provide a useful mechanism for identifying past and 

present environmental parameters (Michener and Lajtha 2007). Humans and animals 

incorporate the isotopic composition of their local environment via eating, breathing, and 

drinking (Schoeninger 1995; Schoeninger and Moore 1992; Schwarcz and Schoeninger 1991). 

Recent advances in archaeological research have demonstrated the reliable potential of using 

isotopic markers in tooth and bone as proxies for variations in climate (Green et al. 2018; Sharp 

and Cerling 1998), ecological change, (Cerling et al. 2015), diet (Knudson et al. 2007; Knudson et 

al. 2015; Tung and Knudson 2018), and geographic movement (Knudson et al. 2009; Tung and 

Knudson 2018).  

 

Oxygen isotopes, presented in delta-notation (δ), are commonly applied to seasonality 

research. δ18O values measured from meteoric waters reflect the amount, source, and season 

of precipitation  (Craig 1961a; Dansgaard 1964; Gat 1996, 2000). In the study area, the austral 

 

8 

 

 

wet season extends from late December to mid-April, corresponding with incoming Atlantic 

Ocean moisture carried across the continent by the easterlies (Garreaud et al. 2009).Therefore, 

it is likely that highland seasonality is predominately reflected in local surface waters as they 

record annual pulses of precipitation. Surface water seasonality must be tested directly to be 

sure regional aquifers do not attenuate the incoming signal from precipitation.  

 

In addition to seasonal fluctuations, δ18O in surface waters and vegetation varies across 

a landscape and often reflects atmospheric processes such as latitude and altitude (Craig 

1961a). Oxygen isotopes have been proposed and applied elsewhere as a promising biomarker 

for identifying movement between the distinctive ecological zones of the Andes (Dufour et al. 

2014; Haas et al. 2017; Knudson 2009). The hypothesized justification states that waters 

available in cold, wet highland locations will demonstrate lower (δ18O) values than those found 

in the warmer, drier lowlands. The elevation trend has been based on interpolated values 

provided by the Oxygen Isotopes in Precipitation Calculator (OIPC) (Bowen 2017; Bowen and 

Revenaugh 2003; Bowen et al. 2005). The calculator is an online tool that predicts the stable 

isotopic composition of precipitation based on location. The actual patterning of δ18O in the 

surface waters of the west Central Andes, and how those data pertain to archaeological 

research questions, is only tentatively known, as extensive and intensive field sampling of 

surface waters is limited (Knudson 2009).  

 

In 2017, Chala-Aldana conducted a stable isotopic study of provenance for the humans 

of Cuncaicha to investigate whether they developed at or travelled to the high-altitude 

Pucuncho Basin. The study utilized radiogenic strontium (87Sr/86Sr) and biogenic oxygen (δ18O) 

data sampled from the enamel carbonates of the five individuals’ teeth. To construct a baseline 

 

9 

 

 

for interpretation Chala-Aldana collected plant and water samples along an elevation transect 

from the Pacific Coast near QJ-280, to the Pucuncho Basin near Cuncaicha. Plants incorporate 

local strontium bedrock values, and those plants in turn are ingested by surrounding fauna and 

humans (Knudson et al. 2010; Scaffidi and Knudson 2020; Scaffidi et al. 2020). Fresh surface 

waters constitute a primary drinking water source for humans and animals. Chala-Aldana’s 

samples represents the first attempt to map differences in environmental isotopic compositions 

between low and high elevations in the Central Andes. 

 

Unfortunately, a local strontium isotope baseline was not possible because there was 

insufficient plant material collected for analysis. Therefore, oxygen isotope data from water 

samples provided a critical secondary line of evidence to the interpretation that the humans of 

Cuncaicha developed at altitude. Yet the preliminary environmental data for oxygen isotopes 

provided conflicting results. Three of the six low elevation (<3500 masl) samples published by 

Chala-Aldana et al. (2018) provided lower stable oxygen values that would be expected at high 

altitude . Simultaneously, 3 of the 14 highland samples presented the reverse trend, reporting 

lower values than expected. At the time of publication, Chala-Aldana (2017) explained the 

anomalous samples as follows: 

In the case of the samples Pu-wa-021, Out-wa-004 and Out-wa-002, which came from 
standing waters above 3500 masl (Table 8), it is likely that high evaporation rates for the 
winter 2015 led to an increase of the δ18O, likely explaining the values obtained. 
Standing waters are more affected by evaporation than flowing waters. In the case of 
the samples outside the basin [at low elevation], rapid flow from high elevation waters 
(Knudson, 2009) likely caused the anomalies in the samples Out-wa-012, Out-wa-013 
and Out-wa- 014. δ18OMW [sic] values of the rest of the samples –after excluding the 
values with anomalies- fit within the expected OIPC trend. (33)  

 

10 

 

 

 

Due to the challenges involved in setting a baseline from plants, the final interpretation 

considered regional strontium values that had been published elsewhere (Chala-Aldana 2017). 

However, a recent regional dataset published by Scaffidi et al. (2020) suggests strontium may 

not account for movement between elevations. The regional bedrock in some areas of the 

Central Andes presents as a series of uniform geologic layers that bisect multiple elevation 

zones. The bedrock geology renders strontium a poor candidate to independently distinguish 

migration of humans or animals between low and high elevation. Considering the incredible 

utility of a biochemical proxy to differentiate between low and highland living, it is crucial to 

resolve whether stable oxygen is demonstrably suited to archaeological provenance research in 

the Andes or whether another isotope will need to be considered. 

1.2 Summary of Research Problems 

 

Based on the density and diversity of the earliest site deposits, Rademaker et al. (2014), 

have postulated that the Cuncaicha rock shelter may have been used as base camp as early as 

the Terminal Pleistocene. Such an interpretation would suggest a more rapid biogeographic 

expansion into high-altitude landscapes than previously proposed (Aldenderfer 2006, 2011; 

Aldenderfer 1998). Capriles et al. (2016) contest Rademaker et al.’s conclusion directly and 

argue instead that the Pleistocene occupations likely correspond to “circuits of seasonal 

logistical mobility” (p. 99).   

 

Determining the seasonality for the Cuncaicha occupation sequences will not resolve 

the debate regarding site function. However, in combination with other archaeological 

measures including sediment micromorphology, faunal and lithic analyses, and paleobotanical 

studies, seasonality would provide critical insights into how humans first began to colonize 

 

11 

 

 

high-elevation landscapes (Rademaker et al. 2016). Provenance data from human remains 

promise the greatest potential for identifying the emergence of extended highland use.  

1.2.1 Baseline Data for a New Seasonality Proxy 

 

The faunal and botanical indicators at Cuncaicha reveal wet season occupations during 

both the Terminal Pleistocene and Early Holocene. Whether the site was also used during the 

dry season remains untested. Establishing whether there was a dry season occupation will 

require a new seasonality proxy. Moore (2016) recommended an incremental stable isotope 

analysis of the archaeological adult faunal teeth at Cuncaicha. As the first phase of this 

research, I sample the surface waters of the high elevation landscape surrounding Cuncaicha to 

investigate the seasonal isotopic ecology of the area.  

1.2.2 An Elevation Transect for Provenance Research  

 

δ18O measured on human and animal tooth enamel have been proposed as a 

meaningful way to study migration between western Andean ecological zones (Knudson 2009). 

The principles proposed by Knudson (2009) have since been deployed as assumptions to 

underly arguments for permanent high-elevation land use by 9,000 to 7,000 years ago (Chala-

Aldana et al. 2018; Haas et al. 2017; Haas et al. 2020). All three studies suggest that the δ18O 

values measured from enamel carbonate (Chala-Aldana et al. 2018) or bone bioapatite (Haas et 

al. 2020; Haas et al. 2017) are consistent with reported values for highland waters. However, 

with the exception of the 20 water samples collected by Chala-Aldana (2017), the actual stable 

oxygen patterning of available drinking waters in the local study areas has yet to be explicitly 

 

12 

 

 

established. An environmental baseline for the west Central Andes will provide the needed data 

to confirm or reinterpret such studies.  

 

1.3 Research Questions 

 

 

The goal of this interdisciplinary thesis is to develop a reliable δ18O baseline for the 

surface waters of the Majes Valley and nearby highlands of southern Peru. I defined the study 

area based on the location of the three interlinked archaeological sites, Quebrada Jaguay 280, 

Pampa Colorada 343, and Cuncaicha. To develop the baseline, I collected 98 samples from 

flowing waters between 36 and 4938 meters in elevation over three austral seasons: July 2018 

(dry), February and March 2019 (wet), and June 2019 (dry).  

 

The intended outcome is twofold: first, an environmental dataset is necessary to justify 

developing a new, exacting method of seasonality for Andean sites; second, an elevation 

transect will reveal whether stable oxygen is a suitable isotope for provenance research, 

including studies that assess movement or residence among different altitudinal zones (Chala-

Aldana et al. 2018; Haas et al. 2017; Knudson 2009). The research conducted for this thesis 

does not include the stable isotopic study of the archaeological materials from Cuncaicha. 

Rather, it initiates the development of an environmental baseline to justify and review 

archaeological research. I address the following research questions: 

1.  Does δ18O measured from the surface waters in the Pucuncho Basin record seasonal 

change?  

I investigate seasonal and annual oxygen variation in the surface waters in the study 

area surrounding Cuncaicha as the first phase in developing an isotopic proxy of 

seasonality. 

 

13 

 

 

2.  Do the surface waters between the Pacific coast and Andes demonstrate distinct δ18O 

values at different elevations?  

I conducted a targeted elevation-transect from the coast to the highlands to identify the 

specific elevation patterning and primary controls for surface water δ18O in the study 

area.   

 

I develop a baseline for surface water δ18O using samples collected from the low 

elevation Majes Valley and highland Pucuncho Basin during austral wet and dry seasons. 

1.4 Overview of Document Structure 

 

To begin, the background in Chapter 2 provides a justification for the intended research 

in terms of both the archaeological questions asked and the environmental sampling approach 

applied. It begins by expanding on the archaeological coast-highland connection in southern 

Peru and describing the research area. The history of investigations and outstanding research 

questions for the three sites are synthesized. This is followed by a review of stable isotopic 

analysis in archaeological and environmental research and sets specific predictions for the 

project based on the regional geography of the research area. Chapter 3 presents the field and 

analytical methods used to construct elevation and seasonal surface water baselines. Chapter 4 

presents the results of three seasons of data collection.  Part 1 summarizes general results. Part 

2 presents the results of seasonal sampling and Part 3 covers the results of the elevation 

transect. Chapter 5 discusses the implications of the stable oxygen isotope data for seasonality 

and mobility studies in the south-Central Andes. In conclusion, Chapter 6 reveals the composite 

limitations and implications of this work to suggest future directions for stable oxygen analyses 

in Andean archaeology. 

 

14 

 

 

CHAPTER 2: BACKGROUND 

 

Chapter 2 presents a regional literature review to justify the proposed isotopic study. I 

first review in depth the archaeological evidence for a Terminal Pleistocene and Early Holocene 

connection between the coast and highlands of southern Peru. I then provide a geological and 

ecological summary of the study area. Next I review the principles of stable isotopic research, 

the applications in archaeology, and specific regional controls as they pertain to the study.  I 

conclude by proposing two predictions for expected seasonal and elevation δ18O baselines in 

the project area.  

2.1 Archaeological Context 

 

 

Recent excavations and collections work at the highland site Cuncaicha and the coastal 

sites Quebrada Jaguay 280 (QJ-280) and Pampa Colorada 343 (PC-343) provide the materials 

necessary to investigate inter-zonal connections more closely (Figure 3). The next phase of 

investigation requires determining whether the presence of non-local raw materials, at the 

coast by the Terminal Pleistocene and in the highlands by the Early Holocene, are the product 

of direct or indirect (trade) procurement (Rademaker 2017). 

 

15 

 

 

Figure 3: Aerial view of the three sites and Alca obsidian sources. Imagery from Google Earth Pro. 

 

2.1.1 Coastal Sites 

2.1.1.1 Quebrada Jaguay 

 

Robust evidence for a Terminal Pleistocene (TP) coast-highland connection first 

emerged at Quebrada Jaguay 280 (QJ-280), a maritime site seven kilometers inland from a 

paleo-shoreline in southern Peru ( 

Figure 4) (Sandweiss et al. 1998). TP and Early Holocene (EH) levels at QJ-280 date between 

13,000 to 8,300 calibrated years BP (Jones et al. 2017; Sandweiss et al. 1998). Both TP and EH 

strata contain subtypes of Alca obsidian sourced ~140 kilometers north to the high-altitude 

core of the Andes Mountains (Sandweiss et al. 1998; Rademaker et al. 2013) ( 

 

16 

 

 

Figure 4). Other non-local artifacts at the site include petrified wood from the Upper Moquegua 

formation ~30 kilometers to the north and Opuntia ficus-indica (prickly pear cactus) seeds, 

which derive from a mid-altitude zone in the Peruvian Pampas (Rademaker 2017; Rademaker 

2012; Tanner 2001). The faunal assemblage at QJ-280, which includes marine shells, fish 

vertebrae, and crab, supports the notion that the people who used the site effectively exploited 

marine resources.  

 

To assess the time of year QJ-280 was occupied, the time of harvest was determined for 

a sub sample of the shells comprising the site middens. δ18O and δ13C were measured 

incrementally from the growth bands of 36 marine shells to reconstruct seasonal sea surface 

temperatures. The analysis provided a means to determine what time of year the shells were 

harvested. The data suggest occupations may have been restricted to the austral spring and 

summer months between February to April (Gruver 2018).These seasonality data are consistent 

with the summer occupations proposed by Sandweiss (1998). Sandweiss hypothesized that site 

use would depend on the availability of freshwater in the area. In the present day, water is only 

observed in the quebrada bed in austral summer. The seasonal flow emerges and dissipates 

with the wet season in the high Andes. If the timing of modern water flow reflects the annual 

availability in the past, it is possible that freshwater resources may have limited tenure at QJ-

280.  

 

Where groups resided outside of this period remains unknown. It is possible that the 

inhabitants of QJ-280 networked along the coast in a year-round occupation of a littoral 

resource patch. Alternatively, coastal groups may have moved inward to exploit the resources 

of interior zones (Gruver 2018; Rademaker 2017; Sandweiss et al. 1998). In the latter scenario, 

 

17 

 

 

the non-local material at QJ-280 points to either mid- or high-altitude zones as potential 

supplemental habitats.  

 

In 1998 Sandweiss argued convincingly that the Terminal Pleistocene dates, seasonal 

nature of water flow, and presence of Alca obsidian at QJ-280 supported inter-zonal mobility 

model of settlement where people moved between the highlands and the coast. While others 

have incorporated the site in discussions of the coastal migration theory, their statements are 

generally tentative, given the paucity of South American costal sites dated to the Pleistocene 

(Braje et al. 2019; Dillehay et al. 2008; Erlandson 2001; Erlandson et al. 2011).   

 

 

A revised chronology from QJ-280 reveals that regionally, coastal occupations may have 

been contemporaneous with or younger than those established for the highlands (Rademaker, 

personal communication). The new data place the Terminal Pleistocene occupations between 

12,400 to 11,500 calibrated years BP and the Early Holocene deposits between 9,700 to 8,200 

calibrated years BP (Rademaker and Gruver 2018). These dates support a reconsideration 

Sandweiss’ original model. It is possible that transhumance between elevation zones followed 

an established migration in the interior. Understanding the settlement systems of the first 

Andeans will require comparing seasonal evidence from contemporary sites in other locations. 

 

18 

 

 

 

Figure 4: Overview of Quebrada Jaguay 280. View to the north during the dry season. 

2.1.1.2 Pampa Colorada 343 

 

Fifteen kilometers west of the Quebrada Jaguay drainage, Pampa Colorada 343 (PC-343) 

is another early maritime site with raw materials from the interior ( 

Figure 5). Two radiocarbon dates on shell reveal an EH age between ~8,920-8,265 calibrated 

years BP (McInnis 2006). PC-343 contains expansive shell middens and intermittent human 

burials. Materials at the site include diverse fauna such as surf clams, fish, chiton, and bird; and 

an assortment of lithic materials including scrapers, awls, projectile points, and debitage. Like 

Quebrada Jaguay 280, stone tools of petrified wood and Alca obsidian suggest the occupants at 

PC-343 were familiar with groups or landscapes beyond the coast. For this site, an isotopic 

study of the human remains would provide the most accurate estimate of human movement 

 

19 

 

 

among zones. Contextual evidence dates the human remains to the Early Holocene. If accurate 

these individuals represent some of the oldest human burials on the Peruvian coast. 

Comparably ancient highland and lowland burials would provide the most direct means of 

studying the difference in lifeways between the Andes and Pacific coast. As of yet, no stable 

isotopic analyses have been conducted to determine the provenance of the PC-343 human 

remains.  

 

Figure 5: View of the shell middens at Pampa-Colorada 343. View to the northeast during the dry 

season. Photo curtesy of Kurt Rademaker.  

2.1.2 A High-Elevation Comparison 

 

The Alca obsidian recovered at Quebrada Jaguay 280 and Pampa Colorada 343 led 

Rademaker (2006) to search the surrounding highlands for outcrops of obsidian and potential 

 

20 

 

 

early sites. Subsequent survey resulted in the characterization of five Alca sub-sources and nine 

highland hunter-gather sites (Rademaker 2012, Rademaker et al. 2013). One site, Cuncaicha, 

revealed intact strata, rich archaeological deposits, and Terminal Pleistocene (TP) and Early 

Holocene (EH) dates comparable to the earliest occupations at QJ-280.  Analyses at Cuncaicha 

are ongoing, as discussed below. 

2.1.2.1 Cuncaicha 

 

The Cuncaicha rockshelter (4480 masl) is located ~150 from the coast in the Pucuncho 

Basin, Peru (Rademaker, Bromley, et al. 2013; Rademaker et al. 2014) (Figure 6).  A robust 

chronology from 38 accelerator mass spectrometry (AMS) bone collagen dates brackets the TP 

and EH occupations with 23 dates spanning between 12,600-11,200 and seven dates between 

9500-9000 calibrated years BP (Rademaker and Hodgins 2018). The composite assemblage 

includes rich deposits, spanning over ten thousand years, of assorted stone tools and animal 

bones, and rare objects such as crystals, seashell, and ochre. Such persistence of place at 

Cuncaicha suggests ongoing importance of the Pucuncho Basin for Andean peoples, from initial 

exploration to contemporary populations (Francken et al. 2018; Rademaker and Moore 2018). 

For the TP assemblage alone, 100 formal tools, nearly 15,000 pieces of debitage, 6.5 kilograms 

of faunal material, and ochre show that diverse activities took place at the site from the earliest 

occupations (Rademaker and Moore 2018; Rademaker et al. 2021 In Prep).  

 

21 

 

 

 

Figure 6: Cuncaicha rockshelter archaeological site on the Andean Plateau. View to the south during the 

dry season. Nevado Coropuna is visible to the east behind the shelter.  

 

The composition of Cuncaicha’s assemblage changes through time. In stark contrast to 

the deposits at the coast, the Terminal Pleistocene levels of Cuncaicha contain no readily 

identifiable non-local materials from a different ecological zone. Most of the TP artifacts are 

local to the surrounding area or sourced to a nearby highland location (Rademaker et al. 2021, 

In Prep). Further, the presence of bone needles and lithic scrapers and awls suggest 

technological adaptions suited to the cold environment. The local origin and design of the 

artifacts suggests the humans occupying the site either had expansive knowledge of the 

regional ecology upon arrival or were quick to learn the surrounding resource structure.  

 

22 

 

 

 

The Early Holocene assemblage expands to include non-local raw materials from lower 

elevations, such as marine shells and petrified wood from the Upper Moquegua formation 

(Rademaker pers comm). The local EH artifacts indicate the sustained use of the highlands in 

the Terminal Pleistocene continued through the Early Holocene. These materials may reflect 

emerging trade networks or expanding resource forays.  

 

In 2015, during the fourth and final season of excavations at Cuncaicha, five human 

burials were recovered. Three EH burials, dating between 9,300-8000 calibrated BP, are 

contemporaneous with the burials found at PC-343 and represent some of the oldest known 

and highest-altitude human remains in the world (Francken et al. 2018; Rademaker and 

Hodgins 2018).  A weighted mean of three dates for each EH Burial place the interments around 

9133-8788, 8986-8691, and 8536-8386 calibrated years BP. The remaining two individuals date 

to the Late Holocene, around 4287-4085 and 3361-3182 calibrated years BP. The expansion of 

raw material types and the interment of human remains signifies a change in site-use starting in 

the Holocene. These records reflect emerging mobility strategies or trade relations, familiar 

with lower elevation resources and people. 

2.1.3 Outstanding Research at Cuncaicha 

2.1.3.1 Site Seasonality  

 

Seasonality at Cuncaicha was assessed using palaeobotanical and zooarchaeological 

measures. Macro botanical methods have identified organic remains that are only available 

when plants flower during the wet season (Furlotte, In Progress). Tooth eruption data from 

Cuncaicha’s juvenile fauna corroborate the botanical assessment of a wet season occupation, 

 

23 

 

 

placing site occupation between March-April (Moore 2016b). Faunal measurements of tooth 

eruption and wear have been applied successfully in the northern Junín region of the Central 

Andes to accurately identify the seasonality and duration of the earliest occupations at 

Panaulauca (Moore 1998). However, applying the method at Cuncaicha may reveal less 

accurate results due to the small sample size and differences in regional seasonality and animal 

behavior (Katherine Moore, personal communication). As such, Moore (2016) proposed an 

isotopic study to determine the time of harvest for the adult animals at the site. Such an 

approach exists in the stable isotopic literature (Balasse et al. 2002; Balasse et al. 2003; Eerkens 

et al. 2014; Green et al. 2018; Knudson et al. 2007; Koch et al. 1989) but will require species- 

and location-specific developments.  

 

Preliminary faunal analyses conducted in 2012, 2015, and 2018 (Kate Moore, personal 

communication) revealed three primary species of suitable fauna: vicuña, guanaco, or taruca. 

Vicuña appear the best suited for a new seasonality proxy for four reasons: (1) vicuña compose 

85% of the faunal assemblage, suggesting they were the primary prey for the residents of 

Cuncaicha (Moore 2013, 2016b); (2) vicuña maintain territories at high-altitude throughout the 

year; (3) vicuña are obligate drinkers and ingest water and plants directly from their 

surrounding habitat (Hillson 2005; Moore 2013, 2016a, 2016b; Tomka 1992); and (4) vicuña are 

the only known ungulate with ever-growing incisors (Figure 7). If there is isotopic variability 

between seasons vicuña incisors should (A) record local isotopic seasonality and (2) reveal an 

isotopic signal that records multiple months preceding their death–enabling an accurate 

estimate of the time of year they died and by proxy, the season of human activity in the area. A 

preliminary scan of the Cuncaicha faunal materials by Milton and Moore in 2018 suggests 

 

24 

 

 

suitable archaeological material may be available for the TP (pending a secure date on dentine) 

and are abundant for the EH. The first stage of developing a new proxy necessitates the 

establishment of a seasonal isotopic baseline for the local environment. Because of the wet and 

dry nature of seasons in the region, such data collection will need to target local surface waters. 

Figure 7: Cross-section of an adult vicuña incisor showing parallel enamel growth lines characteristic of 

an ever-growing tooth. 

 

2.1.3.2 Human Provenance 

 

The introductory section 1.2 reviewed that 87Sr/86Sr and δ18O analyses have been 

applied to the human teeth of the Cuncaicha humans to determine the childhood provenance 

of each individual (Chala-Aldana 2017). In addition to these analyses, three bones from each of 

the five individuals were isotopically analyzed for δ13C, δ15N, and δ34S as a means to assess the 

locality and composition of their adult diet (Haller von Hallerstein 2018). Bone remodels 

throughout the life of an individual, therefore isotopic values sampled from bone reflect the 

last 7-10 years of an individual’s diet (DeNiro and Epstein 1978; Passey et al. 2005; Sponheimer 

 

25 

 

 

and Cerling 2014; Tanz and Schmidt 2010). Comparing both the adolescent values derived from 

teeth and later life stages recorded in bone could reveal whether the humans of Cuncaicha 

represent individuals adapted to and persisting at high-altitude or mobile individuals who 

transitioned between different ecological zones and lived at multiple elevations.   

 

For the dietary analysis, δ13C, δ15N, and δ34S were measured on bone collagen from the 

humans remains, as well as a small sample of modern and archaeological fauna. By sampling 

fauna restricted to high-altitude ecosystems the researchers were able to characterize a local 

and elevation-restricted dietary signal. The initial assays suggest that the humans buried at 

Cuncaicha consumed diets consistent with local resources, including viscacha and camelid, 

during their adult years of life (Haller von Hallerstein 2018).  

 

For both the dietary and provenance studies, some uncertainties for local isotopic 

patterning remain. Confirming the locality of the Cuncaicha humans during adolescence will 

require revised interpretations based on isotopic data compiled from the local environment. 

This research provides the first phase of this work by establishing the δ18O values of surface 

waters in the immediate elevation zones surrounding Cuncaicha.  

2.1.4 Section Summary 

 

Obsidian and petrified wood in the Terminal Pleistocene and Early Holocene 

components of the coastal sites Quebrada Jaguay 280 and Pampa Colorada 343 link the coast to 

interior ecosystems. Seasonality and provenance data from Cuncaicha could help to resolve 

site-specific questions. By comparing the data from Cuncaicha with seasonality and provenance 

data from the coast, it might be possible to reveal the regional settlement strategies active 

 

26 

 

 

during the Terminal Pleistocene and Early Holocene. While both studies have been initiated, 

seasonality data are incomplete, and the provenance study uncertain. For both studies a 

seasonal and elevation oxygen isotope baseline, constructed from the available surface waters 

linking the coastal and highland sites, as presented herein, will provide the data necessary to 

enter the next phase of research. The next section reviews the geologic and ecologic make-up 

of that study area.  

2.2 Study Area  

 

The Cuncaicha rock shelter is located in the Pucuncho Basin ~150 kilometers northwest of 

the city of Arequipa (Figure 8). The basin is a stable landscape of water, fauna, and flora and is 

separated from lower-elevation topography. At the eastern extent of the basin, the north-

facing rock shelter sits atop a hill, overlooking a flat plain and the perennially flowing Quebrada 

Cuncaicha below.  

 

27 

 

 

Figure 8: Map of Cuncaicha rock shelter, Pucuncho Basin, and Alca-1, 4, and 5 sources. 

 

28 

 

 

 

2.2.1 Local Geology 

 

Pucuncho is a high-altitude (~4300-4600 masl) basin at the northernmost edge of the 

Altiplano in southern Peru. The basin appears as a flat plain in the puna ecological zone, 

between three prominent glaciated volcanos: Nevado Firura (5498 masl; 15º14’S, 72º48’W) to 

the north, Nevado Solimana (6093 m.a.s.l; 15 24’S, 72º52’W) to the west, and the Miocene-age 

Nevado Coropuna (6405 masl; 15º 33’S, 72º 93’W) to the south (Bromley 2010). The prominent 

stratovolcano, Nevado Coropuna was last active during the Early Holocene. Nevados Solimana 

and Firura are both considered extinct (Venturelli et al. 1978).  

 

Volcanic and glacial processes characterize the geomorphology of the basin. The basin is 

situated atop an ignimbrite plateau. Initial geologic formation involved volcanic deposits of 

andesite and basalt throughout the Miocene and Pleistocene. Pliocene-age Grupo Barroso 

flows compose the area around Cuncaicha, including the shelter itself  (Galaś 2011). The area 

surrounding Cuncaicha has lacustrine and fluvial features. In front of the site, to the north, is a 

Pleistocene-Holocene alluvial plain. 

 

Extensive glacial activity during the Pleistocene formed the present-day landscape. 

Moraines, boulders, and glacial outwash (sandy gravel and rock) characterize the area 

surrounding the shelter. During the Pleistocene, the glaciers of Coropuna, Solimana, and Firura 

extended 9-12 kilometers further from the termini of present day (Bromley 2010; Bromley, Hall, 

Rademaker, et al. 2011; Bromley, Hall, Schaefer, et al. 2011). Ice from all three mountains 

reached elevations as low as 4540 meters, approximately 1000 meters lower than the present-

day elevations, between ~20-25 thousand years ago (Bromley 2010; Bromley et al. 2011). 

Helium and Beryllium dates from the terminal moraines indicate glaciers began retreating 

 

29 

 

 

shortly thereafter, well before the arrival of humans (Bromley et al. 2009; Rademaker et al. 

2014). 

2.2.2 Hydrology  

 

From west to east, the study area encompasses the Ocoña (15,459.9 km2), Jaguay 

(1,064.3 km2), and Majes (16,935.7 km2) watersheds (Figure 9). Freshwater on the plateau 

consists primarily of seasonal springs and perennial bofedales (wetlands), quebradas (streams), 

and ríos (rivers) (Figure 10). Seasonal precipitation and underground aquifers feed the local 

surface waters (Fonkén 2014). Though the surrounding mountains are glaciated, field 

inspection has revealed minimal direct meltwater inputs into the basin regardless of season 

(Bromley and Rademaker 2018, personal communication). Highland springs and streams 

concentrate into rivers before discharging over the plateau rim. East of Coropuna waters drain 

to the Colca canyon and discharge as the Río Majes. At lower elevations, water bodies comprise 

small order streams and large rivers; canals divert waters to surrounding towns. 

 

30 

 

 

 

Figure 9: Map of Ocoña, Jaguay, and Majes watersheds. 

 

The Pucuncho Basin lies in the Ocoña watershed. The basin shows a similar configuration 

to the perched wetland identified by Sitzia et al. (2019). Sitzia et al. describe a high-altitude 

wetland on an impermeable ignimbrite substrate. The geology results in a near-surface water 

table and perennial wetlands. The comparable surface geology of the Pucuncho Basin to Sitzia’s 

 

31 

 

 

study might explain the abundance of perennial freshwater near the surface. Waters from the 

Pucuncho Basin merge at the Río Arma before discharging over the plateau edge. Below the 

plateau, the Río Arma becomes the fifth-order Río Ocoña, which empties into the Pacific. 

 

Figure 10: The prominent water bodies of the study area. These include high-altitude springs (upper 
right, dry season) and wetlands (upper left, wet season); as well as canals (lower left, dry season) and 

rivers and streams (lower right, dry season) found at all elevations. 

2.2.3 Highland Ecology 

 

At 15ºS, low seasonal fluctuations in temperature and the perennial availability of water 

support a stable high-altitude ecological zone, known as the puna. The puna zone begins 

around 3200-3700 masl and ends below snow line (5000+). Latitudinally the puna zone extends 

 

32 

 

 

from approximately 10-30ºS latitude (Troll 1968). Puna habitats comprise grasses and shrubs. 

The wetter latitudes in the north are warmer and wetter, demonstrating greater primary 

productivity. The arid dry and salt puna to the south receive diminished annual precipitation. 

During the first human occupation at the end of the Pleistocene, wetter conditions may have 

extended a sense of fixed habitats, like those of the wet and dry puna, to the southern salt 

puna (Blard et al. 2011). 

2.2.3.1 Fauna 

 

Local fauna includes medium-bodied camelids, deer, Hippocamelis antisensis (taruca), the 

small mammal Lagidium peruanaum (viscacha), and wild birds. Today, four camelid species 

inhabit the puna: these include the wild Vicugna vicugna (vicuña) and Lama guanicoe 

(guanaco), and the domesticated Lama glama (llama) and Vicugna pacos (alpaca)(Moore 

2016a). Only the wild camelid species would have been present during the Terminal Pleistocene 

and Early Holocene, and important distinctions exist between them. The larger-bodied guanaco 

is dry-adapted, allowing it to move freely between ecological and altitudinal zones (Tomka 

1992). The smaller vicuña (Figure 11) is restricted to high-altitude >3200 masl and is an obligate 

drinker, making it unsuited to occupy arid environments (Franklin 1983; Moore 2016a).  

 

Genetic evidence suggests that vicuña populations rapidly expanded on the Andean 

Plateau at the end of the Pleistocene(Marín et al. 2007; Marín et al. 2017). This may have 

resulted from climatic amelioration, as abundant pasturage and charged aquifers would have 

allowed vicuña to range into the lower latitudes of the modern dry and salt punas (Blard et al. 

2011). Annually fed surface waters have allowed vicuña herds to occupy and maintain highland 

territories year-round (Marín et al. 2017; Moore 2016a; Yann et al. 2016).  

 

33 

 

 

Figure 11: A herd of vicuña on the Andean Plateau, posed in front of Nevado Coropuna. Image shows 

llareta cushion plants and bunch grasses during the dry season. 

 

2.2.3.2 Flora 

 

The flora of the puna includes shrubs, grasses, mosses, and cacti. Microhabitats vary 

across space because of changes in soil type, elevation, and geographic position.  Furlotte 

(2018) prepared an extensive, site-specific report for Cuncaicha and the surrounding Pucuncho 

Basin. Furlotte identified multiple perennial grasses, woody shrubs, and cushion plants. 

Common grasses include ichu bunch grass including Festuca, Stipa, and Calamagrostis; and 

small-to-medium shrubs such as Parastrephia lucida and Parastrephia quadrangularis (Szpak et 

al. 2013). The cushion plant, Azorella compacta, common name, llareta, is the most probable 

fuel for the inhabitants of Cuncaicha (Figure 11).  To the west, the quenual tree may have 

 

34 

 

 

served as a second important resource. The modern elevation limit for local quenual falls 

between 4000-4200 masl. 

2.2.4 Section Summary 

 

The surface waters of the study area transect three prominent watersheds populated by 

various water body types. Available waters support plants and animals in a high-altitude 

ecological zone, the puna, before discharging southwest to the Pacific coast. The next section 

reviews the principles of stable isotopes and outlines the specific controls for oxygen isotopes 

in the region.  

2.3 Stable Isotopes  

 

An isotope refers to the different atomic masses of a single element. The variation in 

weight is driven by the number of neutrons in the atom's nucleus, which is denoted by the 

number to the upper right of an element (Kendall and Caldwell 1998). Unlike radioactive 

isotopes, stable isotopes do not decay over time. Different elements occur in predictable ratios 

throughout the global environment. Typically, the most stable weight of an isotope is found in 

the greatest abundance (Kendall and Caldwell 1998). The relative abundances of certain 

isotopes are controlled by specific geologic, biologic, or atmospheric processes (Luz et al. 2009; 

Luz et al. 1984). Fractionation processes result in the enrichment of one stable isotope over 

another, where lighter isotopes react more readily to kinetic processes. The processes driving 

fractionation differ between elements. 

 

35 

 

 

2.3.1 Stable Oxygen 

 

There are three stable forms of oxygen atoms, each containing eight protons and 

between eight and ten neutrons (Kendall and Caldwell 1998). Oxygen-16 (16O) constitutes 

99.76% of available oxygen, oxygen-18 (18O) 0.2%, and oxygen-17 (17O), the rarest, comprises 

only 0.04%. The standard notation for oxygen, δ18O, expresses the measured ratio of heavy to 

light isotopes in a sample, 18O/16O, as compared to a standard (Craig 1961b). The global 

standard for oxygen isotopes is the Vienna Standard Mean Ocean Water (VSMOW). 

 

 18O varies with meteoric processes including distance from source, latitude, altitude, 

aridity, temperature, and precipitation (Craig 1961a; Dansgaard 1954, 1964; Gat 1996, 2000).. 

Oxygen fractionation occurs during multiple phases of the water cycle including evaporation, 

transportation, and condensation. Craig (1961a) demonstrated a linear relationship exists 

between δD and δ18O in precipitation, known today as the meteoric water line (MWL). Building 

on Craig’s research, Dansgaard (1964) showed that the δ18O of precipitation increases in value 

with increasing temperature and vice versa.   

2.3.2 Oxygen Isotopes for Archaeological Research  

 

The predictability of δ18O across a landscape allows for reconstructions of human and 

animal movement and regional assessments of short- (seasonal) and long-term (decadal) 

climate change (Bowen 2010; Bowen and Revenaugh 2003; Bowen and Wilkinson 2002); 

principles adapted for archaeological research. Living terrestrial organisms assimilate the 

isotopic ratios of their local ecosystems via eating, breathing, and drinking (Passey et al. 2005; 

Sponheimer and Cerling 2014). Advances in paleontological research have revealed the 

 

36 

 

 

potential of using organic and inorganic human and animal tissues as records of ecological 

change (Bocherens et al. 1996; Cerling and Harris 1999; Cerling and Sharp 1996; Drucker et al. 

2008; Fry 1991; Iacumin et al. 2000; Koch et al. 1998), diet (Bocherens et al. 2016; Cerling and 

Harris 1999; DeNiro and Epstein 1978; Díaz-Zorita Bonilla et al. 2016; Drucker and Bocherens 

2004; Passey et al. 2005; Schoeninger and Moore 1992; Somerville et al. 2015; Sponheimer and 

Cerling 2014), health and life history (Smith et al. 2018), and geographic movement (Balasse et 

al. 2002; Chala-Aldana 2017; Dufour et al. 2014; Knudson 2009; Knudson et al. 2009; Price et al. 

2002; Santana-Sagredo et al. 2018; Zazzo et al. 2006).   

 

Metabolized oxygen isotopes are incorporated into tooth and bone apatite (Lee-Thorp 

and Sponheimer 2003). The high organic content of bone is more susceptible to diagenic 

change and heating alterations than the mineral portion of teeth, therefore bones are rarely 

used for stable oxygen analyses (DeNiro 1985; Lee-Thorp 2002; Lee-Thorp and Sponheimer 

2003; Schoeninger et al. 2003; Schoeninger and Moore 1992; Schwarcz and Schoeninger 1991; 

van Klinken 1999; Zazzo et al. 2013). The highly mineralized nature of tooth enamel leaves it 

resistant to diagenic processes, making it the preferred tissue for stable oxygen analysis 

(Bocherens et al. 2011; Lee-Thorp et al. 2003).  

 

The use of oxygen isotopes in archaeological provenance research in the Andes is a 

relatively recent development (Knudson 2009).  Oxygen isotopes are assumed to reflect 

regional seasonal or spatial variations in imbibed waters, as humans and animals primarily 

consume oxygen via drinking (Longinelli 1984; Luz et al. 1984). Archaeologists estimate local 

climate, migration between areas, and provenance during early life using the δ18O of human 

tissues as a proxy for drinking water (Daux et al. 2008; Kennedy et al. 2011; Luz et al. 1984).  

 

 

37 

 

 

2.3.2.3 Analytical Interpretations  

 

Culture and biology present challenges to interpreting the oxygen values derived from 

human tissues.  Water storage, rainwater collection, and preparation techniques (boiling) would 

all alter the composition of local waters and result in a value that does not exactly match the 

local environment. The metabolic processes contributing to the observed δ18O in biologic 

tissues varies with species and individual consumption. The formula adjusts for an observed 

enrichment in δ18O resulting from ingested food. Iacumin et al. (1996) developed a formula to 

convert δ18O enamel carbonate (C) to δ18O phosphate (P). The formula is printed in Equation 1.  

Equation 1: Conversion for Enamel Carbonate to Phosphate 

δ18OP = 0.98 δ18Oc –8.5 

Daux et al. (2008) developed a formula to convert the δ18O of human phosphate (δ18Op) to 

δ18OW (drinking water). The formula is provided in Equation 2.  

Equation 2: Conversion for Enamel Phosphate to Drinking Water 

δ18OW = 1.54 (±0.09) * δ18OP – 33.72(±1.51) (DAUX) 

2.3.3 Section Summary  

 

Stable oxygen values from teeth should reflect the regional landscape, but interpreting 

those values requires environmental mapping and considerations of both cultural adaptations 

and species biology. A baseline needs to contend with both spatial change across elevations in 

the region, as well as seasonal change driven by climate. The next section reviews isotopic 

inputs to the study area to support the predictions for oxygen isotope patterning in the region.  

 

38 

 

 

2.4 Stable Isotopic Inputs to the Study Area 

 

The study area extends from the Pacific coastal sites, Quebrada Jaguay 280 and PC-343, 

transects through the Majes Valley, and ends in the Pucuncho Basin on the Andean Plateau. 

The two research objectives include a stable oxygen elevation baseline for provenance 

analyses, as well as a preliminary seasonality baseline data for the surface waters of the 

plateau. This section reviews the climate literature to set predictions for both studies.  

 

The temporal and spatial controls of stable isotopes in the environment include (1) input 

source(s) and (2) fractionation processes. To develop a targeted sampling strategy for the 

surface waters of southern Peru, I identify potential input source(s) of stable oxygen to the 

study area, the temporal variability of these sources, and the fractionation effects determining 

the δ18O of Andean surface waters. I briefly review the paleoclimate literature to explore the 

suitability of a modern baseline for accurately interpreting materials accumulated at regional 

archaeological sites over the past ~13 thousand years. 

2.4.1 Input Sources for the Pucuncho Basin and Western Slope 

 

The isotopic composition of a substrate reflects various geophysical processes or 

fractionation effects. Water molecules, composed of hydrogen and oxygen, are sensitive to 

atmospheric and hydrological processes. Seasonal and source inputs are important controls for 

the isotopic composition of surface waters in the study area. Seasonal atmospheric change 

manifests as a fluctuation in temperature or precipitation. Oxygen should therefore record the 

change between wet and dry seasons in the study area.  

 

39 

 

 

 

Next, it is important to control for the fractionation processes that control the isotope. 

Multiple sources or fractionation may complicate analysis. As a regional example, competing 

stable oxygen inputs from the Pacific and Atlantic could attenuate a seasonal signal in the high-

Andes.  

 

Climate variability in the study area reflects the specific air circulation active in the 

geographic location. South American climate encompasses tropic, subtropic, and extra-tropic 

zones; the seasonal, annual, and inter-decadal climate regimes vary between regions (Bershaw 

et al. 2010; Blard et al. 2011; Garreaud et al. 2003). The Andes pose a distinct orthographic 

barrier to atmospheric circulation and moisture transport, especially between the latitudes of 

0-40ºS where the Andean mountain belt frequently rises above 4km in elevation (Garreaud et 

al. 2009; Vimeux et al. 2009).  

 

The geographic focus for seasonality is the Pucuncho Basin, a high-altitude, tropical 

basin at 15ºS, the northern edge of the Altiplano. The Altiplano is a meteorologically distinct 

region of the Central Andes between 15-22ºS where the Andes reach their widest lateral extent 

(Garreaud et al. 2003). Often referred to as the Andean Plateau, the elevation averages >3,700 

masl, constituting the highest, uninterrupted territory outside of the Himalayas. The tropics 

experience relatively stable temperatures year-round; precipitation constitutes the 

predominant mode of atmospheric change, and seasons are characterized as wet or dry (Vuille 

et al. 2012). In the Andean sub-tropics, diurnal changes in temperature are more significant 

than annual change (Thompson et al. 2000; Vuille and Bradley 2000).  

 

In the study area, weather station data from the Pucuncho Basin demonstrate small 

seasonal amplitudes in temperature, but considerable change in precipitation. Mean 

 

40 

 

 

temperature in sub-tropical Andes (>3500 masl)  is low (~7ºC) and varies between ~25ºC during 

the day, but only ~1-8ºC over the course of the year, whereas precipitation increases by as 

much as 800 mm during the wet season (Dornbusch 1998). Competing moisture sources could 

complicate predictions for both seasonal and spatial signals (Vuille et al. 2012). Two sources 

dominate the precipitation inputs for the South American continent: the Atlantic and Pacific 

Oceans. I discuss both below. 

2.4.1.1 Pacific Inputs 

 

Pacific moisture does not contribute much to the composition of water in the Andes 

(Abbott et al. 2003; Garreaud et al. 2003). Despite the proximity of the Pacific Ocean to the 

Altiplano, it is not the dominant input source of meteoric waters to the landscape. The 

topography of the western continent restricts lateral transport of Pacific moisture and an 

atmospheric temperature inversion at 800 meters above sea level inhibits vertical movement 

(Rutllant and Ulriksen 1979).The temperature inversion results from subsiding air masses in the 

southeast Pacific and is responsible for the hyper-arid conditions of the Atacama desert on the 

southwestern coast. Sporadically, dry air masses from the Pacific will reach the Altiplano. 

Herreros et al. (2009) identified Pacific marine aerosols and pollen from southern Patagonia in 

an ice core from Coropuna. Both aerosols and pollen are interpreted as chemical indicators of El 

Niño events rather than moisture transported from the Pacific. Together the regional 

temperature inversion and local topography suggest the Pacific does not contribute to the 

moisture in the Pucuncho Basin. This leaves the Atlantic Ocean as the dominant source of 

incoming stable oxygen to the study area.  

 

41 

 

 

2.4.1.2 The Atlantic Source 

 

The South American Summer Monsoon system (SASM) transports Atlantic moisture 

across the South American continent to the high Andes. The SASM is a prominent internal 

mode of climate variability in the southern hemisphere (Vuille et al. 2012). The monsoon 

originates in the west Atlantic, south of the Intertropical Convergence Zone (ITCZ) in the 

Atlantic (Garreaud et al. 2009; Garreaud et al. 2003). During peak activity, the SASM transports 

air masses west across the South American continent, distributing precipitation along the way. 

While fractionation effects spatial change in meteoric compositions of δ18O, various climate 

regimes impact temporal patterns of moisture transport to the area. 

2.4.2 Stable Isotopic Controls 

2.4.2.1 Fractionation  

 

Precipitation constitutes the primary source for Andean surface waters, though 

groundwater and meltwater may also contribute to annual flow. The isotopic composition of 

surface waters reflects upstream processes of air masses prior to final rainout in the 

depositional environment. A Rayleigh Distillation model is used to describe the isotopic controls 

of the water cycle (Vuille et al. 2012). Rayleigh fractionation refers to the preferential expulsion 

of heavier nuclides during condensation and lighter nuclides during evaporation (Dansgaard 

1964). For example, evaporated surface waters exhibit more enriched (higher) values than the 

depleted (lower) values of incoming precipitation. In the low-latitude tropics, the primary 

controls for the stable isotopic composition of water include altitude, distance, and rainout 

effects (Dansgaard 1954, 1964; Gat 1996, 2000). 

 

42 

 

 

In the tropical Andes, ! values of surface waters reflect the amount of isotopic rainout, 

 

or transport efficiency across the monsoon region. The amount effect refers to the decrease of 

stable isotopic values as rainout increases. The continental effect predicts that waters farther 

from the source (Andes) are lower than those closer to the Atlantic coast. Similarly, waters at 

higher elevations present lower values than those at low elevation, because rainout increases 

with elevation. In the research area, rainout increases with distance from the Atlantic source 

and with altitude as air masses move up the leeward side of the Andes (Garreaud et al. 2003). 

Therefore, the stable isotopic composition of precipitation becomes lower as it moves farther 

into the Altiplano. 

2.5 Climate Variability 

2.5.1 Annual Variability, Seasonality 

 

The South American Summer Monsoon is a seasonal phenomenon. Annually, increased 

solar forcing and the displacement of the ITCZ result in the enhanced lofting of air masses over 

the Andes and increased moisture transport across the continent (Garreaud et al. 2003). The 

SASM emerges around October, in the austral spring, reaches peak intensity between 

December and February, and subsides in April, late in the austral fall (Vuille et al. 2012). While 

active, the SASM pushes Atlantic moisture westward across the South American continent. 

During this period, northern latitudes and the eastern Amazon Basin receive abundant rainfall. 

Intra-seasonal oscillations force wet and dry pulses, typically lasting around two weeks (T~15 

days). In the Altiplano, precipitation varies on a northeast to southwest gradient, where 

locations closer to the source receive more moisture (Garreaud et al. 2003; Vuille et al. 2012).  

 

43 

 

 

 

The Pucuncho Basin experiences a wet season during the mature phase of the SASM. 

Seventy percent of local precipitation (600-1000mm) collects between late December to mid-

to-late March.  The Online Isotopes in Precipitation Calculator (OIPC) predicts the incoming 

isotopic value of precipitation for the study area (Figure 12) (Bowen 2017). δ18O in precipitation 

is lower during the wet season months and higher during the dry season. As a reflection of local 

precipitation, surface waters in the Pucuncho Basin should become lower during the wet 

season as depleted incoming precipitation replenishes the local water table. In contrast, surface 

waters should demonstrate lower values during the dry months as discharge lowers the water 

level and evaporation gradually removes lighter isotopes (Bowen 2017). 

Figure 12: Predicted seasonal variation in !18O in precipitation for the Pucuncho Basin based on OIPC 

 

modeled values for water sampling locations of Chala-Aldana et al., 2018. 

 

44 

 

 

2.5.2 Inter-Annual Variability 

 

The primary inter-annual mode of climate variability in the area is the El Niño Southern 

Oscillation (ENSO) phenomenon. ENSO is an internal mode of climate variability. It manifests as 

multi-year reversals in sea surface temperatures and atmospheric circulation across the tropical 

and subtropical Pacific. ENSO effects global climate (Garreaud et al. 2009). Climate records 

suggest the Altiplano is sensitive to significant ENSO activity. The moisture supply to the 

Altiplano is driven by upper-level atmospheric circulation, which correlates with changes in 

Pacific sea surface temperature (SST) driven by latent heat flux (Vimeux et al. 2009). The 

relationship between Pacific SSTs and Atlantic precipitation is indirect (Garreaud et al. 2003). 

 

During ENSO, the Altiplano mirrors the response of the west coast; El Niño events warm 

the troposphere leading to decreases in precipitation, while La Niña events result in cooler and 

wetter conditions (Vimeux et al. 2009). Rainfall and ice core data collected between 1955 and 

2003 indicate the reverse conditions on the Altiplano during significant El Niño events. In 1982-

3 and 1992, El Niño conditions resulted in a 48-100% decrease in precipitation. Similarly, an ice 

core from the saddle of Coropuna showed an enrichment of local pollen and dust for the 1982-

3 and 1992 events. However, during the 1998 event, five out of seven local meteoric stations 

recorded a 5-51% increase in the amount of precipitation (Silverio and Jaquet 2012). The rain 

gauge data indicate a strong ENSO event might temporarily (~T=1-2 years) attenuate a seasonal 

signal driven by incoming precipitation.  

 

45 

 

 

2.6 Proposed Study and Predictions 

 

I have identified two outstanding data requirements for interpreting the archaeological 

at Cuncaicha (1) an exacting provenance reassessment of Cuncaicha’s human burials and (2) the 

development of a stable isotopic method to study records of seasonality in the highlands.   

 

Oxygen isotopes have been proposed as an excellent candidate for both studies for three 

practical reasons: (1) direct and daily ingestion via drinking will reliably capture changes in 

environment and behavior through time, (2) reliable consumption by both humans and animals, 

and (3) preservation in the mineral component of tooth enamel will results in reliable 

preservation. For the study region, oxygen isotopes are bioavailable to humans and animals in 

the surface waters between the coast and highlands. A dataset constructed from observed 

values remains to be developed. 

 

As the first phase of research, I propose to construct this baseline from surface water 

samples collected between the Pacific coast and the high altitude Pucuncho Basin in southern 

Peru. I offer predictions for seasonal and elevation-based results.  

2.6.1. Predictions 

 

Two hypotheses underlie the expected outcome for an elevation transect and seasonal 

sampling. I back each with a set of predictions that lead me to this hypothesis.  

2.6.1.1 Andean Surface Water Seasonality 

Hypothesis 1: the ! 18O of surface waters in the study area will demonstrate lower values 

during the wet season (Dec-April) and higher values during the dry season.  

 

46 

 

 

 

 

I used the oxygen predictors in precipitation calculator (OIPC) to estimate the 

contributions of ! 18O in precipitation to Pucuncho Basin surface waters during the wet season 
(Figure 12). The OIPC suggests ! 18O in precipitation will be 11‰ lower in the wet season; 
therefore, ! 18O in surface waters should be lower during the wet season and higher during the 

dry season. The onset of the wet season is defined by the onset of precipitation. However, 

stream gauge data from the Servicio Nacional de Meteorología e Hidrología (SENHAMI) indicate 

 

 

regional waters do not register incoming rains until mid-January.  

I also predicted that surface waters would demonstrate an attenuated seasonal signal 

when compared to the OIPC, as surface waters average the incoming ! 18O in precipitation with 
existing ! 18O in surface waters.  I further predicted that lower order streams (smaller volumes 
of running water) should be more sensitive incoming ! 18O in precipitation.  

 

 

 

 

47 

 

 

 

 

 

Figure 13: SENAMHI stream gauge data for 2018-2019 in the Majes (top) and Ocoña (bottom) 

watersheds. Each graph shows volume of stream discharge (left axis) over time (bottom axis). Together 
these images show the complete onset, peak, and decline of the 2018-2019 wet and dry seasons in the 

study area. 

 

48 

 

 

2.6.1.2 The Elevation Transect 

higher at the coast—except where lower elevation waters represent large highland 

Hypothesis 2: the !18O in surface waters in the study area will be lower in the Andes and 
The altitude effect predicts that the !18O values of surface waters will become lower with 
increasing elevation. The amount effect predicts that the !18O values of meteoric waters will 

discharges.  

 

become lower with increasing distance from the Atlantic coast, as heavier isotopes rain out (the 

amount and continental effects). On the windward side of the Andes, both effects result in the 

depletion of meteoric waters. On the leeward side, the effects directly conflict until about 800 

meters above sea level (where Pacific inputs may complicate the signal). I predicted that the 

altitude effect would have a stronger control on surface waters along an elevation transect 

based on previous work by Mark and McKenzie (2007).  

 

However, I also predicted that larger water bodies (with greater discharge and faster 

moving waters) might demonstrate a less predictable pattern, as they are less susceptible to 

evaporation processes. 

 

 

 

49 

 

 

CHAPTER 3: METHODS 

3.1 Field Methods and Collection Procedure 

3.1.1 Elevation Sampling 

 

Surface waters were collected from perennial waters throughout multiple elevation 

zones between the Pacific coast and the highland Pucuncho Basin. The survey extent was 

determined by the location of three archaeological sites, Cuncaicha, Pampa Colorada-343, and 

Quebrada Jaguay (Figure 14). The transect represents approximately 150 kilometers from coast 

to highlands and nearly 5000 meters of vertical elevation.  

 

50 

 

 

 

Figure 14: Map of the Survey Area. The survey route (light pink line) extends from Pampa Colorada 343 
and Quebrada Jaguay 280 to Cuncaicha. Sampling crossed three watersheds, the Ocoña, Jaguay, and 

Majes Drainages. 

 

Below the plateau, sampling followed waters from the Ríos Majes and Blanco and 

human-made canals. Río Majes samples extended from the Camaná River Delta (36 masl) to the 

confluence of the Río Colca and Río Capiza (896 masl) (Figure 14). Between the Río Majes and 

the edge of the plateau (~1000 to 3400-3800 masl), most waters derived from the Río Blanco 

 

51 

 

 

and its associated canals (Figure 15). On the plateau, sampling included any viable drinking 

source for humans and animals. Waterbodies included springs, streams, wetlands, and rivers.  

Figure 15: Map of key sampling locations below the plateau. The confluence of Río Capiza (dashed 
yellow line) and the Río Colca (dashed blue line) mark the beginning of the Río Majes (blue area). The 
Río Blanco (dashed pink line) discharges over the plateau rim (dotted white line) around ~3500 masl.  

 

3.1.2 Temporal Sampling 

 

Samples were collected along the elevation transect during months that represent both 

the austral wet and dry season. Seasonal sampling included the collection of wet and dry 

 

52 

 

 

season waters over a broad area, and the testing specific water bodies during both seasons. For 

the general seasonal samples, Dry season samples include any waters collected between June 

to December, wet season samples represent February and March. Samples were not collected 

in January, April, or May. 

 

To test for perceptible seasonal change in surface waters, samples were collected to test 

daily, seasonal, and yearly variation. For each type of comparison, two or more samples were 

collected from a single water source at the same approximate location. Daily and inter-annual 

sampling was conducted to measure the magnitude of observable change in local surface 

waters.  

3.1.2.1 Daily Tests 

 

Daily tests included replicated ‘simultaneous’ samples, and one test to measure diurnal 

variation. Simultaneous pairs consist of two samples taken from a single water source at the 

same time during the same test. A diurnal test was taken for the same location in the morning 

and evening. 

3.1.2.2 Seasonal Tests 

 

Maximum seasonal variation was established through targeted, replicated sampling of 

easily recognized water bodies during a peak wet (March 2019) and dry (July 2018 or June 

2019) seasons.  

3.1.2.3 Yearly Tests 

 

Approximate yearly variability was determined by comparing samples taken from the 

same season (dry) during different years. These included one-, three-, and four-year 

 

53 

 

 

comparisons. One-year tests compared locations I sampled in both July 2018 and June 2019. 

Three- and four-year tests compared my July 2018 or June 2019 samples to published values 

from 2015 that derived from known, easily recognizable locations (Chala-Aldana et al. 2017).   

3.1.3 Sampling Procedure 

 

Waters were collected in 30 mL high-density polyethylene Nalgene bottles and sterile 

15mL polypropylene Corning tubes. Location data were recorded using a Trimble GeoExplorer II 

global positioning system (GPS) set to the Universal Transverse Mercator (UTM) coordinate 

system. Spatial accuracy for X, Y, and Z was <1.5 meters. Water collection followed International 

Atomic Energy Agency (IAEA) protocol (Introduction to Water Sampling and Analysis for Isotope 

Hydrology). Each container was labeled on the side and cap with the water sample number. 

Vials were rinsed in the local water, submerged and capped, then inspected. If an air bubble was 

identified, the sample was recollected. After collection, the vial was dried. Clear tape was 

placed over the label, and electrical tape around the seal. Each vial was then covered in foil and 

stored in a dark box in a cool location.  Location, date, time, temperature, relative humidity, 

weather, water body name (if known), and direction of flow was recorded for each sample. Each 

water source was photographed four times, once in each of the cardinal directions. Vials were 

sealed with electrical tape, covered in foil, and stored in a dark box. Samples were not 

permitted to freeze or sit in the sun at any point. All samples were successfully transported to 

the United States undamaged. 

 

54 

 

 

3.1.3.1 Labeling System  

 

I collected 90 samples during my fieldwork. I labeled each vial with a four-digit (YYXX) 

Water Sample (WS) catalog number. The first two numbers refer to the year I collected the 

sample (2018 or 2019 = YY). The last two numbers reflect the chronological order of sampling 

and provide the individual identifier (XX) for the sample. For example, water sample WS 1801 

was the first sample collected in 2018.  

 

A local highland family collected monthly five samples between August and December 

2018. I include the samples in the analysis but ultimately intend them for a separate study of 

seasonality not discussed here. The monthly samples received labels with the prefix ‘PU’ to 

separate them from transect samples. I labeled the vials with our highland partners for water 

collection. After the family collected each sample, they wrote the date of sampling under the 

number using permanent marker. Samples include 01, 03, 04, 05, and 06. Vial 02 was lost or not 

collected.  

 

Kurt Rademaker collected three samples in December 2018. These samples are labeled 

with his initials (KR) and two digits to indicate order of collection (01, 02, and 03).  

3.2 Laboratory Analysis of Oxygen and Hydrogen 

 

Stable isotopic analysis was conducted at the Environment and Natural Resources 

Institute Stable Isotope Lab (ENRI SIL) at the University of Alaska, Anchorage. Samples 1801-

1815, Pu-XX samples, 2018-KR-XX samples, and 1901-1926 were analyzed on May 5th, 2019. 

Samples 1927-1970B were analyzed on August 10th, 2019.   

 

55 

 

 

 

Water δ2H and δ18O isotope ratios were measured using a Picarro L2130-i WS-CRDS 

analyzer (Picarro, Sunnyvale, CA). Samples were screened to verify that organic content was 

low enough to minimize absorption interference spectra. Each sample was then analyzed six 

times and reanalyzed if the standard deviation of the six replicates was greater than 0.3‰ for 

(cid:0)18O and/or 3‰ for (cid:0)2H, or if the internal standard for the run differed from the accepted value 

by greater than ±0.2‰ or 2‰, for (cid:0)18O and (cid:0)2H, respectively. International reference standards 

(IAEA, Vienna, Austria) were used to calibrate the instrument to the VSMOW-VSLAP scale and 

working standards (USGS45: δ2H= -10.3 ‰, δ18O= -2.24 ‰ and USGS46 : δ2H= -235.8 ‰, δ18O= 

-29.8 ‰) were used with each analytical run to correct for instrumental drift. Long-term mean 

and standard deviation records of a purified water laboratory internal QA/QC standard (δ2H= -

149.80 ‰, δ18O= -19.68 ‰) yield an instrumental precision of 0.93 ‰ for δ2H and 0.08 ‰ for 

δ18O. Water measurements are reported in standard delta (δ) notation as the per mil (‰) 

deviation from Vienna Standard Mean Ocean Water (VSMOW). 

Equation 3: Delta Notation (Craig 1961b) 

! = ((Rsample/Rstandard) – 1) *1000 

R = ratio of heavy to light isotopes (18O/16O) 

3.3 Calculated Change for Temporal Tests 

 

The temporal dataset was divided into subtests based on sampling location. Samples for 

the temporal tests were grouped by test type (simultaneous, daily, seasonal, or inter-annual). 

Entries were ordered by elevation (lowest to highest) then chronological sequence.  

 

56 

 

 

 

The per mil δ18O difference between pairs was calculated to measure temporal change. 

To do this, he smaller of the two δ18O values in a pair was subtracted from the larger δ18O value. 

Once calculated, the difference values were compiled into a new table for analysis in RStudio. 

Targeted temporal tests were compared within groups (daily vs. daily) and between groups 

(daily vs. seasonal) to assess the time frame that results in maximum variation in surface water 

δ18O values.  

3.4 Software Analyses 

 

Measured δ18O and δD values were added to a geospatial database in a Geographic 

Information System (GIS) and mapped in ESRI ArcMap 10.6.1. Hydrological data were derived 

from the HydroSHEDS (Hydrological data and maps based on SHuttle Elevation Derivatives at 

multiple Scales) South American River Networks and HydroBASINS 30 second shapefiles. Data 

plots and statistical summaries were created in RStudio 3.3.2 with the ggplot2 and Tidyverse 

packages.  

3.4.1 Watershed and Stream Order Classifications 

 

Each location datum was assigned to the Ocoña, Jaguay, or Majes watersheds based on 

the HydroBASIN shapefile. Field observation was used to assign each water body as either 

human modified (HUM) or natural (NAT). Stream order classification (1-5) was assigned to each 

sample using the ORD_STRA field in the HydroSHEDS River Networks shapefile (Figure 16). 

Canal samples were assigned a Strahler order based on the nearest order of stream in the GIS. 

Stream order identifications are based on the Strahler classification system, where two streams 

of the same order to must merge to generate the next size class. The Strahler system reduces 

 

57 

 

 

the chance that unmapped or misidentified first order streams will lead to the resulting 

misclassification of higher order streams.  

Figure 16: Map provides an example of stream orders in the study area.  

 

 

58 

 

 

 

 

Chapter 4 reviews the results of the seasonal and spatial studies. Section 4.1 relays !18O 

CHAPTER 4: RESULTS 

 

results in terms of other research investigating meteoric waters in the region. Section 4.2 

reviews the results of temporal testing. Sections 4.3 and 4.4 explore the spatial distribution of 

surface water δ18O in the study area and examine these findings in terms of the archaeological 

research objectives for a provenance marker.  

4.1 General Results 

4.1.1 Results of Sample Collection 

 

The final sample set comprises 98 surface water samples collected over 12 months. 

Ninety-four samples were compiled as part of the elevation transect and temporal testing. Four 

samples represent the high-elevation samples obtained by local collaborators on a month-to-

month basis from September to December 2018. Together, the data represent approximately 

4000 square kilometers of area (Figure 17) collected during two dry seasons (July-December 

2018; June 2019) and one wet season (February- March 2019). Samples derive from the Jaguay, 

Majes, and Ocoña drainage basins at elevations between 36 and 4938 meters above sea level 

(masl).  

 

 

59 

 

 

Figure 17: Map of all sample locations within the survey area. Yellow triangles show dry season testing 

and blue circles indicate wet season sampling. 

 

 

60 

 

 

 

The initial (2018) dry season sampling was restricted to natural streams above 2500 

MASL, because the original project aimed to investigate only seasonal variation for high-

elevation waters. The 2018 dry-season dataset comprises 15 samples (1801-1815) collected 

between 3070 and 4896 MASL. These samples comprise areas on the western and northern 

margins of Nevado Coropuna. The lowest sample from the 2018 dataset came from the Colca 

Canyon (3070 MASL) in late July. It is the only 2018 sample gathered from the Majes 

Watershed.  

 

The four month-to-month samples derive from highland wetlands, or bofedales, which 

reflect the drinking-water source of a single alpaca purchased and monitored for a separate 

study. These samples were included in the general analysis. The precise sampling locations are 

unknown; therefore, the values are excluded from the more detailed seasonal and stream-class 

analyses. 

 

In 2019, the aims expanded to include identifying variation between seasons, in multiple 

elevation zones. The second field collection occurred from February 27th to March 1st, 2019, 

during the peak of the Andean wet season. Twenty-six samples (1901-1926) represent 

elevations between 383 and 4719 MASL. Fifteen of the transect samples constitute the puna 

zone at the southeast edge of the plateau. The other 11 locations reflect waters from below the 

plateau rim between the highland town of Chuquibamba (2950 MASL) and the lower Majes 

River valley town of Aplao (620 m.a.s.l).  

 

Humans have modified the surface waters below Andean Plateau. Waters from the Río 

Grande are redirected by canals to the terraces of the surrounding towns. A network of canals 

feeds these waters from Chuquibamba to Aplao. It was necessary to collect from canals to 

 

61 

 

 

capture a representative sample of the surface water variation between ~1000 and 3000 MASL. 

Canals only divert natural waters, so should record changes in elevation and season.  

 

The 2019 dry season transect comprises 49 samples (1927-1970B) collected over two 

days in June 2019. Collection during the final season extended from the mouth of the Majes 

River at the Pacific coast to the center of the puna. The total range of elevation in 2019 was 

between 36 and 4720 MASL. 2019 sampling targeted former wet and dry season sample points 

and under-sampled elevation zones. For example, the first two collection surveys lacked 

samples from elevations between 3000-4000 MASL. The June 2019 sampling transect targeted 

waters in this elevation range.  

 

A second focus to the June 2019 dry season included retesting the three lowland values 

omitted by Chala-Aldana (2017 et al. 2019), as discussed in Chapter 1. The samples in question, 

Out-wa-011, Out-wa-012, and Out-wa-013, derived from waters below 700 MASL and yielded 

results more depleted than the values predicted using the Oxygen Isotopes in Precipitation 

Calculator (OIPC).  Chala-Aldana also rejected three highland values from his analysis, Out-wa-

002, Out-wa-004, and Pu-wa-021, that were not retested. Each of these values represented 

saline or standing waters, which are not suitable drinking sources for humans.   

 

Figure 18 shows the density of samples collected between seasons and across 

elevations. The graph indicates a bimodal distribution for the dry-season sampling, with nodes 

at both lowest (0-1000 masl) and highest (4000-5000 masl) elevation ranges. The wet season 

plot is unimodal and skewed to the left. For both seasons most samples were collected from 

high elevation (>3500 masl), creating a bias. The data are also biased between seasons. The dry 

season (June-December) comprises 73% of samples.  

 

62 

 

 

 

Figure 18: Density of samples collected for wet (blue line) and dry (yellow line) seasons. The dry season 
shows a bimodal distribution of samples; most were collected below 1000 masl or above 3500 masl. The 

wet season shows a unimodal distribution with most samples collected from above 3500 masl.  

4.1.2 Results of δ18O and δD Analysis 

 

Results of hydrogen and oxygen analysis are reported in delta-notation (δ) (Error! R

eference source not found.) relative to the Vienna Standard Mean Ocean Water (VSMOW) 

 

63 

 

 

standard and displayed in Table 1. δ18O is 18O / 16O and δ2H or δD 2H / 1H.  δ ratios for both 

oxygen and hydrogen are presented in per mil (‰) units. Standard mean machine precision for 

δ18O measurements was 0.03‰ and range from 0.00 to 0.22‰. The mean δD precision was 

0.16‰, with a range of 0.00 to 1.24‰. Overall instrumental precision was good; all samples 

were retained for further analyses. Results of the δD measurements are considered in the 

general results but not considered the seasonal or elevation analyses. 

 

64 

 

 

Table 1: Summary of Results. In order of increasing elevation. Water Sample (WS) number indicates the year the sample was collected.  

WS Number 

Water Body 

Season 

Easting 

Northing 

Elevation 
(masl) 

δ18O 

(VSMOW) 

δ18O  
St. Dev. 

δD 

(VSMOW) 

δD 

St.Dev. 

Strahler 
Order 

1927 

1928 

1929 

1930 

1925 

1926 

1932 

1931 

1933 

1934 

1939 

1935 

1940 

1936 

1937A 

1937B 

1938 

1924 

Majes mouth 

Irrigation canal 

Majes 

Irrigation canal with plants 

Majes (Punta Colorada) 

Majes (Punta Colorada) 

Majes (Punta Colorada) 

Majes canal 

Majes at Aplao 

Majes (bridge of stix) 

Chuiqui stream in canal 

Majes before Castillo 

Quebrada del Castillo canal 

Northeast Majes 

Colca feed into Majes 

Colca feed into Majes 

Rio Capiza (Coropuna Viraco 

input) 

dry 

dry 

dry 

dry 

wet 

wet 

dry 

dry 

dry 

dry 

dry 

dry 

dry 

dry 

dry 

dry 

dry 

741970 

8164671 

742862 

8168020 

761492 

8177561 

768669 

8189857 

772350 

8198886 

772350 

8198886 

772321 

8199060 

772371 

8199063 

770100 

8222763 

769426 

8228514 

770003 

8223428 

771300 

8232016 

767998 

8228684 

772316 

8242386 

774003 

8246601 

774003 

8246601 

773348 

8246630 

36 

67 

184 

296 

383 

383 

383 

387 

631 

707 

711 

731 

800 

866 

896 

896 

937 

Canal Mascapmpa 

wet 

765517 

8235969 

1128 

-16.40 

-16.48 

-16.41 

-15.19 

-16.49 

-16.60 

-16.37 

-16.29 

-16.81 

-16.78 

-15.62 

-16.74 

-16.49 

-16.95 

-17.09 

-17.10 

-13.39 

-12.56 

0.03 

0.03 

0.01 

0.02 

0.00 

0.02 

0.08 

0.04 

0.06 

0.03 

0.03 

0.02 

0.01 

0.02 

0.01 

0.03 

0.00 

0.05 

-123.51 

-124.04 

-124.06 

-116.14 

-128.21 

-129.44 

-124.10 

-123.79 

-126.58 

-126.49 

-118.58 

-125.87 

-124.82 

-127.36 

-128.70 

-128.70 

-95.97 

-92.81 

0.25 

0.15 

0.07 

0.06 

0.13 

0.00 

0.23 

0.13 

0.15 

0.09 

0.05 

0.03 

0.04 

0.04 

0.11 

0.04 

0.23 

0.27 

5 

5 

5 

5 

5 

5 

5 

5 

5 

5 

5 

5 

1 

5 

5 

5 

4 

3 

 

 

65 

Type 

Natural 

Canal 

Natural 

Canal 

Natural 

Natural 

Natural 

Canal 

Natural 

Natural 

Canal 

Natural 

Canal 

Natural 

Natural 

Natural 

Natural 

Canal 

On / Off 
plateau 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

 

WS Number 

Water Body 

Season 

Easting 

Northing 

Elevation 
(masl) 

δ18O 

(VSMOW) 

δ18O  
St. Dev. 

δD 

(VSMOW) 

δD 

St.Dev. 

Strahler 
Order 

Table 1 (Cont’d) 

1923 

1941 

1922 

1942 

1921 

1943 

1920 

1919 

1944 

1916 

1917 

1918 

1966 

1967A 

1967B 

1815 

1962 

1968A 

1968B 

Rio Grande 

Rio Grande 

Canal de Pacaychacra 

Canal de Pacaychacra 

Huacllay canal 

Huacllay canal 

Iray canal 

Iray canal 

Iray canal 

Sequia La Rinconada 

Chucquibamba canal 

Chucquibamba canal 

Sequia La Rinconada 

Río Grande 

Río Grande 

Rio Andagua 

Río Arma at Salamanca 

Río Blanco 

Río Blanco 

wet 

dry 

wet 

dry 

wet 

dry 

wet 

wet 

dry 

wet 

wet 

wet 

dry 

dry 

dry 

dry 

dry 

dry 

dry 

762277 

8238806 

762218 

8238811 

758748 

8242414 

758672 

8242375 

756653 

8244007 

756914 

8243520 

754343 

8245048 

753554 

8246318 

753144 

8246407 

751338 

8247880 

751394 

8247541 

750888 

8247547 

751494 

8248662 

751523 

8250402 

751523 

8250402 

790338 

8283984 

732609 

8284844 

751163 

8253578 

751163 

8253578 

1379 

1379 

1811 

1811 

1960 

1960 

2347 

2521 

2521 

2889 

2902 

2938 

2965 

2980 

2980 

3070 

3221 

3396 

3396 

-12.13 

-9.74 

-12.32 

-10.83 

-11.39 

-10.27 

-12.25 

-11.87 

-11.55 

-11.92 

-10.85 

-11.39 

-12.31 

-12.39 

-12.44 

-16.70 

-17.11 

-13.42 

-13.49 

0.02 

0.03 

0.05 

0.01 

0.10 

0.03 

0.00 

0.04 

0.01 

0.04 

0.03 

0.02 

0.02 

0.00 

0.01 

0.04 

0.00 

0.01 

0.01 

-89.00 

-72.58 

-89.34 

-77.78 

-82.14 

-74.47 

-87.89 

-85.85 

-82.42 

-83.68 

-76.70 

-78.99 

-86.70 

-87.77 

-87.98 

-127.38 

-125.91 

-94.33 

-94.49 

0.55 

0.10 

0.29 

0.03 

0.81 

0.04 

0.23 

0.02 

0.08 

0.25 

0.62 

0.28 

0.01 

0.03 

0.01 

0.01 

0.05 

0.01 

0.01 

3 

3 

3 

3 

3 

3 

1 

1 

1 

2 

2 

2 

2 

2 

2 

4 

3 

2 

2 

 

 

66 

Type 

Natural 

Natural 

Canal 

Canal 

Canal 

Canal 

Canal 

Canal 

Canal 

Canal 

Canal 

Canal 

Canal 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

On / Off 
plateau 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

OFF 

 

WS Number 

Water Body 

Season 

Easting 

Northing 

Elevation 
(masl) 

δ18O 

(VSMOW) 

δ18O  
St. Dev. 

δD 

(VSMOW) 

δD 

St.Dev. 

Strahler 
Order 

Type 

On / Off 
plateau 

Table 1 (Cont’d) 

1963 

1969A 

1969B 

1970A 

1970B 

1909 

1945 

1965 

1802 

1801 

1964 

1803 

1960 

1961 

1946 

1804 

1910 

1806 

1805 

Salamanca Sincha 

Río Blanco 

Río Blanco 

Río Blanco 

Río Blanco 

dry 

dry 

dry 

dry 

dry 

731929 

8282985 

751344 

8256230 

751344 

8256230 

749675 

8259172 

749675 

8259172 

Quebrada de Rata 

wet 

743603 

8257875 

Quebrada de Rata 

Quebrada de Rata Nighttime 

Quebrada Pariaviri 

Quebrada Pariaviri 

Quebrada Pariaviri 

Quebrada Huamantirca 

Río Arma Bofedal 

Río Arma 

Quebrada Palljaruta 

Quebrada Sique Grande 

Quebrada Azufrioc 

Quebrada Sique Grande 

Quebrada Sique Grande 

dry 

dry 

dry 

dry 

dry 

dry 

dry 

dry 

dry 

dry 

wet 

dry 

dry 

743642 

8257759 

743642 

8257759 

737720 

8263072 

737764 

8263279 

737764 

8263279 

735887 

8283367 

739698 

8293581 

739698 

8293581 

740453 

8265580 

737719 

8284287 

740988 

8276666 

737687 

8283282 

737725 

8283340 

3559 

3727 

3727 

3800 

3800 

3922 

3936 

3936 

3966 

3967 

3967 

4048 

4085 

4085 

4116 

4161 

4186 

4205 

4215 

-13.58 

-13.46 

-13.48 

-13.12 

-13.56 

-12.77 

-9.98 

-10.14 

-12.34 

-12.35 

-12.30 

-14.18 

-17.55 

-17.66 

-12.10 

-15.06 

-13.58 

-14.57 

-14.68 

0.01 

0.01 

0.00 

0.01 

0.03 

0.13 

0.01 

0.02 

0.00 

0.02 

0.02 

0.00 

0.05 

0.01 

0.02 

0.04 

0.02 

0.03 

0.00 

-99.87 

-94.61 

-94.74 

-93.20 

-95.11 

-91.45 

-77.01 

-78.18 

-87.45 

-87.82 

-88.97 

-102.07 

-128.57 

-129.19 

-86.30 

-106.60 

-96.04 

-103.53 

-104.05 

0.04 

0.02 

0.02 

0.00 

0.06 

0.99 

0.01 

0.03 

0.06 

0.17 

0.27 

0.05 

0.01 

0.03 

0.05 

0.22 

0.20 

0.03 

0.00 

 

 

67 

1 

2 

2 

2 

2 

1 

1 

1 

1 

1 

1 

1 

3 

3 

1 

1 

1 

1 

1 

Canal 

OFF 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

 

WS Number 

Water Body 

Season 

Easting 

Northing 

Elevation 
(masl) 

δ18O 

(VSMOW) 

δ18O  
St. Dev. 

δD 

(VSMOW) 

δD 

St.Dev. 

Strahler 
Order 

Table 1 (Cont’d) 

1950 

Río Arma 

2018-KR-01 

Rio Arma at bridge 

1814 

1908 

1951 

1959 

1901 

1952 

1955 

1907 

1954 

1902 

1904 

1953 

1906 

1903 

1808 

1958 

1956 

Rio Arma 

Rio Arma 

Río Colunga/ Rio de 

Choriltes 

Río Achala 

Río Colunga/ Rio de 

Choriltes 

Quebrada Chilacala 

Río Grande 

Río Achala 

Auquimancha/Rio Cuncaicha 

Auquimancha/Quebrada 

Cuncaicha 

Chalojmarka 

Chalojmarka 

Río Blanco o Mapa Mayo o 
Quebrada Collpa Huayco 

Rio Grande 

Bofedal Pucuncho 

Capilla; Río Pucuncho 

Río León Huactana 

dry 

dry 

dry 

wet 

dry 

dry 

wet 

dry 

dry 

wet 

dry 

wet 

wet 

dry 

wet 

wet 

dry 

dry 

dry 

742645 

8294544 

742819 

8294579 

744278 

8294697 

744469 

8294780 

748132 

8297085 

747064 

8296039 

748128 

8297072 

748268 

8297238 

748858 

8297801 

747100 

8296023 

748872 

8297784 

748983.1 

8297822 

748658 

8297498 

748660 

8297492 

747760 

8296537 

748807 

8297737 

748708 

8298933 

749218 

8300353 

751297 

8300700 

4272 

4278 

4310 

4310 

4327 

4328 

4333 

4342 

4343 

4345 

4347 

4349 

4349 

4350 

4352 

4354 

4359 

4375 

4383 

-17.25 

-17.11 

-16.97 

-17.99 

-17.21 

-16.28 

-17.39 

-13.88 

-17.44 

-16.06 

-13.94 

-16.98 

-12.94 

-13.86 

-17.04 

-15.47 

-14.84 

-17.30 

-17.14 

0.02 

0.06 

0.02 

0.02 

0.03 

0.01 

0.02 

0.00 

0.02 

0.03 

0.03 

0.01 

0.07 

0.02 

0.07 

0.10 

0.01 

0.00 

0.01 

-127.04 

-125.06 

-125.63 

-133.13 

-129.34 

-123.01 

-130.03 

-112.29 

-127.30 

-122.12 

-113.65 

-126.69 

-104.04 

-112.22 

-125.53 

-120.53 

-114.06 

-126.41 

-126.63 

0.11 

0.06 

0.06 

0.29 

0.46 

0.00 

0.15 

0.12 

0.11 

0.13 

0.02 

0.04 

0.07 

0.05 

0.01 

0.68 

0.02 

0.02 

0.05 

3 

3 

3 

3 

2 

2 

2 

1 

3 

2 

1 

1 

1 

1 

3 

3 

2 

3 

2 

 

 

68 

Type 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

On / Off 
plateau 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

 

WS Number 

Water Body 

Season 

Easting 

Northing 

Elevation 
(masl) 

δ18O 

(VSMOW) 

δ18O  
St. Dev. 

δD 

(VSMOW) 

δD 

St.Dev. 

Strahler 
Order 

Table 1 (Cont’d) 

2018-Pu-01 

Quebrada Collpa Huauco 

2018-Pu-03 

Quebrada Collpa Huauco 

2018-Pu-04 

Quebrada Collpa Huauco 

2018-Pu-05 

Capilla; Río Pucuncho 

2018-KR-02 

Quebrada Collpa Huauco 

2018-KR-03 

Quebrada Collpa Huauco 

2018- Pu-06 

Capilla; Río Pucuncho 

1811 

1905 

1947 

1911 

1912 

1807 

1957 

1913 

1813 

1915 

1948 

1914 

Río León Huactana 

Capilla; Río Pucuncho 

Quebrada Japuchaca 

Quebrada Japuchaca 

Quebrada Japuchaca 

Quebrada Cuncaicha (at 

site) 

Quebrada Cuncaicha (at 

site) 

Quebrada Pucunchitoyocc 

Rio Agua Caliente 

Road below Q. 
Pucunchitoyocc 

Quebrada Pucunchitoyocc 

Bofedal Pucunchitoyocc 

dry 

dry 

dry 

dry 

dry 

dry 

dry 

dry 

wet 

dry 

wet 

wet 

dry 

dry 

wet 

dry 

wet 

dry 

wet 

748949 

8294183 

748949 

8294183 

748949 

8294183 

749003 

8302075 

748949 

8294183 

749028 

8294248 

749003 

8302075 

751509 

8300873 

749003 

8302075 

741375 

8270657 

741356 

8270635 

741387 

8270628 

756111 

8298963 

756557 

8298854 

744016 

8275307 

756611 

8291491 

743995 

8275275 

743994 

8275277 

743670 

8275292 

4390 

4390 

4390 

4390 

4390 

4390 

4390 

4392 

4408 

4430 

4437 

4440 

4457 

4474 

4539 

4567 

4699 

4717 

4719 

-16.47 

-17.41 

-17.58 

-17.23 

-16.43 

-16.32 

-17.36 

-16.96 

-17.42 

-13.90 

-14.29 

-14.00 

-14.73 

-14.39 

-14.65 

-17.57 

-15.01 

-14.24 

-14.57 

0.02 

0.00 

0.05 

0.10 

0.03 

0.05 

0.01 

0.01 

0.10 

0.01 

0.08 

0.00 

0.01 

0.01 

0.04 

0.04 

0.07 

0.01 

0.22 

-124.43 

-128.16 

-128.81 

-125.59 

-123.67 

-122.90 

-125.92 

-124.25 

-125.89 

-97.62 

-102.30 

-99.58 

-111.25 

-114.32 

-105.25 

-128.58 

-109.67 

-100.35 

-105.95 

0.11 

0.28 

0.05 

0.05 

0.21 

0.15 

0.07 

0.03 

0.06 

0.09 

0.69 

0.41 

0.06 

0.04 

1.18 

0.13 

0.24 

0.07 

1.24 

1 

2 

2 

2 

1 

1 

2 

2 

2 

1 

1 

1 

1 

1 

1 

2 

1 

1 

1 

 

 

69 

Type 

Canal 

Canal 

Canal 

Natural 

Canal 

Canal 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

Natural 

On / Off 
plateau 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

ON 

 

Table 1 (Cont’d) 

WS Number 

Water Body 

Season 

Easting 

Northing 

Elevation 
(masl) 

δ18O 

(VSMOW) 

δ18O  
St. Dev. 

δD 

(VSMOW) 

δD 

St.Dev. 

Strahler 
Order 

1949 

1812 

1810 

1809 

Bofedal Pucunchitoyocc 

Quenua Ranra 

Quebrada Ullullo 

Quebrada Ullullo 

dry 

dry 

dry 

dry 

743978 

8275286 

763669 

8284336 

756489 

8285159 

756375 

8284836 

4720 

4830 

4896 

4938 

-13.28 

-17.74 

-16.84 

-16.75 

0.01 

0.02 

0.03 

0.05 

-95.79 

-135.92 

-121.49 

-120.58 

0.04 

0.05 

0.03 

0.07 

1 

1 

1 

1 

Type 

Natural 

Natural 

Natural 

Natural 

On / Off 
plateau 

ON 

ON 

ON 

ON 

 

 

70 

 

4.1.3 Stable Isotopic Variation In Andean Surface Waters 

 

Surface waters in the study area show considerable variability in δ18O. The total range of 

δ18O is between 17.99 to -9.74‰ and -135.92 and -72.58 ‰ for δD. δD and δ18O values for local 

surface waters were plotted with the Global Meteoric Water Line (GMWL) (Figure 19). The 

GMWL reflects the highly correlated (>r2=0.95) linear relationship between oxygen and 

hydrogen in precipitation (Craig 1961a). The formula for the GMWL is: 

Equation 4: Global Meteoric Water Line 

δD = 8 (δ18O) + 10‰ 
δD = slope (δ18O) + d‰ 

 

D is the deuterium excess (d=δD – 8 δ18O) and represents enrichment from evaporative 

effects. Evaporative loss causes waters to plot below or to the right of the GMWL. Vapors 

derived from evaporation plot above and to the right of the line. Waters condensed from the 

evaporated vapors will plot parallel to the line but demonstrate a greater deuterium excess. The 

Local Surface Water Line (LSWL) derived from the entire dataset is: 

Equation 5: Local Surface Water Line 

  δD= 7.86 (δ18O) + 7.08.  

 

The LSWL suggests local δD and δ18O values are highly correlated (r2=0.96). Data plotting 

below the LSWL either indicate direct evaporative effects or sources mixed with other 

evaporated waters (Craig 1961a; Gat 1996, 2000). The plotted data and LSWL show little to no 

evaporative effects affecting meteoric precipitation prior to deposition in Andean surface 

waters.  Seasonal differences in the local meteoric water line are plotted and discussed in 

Section 4.2.1.  

 

71 

 

 

Figure 19: Plot of δ18O and δD for all samples. The black line is the Global Meteoric Water Line (GMWL) 
δD = 8(δ18O) +10‰. The dashed black line reflects the local surface water line (LSWL) δD= 7.86 (δ18O) + 

7.08‰. Class 1-5 indicates stream order, where first order streams are the smallest volume. The 

enriched samples below the line are discussed in Figure 21. 

 

4.2 Results of Seasonal Sampling 

 

In total, seventy-two samples between 36 and 4938 masl represent the austral dry 

season and 26 samples between 383 and 4719 masl represent the wet season. Dry season 

samples include all samples collected from July-December 2018, and June 2019. Wet season 

samples reflect collection during February-March 2019.  

 

72 

 

 

4.2.1 Local Variation in Andean Surface Waters Between Seasons 

 

The range of δ18O values observed in Andean surface waters is between -17.37 to -

9.74‰ during the austral dry season and -17.99 to -10.84 ‰ in the wet season (Table 2). 

Seasonal results of δD and δ18O values were plotted relative to the Global Meteoric Water Line 

(GMWL). The plot suggests more evaporated values may occur during the dry season than the 

wet season (Figure 20), though this may also reflect the dry season sampling bias.  

Table 2: Summary of Sampling for Each Field Season. 

Season 

DRY 

WET 

DRY 

Year 

2018 

2019 

2019 

Samples 

(n =) 

Min  

Elevation 
(masl) 

Max 

Elevation 
(masl) 

23 

26 

49 

3070 

347 

36 

4938 

4719 

4720 

Min δ18O 
(VSMOW) 

Max δ18O 
(VSMOW) 

Average δ18O 
(VSMOW) 

-17.74 

-17.99 

-17.66 

-12.34 

-10.84 

-9.74 

-16.01 

-14.23 

-14.59 

 

 

73 

 

 

 

 
Figure 20: Seasonal datasets plotted with the respective wet and dry season Local Surface Water Lines (LSWL). Dry season samples (yellow dots) 

and LMWL (δD = 7.75 * δ18O + 5.52‰) are presented on the left and wet season samples (blue dots) and LMWL (δD = 8.15 * δ18O + 11.22‰) 

presented on the right. Values that plot below the GMWL and LSWLs reflect enriched waters. 

 

 

74 

 

 
 

 

Seasonal Local Surface Water Lines (LSWL) base on all samples are:  

Dry Season: δD = 7.75 * δ18O + 5.52‰ 
Wet Season: δD = 8.15 * δ18O + 11.22‰ 

LSWL: δD = 7.86 (δ18O) + 7.08.  
GMWL: δD = 8(δ18O) +10‰  

Both the dry season slope (7.75) and intercept (5.52‰) suggest slightly more 

 

evaporated surface waters in the study area during this time. The wet season intercept 

(11.22‰) indicates that the incoming precipitation may reflect previously evaporated waters 

carried through transport to the study area. The increased intercept during the wet season 

(11.22‰) likely reflects the influx of precipitation from evaporative vapors transported from the 

east. 

 

To adjust for the difference in sample size between seasons, the plot was recreated 

using only samples that were replicated during both the wet and dry season (n=24 samples, 

n=12 per season) (Figure 21). 

Dry Subset: (δD = 7.64 * δ18O + 3.14‰) 
Wet Subset:(δD = 8.10 * δ18O + 10.83‰) 

In a subset test, the dry season slope and intercept are both decreased indicating more 

 

 

evaporated waters than the wet season samples from the same locations. Some waters appear 

more affected by evaporation during the dry season than others (Figure 21). For example, 

sample location 2, 4, 6, and 9 all fall below the GMWL during the dry season, indicating waters 

susceptible to evaporative processes. Sample location 1, the Majes River, falls to the right of the 

line during both the wet and dry seasons. Sample locations 2 (Río Grande; 1379 masl), 4 

(Huacllay Canal; 1960 masl), and 9 (Río Cuncaicha; 4348 masl) fall to the right of the GMWL, 

suggesting enrichment occurs at these locations during the dry season.  

 

75 

 

 

 

 

Figure 21: δ18O plotted against δD for sample locations replicated as part of the seasonality study. 

Dashed black line indicates the global meteoric waterline δD = 8.0 * δ18O + 10.00‰. Blue solid line δD = 
8.10 * δ18O + 10.83‰. indicates the local meteoric water line during the wet season. Yellow dashed line 
represents the local meteoric water line during the dry season δD = 7.36 * δ18O + 3.14‰. More samples 

appear to plot below the line during the dry season. 

4.2.2 Results of Temporal Testing 

 

To test for measurable variation in surface waters at multiple time scales, samples were 

collected daily, seasonally, and yearly from the same water bodies. The resulting temporal 

 

76 

 

 

dataset consists of 29 paired samples. Seven simultaneous tests represent samples taken from 

the same locations at the same time. To estimate the degree of diurnal change in surface 

waters, one pairing was collected from the Quebrada de Rata in both the morning and evening. 

Seasonal samples include 12 pairs acquired during the dry seasons of 2018 or 2019 and the wet 

season of 2019. Three paired samples collected during both the 2018 and 2019 dry seasons test 

same-season variation during different years. Finally, six tests compare waters collected in 2018 

or 2019 with published (Chala-Aldana et al. 2018) 2015 dry season values.  

4.2.2.1 Boxplot Assessment for All Temporal Tests 

 

Per mil (‰) differences in δ18O between temporal pairs are presented in Tables 3, 5, 7, 

and 9. Potential outliers for all subsets were identified using the boxplot method to assess the 

change in δ18O across the time scales (Figure 22). The simultaneous test revealed one outlier 

(difference = 0.44‰) from water sample pairs WS1970A and WS1970B from the Río Blanco at 

3800 masl (Table 3) the pair was removed prior to further analysis. This Río Blanco at this 

elevation is a waterfall and presents fast moving, highly aerated waters; this is the likely cause 

of the increased difference between replicate samples. 

 

 

 

 

77 

 

 

 

Figure 22: Box plot assessment for the per mil δ18O difference between paired temporal samples. Mean 

instrumental precision (0.03‰) shown by the dashed black line. Maximum instrumental precision 

(0.22‰) indicated by the dotted grey line.  

4.2.3 Daily Tests 

4.2.3.1 Simultaneous Tests 

 

Results of simultaneous testing are provided in Figure 1. The same-time dataset (n=14) 

represents paired samples of water taken from seven locations between 383 and 4085 masl 

The first same-time pair (SIM-01) was collected from highland waters, at the rim of the plateau, 

 

78 

 

 

from a class-two stream on July 11th, 2018 (1801 and 1802) in the dry season. The second same-

time test (SIM-02) was collected below the plateau from the Majes River at the Punta Colorada 

bridge, west of the town of Aplao, on March 1st, 2019 (1925 and 1926) in the wet season. The 

remaining six simultaneous pairs were collected from rivers in the low- and highlands during 

the 2019 dry season. Replicate test SIM-06 from the Rio Blanco (Table 3) is considered an 

outlier and is not included in the data plots. The observed δ18O values for each daily pair are 

plotted in Figure 24. Per mil difference between daily samples is shown in Figure 25. 

 

Little to no variation was expected for samples taken during the same day, regardless of 

time. The per mil difference between same-day tests, including the outlier, ranges between 

δ18O 0.00‰ to 0.44‰. Excluding the outlier, the range is 0.00‰ to 0.11‰. The mean standard 

machine error for δ18O was 0.03‰, range 0.00 to 0.22‰. The mean is less than the difference 

for four simultaneous samples and all observed daily variation is less than the maximum 

possible standard machine error. 

 

79 

 

 

Table 3: Results of Daily Tests.  

Sample ID 

Pairing ID 

Test Type 

Location 

1925 

1926 

1937A 

1937B 

1967A 

1967B 

1968A 

1968B 

1969A 

1969B 

1970A 

1970B 

1801 

1802 

1960 

1961 

1945 

1965 

SIM-01 

Simultaneous 

Majes River 

SIM-02 

Simultaneous 

Río Colca Feed into Majes 

River 

SIM-03 

Simultaneous 

Río Grande 

SIM-04 

Simultaneous 

Río Blanco 

SIM-05 

Simultaneous 

Río Blanco 

SIM-06 

Simultaneous 

(outlier) 

Río Blanco 

SIM-07 

Simultaneous 

Quebrada Pariaviri 

SIM-08 

Simultaneous 

Río Arma 

DIUR-01 

Diurnal 

Quebrada de Rata 

Time 

15:30 

15:32 

12:00 

12:02 

9:20 

9:22 

12:00 

12:02 

13:00 

13:02 

13:30 

13:32 

9:42 

9:53 

15:30 

15:32 

8:23 

19:20 

Elevation 
(masl) 

383 

 

896 

 

2980 

 

3396 

 

3727 

 

3800 

 

3967 

 

4085 

 

3936 

 

Measured 

δ18O 

(VSMOW) 

-16.49 

-16.60 

-17.09 

-17.10 

-12.39 

-12.44 

-13.42 

-13.49 

-13.46 

-13.48 

-13.12 

-13.561 

-12.35 

-12.34 

-17.55 

-17.66 

-9.98 

-10.14 

δ18O Average 
(VSMOW) 

δ18O Difference 

(‰) 

Date 

-16.55 

0.11 

3-Mar-19 

-17.10 

0.01 

13-Jun-19 

-12.42 

0.05 

15-Jun-19 

-13.46 

0.07 

15-Jun-19 

-13.47 

0.02 

15-Jun-19 

-13.34 

0.44 

15-Jun-19 

-12.35 

0.01 

11-Jul-18 

-17.61 

0.11 

14-Jun-19 

-10.06 

0.16 

14-Jun-19 

 

 

80 

 

 
Figure 23: Map of daily sampling locations. Replicate tests are dark blue: the Río Majes, SIM-01 (383 
masl); The Río Colca, SIM-02 (896 masl); the Río Grande, SIM-03 (2980 masl), SIM-04 (3396 masl), and 
SIM-05 (3727 masl); Quebrada Pariaviri, SIM-06 (3967 masl); and the Río Arma, SIM-07 (4085 masl). The 

diurnal sample (lighter blue), DIUR-01 (3936 masl) derives from the Quebrada de Rata. 

 

81 

 

 

 
Figure 24: Observed δ18O for simultaneous (dark blue triangles) and diurnal (light blue circles) pairs. 
Samples include the Río Majes, SIM-01 (383 masl); The Río Colca, SIM-02 (896 masl); the Río Grande, 
SIM-03 (2980 masl), SIM-04 (3396 masl), and SIM-05 (3727 masl); Quebrada Pariaviri, SIM-06 (3967 

masl); and the Río Arma, SIM-07 (4085 masl). The diurnal sample, DIUR-01 (3936 masl) derives from the 

Quebrada de Rata.  

  

 

 

82 

 

 

 
Figure 25: Change in δ18O between pairs collected for daily tests. Replicate tests (dark blue triangles) 
include the Río Majes, SIM-01 (383 masl); The Río Colca, SIM-02 (896 masl); the Río Grande, SIM-03 
(2980 masl), SIM-04 (3396 masl), and SIM-05 (3727 masl); Quebrada Pariaviri, SIM-06 (3967 masl); and 
the Río Arma, SIM-07 (4085 masl). The diurnal sample (light blue circle), DIUR-01 (3936 masl) derives 
from the Quebrada de Rata. All demonstrate less change than the maximum machine error (0.22‰). 

 

The boxplot for δ18O difference between same-time samples taken at a single location 

revealed one outlier with larger than expected variance. The sample collection, from Location 

6, Río Blanco, utilized identical conical 15 mL Corning vials and the analytical machine precision 

 

83 

 

 

for both samples was good. To understand the impact of the outlying value on the dataset, the 

value was removed and the range and mean recalculated. The adjusted range for same-time 

tests is: δ18O 0.00‰ to 0.11‰ (Figure 24).  

 

Locations 3 (Río Grande) and 5 (Río Blanco) of the same-time tests utilized different vial 

types to test for variation between sampling containers; these pairs utilized one 30 mL Nalgene 

tube and one 15mL Corning vial. No significant variation was expected between vials, and no 

significant variation was identified (0.05‰ and 0.01‰). The statistical summary of results 

provided in Table 4 considers all daily samples to provide the most conservative simultaneous 

baseline variation possible. The range of difference for δ18O was 0.43‰. The interquartile range 

for per mil variation was 0.09‰. 

Table 4: Statistical Results for Simultaneous Tests. 

Pairing Type 

Simultaneous 

Tests 

Samples 

(n =) 

8 

Mean 
Change 

(‰) 

0.10 

Standard 
Deviation 

Minimum 
Change 

(‰) 

Maximum 
Change 

(‰) 

Range 
(‰) 

Q1 

Q3 

0.14 

0.01 

0.44 

0.43 

0.02 

0.11 

IQR 

.09 

4.2.3.2 Diurnal Test 

 

A single test was conducted during June 2019 to examine diurnal variation for a single 

location, as shown in Figure 24. Water sample numbers 1945 and 1965 were collected 

approximately 11 hours apart from the Quebrada de Rata at the edge of the Andean Plateau. 

The test indicates little difference between water collected from a single location at different 

times of day. Variation for the diurnal test was 0.16‰ for δ18O (Figure 24). This value falls 

slightly outside of the range of per mil difference for the simultaneous tests. It is less than the 

range for machine precision error.  Therefore, diurnal variance is negligible. 

 

84 

 

 

4.2.4 Seasonal Tests 

 

Fifteen locations were sampled during both the wet and dry seasons. Three sample pairs 

were excluded from the seasonal analysis due to the uncertainty of their locations after a visual 

inspection in the GIS. Seasonal sample locations represent elevations between 383 and 4720 

masl. The final seasonal test dataset is provided in Table 5 and shown in Figure 26.  

 

Twelve locations were sampled during both the wet (2019) and dry (2018 or 2019) 

seasons (Figure 26). Two seasonal locations incorporated samples collected as part of the daily 

testing (SEAS-01) Río Majes and (SEAS-06) Quebrada de Rata. For these seasonal tests, three 

values were available for comparison. To reduce the comparative sample to two, the samples 

with the greatest difference was selected for comparison. For example, two samples were 

collected from the Río Majes during the wet season (2019), and one during the dry season 

(2019). The two wet season samples provided δ18O values of -16.60‰ (WS1926) and -16.49‰ 

(1925). The dry season value was -16.37‰ (WS1932). Water Samples 1926 and 1932 were used 

for the seasonal analysis because they reflect the greatest difference possible between seasons 

for waters collected at this location.  

 

 

 

85 

 

 

Table 5: Results of Seasonal Sampling.  

WS Number 

Pairing ID 

Location 

Season 

Elevation 
(masl) 

δ18O 

(VSMOW) 

1926 

1932 

1923 

1941 

1922 

1942 

1921 

1943 

1919 

1944 

1909 

1945 

1814 

1908 

1901 

1951 

1902 

1954 

1911 

1947 

1913 

1948 

1914 

1949 

SEAS-01 

 

Majes (Punta Colorada) 

SEAS-02 

 

Río Grande 

 

SEAS-03 

 

Canal de Pacaychacra 

 

SEAS-04 

 

Huacllay canal 

 

SEAS-05 

 

Iray canal 

 

SEAS-06 

 

Quebrada de Rata 

 

SEAS-07 

 

Río Arma 

Quebrada Ocoruru 

SEAS-08 

 

Río Colunga/ Rio de Choriltes 

 

SEAS-09 

 

Auquimancha/Río Cuncaicha 

 

SEAS-10 

 

Quebrada Japuchaca 

SEAS-11 

 

Quebrada Pucunchitoyocc 

SEAS-12 

Bofedal Pucunchitoyocc 

wet 

dry 

wet 

dry 

wet 

dry 

wet 

dry 

wet 

dry 

wet 

dry  

dry 

wet 

wet 

dry 

Wet 

dry 

Wet 

dry 

wet 

dry 

wet 

dry 

383 

1379 

 

1811 

 

1960 

 

2521 

 

3929 

 

4310 

 

4330 

4348 

 

4434 

4628 

4720 

-16.60 

-16.37 

-12.13 

-9.74 

-12.32 

-10.83 

-11.39 

-10.27 

-11.87 

-11.55 

-12.77 

-9.98 

-16.97 

-17.99 

-17.39 

-17.21 

-16.98 

-13.94 

-14.29 

-13.90 

-14.65 

-14.24 

-14.57 

-13.28 

Average 
δ18O  

(VSMOW) 

Difference 

δ18O  
(‰) 

-16.49 

0.23 

-10.94 

2.39 

-11.58 

1.49 

-10.83 

1.12 

-11.71 

0.32 

-11.38 

2.79 

-17.48 

1.02 

-17.30 

0.18 

-15.46 

3.04 

-14.10 

0.39 

-14.50 

0.41 

-13.92 

1.29 

 

86 

 

 

 

Figure 26: Map of the seasonal sample locations: the Río Majes, SEAS-01 (383 masl); The Río Grande, 
SEAS-02 (1379 masl); the Canal de Pacaychara, SEAS-03 (1811 masl); Huacllay Canal, SEAS-04 (1960 

masl); Iray Canal, SEAS-05 (2521 masl); Quebrada de Rata, SEAS-06 (3936 masl); the Río Arma, SEAS-07 

(4310 masl); Río Colunga, SEAS-08 (4333 masl); Río Cuncaicha, SEAS-09 (4347 masl); Quebrada 
Japuchaca, SEAS-10 (4435 masl); Quebrada Pucunchitoyocc, SEAS-11 (4717 masl); and Bofedal 

Pucunchitoyocc, SEAS-12 (4720 masl).  

 

87 

 

 

 

Sample pairs 1-5 were acquired from below the plateau rim. Samples 3-5 represent 

canals that channel waters down to the city of Aplao from the town of Chuquibamba. The 

remaining two samples (pairs 1 and 2) derive from the Majes River (383 masl) and Río Grande 

(1379 masl), Order 5 and Order 4 rivers respectively. The remaining seven pairs (6-12) 

represent waters collected from the Andean Plateau. The boxplot assessments (Figure 22) 

identified no outliers in the seasonal test dataset; therefore all 12 values were retained for 

further assessment. The δ18O for each sample collected as part of a seasonal pairing is 

presented in Figure 27.  

 

All dry season samples demonstrate slightly higher δ18O values than the wet season 

samples from the same locations. All Class 3 streams demonstrate an appreciable amount of 

variation. The degree of δ18O change in is variable among first order streams. For example, 

sample location SEAS-06, Quebrada de Rata, demonstrates a 2.97‰ shift between seasons. But 

sample location SEAS-10, Quebrada Japuchaca, only demonstrates a 0.39‰ change. All 

seasonal samples demonstrate more δ18O change than daily samples (Table 5).  

Table 6: Statistical Results of Seasonal Sampling. 

Pairing Type 

Seasonal 

Tests 

Samples 

(n =) 

12 

Mean 
Change 

(‰) 
1.21 

Standard 
Deviation 

1.01 

Minimum 
Change 

(‰) 
0.18 

Maximum 
Change 

(‰) 
3.04 

Range of 
Change 

(‰) 
2.86 

Q1 

Q3 

IQR 

0.39 

1.49 

1.10 

 

88 

 

 

Figure 27: Plot showing δ18O values for each sample in the seasonal pairs. Data are coded by wet (blue triangles) and dry (yellow circles) seasons. 

The Majes (purple), Jaguay (blue), and Ocoña (coral) drainages are indicated. Location codes reflect: the Río Majes, SEAS-01; The Río Grande, 
SEAS-02; the Canal de Pacaychara, SEAS-03; Huacllay Canal, SEAS-04; Iray Canal, SEAS-05; Quebrada de Rata, SEAS-06; the Río Arma, SEAS-07; 
Río Colunga, SEAS-08; Río Cuncaicha, SEAS-09; Quebrada Japuchaca, SEAS-10; Quebrada Pucunchitoyocc, SEAS-11; and Bofedal Pucunchitoyocc, 

SEAS-12.   

 

 

 

89 

 

 
Figure 28: Plot showing the difference in δ18O (‰) between seasonal sample pairs. Location codes in order of increasing elevation: the Río 

Majes, SEAS-01; The Río Grande, SEAS-02; the Canal de Pacaychara, SEAS-03; Huacllay Canal, SEAS-04; Iray Canal, SEAS-05; Quebrada de Rata, 
SEAS-06; the Río Arma, SEAS-07; Río Colunga, SEAS-08; Río Cuncaicha, SEAS-09; Quebrada Japuchaca, SEAS-10; Quebrada Pucunchitoyocc, SEAS-
11; and Bofedal Pucunchitoyocc, SEAS-12.   All samples demonstrate more change than maximum machine error (0.22‰) and maximum daily 

change (0.16‰). 

 

 

90 

 

4.2.5 Yearly Tests 

 

A second threshold for identifying significant seasonal variation comes from the results 

of yearly sampling, where the same location is sampled during the same (dry) season in 

different years. These tests provide an insight to whether the observed variation in seasonal 

will be greater than inter-annual variation. Results of annual testing are reported in Table 7 and 

locations presented in (Figure 30). Yearly tests are divided into single year (<1 year; “YR1”) and 

multi-year (>3 years; “MYR”) groupings. The amount of time between samples is presented for 

the multi-year groupings. Single year groupings were sampled 11 months apart and represent 

samples from July 2018 and June 2019. Figure 28 demonstrates the per mil difference between 

yearly samples.  

Table 7: Results of Annual Sampling.  

WS number 

Pairing iD 

Location 

Date 

1801 

1964 

1814 

1950 

1807 

1957 

YR-01 

 

Quebrada 
Pariaviri 

YR-02 

 

Río Arma 

 

YR-03 

Quebrada 
Cuncaicha 

July 2018 

June 2019 

July 2018 

June 2019 

July 2018 

June 2019 

Average 
Elevation 
(masl) 

3967 

4291 

 

4465 

δ18O 

(VSMOW) 

Average δ18O 
(VSMOW) 

δ18O Change 

(‰) 

-12.35 

-12.30 

-16.97 

-17.25 

-14.73 

-14.39 

-12.33 

0.05 

-17.11 

0.28 

-14.56 

0.34 

 

 

 

91 

 

 

Figure 29: The δ18O of surface waters sampled during the dry season in multiple years. Annual samples 
(<1 year; yellow circles) include: Quebrada Pariaviri, YR1-01 (3966 masl); Río Arma, YR1-02 (4300 masl); 
and Quebrada Cuncaicha, YR1-03 (4457 masl). Multiyear samples (>3 years; orange triangles) include the 
Río Majes, MYR-01 (383 masl); The Río Grande, MYR-02 (1379 masl); Río Grande, MYR-03 (2980 masl); 

Río León Huactana MYR-04 (4384 masl); Quebrada Cuncaicha (3-year) MYR-05 (4457 masl); and 

Quebrada Cuncaicha (4-year) MYR-06 (4474 masl). 

 

 

92 

 

 

Figure 30: Map of the yearly sample locations. Annual samples (<1 year) include: Quebrada Pariaviri 
YR1-01 (3966 masl); Río Arma, YR1-02 (4300 masl); and Quebrada Cuncaicha, YR1-03 (4457 masl). 

Multiyear samples (>3 years) include the Río Majes, MYR-01 (383 masl); The Río Grande, MYR-02 (1379 
masl); Río Grande, MYR-03 (2980 masl); Río León Huactana MYR-04 (4384 masl); Quebrada Cuncaicha 

(3-year) MYR-05 (4457 masl); and Quebrada Cuncaicha (4-year) MYR-06 (4474 masl). 

 

 

93 

 

 

4.2.5.1 Annual Tests 

 

Annual tests represent sample locations that were replicated during the dry seasons of 

2018 and 2019. The dry seasons occurred in different months (July 2018 and June 2019); 

however, minimal change is expected at such a short time scale. Three annual tests were 

recorded within the dataset, two from second order streams (pairings YR1-01 and YR1-03) and 

one from a fourth order river (pairing YR1-02). All three pairings derive from plateau waters. 

Notable in this dataset is the Quebrada Cuncaicha (pairing YR1-03) that runs in front of the 

archaeological site. All three pairs show remarkably little per mil variation. Results of annual 

statistical tests are reported in Table 8. Measured change between tests ranged from 0.05 to 

0.34‰, with a mean change of 0.22‰. These data suggest annual change in surface waters is 

minimal; however, (1) interannual variation and (2) a larger sample set remain to be tested. 

Table 8: Results of Annual Statistical Tests. 

Pairing Type 

Samples  

(n =) 

One Year 

3 

Mean 
(‰) 

0.22 

Standard 
Deviation 

1.01 

Minimum 
Change  

(‰) 
0.05 

Maximum 
Change 
 (‰) 
0.34 

Range of 
Change 

(‰) 
0.29 

Q1 

Q3 

IQR 

0.17 

0.31 

0.14 

 
4.2.5.2 Interannual Tests 

 

Some water samples taken by Chala-Aldana in 2015 overlap the locations in the 2018-

2019 sample set. Potential interannual samples were identified by converting Chala-Aldana’s 

(2018) sample locations to UTM and comparing them to the 2018-2019 dataset in a GIS. The six 

potential interannual locations identified are presented in Table 9. 

  

 

 

 

94 

 

 

Table 9: Results of Interannual Tests. 

WS number 

Pairing ID 

Location 

OUT-WA-011 

1932 

OUT-WA-014 

1941 

OUT-WA-017 

1967B 

PU-WA-009 

1956 

PU-WA-005 

1807 

PU-WA-005 

1957 

MY-01 

 

Majes (Punta 

Colorada) 

MY-02 

 

Río Grande 

 

MY-03 

 

Río Grande 

  

MY-04 

 

Río León Huactana 

 

MY-05 

 

Quebrada Cuncaicha 

 

MY-06 

 

Quebrada Cuncaicha 

 

Season 
& Year 

Dry 2015 

Dry 2019 

Dry 2015 

Dry 2019 

Dry 2015 

Dry 2019 

Dry 2015 

Dry 2019 

Dry 2015 

Dry 2018 

Dry 2015 

Dry 2019 

 

Elevation 
(masl) 

Oxygen 
(VSMOW) 

Average 
(VSMOW) 

Difference 

δ18O 

(‰) 

371 

1379 

 

2887 

4383 

4463 

4475 

-15.95 

-16.37 

-9.49 

-9.74 

-11.86 

-12.44 

-17.57 

-17.14 

-14.56 

-14.73 

-14.56 

-14.39 

-16.16 

0.42 

-9.62 

0.25 

-12.15 

0.58 

-17.36 

0.43 

-14.65 

0.17 

-14.48 

0.17 

 

Six interannual samples span elevations between 371 and 4475 masl representing waters from 

both on and off the Andean Plateau. The interannual difference between surface waters range from 

0.17 to 0.58‰ with a mean of 0.34‰. 

Table 10: Results of Interannual Statistical Tests.  

Pairing Type 

Per mil 

Difference (‰) 

Samples  

(n =) 

6 

Mean 
Change 

(‰) 
0.34 

Standard 

Deviation (s) 

0.17 

Minimum 
Change 

(‰) 
0.17 

Maximum 
Change 

(‰) 
0.58 

Range 
(‰) 

Q1 

Q3 

IQR 

0.41 

0.19 

0.43 

0.22 

 

95 

 

 

Figure 31: Per mil difference in δ18O between surface waters sampled during the dry season in multiple years. Annual samples (<1 year; yellow 
circles) include: Quebrada Pariaviri, YR1-01 (3966 masl); Río Arma, YR1-02 (4300 masl); and Quebrada Cuncaicha, YR1-03 (4457 masl). Multiyear 
samples (>3 years; orange triangles) include the Río Majes, MYR-01 (383 masl); The Río Grande, MYR-02 (1379 masl); Río Grande, MYR-03 (2980 

masl); Río León Huactana MYR-04 (4384 masl); Quebrada Cuncaicha (3-year) MYR-05 (4457 masl); and Quebrada Cuncaicha (4-year) MYR-06 

(4474 masl). 

 

 

 

96 

 

4.2.6 Results of Temporal Comparisons 

 

Results of the comparison for difference values between daily, seasonal, and yearly 

changes in stable oxygen of surface waters for the study area are depicted in Figure 32. Because 

Andean surface waters may be susceptible to changes in δ18O over multiple time scales, it is 

important to consider both daily and annual change in determining whether seasonal change 

registers at the greatest magnitude in Andean surface waters.  

 

Daily and yearly tests set two rhythmic baselines for non-seasonal variation in the study 

area. Figure 32 plots all temporal data by sample location to demonstrate the measured change 

in δ18O values of surface waters over multiple time intervals. The dashed line indicates the 

greatest change observed in a non-seasonal pair (0.58‰; multi-year). Surface waters in the 

study area appear to reflect the greatest degree of δ18O change at the seasonal level. A final 

test explores different levels of temporal variation for controlled locations. 

 

 

97 

 

 

Figure 32: Per mil δ18O change for simultaneous (dark blue), diurnal (light blue), seasonal (green), one-year (yellow) and multi-year (orange) 
tests. Numbers correspond to sampling locations for each test in order of elevation. Maximum observed machine error indicated by the black 
dashed line (0.22‰).   Maximum daily change (0.16‰) indicated by the blue dashed line. Maximum yearly change (0.58‰) indicated by the 

orange dotted line.  

 

 

 

98 

 

4.2.6.1 Change Over Multiple Time Scales, Controlled for Location 

 

In the first plot, Figure 33, the per mil δ18O daily change is compared to δ18O seasonal 

change for three locations. The Quebrada de Rata (a first order stream) is located in the Jaguay 

drainage at 3936 masl and demonstrates a 0.16‰ change in δ18O diurnally, but a 2.79‰ 

change seasonally. The difference between daily and seasonal change is 2.63‰. Next, the Río 

Arma (a third order river) is located in the Ocoña drainage system at ~4085 masl. Measured 

change in δ18O between replicate simultaneous tests is 0.11‰ while measured seasonal change 

is 1.02‰ The δ18O per mil change between simultaneous and seasonal samples is 0.91‰. 

Finally, the Río Majes (a fifth order river) is located at 383 masl in the Majes drainage. 

Simultaneous change is 0.11‰ and seasonal percent change is 0.23‰. Per mil δ18O change is 

only 0.12‰ for this location. 

 

Comparative daily and seasonal tests from identical locations are limited to three 

samples; any conclusions will be biased by the small sample size. For these locations, seasonal 

change is greater than daily change for all locations, though seasonal change in the Río Majes is 

considered negligible. The percent of seasonal change decreases with increasing stream order 

for this dataset (n=3).   

 

99 

 

 

 

Figure 33: Measured per mil difference in δ18O at the daily (yellow) and seasonal (red) scales for the 

Quebrada de Rata, Río Arma, and Río Majes.  

 
 

In the plot, Figure 34, per mil daily change in δ18O is compared to the seasonal and 

annual change for two locations. The Río Arma (third order) is located in the Ocoña drainage 

system at ~4085 masl. The measured change in surface water δ18O between simultaneous and 

annual samples is 0.27‰ and per mil change between seasonal and annual samples is 0.74‰. 

 

100 

 

 

For all locations, the change in surface water δ18O between seasons is greater than the change 

measured for single-day samples. 

Next, the Río Majes (fifth order) is located at 383 masl in the Majes drainage. For the 

three locations, the Río Majes demonstrates the greatest change in δ18O surface waters at the 

multi-year scale. Observed seasonal change in δ18O for the Río Majes was 0.23‰, while multi-

year variation was measured at 0.42‰. Per mil change in the surface water δ18O of the Río 

Majes, compared at the daily and multi-year levels, is low (0.31‰); and even lower when 

comparing the degree of seasonal and annual δ18O change (0.19‰). Multi-year variation is 

greater than seasonal variation for the Río Majes. The δ18O of Majes surface waters are 

remarkably stable over time. 

 

The Río Grande (third order) at 1379 masl in the Majes drainage provides a third 

location to compare seasonal and multi-year (4 years) variation. Observed seasonal change in 

δ18O for the Río Grande was 2.39‰, while multi-year variation was measured at 0.25‰. The 

per mil difference between seasonal and annual change is 2.14‰. Seasonal change is greater 

than yearly change for this location. 

 

As the final location in this comparative test, the Quebrada Cuncaicha (second order) in 

the Ocoña drainage was measured at the one-(2018-2019), three- (2015-2018) and four-year 

(2015-2019) scales. One-year change is 0.34‰, three-year change 0.17%, and four-year change 

0.17‰. The maximum per mil variation observed for the Quebrada Cuncaicha at the yearly 

scale was 0.34 per mil change between 2018 and 2019. 

 

For these sample locations, smaller order streams demonstrate more change in surface 

water δ18O at the season-year scale than larger order streams.  

 

101 

 

 

 
Figure 34: Per mil change in the δ18O of surface waters at the daily (yellow), seasonal (red), and annual 
(blue) scales for the Río Arma, Río Grande, and Río Majes. Pairings from the Quebrada Cuncaicha reflect 

yearly change at a one- and four-year scale.  

4.3 Spatial Variation in Stable Oxygen in Southern Peru  

 

In this section I investigate whether the δ18O values of surface waters vary predictably 

between the Pacific coast and high Andes. I use an elevation transect to assess whether δ18O 

values are distinct among different elevations. As a brief review, the study area begins at the 

 

102 

 

 

Pacific coast and extends north-northeast up into the high-altitude Andes mountains. The 

primary oxygen isotopic input derives from Atlantic precipitation arriving from the east. In the 

study area, the west coast is the lowest elevation zone, experiences warmest temperatures, 

and is furthest from the Atlantic system. The targeted Andean high-elevation locations are 

closer to the input source, receive more precipitation, and are colder overall.  

 

Published studies assume that regional δ18O values in meteoric waters will have an 

inverse relationship with elevation (Chala-Aldana et al. 2018; Haas et al. 2017; Knudson 2009). 

Specifically, δ18O values should decrease with increasing elevation. I measure surface water 

δ18O values to test the assumption that low elevation waters demonstrate higher δ18O values 

than high-elevation waters. This section reviews whether a relationship between elevation and 

surface water δ18O exists in the Study Area. 

4.3.1 δ18O Variation Across an Elevation Transect 

 

Table 11 and Figure 35 present the results and locations of the elevation transect. The 

δ18O value for each sample was plotted against the elevation of collection (Figure 36). To 

simplify discussion, elevation changes are discussed in 1000-meter intervals (Table 11). 

Table 11: Results of Elevation Subsets. 

Elevation 
Bracket 
(masl) 

< 1000 

1001-2000 

2001-3000 

3001-4000 

> 4000 

Number 

(n=) 

Minimum 
Elevation 
(masl) 

Maximum 
Elevation 
(masl) 

17 

7 

9 

15 

50 

36 

1128 

2347 

3070 

4048 

937 

1960 

2980 

3967 

4938 

 

Minimum 

δ18O 

(VSMOW) 

Maximum  

δ18O 

(VSMOW) 

-13.39 

-9.74 

-10.85 

-9.98 

-12.10 

-17.10 

-12.56 

-12.44 

-17.77 

-17.99 

103 

 

1st Quartile 

δ18O 
(‰) 

3rd Quartile 

δ18O 
(‰) 

-16.78 

-12.22 

-12.31 

-13.53 

-17.25 

-16.37 

-10.55 

-11.55 

-12.35 

-14.44 

Mean 
δ18O 
(‰) 

-16.31 

-11.32 

-11.89 

-13.19 

-15.83 

 

Figure 35: Map of all sampling locations illustrating the δ18O values of the collected surface waters. 

 

 

104 

 

 

Figure 36:Distribution of δ18O across elevation. The plot incorporates all samples. Four outliers include 
WS1962 (pink) in the Ocoña drainage basin, WS1815 (blue) in the Majes drainage basin, and the paired 

samples WS1945 and WS1965 (orange) from the Jaguay drainage basin.  

 

Below 1000 meters (n=17), δ18O ranges from -17.10 + -13.39 ‰, with a mean of -

16.31‰. The first and third quartile range for samples below 1000 meters is restricted to -16.78 

to -16.37‰. Between 1000 to 2000 meters (n=7), δ18O range from -12.56 to -9.74 ‰ and 

 

 

105 

 

 

average -11.32‰. From 2000-3000 meters (n =9) values are restricted to -12.44 to -10.85 ‰, 

with the mean -11.89‰. The range of δ18O values (n=15) in the 3000-4000-meter bracket is the 

largest, between -17.11 to -9.98 ‰. The mean is -13.19‰ and interquartile range -13.53 to -

12.35 ‰. Values above 4000 meters (n=50) are as low as -17.99 and high as -12.10 ‰. The 

mean is -15.89‰. 

 

Three patterns emerge from the initial investigation. First, the plot shows considerable 

overlap in stable oxygen for waters below 1000 masl and above 4000 masl for the δ18O range 

between -17.5 to -15‰. Second, there is no appreciable distinction for δ18O values between the 

elevations of 1000 to 4000 masl. Third, between the lowest sample (36 masl) and 3000 meters, 

the δ18O values of surface waters coarsely demonstrate a pattern opposite of expectations: 

δ18O values increase with elevation.  

4.3.2 Seasonal Variation in δ18O by Elevation  

 

Next, data were examined by season to determine whether δ18O patterning across 

elevations changes annually (Figure 37). 

4.3.2.1 Dry Season  

 

Below 1000 meters, the elevation trends remain consistent with the overall results. 

During the dry season, elevations between 1000-4000 meters demonstrate a clear inverse 

relationship with δ18O values; patterning for this elevation range matches expectations as 

waters become higher as they decrease in elevation.  

 

106 

 

 

4.3.2.2 Wet Season  

 

In the wet season, the small sample size biases the results. Two samples below 1000 

meters derive from the Río Majes, which reflects little seasonal change. Between 1000 to 3000 

meters, the range of δ18O values is more restricted and lower than for the dry season. The same 

locations in this elevation range were sampled during both seasons; therefore, elevation 

patterning of δ18O changes with season. Between 3000-4000 meters, there are no 

representative samples. Above 4000 (n=14) the wet season range is ~2‰ lower than the dry 

season. The shift to lower values reflects the influx in precipitation. Waters from the Quebrada 

de Rata, the two ‘outlier’ samples, shift to become 2.97‰ higher during the dry season. 

 

 

 

107 

 

 

 
Figure 37: Distribution of δ18O of Andean surface waters by elevation. Data are divided by season. Yellow dots (left) reflect the 2018 and 2019 

dry season samples and blue triangles (right) the 2019 wet season samples.

 

 

108 

 

4.3.3 δ18O Variation Across Watersheds 

 

Data were subdivided into the three watersheds: Jaguay, Majes, and Ocoña. Results are 

presented in Table 12. A box plot of δ18O values for each watershed suggests distinct ranges of 

δ18O (Figure 38). The Ocoña drainage basin comprises 54 samples between 3221 to 4938 masl. 

The range of δ18O in sampled Ocoña surface waters was -17.99 to -12.30‰. Majes samples 

comprised 40 samples, from 36 to 3800 masl. Sampled Majes waters ranged from -17.10 to -

9.74‰, with a mean of -14.02‰.  The Jaguay drainage basin contains only four samples. 

Samples came from 3922 to 4116 masl. The range of sampled Jaguay δ18O values was δ18O -

12.77 to -9.98, and the mean -11.25‰. The elevation range varies significantly across drainage 

basins. A plot of δ18O across elevation is considered for each watershed (Figure 39). 

Table 12: Results Divided by Watershed. 

Watershed 

Number (n=) 

Minimum 

Elevation (masl) 

Maximum 

Elevation (masl) 

Minimum δ18O 

(VSMOW) 

Maximum δ18O 

(VSMOW) 

Mean δ18O 
(VSMOW) 

Ocoña Drainage 

Majes Drainage 

Jaguay Drainage 

54 

40 

4 

3221 

36 

3922 

-17.99 

-17.10 

-12.77 

-12.30 

-9.74 

-9.98 

-15.69 

-14.02 

-11.25 

4938 

3800 

4116 

 
 

109 

 

 

 

 

Figure 38: Dot plot of δ18O range for the three sampled drainage basins. The Ocoña expresses the most 

depleted surface water values overall, and the Jaguay the most enriched values.  

 

110 

 

 

 

 

Figure 39: δ18O variation by elevation divided by individual watersheds. Jaguay drainage is on the left, Majes center, and Ocoña right. Data are 

coded by wet (blue triangles) and dry (yellow circles) seasons. 

 

 

111 

 

 

 

4.3.3.1 Jaguay Drainage Basin 

 

The data for the Jaguay drainage basin is limited by a small sample size (n=4). The 

sample reflects only two distinct water bodies: Quebrada de Rata and Quebrada Palljaruta 

(Figure 40). The minimum δ18O value for Jaguay surface waters was -12.77‰ and maximum -

9.98‰. Both values derive from the Quebrada de Rata. The total range is influenced by 

seasonal change, not variation across water bodies or elevation.  

 

112 

 

 

Figure 40: Map of the Jaguay drainage basin and the Quebrada de Rata samples. 

 

113 

 

 

 

4.3.3.2 Majes Drainage Basin 

 

The Majes drainage basin samples were limited to elevations below 4000 meters. 

Waters follow predicted expectations and become higher with decreasing elevation during the 

dry season. If seasonal variation is considered, δ18O values of surface waters below 3000 meters 

exhibit a near complete reversal in this pattern and appear to become lower with decreasing 

elevation. The outlier to this subset is WS1815, a sample from the Río Andagua at 3070 masl I 

collected the sample further north in the drainage than the other waters in this drainage.   

4.3.3.3 Ocoña Drainage Basin 

 

The Ocoña drainage basin sample set is biased toward high elevation locations above 

3000 meters, including the Pucuncho Basin. The range of spread for δ18O in surface waters is 

5.69‰. The Ocoña reflects diverse waterbodies and stream orders, exhibiting the range of 

variation possible within a single watershed. All waters in this sample set drain into the Río 

Arma before discharging over the plateau rim. WS1962, the Río Arma in Salamanca (3221 masl), 

is the lowest water sample collected in the Ocoña drainage. 

4.3.4 Variation in δ18O by Stream Order 

 

Samples are subdivided their Strahler stream order classification and compared across 

watersheds (Table 13). A map stream orders sampled within the study area (Figure 41). A plot 

of δ18O values for stream order organized by watershed is presented in Figure 42. 

 

 

 

 

114 

 

 

Table 13: Summary of δ18O Variation Among Stream Orders.  

Strahler 

Stream Order 

Number  

(n=) 

1 

2 

3 

4 

5 

30 

26 

17 

2 

15 

Minimum 
Elevation  
(masl) 
3922 

Maximum 
Elevation  
(masl) 
4938 

2889 

1128 

937 

36 

4567 

4375 

3070 

896 

Minimum δ18O 

(VMSOW) 

Maximum δ18O 

(VSMOW) 

Mean δ18O 
(VSMOW) 

-17.74 

-17.58 

-17.99 

-16.70 

-17.10 

-9.98 

-10.85 

-9.74 

-13.39 

-15.19 

-14.00 

-14.91 

-14.86 

-15.04 

-16.49 

 

115 

 

 

Figure 41: Map showing stream order for sampling locations. 

 

116 

 

 

 

4.3.4.1 Jaguay Surface Waters 

 

All samples from the Jaguay watershed reflect first order streams. First order streams in 

the Jaguay drainage basin are more depleted than first order streams in the Ocoña watershed. 

4.3.4.2 Majes Surface Waters 

 

Majes samples encompass all five stream orders. Between 2000 to 4000 meters, first- 

and second-degree stream orders indicate higher δ18O values in surface waters as elevation 

decreases, supporting published expectations. Stream orders 3-5 represent all sampled 

locations below 2000 meters elevation. Contrasting the 2000-4000 elevation range, δ18O values 

of surface waters below 2000 meters reverse direction and indicate more depleted values. The 

lowest δ18O values in the Majes drainage occur below 1000 meters elevation. Almost all 

represent a fifth order stream, the Río Majes, or its associated canals. The single exception the 

Quebrada de Castillo Canal. I estimated the source waters of the canal based on its map 

location. It is possible Quebrada de Castillo represents a misclassification and carries waters 

diverted from the Majes.  

 

Stream order data help explain the low value from the outlier WS1815 (-16.70‰). 

WS1815 was derived from the Río Andagua, a fourth-order stream that becomes a fifth-order 

stream at the Río Colca (-17.10‰) before turning into the Río Majes (-16.55‰). The trajectory 

of δ18O values from the Andagua to the Majes demonstrate the transportation of highland 

waters and highland isotopic values from the interior Andes to the coast. 

 

117 

 

 

4.3.4.3 Ocoña Surface Waters 

 

Waters in the Ocoña represent first to third order streams, and both natural and canal 

waters. The plot indicate that first order streams demonstrate the highest δ18O values, δ18O is 

lower in second order streams than first order streams, and third order streams are lowest 

overall. First and second order streams feed into the Río Arma (order =3). WS1962, the Río 

Arma at Salamanca (3221 masl; -17.11‰), is a third order stream below the plateau. The Río 

Arma carries perennial highland waters to the Río Ocoña before discharging to the Pacific 

Ocean. Like WS1815 in the Majes, this sample likely reflects the general composition of waters 

transported from high elevation zones to the coast.  

 

118 

 

 

Figure 42: Stream order δ18O values for the Jaguay, Majes and Ocoña watersheds. 

 

 

 
 

 

 

119 

 

4.3.5 Range of δ18O Across Elevations Throughout the Central Andes 

 

As drainages were not sampled equally across elevations, I compare these data to 

surface water samples (n=241) published for the Central Andean region by Bershaw et al. 

(2016). The δ18O data collected by Bershaw et al. (2016) represent various latitudes between 13 

to 20ºS and elevations between 27 to 4762 masl (Figure 43). All samples were collected 

between May and June, dry season waters. At 15ºS, my study area falls in the center of 

Bershaw et al.’s study area. Some sample locations were replicated between studies. I removed 

one value from Bershaw et al. (279511-05: value = 5.6‰) as an outlier to the dataset. This 

sample likely reflects very enriched waters from a sitting waterbody that is not representative 

of the surface waters considered in my own data. A summary of the retained data is presented 

in Table 14. 

Table 14: Summary of Data from Bershaw et al. 2016.  

Number of 
Samples (n=) 

241 

Minimum  
Elevation 
(masl) 

27 

Maximum  
Elevation  
(masl) 
4762 

Minimum δ18O 

(VSMOW) 

Maximum δ18O 

(VSMOW) 

-19.60 

-2.90 

Mean δ18O 
(VSMOW) 

-14.68 

 
 

The surface waters sampled for this study are plotted with those from by Bershaw in a 

regression for δ18O and δD with the Global Meteoric Water Line (Figure 44). The plot suggests 

the waters collected by Bershaw et al. include more evaporated waters than those presented in 

this thesis.  

 

 

 

120 

 

 

Figure 43: Map of Bershaw et al. 2016 (circles) and Milton 2018-2019 (triangles) water sample locations. 

δ18O values scaled to be consistent across both datasets.  

 

 

121 

 

 

 

 

Figure 44: Bershaw et al. 2016 (dark grey circles) and Milton’s 2018-2019 (light grey triangles) surface 
water samples plotted in a regression of δ18O and δD. The solid black line is the Global Meteoric Water 

Line (GMWL) δD = 8(δ18O) +10‰. 

 

Figure 45 presents the distribution of δ18O values across all sampled elevations for both 

datasets. The data from Bershaw et al. 2016 range from -2.90 to -19.60‰. The interquartile 

range of Q1 -16.30 to Q3 -13.60‰ overlaps with the interquartile range of the 98 surface water 

 

122 

 

 

samples presented here (-16.92 to -13.16‰). The range of stable oxygen in surface waters 

presented in this thesis is indistinguishable with surface water values from the surrounding 

Central Andes. 

 

A sample collected by Bershaw et al. (2016) from the Ocoña river at 27 masl provides a 

low elevation value for the Ocoña watershed. The value is -16.40‰. Assuming stability in 

surface waters over multiple years (supported by Section 4.2), the Ocoña drainage shows less 

than a 1‰ increase in δ18O between the 3221 masl sampled presented here (-17.11‰) and the 

coastal sample published by Bershaw et al. 2016. This example suggests that a “watershed 

effect” drives a lowland-highland overlap in δ18O surface waters for large drainages in the 

western Central Andes. 

 

123 

 

 

 
Figure 45: The elevation distribution of δ18O in surface waters for both Bershaw et al. 2016 (dark grey 

circles) and Milton 2018-2019 (light grey triangles) data.  

4.4 δ18O for Archaeological Human Provenance Research 

 

To situate the dataset within Andean archaeological literature of mobility research, the 

samples were divided into a series of “place” classifications. These trial explorations help 

determine whether oxygen isotopes are suited to identifying the migration history of an 

 

124 

 

 

individual based on the isotopic signals in their biological remains. I consider multiple geologic 

variables to identify underlying processes influencing the δ18O in regional surface waters. 

Second, I consider two arbitrarily defined “zones” of human residence. These zones refer to 

published archaeological studies: the Andean Plateau (Haas et al. 2017) and an abbreviated 

version of Pulgar Vidal’s (1987) ecological zones originally proposed by Knudson (2009) and 

later revised by Scaffidi and Knudson (2020); Scaffidi et al. (2020). 

4.4.1 Waters On and Off the Plateau 

 

For the first model, the study area was divided into plateau/non-plateau values. 

Because the Andean Plateau occurs at high-to extreme elevations, any tenured stay above the 

plateau rim reflects efficient acclimatization, or developmental or genetic adaptation. Samples 

were inspected in the GIS and coded to reflect whether they fell on or below the plateau. 

Results of the on/off plateau test were plotted using a dot plot to demonstrate the total range 

of observed δ18O values (Figure 46). Plateau waters (3727-4938 masl; n=60) range from -17.99 

to -9.98‰. The mean is -15.25‰. Off plateau waters (36-3559; n=38) range from -17.11 to -

9.70‰. The mean is -14.15‰. Surface water δ18O values are only slightly lower on the plateau 

than off the plateau (<0.5‰ difference). If the Quebrada de Rata samples are not considered, 

all plateau waters are less than -12‰, which would suggest up to a ~2.5‰ difference between 

plateau and non-plateau waters.   

 

 

125 

 

 

Figure 46: Plot comparing the spread of possible δ18O values off (light grey circles) and on (dark grey 

triangles) the Andean Plateau.  

 

4.4.2 Abbreviated Vidal Ecozones 

 

Data were divided into an abbreviated version of Vidal’s (1981) ecological zones 

following Scaffidi and Knudson (2020) and Scaffidi et al. (2020).The designation serves to 

situate these results within the broader literature of stable isotopic migration research in the 

 

126 

 

 

Andes, which frequently focus on the movement of humans between elevations. The three 

zones include a coastal zone (<500 masl), a mid-elevation terrestrial Yunga zone (500-2300 

masl), and a high-elevation, Highland zone (2300+ masl). Coastal zones were expected to 

demonstrate the highest δ18O values, and Highland waters the lowest δ18O values (Knudson 

2009) . 

 

Observed δ18O of surface waters do not demonstrate the predicted distributions when 

plotted by zones (Figure 47). The coastal zone (n=8) exhibits the most depleted δ18O values. The 

mean coastal δ18O value is -16.28‰ and the range -16.60 to -15.19 ‰. The Yunga zone (n=16) 

ranges from -17.10 to -9.74 ‰, the mean is -14.14‰. The highland zone (n=74) ranges from -

17.99 to -9.98 ‰, the mean is -14.82‰ (n=74).  

 

 
 

 

127 

 

 

 

 

Figure 47: Dot plot of the spread of δ18O values observed in the surface waters across three Central 

Andean ecozones. Coastal (<500 masl), Yungas (500-2500 masl), and Highland (>2500 masl).   

 

While the coastal zone is distinct, it is completely encompassed by the range of values in 

both the Yunga and highland zones. The only distinct values from the Yunga zone are waters 

with δ18O values > - 10‰. In the highlands, distinct waters include those > -17.10‰.  

 

128 

 

 

4.4.3 The Humans of Cuncaicha  

 

The drinking water (DW) δ18O values for the humans of Cuncaicha are presented in 

Table 155. The measured δ18ODW from the teeth of each individual is plotted against the study 

area surface water δ18O data in Figure 48 and against the Bershaw et al. comparative δ18O data 

in Figure 49. 

Table 15: δ18ODW, 87Sr/86Sr, and δ13CVPDB values from the humans of Cuncaicha. 

Individual 

AS # 

Age 
14C BP 

Age  
Cal BP 

Tübingen Lab 
Specimen ID 

Period 

δ18ODW 
(VSMOW) 

15-04 

15-03 

15-06 

15-05 

15-07 

3107±20 

3361-3182 

CUND 16 

Late Holocene 

3836±21 

4287-4085 

CUND 19 

Late Holocene 

7689±22 

8536-8386 

CUND 20 

Early Holocene 

7999±31 

8986-8691 

CUND 17 

Early Holocene 

8136±32 

9133-8788 

CUND 18 

Early Holocene 

-16.40 

-15.54 

-13.51 

-14.04 

-15.25 

87Sr/86Sr 

0.706607 

NA 

δ13C 
(VPDB) 

-14.45 

-12.17 

0.706433 

-13.00 

0.706437 

-12.60 

0.706427 

-12.44 

 

 

Figure 48, which plots samples collected from the study area, suggests the δ18Odw 

derived from the enamel carbonates of individuals 15-03, 15-04, 15-06, and 15-07 overlaps 

δ18Osw values observed below 1000 masl and above 4000 masl. 15-05 only intersects values 

observed above 4000 masl from this dataset. The δ18ODW value from individual 15-04 (-16.40‰) 

also crosses the Río Andagua sample from 3070 masl. 

 

 

129 

 

 

 

Figure 48: δ18ODW values for the five humans of Cuncaicha plotted against surface water values from the 

study area. 

 

Figure 49, which includes Bershaw’s et al.’s samples, demonstrates that all Cuncaicha 

individuals’ δ18ODW values may be consistent with lower elevation δ18O surface water values 

observed throughout the Central Andes region. Individuals 15-03, 15-04, and 15-07 have 

δ18ODW values that are found within the entire 5000 masl range. The δ18ODW from individuals 

 

130 

 

 

15-05 and 15-06 are consistent with surface water δ18O values observed at elevations between 

1000 to 5000 masl. 

Figure 49: The δ18ODW for the humans of Cuncaicha plotted against the δ18O values in surface waters 

from the Central Andes.  

 

 

 

 

131 

 

 

CHAPTER 5: DISCUSSION AND CONCLUSIONS 

 

The goal of this work was to develop a surface water δ18O baseline for the study area 

between three archaeological sites, Pampa Colorada 343, Quebrada Jaguay 280, and Cuncaicha 

rock shelter. I asked the questions: (1) does the δ18O composition of surface waters in the 

Pucuncho Basin change seasonally? And (2) Do the surface waters between the three sites 

demonstrate distinct δ18O values at different elevations? I discuss my results and synthesize the 

implications for the research objectives identified in Chapter 1. I conclude with some broader 

insights I encountered through this research.  

5.1 General Comments 

 

The instrumental precision for all measurements was good, so all 98 samples were 

retained for the analysis. The local surface water lines for both the wet and dry season in the 

sampling area are consistent with both the regional meteoric water line as well as the global 

meteoric water line, indicating that the δ18O composition of surface waters in the Andes follow 

predicted isotopic behavior of other regions (Figure 21). The data collected for this study 

represent the first seasonal dataset measuring the degree of δ18O change for surface waters in 

the study area. These data also represent the first set of inter-annual data published for the 

area. Without a decadal record it is not possible to determine whether there are long-term 

trends or shifts in the surface water values in the study area.  

5.2 Seasonal δ18O Variation in Regional Surface Waters  

 

In the research area, the greatest degree of temporal variation for δ18O was observed 

for seasonal samples (mean change: 1.21 per mil). The next greatest change was observed at 

 

132 

 

 

the annual scale (mean change: 0.22 per mil). Daily surface water change provided the smallest 

difference between samples (mean change: 0.10 per mil).  

 

Seasonal change varied among water bodies in the study area. The sample size was too 

small to determine the underlying processes contributing to this variation. However, it is likely 

that groundwater contributes an important factor. Streams primarily fed by groundwater may 

be less sensitive to, or attenuate, δ18O arriving through seasonal precipitation. Comparisons for 

daily, seasonal, and annual change at a single location suggest that the volume of the water 

body (stream order) may also be an important driver of temporal change in δ18O at an isolated 

water source. If this is the case, it is likely because lower volume streams, or shallower water 

bodies, are more sensitive to evaporative processes and therefore reflect atmospheric 

seasonality more readily. 

 

It is important to note that some water bodies do not reliably record a degree of 

seasonal change that is greater than the observed daily or annual change. For example, the Río 

Majes demonstrated a 0.23‰ shift between the 2019 wet and dry seasons but recorded a 

0.34‰ change between the dry seasons of 2015 and 2019. It is fair to note that neither the 

seasonal nor annual change registered in the Río Majes are considered significant; further as a 

low-elevation river, the Río Majes would not inform on a seasonal highland proxy.   

 

Figure 28 and Figure 32 show that two locations, the Río Majes at 383 masl (SEAS-

01/Pair 9) and the Río Colunga at 4333 masl (SEAS-08/Pair 16), demonstrate less change 

between seasons than the baseline for daily change observed regionally. The Majes is a 5th 

order stream and Río Colunga a 2nd order, suggesting stream order may not be the underlying 

control for seasonal shift in surface water δ18O.  

 

133 

 

 

 

Three more seasonal pairs, Iray canal at 2521 masl (SEAS-05/Pair 13), Quebrada 

Japuchaca at 4435 masl (SEAS-10/Pair 18), and Quebrada Pucunchitoyocc at 4717 masl (SEAS-

11/Pair 19), demonstrate less change between seasons than that observed interannually.  All 

three are first order streams, which might suggest that smaller order streams are less sensitive 

to seasonal precipitation; however, the greatest magnitude of seasonal change is also observed 

in first order streams. For example, Quebrada de Rata (SEAS-06/Pair 14) provided the largest 

difference in δ18O values between seasons. The Quebrada de Rata represents the only surface 

water location sampled seasonally in the Jaguay watershed. The Jaguay drainage does not 

extend into the central plateau but is perched at the western edge of the Andes. For this study, 

it is the watershed farthest from the incoming precipitation source. A tentative explanation 

may be that the Jaguay watershed has a smaller catchment that receives less seasonal 

precipitation and is therefore more sensitive to evaporative processes during the dry season, as 

reflected by the degree of seasonal change captured by the Quebrada de Rata. Catchment size 

and groundwater availability will constitute an important variable to consider in future studies. 

 

The data provided here do not help predict which water bodies should demonstrate 

seasonal change, beyond the conclusion that higher order streams may be less sensitive than 

smaller order streams. Further testing should provide greater insight into regional seasonal 

change.  

5.2.1 Prospectus for a Seasonal Proxy 

A seasonal proxy utilizing δ18O will face a number of challenges for the following reasons: 

1.  While seasonal variation does appear to be measurable in regional surface waters, the 

magnitude of δ18O change is not consistent throughout the area. Not all surface waters 

 

134 

 

 

demonstrate a seasonal change that is greater than routine temporal (daily/annual) 

change.  

2.  Considering all locations sampled in the Pucuncho Basin, the total range of natural 

variability in δ18O of surface waters encompasses the seasonal change observed in δ18O. 

For example, values between -16.98 and -13.94‰ δ18O were observed for the Río 

Cuncaicha between the wet and dry seasons of 2019 (a 3.04‰ shift); other water bodies 

throughout the Pucuncho Basin annually (wet and dry season considered together) 

demonstrate a range in δ18O values between -17.74 to -9.74‰. The 8‰ difference 

observed between water bodies across the plateau may obscure seasonal variation 

observed within a single water body in the high-elevation study area. Considering that 

vicuñas migrate (where a shell or speleothem does not), any further proxy development 

will also need to contend with the variation observed across a range of multiple water 

bodies.  

 

It may be possible to control for the drinking source using another isotope and careful 

observational studies of animal behavior. Incremental strontium may provide an approach to 

translate these findings to archaeological material. For now, I recommend a second isotope, 

δ13C, to help isolate a seasonal signal in the Pucuncho Basin.   

5.3 Spatial δ18O Variation in the Central Andes 

 

The elevation plots in Section 4.3 indicate that the δ18O values of surfaces waters in the 

study area record limited change between low and high elevations (Figure 36, Figure 37, Figure 

39, and Figure 42). Results indicate the primary predictors for δ18O in surface waters include the 

positioning and size of a targeted watershed, and the stream order transporting water. The 

 

135 

 

 

overlapping highland and lowland δ18O values are a genuine phenomenon driven by 

hydrogeological processes. WS1815 (Río Andagua, 3070 masl) and WS 1962 (Río Arma, 3221 

masl), as well as the Río Ocoña value derived from Bershaw et al. (2016), suggest that the 

overlapping δ18O values in highland and lowland values is likely a consistent phenomenon 

among higher order streams in the western Andes. Higher volumes of water likely move 

through elevation zones between the highlands and the coast more rapidly, providing little time 

for evaporative processes to affect the composition of surface waters. This creates a 

“watershed effect” where coastal values reflect δ18O values of interior waters. This finding 

supports the hypothesis that high-volume lowland waterways preserve highland isotopic 

signatures, as originally proposed by Knudson (2009). Smaller bodies of water appear to express 

a greater range of variation than larger bodies of water within this dataset. The mean δ18O of 

surface waters decreases with increasing stream order.  

 

In summary higher order streams carry more depleted waters, likely reflecting rapid 

transport of depleted highland waters from the east; this, in turn, leads to overlapping δ18O for 

surface waters at different elevations and latitudes in the western Andes. It remains unknown 

whether δ18O in surface waters demonstrate a similar signal throughout all latitudes and 

longitudes of the Central Andes. However, it is highly likely that all rivers on the western (rain 

shadow) side follow this pattern, due to the homogeneity of the topography on the west-

central edge of the continent. The data from Bershaw et al. 2016 suggests this to be so. More 

intensive sampling to the north and south of the research area should reveal the degree of 

overlap in δ18O of surface waters in the area.  

 

136 

 

 

5.3.1 Reconciling the Watershed Effect and Archaeological Aims 

 

In the course of my writing, Scaffidi (and Knudson 2020a; et al. 2020b) published two 

extensive radiogenic strontium isoscapes using archaeological materials and modern surface 

water samples. They conclude that while strontium isoscapes appear promising, they are 

complicated by the lack of distinct bedrock geology in the Central Andes. The results presented 

within this study suggest a similar story for oxygen isotopes: the patterning of δ18O in surface 

waters between the highlands and the coast, or even between drainages, is difficult to resolve. 

Knudson (2009) discussed the challenges of δ18O for Andean archaeological sites, in terms of 

archaeological δ18O data. She states:  

In conclusion, oxygen isotope data from archaeological enamel and bone hydroxyapatite 
carbonate demonstrate that the variability within each site, even among individuals 
identified as local based on strontium isotope signatures, is greater than the variability 
between sites located in different environmental zones. While some regional trends are 
apparent, it would be inappropriate to attempt to identify the environ- mental zone in 
which an individual lived based on her or his oxygen isotope signature alone. However, 
continuing to explore and better understand the complexities of oxygen isotope 
analyses in the Andes, and beyond, through detailed studies of drinking water sources 
and their isotopic signatures will better enable scholars to use oxygen isotope data to 
identify archaeological residential mobility. (185) 
 
In this thesis I have provided evidence to suggest that the complexities of δ18O extend to 

 

drinking water sources. This leads me to suggest reinterpretations of previous studies that have 

employed δ18O to support arguments of residential mobility or provenance in the Andes. In 

Sections 1.2.1 and 4.4 I presented three examples of how stable isotopes have been used for 

archaeological research. Haas et al. (2017) and Haas et al. (2020) used the δ18O of bone apatite 

from multiple Early to Middle Holocene humans to argue for permanent occupation of the 

Andean Plateau by 9,000 to 7,000 years ago. Similarly, Chala-Aldana et al. (2018) have stated 

 

137 

 

 

that the δ18ODW values derived from the tooth enamel of the Cuncaicha humans are consistent 

with high-elevation waters. The data presented herein suggest that while lower δ18O values are 

observed in the highlands, they are also observable at intermediate and low elevations, 

including the coast. The data I collected overlap a study area first established by Chala-Aldana 

(2017). Therefore, I offer a reevaluation of the Cuncaicha human δ18ODW data. All five 

individuals provided δ18ODW values consistent with high elevation–and low elevation– waters. 

δ18O cannot independently resolve whether the individuals from Cuncaicha lived at the coast or 

in the Andes. The same is likely true for Haas et al. 2017 and Haas et al. 2020, and all other 

migration studies employing δ18O in the Central Andes. It would be worthwhile to test the local 

waters surrounding these sites directly.  

 

 

The locality of faunal and lithic materials found with the humans of Cuncaicha, in 

concert with their final interment location, makes it reasonable that all five individuals did 

indeed spend time at high elevations. However, such a conclusion is not, at this time, justifiable 

on the basis of δ18ODW values alone. Archaeological analyses and refined local isotopic baselines 

promise a valuable, interdisciplinary means of resolving this question. For Cuncaicha, five other 

isotopic measurements may provide the answer. The dietary isotopes of δ13C, δ15N, δ34S 

measured on bone collagen provided results that are consistent with expected high-elevation 

data (Haller von Hallerstein 2018). By testing these isotopes in the environment, we may show 

that a strong elevation signal occurs with these isotopes. For example, sulfur and strontium 

isotopes potentially demonstrate “sea spray effects,” that result in distinct lowland isotopic 

signatures (Nehlich 2015). The absence of such a signal in the Cuncaicha human collagen could 

 

138 

 

 

support the statement that any or all of these individuals remained above the Pacific fog or 

garúa zone (~0-1000 masl).  

 

There are two more productive lines of inquiry. First, it will be worthwhile to reexamine 

the existing 87Sr/86Sr measurements from the Cuncaicha human enamel considering the larger 

datasets recently developed by Scaffidi (and Knudson 2020a; et al. 2020b).  

 

Second, while  already published, the δ13CVPDB from the Cuncaicha human enamel 

carbonates have not been formally considered. Dufour et al. (2014) provide data to suggest 

that δ13C may be a secure indicator of high elevation use. Haas et al. (2017) and Haas et al. 

(2020) provided the δ13C values of bone apatite as a secondary measure of highland 

occupation; though they note that preservation of their burials is challenging and apatite is 

susceptible to chemical alterations distinct from collagen (Bocherens et al. 2011; Lee-Thorp and 

van der Merwe 1991; Lee-Thorp and Sponheimer 2003). In the case of all concerned studies, 

Szpak et al. (2013) have recommended that δ13C should be measured locally for Andean 

environments. The next phase of archaeological inquiry into high-elevation residence should 

attempt to develop multi-isotopic and multi-source (environmental and archaeological) 

baseline. I consider this further through the next section.  

5.4 Isotopic Methods 

5.4.1 The Baseline: Isotopes Through Time, Across Space 

 

The results presented in this work renew attention for the assumptions underpinning 

archaeological applications of isotopic methods. In the last few years, a number of isoscapes 

have been established for archaeological research (Bataille et al. 2018; Bowen 2010; Scaffidi 

 

139 

 

 

and Knudson 2020; Scaffidi et al. 2020; Widga et al. 2010). The general idea behind the isoscape 

is to interpolate isotopic values across a landscape to create a map containing distinct isotopic 

zones (Bowen 2010). I acknowledge that the term isoscape has a particular connotation in 

isotopic research. While all isoscapes can serve as baselines, not all isotopic baselines make an 

isoscape. The benefits of applying the isoscape are apparent: they provide a clear map for 

“sourcing” humans or objects based on their isotopic values.  

 

The data provided in this thesis reveal an important question for stable isotopic 

baselines: does averaging isotopic values across space tell us more than working with individual 

values? The baseline I have presented shows the exact δ18O value of surface waters in the study 

area, as they were observed in 2015, 2018, and 2019. These data provide an important insight 

for baselines: local variation in δ18O parallels and perhaps even exceeds regional variation.  

 

Humans and their animal companions may spend days, months, or years within a single 

“interpolated area” of an isoscape, contacting multiple water sources and ingesting any range 

of values from the landscape. The actual water sources utilized cannot be known; therefore, it 

is safe to assume that any average possible, given the total range of values is an area, could be 

revealed when sampling a human’s biological tissue. This is to say: archaeological research 

considering pedestrian migration (pre horse, car, plane, or boat) must consider the spatial 

resolution of all possible values in a given landscape. It is all values within the range of observed 

variation on a landscape that matters, not the mean. I would revise this statement in the case 

of an isotope that does show distinct values between elevation zones–should that be shown 

convincingly.  

 

140 

 

 

 

I would submit that anthropological research concerning Andean populations, pre-

dating motorized transport, should use caution when applying overly simplified or interpolated 

models.  Rather, true variation baselines, that map the total range of variation within a region, 

will provide the most realistic data for understanding the explicit limitations or potentials of a 

given isotope for archaeology.  

5.4.2 Consider the Dentist: Microdrills and Teeth  

 

Tailing the isoscape, I briefly discuss the methods used to sample δ18O from preserved 

tissue. To my knowledge, all three provenance studies cited within utilized a ‘bulk’ sampling 

approach, where a single sample of the tissue substrate was collected using a micro mill. Such 

approaches produce a single value, and in the process average months to years of time. 

Research questions seeking to identify mobility and seasonality must consider the temporal 

resolution of the values derived from the target record (tooth or bone). Much like an isoscape, 

microdrilling potentially masks important variation over time and across space. High resolution 

methods, such as Laser Ablation and SIMS, may provide a means to resolve where humans 

were going at the actual temporal scale being investigated (Blumenthal et al. 2016). At present, 

these methods require significant developments, such as accurate standards that reflect 

biological materials (Dr. John Valley, personal communication). 

5.5 Next Steps 

5.5.1 Seasonality Research 

 

This research confirmed the presence of a seasonally driven signal in the δ18O of Andean 

surface waters. The next phase of investigation requires controlling for the water body that 

 

141 

 

 

provides the signal. Seasonal stable oxygen research necessitates broader spatial and ecological 

considerations. Tests should target a wide range of waters to reflect a more diverse set of 

stream orders and watersheds. I was not able to use a few of my temporal samples in this 

dataset, due to poor spatial resolution during post-processing. In some locations I sampled too 

close to another water body, which generated confusion as I went to review the exact location 

in the GIS. I removed these locations from the temporal dataset to minimize the potential for 

error. If possible, seasonal sampling locations should be discretely marked and tagged, much 

like site datums. This would ensure the researcher could accurately identify and replicate 

specific water bodies, which will be necessary to capturing the total range of complexity for 

seasonality in surface waters of the Andes.  

 

Strontium isotopes may provide the high-resolution provenance marker required to 

isolate movement between individual water bodies in the archaeological record but as 

discussed in Section 5.4.2, it is essential that the isotopic sampling method does not mask 

distinct movements of the organism across a landscape. High resolution strontium sampling 

may be possible with NanoSIMS (personal communication with directors of the Arizona State 

University National Science Foundation SIMS Workshop, 2019).   

 

It important to consider that the seasonal δ18O change in surface waters could be highly 

variable across latitudes. Multiple sites across the Andes are potential candidates for a seasonal 

analysis using vicuña teeth; therefore, I intend to sample more areas for surface water 

seasonality. For example, the Junín region (three degrees to the north) has a longer wet season 

than the Pucuncho Basin. The extended duration of precipitation may completely attenuate a 

seasonal signal in surface waters, paralleling the effect observed in higher order streams. In 

 

142 

 

 

contrast, sites in southern areas of Peru (Haas et al. 2017), Bolivia (Capriles et al. 2016), and 

northern Chile (Osorio et al. 2011), receive less precipitation and may reflect greater 

amplitudes in surface water seasonality, mirroring the Jaguay Watershed. Like the studies 

presented in this thesis, both predictions may be thrown into question through local testing, 

due to regional differences in hydrology.  

 

As an interim consideration, other isotopes could (and should) be explored. Enamel 

carbonates comprise oxygen and carbon, therefore the latter seems a reasonable course of 

investigation. Serial sampling of modern camelid incisors supports this as promising route for 

further study (Author’s personal data 2018 & 2019). Demonstrating that the δ13C values in 

teeth reflects seasonal change (and not a separate behavioral or environmental variable) will 

require the same baseline work as δ18O. A carbon baseline should test the seasonal and spatial 

variation of δ13C in the preferred forage of camelids.  A baseline modeled after the work of 

Blumenthal et al. (2016) and Samec et al. (2018), would provide the necessary data to interpret 

the δ13C captured by camelid tooth enamel. 

5.5.2 Mobility Research 

 

Human mobility research in the study area, and perhaps the Central Andes as a region, 

should transition to isotopes other than δ18O. The collection of rainwater and drink-preparation 

strategies by humans add to the complexity of interpreting δ18O values derived from teeth, as 

discussed thoroughly by Knudson (2009). Past climates may also complicate the use of δ18O for 

archaeological research as even a ~1-2 per mil shift could push the location of a surface water 

into a different “zone”. For example, Buffen (2008) measured the δ18OICE of the Coropuna ice 

cap and identified Terminal Pleistocene values 8-10‰ lower than present day.  

 

143 

 

 

 

These considerations, in combination with the data presented in this thesis, leads me to 

recommend that oxygen isotopes not be considered further for archaeological studies of 

human mobility and migration in this region. I instead recommend the use of targeted isotopes, 

such as carbon, strontium, and sulfur. As discussed, sulfur and strontium have “sea spray” 

signals that may be useful in determining whether humans lived near the coast or in terrestrial 

zones.  

 

In all cases of provenance: more stable isotopic sampling of the environment is needed. 

This research demonstrates the importance of both local and regional data. Sampling strategies 

concerned with imbibed materials should consider any and all possible drinking/eating sources 

within an area and generated spatial baselines to reflect the full degree of regional variability. 

 

As the final consideration for mobility research, I emphasize a final time: the resolution 

of the method matters. Archaeology has revealed again and again the highly mobile nature of 

the South American people over the last 13,000 years. It is important that our methods capture 

the time frame we are interested in.  

5.6 For Posterity  

 

As of 2020, stable isotopic research in archaeology constitutes a destructive analytical 

method. Destroyed materials are not preserved materials and cannot be studied by future 

generations, or by ourselves as better methods become available. Archaeologists must invest in 

constructing adequate baselines for their methods before applying destructive techniques to 

invaluable archaeological materials. Such an approach will ensure adequate material for the 

future. 

 

144 

 

 

5.7 Summary of Conclusions 

 

The data presented in this thesis offer only preliminary insights into the seasonal and 

altitudinal patterning of stable oxygen of surface waters in the Jaguay, Majes, and Ocoña 

drainages; it is evident we need further testing to define the spatial and temporal constraints of 

isotopes in the Central Andes. I summarize the most direct take-aways from my four results 

sections, then discuss my final considerations.  

1.) First, measurements of δ18O in the surface waters of the study area are consistent 

with global meteoric waters and regional water compositions throughout the western 

Central Andes.  

2.) Second, individual water bodies in the study area demonstrate higher δ18O values 

during the dry season and lower δ18O values in the wet season. Seasonal change in the 

study area may be unappreciable when considered at the system-scale. Determining 

whether stable oxygen is suited to a developing a new archaeological record of 

seasonality will require more systematic and widespread sampling, and probably a 

second isotope to isolate the signals from individual water bodies. A second isotope, 

such as δ13C, may be the most direct route of deriving a seasonal signal from an animal 

record. As the next phase of research for the Pucuncho Basin, I suggest incremental 

sampling of modern camelid teeth and a seasonal isotopic analysis of vicuña forage.   

3.) Third, plateau and lowland surface waters both show an appreciable and overlapping 

range of δ18O. It does not appear possible to predict the δ18O values of surface waters 

based on elevation. Rather, surface water δ18O appears to reflect stream order, 

 

145 

 

 

catchment size, and the geographic position of a waterbody within the watershed. 

Differentiating low and high elevation landscapes will require another isotope. 

 

The data presented within do not support δ18O as an efficient or effective isotope for 

assessing residential mobility in human remains in the Central Andes; they do, however, 

suggest that other information, such as paleoseasonality, may be gained from δ18O analyses of 

faunal teeth. Carefully planned field seasons could resolve outstanding questions about 

landscape patterning in short time.  

5.8 A Case for Interdisciplinary Science 

 

Throughout this research I frequently found myself saying, “uh oh”. Few of my results 

matched my expectations or predictions. By working closely with committee members from 

multiple fields, I had the good fortune to encounter perspectives on isotopes not considered in 

most archaeological literature. Both the establishment of environmental baselines and the 

selection of laboratory methods merits closer attention in future stable isotopic studies 

conducted in the Central Andes. Accurate future assessments of migration will require multiple 

isotopes to resolve lateral and vertical questions of movement. I caution that any isotopic suitor 

in the race for a migration proxy be studied with care across a broad transect that encompasses 

multiple ecological and geological areas. The people of South America are known for 

geographically extensive social and trade networks that transect diverse and sometimes harsh 

environments. Researchers must be prepared to untangle the complexities of both human and 

natural systems in the Andes.  

 

The material and post-processing that goes into making archaeological baselines 

matters. Based on my results, I advocate for archaeologists to invest in developing extensive 

 

146 

 

 

baselines constructed from extant materials and forgo archaeological sampling until they are 

sufficiently confident their approach is justified. Ancient human remains and incremental 

structures, like those found in vicuña teeth, are rare–and therefore invaluable. Environmental 

baselines offer an unprecedented opportunity for archaeologists collaborate on 

interdisciplinary teams and with local communities to bring about high-impact, transdisciplinary 

science. Water samples especially may provide data for environmental monitoring, activism, 

and justice. The waters collected for this study will be available to other researchers for use in 

other research areas. Interdisciplinary collaborations serve as a radical approach to ensuring 

the posterity of archaeological curation practices and the longevity and relevance of our 

discipline; and, based on personal experience, prevent an “uh oh” from becoming “oh no.” 

Something we might all applaud.  

 

 

 

 

147 

 

 

 

 

 

 

 

 

 

REFERENCES 

 

148 

 

 

 
 

 

REFERENCES 

 
Abbott, Mark B., Brent B. Wolfe, Alexander P. Wolfe, Geoffrey O. Seltzer, Ramon Aravena, Brian 

G. Mark, Pratigya J. Polissar, Donald T. Rodbell, Harry D. Rowe, and Mathias Vuille 
2003  Holocene paleohydrology and glacial history of the central Andes using 
multiproxy lake sediment studies. Palaeogeography, Palaeoclimatology, Palaeoecology 
194(1-3):123-138. 

2006  Modelling plateau peoples: the early human use of the world's high plateaux. 
World Archaeology 38(3):357-370. 

2011  Peopling the Tibetan Plateau: Insights from Archaeology. High Altitude Medicine 
& Biology 12(2):141-147. 

1998  Montane Foragers Asana and the South-Central Andean Archaic. University of 
Iowa Press, Iowa City. 

 
Balasse, Marie, Stanley H. Ambrose, Andrew B. Smith, and T. Douglas Price 
 

2002  The Seasonal Mobility Model for Prehistoric Herders in the South-western Cape 
of South Africa Assessed by Isotopic Analysis of Sheep Tooth Enamel. Journal of 
Archaeological Science 29(9):917-932. 

 
Balasse, Marie, Andrew B. Smith, Stanley H. Ambrose, and Steven R. Leigh 
 

2003  Determining Sheep Birth Seasonality by Analysis of Tooth Enamel Oxygen 
Isotope Ratios: The Late Stone Age Site of Kasteelberg (South Africa). Journal of 
Archaeological Science 30(2):205-215. 

 
Bataille, Clement P., Isabella von Holstein, C. C., Jason E. Laffoon, Malte Willmes, Xiao-Ming Liu, 

and Gareth R. Davies 
2018  A bioavailable strontium isoscape for Western Europe: A machine learning 
approach. PLoS One 13(5):e0197386. 

 
Bershaw, John, Carmala N. Garzione, Pennilyn Higgins, Bruce J. MacFadden, Frederico Anaya, 

and Herculano Alvarenga 
2010  Spatial–temporal changes in Andean plateau climate and elevation from stable 
isotopes of mammal teeth. Earth and Planetary Science Letters 289(3-4):530-538. 

 
Bershaw, John, Joel E. Saylor, Carmala N. Garzione, Andrew Leier, and Kurt E. Sundell 
 

2016  Stable isotope variations (δ18O and δD) in modern waters across the Andean 
Plateau. Geochimica et Cosmochimica Acta 194:310-324. 

 
Aldenderfer, Mark 
 

 
Aldenderfer, Mark S. 
 

 

 
 

 

 

 

 

149 

 

 

Blard, P. H., Florence Sylvestre, Aradhna Tripati, Christelle Claude, C. Causse, Anne Coudrain, 

Thomas Condom, Jean L. Seidel, Françoise Vimeux, C. Moreau, J. P. Dumoulin, and 
Jérôme Lavé 
2011  Lake highstands on the Altiplano (Tropical Andes) contemporaneous with 
Heinrich 1 and the Younger Dryas: new insights from 14C, U–Th dating and δ18O of 
carbonates. Quaternary Science Reviews 30(27-28):3973-3989. 

 
Blumenthal, Scott A., Jessica M. Rothman, Kendra L. Chritz, and Thure E. Cerling 
 

2016  Stable isotopic variation in tropical forest plants for applications in primatology: 
Kibale Plant Isotopes. American Journal of Primatology 78(10):1041-1054. 

 
Bocherens, Hervé, Dorothée G. Drucker, Miriam N. Haidle, Hansjürgen Müller-Beck, Susanne C. 

Isotopic evidence (C, N, S) for a high aquatic dietary contribution for a Pre-Dorset 

Münzel, and Yuichi I. Naito 
2016 
muskox hunter from Umingmak (Banks Island, Canada). Journal of Archaeological 
Science: Reports 6:700-708. 

 
Bocherens, Herve, Gilles Pacaud, Petr A. Lazarev, and Andre Mariotti 
 

1996  Stable isotope abundances (13C, 15N) in collagen and soft tissues from 
Pleistocene mammals from Yakutia: Implications for the palaeobiology of the Mammoth 
Steppe. Palaeogeography, Palaeoclimatology, Palaeoecology 126(1-2):31-44. 

 
Bocherens, Hervé, Oliver Sandrock, Ottmar Kullmer, and Friedemann Schrenk 
 

2011  Hominin palaeoecology in Late Pliocene Malawi: First insights from isotopes 
(13C, 18O) in mammal teeth. South African Journal of Science 107(3/4). 

 
Bowen, Gabriel J. 
 

Isoscapes: Spatial Pattern in Isotopic Biogeochemistry. Annual Review of Earth 

2010 
and Planetary Sciences 38(1):161-187. 

 

 

 
 

2017  The Online Isotopes in Precipitation Calculator http://www.waterisotopes.org, 
accessed. 

 
Bowen, Gabriel J., and Justin Revenaugh 
 

2003 
Interpolating the isotopic composition of modern meteoric precipitation: 
ISOTOPIC COMPOSITION OF MODERN PRECIPITATION. Water Resources Research 
39(10). 

 
Bowen, Gabriel J., Leonard I. Wassenaar, and Keith A. Hobson 
 

2005  Global application of stable hydrogen and oxygen isotopes to wildlife forensics. 
Oecologia 143(3):337-348. 

 
Bowen, Gabriel J., and Bruce Wilkinson 
 

2002  Spatial distribution of d18O in meteoric precipitation. 

 

150 

 

 

Braje, Todd J., Jon M. Erlandson, Torben C. Rick, Loren Davis, Tom Dillehay, Daryl W. Fedje, 

Duane Froese, Amy Gusick, Quentin Mackie, Duncan McLaren, Bonnie Pitblado, Jennifer 
Raff, Leslie Reeder-Myers, and Michael R. Waters 
2019  Fladmark + 40: What Have We Learned about a Potential Pacific Coast Peopling 
of the Americas? American Antiquity 85(1):1-21. 

 
Bromley, Gordon 
 

2010  A Glacial-Geologic Approach to Studying Late Quaternary Climate and Ice 
Fluctuation in the Southern Hemisphere:164. 

 
Bromley, Gordon R. M., Brenda L. Hall, Kurt M. Rademaker, Claire E. Todd, and Adina E. 

Racovteanu 
2011  Late Pleistocene snowline fluctuations at Nevado Coropuna (15°S), southern 
Peruvian Andes. Journal of Quaternary Science 26(3):305-317. 

 
Bromley, Gordon R. M., Brenda L. Hall, Joerg M. Schaefer, Gisela Winckler, Claire E. Todd, and 

Kurt M. Rademaker 
2011  Glacier fluctuations in the southern Peruvian Andes during the late-glacial 
period, constrained with cosmogenic 3He. Journal of Quaternary Science 26(1):37-43. 

 
Bromley, Gordon R. M., Joerg M. Schaefer, Gisela Winckler, Brenda L. Hall, Claire E. Todd, and 

Kurt M. Rademaker 
2009  Relative timing of last glacial maximum and late-glacial events in the central 
tropical Andes. Quaternary Science Reviews 28(23-24):2514-2526. 

 

 

 

 

 

 

 
Buffen, Aron 
 

2008  Abrupt Holocene Climate Change: Evidence from a New Suit of Ice Cores from 
Nevado Coropuna, Southwestern Peru and Recently Exposed Vegetation from the 
Quelccaya Ice Cap, Southern Peru Master's, Ohio State University. 

 
Capriles, José M., Juan Albarracin-Jordan, Douglas W. Bird, Steven T. Goldstein, Gabriela M. 

Jarpa, Sergio Calla Maldonado, and Calogero M. Santoro 
2018  Mobility, subsistence, and technological strategies of early Holocene hunter-
gatherers in the Bolivian Altiplano. Quaternary International 473:190-205. 

 
Capriles, José M., Juan Albarracin-Jordan, Umberto Lombardo, Daniela Osorio, Blaine Maley, 

Steven T. Goldstein, Katherine A. Herrera, Michael D. Glascock, Alejandra I. Domic, Heinz 
Veit, and Calogero M. Santoro 
2016  High-altitude adaptation and late Pleistocene foraging in the Bolivian Andes. 
Journal of Archaeological Science: Reports 6:463-474. 

 
Cerling, Thure E., Samuel A. Andanje, Scott A. Blumenthal, Francis H. Brown, Kendra L. Chritz, 
John M. Harris, John A. Hart, Francis M. Kirera, Prince Kaleme, Louise N. Leakey, Meave 
G. Leakey, Naomi E. Levin, Fredrick Kyalo Manthi, Benjamin H. Passey, and Kevin T. Uno 

 

151 

 

 

2015  Dietary changes of large herbivores in the Turkana Basin, Kenya from 4 to 1 Ma. 
Proceedings of the National Academy of Sciences 112(37):11467-11472. 

1999  Carbon Isotope Fractionation between Diet and Bioapatite in Ungulate Mammals 
and Implications for Ecological and Paleoecological Studies. International Association for 
Ecology 120(3):347-363. 

1996  Stable carbon and oxygen isotope analysis of fossil tooth enamel using laser 
ablation. Palaeogeography Palaeoclimatology Palaeoecology 126:173-186. 

 
Cerling, Thure E., and John M. Harris 
 

 
Cerling, Thure E., and Zachary D. Sharp 
 

 
Chala-Aldana, Döbereiner 
 

2017  Assessment of Mobility and Highland Occupation Strategies during the Early 
Holocene at the Cuncaicha Rock Shelter through Strontium and Oxygen Isotopes 
Master's, University of Tübingen. 

 
Chala-Aldana, Döbereiner, Hervé Bocherens, Christopher Miller, Katherine Moore, Gregory 

Investigating mobility and highland occupation strategies during the Early 

Hodgins, and Kurt Rademaker 
2018 
Holocene at the Cuncaicha rock shelter through strontium and oxygen isotopes. Journal 
of Archaeological Science: Reports 19:811-827. 

 

 

 

 

 

 
Craig, Harmon 
 
 
 

 
Dansgaard, Willi 
 

 
DeNiro, M.J. 
 

1961a  Isotopic Variation in Meteoric Waters. Science 133(3465):1702-1703. 

1961b  Standard for Reporting Concentrations of Deuterium and Oxygen-18 in Natural 
Waters. Science 133(3467):1833-1834. 

1954  The O18-abundance in fresh water. Geochimica et Cosmochimica Acta 6:241-
260. 

1964  Stable Isotopes in Precipitation. Tellus XVI. 

 
 
 
Daux, Valérie, Christophe Lecuyer, Marie-Anne Heran, Romain Amiot, Laurent Simon, François 

Fourel, François Martineau, Niels Lynnerup, Hervé Reychler, and Gilles Escarguel 
2008  Oxygen isotope fractionation between human phosphate and water revisited. J 
Hum Evol 55(6):1138-1147. 

1985  Postmortem preservation and alteration of in vivo bone collagen isotope ratios 
in relation to palaeodietary reconstruction. Nature 317:806-809. 

152 

 

 

 

 

DeNiro, M.J., and S. Epstein 
 

1978 
et Cosmochimica Acta 42:495-506. 

Influence of Diet on the Distribution of Carbon Isotopes in Animals. Geochimica 

 
Díaz-Zorita Bonilla, Marta, Dorothée G. Drucker, Pascale Richardin, Verónica Silva-Pinto, 

Marcela Sepúlveda, and Hervé Bocherens 
2016  Marine food consumption in coastal northern Chilean (Atacama Desert) 
populations during the Formative Period: Implications of isotopic evidence (C, N, S) for 
the Neolithic process in south central Andes. Journal of Archaeological Science: Reports 
6:768-776. 

 
Dillehay, Tom D., C. Ramírez, M. Pino, M. B. Collins, J. Rossen, and J. D. Pino Navarro 
 

2008  Monte Verde: Seaweed, Food, Medicine, and the Peopling of South America. 
Science, New Series 320(5877):784-786. 

 
Dornbusch, Uwe 
 

1998  Current Large-Scale Climatic Conditions in Southern Peru and Their Influence on 
Snowline Altitudes (Gegenwärtige großräumige klimatische Bedingungen in Südperu 
und ihr Einfluß auf die Höhenlage der Schneegrenze):41-54. 

 
Drucker, Dorothée G., and Hervé Bocherens 
 

2004  Carbon and nitrogen stable isotopes as tracers of change in diet breadth during 
Middle and Upper Palaeolithic in Europe. International Journal of Osteoarchaeology 
14(3-4):162-177. 

 
Drucker, Dorothée G., Anne Bridault, Keith A. Hobson, Elwira Szuma, and Hervé Bocherens 
2008  Can carbon-13 in large herbivores reflect the canopy effect in temperate and 
 
boreal ecosystems? Evidence from modern and ancient ungulates. Palaeogeography 
Palaeoclimatology Palaeoecology 266(1-2):69-82. 

 
Dufour, Elise, Nicolas Goepfert, Belkys Gutiérrez Léon, Claude Chauchat, Régulo Franco Jordán, 

and Segundo Vásquez Sánchez 
2014  Pastoralism in Northern Peru during Pre-Hispanic Times: Insights from the 
Mochica Period (100–800 AD) Based on Stable Isotopic Analysis of Domestic Camelids. 
PLoS ONE 9(1):e87559. 

 
Eerkens, Jelmer W., Alex DeGeorgey, Howard J. Spero, and Christophe Descantes 
 

2014  Seasonality of Late Prehistoric Clamming on San Francisco Bay: Oxygen Isotope 
Analyses of <i>Macoma nasuta</i> Shells from a Stege Mound, CA-CCO-297. California 
Archaeology 6(1):23-46. 

 
Erlandson, Jon M. 
 

2001  The Archaeology of Aquatic Adaptations: Paradigms for a New Millennium. 
Journal of Archaeological Research 9:287-350. 

 

153 

 

 

Erlandson, Jon M., Torben C. Rick, Todd J. Braje, Molly Casperson, Brendan Culleton, Brian 

Fulfrost, Tracy Garcia, Daniel A. Guthrie, Nicholas Jew, Douglas J. Kennett, Madonna L. 
Moss, Leslie Reeder, Craig Skinner, Jack Watts, and Lauren Willis 
2011  Paleoindian Seafaring, Maritime Technologies, and Coastal Foraging on 
California’s Channel Islands. Science 331(6021):1181-1185. 

2014  An introduction to the bofedales of the Peruvian High Andes:13. 

 
Fonkén, M. S. Maldonado 
 
 
Francken, Michael, Judith Beier, Hugo Reyes-Centeno, Katerina Harvati, and Kurt Rademaker 
 

2018  The human skeletal remains from the Cuncaicha rockshelter, Peru. In New 
Perspectives on the Peopling of the Americas, edited by Katerina Harvati, Hugo Reyes-
Centeno, and Jaeger Gerhard, Tuebingen, Germany. 

1983  Contrasting socioecologies of South America's wild camelids: the vicuña and the 
guanaco. In Advances in the Study of Mammalian Behavior, edited by John F. Eisenberg, 
and Devra G. Kleiman. 

1991  Stable isotope diagrams of freshwater food webs. Ecology 72(6):2293-2297. 

2018  Macrobotanical Report for the Cuncaicha-1 Rock Shelter Site, Department of 
Arequipa, Peru: Terminal Pleistocene and Early Holocene Occupation Components. 

 

 
 

 

 
Franklin, William 
 

 
Fry, Brian 
 
 
Furlotte, Brett A 
 

 
Galaś, Andrzej 
 

 
Gat, Joel R. 
 

2011  The extent and volcanic structures of the Quaternary Andahua Group, Andes, 
southern Peru. Annales Societatis Geologorum Poloniae 81:1-19. 

 
Garreaud, René D., Mathias Vuille, Rosa Compagnucci, and José Marengo 
 

2009  Present-day South American climate. Palaeogeography, Palaeoclimatology, 
Palaeoecology 281(3-4):180-195. 

 
Garreaud, René, Mathias Vuille, and Amy C. Clement 
 

2003  The climate of the Altiplano: observed current conditions and mechanisms of 
past changes. Palaeogeography, Palaeoclimatology, Palaeoecology 194(1-3):5-22. 

1996  Oxygen and Hydrogen Isotopes in the Hydrologic Cycle. Annual Review of Earth 
and Planetary Sciences 24(225-62). 

2000  Atmospheric water balance–the isotopic perspective. Hydrological Processes 
14:1357-1369. 

154 

 

 

Green, Daniel R., Tanya M. Smith, Gregory M. Green, Felicitas B. Bidlack, Paul Tafforeau, and 

Albert S. Colman 
2018  Quantitative reconstruction of seasonality from stable isotopes in teeth. 
Geochimica et Cosmochimica Acta 235:483-504. 

 
Gruver, Stephanie Melissa 
 

2018  Occupational Seasonality and Paleoenvironmental Reconstruction at Quebrada 
Jaguay 280 Master's, Northern Illinois University. 

 
Haas, Randall, Ioana C. Stefanescu, Alexander Garcia-Putnam, Mark S. Aldenderfer, Mark T. 

Clementz, Melissa S. Murphy, Carlos Viviano Llave, and James T. Watson 
2017  Humans permanently occupied the Andean highlands by at least 7 ka. Royal 
Society Open Science 4(6):170331. 

 
Haas, Randall, James T. Watson, Tammy Buonasera, John R. Southon, Jennifer C. Chen, Sarah 

Noe, Kevin Smith, Carlos Viviano Llave, Jelmer Eerkens, and Glendon Parker 
2020  Female hunters of the early Americas. Sci Adv 6(45). 

 
 
Haller von Hallerstein, Sophia 
 

 

 

 

2018  Multi-isotopic paleo-diet reconstruction of hunter-gatherer in the Peruvian Puna. 
MSc, Institut für Naturwissenschaftliche Archäologie, Eberhard-Karls-Universität 
Tübingen. 

 
Herreros, J., Isabel Moreno, Jean D. Taupin, Patrick Ginot, Nicolas Patris, M. De Angelis, Marie-

Pierre Ledru, F. Delachaux, and Ulrich Schotterer 
2009  Environmental records from temperate glacier ice on Nevado Coropuna saddle, 
southern Peru. Advances in Geosciences 22:27-34. 

 
Hillson, Simon 
 

2005  Teeth. 2nd ed ed. Cambridge manuals in archaeology. Cambridge University 
Press, New York. 

 
Iacumin, P., H. Bocherens, A. Mariotti, and A. Longinelli 
 

1996  Oxygen isotope analyses of co-existing carbonateand phosphate in biogenic 
apatite: a way to monitor diagenetic alteration of 

bone phosphate? Earth and Planetary Science Letters 142:1-6. 
 
Iacumin, Paola, V. G. Nikolaev, and Michele Ramigni 
 

2000  C and N stable isotope measurements on Eurasian fossil mammals, 40 000 to 10 
000 years BP: Herbivore physiologies and palaeoenvironmental reconstruction. 
Palaeogeography, Palaeoclimatology, Palaeoecology 163(1-2):33-47. 

 
Introduction to Water Sampling and Analysis for Isotope Hydrology 

Atomic Energy Agency, Vienna. 

 

. International 

 

155 

 

 

Jones, Kevin B., Gregory W. L. Hodgins, and Daniel H. Sandweiss 
 

2017  Radiocarbon Chronometry of Site QJ-280, Quebrada Jaguay, a Terminal 
Pleistocene to Early Holocene Fishing Site in Southern Peru. The Journal of Island and 
Coastal Archaeology:1-19. 

 
Kendall, Carol, and Eric A. Caldwell 
 

1998  Fundamentals of Isotope Geochemistry. In Isotope Tracers in Catchment 
Hydrology, edited by C. Kendall and J.J. McDonnell, pp. 51-86. Elsevier Science,, 
Amsterdam. 

 
Kennedy, Casey D., Gabriel J. Bowen, and James R. Ehleringer 
 

2011  Temporal variation of oxygen isotope ratios (delta18O) in drinking water: 
implications for specifying location of origin with human scalp hair. Forensic Sci Int 
208(1-3):156-166. 

 
Knudson, Kelly J. 
 

2009  Oxygen isotope analysis in a land of environmental extremes: the complexities of 
isotopic work in the Andes. International Journal of Osteoarchaeology 19(2):171-191. 

 
Knudson, Kelly J., Arthur E. Aufderheide, and Jane E. Buikstra 
 

2007  Seasonality and paleodiet in the Chiribaya polity of southern Peru. Journal of 
Archaeological Science 34(3):451-462. 

 
Knudson, Kelly J., Christina Torres-Rouff, and Christopher M. Stojanowski 
 

Investigating human responses to political and environmental change through 

2015 
paleodiet and paleomobility: HUMAN RESPONSES TO POLITICAL AND ENVIRONMENTAL 
CHANGE. American Journal of Physical Anthropology 157(2):179-201. 

 
Knudson, Kelly J., Hope M. Williams, Jane E. Buikstra, Paula D. Tomczak, Gwyneth W. Gordon, 

Introducing δ88/86Sr analysis in archaeology: a demonstration of the utility of 

and Ariel D. Anbar 
2010 
strontium isotope fractionation in paleodietary studies. Journal of Archaeological 
Science 37(9):2352-2364. 

 
Knudson, Kelly J., Sloan R. Williams, Rebecca Osborn, Kathleen Forgey, and Patrick Ryan 

Williams 
2009  The geographic origins of Nasca trophy heads using strontium, oxygen, and 
carbon isotope data. Journal of Anthropological Archaeology 28(2):244-257. 

 
Koch, Paul L., Daniel C. Fisher, and David Dettman 
 

1989  Oxygen isotope variation in the tusks of extinct proboscideans: A measure of 
season of death and seasonality. Geology 17:515-519. 

 

 

 
 

 

156 

 

 

Koch, Paul L., Kathryn A. Hoppe, and S. David Webb 
 

1998  The isotopic ecology of late Pleistocene mammals in North America - Part 1. 
Florida. Chemical Geology 152(1-2):119-138. 

 
Lee-Thorp, Julia A. 
 

2002  Two decades of progress towards understanding fossilization processes and 
isotopic signals in calcified tissue minerals. Archaeometry 44:435-446. 

 
Lee-Thorp, Julia A., Matt Sponheimer, and N. H. Van der Merwe 
 

2003  What do stable isotopes tell us about hominid dietary and ecological niches in 
the Pliocene? International Journal of Osteoarchaeology 13(1-2):104-113. 

 
Lee-Thorp, Julia A., and Nikolaas J. van der Merwe 
 

1991  Aspects of the Chemistry of Modern and Fossil Biological Apatites. Journal of 
Archaeological Science 18:343-354. 

 
Lee-Thorp, Julia, and Matt Sponheimer 
 

2003  Three case studies used to reassess the reliability of fossil bone and enamel 
isotope signals for paleodietary studies. Journal of Anthropological Archaeology 
22(3):208-216. 

 
Longinelli, Antonio 
 

1984  Oxygen isotopes in mammal bone phsophate: A new tool for paleohydrological 
and paleoclimatological research? Geochimica et Cosmochimica Acta 48:385-390. 

 
Luz, Boaz, Eugeni Barkan, Ruth Yam, and Aldo Shemesh 
 

2009  Fractionation of oxygen and hydrogen isotopes in evaporating water. 
Geochimica et Cosmochimica Acta 73(22):6697-6703. 

 
Luz, Boaz, Yehoshua Kolodny, and Michal Horowitz 
 

1984  Fractionation of oxygen isotopes between mammalian bone-phosphate and 
environmental drinking water. Geochimica et Cosmochimica Acta 48(8):1689-1693. 

 
Marín, Juan C., C. S. Casey, M. Kadwell, Kaltouma Yaya, D. Hoces, Juan Olazabal, Raul Rosadio, J. 

Rodriguez, Angel Spotorno, Michael W.  Bruford, and Jane C. Wheeler 
2007  Mitochondrial phylogeography and demographic history of the Vicuña: 
implications for conservation. Heredity 99(1):70-80. 

 
Marín, Juan C., K. Romero, R. Rivera, Warren E. Johnson, and Benito A.  González 
 

2017  Y-chromosome and mtDNA variation confirms independent domestications and 
directional hybridization in South American camelids. Animal Genetics 48(5):591-595. 

157 

 

 

 

 
 
 

 

Mark, Bryan G., and Jeffrey M. McKenzie 
 

2007  Tracing Increasing Tropical Andean Glacier Melt with Stable Isotopes in Water. 
Environmental Science & Technology 41(20):6955-6960. 

 
McInnis, Heather 
 

2006  Middle Holocene Culture and Climate on the South Coast of Peru: Archaeological 
Investigation of the Pampa Colorada. Doctor of Philosophy, Anthropology, University of 
Oregon. 

 
Michener, Robert, and Kate Lajtha 
 

2007  Stable Isotopes in Ecology and Environmental Science. Second ed. Ecological 
Methods and Concepts Series. Blackwell Publishing Malden, MA. 

2013  Large Mammal Zooarchaeology at Cuncaicha: Exploration of Research Potential.  

2016a  Early Domesticated Camelids in the Andes. In The Archaeology of Andean 
Pastoralism, edited by Jose M. Capriles, and Nicholas Tripcevich. 

2016b  Report on the 2015 Season of Excavation and Cuncaicha and the 2015 and 2016 
Study Seasons of the Zooarchaeological Samples from the Site. . 

1998  Measures of Mobility and Occupational Intensity in Highland Peru. In Seasonality 
and Sedentism: Archaeological Perspectives from Old and New World Sites, edited by 
Thomas R. Rocek and Ofer Bar-Yosef, pp. 181-197. Peabody Museum of Archaeology 
and Ethnology, Cambridge, Massachusetts. 

1979  Biology of High-Altitude Peoples - Baker,Pt. American Anthropologist 81(3):708-
710. 

2017  Human Genetic Adaptation to High Altitudes: Current Status and Future 
Prospects. Quat Int 461:4-13. 

 
Moore, Katherine 
 
 
 

 
Moore, Katherine M.  
 

 
Moore, Lorna G. 
 

 
Nehlich, Olaf 
 

 
 

 
 

 

 

2015  The application of sulphur isotope analyses in archaeological research: A review. 
Earth-Science Reviews 142:1-17. 

 
Osorio, Daniela, Donald Jackson, Paula C. Ugalde, Claudio Latorre, Ricardo De Pol-Holz, and 

Calogero M. Santoro 
2011  Hakenasa Cave its relevancce for peopling of the southern Andean Altiplano. 
Antiquity 85:1194-1208. 

158 

 

 

Osorio, Daniela, James Steele, Marcela Sepúlveda, Eugenia M. Gayo, José M. Capriles, Katherine 

A. Herrera, Paula C. Ugalde, Ricardo De Pol-Holz, Claudio Latorre, and Calogero M. 
Santoro 
2017  The Dry Puna as an ecological megapatch and the peopling of South America: 
Technology, mobility, and the development of a late Pleistocene/early Holocene Andean 
hunter-gatherer tradition in northern Chile. Quaternary International 461:41-53. 

 
Passey, Benjamin H., Todd F. Robinson, Linda K. Ayliffe, Thure E. Cerling, Matt Sponheimer, M. 

Denise Dearing, Beverly L. Roeder, and James R. Ehleringer 
2005  Carbon isotope fractionation between diet, breath CO2, and bioapatite in 
different mammals. Journal of Archaeological Science 32(10):1459-1470. 

 
Price, T. Douglas, James H. Burton, and Robert A. Bentley 
 

2002  The characterization of biologically available strontium isotope ratios for the 
study of prehistoric migration. Archaeometry 44:117-135. 

 
Rademaker, Kurt 
 

2006  Geoarchaeological investigations of the Wayñuna site and Alca obsidian source, 
Peru. MSc, Quarternary and Climate Studies, University of Maine, Orono, Main. 

 

 

 
 

 

2017  Chronology, seasonality, and inter-zonal connections in a Terminal Pleistocene-
Early Holocene settlement system, southern Peru. NSF, Peru. 

 
Rademaker, Kurt, Gordon R. M. Bromley, and Daniel H. Sandweiss 
 

2013  Peru archaeological radiocarbon database, 13,000–7000 14C B.P. Quaternary 
International 301:34-45. 

 
Rademaker, Kurt, Michael D. Glascock, Bruce Kaiser, David Gibson, Daniel R. Lux, and Martin G. 

Yates 
2013  Multi-technique geochemical characterization of the Alca obsidian source, 
Peruvian Andes. Geology 41(7):779-782. 

 
Rademaker, Kurt, and Stephanie Melissa Gruver 
 

2018  Chronology and seasonality at Quebrada Jaguay:Implications for the early 
settlement of South America. Chicago, IL. 

 
Rademaker, Kurt, and Gregory Hodgins 
 

2018  Exploring the chronology of occupations and burials at Cuncaicha rockshelter, 
Peru. In New Perspectives on the Peopling of the Americas, edited by Katerina Harvati, 
Hugo Reyes-Centeno, and Gerhard Jaeger, Tuebingen, Germany. 

 
Rademaker, Kurt, Gregory Hodgins, Katherine Moore, Sonia Zarrillo, Christopher Miller, Gordon 
R. M. Bromley, Peter Leach, David A. Reid, Willy Yépez Álvarez, and Daniel H. Sandweiss 

 

159 

 

 

2012  Early Human Settlement of the High-Altitude Pucuncho Basin, Peruvian Andes 
Dissertation, University of Maine, Orono. 

2018  Variation in Occupation Intensity of Early Forager Sites of the Andean Puna: 
Implications for Settlement and Adaptation:44. 

 
Rademaker, Kurt M. 
 

 
Rademaker, Kurt, and Katherine Moore 
 

 
Rutllant, Jose, and Pablo Ulriksen 
 

 

 
 

2014  Paleoindian settlement of the high-altitude Peruvian Andes. Science 
346(6208):466-469. 

2016  Cuncaicha Rockshelter, a Key Site for Understanding Colonization of the High 
Andes. Current Anthropology 57(1):101-103. 

1979  Boundary-layer dynamics of the extremely arid northern part of Chile. Boundary-
Layer Meteorology 17(1):41-55. 

 
Samec, Celeste T., Hugo D. Yacobaccio, and Héctor O. Panarello 
 

2018  Stable isotope compositions of South American camelids in the Dry Puna of 
Argentina: A frame of reference for the study of prehistoric herding and hunting 
strategies. Journal of Archaeological Science: Reports 18:628-636. 

 
Sandweiss, Daniel H., Heather McInnis, Richard L. Burger, Asunción Cano, Bernardino Ojeda, 

Rolando Paredes, María del Carmen Sandweiss, and Michael D. Glascock 
1998  Quebrada Jaguay: Early South American Maritime Adaptations. Science 281:4. 

 
 
Santana-Sagredo, Francisca, Julia Lee-Thorp, Rick Schulting, and Mauricio Uribe 
 

2018  Mobility in the Atacama Desert, northern Chile, in the Late Intermediate Period 
(AD 900–1450): A re-evaluation using stable isotope analysis. Quaternary International. 

 
Scaffidi, Beth K., and Kelly J. Knudson 
 

2020  An archaeological strontium isoscape for the prehistoric Andes: Understanding 
population mobility through a geostatistical meta-analysis of archaeological 87Sr/86Sr 
values from humans, animals, and artifacts. Journal of Archaeological Science 117. 

 
Scaffidi, Beth K., Tiffiny A. Tung, Gwyneth Gordon, Aleksa K. Alaica, Luis Manuel González La 

Rosa, Sara J. Marsteller, Allisen Dahlstedt, Emily Schach, and Kelly J. Knudson 
2020  Drinking Locally: A Water 87Sr/86Sr Isoscape for Geolocation of Archeological 
Samples in the Peruvian Andes. Frontiers in Ecology and Evolution 8. 

 

 
Schoeninger, Margaret, K. Hallin, H. Reeser, J. W. Valley, and J. Fournelle 
 

Isotopic alteration of mammalian tooth enamel. International Journal of 

2003 
Osteoarchaeology 13(1-2):11-19. 

 

160 

 

 

1995  Stable isotope studies in human evolution. Evolutionary Anthropology 4:83-98. 

Schoeninger, Margaret J. 
 
 
Schoeninger, Margaret J., and K. Moore 
 

1992  Bone Stable Isotope Studies in Archaeology. Journal of World Prehistory 6:247-
296. 

 
Schwarcz, Henry P., and Margaret J. Schoeninger 
 

1991  Stable Isotope Analyses in Human Nutritional Ecology. Yearbook of Physical 
Anthropology 34:283-321. 

 
Sharp, Zachary D., and Thure E. Cerling 
 

 
Silverio, Walter, and Jean-Michel Jaquet 
 

 

 

 

 

 

1998  Fossil isotope records of seasonal climate and ecology: Straight from the horse's 
mouth. Geology 26(3):219. 

2012  Multi-temporal and multi-source cartography of the glacial cover of Nevado 
Coropuna (Arequipa, Peru) between 1955 and 2003. International Journal of Remote 
Sensing 33(18):5876-5888. 

 
Sitzia, Luca, Eugenia M. Gayo, Marcela Sepulveda, Juan S. González, Lucia Ibañez, Alain 

Queffelec, and Ricardo De Pol-Holz 
2019  A perched, high-elevation wetland complex in the Atacama Desert (northern 
Chile) and its implications for past human settlement. Quaternary Research 92(1):33-52. 

 
Smith, Tanya M., Christine Austin, Daniel R. Green, Renaud Joannes-Boyau, Shara Bailey, Dani 
Dumitriu, Stewart Fallon, Rainer Grün, Hannah F. James, Marie-Hélène Moncel, Ian S. 
Williams, Rachel Wood, and Manish Arora 
2018  Wintertime stress, nursing, and lead exposure in Neanderthal children. Science 
Advances (4):eaau9483. 

 
Somerville, A. D., P. S. Goldstein, S. I. Baitzel, K. L. Bruwelheide, A. C. Dahlstedt, L. Yzurdiaga, S. 

Raubenheimer, K. J. Knudson, and M. J. Schoeninger 
2015  Diet and gender in the Tiwanaku colonies: Stable isotope analysis of human bone 
collagen and apatite from Moquegua, Peru. Am J Phys Anthropol 158(3):408-422. 

 
Sponheimer, Matt, and Thure E. Cerling 
 

2014 
Geochemistry, pp. 341-355. Elsevier. 

Investigating Ancient Diets Using Stable Isotopes in Bioapatites. In Treatise on 

 
Szpak, Paul, Christine D. White, Fred J. Longstaffe, Jean-François Millaire, and Víctor F. Vasquez 

Sanchez 
2013  Carbon and nitrogen isotopic survey of northern peruvian plants: baselines for 
paleodietary and paleoecological studies. PLoS One 8(1):e53763. 

161 

 

 

Tanner, Benjamin R. 
 

2001  Lithic Analysis of Chipped Stone Artifacts Recovered From Quebrada Jaguay, 
Peru Master's, University of Maine, Orono. 

 
Tanz, Nicole, and Hanns-Ludwig Schmidt 
 

2010  delta(34)S-value measurements in food origin assignments and sulfur isotope 
fractionations in plants and animals. J Agric Food Chem 58(5):3139-3146. 

 
Thompson, Lonnie G., Ellen Mosley-Thompson, and Keith A.  Henderson 
 

2000 
Maximum. Journal of Quaternary Science 15(4):377-394. 

Ice-core palaeoclimate records in tropical South America since the Last Glacial 

1992  Vicuñas and Llamas: Parallels in Behavioral Ecology and Implications for the 
Domestication of Andean Camelids. Human Ecology 20(4):407-433. 

 
Tomka, Steve A. 
 

 
Troll, Carl 
 

 

 
 
 

 

1968  The Cordilleras of the tropical Americas: aspects of climatic, phytogeographical 
and agrarian ecology. . Carl Troll. Colloquium Geographicum 9., Bonn, Germany. 

 
Tung, Tiffiny A., and Kelly J. Knudson 
 

2018  Stable isotope analysis of a pre-Hispanic Andean community: Reconstructing pre-
Wari and Wari era diets in the hinterland of the Wari empire, Peru. American Journal of 
Physical Anthropology 165(1):149-172. 

 
van Klinken, Gert J. 
 

1999  Bone Collagen Quality Indicators for Paleodietary and Radiocarbon 
Measurements. Journal of Archaeological Science 26:687-695. 

 
Venturelli, G., M. Fragipane, M. Weibel, and D.  Antiga 
 

1978  Trace Element Distribution in the Cainozoic Lavas of Nevado Coropuna and 
Andagua Valley, Central Andes of Southern Peru. Bulletin Volcanologique 41:213-228. 

 
Vidal, Pulgar J. 
 
 
Vimeux, Françoise, Patrick Ginot, Margit Schwikowski, Mathias Vuille, Georg Hoffmann, Lonnie 

1981  Geografiìa del Perú: Las Ocho Regiones Naturales del Perú, Lima. 

G. Thompson, and Ulrich Schotterer 
2009  Climate variability during the last 1000 years inferred from Andean ice cores: A 
review of methodology and recent results. Palaeogeography, Palaeoclimatology, 
Palaeoecology 281(3-4):229-241. 

162 

 

 

Vuille, Mathias, and Raymond S. Bradley 
 

2000  Mean annual temperature trends and their vertical structure in the tropical 
Andes. Geophysical Research Letters 27(23):3885-3888. 

 
Vuille, Mathias, Stephen J. Burns, B. L. Taylor, Francisco W. Cruz, Broxton W. Bird, Mark B. 

 
Zazzo, Antoine, Marie Balasse, and William P. Patterson 
 

2006  The reconstruction of mammal individual history: refining high-resolution 
isotope record in bovine tooth dentine. Journal of Archaeological Science 33(8):1177-
1187. 

 
Zazzo, Antoine, Mattieu Lebon, Laurent Chiotti, Clothilde Comby, Emmanuelle Delque-Kolic, 

Roland Nespoulet, and Ina Reiche 
2013  Can We Use Calcined Bones for C-14 Dating the Paleolithic? Radiocarbon 55(2-
3):1409-1421. 

 

 

 
 

 

Abbott, L. C. Kanner, Hai Cheng, and Valdir F. Novello 
2012  A review of the South American monsoon history as recorded in stable isotopic 
proxies over the past two millennia. Climate of the Past 8(4):1309-1321. 

 
Widga, Christopher C., J. Douglas Walker, and Lisa D. Stockli 
 

2010  Middle Holocene Bison diet and mobility in the eastern Great Plains (USA) based 
on d13C, d18O, and 87Sr/86Sr analyses of tooth enamel carbonate. Quaternary Research 
73(3):449-463. 

 
Yann, Lindsey T., Larisa R. G. DeSantis, Paul L. Koch, and Ernest L. Lundelius 
 

2016  Dietary ecology of Pleistocene camelids: Influences of climate, environment, and 
sympatric taxa. Palaeogeography, Palaeoclimatology, Palaeoecology 461:389-400. 

163