U-Th Dating of Lacustrine Carbonates
by
Christine Y. Chen
A.B., Princeton University, 2013
Submittedin partial fulfillment of the requirements for the degree of
Doctor of Philosophy
at the
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
and the
WOODS HOLE OCEANOGRAPHIC INSTITUTION
February 2020
02020 Christine Y. Chen. All rights reserved.
The author hereby grants to MIT and WHOI permission to reproduce and to
distribute publicly paper and electronic copies of this thesis document in whole or in
part in any medium now known or hereafter created.
Signature redacted
A uth or .....................................
Joint Program in Oceanography/Applied Ocean Science and Engineering
Massachusetts Institute of Technology
and Woods Hole Oceanographic Institution
January 10, 2020
Signature redacted
Certified by.........................................
David McGee
Thesis Supervisor
Massachusetts Institute of Technology
Signature redacted
A ccepted by .................... ...........
(7 Oliver Jagoutz
OF TECHOOY Chair, Joint Committe for Marine Geology and Geophysics
Massachusetts Institute of Technology/
Woods Hole Oceanographic Institution
LIBRARIES
77 Massachusetts Avenue
Cambridge, MA 02139
MITLibraries http-/Iibrariesmit.-dulask
DISCLAIMER NOTICE
Due to the condition of the original material, there are unavoidable
flaws in this reproduction. We have made every effort possible to
provide you with the best copy available.
Thank you.
The images contained in this document are of the
best quality available.

U-Th Dating of Lacustrine Carbonates
by
Christine Y. Chen
Submitted to the Joint Program in Oceanography/Applied Ocean Science and
Engineering
Massachusetts Institute of Technology
and Woods Hole Oceanographic Institution
on January 10, 2020, in partial fulfillment of the
requirements for the degree of
Doctor of Philosophy
Abstract
Carbonates are prevalent in many modern and ancient lacustrine settings, but reconstruc-
tions of past lake levels or environments from such materials have been hindered by poor
chronology. Uranium-thorium (U-Th) dating has the potential to fill a gap in current
geochronological tools for such archives, but past attempts have been confounded by poor
understanding of the complex makeup of lacustrine carbonates, leading to misguided con-
clusions on both the utility of certain geochronological tools as well as the age of these
deposits. This thesis showcases strategies for the successful application of U-Th geochronol-
ogy to two types of lacustrine carbonates: lake bottom sediments and tufa deposits. Chap-
ter 2 presents a systematic approach to U-Th dating carbonate-rich lake sediments using
the ICDP sediment core from Lake Junin, Peru. Chapters 3-5 seek to demonstrate the
descriptive power of combining precise U-Th dates on tufas and other carbonates with
geologic observations of their depositional context at all scales-from the outcrop to the
microscale. Here, the tufas originate from a transect of closed-basin lakes in the central
Andes of northern Chile. With improved sample selection and leveraging of the incontro-
vertible constraints of stratigraphy and coevality, we are able to test the validity of U-Th
data. Combining quality-controlled geochronological constraints with careful characteri-
zation of different carbonate facies can yield new insight on the character of lake level
changes. These case studies offer frameworks for interpreting scattered geochronologic
data of any size or system. By embracing the noise in our data, we now have a richer
understanding of the controls on uranium in these deposits. Of all the lessons learned,
we hold the following as most important: for the determination of the age of lacustrine
carbonates, geologic context-in the form of sedimentological observations, additional geo-
chemical data, and paleoecological descriptions-is of equal importance to the numerical
accuracy and precision of geochronological measurements.
3
Thesis Supervisor: David McGee
Title: Associate Professor
4
Biography
Christine Yifeng Chen (5-ifA) is the only child of Dr. Xinghao Chen (M 18) and Lanying
Tiana Her parents immigrated to the UnitedState s to continue their education
after rules restricting the age of university students prevented them from doing so in their
home country. Her father first arrived in the United States at the age of 27 by way of
New York City on January 20, 1985; he remembers the date well because it was the day
of President Ronald Reagan's second inauguration. Using most of the money he had on
him, he bought a bus ticket to Newark, New Jersey, to make his way towards Rutgers
University, where he had been accepted for the doctoral program in Electrical Engineering
and Computer Science. Her mother joined him two years later after getting accepted to
the same program. After completing a masters degree in 1990, her mother gave birth to
Christine on March 12, 1991. With the help of grandparents from both sides of the family,
her parents raised Christine while her father completed his dissertation in 1993. Wanting
to give Christine the best childhood possible, both of her parents made the decision to
remain in America, despite being far away from family in China.
In 1994, the family moved to the Greater Binghamton Area as both parents started jobs
at IBM in Endicott, NY. There, they lived in a home with a backyard where Christine spent
the majority of her childhood, exploring the nearby woods and creeks with her best friends
down the street. Together, they wrote stories and made drawings about their adventures
together. Second to spending time outside, going to school was her next favorite activity.
She was always stymied by the question, "What is your favorite subject?" because all of
them were fun and interesting to her in different ways. Christine attended grade school
until graduating from Union Endicott High School in the spring of 2009. That fall, she
enrolled at Princeton University on a full-tuition financial aid scholarship. She stepped onto
campus with the intention to major in Geosciences, not only because of her interest in soil
and agriculture, but also because the course catalog indicated that it was the only major to
offer free field trips. Her first geology trip in October 2009 to southern California marked a
major turning point in her life: scrambling down steep slopes of volcanic craters, crawling
through narrow spaces of water-carved canyons, and ascending the majestic star dunes in
Death Valley, she was the happiest she had ever been. The most lasting impression of the
trip was made by Mono Lake and its oddly sculptured spires and knobs of tufa, exposed
due to the lowering of lake levels from human demands on water. The fact that she devoted
her dissertation in part to the study of tufas, many years later, may be more than a happy
coincidence.
5
Acknowledgments
Like all scientists, I stand on the shoulders of giants. This work would not have been
feasible without the efforts of countless others before me. I have added my piece to the
edifice of human knowledge, and I hope others will find it sound, sound enough to build,
modify, or improve on. Likewise, the fact that I am graduating with a PhD from MIT and
Woods Hole was an improbability only made possible by the collective effort of all those
who have invested in me at different points in my life.
Where do I begin? I suppose from the start. My mother and father both came to
America on student visas in the 1980s to continue their education after some delays in
doing so in their home country. I am grateful to the United States of America for providing
them and other immigrants the opportunity to pursue their dreams of self-betterment and
a better life. As they completed their degrees at Rutgers University, they were also able
to experience a slice of the American way, and they noticed that kids here seemed to
have a lot of fun: they played sports, like soccer or basketball; some had hobbies playing
musical instruments or making art; and kids hung out with other kids, often terrorizing
the neighborhood. In short, kids seemed happy. And so, despite being in a foreign place
thousands of miles away from all other family, they decided to stay in America to raise
me, so that I might have a happy childhood. And I did. I am grateful to my parents for
making that sacrifice.
Growing up, I was lucky to live on a street with two other girls of essentially the same
age as me: Debbie Miller and Laura Jurewicz. I cannot think of any two more perfect
people to have grown up with. Somehow, we found each other. These are my strongest
memories: crash courses on red wagons and black sleds; expeditions up the creek that felt
like our secret; climbing Zarathustra and getting pine needles and sap in our hair; scavenger
hunts and board games created from our imagination; riding our bikes recklessly up and
down and around the steep curve of our street; and lighting things on fire that we probably
shouldn't. There are other memories, less clear, but all warm. When I think about the
kind of person I want to be, it is the person who we would have dreamt of being back then.
I am grateful to them for making me feel like I could be myself and still belong.
I had several teachers in grade school who supported my diverse interests and long after-
school stays, Sadly, some of their names now escape me. But I can remember Mrs. Simon,
Mr. and Mrs. Stanko, Mr. Materese, Mrs. Golden, Ms. Henry, Mr. Friend, Ms. Trupp,
Mr. Johnson, Mr. Hubert, Mr. Rinde, and my 6th grade teacher whose birthday is on June
4 and whose favorite music artist is Pink. I am grateful to them for making me feel at
home.
In college, I spent the majority of my time ensconced in the incomparable Guyot Hall,
a timeless building with gargoyles of fossils and even a geologic feature named after it (yes,
not the other way around). There, in Guyot 16, a dusty and old wood-trimmed room with
models of mineral lattices cluttering the ceiling, is where I first met Adam Maloof, to whom
I owe everything. His sincere interest and commitment to my academic development over
6
OR qFMVNRR*RMWr- I - ,
my four years of undergrad profoundly changed the way I saw the world, and my role in
it. He saw potential in me that I did not know I had. I am forever grateful to him for
believing in me, so that I could believe in myself.
On the second floor of Guyot, I met my second family: Jon Husson, Blake Dyer,
Catherine Rose, Brenhin Keller, Kyle Samperton, Clara Blittler, and Liz Lundstrom.
They are the closest people I will ever have to siblings, and I am forever grateful to them
for taking me under their wings. Everyone has now left that nest, but I am so, so proud
of them and all that they have achieved. I also share many wonderful and unforgettable
memories with Frederik Simons and Blair Schoene, who I thank for being models of kind
professors who genuinely care for their students, remembering each one of us as people
first. I think back on these days with warm fondness, as if a sunbeam were shining on my
heart.
My graduate school experience can only be described as a fulfillment of dreams that
I did not realize I had. I have had the incredible privilege of traveling to far-flung and
remote places where few have ever been, as a true explorer and observer of the unknown
and mysterious. Watching NOVA on PBS every week, my eight-year-old self would hardly
dare to believe that I could one day be such a scientist, let alone ever be granted a title like
"National Geographic Young Explorer." It is an impossible dream made real in this version
of the universe. I am grateful to my advisor, David McGee, for enabling these opportunities,
and I appreciate his patience and trust in granting me the broad autonomy to chart this
path for myself. Jay Quade, Kristin Bergmann, Tim Lowenstein, Andrew Ashton, and
Roger Summons have been generous with their time and supportive as my thesis committee
members, mentors, and collaborators, and with luck, I hope those connections continue.
An acknowledgment of all those who have helped me get here would not be complete
without mentioning the many other students, post-docs, researchers, and people who I am
grateful to have known and learned from. To nama a few: Elena Steponaitis, Christopher
Kinsley, Gabi Serrato Marks, Michaela Fendrock, Irit Tal, Adam Jost, Ben Hardt, Chris
Hayes, Rick Kayser, Charlotte Skonieczny, Francois Tissot, Nick Scroxton, Josh Murray,
Yan Zhang, Zixuan 'Crystal' Rao, Jade Fischer, AJ Iversen, Almanzo Seguin, and Ruth
Tweedy as fellow lab mates and friends; Kim Huppert, Justin Stroup, Roger Fu, Maya
Stokes, Sebastian Jimenez Rodriguez, Francisco Gonzilez Pinilla, Hector Orellana, Matias
Frugone Alvarez, and Kristian Olson as the best field partners one could ask for; Jade Zim-
merinan, Blas Valero-Garc6s, and Claudio Latorre as invaluable facilitators of my research;
Arielle Woods, Sophie Lehmann, Liseth P6rez, Antje Schwalb, Rob Hatfield, Joe Stoner,
Don Rodbell, Pedro Tapia, and Mark Abbott as collaborators on the Junin project; and
Melody Abedinejad, Erin Wedding, Ronni Schwarz, Brian Smith, Michael Richard, Kris
Kipp, Meg Tivey, Megan Jordan, Brandon Milardo, Lauren Hinkell, Jen Fentress, Vicki
McKenna, Julia Westwater, and Lea-Fraser for making sure that the sun rises everyday.
There are also a great many friends and colleagues with whom I have made wonderful
memories: Kai Sheng Tai, Julia Yan, Kevin Zhang, Greg Ely, Sharon Newman, David
Wang, Danielle Gruen, Rohini Shivamoggi, Julia Wilcots, Mara Freilich, Marianna Linz,
7
Catherine Wilka, Luis Lena, Erik Lindgren, Eva Golos, Michael McClellan, Jimmy Bra-
mante, Jeehyun Yang, Marjorie Cantine, Brian Green, Xinqian Cui, Gareth Izon, Athena
Eyster, Sam Goldberg, Mukund Gupta, Ainara Sistiaga, Jorsua Herrera, Fatima Husain,
Mike Eddy, Annie Bauer, Ken Ferrier, Alex Andrews, Dan Amrhein, Ben Mandler, Bryan
Kaiser, Paul Richardson, Dino Bellugi, Mike Sori, Jaap Nienhaus, Seulgi Moon, Jean-
Arthur Olive, Niya Groveza, Ben Klein, Jeemin Rhim, Noah Anderson, Tyler Tamasi,
Martin Wolf, Patrick Beaudry, Stephanie Brown, Ellen Lalk, Tyler Mackey, Rose Palermo,
Lizzy Wallace, Billy Shinevar, Varun Madiath, and so many more.
Graduate school also marked another deep awakening, this time about how the world
really works. I made many mistakes, but in the process, I see more clearly the barriers
to transformational progress in our field and discipline. I have also come to a better
understanding of myself. The burden of seeing is a constant weight on my mind, but it is
better than inadvertently becoming part of the problem. For inspiring me and for lessons
learned, I thank countless people on Twitter for opening my mind to ideas from outside of
my bubble. Attitudes change, but only because brave people jump into the fire and make
it happen: I will continue to challenge the status quo and to pay my good fortune forward
whenever I can.
Funding sources: National Science Foundation, Massachusetts Institute of Technology,
Geological Society of America, MIT MISTI, Comer Foundation, American Philosophical
Society, National Geographic Society, Explorers Club.
8
Contents
1 Introduction 19
2 U-Th dating of lake sediments: Lessons from the 700 kyr sediment record
of Lake Junin, Peru 23
2.1 Introduction ..................................... 23
2.2 B ackground . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.2.1 Basic principles of U-Th dating . . . . . . . . . . . . . . . . . . . . 26
2.2.2 Previous work on U-Th dating of lake sediments . . . . . . . . . . . 30
2.2.3 Background on the lake sediments from Lake Junin . . . . . . . . . 33
2.3 M ethods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.3.1 Core sampling for U-Th dating . . . . . . . . . . . . . . . . . . . . . 38
2.3.2 Sample preparation and chemistry for U-Th dating . . . . . . . . . 42
2.3.3 Estimating the initial 230Th correction . . . . . . . . . .... . . . . . 43
2.3.4 Calculating weighted means and uncertainties of samples with repli-
cate analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.3.5 Other corresponding data . . . . . . . . . . . . . . . . . . . . . . . . 45
2.4 Re su lts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
2.5 Curation of U-Th data using threshold criteria . . . . . . . . . . . . . . . . 47
2.6 Understanding the scatter . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
2.6.1 Detrital contamination . . . . . . . . . . . . . . . . . . . . . . . . . 51
2.6.2 Open system uranium remobilization . . . . . . . . . . . . . . . . . 54
2.6.3 Ostracode and mollusc shells . . . . . . . . . . . . . . . . . . . . . . 58
2.7 Modeling the effects of detrital contamination and uranium remobilization 61
2.7.1 Modeling results for ~75 ka-aged samples . . . . . . . . . . . . . . . 62
2.7.2 Modeling results for ~550 ka-aged samples . . . . . . . . . . . . . . 64
9
2.8 Conclusions: the age-depth model for the PLJ-1 splice . . . . . . . . . . . . 65
2.9 Considerations for future U-Th dating of lake sediments . . . . . . . . . . . 67
2.10 Supplementary Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
2.10.1 Methods of U-Th measurements on materials from 1996 piston core 80
2.10.2 Methods of other datasets used to interpret U-Th data . . . . . . . 80
2.10.3 Failure to build isochrons . . . . . . . . . . . .. . . . . . . . . . . . 82
2.10.4 Calculation of SmUiec . . . . . . . . . . . . . . . . . . : . . . . . 82
2.10.5 Parameters and priors used for Bacon age-depth model . . . . . . . 83
2.10.6 Determination of relative paleointensity tie points. . . . . . . . .. 84
3 U-Th dating of tufas from Agua Caliente I, Laguna de Tara and Salar de
Loyoques, northern Chile 99
3.1 Introduction . . . . . . . .... . . . . . . . . . . . . . . . . . . .... . . . . 99
3.2 Modern climate of the Altiplano-Puna plateau . . . . . . . . . . . . . . . . 101
3.3 Previous Work on Past Changes in the SASM . . . . . . . . . . . . . . . . 104
3.4 Study A rea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
3.5 M aterials and M ethods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
3.5.1 Field sampling and shoreline mapping . . . . . . . . . . . . . . . . . 106
3.5.2 U-Th dating of shoreline tufas and other lacustrine deposits . . . . 108
3.5.3 Determination of mineralogyand stable isotope composition of car-
bonates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
3.6 R esults . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
3.6.1 Observed tufa varieties and other lacustrine carbonates . . . . . . . 110
3.6.2 U-Th dating of tufas and other lacustrine deposits . . . . . . . . . . 119
3.6.3 Stable isotope composition of deposits . . . . . . z . . . . . . . . . . 122
3.6.4 Paleoshoreline features and magnitude of lake area changes . . . . . 123
3.7 D iscussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
3.7,1 Relative temporal constraints on lake carbonate and paleoshoreline
formation in Agua Caliente I . . . . . . . . . . . . . . . . . . . . . . 123
3.7.2 U-Th ages of tufa and carbonate deposits and their implications for
past lake level changes . . . . . . . . . . . . . . . . . . . . . . . . . . 124
3.7.3 Comparison with shoreline and sediment core records from the Titicaca-
Uyuni and Miscanti lake basins . . . . . . . . . . . . . . . . . . . . 127
3.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
10
Now 11.11IR" 19"Rpq
3.9 Supplementary Materials . . .... . . . .... . . . . . . . . . . . . . . . . . 137
3.9.1 Differential GPS measurements of shoreline features and sample lo-
cations . . . . . . . . . . . .... ........ .. .... ...137
3.9.2 Calculation of modern lake areas and paleolake areas . . . . . . . . . 137
3.9.3 Differences in 5 234 Uinitial values of lake carbonates between basins 138
3.9.4 Paleoshoreline features in Agua Caliente I and Salar de Loyoques. 138
3.9.5 Potential avenues for future geomorphological research . . . . . . . . 140
4 Honey calcite: gravitational drip cements of lacustrine origin preserve
evidence of rapid, large-magnitude lake level fluctuations 147
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
4.2 Field and Geologic Context . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
4.3 Stratigraphic coherence of U-Th geochronological results . . . . . . . . . . . 154
4.3.1 Origins of the honey calcite cement . . . . . . . . . . . . . . . . . . . 154
5 U-Th dating of tufas from the Miscanti-Miniques-Pampa Varela lake sys-
tem, northern Chile 161
5.1 G eologic Setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
5.2 M ethods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
5.3 R esults . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
5.3.1 Geologic and geomorphic context of tufa deposits . . . . . . . . . . . 164
5.3.2 Characteristics of various tufa facies . . . . . . . . . . . . . . . . . . 166
5.3.3 Results of U-Th dates on tufa deposits . . . . . . . . . . . . . . . . . 168
A Appendix Tables 175
11
12
List of Figures
2-1 The basic principles of U-Th dating. . . . . . . . . . . . . . . . . . . . . . . 28
2-2 Geologic map of Lake Junin and photo of modern carbonate silt deposition 34
2-3 Comparison of radiocarbon and U-Th dates from the 1996 core and the
P LJ-1 splice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2-4 Stratigraphic column of the PLJ-1 splice and locations of samples for U-Th
da tin g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2-5 Core sampling process for U-Th dating . . . . . . . . . . . . . . . . . . . . . 39
2-6 The CMC and RGA facies: core scanning images and U-Th sample locations 40
2-7 The CP facies: core scanning images and U-Th sample locations . . . . . . 41
2-8 Step-by-step application of thresholding criteria . . . . . . . . . . . . . . . . 49
2-9 Cross-plot and histograms of calcium carbonate content and optical lightness 52
2-10 Comparison of U-Th data with elemental concentrations, facies, and thresh-
old criteria result . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
2-11 Comparison of U-Th data with red-green color reflectance, total organic
carbon, and uranium concentration . . . . . . . . . . . . . . . . . . . . . . . 56
2-12 Comparison of U-Th data with ostracod shell color . . . . . . . . . . . . . . 60
2-13 Modeled pathways of U-Th isotopic evolution for samples found -20-23 m
dep th . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
2-14 Modeled pathways of U-Th isotopic evolution for samples found -70-75 in
depth .......................................... 66
2-15 Age-depth model for the PLJ-1 splice . . . . . . . . . . . . . . . . . . . . . 68
2-16 Comparison of uranium concentration and 8 2 3 4 Uiec for three samples of the
C M C facies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
2-17 Comparison of different Bacon age-depth models run with varying lengths
of the thickness parameter . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
13
2-18 Isochron plot of analyses from sample M15 . . . . . . . . . . . . . 87
2-19 Isochron plot of analyses from sample L7...... . . . . . . . . 88
2-20 Isochron plot of analyses from sample LI . . . . . . . . . . . . . . 89
2-21 Isochron plot of analyses from sample K16 . . . . . .. . . . . . . . . 90
2-22 Isochron plot of analyses from samples J5 and J6 . . . . . . . . 91
2-23 Isochron plot of analyses from sample H7 . . . . . . . . . . . . . 92
2-24 Isochron plot of analyses from sample H6 . . . . . . . . . . . . . . 93
2-25 Isochron plot of analyses from samples G13 and G14 . . . . . . . . 94
2-26 Isochron plot of analyses from samples G7 and G8 . .... . . . . 95
2-27 Isochron plot of analyses from samples G6 . . . . . . . . . . . 96
2-28 Isochron plot of analyses from samples F15 . . . . . . . . . . . . . 97
2-29 Isochron plot of analyses from samples F9 . . . . . . . . . . . . . . 98
3-1 Precipitation map of the South American summer monsoon and a map of
the transect of closed-basin lakes under study . . . 102
3-2 Satellite imagery of Agua Caliente I, Laguna de Tara, and Salar de Loyoques 107
3-3 Cartoon elevational distribution of paleoshoreline features, tufa facies, and
other carbonate deposits from Agua Caliente I . . . . . . . . . . . . . . . . 112
3-4 Field photos of encrusting floret tufa and caliche in Agua Caliente I . . . . 113
3-5 Field photos of cone-shaped tufas and nodular, platey carbonate . . . . . . 115
3-6 Field photos of transformed ikaite deposits in Agua Caliente I, Salar de
Loyoques, and Laguna de Tara . . . . . . . . . . . . . . . . . . . . . . . . 118
3-7 U-Th dates from carbonates in Agua Caliente I, Salar de Loyoques, and
Laguna de Tara . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
3-8 U-Th dates from hand samples of transformed ikaite from Laguna de Tara . 121
3-9 Comparison of various proxy records to U-Th dates on tufas and other lake
carbonates in Agua Caliente I, Laguna de Tara, and Salar de Loyoques . . . 128
3-10 Comparison of number of records in the Global Lake Status Data Base by
continent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
3-11 813C and 8180 of tufas and other carbonates from Agua Caliente I, Laguna
de Tara, and Salar de Loyoques . . . . . . . . . . . . . . . . . . . . . . . . . 142
3-12 Histograms of estimated post-processed accuracy of all dGPS measurements
from Agua Caliente I and Salar de Loyoques . . . . . . . . . . . . . . . . . . 143
3-13 Field context and photographs of reworked pieces of tufa with honey calcite 144
14
_ MM.W I "IMI FIRIMPIRRIIIIIII I
3-14 Comparison of the maximum areal extent of Laguna de Tara and Salar de
Loyoques to modern day areal extent of the lake . . . . . . . . . . . . . . . 145
3-15 Cross-sectional transect of the Agua Caliente I and Salar de Loyoques basins 146
4-1 Field photos of honey calcite in association with cone-shaped tufas . . . . . 150
4-2 Petrographic analysis of the honey calcite infiltrating porous tufa in Agua
Caliente I (Panels A-H) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
4-2 Petrographic analysis of the honey calcite infiltrating porous tufa in Agua
Caliente I (Panels J-K) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
4-3 Outcrop photos of sample site AD1O-233 . . . . . . . . . . . . . . . . . . . . 153
4-4 U-Th dates from sample AD1O-233-10 . . . . . . . . . . . . . . . . . . . . . 155
4-5 Camel plot diagram of all U-Th dates from honey calcite from Agua Caliente 1156
4-6 Comparison of U-Th dates of different tufa and carbonate deposits at dif-
ferent elevations in Agua Caliente I . . . . . . . . . . . . . . . . . . . . . . . 157
4-7 8180 data from sample AD10-233-10 . . . . . . . . . . . . . . . . . . . . . . 160
5-1 Overview map of the Miscanti-Mifiiques-Pampa Varela lake system . . . . . 162
5-2 Field photographs of paleoshorelines and tufa deposits in the MMPV lake
basin system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
5-3 The fibrous mat carbonate facies . . . . . . . . . . . . . . . . . . . . . . . . 167
5-4 The cement encrustations facies. . . . . . . . . . . . . . . . . . . . . . . . . 168
5-5 Preservation of charophyte algae by carbonate cement . . . . . . . . . . 169
5-6 Evidence for open system behavior in MMPV tufas . . . . . . . . . . . . . . 172
5-7 Summary of U-Th dates from the MMPV lake system . . . . . . . . . . . . 173
15
16
List of Tables
2.1 Lake sediment sequences using U-Th dating for age control . . . . . . . . . 31
2.2 Other datasets used in the study for comparison with U-Th data . . . . . . 46
2.3 Average values of various U-Th data for each sample, calculated from repli-
cate analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
2.4 U-Th data for replicate analyses of each sample . . . . . . . . . . . . . . . . 73
3.1 Modern and ancient lake areas for Agua Caliente I, Laguna de Tara, and
Salar de Loyoues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
3.4 Stable isotope data of samples from Agua Caliente I and Laguna de Tara . 136
3.5 Modern 5180 and 5D values of waters from Agua Caliente I and Laguna de
Ta ra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
A.1 U-Th data associated with Chapters 3-5 . . . . . . . . . . . . . . . . . . . . 176
A.2 U-Th data associated with Chapters 3-5 . . . . . . . . . . . . . . . . . . . . 178
A.3 U-Th data associated with Chapters 3-5 . . . . . . . . . . . . . . . . . . . . 179
A.4 U-Th data associated with Chapters 3-5 . . . . . . . . . . . . . . . . . . . . 180
17
18
Chapter 1
Introduction
Geology, at its core, is a discipline of science concerned with the reconstruction of events
and processes throughout the history of time. As such, geochronology is of unparalleled im-
portance to the field, providing constraints on past rates of change of fundamental processes
on Earth (e.g., evolution, tectonics, climate change) and testing hypotheses of causal rela-
tionships of critical events in Earth history (e.g., mass extinctions). Before the discovery
of radioactivity (Becquerel, 1896), geologists told time the old fashioned way: by making
careful qualitative observations of rocks. It began with the principles of superposition and
cross-cutting relationships (Avicenna, 1027; Steno, 1669), followed by the principle of faunal
successions (Smith, 1816), which enabled early geologists to correlate strata from around
the world to formulate a geologic timescale of Earth history. This relative timescale was
then put in absolute terms starting with the first application of uranium decay series in the
measurement of geologic age by Arthur Holmes in his paper, "The Association of Lead with
Uranium in Rock-Minerals and Its Application to the Measurement of Geological Time"
(Holmes, 1911).
Since then, our understanding of isotopes and our ability to measure them have im-
proved, and over the past few decades, the increased sophistication of mass spectromet-
ric techniques has led to the proliferation of geochronological measurements of increasing
precision. However, in the excitement of generating and advancing such quantitative mea-
surements, the application of the aforementioned principles of stratigraphy to form relative
constraints on sequences was relegated and at times considered lesser than the numerical
information generated from isotopic analyses. Research efforts were mainly focused on
decreasing analytical uncertainties and increasing numerical accuracy on measurements.
19
Today, mass spectrometry for uranium series isotopes is at a level of technological
advancement such that the amount of material needed for an analysis is of a scale on par
with the sub-cediineter cross cutting and stratigraphic relationshipscommonly observed in
lacustrine sediments and shorezone deposits. These incontrovertible constraints can now
be leveraged as equally powerful information to assess the accuracy of geochronological
data. Furthermore, the precision of individual analyses is now high enough such that
the scatter of dates itself can represent geologically meaningful information rather than
problems related to the analytical measurement ("dispersion" or ''geologic scatter").
This thesis "aims to recombine the age-old axioms underpinning the field of geology
with modern day uranium-thorium (U-Th) dating techniques for the determination of the
age of lacustrine carbonates. These materials are historically considered to be non-ideal
for this geochronological tool and thus often not worth the investment of time, resources,
and personnel. We apply U-Th dating to two types of lacustrine carbonate deposits: lake
sediments and tufas.
Chapter 2 presents a systematic approach to U-Th dating carbonate-rich sediments
from the ~100-n-long drill core from Lake Junin, Peru. Deep sediment cores from long-
lived lake basins are fundamental records of paleoenvironmental history, but the power
of these reconstructions has often been limited by poor age control. U-Th dating has the
potential to fill a gap in current geochronological tools available for such sediment archives.
The U-Th dating results from the sediment core form the foundation of an age-depth model
spanning -700 kyrs. High uranium concentrations (0.3-4 ppm) of these sediments allow
us to date smaller quantities of material, giving us the opportunity to improve sample
selection by avoiding detrital contamination, the greatest limiting factor to the success of
previous U-Th dating efforts in other lake basins. The dates from 174 analyses on 55 bulk
carbonate samples revealed significant scatter that could not be resolved with traditional
isochrons, suggesting that at least some of the sediments have not remained closed systems.
To understand the source of noise in the geochronological data, we first apply threshold
criteria that screen samples by their U/Th ratio, reproducibility, and 8234 Uiniial value. We
then compare these results with facies types, trace element concentrations, carbonate and
total organic carbon content, color reflectance, mineralogy, and ostracode shell color to
investigate the causes of open system behavior.
We find that the greatest impediment to U-Th dating of these sediments is not detrital
contamination, but rather post-depositional remobilization of uranium. After examining U-
Th data in these contexts, we identify samples that have likely experienced the least amount
20
of alteration, and use dates from those samples as constraints for the age-depth model.
Our work has several lessons for future attempts to U-Th date lake sediments, namely
that geologic context is equally important as the accuracy and precision of analytical
measurements when determining the age of sample materials. In addition, we caution
that significant geologic scatter may remain undetected if not for labor intensive tests of
reproducibility achieved through replication. As a result of this work, the deep sediment
core from Lake Junin is the only continuous record in the tropical Andes spanning multiple
glacial cycles that is constrained entirely by independent radiometric dates. As such, this
record is uniquely poised to offer new insights on past climate and environmental changes
in the tropical Andes, complementing and testing the long but tuned sediment records
from Sabana de Bogota to the north (-5°S; Groot et al., 2011) and Lake Titicaca to the
south (-16°S; Fritz et al., 2004, 2007).
Chapters 3-5 focus on U-Th dating of tufas and other lacustrine shorezone deposits. In
arid regions worldwide, extensive build-ups of porous carbonate rock called "tufa" are un-
mistakable evidence for past landscape occupation by pluvial lakes. These tufas frequently
exhibit a rich diversity of architectural structure, morphology, composition, and texture,
but leveraging these characteristics to reconstruct the conditions and processes responsible
for this diversity is often limited by uncertainties in the timing and rate of tufa formation.
Conversely, past attempts at dating tufas have been confounded by poor understanding of
the complex makeup of these deposits, leading to misguided conclusions on both the utility
of certain geochronological tools for tufas as well as their age.
Chapters 3-5 present data from late Pleistocene lake basins in the central Andes for
insight. We demonstrate the descriptive power of combining (1) precise U-Th dates on
tufas and other lake carbonates with (2) geologic observations of their depositional con-
text at all scales-from the outcrop to the microscale. These analyses inform one another:
taking advantage of our ability to U-Th date small (<10 mg) amounts of powder, we use
petrography to improve sample selection, and then test data for internal consistency by
applying physical stratigraphic and coevality constraints. From this, we observe that U-
Th dates on dense (non-porous) or crystalline materials more often yield data that pass
these tests. We also document open system behavior in deposits that would nominally
pass traditional geochemical criteria for valid U-Th dates. Pairing quality-controlled U-Th
dates with outcrop and petrographic observations, we present more nuanced insights on
the timing and variability of past lake levels associated with tufa formation. Dates calcu-
lated from measurements are only as good as the interpretation of the geologic materials
21
utilized-only then should geochronological data be used to infer the timing or duration of
a specific process or event. By combining the geologic context of lacustrine sediments and
tufas with U-Th geochronological constraints, our interpretations are more closely aligned
with that which is the truth.
22
Chapter 2
U-Th dating of lake sediments:
Lessons from the 700 kyr sediment
record of Lake Junin, Peru
2.1 Introduction
Since the founding of the International Continental Scientific Drilling Program (ICDP) in
1996 (Colman, 1996), scientific teams have recovered dozens of deep lake sediment cores
from nearly every continent in the world. Due to their continuity, resolution, and wide ge-
ographic distribution, these sediment records have provided important long-term perspec-
tives on Earth's terrestrial environmental history. As the spatial and temporal coverage of
such records expand, the next step is to combine these records with complementary studies
from marine and ice cores to address longstanding questions about the linkages and causal
relationships among terrestrial, marine, and atmospheric phenomena. Here, the challenge
lies in comparing the timing, rate, and duration of past land surface and ecosystem changes
to those of past events identified elsewhere in the oceans, atmosphere, and other continen-
tal regions. Thus, the extent to which tests for leads and lags in the climate system are
useful is limited not by the quality of environmental proxy interpretation, but rather by
the quality of the temporal constraints.
While ice and marine cores are often amenable to layer counting or anchoring to globally
synchronous reference timescales (e.g., oxygen isotope "chronostratigraphy" in marine sed-
23
iments, methane gas concentrations in ice cores), determining a reliable age-depth model
for long lacustrine sediment sequences is generally more'problematic. Because lake basins
occupy a broad range of environments, each drilling location often contains a site-specific
accumulation of terrigenous and biogenic sediment as well as a unique post-depositional
alteration history influenced by non-climatic processes like tectonics and volcanism. Thus,
aligning suchrecords to external reference timescales (colloquially known as "tuning") fe-
quires a thorough investigation of how global climate events and more proximal geologic
processes affect local paleoenvironmental proxy variability. Proving such relationships con-
vincingly can be-a formidable undertaking, but in the absence of cther data, tuning is often
the only means available to establish time constraints. As a result, such records are lim-
ited in their ability to address climatic questions that are dependent on the relative timing
of events (e.g., Prokopenko et al., 2006; Nowaczyk et al., 2013; Stockhecke et al., 2014;
Francke et al., 2016).
Therefore, when possible, absolute chronological data from radiometric and paleomag-
netic dating techniques are highly desirable and generally serve as first-order constraints
on age-depth models of sediment cores. The success and utility of such methods is depen-
dent on factors such as the availability and quality of datable materials, the time range
of the method, and the adherence to assumptions underpinning each technique within a
given sediment sequence. When these factors align, the resulting independent chronolo-
gies allow for compelling investigations of forcing relationships (e.g., the radiocarbon- and
tephra-based chronologies of Laguna Potrok Aike in Patagonia [Kliem et al, 2013] and
Lake Pet6n Itzi in Guatemala [Kutterolf et al., 2016]). However, problems commonly arise
when suitable dating materials are absent or the true age of the sediments is outside the
applicable temporal range of a method: for instance, datable tephras are rare in most
environments and the radiocarbon method is generally limited to the last 50 ka.
Currently, there exists a gap in comprehensively tested high-precision geochronological
tools in the time interval between 50 and 780 ka, beyond the limit of radiocarbon dating and
up to the most recent geomagnetic reversal (Brunhes-Matuyama), after which paleomag-
netic reversal stratigraphy can be applied. Here, methods like uranium-thorium (U-Th),
cosmogenic exposure, and optically stimulated luminescence (OSL) dating have potential
(e.g., Roberts et al., 2018). However, these systems have mostly been underexplored in
their broad application to lake sediments or have not been refined since improvements in
instrumentation have opened new doors for sample selectivity. Ideally, data from multiple
complementary chronological tools with different operating assumptions can be used to
24
cross-validate one another (e.g., Colman et al., 2006; Shanahan et al., 2013), and in the
process,.reveal information about the nature of uncertainties, andbiases specific to each
technique.
To this end, we present our efforts to U-Th date-the carbonate-rich sediments from the
deep drill core extracted in 2015 from Lake Junin, Peru. Our strategy for sample selection,
tests for internal consistency that leverage stratigraphic coevality constraints, and use of
other corresponding sedimentological, geochemical, and paleoecological data to inform our
interpretations of the U-Th data can serve as a framework for future attempts to apply U-
Th dating techniques to long cores. Our results also indicate that future work to establish
or refine U-Th-based lake sediment chronologies must include deliberate tests that probe
for possible open system behavior or excess "geologic scatter"-unresolved errors that
can affect the accuracy and precision of dates due to unknown geologic complexities not
accounted for in typical uncertainty calculations and corrections (Ludwig and Paces, 2002).
Without a methodical exploration of U-Th data in context of other geologic information,
age-depth models that contain single, standalone U-Th analyses that, at face value, seem
like valid ages, may in fact hide the existence of geologic scatter and therefore be inaccurate.
The organization of this paper is as follows: We first provide a basic overview of the
principles behind U-Th dating and review previous efforts to apply U-Th geochronology
to lake sediments (Section 2.2). After describing the relevant background of Lake Junin
(Section 2.2.3) and our methods for core sampling, U-Th geochemistry, and isotopic mea-
surement (Section 5.2), we then present the results of 174 U-Th analyses from 55 unique
samples (Section 5.3). Of these, only 72 analyses from 18 samples are used in the final
chronology for the core. We explain our screening procedure for evaluating the validity of
each U-Th date (Section 2.5), and then interpret our analyses alongside other sedimento-
logical, geochemical, and paleoecological data to show that uranium remobilization, not
detrital contamination, is the most likely cause for discrepancies in our data (Section 2.6).
We then simulate the impact of detrital contamination and uranium remobilization on the
isotopic evolution of our samples to further support this conclusion (Section 2.7). Using
the U-Th age constraints that pass our criteria and radiocarbon dates from Woods et al.
(2019), we then describe the construction of the age-depth model for the Lake Junin sed-
iment record (Section 2.8). We end with a discussion on the uncertainties in U-Th age
estimates learned from this study and propose considerations for future U-Th dating of
lake sediments (Section 2.9).
Because terminology is important, hereafter, we distinguish between the terms date
25
and age, adopting the convention followed by other geochronologists (e.g., Schoene et al.,
2013; Dutton et al., 2017): a date is a number calculated from a decay equation and
isotopic measurements, whereas ana ge is an interpretation of a date in the context of
other information and represents a geologically meaningful time.
2.2 Background
Thus far, the application of U-Th dating in continental paleoclimate archives has been most
visible and transformative in unrecrystallized corals and dense carbonate precipitates like
cave stalagmites and groundwater vein calcites (e.g., Winograd et al., 1992; Cheng et al.,
2000; Wang et al., 2001). In comparison, U-Th dating of lake sediments has historically
been less straightforward. To place the challenges of our work in this context, in this
section, we briefly describe the basic principles of U-Th dating, the geologic processes in
lake sediments that can compromise the underlying assumptions of this dating system, and
the strategies used by other studies to overcome or account for these issues.
2.2.1 Basic principles of U-Th dating
There are several "uranium-series disequilibrium" dating methods that make use of the
decay chains of various actinide nuclides (e.g., 238 U, 235U; see Bourdon et al., 2003). Unlike
other notable radiometric chronometers such as uranium-lead or potassium-argon, which
compare the concentrations of a parent nuclide to that of its stable daughter product,
uranium-series disequilibrium dating schemes instead compare the activity-the number
of disintegrations per unit time per unit mass of a material-of a parent nuclide to those
of their series of unstable daughter products. These methods estimate time by measuring
the degree to which different daughter isotopes along a decay chain are out of secular
equilibrium, a steady state in which the activity of both the parent and daughter nuclides
are equal (i.e.,-the number of daughter nuclides forming is equal to the number of daughter
nuclides decaying). Because the half-lives of the parent isotopes are much longer than that
of all intermediate daughter products in these decay chains, a material containing the
parent isotope that has remained unperturbed for several million years will have reached
secular equilibrium (i.e., the activity ratio of the parent nuclide to its daughter product
will be equal to 1).
Disturbances to this equilibrium caused by various natural geochemical processes form
26
the basis of uranium-series disequilibrium dating. For example, because of differences in
the solubility of uranium and thorium complexes in natural waters of near-surface and
surface environments, the highly soluble parent uranium is separated from its effectively
insoluble daughter product thorium in a marine or lacustrine carbonate deposit. Once this
separation occurs, the system will follow the laws of radioactivity, restoring equilibrium
between the parent and daughter nuclides at a rate determined by their respective decay
constants. Thus, the timing of carbonate formation is determined by measuring the extent
to which daughter product growth has restored the system to secular equilibrium (i.e.,
the extent to which the activity ratio of the parent nuclide and its daughter product has
returned to unity). Using measurements of relevant activity ratios, a date can then be
calculated from decay equations and constants.
Of the many uranium-series dating techniques available, in this paper, we use the more
widely applied 230 Th-234 U- 238 U disequilibrium dating method, for which "U-Th dating"
commonly serves as shorthand (Fig. 2-1). U-Th dating has been most widely applied in car-
bonate minerals: not only are they nearly ubiquitous in most continental waters, but they
also contain relatively higher amounts of uranium and are less prone to post-depositional
alteration than other lacustrine precipitates, like halite. As previously mentioned, the
conditions that increase the solubility and mobility of uranium tend also to decrease the
solubility and mobility of thorium. In oxic environments, uranium generally assumes its
highest oxidation state (U'+) in the form of the highly soluble uranyl ion (UO2+) which
easily forms stable complexes with carbonate ions (COt~), further enhancing its solubil-
ity. UO+ is then adsorbed onto or structurally incorporated into carbonate mineral host
phases (Langmuir, 1978; Reeder et al., 2000, 2001; Kelly et al., 2003, 2006). In contrast,
thorium is generally very insoluble and immobile in most aqueous environments where
pH > 3, with some exceptions (Chabaux et al., 2003). The solubility of both uranium
and thorium increase significantly when forming complexes with organic ligands like humic
and fulvic acids (Langmuir and Herman, 1980; Halbach et al., 1980; Murphy et al., 1999;
Lenhart et al., 2000). Thus, in most conditions, except for those that are highly reducing
or organic-rich, fluids are enriched in uranium and depleted in thorium, and this extreme
fractionation is preserved when calcium carbonate forms from such waters.
Two equations take advantage of this behavior to form the backbone of U-Th dating.
The first is the 2 30Th age equation:
27
Atomic Number 4000C
90 91 92 soluble
. -- most surtace
234 238U waters 3500,,.
4d a Gy 41r 1in soluble In
most surface
waters
34 3000
' Pa Other Relevant
Isotopes
2500
20 6Pb T I T
stable kgk.6kyrs stalet
2000
3.5
e3.0 chemiral aprooch to'fra oafticn secuor eq br n 15001-
1000
0.5
0 500
4000
-- 3500 .7
3000
2500 -- 3800
M 2000 .0 1.5 2:0 2.5 3:0 3.5
5 1500 "Th/8U activity
g 1000 -500
500
0
0 100 200 300 400 500 600 700 800 900-1000
Time Since Initial Fractionation (kyrs) -1000I
adsorbed
oSources of V and T h lattice-bound
in Lake Sediments -
00 OOi
hosts/sediment 00 OA
constituents
* uranium (diseq.) pure endogenic pure detrital clays and Fe- and Iow-02
lake carbonate detrital alumino- organic 0 uranium (sec. eq.) Mn-oxides
porewaters
carbonate silicates matter (authigenic U)
A detrital thorium I
IDEAL UNSUITABLE. NOTIDEAL FOR U-THDAING
Figure 2-1: The basic principles of U-Th dating. [A] Schematic of the portion of the 21U decay chain
that is relevant for U-Th dating. The half-lives and type of particle emitted during radioactive decay (an
< or P particle) of each isotope are shown. Ultimately, the decay chain ends with the stable 206Pb. [B]
2 38 32 4 3 4
and [C] Panels illustrating the evolution of 23 0 Th/ U activity and measured U (82 Um)-the two
ratios used for the calculation of U-Th dates-after initial fractionation. The three thick red-shrded lines
represent different pathways towards secular equilibrium based on the value of 324 U1 uai (see legend in
bottom of Panel B). Values shown are within the range of values observed in theLake Juninsediments, but
are otherwise arbitrary and selected purely for demonstration. Panel C plots the same curves in Panel B
but in 2 30 Th/ 23 U activity- 2 3 4Ums pace to show the graphical solution to the age equations. Hollow circles
mark the initial isotopic composition of the.sample. Straight gray lines represent solutions to the 2
30Th
age equation (Eq. 2.1, lines labeled in kyrs with black text) and curved graycontours represent solutions
to the 23 U equation (Eq. 2.2, some curves labeled with their 2 "Uiial in gray). The plot of data in
23 0 Th/ 2 3 8U activity-82 3U4 ms pace originates from Edwards,(1988). [D] Schematic of uranium and thorium
sources in lake sediments. Black circles 'represent uranium in 23 Th/ 23 8 U disequilibrium whereas. white
circles represent uranium in secular equilibrium. To simplify, the detrital carbonate and aluminosilicate
constituents represent bedrock-derived material of old age (>2 Ma). The placement of circles and triangles
within or around boxes represents how uranium and thorium are associated with each host: bound within
the crystal lattice or adsorbed to the substrate surface. The box furthest to the right represents the
low-oxygen porewater uranium sink, where uranium changes from a soluble to insoluble valence state and
accumulates authigenically. 28
- 13 A234A 34  1
238U3eT h =+ A231 238UU  A230A -230A  234 _ e-( p- 2 (2.1)
where square brackets around ratios indicate activity ratios; A symbols are decay con-
stants; and t is the date (Bateman, 1910; Broecker, 1963; see Edwards, 1988 and Ivanovich
and Harmon, 1992 for derivation). The 234U/ 238U activity is more commonly expressed
using delta (8) notation, representing the deviation in parts per thousand (permil; %o) of
23 4U/ 238 U from secular equilibrium: 6234U= ([ 234 U/ 238 U]-1) x 1000. From this equation,
it is clear that measuring three key isotopes- 238U, 234U, and 230 Th-allows us to uniquely
solve for t (Fig. 2-1A).
The term in Eq. 2.1 involving 8 234 U exists to account for the enrichment of 234 U over
2 38 U commonly observed in natural waters (Thurber, 1962). This disequilibrium is caused
by the preferential leaching of 234U during water-rock interactions due to its displacement
inside the crystal lattice of the host mineral by the alpha recoil of its parent 234Th (Kigoshi,
1971; Kronfeld, 1974; Fleischer, 1982; Chabaux et al., 2008). After solving for t using
Eq. 2.1, we can use a second equation to determine the starting value of 8234 U at the time
of fractionation (8 2 3 4 Uinitial):
6 2 3 4 Urn = (6234Uinitial)eA234t (2.2)
where the subscript m represents the present measured value. Thus, these two equations
allow us to solve for two unknowns with the measurement of two isotopic ratios.
Fig. 2-lB shows the expected isotopic evolution of 230 Th/ 238U activity and 82 3 4 Uinitial
over time, provided the system remains closed. Fig. 2-iC shows the graphical solution to
Eqs. 2.1 and 2.2: straight, sub-vertical lines represent solutions to the 230 Th age equation
and curved lines emanating from the y-axis are solutions to the 8234 U equation. From this
figure, two observations about this dating system can be made: (1) for a given analytical
error, as the true age of the sample increases, so too does the error in the age estimate
due to the increasing closeness of the age isolines; and (2) although both2 3 0Th/ 238 U and
234 U/ 238U have not yet returned to their secular equilibrium values even after 1 Myrs,
the age isolines eventually are so closely spaced that current analytical abilities cannot
distinguish between a sample of a finite age and a sample of infinite age. This characteristic
and current mass spectrometry techniques dictate the practical limit of U-Th dating at
29
~700 kyrs (Stirling et al., 2000; Edwards et al., 2003; Cheng et al., 2016; Fig. 2-1A).
2.2.2 Previous work on U-Th dating of lake sediments
With all radiometric chronometers, a date is only interpretable as a meaningful age if the
system meets the following criteria for closed-system behavior: (1) all decay products were
absent at the time of formation, or can be corrected for if present; and (2) there was no
gain or loss of any radionuclides after formation other than by radioactive decay. For U-
Th dating of lake sediments, the most common obstacle is the lack of material that fulfill
these criteria. Even carbonate-rich sediments remain difficult to date, as the carbonates
often contain non-ideal constituents or have experienced post-depositional alteration due
to various weathering, transport, and mixing processes common in lake basins (Fig. 2-1D).
Despite the challenge, geochronologists have devised ways to circumvent these issues.
Table 2.1 is a list of lake sediment studies in which U-Th dating was applied, each with
varying degrees of success. We distinguish between studies working with evaporites and
carbonates, as these are the two most common materials used. Success has been limited
primarily by the incorporation of detrital materials that introduce initial 230 Th, which
increases uncertainties and, if not fully corrected for, potential inaccuracies. Detrital con-
tamination is usually even more problematic for non-carbonate evaporites like gypsum or
halite because these materials typically have very low amounts of uranium derived from
precipitating waters. The most common detrital materials found in carbonates ard evap-
orites are clay minerals (aluminosilicates and bulk limestone; Fig. 2-1D). Attempts to
chemically separate detritus from bulk sediment have been made, but selective acid leaches
meant to isolate endogenic material from the detrital component were found to also differ-
entially fractionate the uranium and thorium isotopes in unpredictable ways (Bischoff and
Fitzpatrick, 1991; Luo and Ku, 1991), making sequential acid leaching techniques for age
determination ineffective in all but the most controlled experimental cases (Ku and Liang,
1984; Schwarcz and Latham, 1989).
Thus, most U-Th dating applications of lake sediments have applied corrections for
detrital contamination by processing a series of coeval samples through total sample disso-
lution and then using "isochron" techniques. Here, the long-lived isotope 2 32 Th (Fig. 2-1A)
acts as a tracer of contamination: assuming that the endogenic material contains no 23 2Th
or initial 230 Th, any 232Th detected is attributed to the detrital component, and the accom-
panying amount of detrital 230 Th is assumed to occur at a particular proportion relative
30
Table 2.1: A list of sites and the associated studies which used U-Th dating to develop an age model for a lacustrine sediment sequence.
Also indicated are other types of age constraints used for age model construction, separated into two categories: absolute constraintsa nd
other tie points. In these columns, '1' indicates that the chronological tool played a primary, first-order role in the age model, whereas '2'
indicates that the tool played a second-order role, used only after application of first-order data. If no number appears beneath a column,
this indicates that the chronological tool was not used to construct the age model at the site. The list of chronological tools in the table
header reflects the tools used amongst the selected sites and is not meant to be a comprehensive of all possible tools.
ABSOLUTE CONSTRAINTS OTHER TIE POINTS
Site Country Duration References C4c U-Tha OSLb pm c tephrad strata
___________________________________________________ carb evap. otherOSbpactehdsrta
Lake Junin Peru ~700 ka this study -1 1
Lake Igelsj6n Sweden 12 ka Israelson et al. 1997 1 1
Babicora Basin Mexico 65 ka Metcalfe et al. (2002) 1 1
Salar de Atacama Chile 106 ka Bobst etal. (2001); Lowenstein 1
et al. (2003)
Lake Balikun China 150 ka Ma et al. (2004) 1 2
Death Valley USA 200 ka Li et al. (1996); Ku et al. (1998) 1
Searles Lake USA 3 Ma Peng et al. (1978); Bischoff et al. 1 1 1 1
(1985); Lin et al. (1998)
Qaidam Basin China 4 Ma Luo and Ku (1991); Phillips etal. 1 1 1
(1993); Wang et al. (2013) and
references therein
Dead Sea Israel 230 ka Kaufman et al. (1992); Haase- 1 1 1 1h
Schramm et al. (2004); Stein and
Goldstein (2006); Torfstein et al.
(2015), references therein
Bear Lake USA 250 ka Colman et al. (2006) 1 1 1 1
Great Salt Lake USA 280 ka Balch et al. (2005) 1 1 1 1 1
Lake Titicaca USA 370 ka Fritz et al. (2004, 2007) 1 1 1 1
Lake Bosumtwi Ghana 450 ka Shanahan et al. (2013) 1 1 1 1 1 F2
'Applications of U-Th dating are separated into three categories based on material: carbonates ('carb.'), which includes minerals such
as calcite and aragonite, and evaporites ('evap.'), which includes minerals such as halite and gypsum.
b'OSL' refers to optically stimulated luminescence dating.
'pmag.' refers to the use of paleomagnetic excursions, reversals, and intensity to determine age constraints.
d'tephra' refers to records that match tephras from the sediment sequence to well-dated tephras from other sites.
e'strata' refers to matching stratigraphic units in the core to equivalent units in other well-dated records.
I'tuning' refers to the alignment of proxy data to specific anchor points in the standardized timescales of external records ("wiggle
matching").
"Diatom silica.
hThe drill core from the Dead Sea Deep Drilling Project was also tuned to the LR04 benthic stack (Lisiecki and Raymo, 2005) and
Soreq Cave (Bar-Matthews et al., 2003).
'For the deeper part of the core, lithologic units were tied to the Devils Hole oxygen isotope record (Winograd et al., 1992)
'Peaks in calcium carbonate content were tied to the Vostok CO2 record (Petit et al., 1999).
kAs a test of the age model, the dust record from Lake Bosumtwi was tied to that of EPICA Dome C (Lambert et al., 2008).
i
to 23 Th - an initial 3 0 Th/ 23 2Th ratio.
For this reason, a sample with a higher measured2 3Th/3 2 Th or 283 U/ 23 2Th ratio
(more 3 8 Uleads to more abundant 23 0Th) is considered more "clean," while a sample
with a lower measured 2 3 0 Th/ 23 2Th or 3 8U/ 23 2Th ratio is considered "dirty" (colloquial
terms used in the literature; e.g., Schwarcz and Latham, 1989; Przybylowicz et al., 1991;
Stein and Goldstein, 2006). By plotting the isotope ratios of several analyzed portions
of a single sample with varying amounts of detritus, an isochron line fit through those
analyses can pinpoint the isotope ratios of the endogenic material, and thus provide a date.
This approach is considered more rigorous than leaching methods, but is only applicable
when there is a single, homogenized source ofdetritus with a consistent 32 0 Th/ 23 2 Th ratio
forming one end member of the sample mixtures. Most studies listed in Table 2.1 apply
this isochron method, but the process is labor intensive because at least three analyses
are required for a date, and many more for one that is statistically rigor6us (Powell et al.,
2002).
In some cases where the measured 3 0 Th/2 3 2Th ratio of sample material is sufficiently
high, single-sample dates are possible by applying an initial 230 Th/ 23 2 Th correction that
generously accounts for the full range of possible detrital 230 Th/ 232 Th ratios. The 230 Th
age equation modified to correct for initial 230 Th is as follows:
230Th -232Th 230Th1-AM 1
238U 238U 232ThJ
- e-23t + 34U A230  )(1- e -(A23 34)t) (2.3)
± 238U A2 3 0 - A234
where i refers to the initial value at the time of fractionation (Edwards et al., 1987).
For impure carbonates, this detrital correction is usually the largest contributor to the
uncertainty of the final date, having greatest effect on samples with low uranium or low
23 0Th/ 23 2Th ratios. The impact of this correction decreases with the age of the sample:
with time, radiogenic 230 Th builds up and any initial 230 Th decays away, making the
proportion of radiogenic 230 Th to initial 230 Th more favorable. Single-sample dating has
thus far only been successful in more recent studies where carbonates with high uranium
concentrations (>3 ppm) are available and inductively-coupled plasma mass spectrometers
allow for smaller amounts of material to be processed, making iteasier to avoid detritus
32
when sampling (e.g., Balch et al., 2005; Fritz et al., 2007;seeTble2.1).
Thus, the presence of initial 230Th in.sample materials has viable workarounds. How-
ever, less directly addressed is the issueof possible post-depositional gain or loss of uranium.
In addition to clays, other sediment-constituents like organic matter and Fe-Mn hydroxides
serve as sources of uranium separate from endogenic materials; here, uranium is adsorbed
to the mineral and solid surfaces of these impurities (Ames et al., 1983a,b,c; Porcelli et al.,
1997; Ku et al., 1998; Schmeide et al., 2000; Chappaz et al., 2010; Fig. 2-8D). In theory,
utilizing these other uranium sources for dating can be satisfactory if the uranium has
remained immobile since their initial incorporation, as they are initially without 230Th
and would accumulate radiogenic 230 Th with time. Indeed, uranium associated with or-
ganic matter and clays enclosed in evaporites has been beneficial to the dateability of such
low-uranium deposits (e.g., Ku et al., 1998). However, adsorbed uranium is far more sus-
ceptible to post-depositional remobilization than uranium bound within the crystal lattice
of carbonates (Alam and Cheng, 2014). Furthermore, organic matter and clays can also
adsorb additional uranium introduced to the materials via fluid flow, for instance when
low-oxygen porewaters render uranium insoluble and cause it to accumulate authigenically
(Fig. 2-iD; Yliruokanen, 1980; Bone et al., 2017).
Due to this capacity for organic matter to uptake uranium, there have been some
attempts to date peats in highly organic-rich sediments that exhibit high uranium concen-
trations of 1-100 ppm (Van Der Wijk et al., 1986; Rowe et al., 1997; Geyh and Mu, 2005;
Frechen et al., 2007), but open system behavior is commonly evidenced by age reversals
or anomalous uranium isotope values in these materials. The sediment sequence at Lake
Junin is interspersed with thick peat and organic-rich mud layers throughout its length,
signaling that the lake has likely experienced considerable changes in lake level and redox
conditions. In the following section, we provide further details about these sediments.
2.2.3 Background on the lake sediments from Lake Junin
Lake Junin (11.0°S, 76.2°W, 4100 m a.s.l.; Fig. 2-2) was targeted as a site for deep drilling
because of its potential to yield the first continuous, absolutely dated record in the trop-
ical Andes that spanned multiple glacial-interglacial cycles. Located on the high plateau
between the eastern and western cordilleras of the central Peruvian Andes, this relatively
shallow (<15 m) lake is the largest water body located entirely within Peru, occupying an
area of ~300 km 2 fringed by marshlands and dense sedge mats (Wright, 1983). Bedrock
33
9. . >250 ka-aged
PN glacialmorames
20°S
80°W 60°W 40°W
LEGEND
o ICDP drill sites
• transect cores
0 1996 piston core
marshlands
floating sedge mats
lake water
200 mcontour interval carbonate mar
Figure 2-2: [A] Map of Lake Junin and its drainage basin, and the location of the three ICDP
drill sites (yellow circles), the nine Livingstone core locations taken along a transect across the
lake (black circles), and the 1996 piston core (white square; Seltzer et al., 2000). The composite
splice for Lake Junin is composed of cores from Site 1 and two transect cores from the center of
the basin (Hatfield et al., 2019). [B] More recently deposited carbonate silt found among the sedge
mats fringing the lake margins. The carbonate silt is subaerially exposed during the dry season
(June-July-August) when lake levels drop -1-2 m. For scale, small white specks in the shallow
water are Chilean flamingos (-1 in). Photograph-taken by Charles Casey from the western shore,
facing approximately northwest across the lake.
consists primarily of Paleozoic-Mesozoic marine carbonates, with some exposure of pre-
Cambrian crystalline silicate rocks along the eastern cordillera (Cobbing et al., 1981). The
lake owes its origin to >250-ka-aged coalescing glacial outwash fans that dam the northern
and southern ends of the lake, respectively (Hansen et al., 1984). Moraine mapping and
cosmogenic exposure ages from boulders on moraine crests indicate that the lake was not
overridden by glaciers or ice at any time in the last 1 million years (Smith et al., 2005a,
2005b), making it one of the few studied lakes in the Andes that predates the maximum
extent of glaciation.
Previously extracted -short (~20-25 m) cores spanning the last. -50 kyrs revealed that
the lake sediments consist of alternating packages of fine-grained glaciogenic silt and endo,
1
genic carbonate silt deposited at a high rate (0.2-1.0 mm yr- ; Hansen et al., 1984; Seltzer
et al., 2000). The carbonate silts are interpreted to have formed similar to the way such
silts form in present day, in which springs and streams supersaturated in calcite enter the
34
fringing wetlands along the western side of the lake and precipitate carbonate on rooted
macrophytes (Flusche et al., 2005; Rodbell et al., 2012). Based on the modern carbonate
production processes observed, it was hypothesized that a longer core would contain more
carbonate-rich sections, coinciding with warm interglacial and interstadial periods when
retreating piedmont glaciers allowed for the formation of marginal wetlands that isolated
the basin center from detrital sediment input (Rodbell et al., 2012).
Proving such a temporal link between carbonate deposition, periods of reduced ice
cover, and past warm intervals in a longer core would rest on the reliability of the age-depth
model. Thus, we conducted a pilot study to determine if U-Th dating could be applied to
bulk samples of the carbonate silts. Success would demonstrate than an independent U-Th-
based chronology could support a long sediment record from this site, providing motivation
for deep drilling. Sample material came from the 1996 piston core taken by G.O. Seltzer and
D.T. Rodbell from the shallow western margin of the lake (Fig. 2-2A). The results from
this initial test were encouraging: most of the U-Th analyses attempted on carbonates
from the upper 10 m of the core were consistent with the chronology produced by 4 C ages
(Fig. 2-3). The three outliers may have been influenced by open system behavior of mollusc
shell fragments, which have been previously shown to post-depositionally uptake uranium
(Blanchard et al., 1967; Kaufman et al., 1971; McLaren and Rowe, 1996). The preliminary
results also revealed that Lake Junin carbonates have high uranium concentrations (0.3-2
ppm) and low detrital content, with ratios of radiogenic 230Th to initial 230 Th that are 10
times greater than sediments from Lake Titicaca (Fritz et al., 2007) and the Great Salt
Lake (Balch et al., 2005).
Following project approval, the uppermost -100 m of sediment was drilled and cored
in eleven holes across three sites in August 2015. This paper focuses only on sediments
recovered from Site 1, the deepest core extracted from the approximate depocenter of the
lake. We work primarily from the PLJ-1 splice, which is comprised of core sections from
four of the five holes at Site 1 and core sections from two Livingstone transect cores close
to the lake depocenter (Fig. 2-2). More specifics regarding the coring operation and the
subsequent generation of the PLJ-1 splice are described in Hatfield et al. (2019); hereafter,
all references to depth in the Lake Junin core refer to the core composite depth below lake
floor (CCLF). For complete information on the radiocarbon dates constraining the first
~50 kyrs of the record, we refer the reader to Woods et al. (2019).
Here we briefly describe the stratigraphy of the PLJ-1 splice; a full description will be
detailed in subsequent publications elsewhere. Broadly, the prediction that a long core
35
--M
25
1996 PU-1
20 1C.
U-Th e
15
-5
0
0 2 4 6 8 10
Depth(m)
Figure 2-3: Comparison of radiocarbon (gray squares)andU-Th(bluecircles)datesfrom the
1996 core and the PLJ-1 splice. Radiocarbon data are from Seltzer et al. (2000) for the 1996 core
and Woods et al. (2019) for the PLJ-1 splice. Note that most U-Th data shown represent a mean
of multiple analyses; see Table 2.3 for details. Circled in red are outlier U-Th analyses from the
1996 core-that contained abundant mollusc shell fragments, and thus may have been affected by
post-depositional uptake of uranium, biasing dates to be younger than the true age. Based on
the PLJ-1 data, the inferred sedimentation rate (not normalized by dry bulk density) at the lake
depocenter over the last, 25 kyrs is ~0.3 m kyrm , which is ~50-60% slower than that of the 1996
core located on the western lake margin (Fig. 2-2).
from Lake Junin would also contain alternating packages of carbonate and glaciogenic
sediment was correct: ~10 m thick packages of cream-colored carbonate silt alternate
with ~10-15 m thick intervals of dark gray, fine-grained carbonate-rich glaciogenic silt
throughout the length of the core until ~85 m, where a thick package of carbonate-rich
sand occurs (Fig. 2-4A). The mean grain size of this bed was incompatible with the drilling
tools during core extraction, preventing deeper core recovery. Peat and organic-rich mud
layers of -1- to 50-cm thickness punctuate both the carbonate and glaciogenic silt intervals
and contain abundant microfossils that suggest that the peats represent times of wetland
encroachment towards the lake center during lake level lowstands (Woods et al., 2019).
Despite this interpretation, there is no stratigraphic evidence of any depositional hiatus
or unconformity throughout the core, suggesting that the drill site has been submerged,
however shallow in depth, for the duration of the record (Rodbell, Abbott, et al., in prep.).
36
0- B STRATiGRAPHIC COLUMN LEGEND
Creamy, laminatedto banded carbonate mud
10- with minor peat and organic-rich mud layers
Variegated, laminated carbonate silt
20-
E
-- Interbedded peat, organic-rich mud layers
0 0 - and gray, banded to massive carbonate silt
o 30 77- 7 -
Grey, banded to massive, carbonate silt with
4- minor peat layers
0 - L Grey, laminated carbonate silt with minor
peatlayers
50-
Mottled (reddish, greenish) laminated
-- E carbonate silt
0.
a: 60-- Mottled (reddish, greenish) fine carbonate
r; sand and silt
0E
0 70 - Mottled (reddish, greenish) carbonate sand0.
N and silt
80 . Depths of core images featured in Figs. 6-7
110nammem=
U-TH SAMPLE DEPTHS90-_
Included in age-depth model
Not included
100-
0246810
# of samples
Figure 2-4: [A] Stratigraphic column of the PLJ-1 splice, described following the core description
scheme proposed by Schnurrenberger et al. (2003). Lithologies in the stratigraphic column legend
are listed in ascending order of mean grain size (smaller at the top, larger at the bottom). [B]
Rectangular bars mark the depths of the cores featured in Figs. 2-6 and 2-7. [C] Stacked histogram
of depths of samples collected for U-Th dating, in which blue represents samples that yielded dates
included in the age-depth model and gray represents samples that yielded dates that were not
included.
37
2.3 Methods
2.3.1 Core sampling for U-Th dating
Within the U-Th geochronology community, there is a common expectation that samples
with the following characteristics, regardless of substrate type( speleothems sediments.,
tufas, et cetera), have greater potential for success: light in color (considered an indicator
of sample purity), non-porous, homogeneous (either as thin laminae or thicker intervals),
and free of shell fragments and other detritus. We took advantage of the opportunity
provided by modern mass spectrometry to process smaller amounts of material by making
a deliberate effort to limit the amount of detritus included during the initial sampling
stage.
Core splitting and sampling took place at the National Lacustrine Core Facility (Lac-
Core) at the University of Minnesota, Twin Cities, in February 2016. After cores were
split lengthwise and the centers were extracted with a plastic "U-channel" for.paleomag-
netic work, sampling for U-Th dating was given first priority on all cores. This order of
operations ensured that the most ideal carbonates would be reserved for dating and not be
under-utilized for other measurements where less ideal materials would suffice.
Cores were visually assessed for material that fulfilled the criteria described above.
Once a carbonate section was identified, we used utility blades, knives from a fruit and
vegetable carving set, and tweezers to cut and extract thin wafers of sediment -0.2-0.5 cm
in thickness (Fig. 2-5). In sections of core containing finely laminated carbonate sequences,
we took care to isolate individual laminae, only sampling the cleanest parts and scraping
away undesirable material when necessary. In addition, when possible, we sampled layers
that appeared to have more detritus immediately adjacent to these cleaner laminae with
the intention of using this material for possible isochron work. We examined smear slides
during sampling in order to petrographically verify that samples identified by eye as being
relatively detritus-free were as such, and made real-time adjustments in sampling strategy
based on results. Fig. 2-4C shows the depths from which U-Th samples were taken and their
relation to stratigraphic units. Sedimentary lithologies were defined following protocols by
Schnurrenberger et al. (2003), including smear slide observations.
During sampling, we also documented the macro-scale sedimentological characteristics
associated with each sample. After observing a variety of carbonate-rich sequences, we
divided them into four lithological facies categories:
sanpe 
other
replicates
one replicate
Figure 2-5: [A] The core sampling process for U-Th dating. Surgical straight edges and blades
from a fruit carving knife set were used to cut out thin (~0.2-0.3 cm) wafers of sediment, which
were then extracted with tweezers. [B] One sample consists of a single wafer of sediment extracted
as shown in Panel A. Replicate analyses are made on separate sections of the sediment wafer: rather
than homogenizing the wafer into a powder and dating the powder multiple times, we date different
sections of the sample in order to assess sample reproducibility.
• Cream-colored massive carbonates (CMC): cream-colored, massive, medium to thick
(-10-50 cm) bedded carbonate silts found in close association with gray, massive,
banded or mottled silt, with some thin (~1-2 cm thick) peat or organic-rich mud
layers (Fig. 2-6)
" Red-green alternating (RGA): red and green laminated sets of carbonate silts that
alternate'in color every -5-20 cm, with some organic-rich peat laminae (Fig. 2-6)
* Cream-colored carbonates with peat beds (CP1, CP2): Cream-colored, faintly
banded carbonate silt interbedded with peat layers, with some associated with thick
(-30-50 cm) overlying peat beds that were laterally continuous across multiple holes
at Site I (CP1), and others associated with thin (-3-5c m), laterally discontinuous
peat beds (CP2) (Fig. 2-7)
Each sample extracted for U-Th dating was subsequently categorized into one of these four
facies.
.39
J,
Dil
D12
D13 C6
D14
CIO
CiI
CMC (cream-colored massive
IRGcaArb.  silt with dark gray carb. silt)U (red-green alternating,
varigated laminated carb. silt)
SAMPLING LOCATIONS C173
- pass
- fail _ 2"U/ 2 2Th dark
- fail - reproducibility gray silt
fail- 6234U,.< (not U-Th dated)
Figure 2-6: Core scanning images and U-Th sample locations of four selected cores that feature the CMC and
RGA facies. The approximate corresponding CCLF is noted in the black rectangular box at the top left of each
core image. The column of gray and white boxes appended to the left of each core image is a ruler representing
alternating blocks of ten centimeters, mimicking the actual ruler used during scanning at LacCore's facilities. Small
rectangles plotted on top of the core image represent sample locations and are labeled by sample ID and clor-coded
by threshold criteria result (see Section 2.5 and Fig. 2-8). 'The columns to the right of each core image represent the
facies that is giver to a sample collected in that depth interval; for example, for the third core image, samples C6,
C10, and C11 are categorized as CMC, while C13 is categorized as RGA. Core scanning images were made using a
Geotek MSCL-CIS at the National Lacustrine Core Facility (LacCore).
40
A5
A6
FACIES DO
cream-colored,
CP1 faintly laminated
carb. interbedded
CP2 with peat layers
CMC
SAMPLING LOCATIONS
- pass
- fail-2 8U/232Th
- fail - reproducibility
fail - 623 4Uiec
Figure 2-7: Core scanning images and U-Th sample locations of four cores that feature the CP
facies, which is subdivided into CP1 and CP2 to differentiate between samples that are associated
with thick (>10 cm) and laterally continuous peat layers (CP1) and those that are not (CP2). Note
that the third and fourth images are of cores from the same CCLF but from different holes at Site
1, shown here to demonstrate the lateral discontinuity of some peat and carbonate beds. Holes at
Site 1 were spaced approximately -20 ma part. Small rectangles plotted on top of the core image
represent sample locations and are labeled by sample ID and color-coded by threshold criteria result
(see Section 2.5 and Fig. 2-8). See caption in Fig. 2-6 for explanation of other symbology used in
the figure.
41
2.3.2 Sample preparation and chemistry for U-Th dating
After core sampling, sediment wafers were frozen and then placed in a vacuum freeze drier
to remove moisture from all material. Most samples'retained their original wafer shape
after this process. A small portion of each sample was then gently disaggregated for micro-
scale sedimentological characterization under a picking microscope. We made qualitative
observations on the following: color; hardness of bulk sediment (friable or compacted);
and the relative abundance of mollusc shell fragments, ostracode valves, organic fibers
(peat. fragments, grasses, seed pods), sponge spicules, siliciclastic grains, or other mineral
precipitates. For subsequent U-Th analyses, we manually removed mollusc shell pieces from
the sample before dissolution, or avoided processing samples containing abundant mollusc
shell fragments that could not be reliably removed. Otherwise, all analyses discussed are
measurements on bulk-samples containing all other aforementioned constituents.
Because U-Th column elusions are time and- resource intensive, a.small set of samples
from different facies were screened for their uranium and thorium concentrations to deter-
mine which facies would most likely yield material viable for dating. Powders of ~2 mg
were dissolved in dilute HNO3 , andanalyses of uranium and thorium concentration were
performed on a VG PQ2+ quadrupole ICP-MS and an Agilent 7900 ICP-MS at MIT. Sam-
ples with higher U/Th ratios were then identified as materials worth further processing as
they are more likely to yield "clean" samples with high 230Th/ 23 2Th ratios (Section 2.2.2).
Replicate analyses were then processed through U-Th column elusions in batches of
5 to 15. When possible, we analyzed at least three replicates from each sample horizon.
Here, we purposefully apportion different aliquots of the original sediment wafer for each
replicate analysis in order to test the reproducibility of dates from stratigraphically coherent
material (Fig. 2-5B). Note that this is an important difference from repeated analyses of
a homogenized powder, which would only provide a measure of internal lab errors or the
quality of sample homogenization. Our original intention in processing samples this way
was not only to test for reproducibility, but also to build isochrons, for which it is necessary
to analyze subsamples that span a range of detrital contamination levels.
After sample selection and preparation, sample dissolution was performed in a clean
laboratory at MIT. Samples of 5-60 mg were dissolved in HNO and spiked with a 2293 Th-
233 U- 236U tracer in Teflon beakers cleaned via a boiling-washing procedure with concen-
trated HNO3 , HCl, and aqua regia. Next, following methods described by Edwards et al.
(1987) and Shen et'al. (2002), uranium and thoriumw ere co-precipitated with ~4-8 mg of
42
Fe oxyhydroxides and-then separated using BioRad AG1-X8 anion exchange resin (100-200
mesh, 0.5 mL column volume). The'isotopic compositions of the resulting uranium and tho-
rium fractions were then measured on a Nu Plasma II-ES multi-collQctor ICP-MS at MIT.
We'introduced sample solutions through a CETAC Aridus II desolvating nebulizer system
coupled to a PFA nebulizer with a 100 pL/min uptake capillary. Each uranium sample
analysis was bracketed by a 5 ng/g solution of the CRM-112a standard (New Brunswick
Laboratories). Each thorium sample analysis was bracketed by an in-house 229 Th2- 3 0Th-
23 2Th standard in order to monitor mass bias and variable SEM yield. 2% HNO 3 solution
blanks also bracketed each sample and standard analysis to determine the background
signal. See Section 2.10.1 for details of U-Th measurements on materials from the 1996
core.
2.3.3 Estimating the initial 23 0 Th correction
As discussed in Section 5.3, the correction for initial 230 Th has a greater impact on impure
sample materials, and so it follows that we must carefully consider this correction for
the lake sediments at Lake Junin. Ten samples processed from the 1996 core yielded
indeterminate ('infinite') dates, in which a unique solution for the 230 Th age equation could
not be found after iteration. These samples all had 23 2Th concentrations that were 20-200
times greater than other samples from the 1996 core that yielded calculable dates (2-7
ppm, compared to 0.04-0.1 ppm), forming a statistically distinct population. Similarly,
these samples also had 238 U/ 232Th ratios that were -50 times lower than that of other
samples (0.3-0.6 ppm). These results suggest that the samples yielding indeterminate ages
had high amounts of detrital contamination that contributed a significant amount of initial
2 30Th at secular equilibrium with 238U, thereby causing apparent infinite dates.
Assuming that the detrital component of the indeterminate samples of the 1996 core is
representative of the isotopic composition of detrital material found in all sediments of the
PLJ-1 splice, we calculated the average 230 Th/2 32Th ratio of the indeterminate samples
and used this ratio for the initial 230 Th correction in our calculations. This estimate has
the effect of counting radiogenic 230 Th accumulated in these samples as detrital, but the
depths of these samples suggest that their true age is no older than 30 kyrs and thus we
do not expect an appreciable proportion of the 2 30 Th to be radiogenic.
The average 230 Th/ 232 Th atomic ratio of the indeterminate samples from the 1996 core
is 7.9 ±0.9 x 10-6. Our starting assumption is that this ratio is invariant through time, but
43
it is entirely possible-if not expected-that the isotopic composition of detritus is variable
due to changes in clastic transport or source regions. To account for these unknowns and
other unknown unknowns, we apply an uncertainty of 50%to this average and-use an initial
230 Th/ 23 T2 h atomic ratio of 8.0 4.0 x19-6 for U-Th data from the PLJ- splice.
2.3.4 Calculating weighted means and uncertainties of samples with repli-
cate analyses
As previously mentioned, we attempted to analyze at least 3-5 replicates for each sample
as a test of the reproducibility of unequivocally coeval material. We calculate a date- for
each individual replicate analysis using Eq. 2.1 and the initial 230 Th/ 232 Th ratio stated
above. We then use these dates to calculate an error-weighted mean (2) and uncertainty
(os)of all replicate analyses in a sample, in which weights are equal to the inverse of the
variance of each date:
- 1 = 1- (2.4)
NN Na
where N is the number of replicate analyses in the sample; x is the individual date of each
replicate; ando2is the variance of the individual dates of each replicate.
We then calculate the degree of agreement between replicate analyses to estimate an
uncertainty that is appropriate for the observed scatter between dates. To do this, we
calculate the Mean Square of Weighted Deviates (MSWD), a measure of the ratio of the
observed scatter around the mean to the expected degree of scatter given the analytical
uncertainties of each data point (McIntyre et al., 1966; Wendt and Carl, 1991). The value
is essentially the chi-squared statistic (goodness of fit) divided by the number of degrees of
freedom (f=N-1), or the "reduced" chi-squared:
1 N 2
MSWD = (x1 2 (2.5)
f
The value of the MSWD tells us if the calculated uncertainties for each date are over-
or underestimated based on the observed scatter in data. A value of ~1 indicates that
the observed scatter is equal to the predicted scatter; values less than 1 indicate that the
observed scatter is less than is predicted by the uncertainties; and values greater than 1
indicate that the observed scatter is more than the predicted scatter. Samples with an
MSWD much greater than 1 are considered to have excess "geologic scatter," suggesting
44
possible biases in the calculated dates, perhaps due to a violation of the assumptions
underpinning the system (e.g., open system behavior). Thus, for any sample with an
MSWD > 1, we expand the uncertainties of the replicate analyses by a factor of '/MSWD
and then recalculate the weighted means with these larger uncertainties using Eq. 2.4.
The IsoplotR program by Vermeesch (2018) also includes this strategy as one option of
treating data with excess geologic scatter (referred to as "overdispersion"). While the
presence of excess scatter is non-ideal and raises concerns about the validity and practical
use of such dates, the data still represent geologically meaningful information and thus
should not necessarily be rejected outright without further consideration (and we will do
much considering, starting in Section 2.5).
Using the MSWD as a black-and-white parameter to evaluate the validity of dates is
generally discouraged, since the highest permissible MSWD is dependent on N (Wendt
and Carl, 1991) and is often subject to interpretation (Powell et al., 2002; Ludwig, 2003).
Thus, we calculate the probability of the observed scatter occurring given the uncertainties
for each replicate analysis (a "probability of fit") by computing the chi-square cumulative
distribution for MSWDxf (the chi-squared statistic) about f degrees of freedom (York,
1968).
Some samples only have 1-2 replicates; these were cases in which early replicate analyses
yielded unfavorable results (i.e., low 238 U/ 232Th ratios) and were thus not further repeated
in the interest of time and resources. For the remainder of this paper, our discussion of U-
Th dates will refer to the weighted means and uncertainties (MSWD-adjusted) of samples
rather than the individual dates of replicate analyses, unless otherwise noted.
2.3.5 Other corresponding data
We use other sedimentological, geochemical, paleoecological, and physical data to interpret
and understand our U-Th data. We provide a list of these datasets in Table 2.2 and their
intended use. More information regarding these methods of measurement can be found
in the Supplementary Materials (Section 2.10.2) and other publications currently being
prepared elsewhere.
45
Dataset Brief Methods Purpose
Elemental sample dissolution, Determine if there exists any relationship between
concentrations ICP-MS U-Th data arid concentrations of Ca and trace el-
ements Mg, Sr, Al, Ti, P, V, Mn, and Fe. Mea-
surements are made on same sample material used
for U-Th dating.
Total inorganic coulometry Determine if there exists a relationship between
and organic U-Th data and carbon content. Only measure-
carbon ments made within 1 cm of the U-Th sample are
paired withU-Th data.
Color reflectance spectrophotometry Determine if there exists a relationship between
on automated core U-Th data and any spectral reflectance wave-
logger length band. Only measurements made within
2.5 mm of the U-Th sample are paired with U-Th
data.
Mineralogy X-ray diffraction Determine the mineral composition of the carbon-
ate phases, and if there are discernible differences
between endogenic and detrital carbonate.
Ostracode picking and iden- Determine if there exists a relationship between
assemblages tification following U-Th data and ostracode color, taphonomy (num-
Prez et al. (2010) ber of broken vs. intact valves; adults vs. juve-
and Karanovic niles), or ecology (benthic vs. swimmer species,
(2012) ornamentation).
Table 2.2: Other datasets used in this study for comparison with U-Th data.
2.4 Results
In total, wegenerated 174"U-Th dates from 55 bulk samples from the PLJ-1 splice. Uraniumh
and thorium geochemical data as well as the number of replicates produced for each sample
(N = 3-8) can be found in Table 2.3. Samples originate from each of the five high (>70%)
CaCO3 intervals that occurred every -10-15 m in the core (Fig. 2-4C). All U-Th dates
from the uppermost 5 m are broadly consistent with radiocarbon dates from terrestrial
macrofossils and charcoal in the same depth interval (Fig. 2-3; Woods et al., 2019). A
sample from ~6.5 m yielded an indeterminate date and had a 230 Th/ 2 32Th atomic ratio of
7.7 ± 0.2 X10~6, consistent with our estimate of the detrital 230 Th/ 23 2Th ratio applied in
corrections. Sample 238 U concentrations are variable and are generally 0.2-4.0 ppm (mean
= 1.5±1.2 ppm, 1-a); 23 2Th concentrations are also variable, ranging 0.02-2.4 ppm (mean
= 0.6±0.5 ppm, 1-a).
46
For the deepest part of the core, the oldest U-Th dates suggest that the record is
no older than ~800 ka. This observation is consistent with the absence of evidence of
the Brunhes-Matuyama magnetic reversal (aged ~780 ka) in the paleomagnetic secular
variation record (Hatfield et al., in prep.). However, the scatter of dates throughout the
entirety of the core is, at first glance, alarming: at 20-25 m, the first high-CaCO 3 section
beyond the interval covered by radiocarbon, U-Th dates already span a range of ~200 kyrs
(Fig. 2-8). The spread of dates increases with depth, reaching -300 kyrs at the bottom
of the record. As is, the scatter of data is too great to build any practical age-depth
model, even after applying outlier analysis. Furthermore, all attempts to reduce scatter
by building isochrons from replicate analyses and adjacent dirty-clean sample pairs failed
(high MSWD and low probability-of-fit; Section 2.10.3).
Here, we arrive at the main crux of this paper. The scatter of data and the failure
to build isochrons is clear evidence that at least some of the dated materials have not
remained closed systems or do not otherwise satisfy the operating assumptions of U-Th
dating. Despite this noise, is there a way to objectively assess the quality of each U-Th
date, and subsequently curate the dataset without biases (avoid "cherry-picking")? In
the following sections, we detail our approach to this question. At times, we will refer to
specific U-Th samples by their sample name, which consists of an alphabetical letter A-P
followed by a number 1-16 (Table 2.4).
2.5 Curation of U-Th data using threshold criteria
Noisy U-Th geochronological datasets are nothing new; in attempts to find clarity in un-
certain data, a common practice is to apply some screening criteria based on uranium
and thorium concentrations. For example, some studies dating corals and carbonate slope
sediments have rejected dates that exceed a certain thorium concentration or do not meet
a minimum uranium concentration because such dates tend to have larger corrections and
errors (Robinson et al., 2002; Henderson et al., 2006; Skrivanek et al., 2018). However,
picking the values for these thresholds can be subjective to an extent, especially if there is
no clear separation between distinct populations within the data.
As a start towards better understanding the scatter in our data, we consider applying
similar thresholds, first by examining the 2 38U/ 23 2Th ratio and the probability of fit of all
dates for a given sample to a single weighted mean (Fig. 2-8A and B). 238 U/ 32 2 Th ratios
ranged between <1 and 30 and probabilities of fit essentially spanned the full domain of
47
possible values, from 0% to 99%. Between 20 and 60 m, we notice that the samples with the
oldest-dates all have 238 U/ 232Th ratios that are <1, including those yielding indeterminate
dates (Fig. 2-8A). One possible explanation for this observation is that these samples have
initial 23 0Th that has not been accounted for with our initial -correction, which would bias
dates to be older than the true age. The effect of this bias would be -greatest in samples
with low radiogenic 2 oTh due to low uranium concentrations. (Note that we later discuss
another explanationfor these data in Section 2.7).
Regarding the probability of fit, deciding how low of a probability is acceptable is
somewhat arbitrary; there is no broad consensus within the geochronology community on
how best to treat such data, especially in cicunistances in which the total number of
subsamples is low (Ludwig, 2012), as is our case. However, most geochronologists would
likely agree that samples with a probability of fit less than 1% (especially those much closer
to 0%) exhibit an amount of excess scatter that is beyond recovery of pratical information
about'the true age of the'sample.
Thus, in the interest of not using too strong of a hand in curating the U-Th data to
begin, we apply two conservative threshold criteria: the 238 U/ 232 Th ratio must be>1 and
the probability of fit >1% (Fig. 2-8A and B). Of the 55 samples, 17 fail the 238 U/ 23 2 Th
criterion and 22 fail the reproducibility criterion. Of the 17 samples that fail the2 3 8 U/ 2 3 2 Th
criterion, eight had more than one replicate analysis, and of those eight, five also fail the
reproducibility criterion. In Fig. 2-8D, we show which criterion each sample fails; for the
purposes of simplifying ensuing explorations into the dataset, the five samples that fail
both aforementioned criterion are categorized as having failed the 238 U/ 2 32Th criterion.
Next, we consider another screening approach adopted for U-Th dating of marine sam-
ples that involves 6234Uinitai. Because the residence time of uranium in the ocean is very
long (400 kyrs; Ku et al., 1977), the 8234 U of seawater is thought to have remained rela-
tively constant for at least the last 400 kyrs (Henderson, 2002; Henderson and Anderson,
2003). Thus, assuming that marine samples reliably preservethe 8234 U values of the waters
in which they formed, dates from marine samples with 8234 U values that deviate signifi-
cantly from modern values are considered potentially inaccurate due to diagenesis (Bard
et al., 1991; Hamelin et al., 1991; Gallup et al., 1994). In contrast, the 8234U of surface
waters is very diverse and has been found to besensitive to basin lithology, basin-specific
weathering mechanics, riverine and groundwater inputs, and climate (e.g., Sarin et al.,
1990; Kronfeld and Vogel, 1991; Plater et al., 1992; Kronfeld et al., 2004; Robinson et al.,
2004; Durand et al., 2005; Grzymko et al., 2007;,Chabaux et al., 2008). Although few
48
M
A U/2Th B reproducibility (probability of fit)
Ind., - + - -- -- - -- - • - Ind. - --- - - - -----------
6001 600 -
500 500l
. 400 .400
<300 4300
2001 .8 200
30 100
1001 100
-ft
IJ'71
0 20 40 60 80 0 20 40 60 80
Depth (m) Depth (m)
c 62"U of Initial endogenic carbonate M threshold criteria results
4000 Ind. r -----------------
35000T 600-
3000 500-
2500
<300
2000 K 00
1500 ~; 1... 
200
1000 * pass
2200 m
100- *fail - U/ h500 - 19004 0Sf ail-reproducibility0ofal -6134U.
U0 20 40 0160 80 0 20 40 60 80
Depth (m) Depth (m)
Figure2-8: Step-by-step application of thresholding criteria: [A] 
238 U/ 23 2Th, [B] reproducibility,
234
and [C] 623 4U of initial endogenic carbonate (8 U c1). [D] shows the data that pass and fail the
three aforementioned criteria. Note that each point represents the weighted mean and standard
deviation of multiple replicate analyses (see Table 2.3). In Panels A-C, blue/red colors represent
values that are more/less ideal for U-Th dating. Values of thresholds are indicated by a red asterisk
(*) in each legend. Sampes that do not satisfy criteria and were thus subsequently eliminated
are colored in dark red. Samples plotted along the uppermost dashed line labeled 'Ind.' refer to
analyses that yielded incalculable U-Th dates or were infinite (indeterminate). In Panel B, analyses
plotted in light gray are those with only one replicate analysis. Panel C only includes data that
pass the 238 U/ 2 32Th and reproducibility thresholds. Shaded gray areas represent the distribution
of 6234Uiec values observed in the Holocene, including data from the 1996 core. For Panel D, note
23 2
that some samples failed both criteria for 2
3 8 U/ Th and reproducibility; in these instances, the
samples were categorized as having failed the 
238 U/ 232Th criteria. All error bars in each panel are
2-a range. Depth refers to the composite core depth below lake floor (CCLF).
49
studies examine the long-term history of internal 8234 U variability in lakes and other sur-
face waters (e.g., Kiro et al., 2018), the range of internal 8 234U variability observed in the
aforementioned river and groundwater studies suggests that the internal 8234 U of lake wa-
ters shouldnot vary significantly without dramatic changes in drainage basin organization.
Since the lithology of a lake basin is invariant over the timescales relevant to this study,
variability in 8234 U is'driven by changes in hydrology. McGee et al. (2012) documented
a 300%c change in the 8234U of lacustrine cave carbonates during the last deglaciation in
Lake Bonneville (Utah. USA), which experienced a ~2x dhange in precipitation.
Thus, we apply a third threshold criterion using the 8234 U of the initial endogenic
carbonate (8 234 Uiec) of each sample, which we calculate by correcting 6234 Uinitial values
for detrital uranium (see Section 2.10.4 for relevant equations). The average'8 234 Uiec of
all samples that yield dates verified by radiocarbon data (including data from the 1996
core) is 2800±300%o. If we compare this average to the 8234Uiec values of the remaining
21 samples, we observe that three samples at -70-75 m have values that fall well below
the average, even outside the range defined by three standard deviations from the mean
(Fig. 2-8C). Because the magnitude of these differences is large, we suggest that these
values are unlikely to reflect real changes in the 8234U of the lake waters, and thus suspect
the validity of these dates. Therefore, we mark these three samples as having failed the
823 4 Uiec criterion (Fig. 2-8D).
While the remaining 18 dates form a visually pleasing line (Fig. 2-8D), this observation
alone does not intrinsically prove that these remaining dates are accurate. However, the
samples generally abide by the rules of stratigraphic order, which is behavior consistent
with closed-system dates. Furthermore, the results of applying the threshold criteria may
vaguely follow our theoretical expectations for normally distributed scatter about the mean.
That is to say: if you were to ask someone to draw a line through the middle of the original
scatter of points, the 18 samples that remain would not stray far from it.
2.6 Understanding the scatter
Having classified the U-Th data into categories that describe the main flaw of each nom-
inally failed sample (Fig. 2-8D), we now explore the underlying causes for poor sample
behavior and determine if the application of threshold criteria is justified. Essentially, we
ask: Is there other evidence that supports our assertion that the threshold criteria failing
samples have not remained closed systems? What is special about the 18 passing samples
50
such th t te xhibitfewer symptoms of open system behavior?
2.6.1 Detrital contamination
The first and most obvious hypothesis for poorly behaving dates is detrital contamination
that is unaccounted for with the initial 23 Th/ 232Th correction. As stated in Section 2.22,
impure sample substrates have been the main obstacle in previous U-Th dating efforts in
lake sediments, and there is no evidence to suggest that Lake Junin would be an exception.
If detrital contamination does indeed play a large role in the scatter of our U-Th data, we
can make certain predictions -for how other sedimentological and -geochemical data would
respond. For example, we would expect that samples with lower CaCO 3 content would
comprise the eliminated dates, especially those that failed the 238 U/ 232Th criterion. We
are able to test this hypothesis directly using co-located measurements of CaCO3 content
(weight %) as well as optical lightness from color reflectance spectra (Table 2.2 and Section
2.10.2). Optical lightness, defined here as the sum of spectra in the visible band of the
electromagnetic spectrum (400-700 nm), has been shown to be a reasonable proxy for
carbonate content in marine sediments (e.g., Nagao and Nakashima, 1992; Mix et al.,
1995; Balsam et al., 1999). Data for the PLJ-1 splice also shows that CaCO 3 content
>50% appears to scale with optical lightness (see gray circles in Fig. 2-9A). Because color
reflectance data were measured on a finer and more regular sampling interval than carbon
data, there are more optical lightness data that correspond to U-Th samples than CaCO 3
content measurements (N = 46 versus 29; Fig. 2-9).
From Fig. 2-9, we notice that all but one of our U-Th samples have >50% CaCO 3
content, with most passing dates having >70% CaCO 3 content. However, many data that
failed threshold criteria occupy the same range in CaCO3 content as passing data. Con-
trary to expectations, each of the five samples with the highest CaCO3 content failed the
threshold criteria (Fig. 2-9B). Likewise, the five samples with the highest optical lightness
also failed (not the same five samples; Fig. 2-9C). Sample C10 is the most extreme case
in this comparison, and as a visual check, we can see its sampling location in Fig. 2-6
and verify its optical lightness value relative to other samples. Again, there is no clear
pattern between passing and failing U-Th dates and optical lightness; samples appear to
exhibit the entire range of optical lightness values observed in the core. If anything, one
could argue that the samples failing the 238U/ 232Th criterion tend to have higher optical
lightness values compared to passing samples, behaving opposite to our predictions. The
51
6- Ii
1%.)
4
N=46
1400 I I *fl
I I 
I 
C109
ther coredata
1200 . pass
I fail- 238 U/ 32 Th
1 fail - reproducibilityI
0 2 ca c6fail - 6& U C;1s-
1000
4A
800-
600- C3
* *4
0
400 N
* ,
0
200 ,60
0 I
0 10 20 30 40 50 60 70 80 90 100 0 24 6 8
CaCO 3(wt %) Counts
Figure 2-9: Cross-plot [A] and histograms of calcium carbonate content [B] and optical lightness
[C], showing the distribution of these values for U-Th samples of each threshold criteria result. In
Panel A, there are 29 colored circles which represent the U-Th samples that have both a corre-
sponding CaCO3 analysis (within 1 cm of the sample location) and a color reflectance measurement
(within 2.5 mm of the sample location). Since color reflectance data were measured on a finer and
more regular sampling interval than carbon data, there are more U-Th samples for which there is
a corresponding color reflectance measurement (N = 46); thus, there are data plotted in the Panel
C histogran that are not shown in Panels A and B. Empty circles with gray outines represent
other pairs of CaCO3 and brightness data throughout the core and are only included if these data
correspond to the exact same core depth. Note that this figure does not include any data from the
upper 6 m of the core.
results are similar when we compare our data with grayscale or luminance (also known as
L*), another "lightness" parameter using the CIE L*a*b* color description system.
Based on these results, we speculate that a high proportion of bedrock carbonate in
the detrital component could explain the high CaCO 3 content offailedsamples;even
the darkest gray silt sections in the core with high magnetic susceptibility had 20-50%
CaCO3 content. The uranium from this detrital carbonate would be at secular equilibrium
(Fig. 2-1D) and would adversely impact our We compare the mineralogy of local carbonate
52
bedrock to core sediments to see if mineralogical differences between these carbonates
cold be used todetect detrital contamination (Sectin 2.10.2). Unfortunately, the results
show that there is no discernible difference between the carbonate berock and carbonate-
bearing lake sediments, even when comparing different grain size groupings. All bedrock
samples were dominated by low-Mg calcite, except one sample, which revealed the presence
of dolomite. Low-Mg calcite was the dominant carbonate phase in all carbonate-bearing
samples from the core, with no evidence of dolomite.
Despite this, elemental ICPMS concentration data support the prediction for detri-
tal contamination in failed samples, in particular, those failing the2 83 U/2 32 Th criterion.
Fig. 2-10A is a biplot of the orthonormal principal component coefficients for Ca, Mg, Sr,
Fe, Mn, Al, Ti, V, and P and the principal component scores for each sample. Here, the
first principal component (PC1) has positive coefficients for elements that are markers for
aluminosilicates (Fe, Mn, Al, and Ti) and explains -44.5% of the variance in concentration
data. Samples failing the 238U/ 23 2Th criterion generally have positive PCI scores, indi-
cating that those samples tend to have relatively higher concentrations of these elements.
Elemental concentration data from a sample of the dark gray carbonate silt also exhibits
higher concentrations of these elements (sample E2; Fig. 2-6, core at 35 in). As most sam-
ples that failed the 238 U/ 232 Th criterion are of the CMC facies (Fig. 2-10B), these data are
consistent with our visual observations of this facies, in which the lighter-colored carbon-
ates visually appear to have semi-gradational boundaries with the surrounding dark gray
glaciogenic silts, suggesting that these samples likely contain some fraction of this material.
In addition, the sample with the highest PC1 score, D11, comes from a boundary between
cream-colored carbonate and dark gray carbonate silt (Fig. 2-6).
Despite the clear relationship between high PC scores and2 83 U/ 2 32Th criterion failing
samples, samples with low PC1 scores are not ubiquitously well-behaved, indicating that
another factor is influencing our data. Interpreting the meaning of the second principal
component (PC2) and the scores for other samples is less clear. PC2 distinguishes among
samples that have high values for Mg, Sr, and Ca. Total organic carbon data suggest that
samples with negative PC2 scores have higher organic carbon content, but there are not
enough available corresponding carbon data to be convincing. Furthermore, there is no
other distinguishable separation of U-Th data by threshold criteria result in the biplot.
53
0.6 A BrC
0.4 CP1
.2 RGA 2
T* ~~'0 10 20 30 40 50 60 /0 80 90 100
0A Cumulative Percentage (%)
0
C [ compacted friable-0.2 
CL pass -
0
0. -0.4 ass fail-6SUi
0al-5 u/1 fail-reprod..
-0.6 0fi-rerdcbIia erd________ 1 1 1 I-0.66 fail-_ U_ fail-238/232
-0.6 -0.4 -0.2 0 -0.2 0.4 0.6 0 10 20 30 40 50 60 70 80 90 100
Component 1 [44.5% of variance] CumulativePercentage
Figure 2-10: [A] Biplot of the orthonormal principal component coefficients for Ca; Mg, Sr, Fe,
Mn, Al, Ti, V, and P concentrations (dashed blue lines, labeled by element) and the principal
component scores for N= 48 samples (circles, color-coded by the threshold criteria result). Light
gray labels are corresponding'sample IDs of samples that are featured in other figures. we encourage
the reader to follow these labels in order to connect these plots with others. [B] Bar chart showing,
for each facies, the relative proportion of samples that are each of the four threshold criteria results.
CMC is cream-colored massive carbonate silt (Fig. 2-6); and CP1 and CP2 are cream-colored
carbonate silts interbedded with peat layers; and RGA is red-green alternating varigated carbonate
silt (Fig. 2-7). See Section 5.2 for further details. The bar colors follow the legend of Panel A.
Numbers within each bar represent the actual number of samples; for example, of the 18 samples
categorized as the CMC facies (top row), 1 passed, 3 failed the reproducibility criterion, and 14
failed the 2 38 U/ 2 3 2 Th criterion. [C] Bar chart showing the relative proportion of samples that are
compacted (light) versus friable (dark) for samples of each threshold criteria result. Note that we
qualitatively assessed sediment hardness for only 21 samples. As with Panel B, numbers within
each bar represent the actual number of samples. Note that Panels B and C do not include U-Th
samples from the upper 6 m of the core.
2.6.2 Open system uranium remobilization
While the data comparisons presented thus far broadly confirm that samples with higher
CaCO3 content are more likely to yield well-behaved U-Th dates, there remain some in-
consistencies with predictions for detrital contamination, mainly that the samples with
the highest CaCO3 content and optical lightness fail the threshold criteria, especially the
238 U/ 232Th criterion. Regarding the other sample data, there were no patterns distin-
guishing passing samples from samples failing the reproducibility and 234Uiec criteria in
Figs. 2-9 and 2-10A. This information leads us to consider the next probable cause for
poor sample behavior: the remobilization of uranium after initial carbonate formation.
54
Using a qualitative assessment of sample hardness, we notice that passing samples were
generally more compacted and dense, whereas failing samples were more friable and soft,
especially those failing the 2 38 U/ 2 2 Th criterion (Fig. 2-10C). This observation fits our in-
tuitive expectation that samples with less porosity would be more impervious to diagenesis
or secondary deposition of uranium from porewater fluid flow (Fig. 2-1D).
Ve now refer back to the facies to which each sample is assigned for further insight.
From Fig. 2-10B, it is clear that facies alone does not dictate how each U-Th sample be-
haves. Instead, there are some broad tendencies: most CMC samples failed the 238 U/ 232 Th
criterion; most RGA samples failed the reproducibility criterion; and most-of the passing
samples'originate from the CP facies (CP1 and CP2). In considering the reasons behind
these patterns, we compare both facies and threshold criteria results with uranium concen-
tration, total organic carbon (TOC), and a*, the red-green color reflectance of sediment,
where +a* values are more red and -a* values are more green (Fig. 2-11). The red or
green color of sediments has long been used as a qualitative indicator of in situ redox
conditions, in which red colors signify oxidizing conditions and green-gray colors suggest
reducing conditions, owing to the strong chromophores associated with ferric and ferrous
iron (Tomlinson, 1916; Lyle, 1983).
Panels A-F in Fig. 2-11 (top half) compare the mean a* of all measurements within
5 cm of the U-Th sample to the mean point-to-point difference in a* across the same
interval (a measure of the 'volatility' of a* around each sample). For example, having
a mean difference of 2 units/cm in a* means that the a* value changes, on average, by
a magnitude of 2 along every cm within the 10 cm range surrounding the U-Th sample.
Such a value would indicate significant volatility in red-green color, given that the total
range of mean a* observed in U-Th samples is ~3. From these panels, we notice that the
CP samples are more red and occupy a relatively narrow range of mean a* values, whereas
the CMC samples tend to be less red and exhibit less volatility in red-green color (Fig. 2-
11D, F), especially those failing the 238 U/ 2 32Th criterion (Fig. 2-11C, E). These results are
consistent with our qualitative observations of the CMC facies, in which the lighter colored
carbonate occurs in ~10-50 cm thick beds that are relatively uniform in color (Fig. 2-6).
Previous studies on sediments have interpreted changes in red-green color intensity
as changes in the input of red iron-bearing materials (e.g., Giosan et al., 2002; Helmke
et al., 2002; Ji et al., 2005), but because these iron-bearing minerals are highly sensitive
to variations in redox environment, reductive diagenesis can subsequently alter sediment
color to be more green (Lyle, 1983; Kbnig et al., 1999; Kdnig et al., 2000). Thus, here we
55
pass CMC
fail - 238/232 0 RGA
Ofail - reprod. * CP1
ofail- 62"U, 0CP2
S2.5
2.0
.0 *A'A6
-0V 1.5 00
1.0 : •~ A0., I-
.0.5
E 0
1 -0.5 0 0.5 1 1.5 2 2.5 -1 -0.5 0 0.5- 1 1.5 2 2.5 3
mean a* mean a
4
0.
7
U
6
5
0 K16,
*C3
414
0*9
1l6,P-4 A
00 2 4 6 8 10 12 14 '160 2 4 6 8 10 12 14 16
TOC(wt%) . TOC(wt%)
Figure 2-11: Cross-plots- and box-and-whisker plots comparing the red-green color reflectance (a*;
from the L*a*b* color space), total organic carbon (TOC) content, and
2 3 8 U concentration of
U-Th samples and showing their relationship to threshold criteria result and facies. In all box-and-
whisker plots, the thick central black lino represents the median; the top and bottom edges of the
box represent the 25th and 75th percentiles, respectively; the whiskers extend to the most extreme
points not considered outliers; and the outliers are plotted as '+' smbols. A point is considered an
outlier if it has a value >1.5x the interquartile range away from the 25th or 75th percentiles. The
top half of the figure compares the mean a* of allmeasurements within 5 cm of the U-Th sample to
the mean point-to-point difference in a* across the same interval. Note that there are some samples
for which there is a 238U concentration measurement but no corresponding TOC; thus, Panels K-L
include data not shown in the cross-plots of Panels G-H. The bottom half of the figure compares
TOC and 2 38U concentration of each U-Th sample. Note that this figure does not include any data
from the upper 5 m of the core.
56
propose that the difference in a* values between the CP and CMC samples is a reflection of
diagenesis: the CP samples, being more red, have not been as altered by interactions with
post-depositional reducing pore fluids and thus better preserve primary isotopic information
to produce passing U-Th dates. In contrast, the CMC samples, which may have originally
appeared more red like the CP samples, have been altered and as a result, have changed a
more green color.
Consistent with this hypothesis is the observation that the CMC and 238 U/ 232 Th thresh-
old failing samples generally have much lower 238 U concentrations and TOC compared to
other samples (Fig. 2-11G-L). Consider the following scenario: a package of endogenic
carbonate containing organic matter is deposited and submerged under oxygenated con-
ditions. After burial, oxygenated porewaters then interact with the organic matter and
begin to degrade it, removing from the sediments any uranium associated with the organic
matter (Section 2.2.2; Fig. 2-1D). This degradation of organic matter may decrease the
local pH of pore fluids such that it begins to alter the endogenic carbonate, leaching ura-
nium originally bound within the crystal lattice. At some later point, the pore fluids are
no longer recycled, and eventually all oxygen is depleted. The now reducing fluids then
begin to reduce the surrounding sediment, shifting its color from red to more green. Any
uranium that was removed from the carbonate into the pore fluid has now precipitated as
authigenic uranium under these reducing conditions, but is no longer lattice-bound and
is thus susceptible to further remobilization (Section 2.2.2; Fig. 2-1D). By the time we
extract the core and measure the isotopic composition of these sediments, they are green
(Fig. 2-11A-F), easy to physically disaggregate (Fig. 2-10C), and have low organic matter
content and uranium concentrations (Fig. 2-11G-L).
If uranium loss has occurred from the CMC facies, preferential loss of 234 U may be
expected, such that replicate analyses produce an inverse relationship between 8324 Uinitial
and 23 8 U concentration (e.g., Robinson et al., 2006). Indeed, such a relationship is observed
for some CMC samples (Fig. 2-16). Furthermore, this proposed mechanism is compatible
with the interpretations by Woods et al. (2019) for Lake Junin sediments of the last 50
ka, in which peat layers represent abrupt,- 25-500 year periods of drought and lake low
stands. Dramatic lake level changes would alter water table gradients and change ground-
water discharge rates through littoral sediments. Thus, for the CMC facies, while detrital
contamination is apparent given the elemental concentration data (Fig. 2-10), the initial
2 3 0Th correction might have compensated to yield an accurate yet imprecise date, were it
not for uranium loss.
57
As for why- the CP facies seems not as affected by such pore fluids: we hypothesize
that the thick peat bds associated with this facies act as a reductive barrier to the- vertical
inovement of such oxygenated porewaters. This explanation is further -bolstered by the
observation that samples bounded on top by a thick (>10 cm) peat layer that is laterally
continuous across multiple holes at the site (i.e., the CP1 facies) yielded more passing dates
with higher 2 38 U/ 2 3 2Th and probability of fit (Fig. 2-10B and 2-7). The cream-colored car-
bonates of the CP1 facies also exhibit faint horizontal banding, possibly representing the
preservation of primary fabric. In contrast, samples from the CP2 facies yielded compar-
atively less ideal U-Th data (Fig. 2-10B). Sedimentologically, the carbonates of the CP2
facies tend to be darker in color and more massive rather than banded in texture. Fig. 2-7
features two depth-equivalent core sections from different holes which were classified as the
CP2 facies. Examining images of these two core sections, it is clear that the uppermost
layer of peat is not laterally continuous. Samples E12 and F4 are the only two samples
of this facies that pass the threshold criteria, and are arguably the most tenuous of the
passing dates: E12 has a 23 8U/ 23 2Th ratio of 1.02 and F4 has a probability of fit only
slightly above 1%, both borderline values.
Thus far, we have multiple lines of evidence that point towards uranium loss as an
explanation for the broad behavior of CMC samples, and the basis for that theory can
explain the acceptable behavior of the CP samples. As for the RGA samples, these sedi-
mnents tend to exhibit higher a* volatility (Fig. 2-11D, F), as one would expect for samples
of a facies defined by alternating beds of red and green laminae (Fig. 2-6). Contrary to
the CMC samples, the red-green color of the RGA facies seems to be controlled by sed-
iment composition, given that the laminae are well-defined and the boundaries between
color changes are very distinct. The sediment color of RGA facies dulled noticeably a few
hours after initial core cutting and exposure, suggesting that iron-bearing minerals again
strongly influence color. To explain the general lack of reproducibility of U-Th data from
RGA samples (Fig. 2-10C), we speculate that the green layers containing ferrous iron may
be reactive enough to remobilize Fe-Mn hydroxides that complex with uranium (Chappaz
et al., 2010), leading to open system'behavior that manifests as poor reproducibility.
2.6.3 Ostracode and mollusc shells
As a demonstration of the utility of paleoecological data for U-Th data interpretation, we
compare our U-Th sample data to measures of ostracode color and mollusc shell abundance.
58
During sample processing, we noticed that the color of ostracode shells often varied from
sample to sample, ranging from translucent to dark gray or black. Because modern pristine
ostracode shells are generally transparent and exhibit only trace pigmentation (Smith and
Delorme, 2010), fossil ostracode shells with dark discoloration or coatings are generally
thought to be altered and are thus avoided for geochemical analyses as a good practice
(Holmes and Chivas, 2002). Many studies have made note of dark coatings on ostracode
shells (e.g., Palacios-Fest et al., 2005; Wrozyna et al., 2012; Mackay et al., 2013), but there
are few systematic studies that attempt to explain the origin and controls on ostracode
discoloration or coatings (e.g., Ainsworth et al., 1990; Schwalb et al., 1995; Holmes, 1998).
The results of our comparison with ostracode shell color are broadly consistent with
our hypothesis that the CMC facies has been altered by reductive diagenesis. Fig. 2-12
compares the threshold criteria result and facies with ostracode color, which we classified
on a 7-point scale from translucent to black. Most CMC samples (5 out of 7) and all
samples failing the 238U/ 232Th criteria had a higher proportion of darker shells. Sample
B5, the only CMC sample that passed threshold criteria, has no ostracode shells with
a color,> 2. All other samples with different threshold criteria results and facies had
ostracode assemblages comprised mainly of light-colored shells, with some exceptions.
Aside from color, there were no conclusive relationships between U-Th data and taphon-
omy (number of broken versus intact shells; adults versus juvenile counts) or ecology (ben-
thic versus swimmer species; ornamentation). However, we noticed that most of the darker
shells (color > 2) belonged to Darwinulidae, a family of benthic ostracodes that are consid-
ered an indicator of groundwater discharge. Schwalb et al. (1995) observed dark coatings
on Darwinula stevensoni valves in Holocene sediments of Williams Lake (Minnesota, USA)
and determined via wavelength- and energy-dispersive (WD/ED) spectrometry that the
coatings were made of irons ulfide. They proposed that the coatings formed during periods
of increased groundwater discharge, in which groundwater supplied additional Fe to the
lake while reactive organic matter and sulfate led to reducing conditions that promoted
iron sulfide formation. This mechanism is analogous to the one we propose to explain the
behavior of U-Th data from the CMC facies. Future investigations should analyze the
composition of the surface coatings with either WD/ED spectrometry or scanning electron
microscope energy-dispersive X-ray (SEM-EDX) analysis.
Regarding mollusc shells, as discussed in Section 2.2.3, extensive attempts to U-Th date
mollusc shells for paleo-sea level reconstructions have shown that this material does not
remain a closed system after burial. We confirm that mollusc shells yield dates that are
59
lighter Ostracode Valve Color darker- 0
200 pm
!0' "0 1 2 4 6 7
0 S F4M 3
M 1
0 0 Li1
0 0 D6
0 0 B13
0 e0 B11B5
0
0 * F13* B15 1175
0 0 G14
0
* A5
0 * A6 it
0 * C13
0 * C3 0 _1 
0 * c10
0 L7 6J~
0 *G6
•0 * D13
0 * C6* D14
0 10 20 30 40 50 60 70 80 90 100
Cumulative Percentage of Ostracode Valves in Sample (%)
Figure 2-12: Relationship between ostracode color and U-Th data. Top row of microscope images
of ostracode valves illustrate the coloration scale and are arranged from lightest to darkest on a
scale of 1 to 7. The bar charts show the relationship between ostracode shell color, threshold criteria
result, and facies for 21 samples. Each row in the chart represents ostracode count data from one
sample. Rows of data are grouped together vertically by threshold criteria and then sorted within
those groups by facies (see colored circles on the left). The length of bars in each row represents the
relative proportion of shells in each sample that are of a particular color. The bar color represents
the ostracode shell color. Numbers within bars indicate the actual number of valves of each color.
From the microscope images (left to right): '(1) Translucent (LV, Cyprididae, sample B5); (2)
White (RV, Limnocytheridae, sample D14); (3) Partly light gray (RV, Darwinulidae, sample D14);
(4) Light gray (RV, Limnocytheridae, sample C6); (5) Partly dark gray (RV, Darwinulidae, sample
C6); (6) Dark gray (RV, Darwinulidae,,sample C6); (7) Black (LV, Limnocytheridae, sample L7).
LV: Left valve external view, RV: Right valve external view.
60
biased young in sample F14 (-8200 ka), in which an analysis comprised purely of mollusc
shell fragments yielded an age -3000 yrs younger that the surrounding bulk sediment
(Table 2.4). Although we visually screened for and manually removed identifiable mollusc
shell fragments from samples before processing, it is possible that smaller unidentified
fragments remained; if differential amounts of mollusc shell fragments were included in
replicate analyses, poorly reproducing U-Th data might be an expected result. There is
some qualitative indication that the RGA facies is more abundant in mollusc shell fragments
than other facies, but the results are not conclusive.
2.7 Modeling the effects of detrital contamination and ura-
nium remobilization
We have examined various sedimentological and geochemical data to evaluate a few hy-
potheses for poorly behaved U-Th data. In this section, we simulate the uranium and
thorium isotopic evolution of samples with various compositions and uranium loss/gain
pathways, and compare these model results with the actual measured isotopic composition
of our samples (Figs. 2-13 and 2-14). Based on these results, we posit that the balance
of evidence from both modeling and the previous data comparisons (Section 2.6) favors
uranium remobilization as the main explanation for the observed scatter, rather than sub-
stantial detrital contamination.
To model the effects of detrital contamination, we calculate the impact of mixing varying
amounts of detrital material with pure endogenic carbonate. We also test the effects
of varying the composition of the detrital material, changing the relative proportions of
marine limestone and aluminosilicate. We assume that all detrital material is isotopically
homogeneous and at secular equilibrium, in which [230Th/ 23 8U] and [23 4 U/ 23 8U] are both
equal to 1. The uranium and thorium concentrations of the detrital components are set to
representaverage values for marine limestone and the upper continental crust (Rudnick and
Gao, 2003): U conc. = 2 ppm and Th conc. = 1 ppm for marine limestone, and U conc.=
2.7 ppm and Th conc. = 10.5 ppm for the aluminosilicate material. Our.measurement
of the uranium and thorium concentration of a dark gray silt sample that has- 30%
CaCO3 content (sample E2; Fig. 2-6) suggests that using these values to simulate the
detritus entering Lake Junin is not unrealistic. Although organic matter is a non-negligible
constituent of our samples (1-16%; Fig. 2-11) and is likely a meaningful uranium source
61
(Fig. 2-1D), we chose to exclude TOC in our model as a necessary simplification for this
exercise. As such, the modeling results should be treated as proofs of concept, rather than
a serious attempt to precisely quantify the manner by which each deviating U-Th date
occurs.
We focus on modeling two groups of data: samples from 20-23 m and samples from
-70-75 m. Taking the passing U-Th dates within these groups at face value, these depth
ranges correspond to samples with nominally true ages of ~75 and- 550 ka, respectively.
The scatter of threshold criteria failing data at ~20-23 m all originate from the CMC facies
and are generally biased older relative to the passing dates, whereas the data at 70-75
mostly consists of the RGA facies and are biased younger (Fig. 2-15B).
2.7.1 Modeling results for ~75 ka-aged samples
Fig. 2-13A compares our U-Th data from the first group (circles) to our simulations of
isotopically evolving samples (colored lines) in 230Th/ 238 U activity-8 234 U space. All models
of samples are evolved for 75 kyrs. To orient the reader: the measured isotopic composition
of the two passing samples in this depth range, K16 and L1, are marked by the pair of
gray circles located adjacent to the pair of blue circles (see legend). The horizontal offset
between the colored and gray circles represents the effect of the initial 230 Th correction, the
magnitude of which is controlled by the 238 U/ 232Th ratio (Eq. 2.3). The gray shaded region
delineates the age range prescribed by the age-depth model (to be shown and discussed
in Section 2.8); for our purposes here, we treat this range as the "true" age range of all
samples shown. Thus, any colored circles that do not fall within the gray shaded region
are samples that yielded apparent dates that are inconsistent with the true age of these
sediments (see straight lines labeled by date). The goal of the subsequent exercises is to
explore what pathways of isotopic evolution can explain the isotopic composition of these
outlying data.
We first approximate the starting isotopic composition of a representative endogenic
carbonate by reverse engineering the isotopic composition of the passing samples. K16 and
Li have an average uranium concentration of 2.6 ppm, carbonate content of ~60%, and
TOC of -10%. These values indicate that non-carbonate detrital material accounts for
-30% of the sample composition. If we set the detrital end member to be made entirely
of aluminosilicate material, we find that a sample consisting of the following material can
roughly match the end isotopic composition of K16 and Li after evolving for 75 kyrs:
62
MEN
0 2500
32000
1000 
10010 2000 -
01[X
21 '1000
PATHWAY
0 
0 25 50 75 ita KEY
Actual Date (kyrs) 500 isotopc
3
2 U-TH DATA isompK
1 0 corrected s4-4 e uncorrected ate r
W or inital"e
0
0 Uconc.(ppm) 3 UST OF PATHWAYS
4 Ucn--500- Pure endogenic carb.-500.---Endogenic carb. with 20/40/60/80% detrital
- 30% detrital - 25/50% Uloss at 65 kyr (late)
C2 - 60% detrital - 50/75% Uloss at 1 kyr (early)0 -10020%de trital - 1% Uloss every 1 kyr (continuous)
0 -100011111
0 0.5 1.0 1.5 0 0.5 1 . 1.5 2 2.5 _3
conc.(pprr ) 2302Th Th"U activity
Figure 2-13: Possible pathways of uranium-thorium isotopic evolution that may explain the outly-
ing data at -20-23 m. [A] Plot comparing U-Th data (circles; see legend) with possible pathways
(colored lines) in 8234 U- 230 Th/ 238 U activity space, following Fig. 2-1C. Corrected U-Th ratios
(color-coded by threshold criteria result) are each paired with their corresponding uncorrected ra-
tios (gray circles with dashed outlines). Thegray triangular wedges represent the expected age
range of samples from this depth range based on the age-depth model, where the dark gray area
is the range of model means (red line in Fig. 2-15A) and the light gray area is the maximum and
minimum of the*uncertainty range (shaded gray area in Fig. 2-15A). Colored lines represent the
isotopic evolution of sample material of mixed composition and uranium loss histories over 75 kyrs
(see list of pathways and pathway key for symbology). The simulated samples are mixtures of.two
isotopically homogeneous end-members: pure endogenic carbonate and detrital material made of
30% limestone and 70% aluminosilicate material. The starting composition of the pure endogenic
carbonate is 2.8 ppm 23 8 U and 2700%o for 62 34 Uinitia. These calculations assume that the endogenic
carbonate contains no initial 230 Th and that uranium loss occurs with no fractionation between
234 U and 2381. [B] and [C] Change in the isotopic evolution as the proportion of detrital material
increases (red-lines). [D] and [E] Change in 2 3 0Th/ 2
38U activity after uranium loss 2 3 4 U isnot
shown because its evolution is no different fromthat without uranium loss (compare dashed yellow
238 232
line with red line of same starting composition in Panel C). [D] and,[E] U and Th concentra-
tions of IU-Th samples in this depth range. K16 and Li are of the CP1 facies; all other U-Th data
featured are of the CMC facies'(see Fig. 2-15B).
63
20-30% detrital material and 70-80% endogenic carbonate with a starting composition
of 8 234Uinitia = 2700%c and uranium concentration = 2.8 ppm (Fig. 2-13A, colored lines
labeled 20% and 30%; see legend). The blue pathway'represents the isotopic evolution of a
pure endogenic carbonate of this composition over the same amount of time, 75 kyrs: with
no 232Th, its initial 23 0 Th/ 238 U activity is zero and its final isotopic composition matches
its true age. Adding detrital material to this pure endogenic carbonate increases the initial
23 0 Th/23 8U activity and decreases the initial 8234Uofthesample(redpathwaysinFig.2-
13B-C), causing the sample to yield an older apparent date after 75 kyrs (red pathways in
Fig. 2-13A). Note that we tested the sensitivity of these simulations to the composition of
the detrital material by varying the proportion of marine limestone, and found that this
detail has a small impact compared to that of the total proportion of detrital material.
From the red pathways in Fig. 2-13A, it is clear that no amount of detrital material is
able to explain the samples with high measured 230 Th/ 238U activities, given the compo-
sition of the end members being mixed. Furthermore, the 23 2Th concentrations of these
samples place an inexact but actionable upper bound on how much detrital material is
reasonable (Fig. 2-13G); for example, a 30% contribution of detrital material that is 50%
limestone already produces a sample with 232 Th concentrations of ~1.7 ppm. Thus, we
invoke uranium loss to explain these data: yellow, pink, and green pathways in Fig. 2-13A
illustrate the impact of continuous, late, and early uranium loss on the sample isotope
composition (see legend and Fig. 2-13D-E). These samples are also the same low uranium
CMC samples failing the 238 U/ 232 Th criterion described earlier. We cannot infer which of
these uranium loss scenarios is most likely at work from isotopic measurements alone, but
our hypothesis for the CMC samples described in Section 2.6.2 would favor early loss. In
addition, the magnitude of uranium loss required to approach the isotopic composition of
the outlying samples is similar to the difference in uranium concentration between passing
and failing samples (Fig. 2-13F).
Note that these modeled uranium loss pathways are simplified to assume that loss
occurs with no fractionation between 234 U and 238 U, but preferential leaching of 234 U is
more likely closer to reality (Section 2.5).
2.7.2 Modeling results for ~550 ka-aged samples
Fig. 2-14 features the U-Th samples from ~70-75 in. At this age, the close spacing of the
age isolines in this regime causes the area defining the "true" age.range of these samples
64
to occupy a much narrower area in 23oTh/ 238 U activity- 23 4U space (shaded gray area
in Fig. 2-14B). Here, we start with a pure endogenic carbonate with a composition of
8 23 4 Uinitai = 2700%o and 238U = 1.8 ppm, values selected via the same reverse engineering
steps described in Section 2.7.1, but using the passing samples in this depth interval.
Detrital contamination has much less impact on the accuracy of the dates at this age, as
demonstrated by the fact that all the modeled pathways of samples with varied percentages
of detrital material still ultimately end in the region defining the true age of the sediments
(gray shaded region in Fig. 2-14B; see Section 2.2.2 for explanation). In order to produce
samples with isotopic compositions that bias dates to be younger, we simulate the impacts
of uranium gain.
Similar to detrital contamination, early uranium loss (or gain) has no effect on the
final sample date at this age. Thus, in Fig. 2-14B, we only illustrate the impacts of late
or continuous uranium gain (green and yellow pathways). In contrast to the pathways
observed for -75 ka-aged sediments, here very small percent gains in uranium can have
measurable impacts. For example, a 1% gain at 540 kyr can cause the sample date to be
-100 kyrs younger (green pathway in Fig. 2-14B). Thus, the sensitivity of the apparent
dates to small alterations in uranium in this regime, in combination with narrowly spaced
age isolines, is likely the cause for the large spread of dates at this depth, as well as the poor
reproducibility of replicate analyses (Fig. 2-8D). Because early uranium gain cannot explain
the young bias of these samples, the hypothesis for post-depositional uranium uptake by
gastropod shells seems less convincing; gains would have to be continuous and gradual,
and it is unclear from where the continuous supply of uranium would come.
Note that although Figs. 2-13-2-14 do not explore compound gain or loss pathways,
such scenarios are not outside the realm of possibility.
2.8 Conclusions: the age-depth model for the PLJ-1 splice
Through the use of threshold criteria that evaluate samples on the basis of their 
23U8 / 232 Th
ratio, reproducibility, and 2 3 4Uiec, we have conducted a methodical curation of the U-Th
data that is justified by comparisons to other sedimentological, geochemical, and paleoe-
cological datasets, as well as modeling of the isotopic evolution of simulated samples. As a
result, we deem 18 of the 55 sample dates as satisfactory for use as age constraints. These
passing samples generally come from sediments of the CP facies, which we hypothesize
have experienced relatively less uranium remobilization because of the thick overlying peat
65
- 70.?-74.9rm
3000 600 .7-
550-
1000
500 -
01
0 0.5 1 1.5 2 2.5
2nThPnU activity
450 PATHWAYcompositkn
stpKEY
isoCtopic
0 400- ef17 b'on
~3000 U ~nin
2000HO 7
1000
, 0 /U-THNDATA
0 100 200 300 400 500 350 cr t Ducr
Actual Date (kyrs) for niti
a la"Th
1.4 300 LIST OF PATHWAYS
- Pure endogenic carbonate
1.3 - Endogenic carbonate with 10% detrital
- Endogenic carbonate with 30%detrital
400 -- 50% detrital - 1% U gain at 540 kyr (late)
- 40% detrital - 0.5% U gain every 20 kyr
300 250
530 535 540 545 550 1.3 1.4 1.5 1.6 1.7 1.8
Actual Date (kyrs) 2"Th"U activity
Figure 2-14: Possible pathways of uranium-thorium isotopic evolution that may explain the out-
lying data at -70-75 m. See caption in Fig. 2-13 for explanation of symbology. [A] Plot showing
initial isotopic composition and evolution of pathways over 550 kyrs. [B] View of the isotopic
evolution of samples within the extent represented by the rectangular box in Panel A. The gray
triangular wedges representing the expected age range of samples from this depth range occupies a
much narrower area in 230 Th/ 2
3 8U activity-8234U space compared to Fig. 2-13. The starting com-
23 234
position of the pure endogenic carbonate is 1.8 ppm 8U and 2700%o for 8 Uinitiai. The detrital
component has the same composition as that from Fig. 2-13,3 0% limestone and 70% aluminosili-
cate material. These calculations assume that uranium gain-occurs with a 8234U ratio equal to the
8234 Uinital of.the sample. [C] and [D] Change in isotopic evolution as the proportion of detrital
material increases. [E] and F] Change in isotopic evolution after uranium gain.
66
beds that act as reductive barriers to post-depositional fluid flow.
Fig. 2-15A shows the age-depth model for the PLJ-1 splice using these U-Th ages and
radiocarbon data. The model was generated using the R-based Bayesian age-depth mod-
eling software program called Bacon (v2.3; Blaauw and Christen, 2011). On'average, the
95% confidence range of this model is -30 kyrs. Trachsel and Telford (2017) tested Bacon
and other age-depth modeling routines (CLAM, OxCal, BChron) on a varved sediment
sequence and found that they all produced mean age-depth models close to the truth, but
each program has its own advantages and disadvantages. In the case of Bacon, the ap-
plication of an accumulation rate prior forces sedimentation rates to be more smooth and
linear than is possibly justified. Thus, while the alternating packages of carbonate and
glaciogenic silt in the core hint at variability in sedimentation rates, this information is not
utilized in the generation of the age-depth model. See Section 2.10.5 for details regarding
the parameters and priors used for the model run.
Fig. 2-15A also compares the radiometric age-depth model to geomagnetic relative
paleointensity (RPI) tie points made between the PLJ-1 normalized remanence record and
well-dated RPI stacks (Hatfield, in prep.). Broadly, these data are consistent with age-
depth model and provide further support for its validity. See Section 2.10.6 for details on
the determination of the RPI tie points.
2.9 Considerations for future U-Th dating of lake sediments
When it comes to U-Th dating of lake sediments, there are no "silver bullets" or easy
answers: no singular facies, carbonate content threshold, color, or any other sedimentolog-
ical or geochemical data could predict the viability of a U-Th date with certainty in these
sediments. In fact, samples that would conventionally be considered ideal were some of the
most poorly behaved samples. One wonders what the outcomes might have been if only
the nominally choicest samples had been processed, and the dateability of the entire core
assessed from those results. Such decision making processes are the norm when less time
and fewer resources are available.
Our concern for overlooked but dateable sediments in other records also extends in
the opposite direction: other studies may be overly reliant on single-analysis U-Th dates
that seem credible but have not been reproduced or tested with stratigraphic coevality
67
.. ... I IE
do
70 0 AGE-DEPTHMODEL
- mean
600 2-a error
envelope
500 INPUTS OTHER
* 4 C *RPI
40 0 U-Th U-T10 (failed)
0 -7
h300
200
1010
_F_ 10 20 30 40 50 60 70 80
Depth (m)
100
-00
4-100 - -- e CMC
4-100
o -100 0 CP2
10 20 30 40 50 60 70 80
Depth (m)
Figure 2-15: [A] and [B] Age-depth model for the PLJ-1 splice generated by Bacon (Blaauw and
Christen, 2011) using radiocarbon (gray squares) and U-Th (blue circles) data. Red line represents
the mean and gray shaded area is the 2-a range error envelope. U-Th data that did not meet
threshold criteria (light gray circles) and relative paleointensity (RPI) tie points (green diamonds)
were not used as inputs for the age-depth model and are plotted as an overlay. Section 2.10.5 for
additional information on the priors and posteriors of the model. Radiocarbon data are from Woods
et al. (2019) and RPI data are from Hatfield et al. (in prep.). [C] Comparison of all U-Th dates
and their departure from the age-depth model mean (thick red line at zero), color-coded by facies.
See Fig. 2-8D to compare with data color-coded by threshold criteria result. [D] Box-and-whisker
plots showing the distribution by facies of the difference between U-Th dates and the age-depth
model mean; see caption ofF ig. 2-11 for description of box-and-whisker plot symbology. All error
bars shown are of 2-o range. Analyses that yielded indeterminant U-Th dates or were infinite are
not'included in this figure.
constraints. Themost glaring example of this can be demonstrated through the results
from sample L7, which had the highest uranium concentration (7 ppm, 2 times higher than
the next highest) and highest 2 30 Th/ 232Th ratio out of all samples (Table 2.3).. Because
68
of this, individual analyses were rather precise, with 400-700 year uncertainties (2-a) on
~100 ka dates. On their own, these dates would be considered excellent, but only after
replication is it revealed that none of the precise dates overlap with one another at the
2-a level. Although a labor intensive strategy, there is no substitute for replication and
reproducibility in assessing the quality of U-Th geochronological data.
Furthermore, while the threshold criteria ultimately decided which data would form
the foundation of the age-depth model, it was the placement of geochronological data in
context of other sedimentological and geochemical information that provided justification
for these thresholds. These data comparisons also provide practical insights oi what other
characteristics to consider for future U-Th dating attempts on lake sediments: for instance,
the aforementioned L7 sample contained some of the darkest ostracode shells categorized
(see image of ostracode shell with color = 7 in Fig. 2-12).
Embracing the noise in our data has led to a richer understanding of the controls on
uranium in these lake sediments. As our ability to resolve this noise increases as the ana-
lytical precision of measurements improves, subtle differences in the noise will become in-
terpretable as information on paleoenvironmental processes themselves. This work demon-
strates the beginnings of what is possible on this front. Although >150 analyses went into
this work, we hope that this number does not intimidate those seeking to apply U-Th
dating to their own lake sediment samples. Rather, we seek to showcase strategies for
interpreting scattered geochronologic data of any size and encourage similar efforts where
better geochronological control would have the most impact. As more high resolution
datasets become paired with drill cores by default (e.g., scanning XRF, color reflectance,
magnetic susceptibility), there will be more opportunities to use such additional data to
test underlying working assumptions for geochronologic tools.
Of all the lessons learned, we hold the following as most important: for the deter-
mination of the age of lake sediments, geologic context-in the form of sedimentological
observations, geochemical data, and paleoecological descriptions-is of equal importance
to the numerical accuracy and precision of geochronologicalmeasurements.
69
Table 2.3: Average values of various U-Th data for each sample, calculated from replicate analyses. Samples that passed all threshold criteria (Section 2.5)
are listed first; then all samples are sorted by depth (CCLF). All uncertainties listed are 2-a. N is the total number of replicate analyses made for each
sample. MSWD and prob. of fit fields are left blank if there is only 1 replicate analysis or marked with an asterisk (*) if any replicate analyses yielded
infinite dates; samples marked with an * are considered to have failed the reproducibility criterion. Weighted Mean Date is reported relative to Before
Present (BP), in which the present is defined as January 1, 1950. 8 2 34 Um is the measured S2 34 U. Meaning of the abbreviations used for facies can be
found in Section 5.2. For threshold criteria result, numbers indicate which threshold criterion each sample failed: 1 = 283 U/ 232 Th; 2= reproducibility;
and 3 2 3 4 Uiac. See Table 2.4 for data associated with replicate analyses and Supplementary Materials for raw unrounded data.
Threshold
Samp.CCLF 23 8U 23 2 Th Prob. Weighted S234 Ujac 23 8 U/ 23 0 Th/ Crite-
N MSWD of Mean Date [2 30 Th/ 23 8U] Facies
ID (m) (ng/g) (ng/g) (%0) (%) 2 3 2 Th 2 3 2 Th ria
Fit (yrs BP) Result
Passing Samples
F9 1.080 3 2911±70 124i25 0.07 0.93 1810+200 0.0760±0.0027 2355.0±1.6 2395±19 24.i4.6 28.9±4.5 - pass
F14 3.530 4 697±58 25.2±5.5 1.38 0.25 8180±150 0.2822±0.0061 2639±11 2727±21 28.5±5.1 127±20 - pass
C F15 3.850 3 980±32 105±13 0.11 0.90 9600±500 0.3665±0.0073 2721±24 2880±53 9.39±0.91 54.5±4.3 - pass
F16 4.330 2 1150±150 250±140 0.87 0.35 10500±1000 0.4251±0.0076 2630±120 2880±130 5.2±2.3 35±15 - pass
C16 4.660 3 1580±250 243±32 0.17 0.84 12900±800 0.4181±0.0089 2085.3±3.8 2257±59 6.51±0.18 43.2±0.46 - pass
D4 4.880 4 1970±71 497±18 0.03 0.99 13200±1200 0.4502±0.0046 1934.3±5.8 2156±95 3.965±0.053 28.34±0.38 - pass
K16 20.806 5 26i0±700 1430±350 0.11 0.98 69000±2000 1.512±0.017 1745±20 2490±260 1.824±0.047 43.8±1.0 CP1 pass
Li 20.823 5 2610±540 1300±380 2.77 0.03 70000±4000 1.535±0.067 1774.7±3.4 2500±260 2.04±0.14 49.6±1.7 CP1 pass
D6 33.839 5 1200±110 1000±160 2.06 0.08 184000±6000 2.176±0.075 1288±74 2800±520 1.211±0.096 41.9±4.7 CP2 pass
M1 45.599 4 2520±52 430±23 3.05 0.03 323000±3000 2.1303±0.0061 936.0±2.0 2440±85 5.87±0.21 198.7±7.6 CP1 pass
M3 45.660 5 1740±130 376±25 0.49 0.75 323000±2000 2.212±0.011 999.5±8.0 2650±160 4.63±0.21 162.8±7.5 CP1 pass
J8 45.686 3 2660±110 341±34 0.79 0.46 317500±1700 2.1828±0.0046 985.3±4.8 2501±68 7.83±0.46 271±16 CP1 pass
J9 45.741 3 3059±61 1179±24 0.59 0.56 311000±3000 1.9583±0.0064 807.5±3.1 2170±160 2.594±0.042 80.7±1.6 CP1 pass
E12 55.318 5 953±39 940±110 1.26 0.28 416000±20000 1.828±0.054 631±31 2700±1100 1.017±0.076 29.5±3.0 CP2 pass
F4 55.587 5 1092±80 120±3.6 3.32 0.01 391000±5000 2.0570±0.0064 816.9±4.1 2550±130 9.10±0.40 297±13 CP2 pass
B5 70.941 5 1699±34 364±17 1.00 0.41 514000±14000 1.6727±0.0098 489.7±3.6 2230±610 4.68±0.20 124.2±53 CMC pass
B11 71.182 4 2650±170 353±34 1.70 0.16 554000±19000 1.6160±0.0082 443.9±4.1 2190±400 7.52±0.28 193.1±7.7 RGA pass
B13 71..190 4 1318i84 134±13 1.69 0.17 580000±20000 1.6572±0.0059 469.5±4.9 2600±1300 9.88i0.41 260±11 RGA pass
Table 2.3 continued from previous page
Threshold
23 8 23 2 Prob. WeightedSamp.CCLF U Th 234um 2 38 2 30(234Uc U/ Th/ Crite-
N MSWD of Mean Date 2 3 0
(ng/g) [ Th/
23
8U] 2 3 2 2 3 2 Facies
ID (in) (ng/g) (%0) (700) Th Th ria
Fit (yrs BP)
Result
P10 6.505 2 749±33 837±52 * * 0.540±0.013 792.8±2.3 0.896±0.028 7.68±0.22 CP2 1,2
L3 21.649 1 619±12 1118±23 - 150000±50000 1.621±0.025 719.6±3.0 2200±1300 0.553±0.028 14.25±0.22 CMC 1
G1 21.725 1 851±17 1230±130 - 70000±50000 1.19±0.12 673.8±2.0 1340±660 0.69±0.11 13.1±1.9 CMC 1
G2 21.878 2 1164±23 1435±70 0.01 0.92 82000±11000 1.700±0.027 1497±11 2840±860 0.812±0.037 21.91±0.64 CMC 1
G5 22.364 1 635±13 1395±28 - 280000±40000 2.162±0.040 966.7±1.7 5200±4200 0.455±0.028 15.61±0.29 CMC 1
G6 22.367 4 351±58 692±81 0.04 0.99 149000±12000 2.40±0.33 1580±360 5300±4600 0.506±0.035 19.2±1.9 CMC 1
G7 22.486 3 418±9 829±17 0.05 0.96 140000±30000 1.794±0.033 932±21 3000±2200 0.504±0.028 14.36±0.19 CMC 1
G8 22.537 1 861±17 2362±48 - 1.986±0.035 621.9±3.9 0.365±0.028 11.50±0.20 CMC 1
M15 23.467 4 323±7 38.3±3 37.0 0.00 102000±3000 2.219±0.043 2302.9±7.0 3175±64 8.46±0.59 298±23 CP2 2
L7- 24.105 7 6800±330 400t27 399 0.00 106000±5000 1.829±0.074 1687.6±3.8 2316±47 17.03i0.68 494±24 CMC 2
D7 33.954 2 838±18 997±46 0.24 0.63 204000±9000 2.331±0.020 1328±25 3500±940 0.840±0.028 31.1±0.92 CP2 1
D11 35.671 2 392±39 870±210 * 0*.00 1.41±0.11 283.9±7.3 0.457±0.091 10.2±1.3 CMC 1,2
D12 35.673 2 262±47 440±22 * * 420000±50000 2.080±0.024 806.1±1.5 4400±1900 0.60±0.14 18.7±5.9 CMC 1,2
D13 35.705 3 168±11 216±15 8.19 0.00 380000±50000 2.078±0.095 823±59 3900±1600 0.778±0.028 25.7±1.5 CMC 1,2
D14 35.733 2 287±6 291±12 * * 2.663±0.040 1019.2±4.4 0.987±0.028 41.71±0.21 CMC 1,2
J5 45.640 4 1360±120 377±83 . 5.51 0.00 277000±5000 2.131±0.043 1002±40 2370±130 3.71±0.60 126i22 CP1 2
J6 45.661 3 1202±80 389±36 14.0 0.00 266000±9000 2.129±0.023 1016.9±5.4 2380±140 3.095±0.095 104.6±4.3 CP1 2
J7 45.683 2 2467±83 389.8±8.9 97.2 0.00 278000±17000 2.162±0.028 1027.3±1.1 2367±88 6.33±0.36 217.3±9.5 CP1 2
3110 45.921 1 1798±36 2117±42 - 327000±20000 1.977±0.016 789.8±0.8 2930±800 0.849±0.028 26.66±0.21 CP2 I
K1 54.873 1 659±13 654±13 - 550000±90000 1.798±0.014 571.8±1.4 3700±2100 1.000±0.028 28.76±0.23 CMC 1
A5. 55.288 2 2360±140 1496±30 72.6 0.00 310000±70000 1.907±0.044 751±19 2290±380 1.578±0.071 47.8±1.0 CP2 2
A6 55z292 3 2520±190 1143±75 105 0.00 320000±50000 1.996±0.064 819.3±6.5 2400±370 2.202±0.028 69.8±2.0 CP2 2
F13 67.741 4 338±7 278±21 0.29 0.84 408000±17000 1.584±0.057 448±45 1820±500 1.219±0.082 30.7±3.1 CP2 3
B2 70.741 1 516±10 615±12 - 470000±70000 1.565±0.017 416.5±2.1 2300±1200 0.839±0.028 20.84±0.22 CMC 1
G13 71.394 4 734±15 159.6±6.0 * * 512000±17000 1.597±0.020 433±10 2060±770 4.60±0.21 117.5±7.3 RGA 2
G14 71.679 2 3084±62 206.0±7.1 2.11 0.15 485000t15000 1.6282±0.0046 464.3±1.8 1850±170 14.98±0.64 387±17 RGA 3
H6, 72.406 3 1188t89 887t33 * * 550000±70000 1.527±0.012 379.2±2.9 2200t1700 1.34t0.12 33.3t1.6 RGA 2
som
*
Table 2.3 continued from previous page
Prob WeigtedThreshold
Samp.CCLF 23 8 U 2 3 2 Th Prob. Weighted 28 3 4 Uiac 2 38 U/ 23 0 Th/ Crite-
N MSWD of Mean Date [2 30 Th/ 238 U] Facies
ID (m) (ng/g) (ng/g) (%o) (%0) 2 3 2 Th 2 3 2 Th ria
Fit (yrs BP) Result
H7 72.409 4 968±51 607±47 * 443000±19000 1.546±0.011 410.5±4.5 1730±460 1.60±0.10 39.7±2.6 RGA 2
B14 73.327 4 1685±35 698±57 12.0 0.00 370000±30000 1.524±0.034 417±11 1420±620 2.43±0.20 58.7±5.2 RGA 2
B15 73.566 4 479±10 161±10 0.19 0.90 305000±5000 1.410±0.015 373.0±4.3 970±120, 2.98±0.15 66.8±3.5 RGA 3
C1 73.573 4 1230±440 330±130 195 0.00 280000±90000 1.440±0.082 379.7±5.6 1270±460 3.8±0.66 87±13 RGA 2
C3 73.704 2 3508±71 673±20 8.62 0.00 520000±60000 1.514±0.012 374.4±2.5 1770±470 5.213±0.054 125.3±2.3 RGA 2
C6 74.222 1 361±7 422.6±8.5 450000±100000 1.346±0.016 258.6±2.8 1360±920 0.855±0.028 18.27±0.22 CMC 1
C10 74.444 8 417±37 275±31 3.69 0.00 361000±14000 1.565±0.057 453±40 1540±340 1.521±0.095 37.8±3.4 CMC 2
C11 74.529 2 1763±36 339±11 8.18 0.00 480000±40000 1.638±0.014 470.1±4.0 1980±330 5.202±0.088 135.3±3.5 CMC 2
C13 74.739 4 1226±42 356±32 7.11 0.00 376000±20000 1.414±0.026 330±10 1100±230 3.46±0.25 77.9±6.8 RGA 2
C14 74.851 2 2135&43 525+12 * * - 1.5881A0.0049 364.4&3.8 4.069&0.088 102.6A2.5 RCA 2
-~1
Table 2.4: U-Th data for replicate analyses of each sample. All uncertainties listed are 2-a. Rep. ID refers to a unique identifier for each replicate
2 3 2 3 2
analysis. Note that sample 'F14(s)' refers to an analysis of gastropod shells from the same horizon as F14. Reported errors for 8U and Th
concentrations are estimated to be ±2% due to uncertainties in spike concentration; analytical uncertainties are smaller. Date Uncorr. indicates that
2 30 0 2 32
no correction has been made for initial Th. Date Corr. are corrected for initial detrital 23 Th assuming an initial 23
0 Th/ Th of 8.0±4.x10-6
2 30
(Section 5.3), and are reported relative to Before Present (BP), in which the present is defined as January 1, 1950. Decay constants for Th and 
2 4
3 U
are from Cheng et al. (2013a); decay constant for 
2 3 8U is 1.55125x1010 yr 1(Jaffey et al., 1971). See Eqs. 2.1--2.3 for equations for the calculation of
2 34
dates. See Eq. 2.2 for the calculation of 8
23 4
Uinitial and Section 2.10.4 for the calculation of 8 Uiac.
Date
2 32 2 3 4 2 3 4 2 3 4
Samp. Rep. 238u Th 8 Um 23h/238U) 23T/232 Unor Date Corr. 8 Uinitia1 8 Uiac
ID ID (ng/g) (ng/g) (o) (yr) (yrs BP) (%0) (%o)(yrs)
Replicate Analyses of PassingS amples
F9 F9(1) 2831±57 97.5±2.0 2356.7±1.1 0.07300±0.00060 33.67±0.29 2392±20 1830±280 2368.9±2.2 2391±13
F9(2) 2940±59 147.4±3.0 2354.87±0.95 0.07805±0.00082 24.71±0.26 2560±27 1740±410 2366.5±2.9 2399±19
F9(3) 2962±59 127.8±2.6 2353.5±1.1 0.07708±0.00082 28.36±0.30 2529±27 1820±360 2365.6±2.6 2394±16
-- I
F14 F14(1) 737±15 26.29±0.53 2635.37±0.68 0.28880±0.00099 128.59±0.49 8942±32 8410±270 2698.7±2.2 2725±15
F14(2) 715±14 29.16±0.59 2640.01±0.82 0.2838±0.0016 110.42±0.62 8769±50 8160±310 2701.5±2.5 2732±17
F14(3) 725±15 28.26±0.57 2652.8±6.3 0.2823±0.0014 114.98±0.57 8690±46 8110±290 2714.3±6.8 2743±21
F14(4) 611±12 17.08±0.34 2625.96±0.69 0.27407±0.00084 155.78±0.42 8494±27 8080±210 2686.5±1.7 2707±12
F14(s) 292±5.8 5.48±0.11 2636.3±3.6 0.1765±0.0012 149.3±1.3 5395±39 5110±150 2674.6±3.8 2688±11
F15 F15(1) 1000±20 119.7±2.4 2694.2±2.2 0.3748±0.0019 49.72±0.23 11524±61 9770±890 2769.5±7.3 2863±53
F15(2) 997±20 99.0±2.0 2741.1±1.0 0.3624±0.0019 57.96±0.30 10980±61 9540±730 2815.9±5.9 2894±44
F15(3) 944±19 96.9±1.9 2728.17±0.97 0.3621±0.0022 55.96±0.33 11010±70 9520±760 2802.4±6.1 2883±46
F16 F16(1) 1259±25 353.7±7.1 2546.1±1.6 0.4304±0.0049 24.33±0.28 13900±170 9600±2200 2616±17 2830±130
F16(2) 1048±21 155.4±3.1 2717.5±6.4 0.4197±0.0025 44.96±0.22 12884±83 10700±1100 2801±11 2919±69
C16 C16(1) 1305±26 205.5±4.1 2084.2±2.5 0.4278±0.0024 43.12±0.23 16021±96 13300±1400 2163.6±9.0 2261±59
C16(2). 1673±33 258.2±5.2 2088.26±0.47 0.4159±0.0023 42.78±0.22 15529±90 12800±1400 2165.2±8.4 2260±57
C16(3) 1773±35 264.5±5.3 2083.6±3.8 0.4105±0.0023 43.69±0.23 15338±95 12700±1300 2159.8±9.0 2251±56
D4 D4(1) 1946±39 500±10 1930.39±0.61 0.4514±0.0034 27.87±0.20 17920±140 13100±2500 2003±14 2154±95
D4(2) 2073±41 520±10 1929.5±2.5 0.4549±0.0033 28.78±0.21 18070±140 13400±2400 2004±14 2151±92
D4(3) 1943±39 489.1±9.8, 1935.47±0.82 0.4508±0.0038 28.44±0.24 17860±160 13200±2400 2009±14 2157±92
Table 2.4 continued from previous page
Samp. Rep. 232Th 823 4 Um 23at DateCorr. 23 4 Ufnta 8234Uiac
ID ID (ng/g) (ng/g) (ng/g) ((%%0o))  1 Th/
3 U 23 Th/23 Th U(nycrosr)r . ((yr rsBPP)  (%70)  (%YO)
D4(4) 1916±38 477.4±9.6 1942±1.2 0.4438±0.0040 28.28±0.25 17520±170 12900±2400 2014±14 2161±92
K16 K16(1) 3861±77 2049±41 1768±1.0 1.5022±0.0068 44.95±0.19 77570i470 68000±5100 2142±31 2500±240
K16(4) 2258±45 1257±25 1750.7±2.2 1.520±0.017 43.35±0.48 79500±1200 69300±5500 2129±33 2510±250
K16(5) 2357±47 1267±25 1713.3±1.9 1.506±0.016 44.48±0.48 80100±1200 70100±5400 2088±32 2450±240
K16(6) 2293±46 1266±25 1746.5±1.9 1.524±0.017 43.81±0.48 79900±1200 69800±5500 2127±33 2500±250
K16(7) 2290±46 1294±26 1746.9±1.8 1.509±0.017 42.39±0.48 78900±1200 68500t5600 2120±34 2510±260
LI L1(2) 3570±71 1984±40 1780.6±1.1 1.648710.0075 47.09±0.21 87550±560 77700±5200 2217i33 2610±260
L1(5) 2358±47 1113±22 1772.4±2.0 1.528±0.015 51.40±0.49 79200±1000 70700±4600 2164±28 2480±210
L1(6) 2314±46 1106±22 1774.2±1.9 1.529±0.015 50.79±0.51 79200±1100 70600±4600 2166±28 2490±210
L1(7) 2381±48 1115±22 1772.8±1.8 1.473±0.015 49.91±0.50 75350±1000 66800±4500 2141±28 2450±200
L1(8) 2403±48 1171±24 1773.6±2.3 1.499±0.015 48.83±0.50 77100±1100 68300±4700 2151±29 2480±220
D6 D6(1) 1091±22 841±17 1347.5±1.2 2.209±0.013 45.48i0.27 193800±2600 180000±7900 2240±50 2840±420
D6(2) 1070±21 806±16 1385.4±1.7 2.291±0.013 48.28±0.26 202300±2700 189400±7500 2364±50 2980±430
D6(3) 1259±25 1141±23 1212.9±2.7 2.151±0.027 37.69±0.47 211600±6500 194000±12000 2100±69 2790±520
D6(4) 1283±26 1104±22 1241.5±2.0 2.111±0.026 38.94±0.47 196100±5400 180000il1000 2061±62 2690±470
D6(5) 1288±26 1103±22 1252.2±1.8 2.117±0.018 39.23±0.33 195000±3800 178700±9800 2074±58 2710±470
M1 M1(1) 2505±50 427.0±8.5 935.7±2.0 2.1324±0.0049 198.65±0.37 327100±3600 324300±3900 2337±26 2451±85
M1(2) 2497±50 411.7±8.2 938.4±1.1 2.1321+0.0046 205.34±0.34 324700±3100 322000±3400 2328±23 2438±78
M1(3) 2496±50 417.4±8.4 936.29±0.6 2.1351±0.0044 202.73±0.29 328400±2900 325600±3300 2347±22 2459±78
M1(4) 2582±52 462.2±9.2 933.7±1.0 2.1213±0.0050 188.15±0.35 321600±3300 318600±3600 2295±24 2412±84
M3 M3(1) 1828±37 373.8±7.5 1004.66±0.81 2.2154±0.0051 172.03±0.31 324600±3200 321300±3600 2488±25 2635±99
M3(2) 1820±36 384.4±7.7 1010.6±2.2 2.2234±0.0051 167.10±0.27 324900±3500 321600±4000 2504±29 2660±110
M3(3) 1854±37 410,1±8.2 991.3±1.2 2.2034±0.0051 158.19±0.28 327600±3300 324100±3800 2474±27 2630±110
M3(4) 1599±32 342.6±6.9 993.9±5.6 2.207±0.011 163.54±0.69 328000±8300 324500±8500 2484±61 2640±160
M3(5) 1596±32 366.7±7.4 997.2±4.0 2.213±0.011 152.93±0.70 328700±7700 325000±8000 2496±58 2660±160
J8 J8(1) 2590±52 315.1±6.3 987.7I1.6 2.1837±0.0041 285.00±0.34 318300±2700 316300±2900 2412±20 2494±62
J8(2) 2607±52 328.0±6.6 988.5±1.6 2.1862±0.0046 275.81±0.43 319300±3000 317200±3100 2420±22 2506±66
J8(3) 2782±56 379.5±7.6 979.8±1.3 2.1785±0.0043 253.53±0.34 321100±2800 318900±3000 2410±21 2503±68
Table 2.4 continued from previous page
Date
Samp. Rep. 2 3 8 U 2 3 2Th [2342Um3 0 /238u] 23 /232 Date Corr. S2 3 4 2 3 4Uinitia 6 Uiac
ID ID (ng/g) (ng/g) (%) 2 ](yrs BP) (%T/0) (%00)
(yrs) (3sP) (/~) ()
J9 J9(1). 3055±61 1156±23 810.79±0.64 1.9655±0.0064 82.48±0.25 320700+4500 313500+5900 1964+33 2190t160
J9(2) 3058+61 1191±24 804.54t0.79 1.9548+0.0060 79.71+0.22 318700i4100 311200i5700 1936+31 2170+160
J9(3) 3064t61 1192124 807.1+1.3 1.9547+0.0058 79.79+0.20 316600+4000 309100+5700 1931+31 2160t160
E12 E12(1) 907±18 845t17 664.1+2.8 1.890+0.014 32.23+0.27 454000+31000 437000+34000 2280+220 3060t810
E12(2) 913+18 821±16 663.9i2.8 1.880i0.012 33.20+0.21 436000+23000 419000+26000 2170+160 2870i650
E12(3) 977±20 1053+21 606.4+3.5 1.779+0.035 26.22+0.52 397000+54000 373000+59000 1740+300 2500t1000
E12(4) 991+20 1024±21 599.3+4.5 1.780+0.034 27.34+0.52 412000+60000 390000+64000 1800+340 2500±1100
E12(5) 976i20 975±20 620.7+3.5 1.809+0.033 28.74+0.52 414000+56000 393000+60000 1880+330 2600t1100
F4 F4(1) 1006i20 116.2+2.3 818.5+2.1 2.0592i0.0054 283.16+0.61 394900+7100 393000+7200 2482+51 2560±110
F4(2) 1003+20 115.9i2.3 811.9+2.2 2.0516+0.0049 281.86t0.51 396300+6800 394400+6900 2471+49 2550t110
F4(3) 1140±23 122.3±2.4 822.0±1.1 2.0557i0.0044 304.29+0.52 385700+5000 384000+5100 2430135 2503±83
F4(4) 1157+23 122.1±2.4 813.5+1.8 2.0545+0.0057 309.03+0.78 397200+7400 395400+7400 2484±53 2560+120
-4
F4(5) 1154+23 123.2+2.5 818.6+1.9 2.0638+0.0064 306.99+0.87 400200+8300 398400+8400 2520+60 2600+130
B5 B5(1) 1693+34 352.7t7.1 494.4+2.1 1.6751i0.0050 127.68+0.30 503000+21000 499000+21000 2020+120 2150i270
B5(2) 1702i34 393.9+7.9 488.2+1.1. 1.6745i0.0052 114.89+0.33 536000+25000 532000+26000 2190160 2340i360
B5(3) 1687134 355.4+7.1 488.7i2.5 1.6726+0.0087 126.04+0.63 525000+41000 521000±41000 2130+250 2250i540
B5(4) 1690+34 354.2i7.1 490.6+2.2 1.6727t0.0088 126.71+0.65 514000+37000 510000+38000 2070230 2200i480
B5(5) 1721±34 362.0+7.3 486.7i3.6 1.6686+0.0098 125.93+0.69 519000+46000 515000+47000 2090+280 2210i610
Bi1 B11(1)- 2501+50 331.3+6.6 447.3+1.2 1.6230+0.0041 194.52i0.37 567000+28000 565000+28000 2200180 2290±370
B11(2) 2501+50 318.5i6.4 445.21+0.84 1.6214+0.0043 202.18i0.46 576000+30000 574000+30000 2250190 2330+400
B11(3) 2787+56 389.2+7.8 445.3+2.8 1.6149+0.0048 183.58+0.47 538000+31000 535000+31000 2020+180 2100+380
B11(4) 2819+56 374.2+7.5' 437.9±1.9 1.6049+0.0053 191.98i0.59 540000+30000 537000+30000 1990180 2070+370
B13 B13(1) 1280+26 131.9+2.6 468.4±1.1 1.6559t0.0041 255.10+0.46 588000+31000 587000+31000 2450+220 2520i450
B13(2) 1218i24 116.1+2.3 471.6+4.9 1.6556i0.0059 275.87+0.65 558000+51000 557000+52000 2270+340 2330+710
B13(3) 1401+28 145.5+2.9 471.25i0.86 1.6568+0.0036 253.31+0.46 568000+23000 567000+23000 2330+150 2400i320
B13(5) 1372i27 141.3i2.8 466.9t3.0 1.6605+0.0053 256.07i0.72 645000±74000 644000+74000 2870+630 3000+1300
Replicate Analyses of Failing Samples
w=
Table 2.4.continued from previous page
Samp. Rep. 23 8 U 2 32 Th 238 4 Um 23 0Th/ 23 8
Date
U] 23 0 Th/ 2 32 Th U r Date Corr. 8234Uinitial 82 34 Uiac
ID ID (ng/g) (ng/g) (%oO) (yrs BP) (%o) (%0)
(yrs)
*P1O P10(1) 726+15 800±16 794.4±1.5 0.543±0.013 7.83±0.18 38300±1000 792±57 1130±360
1000±25000
P10(2) 772±15 874±18 791.2+1.0 0536+0013 752&018 37700£1100 785&59 1130±380
3000±26000
L3 L3(1) 619±12 1118±23 719.6±3.0 1.621±0.025 14.25±0.22 210400±8200 152000±47000 1100±150 2200±1300
G1 G1(1) 851±17 1230±130 673.8±2.0 1.190±0.120 13.1±1.9 121000±21000 67000±46000 810±110 1340±660
G2 G2(1) 1167±23 1485±30 1504.84±0.76 1.720±0.020 21.46±0.25 109200±2000 83000±16000 1900±86 2910±860
G2(2) 1161±23 1386±28 1488.7±2.8 1.681±0.020 22.37±0.26 106600±1900 82000±15000 1875±79 2780±750
G5 G5(1) 635±13 1395±28 966.7±1.7 2.162±0.040 15.61±0.29 321000±25000 275000±43000 2100±260 5200±4200
G6 G6(1) 435.9±8.7 792±16 1041.8±2.3 1.902±0.025 16.62±0.22 195100i5800 150000±32000 1590±150 3200±1700
G6(2) 315.3±6.3 614±12 1772.4±6.1 2.567±0.029 20.93i0.23 182200±4300 150000±21000 2710±160 5800±3400
G6(3) 316.0±6.3 639±13 1777.8±1.8 2.586±0.031 20.30±0.24 184300±4600 151000±22000 2720±170 6100±3900
G.6(4) 336.2±6.7 721±14 1741.8±4.0 2.539±0.033 18.79±0.24 182500i4900 146000±24000 2630±180 6300±4600
G7 G7(1) 414.1±8.3 831±17 955.3±5.9 1.831±0.025 14.48±0.19 199500±6400 144000±43000 1440±180 3200±2200
G7(2) 426.5±8.5 839±17 925.6±3.7 1.768±0.024 14.27±0.19 191400±5800 135000±44000 1360±170 2900±1900
G7(3) 413.7±8.3 818±16 915.1±6.0 1.784±0.024 14.32±0.19 198100±6300 142000±44000 1370±180 3000±2000
G8 G8(1) 861±17 2362±48 621.9±3.9 1.986±0.035 11.50±0.20
M15 M15(1) .327.4±6.5 38.91±0.78 2306.98±0.78 2.2407±0.0058 299.35±0.64 105080±410 103470±910 3089.3±8.0 3193±58
M15(2) 313.5±6.3 35.45±0.71 2310.06±0.98 2.2675±0.0057 318.39i0.64 106800±400 105280±870 3109.2±7.7 3208±56
M15(3) 322.6±6.5 42.20±0.85 2300.3±1.5 2.1937±0.0065 266.23±0.65 102160±450 100400±1000 3053.5±8.9 3166±64
M15(4) 327.3±6.5 36.66±0.73 2294.4±1.5 2.1731±0.0055 308.04±0.61 101030±380 99490+860 3038.2±7.6 3134±54
L7 L7(1) 6960±140 383.1i7.8 1690±1.4 1.8670±0.0044 538.4±2.5 110430±410 109510±620 2302.0±4.4 2337±21
L7(2) 6430±130 372.4±7.6 1686.3±1.5 1.7499±0.0042 479.5±2.0 100410±370 99420±620 2232.4±4.4 2268±22
L7(3) 7340±150 451.5±9.2 1691.47±0.64 1.9770±0.0046 510.6±1.9 120680t450 119670+680 2371.0±4.6 2411±24
L7(4) .6460±130 392.9±8.0 1687.04±0.74 1.8028±0.0041 470.4±1.9 104900±360' 103870+630 2261.7±4.2 2300±23
L7(5) -6920±140 417.7±8.4 1692.2±1.5 1.8034±0.0041 474.2±1.3 1046401370 103620±630 2267.1±4.5 2305±23
L7(6) 6590±130 388.6±7.8 1682.24±0.70 1.8101±0.0040 487.5±1.2 105840±350 104850±610 2261.5±4.0 2298±22
L7(7) 6900±140 393.6±7.9 1684.15±0.78 1.7943±0.0047 499.6±1.4 104340±420 103370±640 2254.6±4.2 2290±22
Table 2.4 continued from previous page
Samp. Rep. 2 3 2238u Th 230Th/23aU 23T/23 UD 23te4 Um Date Corr. 2 3 46 Uinitial 82 3 4 Uiac
ID ID (ng/g) (ng/g) (%o) (yr (yrs BP) (%e) (%)
(yrs)
D7 D7(1) 850&17. 1029i21 1310.5+1.5 2.322i0.020 30.45±0.26 227500i5100 206000±13000 2344i89 3500t940
D7(2) 825±17 965±19 1346.2±1.7 2.339±0.019 31.76±0.25 222000±4600 202000±13000 2378±85 3490±890
Dl D11(1) 365.0±7.3 725±64 289.1±1.1 1.330±0.110 10.7±1.3
D11(2) 419.7±8.4 1025±21 278.7±2.1 1.493±0.039 9.71±0.26
D12 D12(1) 228.8±4.6 456.0±9.2 596.8±1.8 1.829±0.036 14.57±0.29
D12(2) 294.8±5.9 424.9±8.5 806.1±1.5 2.080±0.024 22.91±0.26 447000±42000 424000±46000 2670±360 4400±1900
D13 D13(1) 161.4±3.2 204.2±4,1 862.8±2.5 2.168±0.018 27.20±0.22 467000±35000 448000±38000 3060±340 4700±1600
D13(2) 161.8±3.2 210.5±4.2 851.3±1.8 2.087±0.022 25.46±0.27 378000±23000 354000±28000 2310±180 3600±1200
D13(3) 180.4±3.6 233.1±4.7 755.1±2.2. 1.979±0.021 24.31±0.26 399000±28000 375000±33000 2180±210 3400±1100
D14 D14(1) 291.2±5.8 299.1±6.0 1022.3±1.6 2.691±0.014 41.60±0.21
D14(2) 282.0+5.6 282.1+5.7 1016.1 2.3 2.634±0.014 41.82±0.21
J5 J5(1) 1310±26 411.7±8.2 946.8±1.7 2.0688±0.0067 104.48+0.27 284200±3300 278500±4500 2078±27 2270±130
J5(2) 1383±28 336.9±6.8 1031.87±0.62 2.1567±0.0078 140.58±0.54 277800±3300 273600±4000 2233±25 2390±110
J5(3) 1238±25 285.7±5.7 1032.3±1.5 2.1597±0.0059 148.57±0.32 278800±2600 274800±3300 2242±21 2393±99
J5(4) 1518±30 473.5±9.5 997.5±1.4 2.1404±0.0054 108.97±0.22 289100±2700 283600±3900 2221±25 2430±130
J6 J6(1) 1110±22 348.2±7.0 1019±1.9 2.1544±0.0089 109.03±0.40 283500±4100 278100±5000 2234±32 2440±140
J6(2) 1247±25 402.2±8.0 1021±1.2 2.1230±0.0057 104.53±0.24 269200±2400 263400±3800 2147±23 2350±130
J6(3) 1248±25 416.3±8.3 1010.8±1.6 2.1089±0.0061 100.38±0.25 268300±2600 262300±4100 2119±25 2330±140
J7 J7(1) 2526±51 383.5±7.7 1027.16±0.9 2.1421±0.0041 224.01±0.27 274000±1800 271400±2200 2209±14 2305±62
J7(2) 2408±48 396.1±7.9 1027.4±1.1 2.1816±0.0047 210.61±0.33 291400±2300 288600±2700 2320±18 2429±72
J10 J10(2) 1798±36 2117±42 .789.80±0.77 1.977±0.016 26.66±0.21 351000±14000 327000±20000 1990±110 2930±800
K1 K1(1) 659±13. 654±13 571.8±1.4 1.798±0.014 28.76±0.23 568000±84000 552000±87000 2720±710 3700±2100
A5 A5(1) 2260±45 1479±30 737.1±1.3 1.938±0.013 47.01±0.31 379000±14000 366000±16000 2073±96 2520±380
A5(2) 2464±49. 1513±30 764.48±0.95 1.876±0.010 48.50±0.25 299700±6300 286700±9500 1717±46 2060±260
A6 A6(2) 2442±49 1117±22 812.2±1.9 1.9609±0.0081 68.09±0.28 316400±5600 307500±7400 1935±41 2210±200
A6(3) 2376±48 1086±22 825.1±1.3 2.0697±0.0083 71.91±0.27 396900±9900 389000±11000 2477±76 2830±280
A6(4) 2737±55 1227±25 820.6±2.2 1.9575±0.0087 69.31±0.28 307500±5600 298700±7300 1907±40 2170±190
F13 F13(1) 339.3i6.8 308.316.2 380.6i2.0 1.499t0.013 26.20±0.22 442000±36000 418000±40000 1240±150 1640±500
Table 2.4 continued from previous page
Samp. Rep. 2 3 2238u Th 82 3 4 Uzn Date Corr. 82 3 4 Uinitai 8 2 3 4 Uiac
ID ID (ng/g) (ng/g) (%o) 2 3 Th/23 8 U 2 3 0 Th/ 2 3 2 Th Uncorr.() ()
(yrs)
F13(2) 339.2±6.8 265.6±5.3 467.8±2.6 1.606±0.015 32.57±0.30 417000±31000 398000±34000 1440±140 1830±460
F13(3) 342.1±6.8 278.5±5.6 469.4±1.7 1.611±0.014 31.41±0.26 422000±29000 403000±32000 1460±140 1880±460
F13(4) 330.7±6.6 260.6±5.2 472.6±1.8 1.621±0.013 32.66±0.27 434000±30000 416000±33000 1530±150 1940±480
B2 B2(1) 516±10 615±12 416.5±2.1 1.565±0.017 20.84±0.22 494000±68000 465000±75000 1550±340 2300±1200
G13 G13(1) 728±15 166.4±3.3 428.3±1.8 1.5828±0.0049 109.95±0.31 501000±22000 496000±22000 1740±110 1850±240
G13(2) 727±15 163.0±3.3 426.5±2.1 1.5877±0.0052 112.52±0.32 533000±30000 528000±30000 1900±160 2020±360
G13(3) 742±15 155.6±3.1 445.6±5.7 1.640±0.010 124.22±0.68
G13(4) 737±15 153.6±3.1 445.3±2.1 1.6200±0.0082 123.40±0.60 566000±56000 563000±56000 2180±360 2310±770
G14 G14(1) 3066±61 211.1±4.2 465.4±1.8 1.6273±0.0046 375.25±0.92 477000±16000 476000±16000 1784±84 1820±170
G14(2) 3101±62 201.0±4.0 463.2±1.0 1.6290±0.0037 399.07±0.58 493000±14000 492000±14000 1855±73 1890±150
H6 H6(1) 1115±22 918±18 364.6 1.7 1.648±0.016 31.79±0.30
H6(2) 1288±26 891±18 378.4±2.9 1.523±0.012 34.95±0.27 547000±82000 532000±85000 1700±430 2100±1100
00 H6(3) 1161±23 852±17 380.1±1.6 1.531±0.012 33.11±0.25 580000±110000570000±1200001900±690 2400±1700
H7 H7(1) 893±18 565±11 404.74±0.75 1.607±0.012 40.31±0.30
H7(2) 1003±20 597±12 413.5±1.5 1.551±0.010 41.34±0.26 463000±31000 449000±33000 1470±140 1750±390
H7(3) 988±20 675±14 405.4±1.7 1.5410±0.0110 35.85±0.26 466000±37000.449000±39000 1440±160 1770±460
H7(4) 989±20 589±12 412.8±2.2 1.5452±0.0099 41.20±0.26 448000±28000 434000±30000 1400±120 1680±350
B14 B14(1) 1676±34 739±15 410.00±0.93 1.5156±0.0067 54.60±0.24 392000±12000 380000±14000 1200±47 1360±160
B14(2) 1664±33 614±12 433.5±1.1 1.5425±0.0064 66.35i0.26 385000±11000 376000±12000 1251±43 1390±130
B14(3) 1669±33 727±15 413.2±2.8 1.557±0.015 56.78±0.54 482000±57000 472000±58000 1570±270 1780±620
B14(4) 1730±35 712±14 411.5±4.2 1.480±0.014 57.15±0.53 338000±18000 327000±19000 1035±58 1170±160
B15 B15(1) 477±10 155.4±3.1 373.5±1.9 1.4102±0.0058 68.70±0.30 314400±6400 304500±8300 882±21 968±82
1B15(2) 482±10 171.7±3.4 375.8±2.1 1.4136±0.0061 62.94±0.27 314900±6900 304000±9000 886±23 982±91
B15(3) 471.0±9.4 149.8±3.0 375.3±3.6 1.416±0.011 70.72±0.55 318000±13000 309000±14000 897±37 980±100
B15(4) 488±10 167.1±3.4 367.4±4.3 1.400±0.015 64.94±0.62 311000±16000 301000±17000 859±43 950±120
C1 C1(1) 1628±33 442.8±8.9 375.2±3.3 1.5121±0.0058 88.27±0.30 513000±35000 508000±35000 1570±160 1700±360
C1(2) 1582±32 431.3±8.6 374.6±1.4 1.5083±0.0048 87.83±0.25 500000±22000 494000±23000 1509±98 1630±230
C1(3) 865±17 276.2±5.5 385.3±5.4 1.389±0.012 69.06±0.55 283000±11000 273000±12000 831±31 911±93
Table 2.4 continued from previous page
Date
Samp. Rep. 238u 23 2Th 6234Um f230 Th/ 238  23T/232 Unor   Date Corr. 6
23 4 23 4Ujnitia 8 Uiac
ID ID (ng/g) (ng/g) (%U) L (yrs BP) (%0) (%00)
(yrs)
Cl(4) 839t17 177.9±3.6 383.8i4.9- 1.3509i0.0089 101.12±0.62 256800±6800 250000±7700 777±20 825±58
C3 C3(2) 3557±71 -687±14 372.94±0.95 1.5058±0.0059 123.71±0.48 499000±26000 495000±26000 1510±110 1590±240
C3(3) 3460±69 659±13 375.8±2.5 1.5228±0.0045 126.96±0.27 569000i40000 566000±41000 1850±220 1960±470
C6 C6(1) 361.3±7.2 422.6±8.5 258.6±2.8 1.346±0.016 18.27±0.22 490000±92000 450000±100000 930+280 1360±920
CIO C10(1) 381.9±7.6 267.9±5.4 414.3±1.6 1.525±0.010 34.51±0.22 398000±19000 379000±23000 1208±79 1490±290
C10(2) 382.6±7.7 256.7±5.2 518.2±3.8 1.659±0.012 39.25±0.29 392000±21000 376000±23000 1497±100 1830±340
C1O(3) 379.9±7.6 218.5±4.4 515.5±1.9 1.6498±0.0099 45.53±0.27 384000±15000 370000±17000 1465±71 1740±250
C10(4) 404.4±8.1 271.5±5.5 443.1±1.5 1.552±0.011 36.70±0.27 380000±18000 362000±21000 1232±74 1510±270
C10(5) 431.3±8.6 277.9±5.6 427.2±1.6 1.530±0.011 37.71±0.27 379000±18000 361000±21000 1184±70 1440±250
C1O(6) 431.7±8.6 295.1±5.9 434.9±1.9 1.555±0.012 36.13±0.27 404000±23000 387000±25000 1296±94 1590±320
C10(7) 483±10 325.0±6.5 433.1±2.6 1.528±0.012 36.02±0.28 364000±17000 345000±20000 1147±67 1400±250
C10(8) 443.3±8.9 290.6±5.8 436.4±1.4 1.519±0.011 36.80±0.27 347000±14000 329000±17000 1104±55 1340±230
C11 C11(1) 1781±36 346.6±6.9 467.3±1.1 1.6278±0.0045 132.84±0.32 471000±14000 467000±14000 1747±71 1850±160
C11(2) 1746±35 331.6±6.7 472.91±0.98 1.6480±0.0063 137.73±0.54 514000±26000 510000±26000 2000±150- 2110±330
C13 C13(1) 1229±25 387.3±7.8 321.3±1.9 1.3946+0.0052 70.28±0.24 385000±11000 376000±12000 928±33 1015±98
C13(2) 1281±26 378.0±7.6 322.111.5 1.3899±0.0048 74.77±0.22 374500±9000 366000±10000 904±27 983±84
C13(3) 1214±24 322.1+6.5 341.3±2.6 1.4390±0.0098 86.11±0.58 424000±26000 417000±27000 1106186 1190±190
C13(4) 1180±24 334.8±6.7 336.1±3.6 1.434±0.011 80.24±0.59 429000±31000 421000±32000 11001100 1200±230
C14 C14(1) 2133±43 516±10 367.09±0.91 1.5908±0.0047 104.35±0.25
C14(2) 2136±43 533±11 361.7±1.2 1.5855±0.0049 100.87±0.26
2.10 Supplementary Materials
2.10.1 Methods of U-Th measurements on materials from 1996 piston
core
U-Th measurements on materials from the 1996 core were performed at the University of
Minnesota. For the first set of samples, processed in approximately 1999, mollusc shell
fragments were removed prior to chemical processing. For the second set processed in
2011, mollusc shell fragments were not comprehensively removed. Sample preparation
was identical to the procedures described in Section 2.3.2. Samples for the first set were
analyzed on a Finnigan Element I using methods described in Shen et al. (2002). Samples
for the second set were analyzed using a ThermoScientific Neptune multi-collector ICP-MS
in peak-jumping mode using methods described in Shen et al. (2012) and Cheng et al.
(2013a).
2.10.2 Methods of other datasets used to interpret U-Th data
Elemental concentration data. We measured 55 sediment samples for elemental con-
centrations of Ca, Mg, Sr, Al, Ti, P, V, Mn, and Fe. Samples of ~1-2 mg in weight were
dissolved and diluted in 3% HNO3 and then measured on an Agilent 7900 ICP-MS at
the MIT Center for Environmental and Health Sciences. Sample analyses were bracketed
by a multi-element standard. Two measurements each were also made on two certified
multi-element reference standards, PACS-2 and BCR-2. Data were corrected for blank
intensities. Uncertainties for each element were determined by calculating the average per-
cent difference between recommended values and measured values in PACS-2 and BCR-2,
and then applying the larger percent difference on measured sample values. For example,
the average percent difference between measured and recommended values in Mg (wt %)
was 6% for PACS-2 and 2% for BCR-2; thus, all Mg measurements for samples were as-
signed an uncertainty of.6% of the measured Mg value. Of the 55 samples analyzed, 48
corresponded to U-Th analyses.
Total Inorganic Carbon/Total Organic Carbon. We measured weight percentage
total carbon (TC) and weight-percentage total inorganic carbon (TIC) bycoulometry. For
the measurement of TC, we combusted samples at 1000°C using a UIC 5200 automated
furnace, and analyzed the resultant CO 2 by coulometry using a UIC 5014 coulometer.
Similarly, we measured TIC by acidifying samples with.10% H3 PO4 using an Automate
80
I' WMRPRWM, I _.M.Mrm"
acidification module and measured the resultant CO 2 by coulometry. We calculated weight
percentage total organic carbon (TOC) from TOC = TC-TIC; weight percentage TIC was
converted to percent calcite based on stoichiometry.
Color reflectance spectrophotometry. Color reflectance data were measured using
a Geotek multi-sensor automated core logger (MSCL-XYZ) on split core sections at Lac-
Core. To calculate sediment optical lightness, we took the sum total of light in the visible
region of the electromagnetic spectrum, between 400 and 700 nm, following Balsam et al.
(1999).
Mineralogy. In order to characterize the carbonate mineralogy of the drill core and to
discern possible mineralogical differences between endogenic and detrital CaCO3 , 25 sam-
ples were selected from intervals with variable CaCO 3 abundance (0-85%) from throughout
the core. In addition, 6 samples of carbonate bedrock from within the Junin drainage basin
on both the eastern and western sides of the lake were also analyzed. All samples were
pretreated with 35% H202 and IM NaOH to remove organic matter and biogenic silica,
respectively. Samples were then disaggregated with a solution of NaO 3 P combined with
ultrasonication, and then washed through 53 and 25 im sieves to isolate fractions >53
pm, 25-53 pm, and, <25 pm. These fractions were then scanned on a Phillips PW 1840
diffractometer at 45 kW and 35 mA. Each subsample was scanned twice, wide scans were
conducted at 0.60 (28) per minute from 4.0-70° (28) whereas narrow, more focused, scans
were performed at 0.3° per minute from 28.0-31.0°.
Ostracode assemblage analysis. A total of 22 sediment samples corresponding to
U-Th analyses were selected for ostracode analysis. One 0.25-g aliquot per sample was
removed for most ostracode analyses. Prior to sieving, samples were gently disaggregated
with three freeze/thaw cycles, since sediments were densely compacted. Then, samples
were wet-sieved using a 63 pm sieve. Ostracodes were extracted with fine brushes, identi-
fied and enumerated with respect to numbers per 0.25 g dry sediment. Analysis was done
using a Leica M80 stereo-microscope. Adult and juvenile intact and broken valves were
differentiated. Broken valves were counted if >50% was encountered and when identifi-
cation was still possible. Fossil ostracodes were identified down to family level following
procedures described in P6rez et al. (2010) and Karanovic (2012).
Additionally, we made a brief sediment description that included information of other
fauna, vegetation and minerals found in the observed sediment samples.
We calculated different ratios to facilitate taphonomy interpretations and for a better
understanding of processes such as remineralization and reworking in samples. The bro-
81
ken:intact (B:I) ratio was calculated for each sample to identify samples with relatively
high numbers of broken shells. Similarly the adult:juvenile (A:I) ratio was used to iden-
tify samples with a high number of adult valves, that could indicate transportation of the
lighter juvenile valves to deeper waters. The nektobenthic:benthic (NB:B) was calculated
to evaluate shifts in the relative abundance of bottom-swimming versus bottom-dwelling
individuals.
Shell coloration was taken into account as well when enumerating ostracode shells. We
were able to distinguish 7 different shell colorations: 1. Translucent, 2. White, 3. Partly light
grey, 4. Completely light grey, 5. Partly dark grey, 6. Completely dark grey, 7. Completely
black. Additionally, we used a Scanning Electron Microscope (SEM) TM3000 Hitachi
with BSE Detector II for taking pictures of uncoated specimens to facilitate ostracode
identification and to detect elements of ostracode shells using EDX analysis. All ostracode
analyses were conducted at the Institut fur Geosysteme und Bioindikation (IGeo) of TU-
Braunschweig.
Comparison of other data to U-Th data. All comparisons between different
datasets were carried out with MATLAB scripts written by CYC.
2.10.3 Failure to build isochrons
As mentioned in Section 5.3, determining dates from isochron plots failed (high MSWD
and low probabilities of fit). We used the Isoplot program by Ludwig (2012) to generate
isochron plots of various replicate analyses from bulk sample material, as well as analyses
from adjacent clean-dirty sample pairs. The failure to build isochrons is further evidence of
the existence of open system behavior occurring in these sediments. Figs. 2-18-2-28 show
the results of our isochron building attempts.
2.10.4 Calculation of 5211Uc
We assume that the detrital component has activity ratios essentially at secular equilibrium:
[234 U/ 23 8U]det = 1.0±0.5 and [232 Th/ 238 U]det = 1.2±0.6, following those used by Dutton
et al. (2017). First, we calculate the fraction of uranium that is detrital in each sample
(fdf2):
fdet= [2 32 Th]samp X [ 2 h1/U X [23 8U] samp
82
where brackets indicate activities, samp refers to the sample data, det refers to the detrital
component. We then convert the initial 2[ 34 U/ 23 8U] of the sample to 62 34Uiec with the
following formula:
234 U/ 238U]nit - [23 4 U/
238U]det
62 3 4U [iec = ( amp- - fdet Xf dt i)X1000
We propagate the uncertainties of the original measured [238 U]samp, [232Ujsamp, and detrital
activity ratios for the uncertainty of 8234Uiec.
2.10.5- Parameters and priors used for Bacon age-depth model
We used the following parameters and priors for our Bacon age-depth model: thickness
= 50 cm; acc.mean = 80 yr/cm; acc.shape = 2.0; mem.strength = 15; and mem.mean=
0.8. The age-depth model was generated by executing the following command:
Bacon(core = "PLJ_datesd234U_50", acc.mean = 80, acc.shape = 2, mem.mean
= 0.8, mem.strength= 15, thick = 50, ssize = 10000, burnin = 2000, suggest
= FALSE, depths.file= TRUE, yr.max = 800000, MaxYr=781000, d.max = 8800);
The Baconvergence test was run and yielded a Gelman and Rubin Reduction Factor of
1.031463, which fell below the 1.05 safety threshold and indicates robust mixing of Markov
Chain Monte Carlo iterations.
Trachsel and Telford (2017) tested Bacon and showed that the thickness parameter
(the segment length) had an unpredictable effect on the size of the error envelope. They
also found that the impact of different values for thickness was dependent on acc. shape,
the accumulation shape prior. As Blaauw and Christen (2011) did not explicitly make
any recommendations for how to choose an appropriate value for thickness, Trachsel and
Telford (2017) suggested that the length be shorter than the mean distance between dated
intervals and to choose a value that allowed for faster model convergence.
As can be seen from Fig. 2-15, the chronological constraints of the PLJ-1 splice are not
equally spaced. The average-distance between radiocarbon data in the upper 20 m of the
core is 24 cm; when considering all radiometric data (radiocarbon and U-Th), the average
spacing is 72 cm.
Thus, we carried out a comparison of age-depth models generated using different lengths
83
for the thickness parameter, keeping the accumulation shape prior constant. Fig. 2-17
shows that while the difference in the mean of the age-depth models does not vary by more
than 8 kyr at any point in the record, there are 5-30 kyr differences in the width of the
error envelope.
Unfortunately, there is no rule of thumb or other external information that can help us
determine which length for the thickness parameter is most appropriate. At this point,
the decisions for how to generate the age-depth model are, regrettably, more of an art.
Thus, we collectively settled on using 50 cm as the length for the thickness parameter,
for no other reason other than it seeming "reasonable." We encourage others to use the
chronological constraints generated in this study and others at Lake Junin to create better
age-depth models with improved estimates of uncertainties.
2.10.6 Determination of relative paleointensity tie points
As paleomagnetic reconstructions perform best within intervals of high lithic flux-(glacial
periods in Lake Junin), the U-Th ages in Fig. 2-15A were first used to seat the paleomag-
netic record during the carbonate-dominated interglacial sediments. Following interglacial
anchoring, the Lake Junin normalized remanence record was compared to a regional RPI
stack from the North Atlantic (Xuan et al., 2016) and the global paleointensity inversion
PADM2M of Ziegler et al. (2011). The generally good agreement between these records
and the Lake Junin normalized intensity record allowed additional tie points to be identi-
fied that improved the correlation between the PLJ-1 record and the two well-dated RPI
targets. A somewhat arbitrary 10 ka uncertainty (2-a) was prescribed to account for alias-
ing during the tuning process and the chronological uncertainty of the target stacks (see
Hatfield et al., in preparation for further details). In general, the RPI picks center within
the 2-a error envelope of the PLJ-1 Bacon-derived age model, however, picks falling off
the mean age-depth model likely infer variations in sedimentation rate between glacial-
interglacial sediments (see Woods et al., 2019, for examples within the radiocarbon era)
that might be expected in heterogeneous sedimentary environments but are not captured
in the interglacial-only U-Th age-depth model (see Hatfield et al., in prep).
84
il
2200 I I.I I
Sample C10
2000
1800
1600
LO1400
1200
380 400 420 440 460 480 50 0
Uranium concentration (ng/g)
1400 Sample Cl 3
1300
1200
(l1100
1000
900
1160 1180 1200 1220 1240 1260 1280 1300 1320
Uranium concentration (ng/g)
2450
I-- 2400 -
L0 2350 -
I I
2300 - I T
2250
Sample L7
I I
2200 I 
6200 6400 6600 6800 7000 72 00 7400 7600
Uranium concentration (ng/g)
Figure 2-16: Comparison of uranium concentration and 832 4 Uec for three samples of the CMC
facies: C10 (top), C13 (middle), and L7 (bottom). Samples C10 and C13 show some evidence of
the inverse relationship between uranium concentration and2 4 Uc, as would be predicted if there
was preferential loss of 23 4U (Robinson et al., 2006). However, L7-the sample with the highest
mean uranium concentrations out of any sample analyzed-shows atrend that is more consistent
with the expectations of preferential 2 4U gain. All errors are at the 2-a level.
85
Comparison of dfferent ickns parmtrs with tick =30 m
150
10
en0Ma 95 a onfdne-a e
ao
30 CM mefang of X cm) - (ean of 30 cm))
au-Th
0 10 20 30 40 50 60 70 80 90
Depth (CCLF)
Figure 2-17: Comparison of the size of the 95% confidence range and mean of Bacon age-depth
models run with varying lengths of the thickness parameter. Middle and bottom plots compare
the age-depth models to the results of the model where thickness = 30 cm. The top plot shows
the size of the 95% confidence range for the age-depth model where thickness = 30 cm.
86
MIS
data-point eror eflipses are 2
2.30
2.28
2;28
2.24
2.20
2.18
UAge 110.8244670t,000032 ka
2.16 inIbaal "UPU = 4.18±0.29
MSWD 178, probablity .O000
2.14
0.036 0.038 0.040 0.042 0.044 0.046
232Th/2=U
Figure 2-18: Isochron plot of analyses from sample M15.
87
L7
data-ponte noren lpsesare 2
1.88
1.84
O
1.80 0
1.76
nf /UAge  = 318.66703*t.00052 ka
initial MUPU a 5.29064555514899J*0
MSWD -128, probabuity -0000
1.72
0.0 18 0.019 0.020 0.021
232Th/2asU
Figure 2-19: Isochron plot of analyses from sample L7.
88
data-point error ellipses are 2n
1.70
1.68
Fq 1.66
1.64
MoT/ Age a 216 +460 -210 ka
IniW 2al gpU = 5 +9.4 -2.7 f
MSWD - 1.6, probabilty - 0J21
1.82 L
0.184 0.186 0.188 0.190 0.192
232Th/23sU
Figure 2-20: Isochron plot of analyses from sample L1.
89
K16
data-point error ellpses are 2a
1.7 (D
1.6
0
1.5 0
ratu Age =17.330398tO.OOOOS ka
inwalt 2UmU w 2.99398748895881*0
MSWD = 49, probabliy = 0,000
1.4
0.17 0.18 0.19 2 380.20 0.21 0.22M ThP U
Figure 2-21: Isochron plot of analyses from sample K16.
90
data-point error elipses are 2i
2.18
2.16 0
2.14
C 2.12
oo~
2.10
2.08
'hU Age -290.97382 ±0.00075 ka
2.06f InitialzUP% 4.8 i.3
MSD=797, probabtit - 0,000
2.04
0.07 0.08 0.09 0.10 0.11232Th/238U
Figure 2-22: Isochron plot of analyses from samples J5 and J6.
91
data-point erroempse are 2n
1.57
1.55
1.53
1.51
Th/U Age= 399+360 -160 k
Inii UPMU = 2.6 +2.3 -0.66
MSWD = 0.23, probability 0.80
1.49
0.19 0.20 0.21 0.22 0.23 0.24
232Th/2aeU
Figure 2-23: Isochron plot of analyses from sample H7.
92
-M
data-pointenorellips are 2n
1.72
1.68 0
1.64
I~K
M 1.60 ~KK ~>
S 1.58 /
1.52 0~0 mThAU Age = 44.33223 ±0.00017 ka
1.48 Iniial 4tJP0 - 1.54371735525610
MSWD = 10.5. probability 0.000
1.44
0.22 0.24 0.26 0.28 0.30
232Th/23U
Figure 2-24: Isochror plot of analyses from sample H6.
93
G13, 614
data-ponterro~hpm ar2o
1.65
1.63 fj~. 2
1,61
1.59[
MTh/UA ge 468.7988052±0.0000044 ka
Initial mUi/2Uz - 2.1161575414286 0
MSWD - 3. 1, probabflty - 0.014
1.57
0.0 1 0.03 2 0.05 0.07 0.093aTh/2ssU
Figure 2-25: Isochron plot of analyses from samples G13 and G14.
94
G7,G8
data-point error ellipses are 2a
2.1
2.0
1.9
0
1.8
mThAJ Age = 64.761899:0.000017 ke
Inigal 0U/mU -3.10 *0.22
MSWD = 32, probabilty=  0000
1.7
0.6 0.7 0.8 0.9 1.0
232Th/2wU
Figure 2-26: Isochron plot of analyses from samples G7 and G8.
95
G6
data-point eror eUipses are 2c
2.6
2.4
2.2
2.0
. **ThUAe = 203.340-160 ka
inIa asypay a 2.6 +3 -2.1
MSWD = 1.01, probabifty 0,40
1.8
0.5 - 0.6 0.7 0.8 0.9
2=2Th/2asU
Figure 2-27: Isochron plot of analyses from samples G6.
96
Ift, 4
F15
data-pointerrorellipsesare2.
0.382
0.378 0
0.374
0.370
0.368
0.362 00 ThU Agea  8.411774570.0O00012 ka
Inhaa UpU - 4.03 t0.1 2
MSWD = 4.3, probablity - 0.0144
0.358
0.032 0.034 0.036 0.038 0.040 0.042
232Th1238U
Figure 2-28: Isochron plot of analyses from samples F15.
97
F9
data-point eorelupae are 2n
0.080
0.078 0
0.076
0.074
2*hJAge 2.0212775968t0.0000000047 ke
2* MUU - 3.374 10.067
D=5.9, probability =0.003
0.072
0.010 0.012 0.014 0.018 0.018
232Th/23SU
Figure 2-29: Isochron plot of analyses from samples F9.
98
Chapter 3
U-Th dating of tufas from Agua
Caliente I, Laguna de Tara and
Salar de Loyoques, northern Chile
3.1 Introduction
The South American summer monsoon (SASM) is the main source of precipitation for
much of South America, impacting globally significant ecosystems ranging in diversity
between the world's largest tropical rainforest in the Amazon basin to the world's driest
desert in the Atacama (Nogu6s-Paegle et al., 2002; Jones et al., 2012). Future SASM
changes are of particular concern in the Central Andes, where anomalous decreases in
streamflow and water availability over the past three decades have rendered both human
and ecological communities highly vulnerable to future anticipated climate change (Messerli
et aL, 1997; Aravena et al., 1999; Vicuia et al., 2012; Magrin et al., 2014; Morales et al.,
2015). Despite societal significance, there is substantial uncertainty and poor agreement in
elimatefmodel projections of future precipitation patterns over South Arerica (Marengo
et al., 2012; Rowell, 2012; Knutti and Sedli6ek, 2013; Shepherd, 2014). The short and
sparse instrumental record and disagreement among reanalysis products make even more
recent changes in the magnitude and spatial extent of SASM ambiguous (Grimm, 2011;
Silva and Kousky, 2012).
To improve forecasts of future hydrological change in monsoonal regions, we must
99
examine natural archives that record changes in precipitation during Earth's varied climatic
past. Reconstructions of monsoon responses to past forcings and boundary conditions have
been produced from compilations of pollen and lake level data (Qin et al., 1998; Kolifeld and
Harrison, 2000; Yu et al., 2001; Bartlein et al., 2011) and used for data-based tests of climate
model performance. Such proxy-model comparisons have revealed that models consistently
underestimate past monsoon variations, but proxy data representation is heavily skewed
towards the low- to mid-latitude regions of the Northern Hemisphere (Joussaume et al.,
1999; Coe and Harrison, 2002; Braconnot et al., 2007, 2012; Roehrig et al., 2013; Perez-
Sanz et aL, 2014). Coverage in South'America in both pollen and lake level compilations
is much more sparse, making it impossible to determine whether models are adequately
representing the magnitude and spatial extent of the SASM under past climate conditions
(Figure 3-10) (Qin et al., 1998; Kohfeld and Harrison, 2000).
Furthermore, amongst the paleoclimate data that do exist in South America, there is lit-
tle consensus on how precipitation responded to climate perturbations over the Pleistocene.
In the Bolivian Altiplano of the Central Andes, sediment core records from the Titicaca-
Uyuni basin suggest that lake levels mainly follow insolation and glacial-interglacial varia-
tions, with the wettest conditions occurring during austral summer insolation maxima and
glacial periods (?). However, studies on preserved paleoshoreline deposits from the same
basin suggest that North Atlantic cooling events are the primary drivers of wet conditions
rather than insolation changes (Placzek et al., 2006b, 2013; Blard et al., 2011).
In this paper, we present new U-Th dating constraints on lake level variations from three
small (<40 km 2), high-altitude closed-basin paleolakes on the Altiplano-Puna plateau of
the Central Andes (23°S, 67°W; 4200-4300 meters above sea level). Because this area
experiences a strong seasonal cycle, receiving 50 to 90% of its annual precipitation amount
during austral summer (December-January-February, DJF; Nishizawa and Tanaka, 1983;
Gandu and Silva Dias, 1998; Garreaud et al., 2003; Vuille and Keimig, 2004; Wade, 2014),
the Altiplano-Puna region is ideally suited to represent a pure response to changes in
the SASM (Figure 3-1A). In each of these basins, evidence for previous intervals of much
wetter conditions is evident in spectacularly preserved paleoshorelines, which even a casual
observer can- identify via satellite imagery. Many of these paleoshorelines are encrusted
with carbonate "tufa" deposits, which are essentially fossilized calcareous remains of algal
reefs. Because modern analogues of tufa-forming organisms grow within the photic zone,
the location of these tufa deposits approximates the elevation of past lake levels. Earlier
radiocarbon ( 4C) dating efforts to find the ages of these deposits have been limited due to
100
I
a lack of terrestrial organic matter and large uncertainties in the reservoir effect in these
lake basins. Thankfully, due to high U concentrations and low detrital Th content, we
are able to apply U-Th dating techniques to these materials. By U-Th dating these tufas
and other lacustrine carbonates, we can pair an age to a quantitative constraint for the
magnitude of lake area expansion calculated from differential GPS (dGPS) measurements
of the elevations of shoreline features.
These data are the first-results from a planned north-south transect of six lake basins
spanning the subtropics of the Central Andes (20-33°S; Figure 3-1B) aimed at filling a key
geographical gap between existing lake level chronologies in the tropics and mid-latitudes,
as well as providing valuable constraints on past spatial variations in SASM extent. The
paper is organized as follows: In Section 3.2, we summarize current knowledge on the mod-
ern climate of the Altiplano-Puna plateau. In Section 3.3, we describe previous work that
demonstrates past linkages between local summer insolation and North Atlantic cooling
events to hydrological changes in South America and the Central Andes. We discuss the
setting of the lake basins in our study in Section 3.4 and describe our methods for U-Th
dating and mapping ancient shorelines in Section 3.5. The results and implications of
our findings are discussed in Sections 3.5-3.7. Finally, we outline the future direction of
this research (for instance, a water balance model to provide quantitative constraints on
precipitation and evaporation changes) and other possible avenues in Section 3.8.
3.2 Modern climate-of the Altiplano-Puna plateau
The Altiplano and Puna plateaus of the Central Andes are arid to semi-arid internally-
drained basins situated between cordilleras to the east and west. Together, they form the
second highest continental plateau in the world (average elevation 3700 m), stretching for
1800 km and varying between 350 and 400 km in width (Isaks, 1988; Allmendinger et al.,
1997; Kay and Coira, 2009), Mean annual temperatures from the few available weather
stations on the plateau average to9 C after elevational differences are normalized (Blard
et al., 2011). Based on modern 6180 of rainfall data from the IAEA-GNIP database,
average-5180 values of precipitation on the plateau are -12.1 ± 6.1%o VSMOW (-41.7±
-24.0 VPDB; IAEA/WMO, 2015). Despite the overall aridity of the region (<200 mm yr-1
precipitation in the south and ~1000 mm yr-1 in the north; Garreaud et al., 2003; Vuille
and Keimig, 2004), the plateau is the main source of water for the neighboring hyper-arid
Atacama Desert resting at the base of the western flank of the Andes. Evaporation rates
101
decrease while precipitation rates increase from west to east across the Andes such that open
water bodies are generally only observed at elevations greater than 4000 masl (Magaritz
et al., 1989; Vuille and Baumgartner, 1993; Aravena et al., 1999). This observed west-to-
east gradient in rainfall across the Central Andes and within the plateau is consistent with
an eastern continental moisture source (Garreaud et al., 2003). Moist, cool air from the
Pacific ocean rarely.reaches the plateau due to coastal topography and a thermal inversion
layer over the subtropical southeast Pacific caused by large-scale subsidence (Rutllant and
Ulriksen, 1979).
As mentioned earlier, rainfall over the Altiplano-Puna plateau is highly seasonal and
heavily influenced by the behavior of the SASM. Similar to other classic monsoon systems
70°W 65°W
Trade Winds Targeted
Lake Systems
sSalar del Huasco
2Laguna de Tara
Agua Caliente I
S& Salar de
Loyoques
Lagunas Miscanti
&Mirhiques
*Lagunas Bravas
*Laguna Bebedero
Westerlies Altitude (m)
£4 2000-2500
2500-3000
3000-3500
______________________M E 3500-4000
S4000-4250
100 150 200 250 300 +350 4250-4750
mean DJF precipitation 0 200 km +4750
(mm/month) 65°w
Figure 3-1: [A] Austral summer (DJF) precipitation in South America. The box shows
the region highlighted in panel B, which lies on the western and poleward edge of monsoon
rains. Data are from the CPC Merged Analysis of Precipitation (CMAP) from 1979-2015.
[B] Sites involved in a planned meridional transect of closed-basin lakes, and other lakes
mentioned in the text.
102
(e.g., the Indian summer monsoon), large-scale upper tropospheric circulation of the SASM
is driven by seasonally-varying differential sensible heating of land and ocean that then
triggers moist atmospheric processes which redistribute energy via latent heat transfer
(Zhou and Lau, 1998; Vera et al., 2006; Marengo et al., 2012). Throughout most of the year,
dry air from the subtropical Pacific is brought over the plateau by prevailing subtropical
westerly winds, resulting in very low precipitation rates. During austral winter, the SASM
is at its weakest, and South America receives most of its rainfall north of the equator in line
with the location of the local Intertropical Convergence Zone (ITCZ), centered at ~5°N.
While the local ITCZ is north of the equator, the interior of the continent undergoes its
dry season.
During austral summer at the peak of the SASM, a southward shift of the ITCZ
strengthens the easterly trade winds that advect moisture from the tropical Atlantic
ocean onto the continent, causing heavy convective precipitation in the southern Ama-
zon basin and northern Argentina (Figure 3-1A). These summertime convective thunder-
storms involving cumulonimbus clouds (Houze, 1997) release latent heat (condensational
heating) over the Amazon basin that causes the formation of the Bolivian High, an upper-
troposphere high-pressure cell (Lenters and Cook, 1997; Garreaud et al., 2003; Garreaud
et al., 2009). This warm upper-level anti-cyclonic circulation cell channels mid- and upper-
level easterly winds over the Central Andes, which provide the moist continental lowland
air necessary for the generation of deep convection on the plateau (Garreaud et al., 2003;
Vuille and Keimig, 2004; Falvey and Garreaud, 2005). Deep, moist convection is then
triggered by destabilization of the local lower troposphere via solar-driven sensible heat-
ing of the plateau surface (Garreaud, 1999; Vuille, 1999), releasing moisture and latent
heat as afternoon and early evening rainfall (Fuenzalida and Rutllant, 1987; Garreaud and
Wallace, 1997).
The Bolivian High is also associated with the formation of a low-pressure system cen-
tered over the Gran Chaco region of Argentina (Chaco Low, 25°; Seluchi et al., 2003),
which steers low-level (below ~1500 rn altitude) easterly winds flowing over the Amazon
southwards along the eastern slope of the Andes, funneling moisture from the Amazon
basin into the subtropics (Saulo et al., 2000; Marengo, 2004; Garreaud et al., 2009). En-
hanced specific humidity in the lower troposphere of these subtropical continental lowlands
appears correlative with wet conditions in the southern part of the Altiplano-Puna plateau,
but a causal link between these two conditions, and why the neighboring lowlands do not
matter for conditions on the northern Altiplano-Puna plateau, is unclear (Garreaud et al.,
103
2003; Vuille and Keimig, 2004).
While interannual and decadal-scale fluctuations in precipitation over the Amazon
basin are well-studied, such variations are <15% of the region's annual mean rainfall
amount (Garreaud et al., 2009). In contrast, year-to-year precipitation variations over
the Altiplano-Puna plateau are very strong, alternating between severe drought and very
wet summer conditions (e.g., 11 mm to 277mm in consecutive rainy seasons; Garreaud
and Aceituno, 2001; Garreaud et al,, 2003). Interannual and decadal-scale precipitation
on the plateau shows strong connections with Pacific sea surface temperature (SST) pat-
terns, with a warm eastern equatorial Pacific (El Nio-like conditions) corresponding to
drier conditions (e.g., Lenters and Cook, 1997; Vuille, 1999; Vuille et al., 2000; Garreaud
et al., 2003). During El Nio years, the anomalously warm eastern equatorial Pacific Ocean
causes a warming of the troposphere above, causing large-scale, upper-level zonal flow to
take on a more westerly direction. This strengthening of westerly winds across the Central
Andes transports dry air from the western slopes of the Andes, thus bringing about a dry
spell (Garreaud and Aceituno, 2001; Garreaud et al., 2003; Vuille and Keimig, 2004). An
opposite response occurs during most La Nifia episodes: anomalously cool eastern Pacific
SSTs cool the tropical troposphere and decrease meridional temperature gradients, result-
ing in a poleward shift of weakened westerly winds over the Central Andes, allowing greater
westward penetration of moist air.
Interestingly, there is no evidence for a tropical Atlantic influence on austral summer
rainfall in the Altiplano-Puna plateau on modern interannual timescales (Vuille et al.,
2000). However, given the short and often incomplete instrumental record in the Central
Andes, it is possible that Atlantic teleconnections may only play a role in Central Andes
precipitation on longer timescales.
3.3 Previous Work on Past Changes in the SASM
Records of millennial climate variations from South America are comparatively scarce,
but those that do exist show clear climatic shifts contemporaneous with Greenland during
the last glaciation (e.g., Lowell et al., 1995; Arz et al., 1998; Behling et al., 2000; Lamy
et al., 2000; Peterson et al., 2000; Peterson and Haug, 2006). Oxygen isotope records from
various speleothems in South America (e.g., Cruz et al., 2005, 2007; Wang et al., 2004,
2006, and 2007; Mosblech et al.; 2012; Cheng et al., 2013b; Strikis et al.,. 2015) suggest
that SASM precipitation was influenced by not only changes in local summer insolation, but
104
NORM WAMMMIN
also abrupt millenial-scale North Atlantic cooling events (e.g., Bard et al., 2000) which had
a significant impact on the interhemispheric temperature difference (Shakun et al.,'2012).
These anomalously cool North Atlantic events are called Heinrich events, and are associated
with marine deposition of iceberg-derived coarse-grained sediments. The South American
speleothem records suggest that periods of SASM intensification coincide with Heinrich
events by showing more negative 6180 values during these periods, which is interpreted to
indicate increased rainout upstream. Although these 6180 speleothem records are robust
and possess excellent temporal resolution and precision in ages, the magnitude and spatial
pattern of these precipitation changes cannot be gleaned from such proxies.
Lake records from the Central Andes have already proven to be a powerful and necessary
complement to these South American speleothem records. The shoreline-based records
from the Titicaca-Uyuni lake basin in Bolivia suggest that lake levels were primarily driven
by Heinrich events (Placzek et al., 2006b, 2013; Blard et al., 2011), and corresponding
hydrologic modeling suggests that a ~2-3 factor increase in precipitation relative to present
rates was necessary to maintain the lake at its highstand, called the Tauca phase, during
Heinrich Event 1 (H1; ~18-14.6 ka). Increased Central Andes rainfall associated with
HI is also observed in paleowetland deposits showing elevated groundwater tables (Quade
et al., 2008), fluvial terraces and pack-rat middens showing increased stream discharge
(Latorre et al., 2006; Nester et al., 2007; Gayo et al., 2012), and glacial moraines showing
substantial ice cover expansion (Smith and Delorme, 2010). In a sediment record from
Laguna Miscanti (Figure 3-1B) spanning the last -22 ka (Grosjean et al., 2001), two
periods of higher lake levels during the last deglaciation are inferred from aquatic pollen
assemblages and sediment lithology. Unfortunately, large 4 C reservoir effects for aquatic
organic materials of 2200-4000 years greatly limited the chronology of this record.
In this study, we aim to not only lay the foundation for quantifying past precipitation
changes, but also test the two contradictory interpretations of the Titicaca-Uyuni lake
system. The past extent of lake stages prior to >25 ka in this system is disputed due
to complex basin geometry and uncertainty over past sill elevations (Placzek et al., 2013;
?). By examining these pluvial events in smaller, simpler paleolake basins, we can assess
how consistent our reconstructed lake level histories are with each interpretation of the
Titicaca-Uyuni lake level record.
105
3.4 Study Area
Our study area. focuses on three neighboring lake basins in northern Chile, located south
of Salar de Uyuni (Figure 3-1B): Agua Caliente I (23.13°S, 67.41°W), Salar de Loyoques
(23.25°S, 67.29°W), and Laguna de Tara (23.04 0S, 67.28°W) (Figure 3-2). The landscape
is dominated by large volumes (>10,000 km3 ) of primarily silicic ignimbrites extruded
between 10-1 Ma (Quade et al., 2015). Mean annual temperatures of 0°C and annual
precipitation rates of 150-180 mm/yr make this area a very cold and dry place (Direcci6n
General de Aguas, 1987). At present, freshwater springs breaching the surface a few tens
of meters above the modern lake or salar feed these lakes; surface flow of water is confined
to such spring discharges. Agua Caliente I is also fed by thermal springs in the southwest
part of the basin, as is implied by its namesake. An inactive stream channel located at
the foot of converging alluvial fans from opposite sides appears to link the Agua Caliente
I basin with Salar de Loyoques, but no such surficial linkages exist with Laguna de Tara,
its neighbor to the north.
3.5 Materials and Methods
3.5.1 Field sampling and shoreline mapping
In order to obtain constraints on past lake water elevations, we identified, sampled, and
recorded the present-day elevation of sequences of tufa formations and lacustrine carbonate
deposits that provide information on paleowater depths. When combined with paleoshore-
line observations and geochronologic tools such as 14 C and U-Th dating, this strategy is
successful in reconstructing lake level histories in both large and small lake basins (e.g.,
Oviatt et al., 1994; Benson, 1994, 1995; Placzek et al., 2006b). When sampling, we focused
primarily on deposits associated with shoreline features, but we also sampled material from
road cut and stream channel exposures. When the entire thickness of a unit could not be
sampled, we took samples from the top (outer part) and bottom (inner part) of deposits to
capture the entire duration of lake episodes. Our sampling approach also took advantage of
several cross-cutting relationships and disconformities to find the relative timing of distinct
depositional events.
We collected precise location and elevation data of shoreline features and samples using
a Trimble Geo 7x handheld device, a high-accuracy Global Navigation Satellite System
(GNSS) receiver. After post-processing with proprietary software, we are able to achieve
106
horizontal and vertical location measurements of submeter accuracy (Section 3.9.1 and
Figure 3-12). For shoreline features, we measured the elevations of the crests of gravel
barriers, the base of incised alluvial fan scarps, and the tops and bottoms of abrasional
platforms, following methods described by Chen & Maloof (in revision). We also drove
through the stream channel connecting the Agua Caliente I and Salar de Loyoques basins,
taking regular dGPS elevation measurements of the channel bottom (thalweg) to estimate
the potential overflow (sill) elevation.
The locations and elevations of the highest shoreline features in each basin were then
Figure 3-2: Annotated satellite imagery of the three lake basins examined in our study.
Dashed.l ines serve as a guide to the reader to identify the modern extent of lakes and salars
within each lake basin.
107
used for paleolake area calculations, which were determined with geographic information
systems (GIS), satellite irnagery, and a digital elevation map (DEM) (Section 3.9.2).
3.5.2 U-Th dating of shoreline tufas and other lacustrine deposits
Tufa and other lake carbonate samples were slabbed along the axis of primary growth or
deposition and cleaned with ultrapure water prior to sampling. Using a vertical milling
machine with a tungsten carbide-tipped drill bit, we drilled carbonate powders weighing
2-10 mg from these slabbed surfaces, targeting areas with primary textures and the least
amount of siliciclastic detritus and post-depositional chemical or physical alteration. For
Bolivian Altiplano lake carbonates, Placzek et al. (2006a) found that white-colored car-
bonates tended to have lower initial Th content compared to material with faint pink'or
orange coloration; thus, we also preferentially sampled white-colored material.
Preparation of these aliquots for U-Th dating was then performed in a clean laboratory
at MIT. Aliquots were dissolved in HNO and spiked with a 229Th-233 U- 2363 U tracer in
Teflon beakers cleaned via a boiling-washing procedure with concentrated HNO 3 and HCl.
Next, following methods described by Edwards et al. (1987) and Shen et al. (2002), U
and Th were co-precipitated with ~4 mg of Fe oxyhydroxides and then separated using
BioRad AG1-X8 anion exchange resin (100-200 mesh, 0.5 mL column volume). The isotopic
compositions of the resulting U and Th fractions were then measured on either a Thermo
Scientific Neptune Plus multi-collector ICP-MS at Brown University or a Nu Plasma II-ES
multi-collector ICP-MS at MIT. At both locations, we introduced sample solutions through
a CETAC Aridus II desolvating nebulizer system coupled to a PFA nebulizer with an 100
RL/min uptake capillary. Each U sample analysis was bracketed by a 5 ng/g solution of
the CRM-112a standard (New Brunswick Laboratories). Each Th sample analysis was
bracketed by an in-house 2 29Th-23 Th-232Th standard in order to monitor mass bias and
variable SEM yield. 2% HNO3 solution blanks also bracketed each sample and standard
analysis to determine the background signal. One sample set of chemistry was analyzed at
MIT in peak-jumping mode on the ion counter following methods by Shen et al. (2002).
See Steponaitis et al. (2015) for more details on the mass spectrometry procedure used at
Brown University.
Total procedural blanks were included with each batch of 5-10 samples and were on
average 0.2 0.2 fg 230 Th, 1.2 ± 2.0 fg 234 U, 1.8 ± 0.9 pg 232 Th, and 8.7 ±8.2 pg 23 8U.
These averages do not include anomalous blanks, which had high values of 4.8 fg 230 Th,
108
I
11 fg 234 U, 47 pg 232 Th, and 107 pg 238 U. However, even with such large blanks, the
magnitude of the blank correction for ages we trust in our study was at most 5%. Data
with 230 Th/ 23 2Th ratios >100 ppm (atomic ratio) and <5% blank correction were deemed
acceptable. Replication is necessary and planned for samples where blank corrections were
higher than 5%. After making blank, background, tail, mass bias, and yield corrections,
we calculated U-Th ages using the 238 U half-life measured by Jaffey et al. (1971) and the
230 Th and 234 U half-lives determined by Cheng et al. (2013a). All ages were calculated
with an estimate of the detrital 230Th/ 232Th atomic ratio, (4.4 ± 2.2) x 10-6.
In the initial project stages prior to any U-Th dating of material, we pre-screened
samples to determine which sample materials would be most viable for U-Th dating. In
lacustrine settings, the initial 230 Th/ 232 Th is often high (e.g., Haase-Schramm et al., 2004;
Israelson et al., 1997), causing samples with even modest concentrations of 232 Th to have
large age corrections. Thus, samples with low 2 32Th concentration are considered more
favorable for U-Th dating. Sample powders of ~2 mg were dissolved in dilute HNO 3 , and
analyses of 2 38 U and 232Th concentration were performed on a VG PQ2+ quadrupole ICP-
MS at MIT. A 238 U/ 2 32Th ratio of 100 was our threshold for acceptable material for U-Th
dating.
3.5.3 Determination of mineralogy and stable isotope composition of car-
bonates
For mineralogical determination, powders weighing 1-3 mg were drilled from the same sam-
ple locations used for U-Th dating and non-destructively analyzed using attenuated total
reflectance Fourier transform infrared (ATR-FTIR) spectroscopy on a Perkin Elmer Spec-
trumOne at Harvard University. Some of these same powders were used to determine the
stable isotope composition of samples. The 6 13C and 6180 of carbonates were measured
at the University of Arizona (UofA) and at the Woods Hole Oceanographic Institution
(WHOI). At. UofA, measurements were performed on a KIEL-III automated carbonate
preparation device coupled to a Finnigan MAT 252 gas-ratio mass spectrometer. Pow-
dered samples of -0.1-0.2 mg were dissolved in dehydrated phosphoric acid at 70°C. The
2-o uncertainty of isotope ratio measurements, based on repeated measurements of car-
bonate standards NBS-18 and NBS-19, is± 0.22%o for 6180 values and± 0.16%0 for 5 13 C
values. At WHOI, stable isotope measurements were made on a Thermo Scientific MAT
253 mass spectrometer. Powdered samples of 3.5-5.5 mg were introduced into a custom-
109
made, automated acid reaction and gas purification line built for carbonate clumped isotope
analysis (see Supplementary Materials in Thornalley et al. 2015 for method details). We
repeatedly analyzed NBS-19 carbonate standards to determine the precision of the isotope
ratio measurements, which is ± 0.26%o for 6180 values and ± 0.12%o for 51 3C values at
2-o- uncertainty. Each reported measurement is an average of 3-4 replicate analyses of a
particular sample; thus, the measurements made at WHOI may be more representative of
the average 5180 and 5 13 C value of each drilled location in a sample.
As a preliminary step towards investigating the microfacies of these lacustrine carbonate
deposits, one large format (40 mm x 60 mm) thin section was prepared'at Spectrum
Petrographics (Vancouver, WA) and examined with an Olympus BX51 light polarizing
microscope.-
3.6 Results
Field work and sample collection occurred in 2009, 2010, and 2015. In Agua Caliente I,
samples span an elevational range of 0 to 25 in above the modern salar. In Salar de
Loyoques, the elevational range of samples is 0 to ~70 m above the modern salar. In
Laguna de Tara, samples are 10-30 m above the modern salar. See Table 3.1 for the
elevation of the modern salar or lake in each basin.
3.6.1 Observed tufa varieties and other lacustrine carbonates
Of the three lake basins in this study, Agua Caliente I contained the most unique types of
calcareous tufa and lacustrine carbonate. Figure 3-3 is a schematic representation of the
elevational relationship between paleoshoreline features, different tufa varieties, and other
Lake Modern Highest Shoreline Modern Ancient Factor
Basin Salar/Lake Elev. Above Area Area Increase
Name Elev. (m) Modern (m) (km2 ) (kin 2 )
Agua Caliente It 4219 34 6.6 26.9 4
Salar de Loyoquest 4182 70 10.7 201 19
Laguna de Tara 4322 30 36 179 5
Table 3.1: Table of modern and ancient lake areas for each basin. *Ancient area calculations are
for lake highstands. tCalculation of the ancient lake highstand is based on the assumption that
Agua Caliente I did not overflow into Salar de Loyoques. The present-day elevation of the sill is
below that of the highest paleoshoreline in Agua Caliente I. Themodern day lake area of Salar de
Loyoques includes the modern day lake area of Agua Caliente .
110
0,1011 OWN"
lake carbonate deposits in Agua Caliente I. We now describe the morphology and character
of each type of deposit observed in the field.
ENCRUSTING FLORET TUFA.-By volume', this porous, whitish-beige-colored tufa va-
riety is the most abundant lacustrine carbonate observed, covering many surfaces -5-28
m above the modern lake throughout the basin (~4223 and 4246 m elevation; Figure 3-
4A). These deposits vary in thickness from 3 to 20 cm and appear thickest on smooth,
hard substrates such as boulders on hills made of exposed volcanic bedrock (Figure 3-4B).
Frequently, multiple sequentially-formed encrustations separated by disconformities can be
found on large boulders, with cumulative thicknesses of up to 50 cm. Neighboring boulders
are often cemented together by this deposit.
Where coating boulders thickly, the encrustations can vary laterally in thickness by
several centimeters, forming broad, mound-like, domal buildups that tend to be thickest
towards the boulder tops. In cross-section, the tufa has the appearance of closely-packed
"floret"-like growths of uniform thickness radiating away from their nucleation surface (Fig-
ure 3-4C). Preservation of this growth structure varies basin wide and between sequentially-
deposited encrustations on a singular boulder.
The boulder encrustations are primarily found in the northern and eastern parts of the
basin where small bedrock hills are located. Interestingly, the deposits are most thickly
developed on the lakeward side of these hills, even when elevation data indicates that the
landward side of hills must also have been in contact with lake waters. In the southern part
of the basin where there are no such hillsides and the bathymetric slope is much shallower,
the encrusting floret tufa assumes a thinner, more flat-lying, tabular shape with more
uniformity in lateral thickness (Figure 3-4D). These tufa also tend to be more porous and
weathered, incorporating a larger amount of detrital sediment. These differences in porosity
and morphology within a single tufa variety may reflect different wave energy environments,
in which the more tabular-shaped tufas indicate higher-energy wave conditions (James and
Bourque, 1992), possibly due to larger fetch and a shallower wave run-up slope.
Although lateral discontinuities are present, the floret tufas do not exhibit any lami-
nations in their internal structure in any location. A thin section taken along the growth
axis of an encrusting floret tufa shows dark microbial peloids (microcrystalline carbon-
ate grains) forming radially-oriented branching structures that are surrounded by micrite
(microcrystalline calcium carbonate) containing tiny growth-oriented filaments and dark
lenticular tube-like cyanobacterial-algal microfossils (Figure 4-2D). Trapped detrital mate-
rials consisting of skeletal ostracod (benthic micro-crustaceans encased in a calcitic bivalved
111
=====
425 HONEY CALCITE-- ABRASION INFILTRATING.-
4250 PLATFORMS HONEYCALCITE IGNIMBRITIC
CAALICHHE  a LINCUISED COLLUVIUMILFNGRAVEGLRA IBNARFRIEIRLS TRBAEATCHIN SAGND 
o -c> SCARP
42456
THICK ENCRUSTING
-- 'FLORET" TUFA
SILL
4240 cowP S
ENCRUSTATIONS
C 423 - ~~ --- TRANSFORMED KAITE
23 - ,-
-CONE TUFA
4230 -
FLORET TUFA WITH
CARBONATE MUD CAP
4225
CARBONATE MUD FLAKES
- AND NODULES
4220 MODERN SALAR
4215
Figure -3-3:- Schematic representation of the elevational relationship amongst paleoshoreline features, tufa varieties,
other lake carbonate deposits, and overflow points (sills) in Agua Caliente I. The morphology of the tufa varieties are
simplified, and their complex internal structures are not illustrated. Colored bars on the left represent the eleational
range where each tufa and lake carbonate variety is observed: green = caliche; gray = floret tufa; orango = honey
calcite cement; yellow = transformed ikaite; blue/gray = floret tufa with carbonate cap; blue = carbonate mud flakes
and nodules; red = cone-shaped tufa.
Figure 3-4: Photographs of encrusting floret tufa and caliche observed in Agua Caliente I. [A]
Photo looking northeast from a tufa-encrusted bedrock hillside in the northern part of the basin.
By volume, most tufa in this basin is of the floret variety. [B] The floret tufa drapes over all
hillside surfaces,.including boulders. [C] A close-up, cross-sectional view of a thick deposit of floret
tufa coating a basalt boulder, showing a faint radial fabric. [D] The morphological'expression of
the floret tufa in the southern part of the basin, where the slope is more gently inclined and no
basalt boulders are present. Here, the floret tufa geometry appears more tabular. [E] An abrasion
platform carved into a hillside of exposed volcanic bedrock in the northern part of the basin. The
platform is -33 min length, measuring from the top of the platform (4247.2 ± 0.8 m) to the
bottom (4245.9 ± 0.8 in). Here, exposed caliche is observed at the base of the platform. [F] A
close-up photo of the caliche observed at the location shown in Panel E. The carbonate crust only
penetrates a few centimeters before reaching sand. Red pen is 15.5 cm in length.
113
shell) fragments and possibly carbonate fecal pellets are also observed in intact primary
pore spaces (Figure 4-2E).
CALICHE.-A matrix-supported conglomerate can be found as a calcified crust in de-
pressions between boulders of volcanic bedrock in flat or gently-sloping areas at the base
of abrasion platforms (Figure 3-4E) and at the tops of bedrock hills at ~4242-4246 m el-
evation. The carbonate matrix binding poorly sorted, sub-rounded volcaniclastic pebbles
and sand is fine-grained and patchy in color, from white to light brown (Figure 3-4F). The
carbonate is also concretionary, forming coatings around grains and clasts. Accumulations
of volcaniclastic sand and silt often cover the crust, which seems to be only a few centime-
ters thick. We hypothesize that these deposits are caliches (i.e., a soil carbonate formed by
meteoric waters) based on these observations; however, stable isotope analysis is needed
to confirm its origin.
CONE TUFAS.-Reef-like colonies of cone-shaped tufas that are -10-20 cmin height and
diameter are found along a narrow range of elevations (~4229-4234 m) in the basin. These
cone tufas have a fan-like internal growth fabric of concentric bands linked by radially-
oriented structures stemming from a narrow, mat-type holdfast. In some places, we'find
cones on top of floret tufa buildups, growing from holdfasts embedded in the uppermost
layer of the floret tufa encrustation (Figure 3-5A). Cone spacing is irregular: some groups of
cones appear to grow from holdfasts directly adjacent to one another to form an aggregate
cone composed of 4+ smaller individual entities, whereas other cones are distributed more
evenly across the substrate such that only their wide tops make contact. A gently undulated
surface results from numerous cone tufas joining at their tops. Large gaps between cones
have often been infiltrated with platey carbonate mud deposits.
In other locations, cone tufas are found on sandy substrate and do not form prominent
reef-like colonies. This observation may reflect preservation issues rather than real differ-
ences in the megastructure of the cone tufas. We also observe that the aspect ratio of cones
is variable. For instance, in the southernmost part of the basin where the bathymetric slope
is more shallow, the cones tend to'be shorter and wider (bowl-shaped). Like the tabular-
shaped floret tufas, shorter and wider cones may also be indicate that the southern part of
the basin was host to a higher-energy wave environment (James and Bourque, 1992). Due
to their external geometry, these cone tufas may also indicate that water roughness was
lower than the roughness of the preceding wave environment in encrusting floret tufa.
Conical tufas with clear biological textures are also described in the Uyuni basin of the
Bolivian Altiplano (Rouchy et al., 1996) andthe Miocene Ries Crater lake in Germany
114
MW11 pip 10P, .M WIPPIP M""T""Niq 11 11 RP
aggregate of *s
49 conesone tufa
44
encrus ting d et tu-
- lorete
7 u7a
r p 
Figure 3-5: (Caption on the top of next page.)
115
Figure 3-5: (Figure on previous page.) [A] Photo of cone-shaped tufas on top of two
distinct deposits of encrusting floret tufa. {B] Top-down view of a cone tufa, showing
concentric growth structure. [C] Lighter-colored carbonate iud flakes stand out against
darker volcaniclastic material, tracing a contour line of elevation. [D] The two flakes to
the right of the hammer have been flipped over, showing that some undersides of mud
flakes exhibit wiggly, possibly microbial textures. [E] Nodular forms of carbonate mud.
[F] Layers of platey carbonate in association with encrusting floret tufa. [G] A continuous
layer of carbonate mud capping floret tufa. Dashed circle highlights a few overturned pieces
showing that the floret tufa is directly underneath. [H] A closer view of the carbonate-
capped floret tufas, which look like mushrooms.
(Riding, 1979; Arp, 1995), but their external morphology and size differ from the cones
observed in our study area. Regardless, we interpret the cones to also be of algal origin
due to their structure.
PLATEY CARBONATE MUD FLAKES AND NODULES.-Broken plates of lighter-colored
lithified carbonate marl stand out against the darker backdrop of volcaniclastic materi-
als, and are often found tracing continuous elevational contour lines between 4218 and
4231 m elevation (Figure 3-5C). Sometimes, the undersides of these flakes exhibit a wiggly
embossment, possibly of microbial origin (Figure 3-5D). In other locations, the lithified
carbonate mud is more nodular in form (Figure 3-5E). It is unclear what conditions cause
one morphology to form over the other, but the carbonate nodules often contain dendritic
manganese oxide stains and tend not to incorporate much detrital material compared to
the platey flakes, suggesting that the nodule morphology may be a resuit of carbonate
replacement, rather than lithification, of muds and lake sediment. However, the nodu-
lar carbonates are rich with ostracod remains <1 mm in size, and may even be entirely
ostracod-supported, suggesting that primary material in these carbonates still exists, even
if they have been diagenetically altered.
Platey carbonates are also observed in association with other tufa varieties at elevations
up to 4242 m (Figures 3-5F). Clear sedimentological cross-cutting relationships indicate
that most, if not all, platey carbonates found in these areas were deposited at a time during
or after the formation of encrusting floret and cone tufas. This temporal relationship
is perhaps best demonstrated by the mushroom-like carbonates consisting of floret tufa
capped by lithified carbonate mud (Figure 3-5G & H). When slabbed to expose a cross-
sectional surface, these samples show-that the carbonate capping material fills gaps between
floret growths. Although this particular field relationship is apparent, we note that these
platey carbonate flakes and nodules are diachronous, with some deposits likely forming
116
more recentl'than thers, especially those at lower eevations closerto the modern lake.
We a1so found platey carbonate nodules in Salar de Loyoques.
TRANSFORMED IKA ITE.-Aggregates of elongate prismatic crystal blades of calcium
carbonate are found at many elevations in Agua Caliente I, from 4216 to 4242 m. Based on
their crystal habit, these deposits are interpreted as being pseudomorphs of the metastable
mineral ikaite, CaCO 3 .6H 20 (Swainson and Hammond, 2001). Modern naturally-occurring
ikaites are found mainly in marine environments (e.g., Buchardt et al., 1997, 2001) and
less commonly in continental lacustrine and spring settings (e.g., Bischoff et al., 1993b;
Oehlerih etal. 2013), but they are always observed in strict association withniear-freezing
watertemperatures.Whenwatersaewarred to temperatures above ^44C, ikaite crystals
rapidly decompose to calcite plus water within minutes to hours (Suess et al., 1982; Jansen
et al., 1987), producing a fragile, highly porous crystal mesh due to ~70% volume loss
during transformation (Shearman and Smith, 1985; Larsen, 1994). Thus, early diagenetic
precipitation of a carbonate cement is needed to preserve the original ikaite crystal habit
(Selleck et al., 2007). Transformed ikaites similar in morphology and setting to those
observed in our study area include the thinolites described in the Lahontan and Mono
Lake basins in the western United States, where they are interpreted to have formed at
or below the sediment-water interface (Dunn, 1953; Shearman et al., 1989; Council and
Bennett, 1993; Bischoff et al., 1993b, 1993a). Thus, the elevation of a transformed ikaite
represents a minimum constraint on the elevation of lake level at the time of its formation.
In Agua Caliente I, transformed ikaites manifest as aggregates of irregularly arranged
0.5- to 10-cm-long pyramidal crystals and are generally found filling spaces that may have
been eroded hollows within the floret tufa (Figure 3-6A & B). Carbonate-cemented detrital
grains fill many gaps between individual prisms. In many instances, transformed ikaites are
also found growing on top of and below deposits of platey carbonate. Thus, it is apparent
that ikaite formation occurred during or after the deposition of the floret tufas and platey
carbonates in these areas. However, like the platey carbonate mud flakes and nodules,
the formation of the transformed ikaites in this basin is not necessarily synchronous. We
also found one occurrence of a friable and highly porous transformed ikaite thinly buried
beneath sediment at 4216 m elevation in the modern salar.
Transformed ikaites were also found at Salar de Loyoques and Laguna de Tara. In
Salar de Loyoques, the pyramidal crystals tend to be better preserved and larger in diam-
eter, exhibiting square prismatic morphology in some instances (Figure 3-6C). Although
weathered on the outside, the insides of these transformed ikaites are very white and
117
Figure 3-6: Photographical field context of the ikaite pseudomorphs observed in Agua
Caliente I (A, B). Salar de Loyoques (C), and Laguna de Tara (D).
micro-crystalline. Interestingly, some transformed ikaites contain abundant ostracod re-
mains. We also observe small, <2-mm-sized transformed ikaites embedded in carbonate
nodules with sigmoidal crystal habit, exhibiting both square-prismatic and pyramidal faces.
In Laguna de Tara, transformed ikaite is best preserved between crevices on the undersides
of exposed bedrock overhangs (Figure 3-6D). Here, a secondary carbonate cement obscures
the characteristic prismatic crystal habit of ikaite (Figure 3-8A). A slabbed surface reveals
that the transformed ikaite crystals are white in color with few detrital grains, whereas
the carbonate cement is pink and full sand-sized detrital material. The pink color of the
cement likely comes from the'pink-colored ignimbrite unit in which the Laguna de Tara
basin sits. Transformed ikaites were the only variety of lacustrine carbonate collected from
Laguna de Tara.
HONEY CALCITE CEMENT.-The origin of this carbonate phase, observed only in Agua
Caliente I of the three basins in this study, is perhaps the most unique of all the encountered
varieties of lacustrine carbonate in our study area. The honey calcite is named so because
of its golden cloudy-translucent color. We describe it in more detail in Chapter 4.
118
3.6.2 U-Th dating of tufas and other lacustrine deposits
Our preliminary U-Th age dataset consists of 47 ages of carbonate deposits from Agua
Caliente I (35 out of 47), Salar de Loyoques (6 out of 47), and Laguna de Tara (7 out of
47) (Table 3.2 & 3.3). Spectra from ATR-FTIR spectroscopy indicate that all lacustrine
carbonates examined thus far from Agua Caliente I and Laguna de Tara are composed of
calcite; mineralogical analyses have not yet been run for samples from Salar de Loyoques.
High U concentrations ranging between 1 and 22 g/g (average 7.1 ± 5.6 g/g) and high
2 30Th/ 232Th ratios between 60 and 50,000 ppm (atomic ratio; average 3,100 ± 7,900 ppm)
allow us to date deposits in our study area at high, precision, with 2- uncertainties <200
years at 10-20 ka and <1000 years at 115 ka. These U concentrations are >2 times
greater than the U concentrations of lacustrine carbonates from paleolakes in the Bolivian
Altiplano (Placzek et al., 2006a).
Of these data, the most robust ages come from the honey calcite cement in Agua
Caliente I (Figure 3-7A-E). We refer the reader to Chapter 4 for more details.
Unfortunately, we are not able to reliably reproduce an age for the encrusted floret tufa
due to the pervasive infiltration of the honey calcite cement (Figure 4-2F). We obtained 15
U-Th ages from various spots on slabbed specimens of encrusting floret tufa that differ in
their proportions of honey calcite and tufa. Although we do not know the exact fractional
contribution of each phase in the sample powders, we can guess at their relative proportions
through a qualitative assessment of the color of the drilled location. The sample that best
isolates the floret tufa phase yields a U-Th age of 23,415 t 64 yr BP with a high U
concentration of ~15 g/g (Figure 3-7D; 4233.5 ± 0.8 m elevation). All other sample
powders are mixtures with greater proportions of the honey calcite cement and yield ages
that fall between this age and the 15.2-15.6 kyr BP age of the isolated honey calcite
cements. Based on this data, the -23.5 kyr BP age for the encrusted floret tufa at this
elevation'should be considered a minimum age.
We also obtained U-Th ages from a carbonate nodule capping a floret tufa, a trans-
formed ikaite, and pristine white carbonate within caliche from Agua Caliente I. However,
such ages have not been reliably replicated yet due to possible sample heterogeneities or
a low 2 30 Th/ 232Th ratio (-150 ppm atomic). The carbonate nodule has the highest U
concentration for any sample measured thus far in Agua Caliente I, at ~20 g/g. Two
preliminary ages from the caliche suggest that it is a more recently formed deposit (mid-
to late-Holocene).
119
Figure 3-7: [A] Aerial imagery of Agua Caliente I and Laguna Loyoques, showing location of carbonate
samples collected in January of 2015 and an approximate outline for the lake highstands, based on imagery
and field observations. Numbered sites indicate locations of samples featured in Panels B-G. [B] Field
photo of tufa-encrusted basalt boulders at Site #1. White box indicates the location of sample featured
in Panel D. [C] Field photo of road cut exposing lacustrine carbonate material at Site #2. White boxes
indicate location of sample featured in Panel E. [D] Cross-section of slabbed hand sample of tufa collected
at location featured in Panel B, showing two phases of carbonate growth/deposition and their respective
U-Th ages. [E] Hand sample of lacustrine carbonate (calcite) material and associated U-Th ages. [F] Field
photo of platey carbonate found at Site #3. White box is[ C], which is a cross-sectional view of the sample
with its U-Th age. [H] Aerial imagery of gravel barriers at Site #4. White box indicates location of [J],
which shows an outcrop of carbonate-cemented beach rock. White box indicates sampling location for [K],
where carbonate cements were U-Th dated. The isochron method for calculating U-Th ages will need to
be applied for the beach gravel cements in [K]. 120
.66 @0
Figure 3-8: [A] Transformed ikaites from Laguna de Tara, showing that the prismatic
crystal habits are obscured by a secondary cement. The remaining panels show U-Th ages
from various spots on slabbed transformed ikaites. The transformed ikaite crystals are
white, whereas the carbonate cements are pink due to incorporation of pink ignimbritic
detrital material. Each sample shown was collected from different elevations: [B] 4334 m,
[C] 4349 m,arid [D] 4352 m. Note that elevations were measured with a Garmin GPS,
not the TrimbleGeo 7x. Thee levation of the modern lake is -4322 m according to Google
Earth.
In Laguna de Tara, all U-Th ages are from three samples of transformed ikaite spanning
an elevational range of ~20 m (4334-4352 m elevation). Of all the tufa varieties in this
study, these samples have the lowest 23 0Th/ 232 Th ratios (60 to 550 ppm, atomic ratio).
The 230 Th/ 232 Th ratios of the pink cements were consistently 25-50% lower than adjacent
carbonate cements, consistent with the observation that these cements incorporate a signif-
icant amount of pink ignimbritic detrital material. For the two samples lowest in elevation,
the U-Th ages fall within the range of 15 and 16 kyr BP for both thetransformed ikaite
crystals'and the surrounding cements (Figure 3-8B, C). The highest sample yields an age
of around ~10.0-10.3 kyr BP for the transformed ikaite and -15.2 kyr BP for the cement
(Figure 3-8D).
121
With the available geochronological data, it is unclear if the cement and transformed
ikaites in the highest two samples formed simultaneously, or if the cement arrived shortly
after ikaite formation. If their formation was coincident, it may suggest-that carbonate
replacement of the original ikaite occurred while the crystals were still below the sediment-
water interface. The replicated ~10.0-10.2 kyr BP age for the transformed ikaite at lower
elevation suggests that there are multiple stages of ikaite formation in this basin. The U-Th
age for the cement in this sample is closely aligned with the age of the cements for the other
two'samples, suggesting that cement formation may have occurred simultaneously over the
20-in elevational range covered by these samples. More replication of ages is needed- to
determine the precise sequence of carbonate deposition in Laguna de Tara.
In Salar de Loyoques, we obtained U-Th ages from transformed ikaites (Figure 3-6C),
platey carbonate nodules (Figure 3-7F & G), and carbonate cements from beach gravels
(Figure 3-7H-K). The transformed ikaites are very clean, with high2 30Th/ 2 32Th ratios
(>10,000 ppm atomic). Similar to the carbonate nodule in Agua Caliente I, the carbonate
nodule dated in Salar de Loyoques has a high U concentration of -22 .g/g. Our preliminary
U-Th ages fall between 110 and 140 kyrs BP, but these data have not yet been reliably
replicated. More work is needed to determine if age discrepancies within samples are due
to true sample heterogeneities, incorrect assumptions for the initial detrital Th correction,
or problems with procedural blanks or cross contamination in chemistry. The age from the
beach gravel cement will be of particular interest, given that it was sampled from a gravel
barrier located ~70 m above the modern salar. Furthermore, the beach gravel cement is
one of the few lacustrine carbonate samples in our study that is directly associated with
a paleoshoreline feature, making its age a robust indicator of the timing of lake expansion
in this basin. We anticipate using the isochron method for obtaining a U-Th age of the
beach gravel cement material due to its low 230 Th/ 232Th ratio.
We briefly describe differences in the 6 2 34Uinitial values of lake carbonates between the
basins in Section 3.9.3.
3.6.3 Stable isotope composition of deposits
Table 3.4 lists the stable isotope compositions of samples from Agua Caliente I and Laguna
de Tara. In Agua Caliente I, the 6 13 C and 5180 of encrusting floret tufas, transformed
ikaites, and the honey calcite cement fall within a narrow range of values: +0 to +3%o
(VPDB) for 613C and +1.0 to +2.5%c (VPDB) for 180 (Figure 3-11). The platey carbonate
122
I
nodules are the only samples that yield negative > 13 C values (-3 to -1%o, VPDB). The stable
isotope composition of material that sampled both the honey calcite and encrusted floret
tufa phases generally fall within a 'mixing line' between two end member compositions,
with some nuances. The transformed ikaites from Laguna de Tara occupy an even narrower
range of values: +2.5 to +3.5%c (VPDB) for 513 C and +0.6 to +1.1%o (VPDB) for 5180.
We do not yet have information on the stable isotope composition of the carbonate in the
caliche or cone tufas from Agua Caliente I, nor for any samples from.Salar de Loyoques.
Based on the negative 6180 values for modern precipitation (IAEA/WMO, 2015) and
nearby groundwater (Rissmann et al., 2015), the positive 6180 values of the carbonates
analyzed thus far are an indication of lacustrine origins.
3.6.4 Paleoshoreline features and magnitude of lake area changes
We refer the reader to Section 3.9.4 for details regarding the paleoshorelines features ex-
amined in our study. In each basin, the elevations of the highest of these paleoshorelines
were used for paleolake area calculation. Table 3.1 lists our estimates for modern lake and
paleolake areas for our three lake basins (see Figure 3-14 for area outlines on a map). Agua
Caliente I and Laguna de Tara experienced ~4-5 factor increases in lake area associated
with 30-35 m elevational increases in lake level. The expansion observed in Salar de Loy-
oques was much greater: Based on the elevation of the highest gravel barrier identified in
the southern part of this basin via satellite imagery, Salar de Loyoques increased in area by
a factor of -19 relative to present, rising 70 m and merging with the Agua Caliente I lake
basin. Of further note: Paleoshoreline evidence of this large paleolake in Salar de Loyoques
only exists on the eastern margin of the basin. The contour of elevation representing this
lake overlaps alluvial fafis showing no evidence of lake incision all along the southern and
western parts of the lake (Figure 3-14). Thus, accumulation of these alluvial fans must
have occurred after the lake in Salar de Loyoques regressed to lower elevations.
3.7 Discussion
3.7.1 Relative temporal constraints on lake carbonate and paleoshoreline
formation in Agua Caliente I
As of yet, there are not enough U-Th age constraints to determine the exact sequence of
carbonate and paleoshoreline formation inAgua Caliente I. However, some relative con-
123
straints can be made purely from field observations. For example, at least two generations
of encrusting floret tufa must be older than the cone tufas (Figure 3-5A), and the honey
calcite, which is likely synchronous throughout the basin, must be younger than both of
these tufa varieties (Figures 4-2 & 4-1). Our U-Th ages confirm this relative relationship
between the honey calcite and the floret tufa (Section 3.6.2). All ikaites observed thus
far formed after the encrusting floret tufas (Figure 3-6A & B), and the nodular carbonate
caps are clearly younger than their floret tufa counterparts (Figure 3-5G & H). However,
it is clear that not all ikaites in the basin are the same age: The honey calcite cement
can be found in association with some transformed ikaite formations and not others. Like-
wise, the various carbonate mud flakes and nodules throughout the basin are most likely
diachronous.
In relation to various paleoshorelines, reworked floret tufas with carbonate caps were
found on top of a gravel barrier in the northeast part of the basin (Section 3.9.4). When
linking paleoshoreline features with specific tufa or carbonate deposits, we must remember
that the abrasional platforms and alluvial fan incisions represent the cumulative occupation
of a lake at that elevation; the lake could have risen to the elevation of those features
multiple times in the past. The gravel barriers most likely represent regressive or fluctuating
lake levels, since constructional transgressive features are rarely preserved due to reworking
of sediments upon inundation. One observation of interest is that the cone-shaped tufas
occupy the same elevational range as the lower sets of gravel barriers and the lower alluvial
fan incision with residual salt deposits (Figure 3-3 & Section 3.9.4). It is possible that
these features are synchronous.
Only further careful U-Th dating of samples with clear stratigraphical relationships
and additional field observations will clarify the temporal relationship of carbonate and
shoreline features in this lake basin.
3.7.2 U-Th ages of tufa and carbonate deposits and their implications
for past lake level changes
Due to the ambiguous temporal relationships between deposits as described above and the
lack of independent age constraints on paleoshoreline landforms, linking the U-Th ages of
tufa and carbonate deposits to a paleolake area is challenging. Although the encrusting
floret and cone-shaped tufas are likely of algal origin, the absolute water depth of tufa
formation is uncertain.given that it is extremely difficult, if not impossible, to identify a
124
modern analogue of the species of algae responsible for forming the tufa. The elevational
range of the encrusting floret tufa is also quite large at 20 m, and it is unclear if each
generation of floret growth covered part or the whole of this range (Figure 3-3). Although
potentially useful as a paleosalinity indicator, species identification of the ostacod shells
found in floret tufa pore spaces (Figure 4-2G & H) and within carbonate nodules may
not be helpful for interpreting paleowater depths, given that a single species can be found
in a wide range of depths in both lacustrine and marine settings (e.g., Benson, 1984; De
Deckker, 2002).
There is also evidence for lake level fluctuations that are not captured by a shoreline
carbonate deposit. For instance, reworked pebble-sized pieces of tufa with the honey
calcite cement were found at the base of the alluvial fan incision in the southeast (Section
3.9.4 and Figure 3-13D), indicating that water levels must have risen to that elevation at
some point after -15 ka, a period from which we have yet to find material. We must
recognize that carbonate and tufa formation represent a discontinuous record and are not
only dependent on lake levels, but also influenced by factors such as the concentration
of calcium and carbonate in lake waters and groundwater inputs; salinity; mean annual
and seasonal temperatures; biological productivity and nutrient supply; and other lake
chemistry characteristics (Gierlowski-Kordesch, 2010).
Consider the observation that biologically-mediated tufas are not conspicuously forming
in Agua Caliente I today, even though a lake clearly exists in the basin. At least two condi-
tions are necessary for tufa development: (1) the existence of an algal or bacterial mat, and
(2) sufficient concentrations of calcium and carbonate ions to allow for microbially-induced
precipitation of carbonate during photosynthesis (Golubic, 1973), as well as trapping of
clastic carbonate particles. It is possible that tufa formation in this basin occurs only when
the waters from the hot springs in the southwest, where the only obvious signs of modern
microbial activity occur, mix with the waters from the freshwater springs in the north. In
Pyramid Lake, Nevada, similar mixing zones between nutrient-rich thermal spring waters
and alkaline lakes are prime areas of microbialite formation (Arp et al., 1999). Presently,
the waters from the hot springs and the spring-fed lake are separated into two sub-basins.
Although we have no water chemistry data to test our hypothesis, it is possible that the
spring-fed lake and thermal spring waters possess the condition that is lacking in the other
water body that is necessaryfor tufa development. Thus, when these two water bodies
combine, tufas may be able to form. This hypothesis is consistent with the fact that no
tufas of clear biological origin have been found thus far in Salar de Loyoquesand Laguna
125
de Tara, basins for which we have not yet found any evidence for past or present hot spring
activity. However, we know that calcium carbonate concentrations in the lake must have
been high enough in the past for ikaite and carbonate nodule formation. These observa-
tions support the idea that the hot springs in Agua Caliente I is unique, long-lived, and
critical for tufa formation. If the hot springs were to shut off, higher lake levels may not
necessarily induce tufa formation. Other confounding factors such as salinity and temper-
ature could prevent tufa formation even if the thermal spring and lake waters are able to
mix.
We must also consider that the role of the presently inactive stream channel and sill
elevation remains unclear, given that we have not found deposits in Agua Caliente I that
date to >100 ka, nor material in Salar de Loyoques with ages from the last 30 ka. There is
strong evidence showing that the sill elevation between these two basins has changed over
the last 200 ka: According to satellite imagery, accumulation of alluvial fans has occurred
since the regression of the large >100 ka lake in Salar de Loyoques (Section 3.6.4 and
Figure 3-14). The stream channel incision at the foot of these younger, converging alluvial
fans is the most recent change to the sill elevation. At its present elevation of 4240.5 ±
1.1 m (Figure 3-3 & Section 3.9.4), the sill is lower than some encrusting tufa deposits in
Agua Caliente I. Observations and elevational measurements of the stream channel base
indicate that channel incision occurred via flow into the Salar de Loyoques basin (Section
3.9.4); however, our current body of evidence is not sufficient to prove that a higher lake
in the Agua Caliente I basin overflowed into Salar de Loyoques to create this channel,
and we do not know to what elevation Salar de Loyoques would have filled from such an
overflow. Thus our estimate of the highstand lake area for Agua Caliente I, which assumes
no overflow, is a minimum estimate of lake expansion.
Despite these uncertainties, we are still able to make reasonable assumptions on lake
level changes in each basin. The carbonate cements from the beach gravels in Salar de
Loyoques are perhaps the only dateable material that is in direct association with a pale-
oshoreline feature; thus, we are certain that the lake that existed at some point before 100
ka represents a large lake -19 times the size of the water bodies currently occupying the
Salar de Loyoques and Agua Caliente I basins. Both the tufa formations and transformed
ikaites provide a minimum bound on former lake levels, with ikaites growing at or directly
beneath the sediment-water interface and algal reefs forming at or below the water surface.
In Laguna de Tara, U-Th ages for the transformed ikaite crystals and surrounding
cements align with H1 and the earliest Holocene, shortly after the Younger Dryas (YD;
126
Figure 3-9D). During both of these periods, Laguna de Tara expanded to at least 5 times
its modern area. In Agua Caliente I, the honey calcite cement also yields U-Th ages
that coincide with H1. Although the nature of this phase is complex, our present working
hypothesis is that the honey calcite is a "spray-zone" deposit precipitating from lake waters
that intermittently spray over the tufa formations, beach sands, and gravels via wave action
(Section 4). Although the exact water level represented by these deposits is unclear, the
area of the H1 lake in Agua Caliente I must have been at least 1.5-2 times the area of the
modern lake, based on the elevation of the lowest occurrence of the honey calcite cement.
The ~23.5 ka minimum age from the best isolated encrusting floret tufa material may
indicate the presence of a higher lake during H2 and may be related to the uppermost
paleoshoreline features observed in Agua Caliente I, but more U-Th age replication is
necessary before drawing these conclusions.
Although the expansion observed in Salar de Loyoques is impressive, we note that we
must calculate the ratio between the lake area and corresponding drainage basin area for
each lake before making comparisons. This ratio normalizes lake systems by their size and
is directly proportional to total annual precipitation (Hudson and Quade, 2013).
3.7.3 Comparison with shoreline and sediment core records from the
Titicaca-Uyuni and Miscanti lake basins
The HI and possible H2 lake level increases in the Agua Caliente I and Laguna de Tara
basins coincide with the Tauca highstand and Sajsi lake stage in the Titicaca-Uyuni basin,
which is recorded in both lake sediment records (Baker et al., 2001) and shoreline studies
(Placzek et al., 2006b, 2013; Blard et al., 2011; Figure 3-9C). The -10.0-10.3 kyr.BP age of
transformed ikaite crystals in the highest sample from Laguna de Tara suggest that the lake
was also larger shortly after the YD. No robust U-Th ages from Agua Caliente I dating
to around the YD are available, though some preliminary U-Th dating of transformed
ikaite from Agua Caliente I may have formed at this age. Our preliminary data further
support the finding that the highest. lake levels on the Altiplano-Puna plateau coincided
with Heinrich events, not maxima in local summer insolation. It is of great interest to
determine if the lake highstand in Salar de Loyoques coincided with H11 or with one of
the two highest maxima in summer insolation over the last 200ka. It is possible that the
dominance of influence by Heinrich events in the last 100 ka is due to the modulating effect
of eccentricity on the amplitude of precessional insolation changes, which has been small
127
YD H1 H2 H11
tA E -5
500 -4
U-- C 480
4. 460
440 --
0.5
E°- 0
-0.5
Z3 mE a. -1.0
-1.5
3770 - U aT FW1
3740 U/ThC
E0 3710
'U 3680
3650
30 -- o
25 - -- -
- 2 1 0 2
20
E " Agua Caliente I
" S. de Loyoques
0 15 " LagunadeTara
7E N shoreline deposit
10 *min.lake level
5
0'-
.0 W
'U.W
a0
104 108 112 116 120 124 128 132 136 140
kyrs before present
Figure 3-9: [A] Mean DJF insolation at 22
0 S (black; Laskar et al., 2004) and the 5180 record
Botuveri Cave (purple; Cruz et al., 2005; Wang et al., 2007). [B] Interhemispheric temperature
gradient, with uncertainty (Shakun et al., 2012). [C] Lake level history for the.Poopo, Coipasa, and
Uyuni basins in southern Bolivia (15-22°S; Placzek et al., 2006b). [D] Preliminary U-Th dates of
shoreline tufa (squares) and other lacustrine deposits (diamonds) from Agua Caliente I (red), Salar
de Loyoques (blue), and Laguna de Tara (green) (22-23°S), plotted against their elevation relative
to the modern lake or salar. [E] Laguna Bebedero C dates of shoreline tufas (33°S). Period for
H11 is defined by Cheng et al. (2009).
128
over this interval of time.
We present these conclusions tentatively, for there are several caveats to acknowledge.
As discussed previously, the tufa formations and other shoreline deposits may not neces-
sarily capture every lake level fluctuation that occurs in a basin. The growth of algal or
microbial mats and subsequent preservation of such organisms as tufa formations may only
occur under a limited set of conditions involving temperature, salinity, dissolved inorganic
carbon and calcium ion concentrations, et cetera. Such caveats and uncertainties provide
motivation for further work combining shoreline studies with continuous records from lake
sediment cores.
Assuming that these Central Andes lakes are truly only responding to Heinrich Events,
contrasting records at lower elevations which are affected by both precessional insolation
cycles and these millenial-scalecold events, then the following question arises: Why? One
possibility is that the Central Andes may only be sensitive to the most extreme southward
displacements of the ITCZ, given that it is located at the modern day southern edge of
SASM. Indeed, Heinrich Events and other cold North Atlantic events that dramatically
increase the interhemispheric temperature gradient are associated with the largest shifts
in ITCZ position (Donohoe et al., 2013; McGee et al., 2014). A second possibility is that
these lake level changes are responses to changes in Southern Hemisphere westerly jet
dynamics. Chiang et al. (2014) suggest that North Atlantic cooling may cause weakened
and more zonally symmetric Southern Hemisphere subtropical westerly winds, which also
move poleward. This weakening and poleward shift of the subtropical westerly finds then
allows for more convection and advection of moisture onto the Altiplano-Puna plateau.
3.8 Conclusions
Our preliminary U-Th data and shoreline mapping suggests that lake levels in Agua
Caliente I and Laguna de Tara were significantly higher than present levels during H1.
Agua Caliente I. may have also been higher during H2, but further replication of U-Th ages
is needed. In Salar de Loyoques, we have evidence for a large lake that incorporated the
Agua Caliente I lake basin existing at some time before 100 kyrs ago. Further U-Th dat-
ing will. determine if this lake corresponds to local summer insolation highs or H11. With
increases in lake level coinciding with H1, we have extended the known region of SASM
influence southwards from the Titicaca-Uyuni lake system.
We have yet to take full advantage of the numerous observed field relationships between
129
different tufa varieties in Agua Caliente I. Applying U-Th dating to carbonates with clear
stratigraphical relationships will allow us to develop a more nuanced picture of lake level
changes in these basins. After better understanding these field relationships and their
implication for past lake levels, as well as further replicating U-Th ages, the next step will be
to develop basin-specific water balance models to determine the magnitude of precipitation
and evaporation changes that would be necessary to create such large lakes. We plan to
follow and adapt the methods described in other papers for GIS watershed analysis (Hudson
and Quade, 2013) and the water balance model (Blard et al., 2009; Placzek et al., 2013).
Such a model will need to pay special attention to the impacts of groundwater flow into and
out of the basins. Temperature constraints necessary to determine evaporation rates will be
gleaned from studies on local glaciers and possibly pollen records, as well as experimental
clumped isotope analyses on the lake carbonates in our study. Such quantitative constraints
on past rainfall will be useful for proxy-model comparisons of general circulation models
that are currently used for future projections of precipitation under a warming climate.
Although there remain many unknowns, we feel confident that applying the these same
methods to a series of 6 lakes along a north-south transect in northern Chile will ultimately
allow us to create a spatio-temporal map of water balance changes over the late Pleistocene.
There are also several potential avenues for geomorphology-inclined research pertaining
to the paleoshorelines and alluvial fans in these basins, which we describe in Section 3.9.5.
We also plan to conduct radiocarbon ("C) analyses on lake carbonate materials from Agua
Caliente I and Laguna de Tara to constrain the reservoir effect in these basins, which will
be useful for any future work on lake sediment cores.
Acknowledgements
We thank Rick Kayser at MIT and Soumen Mallick at Brown University for their help
with mass spectrometry; Ashling Neary for running the ATR-FTIR carbonate mineralogy
measurements; Elena Steponaitis, Irit Tal, and Ben Hardt for lab chemistry assistance;
and Tim Grove and Ben Mandler for petrographic scope assistance. Weifu Guo assisted
C.Y.C. in making the stable isotope measurements at WHOL. Discussions with Kristin
Bergmann, Adam Hudson, and Elena Steponaitis on tufa microfacies interpretation were
insightful. We also thank Kim Huppert, Roger Fu, Justin Strup, Francisco (Pancho)
Gonzales, Hector Orellana, and Marty Pepper for field assistance. Claudio Latorre and
Blas Valero Garces both provided crucial advice on conducting field work in Chile. This
work was supported by the NSF Graduate Research Fellowship, MIT Ida Green Fellowship,
130
"I PI I IN R I IN 1ROINN 1111111 1 -,---..- .
MIT EAPS Grayce B. Kerr Fellowship, MIT EAPS Callahan-Dee Fellowship, MIT EAPS
Student Research Fund, MIT International Science and Technology Initiatives (MISTI),
WHOI Ocean Ventures Fund, and the Coner Science and Education Foundation.
131
Table 3.2: Locations and descriptions of tufa and other lacustrine carbonate samples from Agua Caliente I, Salar de Loyoques, and Laguna de
Tara. Sample Name is the original name of the sample and is referred to in the text of the paper. Corresponding ID numbers are referenced
in figures. Reported coordinates and elevations of samples from Agua Caliente I and Salar de Loyoques were measured-using the Trimble CEO
7x device. Coordinates and elevations of samples from Laguna de Tara were measured using a less precise GarminG PS device. Mineralogy
was determined using ATR-FTIR spectroscopy. Empty spaces indicate that a measurement has not yet been made for that attribute on that
sample.
ID Lat udedLo e dGP I Vert.
Sample Name No. ) ) Eev. m) Prec. (m) Mineralogy Description
Agua Caliente I
AD10-225(B) 1 -23.084303 -67.401078 4233.5 0.8 calcite white porous tufa
AD09-98(D) 2 -23.083391 -67.407970 4240.6 0.8 white porous tufa
AD10-233a(A) 3 -23.174139 -67.399202 4243.0 0.9 calcite honey calcite
AD10-233a(A) 4 -23.174139 -67.399202 4243.0 0.9 calcite honey calcite
AD10-233a(B) 5 -23.174139 -67.399202 4243.0 0.9 calcite honey calcite
AD10-225(A) 6 -23.084303 -67.401078 4233.5 0.8 calcite honey calcite in tufa
CYC15-049(A) 7 -23.160174 -67.397851 4242.7 0.9 honey calcite cementing beach sand
AD09-98(A) 8 -23.084021 -67.406679 4234.0 2.0 calcite mixture of honey calcite and white tufa
AD09-98(A) 9 -23.084021 -67.406679 4234.0 2.0 calcite mixture of honey calcite and white tufa
AD09-98(B) 10 -23.084021 -67.406679 4234.0 2.0 mixture of honey calcite and white tufa
AD09-98(C) 11 -23.084021 -67.406679 4234.0 2.0 mixture of honey calcite and white tufa
AD10-228(A) 12 -23.085568 -67.402128 4236.6 0.8 calcite mixture of honey calcite and white tufa
AD10-228(C) 13 -23.085568 -67.402128 4236.6 0.8 calcite mixture of honey calcite and white tufa
AD10-228(D) 14 -23.085568 -67.402128 4236.6 0.8 mixture of honey calcite and white tufa
ADO9-95(A) 15 -23.083391 -67.407970 4240.6 0.8 calcite mixture of honey calcite and white tufa
AD09-95(B) 16 -23.083391 -67.407970 4240.6 0.8 calcite mixture of honey'calcite and white tufa
AD09-95(C) 17 -23.083391 -67.407970 4240.6 0.8 calcite mixture of honey calcite and white tufa
AD09-95(D) 18 -23.083391 -67.407970 4240.6 0.8 mixture of honey calcite and white tufa
AD09-95(E) 19 -23.083391 -67.407970 4240.6 0.8 mixture of honey calcite and white tufa
AD09-100(A) 20 calcite mixture of honey calcite and white tufa
ADO9-100(B) 21 mixture of honey calcite and white tufa
AD09-221(A) 22 mixture of honey calcite and white tufa
Continued on next page
Table 3.2 - continued from previous page
ID Latitude Lon tude dGPS Vert.
Sample Namd No. () ) Elev. (in) Prec. (m) Mineralogy Description
AD10-226(A) 23 -23.082755 -67.401042 4237.2 3.6 calcite ikaite pseudomorph
AD10-226(A) 24 -23.082755 -67.401042 4237.2 3.6 calcite ikaite pseudomorph
ADIO-226(B) 25 -23.082755 -67.401042 4237.2 3.6 ikaite pseudomorph
AD09-99(A) 26 -23.084264 -67.406469 4228.8 0.8 calcite platey carbonate nodule
AD09-99(B) 27 -23.084264 -67.406469 4228.8 0.8 platey carbonate nodule
AD09-221(B) 28 calcite platey carbonate nodule
AD09-96(A) 29 platey carbonate nodule(?)
AD09-96(A) 30 platey carbonate nodule(?)
AD09-96(B) 31 calcite platey carbonate nodule(?)
CYC15-016A(A) 32 -23.085533 -67.400864 4244.8 0.9 white, clean calcite infilling(?)
CYC15-016B(A) 33 -23.085533 -67.400864 4244.8 0.9 white, clean calcite infilling(?)
AD09-101(A) 34 calcite ambiguous
Salar de Loyoques
CYC15-047(A) 35 -23.301139 -67.264143 4251.0 1.7 carbonate cementing beach gravel
CYC15-047(B) 36 -23.301139 -67.264143 4251.0 1.7 carbonate cementing beach gravel
CYC15-025AA(A) 37 -23.202370 -67.270781 4192.6 1.0 ikaite pseudomorph
CYC15-025AA(B) 38 -23.202370 -67.270781 4192.6 1.0 ikaite pseudomorph
CYC15-022A(A) 39 -23.202575 -67.272532 4205.5 1.1 ikaite pseudomorph
CYC15-021A(A) 40 -23.202267 -67.272589 4208.5 2.5 platey carbonate nodule
Laguna de Tara
AD10-246(A) 41 -23.030042 -67.347418 4334 - calcite ikaite pseudomorph
AD10-246(B) 42 -23.030042 -67.347418 4334 - calcite ikaite pseudomorph
AD10-245(A) 43 -23.029639 -67.348633 4349 - calcite ikaite pseudomorph
AD10-245(B) 44 -23.029639 -67.348633 4349 - calcite ikaite pseudomorph
AD10-244(A) 45 -23.027097 -67.352499 4352 - calcite ikaite pseudomorph
AD1O-244(A) 46 -23.027097 -67.352499 4352 - calcite ikaite pseudomorph
AD10-244(C) 47 -23.027097 -67.352499 4352 - calcite ikaite pseudomorph
Table 3.3: U-Th dating results of carbonates examined in this study. In the ID number column, a '*'symbol indicates that the
powder dated is a mixture of two phases of carbonate and does not represent an age linked to a specific geologic or hydrologic
event. A 't' symbol indicates that the reliability and interpretation of the age as being representative of a specific event is
unclear due to factors such as ambiguous sample context, high blank corrections, or low 230Th/ 232Th ratios, and requires further
investigation and replication. 'Reported errors for 238U and 232Th concentrations are estimated to be ±1% due to uncertainties
in spike concentration; analytical uncertainties are smaller. b523 4U = ([2 34 U/ 2 38 Ulactivity - 1) X 1000. 2 3c 0Th/ 238Uactivity
1 - 234 eA23nT + (5 fmeasured/1000)[A230/(A230 - A23 4 )](1 - e-( 23o-A234)T), where T is the age. "Uncorrected" indicates
that no correction has been made for initial 2 30 Th. dAges are corrected for detrital 2 30 Th assuming an initial 230/ 23 2 Th of
(4.4±2.2)x106. e5234Uinitial corrected was calculated based on 2 30 Th age (T), i.e., 23 46 Uinitial = 25 34 Umeasured x e A2 3 4 T,
where T is the corrected age. fB.P. stands for "Before Present" where the present is defined as January 1, 1950 C.E. Decay
constants for 2 30 Th and 2 34 U are from Cheng et al. (2013a); decay constant for 23 8U is 1.55125x10-10y r-F (Jaffey et al.,
1971).
23o (acti i2t38) I (ppmb ,ujhAt3gMe " ~~\ ,Aekjr d 2 3 45 Ui e Az(v B.P)f
2 2 
1O. (ng,/ 
23
(pggy at ity)U OT h t(cu n rtcorrece (% corWNe"o)e Ar )ri/ect)
Agua Caliente I
1 15100 300 30900 ± 1100 1241.7 1 0.4433 0.001 3440 100 23550 ± 60 23520 ± 60 1326.9 1.1 23410 ± 60
2* 9700 190 62000 ± 1300 1192.4 0.9 0.3829 0.0011 951 7 20570 ± 70 20490 ± 80 1263.4 1 20380 ± 80
3 4000 80 1900 ± 200 1262.9 1.3 0.3 ± 0.005 10100 1200 15300 300 15300 300 1318.7 1.7 15200 300
4 4030 80 3330 ± 120 1281.2 1.2 0.3082 ± 0.0009 5930 180 15620 50 15610 50 1339 1.2 15500 50
5 4080 80 16500 800 1272.7 1.2 0.3052 ± 0.0013 1200 ± 50 15520 70 15470 80 1329.5 1.3 15360 80
6 3183 19 6765 5 1273 3 0.304 ± 0.002 2400 ± 20 15450 140 15420 140 1329 3 15310 140
7 4320 90 141000 3000 1137.5 1.8 0.2776 ± 0.0017 135.2 ± 0.8 14990 100 14500 200 1185 2 14400 200
8* 7602.6 1.8 181350 160 1227 2 0.3536 ± 0.0014 246.7 ± 1.1 18570 80 18270 170 1292 3 18160 170
9* 5120 100 72000 2000 1217.4 1.2 0.318 ± 0.005 357 10 16700 300 16500 300 1275.4 1.7 16400 300
10* 4210 80 10200 400 1220.6 1.6 0.285 ± 0.006 1870 80 14800 300 14800 300 1273 2 14600 300
11 4520 90 11800 500 1227 1 0.2954 ± 0.0012 1790 70 15320 70 15290 70 1281.1 ±1.1 15180 70
12* 3920 30 16834 13 1252 3 0.311 ± 0.003 1213 10 16010 180 15960 180 1310 3 15850 180
13 5780 120 38000 800 1233.1 1 0.2998 ± 0.0007 724 2 15520 ± 40 15430 60 1288 1.1 15320 ± 60
14 5030 100 11800 400 1233.4 0.9 0.2992 0.001 2020 ± 60 15490 ± 60 15450 60 1288.3 1 15340 ± 60
15* 7922.6 1.8 21910 ± 20 1279 2 0.3203 0.0013 1942 ± 14 16290 ± 70 16260 80 1339 2 16150 80
16 2420 40 11025 ± 11 1260 5 0.308 0.006 1140 ± 16 15800 300 15700 300 1317 5 15600 300
17* 8570 ± 170 18000 400 1253 1.2 0.3366 0.0008 2542 ± 16 17390 40 17360 50 1315.9 1.3 17250 50
18 4980 ± 100 2600 300 1262.8 0.9 0.3028 0.0013 9200 ± 1200 15470 70 15460 70 1319.1 0.9 15350 ± 70
'19* 3560 1 70 6200 300 1245 0.9 0.3113 0.0013 2830 1 150 16060 70 16040 70 1302.7 1 15930 ± 70
20* 1260 ± 20 22200 ± 20 1268 6 0.314 0.008 298 ± 5 16000 400 15800 400 1326 ± 7 15700 ± 400
Continued on next page
Table 3.3 - continued from previous page
._ (gh) (g(g g (%0) (atc t y U (ppTmh / 20 c (u b r(9r ted) (corr o, coreN'ed) Igor_____ orete)
21t 7470 ± 150 80600 ± 1700 1234.3 ± 0.9 0.3282 0.0011 483 ± 3 17080 ± 60 16940 90 1294.7 ± 1 16830 90
22* 16857 ± 5 271700 ± 200 1195.6 ± 1.8 0.3137 0.001 323.8 ± 1.2 16590 ± 60 16380 120 1252.1 1.9 16270 120
23t 2059 ± 15 10718 ±'9 1193 ± 3 0.269 0.003 864 ± 7 14110 150 14040 160 1242 3 13940 160
24t 2190 ± 40 18600± 800 1172.3 2 0.24 0.02 450 40 12700 1100 12600 1100 1215 4 12500 1100
25t 2450 ± 50 21400 400 1162.9 1.2 0.3944 0.001 715 3 21560 60 21450 90 1235.4 1.3 21340 90
26* 20654 ± 5 735400 600 1206.7 1.7 0.3345 0.0012 156.2 0.6 17670 70 17200 200 1267 2 17100 200
29 2015 13 12172 12 - 1189 3 0.277 0.003 765 6 14570 150 14490 150 1239 3 14380 ±150
30* 8610 170 138000 3000 1171.7 1.2 0.2693 0.0009 266.1 1.8 14270 ± 50 14060 120 1219.1 1.3 13950 ± 120
31t 10900 200 69200 1700 1183.5 1.2 0.2819 0.0015 708 11 14890 ± 90 14810 100 1234 1.3 14700 ± 100
32t 13500 300 600 300 1212.8 1.1 0.1464 0.0004 50000 20000 7420 ± 20 7420 20 1238.5 1.1 7310 ± 20
33t 14800 ±.300 6500 700 1197.3 1.6 0.1098 0.0006 4000 400 5560 ± 30 5560 30 1216.2 1.6 5450 ± 30
34* 254 ± 8 5812 6 1226 9 0.315 0.012 . 230 5 16400 ± 700 16100 700 1283 10 16000 ± 700
Salar de Loyoques
35t 1550 ± 30 263000 5000 723 6 1.33 0.03 125.2 ± 0.7 139000 ± 5000 137000 5000 1063 18 136000 ± 5000
36t 1520 ± 30 196000 4000 708.8 i- 1.6 1.252 0.006 153.6 ±0.8 126500 t 1100 124500 1500 1007± 5 124400 ± 1500
37t 5690 ± 110 7800 200 752.1 ± 1.4 1.214 0.003 14100 ± 300 114900 ± 500 114900 500 1040 2 114800 500
38f 5430 ± 110 10400 300 766.3 ± 1.3 1.259± 0.003 10470 ± 150 120200 ± 500 120100 500 1076 2 120000 500
39t 10100 ± 200 167000 3000 705.2 ± 1.3 1.324 ± 0.01 1272 ± 12 140000 2000 140000 2000 1047 6 140000 2000
40t 21700 ± 400 293000 6000 713.4 ± 1.3 1.1825 ± 0.002 1389 ± 4 114500 300 114300 400 985 2 114200 400
Laguna de Tara
41t 2956 ± 19 46380 40 889 ± 3 0.25 0.002 265.7 ± 1.9 15340 160 15110 200 928 3 15000 200
42t 2500 ± 50 87600 ±1800 930.8 ± 1.4 0.2694 0.0008 122.1 ± 0.3 16200 50 15700 300 972.9 1.6 15600 300
43t 2836 ±19 157290 140 890 ±3 0.276 0.003 82.9 ± 0.8 17000 200 16200 500 932 3 -16100 500
44t 3720 ±70 263000 5000 921 ±1.3 0.2802 0.0008 62.92 ±0.13 16980 50 15900 600 963 2 15800 600
45t 18009 ± 7 166420 190 1041.1 ± 1.7 0.1886 0.0006 339.7 1.4 10490 40 10360 70 1072 1.8 10260 70
46t 10700 ±200 55700 ± 1700 1031.4 ±1.3 0.182 0.003 554 16 10150 190 10070 200 1061.1 1.5 9960 200
47t 3300 ± 70 65700 ±1300 954 ±1.6 0.2637 0.0012 210.5 1.1 15630 80 15330 170 996.2 1.7 15220 170
Table 3.4: Stable isotope data of samples from Agua Caliente I and Laguna de Tara. See Table 3.2 for ID number references.
See Section 3.6.3 for a description of the method of data acquisition.
ID 1 3C UofA 1 U 5'80 UofA 1a 513c 51o
No. (%0) (%0) (Poo?, __ _ __ j
Agua Caliente I
1 0.34 0.03 1.56 0.03 0.58 0.01 1.21 0.02
3 1.75 0.01 1.60 0.01 1.91 0.01 1.83 0.04
4 1.75 0.01 1.60 0.01 1.91 0.01 1.83 0.04
6 2.47 0.00 2.09 0.04 2.41 0.01 2.03 0.01
8 2.16 0.01 2.03 0.04 2.24 0.02 1.96 0.02
9 2.16 0.01 2.03 0.04 2.24 0.02 1.96 0.02
10 2.62 0.00 1.67 0.02
12 1.99 0.04 1.76 0.03 2.53 0.26 1.55 0.04
15 2.00 0.05 1.24 0.07 2.00 0.07 1.05 0.11
16 2.00 0.00 1.41 0.01
20 2.35 0.00 1.38 0.03 2.17 0.02 1.81 0.01
22 1.14 0.02 1.25 0.05
24 3.06 0.02 1.70 0.06 2.89 0.01 1.51 0.02
26 -1.28 0.02 1.74 0.04 -1.43 0.09 1.85 0.05
27 -2.35 0.02 2.34 0.07 -2.28 0.01 2.36 0.02
28 -1.45 0.02 2.19. 0.02
31 1.92 0.02 2.42 0.04 1.57 0.06 2.03 0.04
34 1.96 0.01 2.34 0.02 2.26 0.24 2.29 0.03
Laguna de Tara
41 3.21 0.01 1.02 0.04
43 2.61 0.01 0.67 0.06
45 2.80 0.02 0.77 0.05
46 2.80 0.02 0.77 0.05
47 2.63 0.05 0.78 0.01 1
3.9 Supplementary Materials
3.9.1 Differential GPS measurements of shoreline features and sample
locations
A GPS receiver (rover) calculates distance using the travel time and velocity of radio signals
from orbiting satellites. These radio signals may be delayed by atmospheric disturbances
in the troposphere such as cloud cover or charged particles in the ionosphere. Discrepancies
between satellite and receiver clocks, varying levels of satellite connectivity, and inaccurate
monitoring of satellite positions also can lead to errors. We use dGPS to correct for these
errors by relying on the coordination of two receivers: a stationary base station and a
roving receiver making dGPS measurements. The coordinates of the known locations of
base stations serve as local points of reference, which then are used for differential correction
of rover data in post-processing. The base stations run around the clock collecting data on
their location, quantifying drift by comparing their current measured GPS location with
its single known location. In post-processing, the proprietary Trimble Pathfinder Office
software differentially corrects rover data according to this calculated amount of drift at
the time of measurement by comparing time stamps between rover and base station data.
For the differential correction of all dGPS data collected in Agua Caliente I and Salar
de Loyoques, we used hourly data from a Scripps Orbit and Permanent Array Center
(SOPAC) base station in Cordoba, Argentina, located -950 km air distance away from
the site. This base station is the closest station to our field site for which data is easily
accessible through the proprietary software. For the 197 dGPS measurements taken, the
estimated post-processed accuracy of the 197 dGPS measurements taken falls between 0.8
and 1.0 m for 57.9% of data; 1.0 and 2.0 m for 33.0% of data; and >2.0 m for 9.1% of
data. See Figure 3-12A for histogram of estimated post-processed accuracy of all dGPS
measurements. All dGPS measurements were made in the WGS84 datum (EGM96 geoid),
with elevations reported relative to mean sea level (MSL).
3.9.2 Calculation of modern lake areas and paleolake areas
We used ESRI ArcGIS 10.2 software to determine the area of modern lakes and paleolakes.
Using the dGPS elevation measurements and satellite imagery of the highest paleoshoreline
features in each basin, we generated contour lines of elevation from the1 -arc second (-30
m) digital elevation model from the Shuttle Radar Topography Mission (SRTM DEM)
137
and used these outlines as approximations of the perimeter of highstand paleolake areas,
accounting for the -3.5 m average difference between dGPS elevation measurements and
the 30-m SRTM DEM (3-12). The following contour lines of elevations are used to represent
the highstands of each lake basin: 4248 m for Agua Caliente I; 4253 m for Laguna de Tara;
and 4270 m for Salar de Loyoques. To estimate the outline of the modern lake in each lake
basin, we identified 'flat patches' within the SRTM DEM, taking advantage of the fact that
the SRTM DEM records water surfaces as horizontal planes. These outlines for paleolake
highstands and modern lakes were then used for lake area calculations, using an equal-area
conic projection centered on South America.
In the future, we plan to apply the methods described in Hudson and Quade (2013) to
better estimate modern lake areas. Our current methods use the lake areas of 2009 (the
year of the SRTM) to represent modern.
3.9.3 Differences in 523 4 Uinitial values of lake carbonates between basins
The carbonates from Agua Caliente I analyzed thus far vary within a narrow range of
6234 Uinitial values, from 1180 to 1340 %o (average = 1280 ± 40%o). In Salar de Loyoques,
the 6 234 Uinitial values of carbonates also fall within a narrow range between 980 and 1080%o
(average = 1040 30%o). In Laguna de Tara, 5 234 Uinitial values range between 930 and
1070%o (average= 990 ± 60%o).
Unpaired student t tests of the 5234Uinitial values in each basin indicate that Agua
Caliente I 5 234Uinitial values are statistically distinct from the other two basins, suggest-
ing that the Agua Caliente I basin's water source spends a statistically significant dif-
ferent amount of time interacting with bedrock before entering the lake. The difference
in 5 234 Uinitial value between samples from Salar de Loyoques and Laguna de Tara is not
statistically significant.
3.9.4 Paleoshoreline features in Agua Caliente I and Salar de Loyoques
We measured the elevations of various paleoshoreline features in Agua Caliente I and
Salar de Loyoques. In Agua Caliente I, the best preserved features were located in the
northeastern corner of the basin. At least three distinct abrasion platforms in volcanic
bedrock hillsides are present between 4244 and 4252 m elevation. The elevation of the
base of the alluvial fan scarp varies between 4243 and 4245 m. Satellite imagery of the
fan incision in the northwest part of the basin shows that there is-a noticeable amount of
138
salt deposit "staining" the landscape at ~4235 m, suggesting that the water levels may
have fluctuated at approximately that elevation for a long enough duration to leave such
salt deposits. At the base of the incised alluvial fan scarp in the southeast, we found
pebble-sized pieces of reworked white tufas with the honey calcite cement (Figure 3-13),
suggesting that a lake must have existed at this elevation at some point after the deposition
of the honey calcite.
Three separate sets of gravel barriers occupy a narrow range of elevations. The highest
set consists of at least 6 individual barriers that occupy a narrow, -3 im range of crestal
elevations (4244-4247 m) across a lateral distance of ~200 m. The relief of these gravel
barriers is minimal, with crestal heights of <0.5 m relative to lows between individual
barriers. Although these barriers are obvious in satellite imagery, the individual crests
of gravel barriers are only distinguishable on the ground by the relative abundance of
pink disk-shaped ignimbritic cobbles, which are more abundant on crests. Cross-sectional
exposure of these gravel barriers via the channel of spring discharge reveals that beach
cobbles and pebbles are oriented lake wards for at least several meters below the surface.
The lowest set of gravel barriers, clustering around ~4230 in elevation, are more promi-
nent in relief (3 in). These barriers also contain reworked floret tufas with the nodular
carbonate cap. Unfortunately, we did not examine the intermediate set of gravel barri-
ers, but we estimate their elevation to be approximately 4233-4235 in based on satellite
imagery and the DEM (Section 3.9.2). It is possible that this set of barriers should be
grouped with the lowest, set of barriers.
For our measurements of the base of the inactivestream channel connecting the Agua
Caliente I and Salar de Loyoques basins, the highest elevation measured was 4240.5 ± 1.1
in. We tentatively call this elevation the "sill" or spillover elevation for Agua Caliente I.
Taking advantage of color differences between different alluvial fans along the length of
this inactive stream channel, it appears that the direction of flow is predominantly from
northwest to southeast based on the color of entrained materials into the channel, i.e., from
Agua Caliente I to Salar de Loyoques. This observation is consistent with the monotonically
decreasing elevation of the stream channel base towards the southeast, as well as an alluvial
fan feature at the juncture of the stream channel and the Loyoques basin. However, recent
human bulldozing activity at both ends of the stream channel make it difficult to entirely
rule out a pathway of water into Agua Caliente I. The relationship between the elevations
of the sill, the paleoshoreline features, and the tufa and other carbonate deposits found
within Agua Caliente I is illustrated in Figure 3-3.
139
In Salar de Loyoques, constructional paleoshoreline features are only well-preserved
within two embayments in the southeast part of the basin. At least 9 individual gravel
barriers exist, spanning an elevational range of 4225-4270 m across a distance of ~1.3 kin.
Faint traces of gravel barriers in the northern part of the basin are between 4191-4192 m
elevation.
3.9.5 Potential avenues for future geomorphological research
The following are potential pathways for future geomorphological research in these lake
basins:
• The U-Th ages of shoreline deposits could be compared to optically simulated lumi-
nescence (OSL) of beach gravels and detrital material beneath the caliche, as well as
cosmogenic 3He exposure dating of the abrasion platforms of basalt. Any statistically
significant difference between such dating results could constrain "growth" rates of
the encrusting floret and cone-shaped tufa deposits, or help calibrate the OSL and
3He exposure dating systems.
" Due to color differences in different volcanic deposits in the Laguna de Tara basin,
the cumulative amount of along-shore transport is recorded by lighter-colored ign-
imbritic deposits settling overdarker colored sediments of basaltic composition in the
northwest and southeast parts of the basin. Given basin geometryand an average
wind speed and direction, process-based modeling could determine the cumulative
amount of time that the lake occupied the elevation of these paleoshorelines based
on the cumulative amount of along-shore transport indicated by such deposits.
• The alluvial fans in the western and southernmost parts of Salar de Loyoques formed
after the regression of the large lake that existed at some time before 100 ka. Once
a robust U-Th age is determined for the carbonate cements of the beach gravels, we
can calculate a. lower bound on the accumulation rate of these fans.
140
Sample Elev. Lon. Lat. 580 5D Description
Name (m) (°W) (°S) (%0) (%o)
AD09-103 4238 67.4081 23.0885 2.6 -31 spring, ACI
AD10-222 - 67.4002 23.0867 -7.6 -77 main outflow channel of ACI
AD10-229 4225 67.4173 23.1473 0.6 -36 steaming hot spring, SW corner of ACI
AD1O-240 4329 67.3353 23.0100 -7.6 -77 small spring, L. de Tara
AD10-243 4332 67.3388 23.0180 -9.1 -83 major creek flowing into NW corner of Tara
Table 3.5: Table of 6180 and6 D values of various waters from Agua Caliente I (ACI) and Laguna
de Tara. Elevations and coordinates were measured with a a handheld Garmin GPS.
4.0
Asia
3.5 Africa
%A N. America
3.0- S. America
Europe
-012G 2.5- Australia
_ 0 over-represented
2.0-
I-. 1.5-0.~
South America
0- 1.0- .......- N=210
under-represented
0.5-
0 j I I I I I I I I I I
0 2 4 6 8 10 12 14 16 18 20
kyragoB.P.
Figure 3-10: [A] Data representation by continent in the Global Lake Status Database
(Qin et al., 1998; Kohfeld and Harrison, 2000). A value of 1.0 indicates that a continent
has the same data density as the global average for that timestep. [B] and [C] Comparison
of lakes with status information in the database from the past 30 ka in South America and
Africa, respectively.
141
2.6-
T 31
2.4--
27* 2
28
2.2--
31 ' 2productivity
31) 7 6
2.0--
2
1.8- 26
0 1.6 - 0 24
Go 11
1.4 -- *white shoreline tufa
honey calcite 22 1 5
1.2 honey calcite in tufa
*mixed hney calcite 15
1.0-
ikaite pseudomorph
0.8 -- platey carbonate 5,6
X ambiguous
0.6- I . . I . .
-3 -2 -1 0 1 2 3 4
613C (%o, VPDB)
Figure 3-11: Cross-plot of 6 13 C and 51 80 values of carbonates samples from Agua Caliente
I (red) and Salar de Loyoques (green). Numbers correspond to the ID number of the sample
(see Table 3.2). Error bars indicate 2-- uncertainty. See Table 3.4 for stable isotope values.
142
120
N= 197 N = 197
p = 1.2 m m
100 a= 0.6 m 60 y1 m =-3.5  a = 2.5 m
50
80
401
60 0
30
40
20-
20
10
0
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 -12 -10 -8 -6 -4 -2 
0 2 4
dGPS Post-Processed Estimated Vertical Accuracy (m) dGPS Elevation - SRTM 30m DEM Elevation (m)
Figure 3-12: [A] Histogram of the estimated post-processed vertical accuracy of all dGPS
measurements from Agua Caliente I and Salar de Loyoques. Mean accuracy is ± 1.2 m (1-
u-). [B] Histogram comparing all dGPS elevation measurements from Agua Caliente I and
Salar de Loyoques to the SRTM 30 m DEM. Mean difference is -3.5 ± 2.5 m. Red dashed
lines represent the mean, with the gray shaded areas and black dashed lines representing
one standard deviation from the mean.
143
Figure 3-13: Panels showing the location of reworked pieces of tufa with honey calcite,
which were found in the float of the scarp of an incised alluvial fan [B] in the southeast
corner of the Agua Caliente I lake basin [A]. [C] Kim Huppert stands at the base of the
scarp while Justin Stroup stands at the top. The elevational difference between the base
and top of the scarp is- 4.5 m. [D] Photo of the reworked pieces of tufa with honey calcite.
Chilean peso for scale is 2.7 cm in diameters.
144
/
I
Figure 3-14: [A] Comparison of the maximum areal extent of Laguna de Tara (green
line) to modern day extent of the lake (cram yellow polygons). [B] Comparison of the
maximum arpal extent of Agua Caliente I (red line) and Salar de Loyoques (blue line) to
modern day extent of the lake (cream yellow polygons). The outline for Agua Caliente I is
drawn underthe assumption that no overflow occurred into the Salar de Loyoques.
145
modern freshwater spring highest elevation
4300- surface discharge point in stream cut highest measurec
4280- highest erosional platform 
grave/barrier
highest gravel barrier highest hico hest
E 4260- bo carbonate-cementedbottomofftomaof ninnciisinonc, sinplaptaerbayte  beach gravel sample
highest tufa samples carbonate.94240 collected , * highest diatomite on top
ikaite of reworked carb.
>4220
lowest"/,spring honey calcite andikaite
4200- clooelclestc ttued  in-place travertineincolluvium 
lowest in-place ikaite
ikaite* deposit
4180-
0 5 10 15 20 25 30 35 40
Cross-sectional Transect (km)
Figure 3-15: Cross-sectional transect of the Agua Caliente I and Salar de Loyoques basins,
showing the elevations of major paleoshoreline features and lacustrine carbonate samples.
The actual distance listed on the x-axis is not meaningful.
146
Chapter 4
Honey calcite: gravitational drip
cements of lacustrine origin
preserve evidence of rapid,
large-magnitude lake level
fluctuations
4.1 Introduction
Porous carbonate build-ups of lacustrine origin called "tufas" are widespread in desert
landscapes, but determining the age of their formation, and thus the timing of the lakes
that formed them, has been notoriously difficult. Early researchers attempting to date
such deposits with carbon-14 techniques remarked on their internal complexity, noting that
multiple carbonate deposition events were often present on sub-centimeter scales (Broecker
and Orr, 1958; Kaufman and Broecker, 1965). Because identification of such composite or
altered samples was not always obvious, and radiometric dating tools at the time could
not operate on sample sizes smaller than the scale of cross-cutting relationships, tufas were
generally considered problematic for geochronological purposes. These sentiments were
further ingrained as subsequent research revealed the problem of carbon reservoir effects in
precipitated carbonates, and even moreso when initial promise for U-Th dating techniques
147
was dampened by difficulties with detrital constituents in tufas (Ku and Liang, 1984).
Here, we report the existence of a distinct carbonate cement found in close association
with tufas and other lake deposits that may reignite optimism for geochronological control
of these materials. This "honey calcite" is named for its golden color and cloudy translu-
cency, and is documented in two geographically disparate Pleistocene-aged lake basins in
northern Chile and the southwestern United States. We suspect its occurrence in closed-
basin lake tufa and shoreline deposits is more common than previously known. We show
that U-Th dates of this deposit are reproducible and stratigraphically coherent, suggesting
that the data are viable as geologically meaningful and interpretable age constraints. The
significance of this deposit is even more compelling when considering its morphology: it
appears within cracks and void spaces as a gravitational "dripstone" or pendant cement
characterized by thickening along the undersides of tufa or other substrates. This morphol-
ogy is considered a classic indicator of the vadose zone, where cements of uneven thickness
precipitate from excess droplets of saturated water and are therefore orientated parallel to
the gravity vector (Mller, 1971; Longman, 1980; Schoelle and Ulmer-Schoelle, 2003).
Because of the diagnostic morphology, field context, and the lateral and elevational
extent of the cement, we interpret the honey calcite to represent a "spray-zone" deposit
formed by high-magnitude lake level fluctuations. The presence of fine laminations in the
cement and U-Th ages suggest that these fluctuations may be rapid, occurring on century
to decadal timescales. We present U-Th geochronological data, petrographic analysis,
and stable isotope results to support our argument. At the site in northern Chile, the
duration of fluctuations captured by the honey calcite is confined to the last -1000 years of
Heinrich Event 1 (18-15 ka). We speculate that this may suggest that the time immediately
preceding the termination of this north Atlantic winter cooling period was characterized
by high variance (unstable) and large magnitude local climate change, similar to findings
by Bakke et al. (2009) from high-resolution northern Atlantic marine sediments and Pigati
et al. (2019) from desert wetlands during the Younger Dryas cold period.
4.2 Field and Geologic Context
The most pristinely preserved instance of honey calcite found thus far is from Agua Caliente
I (23.13 0S, 67.41°W, 4200 meters above sea level; Figs. 3-2 and 3-7A) in northern Chile. In
this basin, the cement is observed in two distinct situations: (1) coating the exterior of or
infiltrating primary void spaces within porous tufas and (2) cementing sands and gravels
148
of colluvium or gravel barrier deposits. In all observed cases,. the honey calcite is dense
(non-prous):,hard, and crystalline.
For (1), throughout the basin, the honey calcite is found associated with the encrust-
ing floret and cone-shaped tufas (see orange vertical bar in Fig. 3-3). In these contexts,
the honey calcite appears thickest along the exterior sides of cones (Figure 4-1) and other
surfaces that were exposed at the time of honey calcite deposition. Fig. 4-2 features pet-
rographic images of the honey calcite infiltrating an encrusted floret tufa sample collected
from -14 m above the modern salt flat. The sample shared a contact with an igneous
bedrock boulder and was cross-sectionally slabbed along a plane normal to the contact
surface. Fig. 4-2B shows a gradient in the amount of honey calcite present: there exists
more cement along the outer parts of the sample compared to the inner parts. Petro-
graphic analysis of this thin section made along this plane show the characteristic pendant
morphology of the honey calcite, especially when infilling larger primary pore spaces in the
outer part of the floret tufa (Figure 4-2J-K). Fine laminations of a few tens of microme-
ters thick are also discernible, in which layers are bounded by a fine layer of dark micritic
material. Despite the laminations, the bladed crystals comprising the honey calcite appear
to grow across multiple laminae (Figure 4-2K).
For (2), along a road cut exposing colluvium at the southern margin of the lake basin
(4243.0 ± 0.9 m elevation; Figure 4-3), we find honey calcite cements up to 1 cm in thickness
predominantly coating the undersides of cracks. The cements are only found within a
resistant belt of outcrop that is laterally continuous at roughly the same elevation for ~100
m, though their own presence throughout this resistant layer is not uniform throughout.
Like that observed within void spaces of tufa, the honey calcite here is laminated and hangs
as smooth and broad overlapping pendants from the roof of cracks. Where the cements do
not entirely fill spaces within cracks, a thin <1-mm-thick veneer of opaque white carbonate
coats the bottom of these cements where exposed to air.
The honey calcite also occurs as sub-horizontal discontinuous "sheets" upon and incor-
porating detrital sediment. In one location at 4242.7 ± 0.9 m elevation, the honey calcite
fully infiltrates sediment within a spit-like formation to form sub-horizontal layers of honey
calcite-supported beach sands. In another location, the honey calcite forms around lenses
of silt-sized particulate sediments, but leaves fenestral pore spaces between overlapping
sub-horizontal sheets.
Another key observation is that, in hand sample, the honey calcite appears to have three
packages differentiable by color, in which the oldest package is darkest yellow in color and
149
Figure 4-1: Field photographs showing that, in the context of shoreline tufas, the honey calcite
appears thickest along the sides of cone-shaped tufas where primary gaps and spaces between cones
exist. [A] Cone-shaped tufas on top of encrusting floret tufa on a volcanic boulder at 4233.9 ± 1.5
m elevation. Panels [B], [C], and [D] are close-up photographs of locations wherethicker deposits
of honey calcite are found, indicated by ovals of black dashed-lines and arrows. Red pen is 15.5 cm
in length; pen cap is 6.5 cm.
150
Figure 4-2: (Caption and Panels J and K on the following page.)
151
Figure 4-2: An examination of the two distinct phases of carbonate formation in Agua Caliente
I, as seen in sample AD10-225 (ID #1 and #6). [A] Photograph showing the field context of the
sample. AD10-225 was attached to a volcanic boulder in the area indicated by the shaded region
outlined in white. The dashed black line indicates the orientation of the planar surface pictured
in [B] and the thin section in [C]. The black orientation arrow is pointing radially away from the
tufa-boulder contact in the direction of growth. This orientation arrow appears in all other panels
of this figure. Red pen is 15.5 cm in length. [B] Flat slabbed surface of AD10-225 with U.S. penny
(~1.9 cm in diameter) for scale, showing two distinct carbonate phases: a dense, honey-colored,
translucent calcite, and a porous, opaque white material. The box represents the approximate area
made into a thin section (40 mm x 60 mm). [C] Close-up of the sample embedded in resin, showing
the location of areas featured in remaining thin section panels. The thin section plane is oriented
parallel to the growth direction. Note that in this panel and all others, gravity is oriented parallel
to the vector going into and out of the page. [D] Thin section photomicrograph in plane polarized
light showing that the white porous carbonate phase consists of dark, branching, microbial peloids
(1) surrounded by micrite containing thin, growth-oriented filaments and dark, lenticular tube-like
cyanobacterial-algal microfossils (2). The honey calcite phase is most easily observed filling or
coating primary pore space (3). [E] Photomicrograph in plane polarized light showing trapped
detritus in primary pore space (1). Trapped materials include skeletal fragments of ostracods (2),
microbial peloids, and possibly carbonate fecal pellets. [F] Photomicrograph in plane polarized
light. The honey calcite possesses fine-scale layering and most obviously fills primary pore space
(1), but it also infiltrates the previously-deposited white porous carbonate phase (2). There are
few places where the white porous phase is untouched by the honey calcite (3). [G] and [H]
Photomicrographs in cross polarized light showing good preservation of ostracod shells (1), which
have been infilled by honey calcite. The honey calcite frequently exhibits an undulose extinction
pattern (2). [J] and [B] Photomicrographs in plane and cross polarized light, respectively, showing
that the honey calcite exhibits pendant-like morphology in areas where it fills primary pore space.
The morphology suggests that the honey calcite phase is a gravitational cement. -Thin, micrometer
thick layers are separated by bands of thin, dark-colored micritic material.
152
Figure 4-3: [A] Image of sample location AD10-233, where a road cut exposes an outcrop of
honey calcite cementing colluvium made of ignimbritic and igneous materials. [B] Closer-up
view of the outcrop with a Sharpie marker for scale.
153
the youngest package is a translucent white color. These three differently-colored packages
of honey calcite are observable elsewhere in the basin, such as within tufas. On top of the
outermost layer, a thin veneer of opaque white carbonate can be found.
4.3 Stratigraphic coherence of U-Th geochronological results
The preparation and chemical procedure for U-Th dating of these deposits is described in
previous chapters. We drilled powders from the tops and bottoms (start and end) of the
three distinct honey calcite packages across multiple samples from both the road-cut out-
crop as well as from a thick honey calcite deposit found along the outside of a cone-shaped
tufa. The initial 230 Th/ 232 Th ratio used to correct the data for detrital contamination was
232
determined by calculating dates for a broad range of initial 230Th/ Th ratio values and
deciding which ratio brought the U-Th data into stratigraphic coherence. After running
these tests, a initial 230Th/ 232Th ratio of 4.5 ± 3.0 ppm atomic was chosen.
Fig. 4-4 shows an example of U-Th data from a sample from the road-cut outcrop.
Here, the dates fall in stratigraphic order.
Fig. 4-5 combines all U-Th dates from honey calcite in Agua Caliente I and displays
them as "camel" plots, or probability density functions. With some exceptions, the vast
majority of the U-Th dates from the tops and bottoms of the three honey calcite packages
are consistent with the constraints of stratigraphic order. When comparing the mean ages
of dates grouped together by coevality (same layer), the data indicate that the depositional
period of the honey calcite occurs at the end of HE1 for ~1000 years.
We hypothesize that the good behavior of U-Th dates owes itself to the purity of the
honey calcite; the cement contains few inclusions and peloidal grains. Our observations of
the extent of honey calcite infiltration from thin section also explains the lack of repro-
ducible data from the encrusting floret tufas: the honey calcite has penetrated the original
underlying tufa to a degree that makes isolating the floret tufa phase on its own practically
impossible (Figs. 4-2F, 4-6).
4.3.1 Origins of the honey calcite cement
Our current working hypothesis for the origin of the honey calcite cement is that it is
an upper-littoral "spray-zone" deposit, precipitating from lake waters that intermittently
spray over the tufa formations, beach sands, and gravels. This hypothesis is consistent
154
Figure 4-4: U-Th dates from sample AD10-233-10. Locations of drilling are indicated with
red lines and are annotated by alphabetical letter in order of stratigraphically oldest to
youngest (i.e., A is the oldest layer drilled).
155
Mean Ages
(kyrs BP)
start( base)
I sat15.8 ±0.4
end (top) end 15.7 ±0.2
4-J
start start 15.5 ±0.2
C
end 2-a range end 15.3 0.1
sart 15.2 ±0.1
15.0 ±0.2
end j end 14.9 ± 0.2
14.6 14.8 15.0 15.2 15.4 15.6 15.8 16.0 16.2 16.4
U-Th date (cal kyrs BP)
Figure 4-5: Camel plot diagram of all U-Th dates from honey calcite at sample sites
AD1O-233 and CYC15-019. Each date is represented by a circle color-coded by honey
calcite package (see 'Marker Legend' on the right). Top two rows show data from the start
and end (top and bottom) of the innermost, stratigraphically oldest honey calcite package;
middle two rows show data from the start and end of the middle honey calcite package; and
the bottom three rows show data from the start, middle, and end of the outermost honey
calcite package. Error bars represent the uncertainty of each date (2-a range). Each date
is then represented by a probability density function, in light gray. The dark gray curve
represents the cumulative probability density function of dates in each row. The mean and
standard deviation of dates in each row is listed to the'right of each plot.
156
YD HE1
yD HEl I I I26
.
24 If loret w/ honey calcitehoney calcite
P platey carbonate
E
22 - transformed ikaite
2-sigmauncertainty
20
E mixtureso f
honey calcite and
181-- - o-th-e nd mme
of unknown age
16 -
U
-- -* - ~ -- ~- ~~14 F- ---------minimumage
of floret tufa
12
12 14 16 18 20 22 24
U/Th dates (cal kyr BP)
Figure 4-6: Comparison of U-Th dates of different tufa and carbonate deposits at different
elevations in Agua Caliente I.
157
with the following observations:
1. The smooth laminations, lack of inclusions and peloidal particles, and positive 1 3 C
values indicate that the honey calcite was physio-chemically deposited, rather than
biologically formed. Inorganic precipitation could have been assisted by evaporation
of spray waters, raising the carbonate saturation state of the fluid to encourage
precipitation of the cement.
2. The pendant morphology observed in both the floret tufa and the road cut of exposed
colluvium suggests that the honey calcite is a gravitational cement reflecting vadose
conditions (Muller, 1971; James and Choquette, 1984). Such pendant, microstalactic-
like morphologies are observed in Holocene beach sediments and ancient limestones
elsewhere, and are interpreted as indicators of supratidal environments, areas where
seawater regularly splashes but does not submerge (e.g., Inden et al., 1996; Schoelle
and Ulmer-Schoelle, 2003).
3. If the honey calcite were a spray deposit, we would expect the cement to be thickest
where spray waters would accumulate most, on the outer tops'and exterior sides of
surfaces. We observe that the honey calcite cement more completely fills the pore
spaces in the outer part of floret tufas (Figure 4-2B-C), leaving open primary pore
space filled with trapped detritus in the inner part of the tufa (Figure 4-2E). The
honey calcite also is also thickest along the sides of cones (Figure 4-1).
4. Both the positive 6180 value and the ubiquitous presence of the honey calcite through-
out the lake basin strongly point towards a lacustrine origin for this deposit. The
5 234Uinitiav alue of the honey calcite is very similar to that of other lacustrine de-
posits in the basin, suggesting that the source of the honey calcite cannot have been
purely groundwater.
5. Good preservation of delicate ostracod remains, which have also been infiltrated by
the honey calcite cement (Figure 4-2G-H), and the lack of dissolution textures at
the contact between separate phases are strong indications that the honey calcite
cement is syn-sedimentary cement, rather than a cement formed via dissolution and
reprecipitation of the original encrusting floret or cone-shaped tufas.
6. Some evidence suggests that the honey calcite cements are greater in percent abun-
dance at lower elevations than at higher elevations. Figure 4-2F shows that the
honeyealcite does not merely coatthesurface of the floret tufa, but lso Idn-rates
158
it, causing the original white porous material to take on a more beige-colored appear-
ance. Samples collected at lower elevations tend to be more completely infiltrated by
the honey calcite cement, with few white-colored areas. More field observations are
needed before this observation can be confirmed.
7. The abrasion platforms on volcanic bedrock hillsides indicate that significant wave-
eroding action must have been present at some point in the past, despite the small
fetch length of the lake.
Such gravitational cements have also been described in beachrock settings in the marine
environment, which may be materials that have yet to be taken advantage of in terms of
geochronological control. Reconstructions of past sea level could benefit from finding a
new type of material for dating.
159
Figure 4-7: 5180 data from sample AD10-233-10. Image on left shows the location of drill
holes; plot on right compares the 5180 values in stratigraphic order. Note that the vertical
axis does not represent depth quantitatively. Annotations on plot to the right represent
the current working hypothesis for the behavior of lake levels at the time of honey calcite
formation.
160
Chapter 5
U-Th dating of tufas from the
Miscanti-Miniques-Pampa Varela
lake system, northern Chile
5.1 Geologic Setting
Lagunas Miscanti and Mifniques are two high-altitude (-4000-4200 meters above sea level)
permanent shallow lakes located in the Altiplano-Puna plateau of northern Chile (Fig. 5-
1). Together with Pampa Varela, these basins fall immediately to the east of a major
northeast-striking fault that separates the Andean cordillera from the lower elevation pre-
cordillera to the west. The basin lithology consists primarily of ignimbrite and other
clastic igneous rocks associated with the now dormant volcanoes that share the same name
as these two water bodies below them (Fig. 5-1B). In the present day, the water budget
is predominantly driven by groundwater flow from a large catchment area (~320 km 2 )
and evaporation (Valero-Garces et al., 1996). Today, this lake system is part of the Los
Flamencos National Reserve and is maintained by the Indigenous Atacamefios people in
Socaire in partnership with the National Forest Corporation of Chile (CONAF).
Previous work on lake sediments and tufa deposits in the basin has indicated that
these basins experienced higher lake levels at some time during the last deglaciation and
Holocene (Grosjean et al., 1995; Valero-Garc6s et al., 1996, 1999; Grosjean et al., 2001). In
a sediment record from Laguna Miscanti spanning the last -22 ka (Grosjean et al., 2001),
two periods of higher lake levels during the last deglaciation are inferred from aquatic
161
70°W 650W
Salard e
Uyuni Miscanti i
Lcoo
4d
2000
2500
3000 oPampa AA.
3500 Varela 5mhigh
4000i c---eshorane
4250 3 kkmtoverflow4750 
o tufasampless s
Figure 5-1: [A) Overview map of the Altiplano-Puna plateau in the central Andes. White
rectangle marks the location of the Miscanti-Mifiques-Pampa Varela (MMPV) lake system.
[B] Satellite imagery of the MMPV lake system. Blue outlines are contours of 
elevation
that approximate the high shoreline in each sub-basin. Contour outline were generated
using the SRTM DEM. Dark blue dashed lines with arrows trace the pathway of overflow
in between each sub-basin. Yellow circles mark the locations of tufa samples discussed in
this paper. This site is approximately ~10 kmn from the town of Socaire in northern Chile.
pollen assemblages and sediment lithology. Unfortunately, interpretation of these records
was hampered from poor chronological control due to a reservoir effect of unconstrained
magnitude on radiocarbon dates. Since terrestrial material suitable for radiocarbon dating
was non-existent in the core, radiocarbon dates were made on aquatic organic matter
1 2
and bulk carbonates formed in waters with 
14 C/ C ratios that may have been out of
equilibrium with that of the atmosphere. This is especially a likely scenario in the active
162
volcanic setting of the central Andes, in which volcanic CO 2 is present in groundwaters.
Slow recharge times are also typical in modern lakes in this area and contribute to the
reservoir effect. Modern measurements from the nearby Laguna Lejia suggest that the
modern reservoir effect is ~2000 years for aquatic organic matter and as high as- 8000
years for carbonates (Grosjean et al., 1995).
In this chapter, we continue our exploration of the link between Heinrich Events and the
hydroclimate of the Altiplano-Puna plateau of the Central Andes by U-Th dating lacustrine
tufa deposits in the Miscanti-Mihiques-Pampa Varela (MMPV) lake basin system. Vari-
ous studies in this region-including work on carbonate-encrusted paleoshorelines showing
lake expansions (Bills et al., 1994; Sylvestre et al., 1999; Placzek et al., 2006a, 2006b;
Placzek et al., 2009; Blard et al., 2011; Placzek et al., 2013; Chen et al., unpublished data),
cave stalagmites indicating increased rainfall amounts (Kanner et al., 2012), paleowetland
deposits showing elevated groundwater tables (Quade et al., 2008), fluvial terraces and ro-
dent middens showing increased stream discharge (Latorre et al., 2006; Nester et al., 2007;
Gayo et al., 2012), and glacial moraines showing substantial ice cover expansion (Smith
and Rodbell, 2010)-all indicate that this region experienced the wettest conditions of the
last deglaciation during Heinrich Event 1 (-15-18 ka). The regional synchroneity of these
wet phases is so prominent that the local phenomenon is often revered to as a "Central
Andean Pluvial Event" (CAPE).
Because Miscanti and Miniques overflow into the Pampa Varela basin upon breaching
the elevations of their sills, the timing of the lake creating the high shoreline in Pampa
Varela represents the hydrologic maximum of the drainage basin.
5.2 Methods
We mapped paleoshorelines and collected carbonate samples over two days in May 2016.
We used a Trimble GEO 7x handheld receiver to collect precise location and elevation data
of shoreline features and tufa samples, which yields sub-meter accuracy after differential
correction using proprietary software.
Carbonate samples were then prepared for U-Th dating following the same procedures
described in the previous chapter. We used an initial 230Th/ 23 ?Th ratio of 4.5 ± 3.0 ppm
atomic for the calculations of dates. Petrographic images of thin sections were made using
a Zeiss AX10 microscope.
Several powders drilled from tufas were also analyzed for their mineralogy using a Nico-
163
let iS50 attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectrometer
at the Center for Nanoscale Systems at Harvard University. To test the reliability of the
results, we prepared three mixtures consisting of known proportions of clean calcite and
aragonite standards and determined if the measurements could accurately reconstruct these
relative proportions. Measurements of these mixed standards were all within 5% of true
proportions of calcite and aragonite.
5.3 Results
5.3.1 Geologic and geomorphic context of tufa deposits
Due to the short time on site, we were unable to survey the basins and their tufa deposits
comprehensively, and thus recognize that this work is limited by our sampling. Regardless,
we report our findings on the most conspicuous deposits, generally located on the northern
parts of each basin (Fig. 5-1B, yellow circles). This spatial distribution may be related to
pathways of groundwater flow and spring discharge: in both Miscanti and Miiques, warm
spring waters are observed seeping into the modern lake along the northern shoreline.
In Miscanti, the high shoreline in the north is mainly expressed as a salient wave-cut
terrace eroded into igneous bedrock (Fig. 5-2A). The promontory in the north-central part
of the coastline consists of a broad platform with outcrops of smoothed basalt boulders on
its lakeward edge. These boulders act as the substrate for a thick (~20-35 cm) and laterally
continuous (200 in) deposit of tufa (Fig. 5-2B) that runs parallel to the paleoshoreline.
Although much of this tufa remains in situ (Fig. 5-2C), it is very porous and friable; as a
result, many fragments of tufa are found in float in the surrounding area. The extent of
this re-deposition of tufa is illustrated by the color contrast between the eroded pink-beige
tufa and dark purple-brown basalt, which is easily observable in both the field (Fig. 5-2B,
see 'tufa in float') and satellite imagery.
In contrast, the smaller Mifniques basin does not have any broad platforms. Here,
tufa deposits are found encrusting basalt boulders along the more steeply sloping bedrock
hillsides (Fig. 5-2D). In Pampa Varela, the only instances of in situ tufa that we encountered
were found encrusting boulders in protected spaces (e.g., undersides of boulders).
In all three basins, pebble- to cobble-sized fragments of tufa were found redeposited
at the crests of the constructional gravel berms at the high and intermediate shorelines.
In fact, in the northeastern corner of Mifiques, an cross sectional exposure reveals that a
164
Figure 5-2: Field photographs, of paleoshorelines and tufa deposits. [A] In the northwest embay-
ment of Miscanti, the high shoreline is expressed as a distinct wave cut terrace eroded into basalt
bedrock(dGPS elev.= 4164±1.0 m) and a constructional gravel berm. Photo taken looking east.
[B] The promontory in the north-central coastline of Miscanti. The outer, lakeward edge of this
platform has outcrops of smooth basalt boulders that are coated with a thick (~20-30 cm) and
laterally continuous (200. m) deposit of tufa (see 'tufa in place'); pieces of this tufa can be found
on the landward, backshore side of this platform in float. The extent of this redeposition of tufa
is illustrated by the contrasting colors of the pink-beige tufa and the dark purple-brown of the
basalt (see 'tufa in float'). Photo taken looking southeast. [C] Close up view of the thick tufa
deposit coating bedrock featured in Panel B as 'tufa in place'. Note the same bedrock headland in
the backgroundsgof Panel B and C. [D] In Mifiques, tufa is found on top of more steeply-sloping
hillsides of bedrock boulders.
165
~10-15 cm thick layer made entirely of eroded tufa fragments lies at the crest of the high
shoreline gravel barrier. In many instances, the hydrodynamical differences between tufa
fragments and other clastic igneous rocks created a 'highlighting' effect, whereby geometry
of berms was accented in color by the spatial distribution of the tufa fragments. These
observations indicate that there was at least one occurrence of higher lake levels after the
original formation of the tufa deposits.
5.3.2 Characteristics of various tufa facies
The tufas in the MMPV lake basin system can be broadly categorized into two categories:
(1) fibrous mats and (2) carbonate cement encrustations. We now describe these two facies
categories in Figs. 5-3-5-5.
The fibrous mat tufa facies consists of fine and elongate calcite crystals that are arranged
radially relative to the surface of the substrate (Figs. 5-3B and C). The thick, continuous
deposit of porous tufa described in Fig. 5-2B and C consists primarily of this facies. These
tufas are not well-lithified and are highly weathered, with many exhibiting clear signs of
recrystallization and/or diagenesis (Fig. 5-3B, white arrows). In exposed cross sections
of outcrop, different beds of fibrous tufa with slightly varied macro-scale morphology are
discernible, with some exhibiting a more splayed, feather-like fabric and others exhibiting
a more classic radial growth fabric. On a sub-centimeter scale, banding can be observed
within the tufas and appears to be related to porosity. In plan view, weathered surfaces
exhibit closely packed polygonal shapes, suggesting that the internal structure of these
tufas consists of a composite of smaller inverted cone shapes of varying sizes. In thin
section, abundant diatoms are observed oriented parallel to the fabric of these tufas.
The cements are the second most common form of carbonate deposition in the tufa
deposits. These -1-3 mm thick cements are generally isopachous, finely laminated and
are found indiscriminately coating various substrates. They are most well-preserved when
coating cobbles of clastic igneous rocks (Fig. 5-4). The coatings vary in color from being
beige to mostly translucent white.
Another common substrate of the carbonate cements are -the remains of charophyte
algae. Many tufas found in situ in Miiques exhibited excellent structural preservation
of macroscopic charophyte algae of the order Charales, commonly known as "stoneworts"
(Fig. 5-5). Charales are morphologically complex filamentous green algae and are found
in shallow, calm, fresh and brackish waters worldwide (Wood and Imahori, 1959; Bold
166
setina vewofa amleofthfbrusma facie (C C607,fon infat.P k
shapes~ ~ ~ ~~~~A ind1at th4rliglctin*fpwe sd o -hdtn.Anoae ae r
in nit o yefsbefre195. Nranisae2arn .Whtaroslbed''m k
ares hwic ae eenrecysallz, whcarthexodsufesftesmp.Arw
reprsefn s rtaht -idreciorn apic p [C Clse p viw o intrna stuctne o th
sample.
sectionalvieofasampleofhefibrousmafabrosCC60matunifoa)Pn
samle.B
Figres5-3:Tshedfirtosraticarboaphifcies.[C]COtrpooloseuppviewofnenltutroftheo
167
and Wyynne, 1978). Living charophytes are found at the sediment-water interface today
(Fig. 5-5A). Finding a direct modern analogue of fossil tufas is very uncommon; if the
depth constraints on the modern occurrence of charophytes could be determined, this
information would allow for better interpretations of the paleolake level represented by the
tufa deposits.
A carbonate-encrusted sample of charophyte algae was found in float in Miscanti, but
we were not able to locate the corresponding in situ outcrop in this basin.
5.3.3 Results of U-Th dates on tufa deposits
The uranium concentrations of these deposits were between 10 and 300 ppm, -2-3 orders of
magnitude higher than concentrations observed in tufa deposits from the Bolivian Altiplano
(Placzek et al., 2006a) and elsewhere in northern Chile (Chen et al., unpublished). In
several instances, tufas are comprised of both the fibrous mat facies and carbonate cements.
In these situations, we are able to leverage the incontrovertible constraints of cross-cutting
relationships and stratigraphic order to test the viability of our U-Th dates. For example,
the cement in sample CYC16-025A featured in Fig. 5-5D has clearly formed around both
the charophyte algae as well as the fibrous mat (see white triangle labeled '3'), and thereby
A YCI2A 9,550 
190
11,000 i 00
cobble
cement coating fibrous mat o
isopachous cement e 11,110 130 10f780 i110
Figure 5-4: The cement encrustations facies. [A] Hand sample consisting of largepebbles
and cobbles that are coated in a- 1-3 mm thick rind of isopachous carbonate cement. [B]
Exposing a cross section of the sample featured in Panel.A (CYC16-029AA) reveals that
one cobble is made not of igneous rock, but rather a fibrous mat tufa. The cement has
infiltrated the pore spaces of the fibrous mat tufa cobble.
168
11,800±2200
D 11,440 80 11,310 90 0 fibrous mat
11,600 200 W Isopachous cement
C A 11,800±200 10,000±300
Figure 5-5: Preservation of charophyte algae by carbonate cement. [A] Modern living charophytes
living a5the sediment-water interface, extracted during sediment coring from Miniques in April
2013. Photo taken by Matias Frugone. [B] Pristine charophyte preservation by carbonate encrus-
tation on the hiliside in the Miniques basin featured in Panel Dof Fig. 5-2.[0] Petrographic thin
section image of acarbonate-encrusted charophyte tufa sample. 1: Isopachous, radial calcite cement
with many inclusions. The area immediately surrounding the original charophyte is darker due to
a higher proportion of dark microbial peloids. 2: Holes indicate locations where the charophyte
used to be. 3:This area iseithethehprimarypreservationofthe original carbonate skeleton of
the charophyte, or asecondary infilling. The carbonate here contains an abundance of diatoms.
[D] Hand sample of acarbonate-encrusted charophyte algae associated witha ~2cm thick fibrous
mat tufa. 1: An instanceinthis sample of awell-preserved charophyte stem. 2: Fine layers in
the fibrous mat tufa. 3: The isopachous cement is coating the fibrous mat, indicating that the
formation ofthecement occurred after the formation of the fibrous mat.
169 :
constrains its formation to be after that of the fibrous mat. For the cements, we are also
able to test the reproducibility of coeval layers within a sample.
Fig. 5-3B shows analyses from two separate layers in a fibrous mat tufa sample. Here,
we observe that the dates within each layer are not reproducible-as in, the analytical
uncertainties calculated from measurements do not overlap with one another. We also
notice that the dates from the stratigraphically older layer are older than the dates from the
stratigraphically younger layer. Note that while the existence of layers implies a sequence
in time, because each layer in these fibrous mat tufa samples is delineated by a change in
porosity rather than any layer representing a depositional hiatus, it is possible that the
entire thickness of the sample formed within a narrow window rather than over a long
enough period of time to allow for U-Th dates to differentiate between the top and bottle
of the sample.
Fig. 5-4B shows analyses from carbonate cements coating cobbles of igneous rock and
one fibrous mat tufa. The uniform color and thickness of the cement surrounding these
cobbles suggests that all the cement in this hand sample formed at the same time. However,
the two analyses of the cement from different locations in the sample yield dates that differ
by -1.5 kyrs and do not overlap in uncertainty. In addition, one of the dates on the
fibrous mat tufa cobble are inconsistent with stratigraphic order constraints. The latter
observation could be explained by the fact that the cement has infiltrated most of the pore
spaces in the fibrous mat tufa, so much so that attempts to isolate powder from the original
fibrous mat cobble is not practically feasible, similar to the effect of the honey calcite in
Agua Caliente I.
Fig. 5-5 shows analyses from a sample that exhibits cross-cutting relationships between
fibrous mat tufa and cements encrusting a charophyte algae. Here again, the two dates
on the isopachous cement do not overlap in uncertainty. Analyses along the uppermost
part of the fibrous mat tufa are all within the same ~500 year range but some analytical
uncertainties of analyses do not overlap.
These results can be explained a number of different ways: (1) It is possible that the
formation of the cement occurred very slowly, such that small differences in'the location
of drilling can lead to large differences in calculated dates; (2) the-initial 230Th/ 232Th
ratio used for the calculation of dates does not adequately account for the true amount
of initial 2 3 0Th in these samples; (3) there exists hydrogenous Th in the lake system that
was incorporated into the tufa at the time of formation; and/or (4) these tufas have not
remained closed systems with respect to uranium.
170
Some data from fibrous mat tufas in both Miscanti and Pampa Varela support the hy-
pothesis for open system behavior (Fig. 5-6). In sample CYC16-018A, powders were drilled
from the bottom and top of the sample in an attempt, originally, to determine the dura-
tion of tufa formation, especially given that this sample formed directly on top of bedrock.
Again, the analyses along seemingly coeval locations yield dates that do not overlap in
uncertainty. However, we notice that dates of powders drilled from portions that appear
altered (recrystallized) have elevated 6 234 Uinitial values and low uranium concentrations
(Fig. 5-6B), a pattern that is consistent with uranium loss.
We also observe elevated 2 3 4Uinitial values in analyses from the outermost part of the
sample CYC16-038A from Pampa Varela (Fig. 5-6C and D). These analyses yielded dates
that were old and inconsistent with stratigraphic order constraints.
Thus, we have good evidence to suggest that these analyses shown in gray in Fig. 5-
6 have experienced uranium loss and should not be considered data that represent any
geologically meaningful event. However, there are still inconsistencies in the remaining
data, primarily the lack of reproducibility of dates from coeval layers and lack of adherence
to the constraints imposed by stratigraphic order. Here, dates from coeval layers span a
range of -5-10 kyrs or more, as opposed to data previously shown in Figs. 5-3-5-5, which
only span a range of <2 kyrs.
Fig. 5-7 plots all U-Th analyses from tufa samples in the MMPV lake system. The
scatter of dates of samples of the fibrous mat facies in Pampa Varela makes it difficult
to make many interpretations on the timing of tufa formation in this basin. We are only
able to say with confidence that a lake of deglacial age once existed in Pampa Varela. In
contrast, there is more coherence in data from Miscanti and Mifiiques: We notice that the
fibrous mats in both basins generally occupy the same ~3-kyr-long time range. There also
exists dates of carbonate cement encrusted charophytes in both basins between 16-13
kyr ago. Other carbonate cements yield dates that are consistent with those from fibrous
mat tufas, with some dates being younger than the youngest fibrous mat data, consistent
with stratigraphic relationships.
Thus, there is evidence from U-Th dating that there existed a higher lake levels from
16-9 ka, broadly consistent with the CAPE I and CAPE II intervals. However, time
intervals in this plot with no data do not necessarily prove that higher lake levels did not
exist during these times: again, we are limited by our sampling bias. In other words, the
absence of data does not prove the absence of a lake.
Future work (imminently) will compare these results to past data from the Miscanti
171
A cYC648A * fibrous
" recrystallized 370
360
- 350
340
330
0 50 100 150 200 251
U concentration (ppm)
c CYC16-038A 121,800 i 800
126;200 t 80
4
8,280 ±30 520|-
32,220 160 12,000 ± 50
3 1;6 13,370 ± 50 480
16,920 ±110 18,370 ±90h---19,210 ±90
17,850 ±100 20,870±90 440.
2 0
16,540 ± 120 k 400-18,760 ± 1001 outermost layer 1 0 3
1 e otheranalyses 360'
16,200 ± 400 10 20 30 40 50 60 
70 80
17,500 ± 300 U concentration 
(ppm)
Figure 5-6: Evidence of open system behavior from analyses of recrystallized portions
of tufa [A, gray] and the outermost layer of porous tufa [C, gray]. All dates are in units
of years before 1950. Dates listed on the left side of Panel C correspond to the layers
annotated on the sample (1-4, 1 being stratigraphically the oldest). [B] and [D] Com-
parison of 8 234 Uinitial and uranium concentration of samples featured in Panels A and C,
respectively. Note that there are analyses featured in these panels that are not shown in
the sample images; additional analyses come from different cross-sectional exposures of the
same sample. In Panel D, the groups labeled 1-3 correspond approximately to analyses
from layers 1-3 in Panel C.
sediment core and attempt to provide another estimate of the reservoir correction of this
lake.
172
0fibrous mat 0 cement 0 cemented charophyte * other 0 altered analyses
4170 --- . -------------- -
4160 - -
4150 - 0
4140
4160 --- ---------------------------------------------------------- - - - ---------- - - --- -- -- ------
440150 - - - - - - - -
4140 -
4130
4060 ---- - - - - - - - - ---------------- - - - - - - - - - - -
E 4040-
r4020-
j4000 -
S3980 -0
8 13960
8 10 12 14 16 18 20 22 24
U-Th Date (kyrs BP)
Figure 5-7: Summary of U-Th dates on tufas from Miscanti (top), Miiques (middle),
and Pampa Varela (bottom), plotted against elevation of sample location. Dashed blue
line marks the elevation of the highest shoreline in each basin. Solid brown line marks the
elevation of the modern lake or lowest point in the basin. Errors bars are 2-. See legend
above all plots.for symbology.
173
174
Appendix A
Appendix Tables
175
Table A.1: U-Th data associated with Chapters 3-5. Reported errors for 238 U and 
23 2 Th concentra-
tions are estimated to be± 1% due to uncertainties in spike concentration; analytical uncertainties
are smaller. 8234 U= ([ 2 4 U/ 23 83 Ujactivity - 1) x 1000. Ages are corrected for detrital 
23 0 Th as-
suming an initial 23 0Th/ 2 32Th of 4.5 ± 3.0 ppm atomic. 2 3 4 Uinitial corrected was calculated based
on 2Th age. B.P. stands for "Before Present" where present is defined as January 1, 1950 in the
23 0 23 4 U are from ?; decay constant for 23 8Common Era. Decay constants for Th and U is 1.55125
x 1010 yr-1 (Jaffey et al., 1971).
4lt01 0/ 230/ Age 4234 0 Ag e
14~sam Sapest attll(' intuer UMamlelae23U ge 23 tel434 3ie)U38) 1h232 m {2n ) 64 g~ 10)
ag/s parel tiM"y M pndN  Pps/ ato46
AP~ACIlI AD09100 JIN NON A009-100(A) 1264 24 22202 23 1268.4 6.2 03297 0.0077 293.2 4.7 16010 420 1326.2 6.7 13500 400
AD09100 ftN NaN A009-100(6) 7030 140 76200 1500 1225.3 1.3 03292 0.0011 482.5 2.1 17179 ,61. 1289.4 
1.4 17030 110
AD09101 NN NaN A009-101(A) 254.1 78 111.9 8.2 1226.1 8.9 0.315 0.012 229.8 1.2 16390 670 1253.1 1.7 16100 700
AusCaetI AD06-90 -2306339 4740797 A00945() 2425 31 11025 11 1219.0 5.1 0.3081 0.0018 1140 _13 11780 320 1317 1.4 11700 300
AguaCial4Il AD0945 23.0339 47.40797 A09-9(C) 8070 160 17020 350 1214 4.2 0.337 0.009 2543 13 17469 62 1334.3 4.4 17440 60
Agu Cientel AD0943 .23.08339 47.40797 AD09-91D) 4678 94 2460 230 1260.4 2.4 0,3045400013 9200 910 15189 73 1317.1 2.6 13330 70AsaCialnIl
AguCaItag AD0945 43.08339 4740797 A0094E) 33S3 67 3970 260 1243.7 1.6 0.3123 0.0014 2830 110 16128 80 1301.3 1.9 16100 80
ADgguusaCc,lae1ttal
AD09-96 Nn NAN AD09-96( 2015 13 12172 12 1189.5 3.1 0,2769 0027 764. 6.1 14370 150 1239.1 3.3 14490 160
A00946 NOR NaN AD0996(A) 8070 160 1161093200 1167.1 1.6 0.2594 0.0063 28 21 13750 350 1213 2.1 13300 400
AguaCaNntal A009-46 NoN NaN ADO96(6} 10270 210 48300 9300 1179.8 2.6 0.2732 0.006 820 16" 14430 340 1228.7 2.9 14400 300
AguLIACalinte I AD09498 -3.08400 47.40674 AD09-98A 4820 96 67200 2100 1216 1.3 0.3192 0.0032 384 11 10730 290 1274.1 1.9 16500 300
AguaCalental A09.98 -23.08400 47.40674 AD09-98)) 3965 79 8420 860 12193 1.6 0.2157 0.006 2140 220 146150 330 1271.5 2 14800 300
Agua galientlI 
AD0948 43.08400 47.40674 AD09-4(C) 4263 53 11190 390 1227.9 2.5 0.2959 0.0012 1790 49 15342 69 122.1 1.6 13310 70
20340 100
AguoCaontl AD09- -2308400 47.40674 A009-93)) 9130 130 5600 1200 1193 1.9 0.343 0.0013 950.7 5.9 20629 78 1266.4 2.1
AguaCalteIl AD10423 -23.06430 4740108 A010225(A) 3183 19 6784.7 3.3 1272.8 2.7 0.3038 0.0025 2401 22 15450 140 1329.4 2.9
19420 140
AD102-5 -2308430 47.40108 AD10-225()8 14160230 13400 8600 1227.1 6.9 0.4405 6.033 7400 400 23350 320 1311.4 7.1 23100 300
AguaciueteI
AcuaCtoalt I AD10-225 43.08430 47.40100 A0-223) 7870 160 27190 350 1137.3 4.1 0.361 0.0014 1659 5.9 18359 89 1253.9 4.3 19310 90
AgaCalienta A010-225 -23.09430 -67.40108 AD1-2236)A) 5010 100 129700 2600 1152.7 2 0.3231 0.0014 198-2 0.81 17485 82 1209.6 2.2 170O 300
AgusCalentlI AD10-225 -23.09430 47.40108 A010-2253(B) 6000 120 64100 1300 11593 1.7 0.3442 0.0014 111 .2.1 14632 84 1221.4 1.8 18100 130
AD10-225 -23.06430 47.40108 AD10-225(C) SO 120 110300 2200 1135A 2.1 .0.4063 0.0016 354.3 1.3 22364 29 1229.8 2.3 22110 200
AguaCalleta AD10421 43.08430 47.40108 AD10-225(D 5630 110 34150 690 1159.3 1.6 0.3236 0.0014 847.2 3.3 17454 83 1297.6 1.7 17370 100
Agusa~lentel A010-22 -23.08430 47.40108 AD10-225(E) 5770 120 75400 1600 1168.6 1.6 *0.3077 0.0012 359.5 1.3 15492 69 1220.5 1.0 16300 140
AD10426 23.08276 47.40104 AD10-226(A} 2059 35 10718 9.2 1193.5 3.1 0.2691 0.0023 864 7.1 14110 350 12417 33 14040 160
AocWiat.I
AAgg0u4aCaaMlleinnttetl AD10226 -2309276 47.40104 AD10-226(A} 2067 41 15500 1300 1169.7 8.1 0.243 0.02 313 65 12800 1100 1212.4 9.2 12700 1100
AguaCallent I A10-226 43.08276 47.40104 AD10-226(B) 2307 46 20240 410 1160.3 2.3 0.3949 0.0011 714.7 2.8 21627 68 1232.9 23 21500 110
AD 226 -23.08276 47.40104 AD10-226SA6 2758 35 3410 690 0161.2 3.4 0.3247 0.0014 416.4 1.9 17504 83 1219.4 3.6 17330 140
Agua CallentelI AD10-226 -23.09276 47.40104 AD10-226A3l(B) 3194 64 31410 640 1170.6 7.1 0.254 0.0014 460.7 2.6 13181 94 1221.4 7.4 13010 130
AD0-226 -2309276 47.40104 AD10-226A(C) 3362 67 16470 340 118.7 1.6 0.3631 0.0014 1177 7 19467 4 1233.6 1.7 19400 90
AguaCentaI A010426 -23.09276 47.40104 AD10-226A(0) 2129 43 9110 190 1177.8 1.9 0.2733 0.0012 1014 7.1 14451 67 1226.7 2 14390 80
AguaCaintaI A310-226 23.08276 47.40104 AD10-226A8(E) 2546 5-7 _29110 530 1158.6 1.7 0.2639 0.0012 419.1 2.2 14401 69 1206.2 1.0 14260 120
AguaCllenttI AD10-226 -23.09276 4740104 A010-2260(A) 2496 30 6770 140 11533 2.1 0.2631 0.0012 1111 8.9 13398 70 1230.7 22 13920 70
A010-226 -2308276 47A0104 AD10-226DAB) 2350 01 19730 400 1178.2 1.9 0.297 0.0014 194.1 2.6 15371 78 1230.1 2 15260 110AguaCaliental AD10426 -23.06276 47.40104 AD10-2260(W 2283 46 ,1090 360 1183 1. 0.2647 0.0017 5303 3.4 13823 9S 1232.1 1.6 13810 120
AguCoentel AD10-226 -23.08276 47.40104 A010-226DOB) 2389 48 8740 180 .110A 1.4 0.2413 0.0008 1047 3.8 12671 47 1223.2 1.3 12620 60
Agua~almll AD10-226 -23.08276 47.40104 AD10-22608(C) 2171 43 22410 400 11703 1.3 0.2625 0.001 403.6 1.5 13463 97 1222.6 
1.7 13720 110
AgaCifintel AD10-225 2308339 47.40209 AD10-228(AW 3916 31 16834 13 12323 3.1 .0.3114 0.0033 1213 10 16010 11960 190
Agus aldientel A010-228 -23.08339 47.40209 AD10-22() 5450 110 30940 720 1232 1.7 0.3003 0.0008 722.8 2.2 1557 48 0237 1.0 15470 s0
Agucghantel AD10-228 23.08559 47.40208 AD10-228(0) 4723 94 11170 350 1229.7 3.2 0.3011 0.0011 2022 49 1519 63 1231 3.3
15590 70
AuaCal.Il AD10-231 -23.17414 47.39920 AD10-233-10(F) 3973 79 63800 1400 :1232.6 1.8 03126 A.0012 262.9 1.1 16086 69 1309.9 2 13830 170
AD10403 -2317414 47.3920 AD10-233-10() 4290 86 51600 1000 1261.9 2.1 0.3111 0.0016 410.3 2.2 11891 87 1323.4 2.2 11730 140
Asuaalente AD10-233 3.17414 47.39920 AD10.233-10() 4141 83 47710 960 128.8 2.2 0.308 0.0012 424.4 1.6 15703 67 1321.5 2.3 15330 120
Agualentel AD10-233 -23.17414 47.16090 A010-233-10() 3261 65 32300 1000 1173.7 1.6 0.2901 0.0011 287.0 1.1 15410 65 1227.1 1.8
13190 160
AD10-233 -23.17414 4739920 AD10-233-14() 3936 79 216500 4300 12403 1.6 0.3208 0.0015 92.6 I1.420 338 33 13053 2.3 13800 00
AgaCaientl A010-233 43.17414 47.39920 A010-233-10() 4061 S, :2120 540 1263.8 1.6 0.3112 0.0009 748.2 2.1 13910 50 13216 1.7
11330 80
Aguacaete AD10433 -23.17414 -9739920 AD10-23341(M} 3520 70 6370 130 1180.2 2.1 0.2892 0.001 2530 13 13271 s9 1232.1 2.2 11250 60
Agutlent l AD10433 23.17414 47.39920 A010-233--2(A-1 4145 63 16000 1300 1269.4 2.2 0.3123 0.0012 361.8 2.2 15938 68 1326.1 2.4 10760 140
A010-233 23.17414 47.39920 AD10-233-542(A-4113 83 39800 1300 1271. 3.9 0314 0.0013 346.4 1.2 16010 7 1329.2 4.1 15820 IS0
AAgauCaCAlell ntal AD10433 -23.17414 47.39920 AD10-233-34(8) 4418 86 52100 1100 1284.6 1.7 0.3139 0.0011 422.7 2.7 15904 60 1343 1.8 15750 120
AD10.31 43.17414 47.39920 A010-2334-2(C) 3170 64 3200 300 1179.8 2 02849 0.001 4460 400 13072L 37 1231 2.1 13060 60
AD10-233 -23.17414 4739920 A010-233-9-2(0) 2775 56 2320 140 1179.7 2.6 0.2893 0.0011 4s20 210 15339 63 12)1.8 2.8 13320 70
A010433 23.17414 47.39920 AD10-23300A) 3797 76 970 570 1263.2 4.3 0.301 00047 19000 11000 15370 260 1319.2,4.8 15400 300
Agu.Clletal
AD10-233 43.17414 47.39920 AD10-233a(3) 3802 76 3146 96 1272, 2.5 0.309 0.001 3930 140 15732 39 5329.7,6 11720 60
AAgguuaaCCAalIinnttal l AD10-233 43.17414 47.39920 AD10-2334) 3795 76 3400 6600 1269.6 2 0.289 0.011 100 99 14660 600 1323.2'3 14600 00
Agua1altntel
A010-233 43.17414 47.39920 A010-233b-1) 2473 49 30140 620. 1171.2 1.6 0.2841 0.002 365 2.7 13100 120 1221.9 1.8
14930 160
AguaCiUuntI A010-233 .23.17414 47.39920 AD10-233b4-(5) 4681 94 1500 330 1229.7 3.1 0.3003 0.001 1345 4.2 11373 58 1264.8 3.3 1350 70
30  
AD10433 -23.17414 47.39920 A010-233b4(Q 4094 82 65 1 1224.-!.7 0.296 00009 2945 16 13371 48 '1279.1 1.3 10350 10
U3aI0.sI 11240 90
AD1033 23.17414 4739920 AD10-233b-20) 4415 8 13340 280 1263 1.3 0.2996 0.0013 1075 12 13276 81 1320. 1.6
176
LLT
00111 *6 S 11)6 091 00161 61T I19M 01000 0010 WE 633 6t WE09 61 916 (0)900z-0100 rlpf010 000606)t- 9000000w
00T10 S16 it-to 11000 00010 C1 0063 0011 001930 3s 1390 (MOSPI-oOoa 6080619* 000)06)- 030-010 amiapeuok100901
800)1 Vt. V666 100-0 TW0O 91t 106 30 0001 0 061SOf0 60 mo)110-0100 0990119 0000061- 00)0100
v3L op arll1
01 11C6 01) 01390 01-0 1"0 60000O 3110 10 1306 0000 00930) a1 0010 134000Z-010V 0030019 096)001t St0)010
&Me.P6W1669
003
00100 Net 6063O 39t
009 0IT010 ' 1a010 31 O V00 LVO) 51 E6069 001 06011 61 00I) M001-0100 6930619 
091)66- SWI-010
00)90 01091 01t 1091 11000 11110 III 03296 006i 00109 60 Vitt . 1)gtO)IV 31)6-0 011)01) 000-0100
00601
60t 1700 60, M6ET 01 110 1000NW0 100)0o rot 31001 006 06010 061 006 00,001-0109 0010619g 0u11106f- 00)0100
01161 on.Lopeonlal
0).399001 000 03011 01 v01t 111)00 100)0 Il T* 0006 00o1 001) I9 ME1 NIFIRZ-0109 00)0019g 001)00- 0010100v
00011 m6wL3030n*I
01 9*1 0)3101 tip 6MIT V1 Ut360 10000 60010 01i £0001 0001 03061 001 0103 110)601-010 0006619v 0110000-. 
00-OIOYv
0ME0I
51 310901 go 00611 01 I 11 LV 000-0 00100 V,1 610o 0011 00001 011 0101 Z-Mt0010100 01)0619O 011)00)t 000-0100
emlopeoookl
MIT0 01l 0100 0000DI 00009 061 0131 110)001-0109 00)60£0- 011)061 00)0100
001 01611 01I  00 £3000 Loon 1)0
00111
vs 0901 00) 06101 01 its 0000o tz1110
0601 60'0 600 00I  SO O LIZO I 0101 00t 000011 01 0)00 M600-110 0313619- 1109111- 60S01)3 al±00103n 
001o1 rof 11000 60110 &I3±OW3 wr*101 100M 
0000 I't 00111000OW 60610 90Z 10111 0001 0001 091 0113 lo3)oO-011)3 000011. 9- H010911)- 0P30)0 )3113331)346
06 00001 11 310)1t 0£ 0OUST *It 01 t1000 9ta6)0 91T M691 0011 00013 01 lost 10)0000-00) H906019 010011- 900-11)33
01t 06011 61 16111 051. 01000 VS 00r1 11000O 6589'0 1T C6611 00 0010 DOM 00M M t10 61 0111 000019- 169601- 310-010
0N 091ST CRZ 3001 t 6£111 itL 001t 01000o 1910 0)Z 10111T 01l 0131 00 011) 1-610-011)03 0000019 Z2990't- 600-0(1)0
01 00111 21) 16)01 is ass1 oft 0106 0900 66010 3)Z V)I1 01R 030 010 000)1 49)61-11)3 030 19v*g 0003000 
310-11)3
0111 3)t 11001 10
8ZVL 0000 00061 00000 09000 91 10100t 31 16£ 00) 01011 M09'10-01 ) 3100019- 001201)- 310-01)01 1013)101)en*
ST90 "Z 3 OIL
06 90900 00. V16 01000 610* 01 001)1T 0011 0690 000 06010 101S1)011 31 6610019q 0631) 
- 01911)11
1*1vmt31)wnft
0611 60f 6001 0£
001 ISM1 00S M16 10000O 966)0 01 600) 0031 0060 09 0061 (2(03610011)1 
3600V01 0603001t 610-6)3o
01 001)1t 96
06 10161 5-1 C90 60000 1660 11 9011 000i 
00901 000 0100) (990311o-O1)o1 3610019- ES60000- 019112)M3
6E 00101 Ott
01101 100 0010S 6o000 rC000 '1 90111 00)11 000009 001 0600 (H949O110-11) 8610019- 660201- 110-101)0006 00"01
00001 00T E1611 V0000 0100 6Z £161T 00E1 00000 006 00318VuT l-1o 361001 6!6 1300)t 919611300tt Mot0
061 0966 0000 6*1 3I06 10000o 00160 6,0 31)1 OL 001 0131 0OT 0001 0o03Oo-ooT3 3610019, 160000)Z- 6191113
16166031)009
000 00301 01100 11t 0301 6100,0 00160o 1) 01)11 0013 
001000 DB 06 E Mill 00900911)03 3610019 0663011 11911)61 13aup6301)04
I**o10)6 93 60M0 060 0CUT 01000 9)610 S1 61611 0090 001011 001 0011 003010-01)01 6610019 064006) 100-111
16300901)346j
00)
06 1301 00 OIl 1TW0 661 V0 10Z 0011 0011 0000 061 0016 
(2)60110126.1 3610019- f6Eeo300 110-5)1)01 13100901)01*t
101000 3v4
0011 6) 901 t1000 61100
DOT 00691 CE 669 91000 6060 60 1661 069 00606E 3N 0000 j0)9wloSI1)03 36100S9 0661061 6M19)01
Ii ual3 onto4
0000 DOM0 060 00s 100 0PIC0 z 1132 00000 90I0 ME 0000 (awst"00 £11)£1 3600m 663000- 
0M191)1 I aum".13 qn0
09M6 911 0001T 31 0060 6's 0'0 19 100 0W60)0 1s 00f601 00) 0066 6£t 0060 (1)0w11~0-610 3610019, 066300)- S 
S161)
09
1030 U6 $'fit 1000 0)9)0
9,1 10311T 001 00066 01 061 (oOWVll0-SIX0) 3610019 06006)Z- 619111)01)
011 fro0t10 ILT 0116 1100 11)
9C1 1-0011 00 1t 0)611 00U1 00001 9 0010V (0)06000-111)01 361001 161301)- SMS0-OIY
006V0 01 0111f 11000 00110
06 101030131)3461l 60601Z 00661 00l 631t 6000*0 600)0 31 11311t 00 01N1 031 061 (3)MVI00-011)01) 16009- 166300)- 010-011)0
01901 901 6)01 ES 1610 I 0101 60000 £010 61T 6111 001) 009000 001 0011 (0)309111)3 £610V'i9 1660) 010-11)3
V1 10111 00 10S0 91 0M0 1000 060)0 31T 062111 010 0011 001 000£ 2)63010-111)0 Lfi£0019- 166600r 019011)01MST0
09 831-11c1) or*
10T 610)0T Z9 9061 910 691 1000 10060O 91T 31311 006 0001 01 06001 (1010011)03T 1610L9- 160600)- 
010-11)3
01
I W0)01 60 MIT) 01 6IVE 60000 616)0 r 01311t 01) 01101 061 0)03 (0)300-01)3 1610019 '603060v- 000-1201
8010 of%*04
0£ P130t V6,0 0I00 Z1000 13100o 66G 6)7601t 0001 0000 £0 
011 (0)000-111 1610019w 166306)- 000-111)
60r 00000 01
00101 01 0601 11000 131)0 0 3111 000 01101 011 0)15 11£JKOv -MA3o1 1610019- 166910 V00-61tW
Olt
Of111 S900 01 00U0 TW000 S0 6 61T 9111 006 01601 0)6 0130 1-0001-110 1310019- 160601)-t 010-111)
00161 grist 6) 1010 01000o 013)0 I1) 06301t 001 00136 011 0006 110"Pvl T0 1 610019 16130'11- 
010-11)01
06
06 91t 00001 10 06M0 01L 6001 01000 096)0 60 10111 060 06361 001 0016 3)WVT&S1)3 100 160306) Lf***- 010-0111
001 01601 11610 6S 3011 01000O 66010
1103W101)346
V? 6111 000 0)300 001 0301 )-43(volo-SL-) 1610- £61000- 010-11)01
of 00301 06301 B80 3061 00000o 036)0 S1 06111 001 00100 011 0001 1-(30000O-11)0 16100L1 166300) 01061201)
06 09901 I1 6)01 61 06961 11 011 3000,0 L11610 11 611911 01) 01061 01l 0)16 1-400000-010 0)66149 001 160i*E- 060-0100
06 00001 US 6BUT 1000*0 000)00 11t 30011 00) 019) ES 0410 040(4100-010 0)6661, 0191160-. 600-0100 16103101)34
0011
31 610)1l Vs 66661T 91S 6101 30000 60)0 61t 66611t DO6 00101 011 0016 t-l)'96Z-OTo 00661t9- 010110)- 60-0100 Ia lwp101)04
05 0)061 I0S 1611_ 6000 911)0 9'1 069011 00N 06000 I3 1100 0-404600010 00661'19- 010116)- 600-0100 0603101)346v
00001 13660 613: 6031 tT1000 61100 El) 09)1r 000) 00)611 33 't)0p (0(-,410-010 0060119- VULVill) 610-010001 133031101)04
MS06 11000 001 060) 01000 13910 IO~ t O060£1 ) 006 01 010(S 61019 00z6~06i1f0tvq  0101111r 060-010009 16106031)346
10C 6011 01
0601 00131 61 10rt' 11000 1060 LIE 6C0I1 010, 00)0 06 6600 10)-4001-010 006609!) 010116)f- 6010100V e3103-ft4
00611 11T U9611I  60 6161 Z'S' "IT0 10000D 30 3'1 10011 010t 06)00 033 00)3 ZJ43014000-0100V 0006C11 0T10110)- frz-0100 ........ 
DOM0 Olt.00 0E0OR 00161 £0 ZT 09 000 00000o 61l 16611 001 0011 01) 0106 119H)1-4061010 0)66619j 010110)- 000-0100
00
00011 61 0t 0 9 9oo0t0 01 0011 Iwo0 30110 31I 06111 000 0001 60 066 ()19 )-019 66619 0101110 111-010 3103101)34
11101 011 0616 11000 0011 r) 60)11 13 I00r 03 HOW WZ£))-q60)- V 00660'19" 010110)t- 111010001)01
0) FUZZ1 SO
0Ous1 00061 to 0300 T1000 606)0O
.0..3....- -
moA Nuned 003 Ipmi0 0636 Ifd SAWA 06 
(to g ( (so1 )( (og0 n4lop (sgl (ad0T (gsO nitl) 39ua3eamu fi (J300o3n n ( Jo tesnh3 5o03dwe9o ot nl tzs (l /Ott /0o 
Table A.2: U-Th data associated with Chapters 3-5 (continued).
LAkim Smpi.1e5 La.itdef) Lnongitude) U-1hSAmeIm 2333U t(7.) 1W233 A) t434U / 1(26) 
53238) Wage) Ae () *2) o(20153.32 Out-)v INS1 (or)
Vr*P
LagunadeTara AD1046 -21.03004 67.34742 AD10-246(B) 23S1 47 82800 1700 922.7 7.2 0.2691 0.0015 121.3 0.52 16190 110 970.6 7.6 15600
LagunadeTara CvCi64045 -23.00291 47.31172 CC16-04SA(6( 1874 37 116500 2300 943. 1.7 0.2903 0.0017 74.13 0.43 17420 110 988.4 2.1 16400
LagunadeTara C1C16-04S -23.00291 -67.31172 175 35 72200 1400 944 1.9 0.278 0.0016 107.1 0.62 X6550 110 927.3 27.3 13900
LgunadeTar C7C164145 -23.00291 47.31172 CYCC16.404504(D8) 2296 46 1790003600 964.8 2.1 0.2953 0.0024 60.12 0.69 17530 160 1010.3 3.2 16300
Lagunadelara CC6045 -23.00291 4731172 CYCI64458(C1 1420 110 91300 1800 956.7 1.6 0.1941 0.0005 183 0.76 11941 54 861.4 1.7 11660
Lagunm1deTar CYC16-045 -23.00291 47.31172 2037 41 111400 3000 947.7 1.8 0.287 0.0024 61.32 0.51 17160 160 991.5 2.9 16000
tagunadeTara CYC16045 ,2300291 47.31172 C1C164458E} 2013 40 243800 4900 939.2 1.5 0.291 0.003S 38.72 0A6 17770 230 982.2 3.9 15900
LagunadeTara CYC16445 -23.00291 47.31172 M~16-04SOM3 5240 100 103200 2100 854.8 1.6 0.1932 0.001 18 0.87 11896 67 883.1 1.7 11600
LagunadeTa CVC16.045 -23.00291 47.31172 CVC160458(F) 5310 110 352100 7100 807.2 1.6 0.196 0.0021 47.33 0.51 12060 140 984.1 2.5 11000
LagunadeTarm CC16..04S -23.00291 47.31172 CVC16.045(G) 3615 72 263100 5300 850.2 1.7 0.2125 0.0026 46.34 0.37 13180 170 379.4 2.7 12000CVC16-048(H)
LagunadoTara CC16-048 -23.00608 47.33198 1140 230 329600 6600 38.9 1.8 0.2064 0.0021 113.9 1.1 12720 140 89.1 2 12300
LagunadeTar CMC1644 -23.00606 47.33198 CVC164488( 12230 240 362400 7300 58.8 1.6 0.2077 0.0014 111.3 0.75 12808 93 89.1 1.9 12300
LagunadeTar CIC164048 -23.00608 47.33198 11340 230 672000 13000 837.1 1.6 0.2123 0.0021 00.98 0.37 13150 140 887.1 2.4 12200
Launade7ara CVC16.048 -23.00608 47.33198 C1C16448C(A) 1050 210 445600 9000 860.9 1.7 0.2122 0.0015 :0.11 0.57 13088 97 891.2 2.2 12400
LagunadeTa CYC16448 .23.00608 47.33198 CMC26.048C(C) 1800240 410100 8480 805.4 1.6 0.2096 0.0017 94.61 0.78 12950 110 31.9 2 12400
LagunadeTara CC16448 .23.0008' 47.33198 12230 240 369200 7400 852.1 2.9 0.2064 0.0014 100.6 0.70 12771 93 882.1 3.1 1200
LagunaKniques C7C1603 -23.75992 47.78862 CCI6-023A(8)
CVCI6.23A(B3 6510 130 9720 200 446.6 1.7 0.1286 0.0005 1367 4.9 10101 40 459.1 1.8
10070
Laguna ufniques CC16023 -23.7992 47.78862 CYC16-023A70) 19970 400 10S40 210 461.2 1.9 0.1425 0.0003 4285 9.9 11132 32 475.9 2 11120
LagunaNnique CYC6-023 -23.75992 47.72862 CC16423A(D) 16230320 29300 590 434.9 2.4 0.1288 0.0004 1133 4.1 10063 38 467.9 2.4 10020
Lagunanniques_ CC16423 23.75992 47.78862 CC16423(E) 15320 310 11600 240 402.9 2 0.1486 0.0006 3117 23 11619 50 478.3 2.1 11600
Laguna mique CC16.023 -23.75992 47.70862 CYC16423A(F) 6046 180 16010 370 430.5 1.8 0.1463 0.0005 1169 6.3 11329 44 465.4 1.8 11490
Laguna uniques 4 C6C16425 .23.75985 47.78864 CVC16-025AA) 4911 93 80600 1600 442.3 1.3 0.1520 0.0006 147.7 0.61 12123 54 457.2 1.6 11800
Louna4 iqd4ue CVC16425 -23.75985 47.7864 CvC1642sA#AB) 5440 110 122700 2500 417.1 2.1 0.1304 0.0006 91.8 0.42 10475 35 429 2.2 10000
lagunanniques C1C16023 -23.75985 47.78864 C1C1642l5A4Q 3804 76 58300 1200 43.1 1.8 0.1537 0.0007 159.3 0.47 12120 so 468.8 1.9 11800
Laguna unique C9C136425 -23.7S985 47.78864 CVC16025AND) 5160 100 11670 230 447 2 0.1454 0.0008 1021 5.8 11490 71 461.7 2.1 11440
Lagunatniques C1C16425 .23.71985 47,78804 CC1642SAM) 5360 110 27730 560 450.4 2 0.1449 0.0006 445.1 1.7 11421 1 465.1 2.1 11310
Laguna niques CYC16425 -23.7590, 47.78864 4947 99 77200 1500 450 2 0.1314 0.0007 154.2 0.7 11963 62 465 2.1 11600
Laguna iques CYC16428 -23.76062 47.78827
CYC16-.02363(1A4.1 7320 130 76200 1500 439.7 2 0.0019 0.0002 5.93 0.33 438 16 460 2 220
Lagunafniques CVC16.028 -23.76082 47.78827 3350 170 16100 1100 405.1 1.3 0.004S 0.0002 10.03 0.36 339 11 453.4 1.8 200
Lagunapaniques CVC164029 -23.76124 47.78737 C0C16-028AMM 3761 75 49020 980 468.7 1.9 0.1271 0.0007 154.8 0.83 9822 58 481.5 2 9550
Laguna niqus CYC16-029 -2376124 47.78737 CYC16429AMA) 2726 SS 101500 2000 401.1 1.8 0.1506 0.001 64.22 0.43 11762 86 4793. 2 11000
Lagunamniques CC16429 -23.76324 47.78737 CYC16-029AA(Q 4001 60 10260 220 419.8 2.1 0.1428 0.0015 382 11 11170 130 474.4 2.7 11110
LagunaJA5niques CC16-029 .23,76124 47.78737 CVC16-29AAMD) 4348 87 12010 20 456.6 2 0.1385 0.0013 796.2 8.7 10840 100 470.7 2 10730
LgunfAniques CC16429 -23.76124 47.78737 399.7 8 10830 230 495.6 9.4 0.8077 0.0066 473.4 4.4 80400 1200 621 12 79900
Lagunan4ques CC16-029 .23.76124 47.78737 CYCI64298BA41 402.1 8 81300 1600 456.2 2.9 0.9176 0.0071 72.03 0.43 101300 1300 600.4 6.3 97000
Lagna knqus CVCI6429 .2376324 47.78737 C018429N(4 359.5 7.2 91300 1800 632.3 3.4 0.2941 0.0051 18.34 0.3 21390 410 62.1 7.2 17000
Lagunah6nniques CYC16429 .2376124 47.78737 CYCI6.029CA(B) 2474 49 25340 510 465.3 1.7 0.1406 0.0007 217.9 1.1 10943 59 479.6 1.7 10730
Lagunanques CVCI6429 -23.76124 47.78737 CYC16429C4C) 2574 51 21900 440 463.4 1.6 0.1437 0.0006268.1 1.1 11214 53 478.1 1.7 11040
Laguna6n4iques CVC16429 -23.76124 -67.78737 CVC164290(4-1 3208 64 106100 2200 465.1 1.8 0.1353 0.0004 63.S3 0.38 10513 64 478.1 2 9800
Laguna 6UnIque CVC16-029 .23.76124 47,78737 C7C16.0290AV)2 3097 62 109100 2200 467.1 6 0.1344 0.0008 60.57 0.33 10427 80 490 6.2 9700
Laguna Wnique CC16029 -13.76124 47.78737 C9C364290A53 3094 62 90000 1800 466.6 1.7 0.134 0.0007 73.14 0.33 10398 18 479.7 1.9 800
Lagunanniques CYC16.030 23.76122 47.78744 CYC16430AAM 12420 370 9430 210 367 1.1 0.1431 0.0004 4427 39 12000 35 379.7 1.6 11990
Lagunahdniques CYC16434 -23.76589 47.78331 CK14-041i 7920 160 60400 1200 482.1 1.6 0.0403 0.0003 84 0.69 3004 26 486 1.7 2840C1C16434M8
Laguna Pnlqum CV1C1634 .23.76189 47.78331
C1C16-034A4C) 1201 24 49390 990 460.8 2.5 0.2124 0.0012 81.98 0.47 16750 110 102.6 2.8 15900
Lagunan ue3 CYCl-034 -23.76589 47.78331 1241 25 60600 1200 401.1 1.8 0.1848 0.0013 60.24 0.41 14430 110 500.1 2.1 13400
Lagunagnques CVC164034 -23.76589 47.73331 CYC16034AND) 1739 35 93100 1900 495.8 2.1 0.2129 0.0015 63.16 0.45 16610 130 518 2.5
Laguna iniques CC16-034 -23.76589 47.78331 CVC16-034AB4W .180 36 87200 1700 485.8 2.1 0.1706 0.0016 56.2 0.52 13210 S130 2.9 2.4 12200
Laguna 4 A8ati3 CYC16-009 .23.71192 47.74437 5650 110 30700 610 264.1 1.5 0.1641 0.0006 479.6 1.7 15092 63 2751. 1.6 14960
Laguna stanti CYC16409 43.71192 47.74437 CC16409A(B)
CYC164094C} 3729 75 34120 700 261.6 2.2 0.1067 0.0009 272.2 1.6 14543 90 276.6 2.3 14320
Lagunahscanti CYC16409 .23.71192 47.74437 41gg 84 37440 750 24.4 1.9 0.1434 0.0007 2948 1.3 13074 71 274.7 2 1260
Laguna ceminti CVCI6009 .23.71192 47.74437 CYC16-0094D) 4279 86 152100 3000 282.6 2.5 0.157 0.0011 70.1 0.47 14170 110 283.4 2.6 13300
LagunansfAnt6 CVC16417 -23.71005 47.76269 CVC16.0174.(8x.1 52200 1000 19260 620 331.2 1.5 0.1466 0.0006 6300 270 1261 S4 343.2 16 12650
Lagunauscant CC16417 23.71005 47.76269 CVC164017-24,C.1 17400 1100 22330 610 326.8 1.6 0.1143 0.0004 4662 88 9785 31 336 1.7 9730
Lagunalwacani C106.017 43.71005 47.76269 C1CIS-0174D.A03. 1 36340 730 13630 S00 313.6 1.7 0.1208 0.0004 110 160 10460 41 321 1.7 10450
Lagunauscanti CVC16.017 43.71005 47.76289 C7C16417 2 58300 1200 11460 230 335.3 3 0.1314 0.0004 10609 S1 11247 45 346.6 3.1 11240
Laguna.stanti C176.017 -23.71005 67.76269 CYCI6417W2 5600 1100 10900 240 331.5 3.3 0.127 0.0015 10440 150 10880 130 341.8 3.6 10380
Laguna acanti CVC16417 .23.71005 47,76269 CYC16417(AJ3 61200 1200 10430 220 330.2 2.2 0.1277 0.0008 11880 110 10960 76 340.1 2.2 10960
Laguna5 W0ant CYC16417 .23.71005 47.76269 C0C164174A(91 22700 430 61800 1300 322.7 3.4 0.1119 0.0005 612.4 3.7 9609 51 331.5 3.5 9550
Lagunans3ant CYCI6017 43.71005 47.76269 CVCI6417A1334 23660 470 66600 1400 320.8 2.6 0.1343 0.0005 645.4 4 9812 50 329.8 2.7 9790
178
Table A.3: U-Th data associated with Chapters 3-5 (continued).
0Th230/ 1h230/ Age d234U Asa
LAM 914mm lWeSet Ladkudr) Longtuddr) U n.hsampleName 238U 1PP) 7h232 i2o) d234U 1P2) 1(2.) 7 )3.0 
Alp 5(3) All
Ulm)(2) ThM3 Mal *wo) t 3,43ni ordw 
1a
wOp
nss/ pI/V ppon peemn 
Lagunabcantl CC16.418 .23.71191 -47.76099 CYC16418A). 1E+00 2500 30940 620 323.9 0,1293 0.0004 8221 22 11160 334.2 
2.3 1110 40
-23.71191 47.76099 CYC6-016AA2 1E.05 2600 32080 850 327.1 01274 0.0004 $170 140 10958 337.3 2 10950 40
Laguna Me.nt CYC16018 -23.71191 47.76099 CYC164018A443 2E405 4500 00200 1200 321.4 0.1278 0.0004 9170 130 11049 331.A 1.5 11040 40
Lagunastnti CYC-16010 .23.71191 -67.76099 CYCI64018AB) IE+05 2400 27900 1900 325.6 0.124 0.0001 3440 040 10670 335.6 1.0 10660 50
LagunaM ant0 i CYC16.018 .23.71191 47.76099 CYC1641SA(C) 1E+05 2300 25200 2200 325.4 0.123 0.0006 8630 770 10832 335.3 1.6 10580 60
Lagurna6xanti CYC16418 -23.71191 47.76099 CYCI016 D) 0E05 2100 45600 1900 328.0 0.1172 0.0006 4200 150 10030 338 1.7 10020 60
Lagunaf.icanti CYC16418 -2.3.71191 47.76099 C7C164016 E) 0E405 2900 38700 2600 331 0.1222 0.0006 7210 470 10457 340.9 1.5 10450 
60
Lagunawstnt CVC16418 23.71191 47.76099 CVC164185n 59700 1200 26100 1600 352.5 0.2114 0.001 7140 430 18400 371.3 1.9 16390 100
CC6-018 -23.71191 47.76099 CYC16-013AG) 31620 630 13300 2000 349.6 0.141 0.001 5320 790 11975 361.6 2.7 11970- 90Lagunaux4anti 
LagunauAscnti CC(16.018 -23.71191 47.76099 CYCI6018AH} 10510 210 15700 1600 348 0.137 0.0015 1460 150 11640 359.6 3.4 11600 140
Lagunalixanti CVCI6Ola -23,71191 47.76099 C7C1601B0W 56100 100 22240 40 345.8 0.1376 0.0005 t1 21 11709 357.4 2.3 11700 
SO
LagunauxhwAnti CVC16018 23.71191 47.76099 CYC164180(8) 66700 1300 27280 550 332.6 0.1267 0.0005 4922 21 10854 343 1.5 10840 s0
PamnaV4a CCI6038 -23.83329 47.81684 CYC16438A.2K1 23520 470 156900 3200 362.7 0.1996 0.0006 47S.3 1.4 17155 380.6 2.3 17010 
120
PampaVarda CC16038 -23.83129 -67.81684 CYC164038A2(B1 44640 880 03200 4100 360.1 0.2355 0.0007 821.6 3.2 20564 301.6 1.8 30460 300
90
PampaVarl CC16038 .23.8329 47.81684 CC16y438A 1 24740 490 81500 1400 360 0.2395 0.0008 1155 3.8 20944 331.8 
1.9 20870 
PampaVarel CVC16435 -23.83829 47.81644 C7C16038A142 20740 510 9100 1800 362.7 0.2136 0.0008 958.7 3.6 18452 382 2.2 
18376 90
PampaVarl CYC16-038 .23.5399 CYC16-038A(A3 25990 520 95700 1900 361.6 0.2223 0.0007 958.3 3.1 19288 381.8 1.6 
19210 90
PaompoVwar CYC16438 -23.83029 47.81684 CYC16.4M3VA(B 51000 1000 63600 1300 360.8 0.1531 0.0005 2007 5.6 1.3379 374.7 1.7 
13370 50
-23.33829 47.8614 CYC26-0386362 74800 1500 53600 1100 361.1 0.1 0.0003 2216 13 8294 369.6 1.4 
420 30
PampaVarda CC16,038
PamnpaVaras CVC164038 -33.3329 47.81684 CKC16.038A04.3 5700 1100 09900 1200 357.9 0.1425 0.0005 2100 11 12029 070.3 1.6 
12000 50
PampaVarda CC16.438 -23.8339 47.831684 CVC164033A(CF1 2200 460 93000 1900 366.3 0.958 0.0031 3729 12 121890 016.5 
3 121800 800
518.8 3.8 124200 800
PampaVarela CYC16038 -23.83829 -67.81684 CYC16)4(3C2 24930 500 104800 2100 363.4 0.9752 0.0027 3681 9.3 126260
CVC16-030 47.31684 CYCl6-03A4Dj 15620 310 391000700 363.8 0.19S9 0.001 124.2 0.6 16793 380.8 1.7 16200 400PampaVaraa -23.8329
67.81684 17020 000900 363.5 0.2001 0.0008 160.2 0.63 17921 381.8 1.9 17500 300PampaVardo CVC16038 -23.83829 CYC16438E) 340 7000
300 1.5 16540 120
PampaVada CVC16038 -23.83829 47.31694 CYC16.03A48 33820 680 21900 4400 362.7 0.1945 0.0007 475.1 1.8 16643
18760 100
PampaVarde CVC16.038 23.832639 47.81684 CVC16038M) 30760-620 137100 2800 362.8 0.213 
0.0008 776.9 2.8 18364 382.5 2.2 
CYC16038 -23.833629 .67.81684PampaVarda CvC1603SAIN) 39000 70 220300 4400 360.4 0.1981 
0.0007 556.7 1.8 17045 378 2.7 16920 110
CC16-038 -23.83929 47.81684 CKC16-038Mi) 39130 780 172500 3500 362.7 0.2002 0,0008 749.7 2.8 17949
381.4 2 1730 100
Pampavarda 
-23.83829 47.81684 CYC16-033(K) 38110 760 190400 3800 364.2 0.3542 0.0012 1126 4 32330 398.9 
2.5 32220 160
PampaVrela CYC16038 382.3 1.9 16790 170
PampaVarel CYC16033 -23.83829
47.83164 CYC160384(A) 19610 390 215500 4300 364.4 0.1936 0.0006 206.5 0.69 17035
47.81684
PampoVala CYC16.038 -23.83929 CC16-038(A62 17030 340 204600 4100 364 0.211 
0.0007 279 0.72 18195 382.9 2.2 17930 .190
PamnpoVardia CVC1638 -23.83829 47.81684 CYC160368(8)I 27550 550 16800 3400 360.9 0.1893 0.0006 490.7 1.2 16231 377.6 1.5 
16090 110
47.81684
PwampaVade CC16438 -23.8329 CVC16-088(082 30710 610 201900 4000 361.2
0.1685 0.0006 455.3 1.2 16148 377.9 24 16000 120
15750 80
Pampa Varl C*C8C16438 .23.83329
-67.81664 CYC16438B(5C1 34070 660 12100 3000 361 0.1852 0.0005 658.8 1.5 15854 377.4 1.0 
Varela CYC16038 -23.83829 47.81684 CVC1603880(C2 3E06 35000 10700 3000 359.9 0.0044 2E05 723.7 2.4 
3S32 360.2 1.4 353 2
Pampa 
-23.83829 47.81684 CC643881(D1 2E+0 3800 135600 2700 363.5 0.09'33 0.0002 3054 4.2 7706 371.5 
IA 7690 20
PampaVarda CC16438
47.16841
PampaVarela CVC16438 -23.23829 CC160388(E) 34050 600 162500 3300 362 0.1881 
0.0008 626 2.7 16102 378.8 1.8 16000 100
CYC16438 23.838329 47.81684 CYC16-0388(F) 25870 120 107900 3400 362.3 0.1977 0.0007 463.6 1.6 16979
380.1 1.7 1630 120
Pampavarea 
PampaV"Ard CVC16438 -23.63829
47.81684 CVC164038C(21B.1 32990:660 064000 1100 0 360 . 0.1968 0.0003 182.6 0.77 16922 3777 1.5 16500 300
47.81684
Pamp&Varda CVC164038 -23.83829 CYC16438C2K()-1 16300 330 288000500 364.7 0.2025 0.0008 
181.3 0.74 17394 382.7 1.7 17000 300
47.81684 1362 1.4 19309 381.8 1.6 19250 70
PampaVared CC16-038 -23,3829 CVC16-038C-3(A1 53900 1100 139900 2800 361.7 0.2226 0.0006 47.81684
PampaVarel C(C16038 -23.83829 CC16-038C4A1 26740 530 2238800 4600 361.8
0.1657 0.0005 307.0 0.9 14066 376.2 1.6 19870 140
PampaVra CC16.0438 -23.833829
47.816034 &C16038C4(BF1 53300 1100 204300 4100 361.9 0.1713 0.0006 709.2 3 14072 377 1.8 14490 80
3768 10 14831 381.2 1.6 14860 50
Pampavarea CYC16041 -23.83990
47.81603 CVC16041AA-1 66400 1300 49010 980 365.5 0.17S2 0.0005 
60
PampaVarv CYC16441 -23.83990
47.91603 CC16041A(A*2 76800 1300 9700 1200 365.1 0.1766 0.0006 3608 12 15018 380.8 1.7 15000 
47.81603 1250 3.6 13837 370.9 1.4 13840 60
PampaVara CC164041 .33.83990 CKC16041A(B1 3460 700 72200 1400 356.7 0.1631 
0.0005 
-23.83990 47.81603 CVC16-0401B4A4 40650 810 03700 1100 362 0.2955 0.0009 3503 15 26379 390 1.6 
3630 100
PwmpaVarea CYC16041 47.816039 18560 80
PampaVarea CYC1641 -23.83990 CYC16041881 21010 420 3930 790 357.3 0.2143 0.0006 
1816 7.3 01599 376.5 1.9 
Pampa Vara CYC1441
.23.33990 47.81603 CC164041(C)1 26780 040 92700 1900 351.4 0,1538 0.0005 705.7 1.9 13105 364.5 
1,8 13030 70
500
SaadeLovoqus CC150421 .23.20227 47.2319 C(C15142LAA 10240 360 246600 4900 711.2 :1.180 0.003 1386 3 114390 981.5 3.1 114200 
SaLar de Lovoque CYC154021 43.20227 47.27259 CVC1502184A 29100 590 471300 9400 725.1 1.1156 0.0029 1109 2.5 103240 969.7 2.7 103000 500
Saar do e~oqes CKC15421
.23.20227 47.27259 CYC15421( 19420 390 363800 7300 699.4 .1.2516 0.0033 1061 2.4 1277B0
1002.3 2.9 127000 600
Solar doeLoaoqu CVC15-022
-23.20257 47.2723 CVC15422A(A) 8480 170 140100 2800 703.7 1.3214 0.0044 1270 4.2 139970 1043.8 4.3 139700 900
Waar de Lovoque CVCI5022 .2;.20257 47.27253 67 
600
CYC15022o6) 4358 49070 980 717 1.2125 0.0036 1710 4.6 1160 100232 .3 110700 
SAla do Lo~ogta CKCS5022
23.20257 47.27253 CC15.022B(B) 8630 170 03600 1100 699.9 1.2968 0.0033 3313 7.2 135900
1026.8 3.8 135300 700
Salar do vfques CYC15425 -23.20237 4727073 CYC1S-025A6A0A 4779 96 6720 150 749.5 1.2115 0.0032 13670 130 114820 1036.3 3.1 114800 500
Salar doLo~oque CYC1525 .23.20237 47.27078 CYC01525AA(8) 4532. 91 3830 180 752.1 1.2578 0.0033 10290 S 121800
1060.5 2.8 121800 600
50iavl4a.3rdoo  yvoocqtuess  CYVC155-402457
Salar dolooqu CVC15.025 -23.20237 47.27078 CC10425AA(C) 6940 140 12450 250 753. 1.1759 0.0031 10412 27 109010
1025.4 2.8 109000 500
Saaet.oyoques CC15702S -23.20237 47.27078 CYC15025AA(0) 5450 110 980 200 742.7 1.1982 0.0028 10079 24 113560 1023.2 2.6 113500 500
-23.30114 47.26414 C7C15-047(A) 1312 26, 220900 4400 710.3 1.3233 0.0069 124.8 0.62 139300 1700 1045 12 
137000 2000
179
Table A.4: U-Th data associated with Chapters 3-5 (continued).
Lake SWAtm 4ple Set La Latitu ) Lengtude 6U-ThiCmpi a4me 238U t426) 16232 03e) d234U file) (9)2301 162 6a 4249 Il 162IN32 fumwr6 IN kwp6.)
nw/s P/8 Parma acth~ty P r, y permod WIP
s4wde Lofoqu CC5047 -23.30114 47.26414 CYCIS.47(6) 1273 25 164700 3300 711.4 1.2492 0.0044 153.3 0.53 125710 530 1008.6 36 123700 1600
aardeLoyoques CyCis547 -3.30114 .67.26414 CYCIS1047(C) 1407 28 3S1700 7100 702.3 1.2072 0.0054 75.81 0.31 119800 1000 974 10 116000 3000
Solardetofoque CyCI5447 -23.30114 47.26414 CYC1S047(D) 1412 28 400700 000 702.1 1.2106 0.0015 67.72 03 120430 940 973.8 9.3 116000 1000
SalardeLayoqum CyC15447 23.30114 4726414 CVC15447(6) 1395 2 191300 3800 702.1 1.2038 0.003S 139.4 .033 119320 610 977.1 4.8 117200 1600
Solardetofaques CC15447 -23.30114 47.26414 CVC15447(F) 1412 30 327500,6600 604.1 1.1503 0.0046 82.64 0.29 111750 740 941.9 7.2 106000 3000
Salardopeque CC15.047 -23.30114. 6726414 CVC15447(6) 1767 35 9000012000 680.2 1.1142 0.006 52.94 0.31 107700 960 907 11 102000 4000
SSadoyqGe CYCI5-047 -23.30114 47.26414 C4C15447(H} 1743 35 511000 17000 672.7 1.1964 0.0061 40.83 0.33 121900 1400 929 15 114000 6000
S1adeLoyoque CYC1S.047 .23.30114 47.26414 CVC1S447(2) 2213 46 25.06 49000 516.6 0.9925 0.0099 1142 0.18 98900 1500 736 31 79000 1SOOO
5alar doL oque CC154047 -23.30114 4726414 CYC15.47(K) 2106 42 962000 20000 670.6 1.023 0.0042 34.93 0.17 05740 600 659 14 88000 6000
SawdeLoyoque CyCi047 -23.30114 47.26414 CYC1MS47(LU) 2417 48 630000 14000 689.6 0.5416 0.0036 52.89 0.24 83100 450 860.1 5.2 75000 3000
Solar do Loyoqu CyCIS-047 -23.30114 -67.26414 CMC15447t66 1260 37 1E.06 30000 647 1.1083 0.0062 22.23 0.14 110500 1000 050 23 97000 10000
Slar del Humo CYCIB6461 .20.30864 48.80093 CvC164615(A)41 467.7 9.4 14320 290 1123.2 0.2587 0.0009 134.1 0.45 14011 so 1167 1.6 13600 300
Slw dal NaO CCI646I .20.3064 45.60093 CYC16.0615(A)2 463 9 4 14360 290 1121.7 0.29 0.001 134 0.1 14039 56 1165.6 1.7 13600 300
Sar dal Huesca CYC164O61 .20.30864 .68.0093 CC164615(8)-1 637 13 77600 1600 521.9 12.3269 0.0048 173 0.65 124110 a00 1160.8 5.5 122400 1400
Solar did Humac CC16461 .20.30864 46.80093 CYC164616(C) ?26 15 195100 3900 852.7 1.3052 0.0059 77.06 0.32 117220 900 1174.1 9.6 113000 3000
Salar del Husco CVC16-61 -20.3064 4880093 CC1604616(DI 316.1 6.3 39800 800 1097.1 0.3596 0.0029 45.35 0.34 20180 180 155.5 4.7 1600 1200
Solardel Hume CfCI641 .20,3084 45.0093 CVC16461a(0) 500 10 16270 330 1054.7 0.2341 0.0064 114.7 3.1 13100 350 1093 2.S 12600 500
Solardel Humao C160461 -20.3064 4.80093 CVC16-061BF) 509 10 22940 460 11461 0.2379 0.0013 83.88 0.43 12683 73 185.7 3 12000 400
SAar de Hascm CYC16461 -20.30864 468.80093 CC16401(G) 532 17 25680 520 1199.1 0.2692 0.0037 140.6 1.9 14080 210 1246.3 4.2 13700 300
Salar d Humco CYC16462 .20.30870 48.50090 CC16462A(4 304.5 6.1 47610 960 1016.1 0.3308 0.0031 33.54 0.33 190 1137.1 16 16300 1600
SWr dd Humao CCI-062 -20.30870 46.50090 CYC16062A()1 465.7 9.3 19260 390 1005.1 0.2772 0.002 106.4 0.77 16030 130 1049.8 4.7 15400 400
SaardalHuaco CYC16062 .20.30670 46.80060 CVC16462A(C)1 610 12 17720 360 11511 0.258 0.0015 141 0.92 13752Z 84 1199.4 2.1 13300 300
Solar del HUNDcO tCC16464 -20.3Oa8 48.60104 CC16464A(AI1 310 6.2 61000 1200 1099.7 0.31SS 0.003 25.49 0.25 17510 10 1146.1 7 14600 1900
Sade4uamo CYCI6.071 -20.30660 48.40030 CYC16071A(A)-1 454.6 9.1 42060 540 1093.9 0.2575 0.0017 44.26 0.28 14370 110 1134.2 6 12500 900
Sar dal Huac : CYC16071 .20.30560 48,80030 CVCI6-71A(A.2 447.4 8.9 41750 540 1097.1 0.2575 0.0016 43.81 0.26 14129 94 1137.4 3.5 12800 900
Saoar delruaso CVC6471 .2030840 4.000 CYCI6.071A(PY1 643 13 59700 1200 1085.5 0.2971 0.0018 10.77 0.31 16540 110 1133.1 4.8 15200 900
SlardelHuao CYC16471 -20.30860 48.50030 CVC16471A(12 623 12 45900 1000 10914 0.2964 0.0015 56.67 0.36 16450 110 1139.8 4.1 15300 200
180
Bibliography
Ainsworth, N.R., Burnett, R.D., Kontrovitz, M., 1990. Ostracod colour change by thermal
alteration, offshore Ireland and Western UK. Marine and Petroleum Geology 7, 288-297.
Alam, M.S., Cheng, T., 2014. Uranium release from sediment to groundwater: Influence of
water chemistry and insights into release mechanisms. Journal of Contaminant Hydrology
164, 72-87. doi:10.1016/j.jconhyd.2014.06.001.
Allmendinger, R.W., Jordan, T.E., Kay, S.M., Isacks, B.L., 1997. The Evolution of the
Altiplano-Puna Plateau of the Central Andes. Annual Review of Earth and Planetary
Sciences 25, 139-174. doi:10.1146/annurev.earth.25.1.139.
Ames, L.L., McGarrah, J.E., Walker, B.A., 1983a. Sorption of Trace Constituents from
Aqueous Solutions onto Secondary Minerals. II. Radium. Clays and Clay Minerals 31,
321-334.
Ames, L.L., McGarrah, J.E., Walker,.B.A., 1983b. Sorption of Uranium and Radium by
Biotite, Muscovite, and Phlogopite. Clays and Clay Minerals 31, 343-351.
Ames, L.L., McGarrah, J.E., Walker, B.A., Salter, P.F., 1983c. Uranium and ra-
dium sorption on amorphous ferric oxyhydroxide. Chemical Geology 40, 135-148.
doi:10.1016/0009-2541(83)90095-5.
Aravena, R., Suzuki, 0., Pefia, H., Pollastri, A., Fuenzalida, H.,.Grilli, A., 1999. Isotopic
composition and origin of the precipitation in Northern Chile. Applied Geochemistry
14, 411-422. doi:10.1016/S0883-2927(98)00067-5.
Arp, G., 1995. Lacustrine bioherms, spring mounds, and marginal carbonates of the Ries-
impact-crater (Miocene, Southern Germany). Facies 33, 35-89. doi:10.1007/BF02537444.
Arp, G.,Thiel, V., Reimer, A., Michaelis, W., Reitner, J., 1999. Biofilm exopolymers con-
trol microbialite formation at thermal springs discharging into the alkaline Pyramid Lake,
Nevada, USA. Sedimentary Geology 126, 159-176. doi:10.1016/S0037-0738(99)00038-X.
181
Arz, H.W., Pitzold, J., Wefer, G., 1998. Correlated Millennial-Scale Changes in
Surface Hydrography and Terrigenous Sediment Yield Inferred from Last-Glacial
Marine Deposits off Northeastern Brazil. Quaternary Research 50, 157-166.
URL: http://www.sciencedirect.com/science/article/pii/S033589498919925,
doi:10.1006/qres.1998.1992.
Avicenna, 1027. The Book of Healing.
Baker, P.A., Seltzer, G.O., Fritz, S.C., Dunbar, R.B., Grove, M.J., Tapia, P.M., Cross, S.L.,
Rowe, H.D., Broda, J.P., 2001. The history of South American tropical precipitation for
the past 25,000 years. Science 291, 640-643. doi:10.1126/science.291.5504.640.
Bakke, J., Lie, 0., Heegaard, E., Dokken, T., Haug, G.H., Birks, H.H., Dulski, P., Nilsen,
T., 2009. Rapid oceanic and atmospheric changes during the Younger Dryas cold period.
Nature Geoscience 2, 202-205. URL: http://www.nature.com/articles/ngeo439,
doi:10.1038/ngeo439.
Balch, D.P., Cohen, A.S., Schnurrenberger, D.W., Haskell, B.J., Valero-Garc6s, B.L., Bek,
J.W., Cheng, H., Edwards, R.L., 2005. Ecosystem and paleohydrological response to
Quaternary climate change in the Bonneville Basin, Utah. Palaeogeography, Palaeocli-
matology, Palaeoecology 221, 99-122. doi:10.1016/j.palaeo.2005.01.013.
Balsam, W.L., Deaton, B.C., Damuth, J.E., 1999. Evaluating optical lightness as a
proxy for carbonate content in marine sediment cores. Marine Geology 161, 141-153.
doi:10.1016/S0025-3227(99)00037-7.
Bar-Matthews, M., Ayalon, A., Gilmour, M., Matthews, A., Hawkesworth, C.J., 2003. Sea-
land oxygen isotope relationships from planktonic foraminifera and speleothems in the
Eastern Mediterranean region and their implication for paleorainfall during interglacial
intervals. Geochimica et Cosmochimica Acta 67, 3181-3199.
Bard, E., Fairbanks, R.G., Hamelin, B., Zindler, A., Chi Track
Hoang, 1991. Uranium-234 anomalies in corals older than 150,000
years. Geochimica et Cosmochimica Acta 55, 2385-2390. URL:
https://linkinghub.elsevier.com/retrieve/pii/001670379190115L,
doi:10.1016/0016-7037(91)90115-L.
Bard, E., Rostek, F., Turon, J.L., Gendreau, S., 2000. Hydrological Impact of
Heinrich Events in the Subtropical Northeast Atlantic. Science 289, 1321-1324.
doi:10.1126/science.289.5483.1321.
Bartlein, P.J., Harrison, S.P., Brewer, S., Connor, S., Davis, B.A.S., Gajew-
ski, K., Guiot, J., Harris6n'Prentice, T.I., Henderson, A., Peyron, 0., Pren-
tice, I.C., Scholze, M., Seppi, H., Shuman, B., Sugita, S., Thompson, R.S.,
182
Viau, A.E., Williams, J., Wu, H., 2011. Pollen-based continental climate recon-
structions at 6 and 21 ka: a global synthesis. Climate Dynamics 37, 775-802.
URL: http: //link. springer. com/10.1007/s00382-010-0904-1, doi:10.1007/s00382-
010-0904-1.
Bateman, H., 1910. The solution of a system of differential equations occurring in the
theory of radio-active transformations. Proc. Cambridge Phil. Soc. 15, 423-427.
Becquerel, H., 1896. On the rays emitted by phosphorescence. Compt. Rend. Hebd. Seances
Acad. Sci. 122, 420-421.
Behling, H., W. Arz, H., Pitzold, J., Wefer, G., 2000. Late Quaternary vegetational
and climate dynamics in northeastern Brazil, inferences from marine core GeoB 3104-1.
Quaternary Science Reviews 19, 981-994. doi:10.1016/SO277-3791(99)00046-3.
Benson, L., 1994. Carbonate deposition, Pyramid Lake subbasin, Nevada: 1. Sequence of
formation and elevational distribution of carbonate deposits (Tufas). Palaeogeography,
Palaeoclimatology, Palaeoecology 109, 55-87. doi:10.1016/0031-0182(94)90118-X.
Benson, L., Kashgarian, M., Rubin, M., 1995. Carbonate deposition, Pyramid Lake
subbasin, Nevada: 2. Lake levels and polar jet stream positions reconstructed from
radiocarbon ages and elevations of carbonates (tufas) deposited in the Lahontan
basin. Palaeogeography, Palaeoclimatology, Palaeoecology 117, 1-30. doi:10.1016/0031-
0182(94)00103-F.
Benson, R.H., 1984. Estimating greater paleodepths with ostracodes, especially in past
thermospheric oceans. Palaeogeography, Palaeoclimatology, Palaeoecology 48, 107-141.
doi:10.1016/0031-0182(84)90093-2.
Bills, B.G., de Silva, S.L., Currey, D.R., Emenger, R.S., Lillquist, K.D., Donnellan, A.,
Worden, B., 1994. Hydro-isostatic deflection and tectonic tilting in the central Andes:
Initial results of a GPS survey of Lake Minchin shorelines. Geophysical Research Letters
21, 293-296. doi:10.1029/93GL03544.
Bischoff, J.L., aA. Fitzpatrick, J., Rosenbauer, R.J., 1993a. The Solubility and Stabilization
of ikaite (CaCO3 6H20) from 0' to 25°C: Environmental and Paleoclimatic Implications
for Thinolite Tufa. doi:10.1086/648194.
Bischoff, J.L., Fitzpatrick, J.a., 1991. U-series dating of impure carbonates: An isochron
technique using total-sample dissolution. Geochimica et Cosmochimica Acta 55, 543-554.
doi:10.1016/0016-7037(91)90011-S.
Bischoff, J.L., Rosenbauer, R.J., Smith, G.I., 1985. Uranium-Series Dating of Sediments
from Searles Lake: Differences Between Continental'and Marine Climate Records. Sci-
ence 227, 1222-1224.
183
Bischoff, J.L., Stine, S., Rosenbauer, R.J., Fitzpatrick, J.a., Stafford, T.W., 1993b. Ikaite
precipitation by mixing of shoreline springs and lake water, Mono Lake, California, USA.
Geochimica et Cosmochimica Acta 57, 3855-3865. doi:10.1016/0016-7037(93)90339-X.
Blaauw, M., Christen, J.A., 2011. Flexible paleoclimate age-depth models us-
ing an autoregressive gamma process. Bayesian Analysis 6, 457-474. URL:
http://projecteuclid.org/euclid.ba/1339616472, doi:10.1214/11-BA618.
Blanchard, R.L., Cheng, M.H., Potratz, H.A., 1967. Uranium and Thorium Series Dise-
quilibria in Recent and Fossil Marine Molluscan Shells. Journal of Geophysical Research
72, 4745-4757.
Blard, P.H., Lav6, J., Farley, K.A., Fornari, M., Jim6nez, N., Ramirez, V., 2009. Late local
glacial maximum in the Central Altiplano triggered by cold and locally-wet conditions
during the paleolake Tauca episode (17-15 ka, Heinrich 1). Quaternary Science Reviews
28, 3414-3427. doi:10.1016/j.quascirev.2009.09.025.
Blard, P.H., Sylvestre, F., Tripati, A.K., Claude, C., Causse, C., Coudrain,
A., Condom, T., Seidel, J.L., Vimeux, F., Moreau, C., Dumoulin, J.P.,
Lav6, J., 2011. Lake highstands on the Altiplano (Tropical Andes) con-
temporaneous with Heinrich 1 and the Younger Dryas: New insights from
14C, U-Th dating and 6180 of carbonates. Quaternary Science Reviews
30, 3973-3989. URL: http://dx.doi.org/10.1016/j.quascirev.2011.11.001,
doi:10.1016/j.quascirev.2011.11.001.
Bobst, A.L., Lowenstein, T.K., Jordan, T.E., Godfrey, L.V., Ku, T.L., Luo, S.,
2001. A 106 ka paleoclimate record from drill core of the Salar de Ata-
cama, northern Chile. Palaeogcography, Palaeoclimatology, Palaeoecology 173, 21-
42. URL: http://linkinghub.elsevier.com/retrieve/pii/S003101820100308X,
doi:10.1016/S0031-0182(01)00308-X.
Bold, H.C., Wyynne, M.J., 1978. Introduction to the algae: Structure and reproduction.
Prentice Hall.
Bone, S.E., Dynes, J.J., Cliff, J., Bargar, J.R., 2017. Uranium(IV) adsorption by natural or-
ganic matter in anoxic sediments. Proceedings of the National Academy of Sciences 114,
201611918. URL: http://www.pnas.org/lookup/doi/10.1073/pnas.1611918114,
doi:10.1073/pnas.1611918114.
Bourdon, B.,.Henderson, G.M., Lundstrom, C.C., Turner, S.P. (Eds.), 2003. Uranium-
Series Geochemistry. volume 52. The Mineralogical Society of America.
Braconnot, P., Harrison, S.P., Kageyama, M., Bartlein, P.J., Masson-Delmette,
V., Abe-Ouchi, A., Otto-Bliesner, B., Zhao, Y., 2012. Evaluation of climate
184
models using palaeoclimatic data. Nature Climate Change 2, 417-424. URL:
http://dx. doi. org/10.1038/nclimate456, doi:10.1038/nclimate456.
Braconnot, P., Otto-Bliesner, B., Harrison, S., Joussaume, S., Peterchmitt, J.Y., Abe-
Ouchi, A., Crucifix, M., Driesschaert, E., Fichefet, T., Hewitt, C.D., Kageyama, M.,
Kitoh, A., Lain6, A., Loutre, M.F., Marti, 0., Merkel, U., Ramstein, G., Valdes, P.,
Weber, S.L., Yu, Y., Zhao, Y., 2007. Results of PMIP2 coupled simulations of the Mid-
Holocene and Last Glacial Maximum-Part 1: experiments and large-scale features.
Climate of the Past 3, 261-277. doi:10.5194/cp-3-261-2007.
Broecker, W.S., 1963. A Preliminary Evoluation of Uranium Series Inequilibrium as a Tool
for Absolute Age Measurement on Marine Carbonates. Journal of Geophysical Research
68, 2817-2834.
Broecker, W.S., Orr, P.C., 1958. Radiocarbon chronology of Lake Lahontan and Lake Bon-
neville. Bulletin of the Geological Society of America 69, 1009-1032. doi:10.1130/0016-
7606(1958)69[1009:RCOLLA]2.0.CO;2.
Buchardt, B., Israelson, C., Seaman, P., Stockmann, G., 2001. Ikaite tufa towers in Ikka
Fjord, Southwest Greenland: Their formation by mixing of seawater and alkaline spring
water. Journal of Sedimentary Research 71, 176-189. doi:10.1306/042800710176.
Buchardt, B., Seaman, P., Stockmann, G., Vous, M., Wilken, U., Duwel, L., Kris-
tiansen, A., Jenner, C., Whiticar, M.J., Kristensen, R.M., Petersen, G.H., Thorb-
jorn, L., 1997. Submarine columns of ikaite tufa. Nature 390, 129-130. URL:
http: //dx. do i. org/10 . 1038/36474, doi:10.1038/36474.
Chabaux, F., Bourdon, B., Riotte, J., 2008. Chapter 3 U-Series Geo-
chemistry in Weathering Profiles, River Waters and Lakes, in: Ra-
dioactivity in the Environment. volume 13, pp. 49-104. URL:
https: //linkinghub.,elsevier. com/retrieve/pii/S1569486007000034,
doi:10.1016/S1569-4860(07)00003-4.
Chabaux, F., Riotte, J., Dequincey, 0., 2003. U-Th-Ra Fractionation During Weath-
ering and River Transport. Reviews in Mineralogy and Geochemistry 52, 533-576.
URL: https://pubs.geoscienceworld.org/rimg/article/52/1/533-576/87464,
doi:10.2113/0520533.
Chappaz, A., Gobeil, C., Tessier, A., 2010. Controls on uranium distribution in lake sedi-
ments. Geochimica et Cosmochimica Acta 74, 203-214. doi:10.1016/j.gca.2009.09.026.
Cheng, H., Adkins, J., Edwards, R., Boyle, E.A., 2000. U-Th dating
of deep-sea corals. Geochimica et Cosmochimica Acta 64, 2401-2416.
URL: http://linkinghub.elsevier.com/retrieve/pii/S016703799004226,
doi:10.1016/S0016-7037(99)00422-6.
185
Cheng, H., Edwards, R.L., Broecker, W.S., Denton, G.H., Kong, X., Wang, Y., Zhang, R.,
Wang, X., 2009. Ice age terminations. Science 326, 248-252. doi:10.1126/science.1177840.
Cheng, H., Edwards, R.L., Sinha, A., Spdtl, C., Yi, L., Chen, S., Kelly, M.,
Kathayat, G., Wang, X., Li, X., Kong, X., Wang, Y., Ning, Y., Zhang, H., 2016.
The Asian monsoon over the past 640,000 years and ice age terminations. Na-
ture 534, 640-646. URL: http: //www.nature. com/do if inder/l10. 1038/nature18591,
doi:10.1038/naturel8591.
Cheng, H., Lawrence Edwards, R., Shen, C.C., Polyak, V.J., Asmerom, Y., Wood-
head, J., Hellstrom, J., Wang, Y., Kong, X., Sptl, C., Wang, X., Calvin Alexan-
der, E., 2013a. Improvements in 230Th dating, 230Th and 234U half-life val-
ues, and U-Th isotopic measurements by multi-collector inductively coupled plasma
mass spectrometry. Earth and Planetary Science Letters 371-372, 82-91. URL:
http://dx.doi.org/10.1016/j.epsl.2013.04.006, doi:10.1016/j.epsl.2013.04.006.
Cheng, H., Sinha, A., Cruz, F.W., Wang, X., Edwards, R.L., D'Horta, F.M.,
Ribas, C.C., Vuille, M., Stott, L.D., Auler, A.S., 2013b. Climate change pat-
terns in Amazonia and biodiversity. Nature Communications 4, 1411. URL:
http://www.ncbi.n1m.nih.gov/pubmed/23361002, doi:10.1038/ncomms2415.
Chiang, J.C., Lee, S.Y., Putnam, A.E., Wang, X., 2014. South Pacific Split
Jet, ITCZ shifts, and atmospheric North-South linkages during abrupt climate
changes of the last glacial period. Earth and Planetary Science Letters 406, 233-
246. URL: http: //linkinghub. els evier. com/retrieve/pii/S12821X14005676,
doi:10.1016/j.epsl.2014.09.012.
Cobbing, E.J., Pitcher, W.S., WIlson, J.J., Baldock, J.W., Taylor, W.P., McCourt, W.,
Snelling, N.J., 1981. The geology of the Western Cordillera of northern Peru. volume 5.
Great Britain Institute of Geological Sciences, Natural Environmental Research Council.
Coe, M.T., Harrison, S.P., 2002. The water balance of northern Africa during the
mid-Holocene: an evaluation of the 6 ka BP PMIP simulations. Climate Dynam-
ics 19, 155-166. URL: http://link.springer.com/10.1007/s00382-001-0219-3,
doi:10.1007/s00382-001-0219-3.
Colman, S., Kaufman, D., Bright, J., Heil, C., King, J., Dean, W., Rosen-
baum, J., Forester, R., Bischoff, J., Perkins, M., McGeehin, J., 2006.
Age model for a continuous, ca 250-ka Quaternary lacustrine record from
Bear Lake, Utah-Idaho. Quaternary Science Reviews 25, 2271-2282.
URL: http: //linkinghub. elsevier. com/retr ieve /pii/S0277379106001806,
doi:10.1016/j.quascirev.2005.10.015.
186
Colman, S.M., 1996. Continental Drilling for Paleoclimatic Records: Recommendations
from an International Workshop. Technical Report. GeoForschungsZentrum. Potsdam,
Germany.
Council, T.C., Bennett, P.C., 1993. Geochemistry of ikaite formation at Mono Lake,
California: Implications for the origin of tufa mounds. Geology 21, 971-974.
doi:10.1130/0091-7613(1993)021<0971.
Cruz, F.W., Burns, S.J., Jercinovic, M., Karmann, I., Sharp, W.D., Vuille, M., 2007.
Evidence of rainfall variations in Southern Brazil from trace element ratios (Mg/Ca and
Sr/Ca) in a Late Pleistocene stalagmite. Geochimica et Cosmochimica Acta 71, 2250-
2263. URL: http: //linkinghub. elsevier. com/retrieve/pii/S001670370700
0 8 04 ,
doi:10.1016/j.gca.2007.02.005.
Cruz, F.W., Burns, S.J., Karmann, I., Sharp, W.D., Vuille, M., Cardoso, A.O., Fer-
rari, J.A., Silva Dias, P.L., Viana Jr., 0., 2005. Insolation-driven changes in atmo-
spheric circulation over the past 116,000 years in subtropical Brazil. Nature 434, 63-66.
doi:10.1038/nature03365.
De Deckker, P., 2002. Ostracod Palaeoecology. The Ostracoda: Applications in Quaternary
Research, Geophysical Monograph 131 , 121-134doi:10.1029/131GM06.
Direcci6n General de Aguas, 1987. Balance hidrico de Chile. URL:
http://sad.dga.cl/ipac20/ipac.jsp?session=1349N4543565L.131053&profile=cirh&uri=link=3100006
Donohoe, A., Marshall, J., Ferreira, D., McGee, D., 2013. The relationship between ITCZ
location and cross-equatorial atmospheric heat transport: From the seasonal cycle to
the last glacial maximum. Journal of Climate 26, 3597-3618. doi:10.1175/JCLI-D-12-
00467.1.
Dunn, J.R., 1953. The origin of the deposits of tufa in Mono
Lake. Journal of Sedimentary Research 23, 18-23. URL:
http://ar chives. datapages. com/data/sepm/journals/vol-32/data/023/023001/0018.htm.
Durand, S., Chabaux, F., Rihs, S., Duringer, P., Elsass, P., 2005. U isotope ratios as tracers
of groundwater inputs into surface waters: example of the upper rhine hydrosystem.
Chemical Geology 220, 1-19.
Dutton, A., Rubin, K., McLean, N., Bowring, J., Bard, E., Edwards, R.L., Hen-
derson, G.M., Reid, M.R., Richards, D.A., Sims, K., Walker, J.D., Yokoyama, Y.,
2017. Data reporting standards for publication of U-series data for geochronol-
ogy and timescale assessment in the earth sciences. Quaternary Geochronol-
ogy 39, 142-149. URL: http://dx.doi.org/10.1016/j.quageo.2017.03.001,
doi:10.1016/j.quageo.2017.03.001.
187
Edwards, R., Gallup, C., Cheng, H., 2003. Uranium-series Dating of Ma-
rine and Lacustrine Carbonates. Reviews in Mineralogy and Geochemistry 52,
363-405. URL: http://rimg.geoscienceworld.org/cgi/doi/10.2113/0520363,
doi:10.2113/0520363.
Edwards, R.L., 1988. High Precition Thorium-230 Ages of Corals and the Timing of
Sea Level Fluctuations in the Late Quaternary. Ph.D. thesis. California Insititute of
Technology.
Edwards, R.L., Chen, J., Wasserburg, G., 1987. 238U-234U-230Th-232Th systematics
and the precise measurement of time over the past 500,000 years. Earth and Planetary
Science Letters 81, 175-192. doi:10.1016/0012-821X(87)90154-3.
Falvey, M., Garreaud, R.D., 2005. Moisture variability over the South American Altiplano
during the South American low level jet experiment (SALLJEX) observing season. Jour-
nal of Geophysical Research: Atmospheres 110, 1-12. doi:10.1029/2005JD006152.
Fleischer, R.L., 1982. Alpha-recoil damage and solution effects in minerals: uranium
isotopic disequilibrium and radon release. Geochimica et Cosmochimica Acta 46, 2191-
2201. URL: https://inkinghub.elsevier.com/retrieve/pii/0016703782901946,
doi:10.1016/0016-7037(82)90194-6.
Flusche, M.A., Seltzer, G., Rodbell, D., Siegel, D., Samson, S., 2005. Con-
straining water sources and hydrologic processes from the isotopic analysis of wa-
ter and dissolved strontium, Lake Junin, Peru. Journal of Hydrology 312, 1-
13. URL: https://linkinghub.elsevier.com/retrieve/pii/S022169405000727,
doi:10.1016/j.jhydrol.2005.02.021.
Francke, A., Wagner, B., Just, J., Leicher, N., Gromig, R., Baumgarten, H., Vogel, H.,
Lacey, J.H., Sadori, L., Wonik, T., Leng, M.J., Zanchetta, G., Sulpizio, R., Giac-
cio, B., 2016. Sedimentological processes and environmental variability at Lake Ohrid
(Macedonia, Albania) between 637 ka and the present. Biogeosciences 13, 1179-1196.
doi:10.5194/bg-13-1179-2016.
Frechen, M., Sierralta,' M., Oezen, D., Urban, B., 2007. Uranium-series dating of peat
from central and Northern Europe, in: Sirocko, F:; Claussen, M., Gofli, M.F.S., Litt, T.
(Eds.), Developments in Quaternary Science: The Climate of Past Interglacials. Elsevier.
volume 7. chapter 8, pp. 83-117. doi:10.1016/S1571-0866(07)80033-9.
Fritz, S.C., Baker, P.a., Lowenstein, T.K., Seltzer, G.O., Rigsby, C.a., Dwyer, G.S., Tapia,
P.M., Arnold, K.K., Ku, T.L., Luo, S., 2004. Hydrologic variation during the last 170,000
years in the southern hemisphere tropics of South America. Quaternary Research 61,
95-104. doi:10.1016/j.yres.2003.08.007.
188
Fritz, S.C., Baker, P.A., Seltzer, G.O., Ballantyne, A., Tapia, P., Cheng, H., Edwards, R.L.,
2007. Quaternary glaciation and hydrologic variation in the South American tropics as
reconstructed from the Lake Titicaca drilling project. Quaternary Research 68, 410-420.
doi:10.1016/j.yqres.2007.07.008.
Fuenzalida, H., Rutllant, J., 1987. Origen del vapor de agua que precipita sobre el Altiplano
de Chile.
Gallup, C.D., Edwards, R.L., Johnson, R.G., 1994. The timing of high sea levels over the
past 200,000 years. Science 263, 796-800.
Gandu, A.W.,,Silva Dias, P.L., 1998. Impact of tropical heat sources on the South American
tropospheric upper circulation and subsidence. Journal of Geophysical Research 103,
6001. doi:10.1029/97JD03114.
Garreaud, R., Vuille, M., Clement, A.C., 2003. The climate of the Altiplano: Observed
current conditions and mechanisms of past changes. Palaeogeography, Palaeoclimatology,
Palaeoecology 194, 5-22. doi:10.1016/S0031-0182(03)00269-4.
Garreaud, R.D., 1999. Multiscale Analysis of the Summertime Precipitation over
the Central Andes. Monthly Weather Review 127, 901-921. doi:10.1175/1520-
0493(1999)127<0901:MAOTSP>2.0.CO;2.
Garreaud, R.D., Aceituno, P., 2001. Interannual Rainfall Variability over
the South American Altiplano. Journal of Climate 14, 2779-2789. URL:
http://jcurnals.ametsoc.org/doi/abs/10.1175/1520-0442(2001)014%3C2779:IRVOTSX/3E2.0.CO;2,
doi:10.1175/1520-0442(2001)014<2779:IRVOTS>2.0.CO;2.
Garreaud, R.D., Vuille, M., Compagnucci, R., Marengo, J., 2009. Present-
day South American climate. Palaeogeography, Palaeoclimatology, Palaeoecol-
ogy 281. 180-195. URL: http://dx.doi.org/10.1016/j.palaeo.2007.10.032,
doi:10.1016/j.palaeo.2007.10.032.
Garreaud, R.D., Wallace, J.M., 1997. The Diurnal March of Convective Cloudiness
over the Americas. Monthly Weather Review 125, 3157-3171. doi:10.1175/1520-
0493(1997)125<3157:TDMOCC>2.0.CO;2.
Gayo, E.M., Latorre, C., Jordan, T.E , Nester, P.L., Estay, S.A., Ojeda, K.F.,
Santoro, C.M., 2012. Late Quaternary hydrological and ecological changes in
the hyperarid core of the northern Atacama Desert (-21°S). Earth-Science Re-
views 113, 120-140. URL: http://dx.doi.org/10.1016/j.earscirev.2012.04.003,
doi:10.1016/j.earscirev.2012.04.003.
189
Geyh, M.A., Mu, H., 2005. Numerical 230Th/U dating and a palynological review
of the Holsteinian/Hoxnian Interglacial. Quaternary Science Reviews 24, 1861-1872.
doi:10.1016/j.quascirev.2005.01.007.
Gierlowski-Kordesch, E.H., 2010. Lacustrine carbonates. Developments in Sedimentology:
Carbonates in Continental Settings: Facies, Environments and Processes 61, 2-70.
Giosan, L., Flood, R.D., Aller, R.C., 2002. Paleoceanographic significance of sediment
color on western North Atlantic drifts: I. Origin of color. Marine Geology 189, 25-41.
URL: https://linkinghub.elsevier.com/retrieve/pii/S025322702003225
https://linkinghub.elsevier.com/retrieve/pii/S0025322702003213,
doi:10.1016/S0025-3227(02)00321-3.
Golubic, S., 1973. Relationship between blue-green algae and carbonate deposits. The
Biology of Blue-Gren Algae, Botanical Monograph 9 , 434-472.
Grimm, A.M., 2011. Interannual climate variability in South America: Impacts on seasonal
precipitation, extreme events, and possible effects of climate change. Stochastic Envi-
ronmental Research and Risk Assessment 25, 537-554. doi:10.1007/s00477-010-0420-1.
Groot, M.H.M., Bogoti, R.G., Lourens, L.J., Hooghiemstra, H., Vriend, M., Berrio, J.C.,
Tuenter, E., Van Der Plicht, J., Van Geel, B., Ziegler, M., Weber, S.L., Betancourt,
A., Contreras, L., Gaviria, S., Giraldo, C., Gonzalez, N., Jansen, J.H.F., Konert, M.,
Ortega, D., Rangel, 0., Sarmiento, G., Vandenberghe, J., Van Der Hammen, T., Van Der
Linden, M., Westerhoff, W., 2011. Ultra-high resolution pollen record from the northern
Andes reveals rapid shifts in montane climates within the last two glacial cycles. Climate
of the Past 7, 299-316. doi:10.5194/cp-7-299-2011.
Grosjean, M., Geyh, M.A., Messerli, B., Schotterer, U., 1995. Late-glacial and early
Holocene lake sediments, ground-water formation and climate in the Atacama Altiplano
22-24 OS. Journal of Paleolimnology 14, 241-252. doi:10.1007/BF00682426.
Grosjean, M., van Leeuwen, J.F.N., van der Knaap, W.O., Geyh, M.A., Ammann, B.,
Tanner, W., Messerli, B., Ndfiez, L.A., Valero-Garc6s, B.L., Veit, H., 2001. A 22,000
C-14 year BP sediment and pollen record of climate change from Laguna Miscanti
(230S), northern Chile. Global and Planetary Change 28, 35-51. doi:10.1016/SO921-
8181(00)00063-1.
Grzymko, T.J., Marcantonio, F., McKee, B.A., Mike Stewart, C., 2007. Tem-
poral variability of uranium . concentrations and 234U/238U activity ratios
in the Mississippi river and its tributaries. Chemical Geology 243, 344-
356. URL: https://linkinghub.elsevier.com/retrieve/pii/S009254107002434,
doi:10.1016/j.chemgeo.2007.05.024.
190
Haase-Schramm, A., Goldstein, S.L., Stein, M., 2004. U-Th dating of Lake
Lisan (late Pleistocene dead sea) aragonite and implications for glacial east
Mediterranean climate change. Geochimica et Cosmochimica Acta 68, 985-
1005. URL: http://linkinghub.elsevier.com/retrieve/pii/S0016703703005647,
doi:10.1016/j.gca.2003.07.016.
Halbach, P., Von Borstel, D., Gundermann, K.d., 1980. The uptake of uranium by organic
substances in a peat bog environment on a granitic bedrock. Chemical Geology 29,
117-138.
Hamelin, B., Bard, E., Zindler, A., Fairbanks, R.G., 1991. 234U/238U
mass spectrometry of corals: How accurate is the UTh age of the last in-
terglacial period? Earth and Planetary Science Letters 106, 169-180.
URL: https://linkinghub.elsevier.com/retrieve/pii/0012821X9190070X,
doi:10.1016/0012-821X(91)90070-X.
Hansen, B.C.S., Wright, H.E., Bradbury, J.P., 1984. Pollen studies in the Junin
area, central Peruvian Andes. Geological Society of America Bulletin 95, 1454. URL:
https://pubs.geoscienceworld.org/gsabulletin/article/95/12/1454-1465/186834,
doi:10.1130/0016-7606(1984)95<1454:PSITJA>2.0.CO;2.
Hatfield, R., Woods, A., Lehmann, S., Weidhaas, N., Chen, C., Kiick, J., Pierdominici, S.,
Stoner, J., Abbott, M., Rodbell, D., 2019. Stratigraphic correlation and splice generation
for sediments recovered from a large-lake drilling project: an example from lake junin,
peru. Journal of Paleolimnology , 1-18doi:10.1007/s10933-019-00098-w.
Helmke, J.P., Schulz, M., Bauch, H.A., 2002. Sediment-Color Record from
the Northeast Atlantic Reveals Patterns of Millennial-Scale Climate Variabil-
ity during the Past 500,000 Years. Quaternary Research 57, 49-57. URL:
https://www.cambridge.'org/core/product/identifier/S033589400010565/type/journal-article,
doi:10.1006/qres.2001.2289.
Henderson, G.M., 2002. Seawater (234U/238U) during the last 800 thou-
sand years. Earth and Planetary Science Letters 199, 97-110. URL:
https: //linkinghub. elsevier. com/retrieve/pii/S0012821X2005563,
doi:10.1016/S0012-821X(02)00556-3.
Henderson, G.M., Anderson, R.F., 2003. The U-series Toolbox for Paleo-
ceanography. Reviews in Mineralogy and Geochemistry 52, 493-531. URL:
https://pubs. geoscienceworld.org/rimg/article/52/1/493-531/87463,
doi:10.2113/0520493.
Henderson, G.M., Robinson, L.F., Cox, K., Thomas, A.L., 2006. Recognition of
non-Milankovitch sea-level highstands at 185 and 343 thousand years ago from
191
U-Th dating of Bahamas sediment. Quaternary Science Reviews 25, 3346-
3358. URL: https: // linkinghub. elsevier. com/retrieve/pii/S277379106001363,
doi:10.1016/j.quascirev.2006.03.003.
Holmes, A., 1911. The association of lead with uranium in rock-minerals, and its application
to the measurement of geological time. Proceedings of the Royal Society of London. Series
A, Containing Papers of a Mathematical and Physical Character 85, 248-256.
Holmes, J., 1998. Chapter 10: The palaeoenvironmental significance of iron and manganese
in non-marine ostracod shells: a preliminary analysis, in: Holmes, J., Lynch, K. (Eds.),
The Kingston Papers: a geographical perspective on the environment, economy, and
society. Kingston University School of Geography. chapter Holmes, pp. 198-212.
Holmes, J.A., Chivas, A.R., 2002. Ostracod shell chemistry - overview. Geophysical Mono-
graph Series 131, 185-204. doi:10.1029/131GM10.
Houze, R.A., 1997. Stratiform Precipitation in Regions of Convection: A Meteoro-
logical Paradox? Bulletin of the American Meteorological Society 78, 2179-2196.
doi:10.1175/1520-0477(1997)078<2179:SPIROC>2.0.CO;2.
Hudson, A.M., Quade, J., 2013. Long-term east-west asymmetry in monsoon rainfall on
the Tibetan Plateau. Geology 41, 351-354. doi:10.1130/G33837.1.
IAEA/WMO, 2015. Global Network of Isotopes in Precipitation. Technical Report. URL:
http://www.iaea.org/water.
Inden, R.F., Coalson, E.B., Field, C., 1996. Phosphoria Formation (Permian) Cycles in the
Bighorn Basin, Wyoming, with Emphasis on the Ervay Member, in: Longman, M.W.,
Sonnenfield, M.D. (Eds.), Paleozoic Systems of the Rocky Mountain Region. Society for
Sedimentary Geology, pp. 379-404.
Isaks, B.L., 1988. Uplift of the Central Andean Plateau and bending of the Bolivian
orocline. Journal of Geophysical Research 93, 3211-3231. doi:10.1029/JB093iB4p3211.
Israelson, C., Bjdrck, S., Hawkesworth, C.J., Possnert, G., 1997. Direct U-Th dating
of organic- and carbonate-rich lake sediments from southern Scandinavia. Earth and
Planetary Science Letters 153, 251-263. doi:10.1016/S0012-821X(97)00184-2.
Ivanovich, M., Harmon, R., 1992. Uranium-series-Disequilibrium: Applications to Earth,
Marine, andEnvironmental Sciences. Oxford University Press, New York.
Jaffey, A.H., Flynn, K.F., Glendenin, L.E., Bentley, W.C., Essling, A.M., 1971. Precision
measurement of half-lives and specific activities of U235 and U238. Physical Review C
4, 1889-1906. doi:10.1103/PhysRevC.4.1889.
192
James, N.P., Bourque, P.A., 1992. Reefs and mounds. Facies Models: Response to Sea
Level Change , 323-347.
James, N.P., Choquette, P.W., 1984. Diagenesis 9. Limestones-the me-
teoric diagenetic environment. Geoscience Canada 11, 161-194. URL:
http://journals.hil.unb.ca/index.php/GC/article/viewArticle/3395.
Jansen, J.H.F., Woensdregt, G.F., Kooistra, M.J., Van Der Gaast, S.J., 1987. Ikaite
pseudomorphs in the Zaire deep-sea fan: an intermediate between calcite and porous
calcite. Geology 15, 245-248. doi:10.1130/0091-7613(1987)15<245:IPITZD>2.0.CO.
Ji, J., Shen, J., Balsam, W.L., Chen, J., Liu, L., Liu, X., 2005. Asian monsoon oscillations in
the northeastern Qinghai-Tibet Plateau since the late glacial as interpreted from visible
reflectance of Qinghai Lake sediments. Earth and Planetary Science Letters 233, 61-
70. URL: https: //linkinghub. elsevier. com/retr ieve/pii/iIS12821X05001275,
doi:10.1016/j.epsl.2005.02.025.
Jones, C., Carvalho, L.M.V., Liebmann, B., 2012. Forecast Skill of the
South Aimerican Monsoon System. Journal of Climate 25, 1883-1889.
URL: http://journals.ametsoc.org/doi/abs/10.1175/JCLI-D-11-00586.1,
doi:10.1175/JCLI-D-11-00586.1.
Joussaume, S., Taylor, K.E., Braconnot, P., Mitchell, J.F.B., Kutzbach, J.E., Harrison,
S.P., Prentice, I.C., Broccoli, A.J., Abe-Ouchi, A., Bartlein, P.J., Bonfils, C., Dong,
B., Guiot, J., Herterich, K., Hewitt, C.D., Jolly, D., Kim, J.W., Kislov, A., Kitoh,
A.,.Loutre, M.F., Masson, V., McAvaney, B., McFarlane, N., de Noblet, N., Peltier,
W.R., Peterschmitt, J.Y., Pollard, D., Rind, D., Royer, J.F., Schlesinger, M.E., Syk-
tus, J., Thompson, S., Valdes, P., Vettoretti, G., Webb, R.S., Wyputta, U., 1999.
Monsoon changes for 6000 years ago: Results of 18 simulations from the Paleoclimate
Modeling Intercomparison Project (PMIP). Geophysical Research Letters 26, 859-862.
doi:10.1029/1999GL900126.
Kanner, L.C., Burns, S.J., Cheng, H., Edwards, R.L., 2012. High-Latitude Forcing of
the South American Summer Monsoon During the Last Glacial. Science 335, 570-573.
doi:10.1126/science.1213397.
Karanovic, I., 2012. Recent freshwater ostracods of the world: Crustacea, Ostracoda,
Podocopida. Springer Science & Business Media.
Kaufman, A., Broecker, W., 1965. Comparison of Th230 and C14 Ages for Carbonate
Materials from Lakes Lahontan and Bonneville. Journal of Geophysical Research 70,
4039-4054.
193
Kaufman, A., Broecker, W.S., Ku, T.L., Thurber, D.L., 1971. The status of U-
series methods of mollusk dating. Geochimica et Cosmochimica Acta 35, 1155-1183.
doi:10.1016/0016-7037(71)90031-7.
Kaufman, A., Yechieli, Y., Gardosh, M., 1992. Reevaluation of the Lake-Sediment Chronol-
ogy in the Dead-Sea Basin, Israel, Based on New Th-230/U Dates. Quaternary Research
38, 292-304.
Kay, S.M., Coira, B.L., 2009. Shallowing and steepening subduction zones, continental
lithospheric loss, magmatism, and crustal flow under the Central Andean Altiplano-
Puna Plateau, in: Kay, S.M., Ramos, V.A., Dickinson, W.R. (Eds.), Backbone of the
Americas: Shallow Subduction, Plateau Uplift, and Ridge and Terrane Collision: Geo-
logical Society of America Memoir. The Geological Society of America. volume 204, pp.
229-259. doi:10.1130/2009.1204(11).
Kelly, S.D., Newville, M.G., Cheng, L., Kemner, K.M., Sutton, S.R., Fenter, P., Sturchio,
N.C., Sp6tl, C., 2003. Uranyl Incorporation in Natural Calcite. Environmental Science
& Technology 37, 1284-1287. URL: https: //pubs. acs. org/doi/10.102/es25962f,
doi:10.1021/es025962f.
Kelly, S.D., Rasbury, E.T., Chattopadhyay, S., Kropf, A.J., Kemner, K.M., 2006. Evi-
dence of a Stable Uranyl Site in Ancient Organic-Rich Calcite. Environmental Science
& Technology 40, 2262-2268. URL: https://pubs. acs. org/doi/10. 1021/es051970v,
doi:10.1021/es051970v.
Kigoshi, K., 1971. Alpha-Recoil Thorium-234: Dissolution into Water and
the Uranium-234/Uranium-238 Disequilibrium in Nature. Science 173, 47-
48. URL: http://www.sciencemag.org/cgi/doi/10.1126/science.173.3991.47,
doi:10.1126/science.173.3991.47.
Kiro, Y., Goldstein, S., Kushnir, Y., Stein, M., Lazar, B., 2018. Droughts and tropical-
induced rainfall in the east mediterranean during warm periods, in: Goldschmidt Ab-
stracts.
Kliem, P., Enters, D., Hahn, A., Ohlendorf, C., Lis6-Pronovost, A., St-Onge, G.,
Wastegird, S., Zolitschka, B., the Pasado Science Team, 2013. Lithology, radio-
carbon chronology and sedimentological interpretation of the lacustrine record from
Laguna Potrok Aike, southern Patagonia. Quaternary Science Reviews,71, 54-69.
doi:10.1016/j.quascirev.2012.07.019.
Knutti, R., Sedlieek, J., 2013. Robustness and uncertainties in the new
CMIP5 climate model projections. Nature Climate Change 3, 1-5. URL:
http://dx.doi.org/10.1038/nclimate1716, doi:10.1038/nclimatel7l6.
194
Kohfeld, K.E., Harrison, S.P., 2000. How well can we simulate past climates? Evaluat-
ing the models using global paleoenvironmental datasets (Kohfield and Harrison 2000).
Quaternary Science Reviews 19, 321-346.
K6nig, I., Haeckel, M., Drodt, M., Suess, E., Trautwein, A.X., 1999. Reactive
Fe(II) layers in deep-sea sediments. Geochimica et Cosmochimica Acta 63, 1517-
1526. URL: https://linkinghub.elsevier.com/retrieve/pii/S016703799001040,
doi:10.1016/S0016-7037(99)00104-0.
K6nig, I., Lougear, A., Bruns, P., Grntzner, J., Trautwein, A., Dullo, W.C., 2000. Iron
oxidation in sediment cores (Site 1062) during six months of storage in the Ocean Drilling
Program archive. Proceedings of the Ocean Drilling Program, Scientific Results. volume
172. pp. 1-11.
Kronfeld, J., 1974. Uranium deposition and Th-234 alpha-recoil: An
explanation for extreme U-234/U-238 fractionation within the Trin-
ity aquifer. Earth and Planetary Science Letters 21, 327-330. URL:
https://linkinghub.elsevier.com/retrieve/pii/0012821X74901691,
doi:10.1016/0012-821X(74)90169-1.
Kronfeld, J., Godfrey-Smith, D., Johannessen, D., Zentilli, M., 2004. Uranium series
isotopes in the avon valley, nova scotia. Journal of Environmental Radioactivity 73,
335-352.
Kronfeld, J., Vogel, J.C., 1991. Uranium isotopes in surface watersfrom southern Africa.
Earth and Planetary Science Letters 105, 191-195. doi:10.1016/0012-821X(91)90130-A.
Ku, T.L., Knauss, K.G., Mathieu, G.G., 1977. Uranium in open ocean: concen-
tration and isotopic composition. Deep Sea Research 24, 1005 - 1017. URL:
http: //www. sciencedirect. com/science/article/pii/0146629177905719,
doi:https://doi.org/10.1016/0146-6291(77)90571-9.
Ku, T.L., Liang, Z.C., 1984. The Dating of Impure Carbonates with Decay-Series Isotopes.
Nuclear Instruments and Methods in Physics Research 223, 563-571. doi:10.1360/zd-
2013-43-6-1064.
Ku, T.L., Luo, S., Lowenstein, T.K., Li, J., Spencer, R.J., 1998. U-Series Chronology of
Lacustrine Deposits in Death Valley, California. Quaternary Research 50, 261-275.
Kutterolf, S., Schindlbeck, J.C., Anselmetti, F.S., Ariztegui, D., Brenner, M., Cur-
tis, J., Schmid, D., Hodell, D., Mueller, A., P6rez, W., Schwalb, A., Frische, M.,
Wang, K.L., 2016. A 400-ka tephrochronological framework for Central America from
Lake Pet6n Itz6 (Guatemala) sediments. Quaternary Science Reviews 150, 200-220.
doi:10.1016/j.quascirev.2016.08.023.
195
Lambert, F., Delmonte, B., Petit, J., Bigler, M., Kaufmann, P., Hutterli, M., Stocker, T.,
Ruth, U., Steffensen, J., Maggi, V., 2008. Dust-climate couplings over the past 800,000
years from the EPICA Dome C ice core. Nature 452, 616-619. doi:10.1038/nature6763.
Lamy, F., Klump, J., Hebbeln, D., Wefer, G., 2000. Late Quaternary
rapid climate change in northern Chile. Terra Nova 12, 8-13. URL:
http://doi.wiley.com/10.1046/j.1365-3121.2000.00265.xX/5Cnpapers2://publication/doi/10.
doi:10.1046/j.1365-3121.2000.00265.x.
Langmuir, D., 1978. Uranium solution-mineral equilibria at low temperatures with appli-
cations to -sedimentary ore deposits. Geochimica et Cosmochimica Acta 42, 547-569.
Langmuir, D., Herman, J.S., 1980. The mobility of thorium in natural waters
at low temperatures. Geochimica et Cosmochimica Acta 44, 1753-1766. URL:
papers2://publication/uuid/5E9E5D34-242B-4117-8B39-BF4464A10B4B.
Larsen, D., 1994. Origin and paleoenvironmental significance of calcite pseudomorphs after
ikaite in the Oligocene Creede Formation, Colorado. Journal of Sedimentary Research
A64, 593-603.
Laskar, J., Robutel, P., Joutel, F., Gastineau, M., Correia, A.C.M., Levrard, B., 2004. A
long-term numerical solution for the insolation quantities of the Earth. Astronomy and
Astrophysics 428, 261-285. doi:10.1051/0004-6361:20041335.
Latorre, C., Betancourt, J.L., Arroyo, M.T.K., 2006. Late Quaternary vegetation and
climate history of a perennial river canyon in the Rio Salado basin (22S) of Northern
Chile. Quaternary Research 65, 450-466. doi:10.1016/j.yqres.2006.02.002.
Lenhart, J.J., Cabaniss, S.E., MacCarthy, P., Honeyman, B.D., 2000. Uranium (VI) com-
plexation with citric, humnic and fulvic acids. Radiochimica Acta 88, 345-353.
Lenters, J.D., Cook, K.H., 1997. On the Origin of the Bolivian High and Related Circula-
tion Features of the South American Climate. Journal of the Atmospheric Sciences 54,
656-678. doi:10.1175/1520-0469(1997)054<0656:OTOOTB>2.0.CO;2.
Li, J., Lowenstein, T.K., Brown, C.B., Ku, T.L., Luo, S., 1996. A 100 ka record of water
tables and paleoclimates from salt cores, Death Valley, California. Palaeogeography,
Palaeoclimatology, Palaeoecology 123, 179-203.
Lin, J.C., Broecker, W.S., Hemming, S.R., Hajdas, I., Anderson, R.F.,
Smith, G.I., Kelley, M., Bonani, G., 1998. A Reassessment of U-
Th and 14C Ages for Late-Glacial High-Frequency Hydrological Events
at Searles Lake, California. Quaternary Research 49, 11-23. URL:
http://www.sciencedirect.com/science/article/pii/S033589497919499,
doi:10.1006/qres.1997.1949.
196
Lisiecki, L.E., Raymo, M.E., 2005. A Pliocene-Pleistocene stack of 57
globally distributed benthic 180 records. Paleoceanography 20. URL:
http://doi.wiley.com/10.1029/2004PA001071, doi:10.1029/2004PA001071.
Longman, M.W., 1980. Carbonate Diagenetic Textures from Nearsurface Diagenetic En-
vironments. American Association of Petroleum Geologists Bulletin 64, 461-487.
Lowell, T.V., Heusser, C.J., Andersen, B.G., Moreno, P.I., Hauser, A., Heusser, L.E.,
Schliichter, C., Marchant, D.R., Denton, G.H., 1995. Interhemispheric Correlation of
Late Pleistocene Glacial Events. Science 269, 1541-1549.
Lowenstein, T.K., Hein, M.C., Bobst, A.L., Jordan, T.E., Ku, T.L., Luo, S., 2003. An
assessment of stratigraphic completeness in climate-sensitive closed-basin lake sediments:
Salar de Atacama, Chile. Journal of Sedimentary Petrology 73, 91-104.
Ludwig, K.R., 2003. Mathematical-Statistical Treatment of Data and Errors for
230Th/U Geochronology. Reviews in Mineralogy and Geochemistry 52, 631-656.
doi:10.2113/0520631.
Ludwig, K.R., 2012. User's Manual for Isoplot Version 3.75-4.15: A Geochronological
Toolkit for Microsoft Excel. Technical Report 5. Berkeley Geochronological Center Spe-
cial Publication.
Ludwig, K.R., Paces, J.B., 2002. Uranium-series dating of pedogenic silica and carbonate,
Crater Flat, Nevada. Geochimica et Cosmochimica Acta 66, 487-506. doi:10.1016/S0016-
7037(01)00786-4.
Luo, S., Ku, T.L., 1991. U-series isochron dating: A generalized method employing total-
sample dissolution. Geochimica et Cosmochimica Acta 55, 555-564. doi:10.1016/0016-
7037(91)90012-T.
Lyle, M., 1983. The brown-green color transition in marine sediments: A marker of the
Fe(III)-Fe(II) redox boundary. Limnol. Oceanogr. 28, 1026-1033.
Ma, Z., Wang, Z., Liu, J., Yuan, B., Xiao, J., Zhang, G., 2004. U-series chronology of
sediments associated with Lake Quaternary fluctuations, Balikun Lake, northwestern
China. Quaternary International 121,89--98.
Mackay, A.W., Bezrukova, E.V., Boyle, J.F., Holmes, J.A., Panizzo, V.N., Piotrowska,
N., Shchetnikov, A., Shilland, E.M., Tarasov, P., White, D., 2013. Multiproxy
evidence for abrupt climate change impacts on terrestrial and freshwater ecosystems
in the 0l'khon region of Lake Baikal, central Asia. Quaternary International
290-291, 46-56. URL: http://dx.doi.org/10.1016/j.quaint.2012.09.031
https: //linikinghub. elsevier. com/retrieve/pii/S1040618212032041,
doi:10.1016/j.quaint.2012.09.031.
197
Magaritz, M., Aravena, R., Pefia, H., Suzuki, 0., Grilli, A., 1989. Water chemistry and
isotope study of streams and springs in northern Chile. Journal of Hydrology 108, 323-
341. doi:10.1016/0022-1694(89)90292-8.
Magrin, G.O., Marengo, J.A., Boulanger, J.P., Buckeridge, M.S., Castellanos, E., Poveda,
G., Scarano, F.R., Vicufia, S., 2014. Central and South America, in: Barros, V.R., Field,
C.B., Dokken, D.J., Mastrandrea, M.D., Mach, K.J., Bilir, T.E., Chatterjee, M., Ebi,
K.L., Estrada, Y.O., Genova, R.C., Girma, B., Kissel, E.S., Levy, A.N., MacCracken,
S., Mastrandrea, P.R., White, L.L. (Eds.), Climate Change 2014: Impacts, Adaptation,
and Vulnerability. Part B: Regional Aspects: Contribution of Working Group II to the
Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge
University Press, Cambridge, United Kingdom. volume 2, pp. 1499-1566.
Marengo, J., 2004. Interdecadal and long term rainfall variability in the Amazon basin.
Theoretical and Applied Climatology 78, 79-96.
Marengo, J.A., Liebmann, B., Grimm, A.M., Misra, V., Silva Dias, P.L., Cavalcanti, I.F.A.,
Carvalho, L.M.V., Berbery, E.H., Ambrizzi, T., Vera, C.S., Saulo, A.C., Nogues-Paegle,
J., Zipser, E., Seth, A., Alves, L.M., 2012. Recent developments on the South American
monsoon system. International Journal of Climatology 32, 1-21. doi:10.1002/joc.2254.
McGee, D., Donohoe, A., Marshall, J., Ferreira, D., 2014. Changes in ITCZ location
and cross-equatorial heat transport at the Last Glacial Maximum, Heinrich Stadial
1, and the mid-Holocene. Earth and Planetary Science Letters 390, 69-79. URL:
http://dx. doi. org/10.1016/j. epsl.2013.12.043, doi:10.1016/j.epsl.2013.12.043.
McGee, D., Quade, J., Edwards, R.L., Broecker, W.S., Cheng, H., Reiners, P.W., Evenson,
N., 2012. Lacustrine cave carbonates: Novel archives of paleohydrologic change in the
Bonneville Basin (Utah, USA). Earth and Planetary Science Letters 351-352, 182-
194. URL: http://linkinghub.elsevier.com/retrieve/pii/S0012821X1200386X,
doi:10.1016/j.epsl.2012.07.019.
McIntyre, G., Brooks, C., Compston, W., Turek, A., 1966. The statistical assessment of
Rb-Sr isochrons. Journal of Geophysical Research 71, 5459-5468.
McLaren, S.J., Rowe, P.J., 1996. The reliability of uranium-series mollusc dates
from the western Mediterranean basin. Quaternary Science Reviews 15, 709-717.
doi:10.1016/0277-3791(96)00032-7.
Messerli, B., Grosjean, M., Vuille, M., 1997. Water Availability, Protected Areas, and
Natural Resources in the Andean Desert Altiplano. Mountain Research and Development
17, 229-238.
198
19 -- Ill - 1 1 1 W, -', 1T 'Ift. 111, I'll"-nm"R""Rm"ppmgmmwmm"mmmmm,,W'T''I 1 'j9' - Imm"M
Metcalfe, S., Say, A., Black, S., McCulloch, R., O'Hara, S., 2002. Wet conditions during
the last glaciation in the Chihuahuan Desert, Alta Babicora Basin, Mexico. Quaternary
Research 57, 91-101. doi:10.1006/qres.2001.2292.
Mix, A.C., Harris, S.E., Janecek, T.R., 1995. Estimating lithology from nonintrusive
reflectance spectra: Leg 138. Proceedings of the Ocean Drilling Program, Scientific
Results 138.
Morales, M.S., Carilla, J., Grau, H.R., Villalba, R., 2015. Multi-century
lake area changes in the Andean high-elevation ecosystems of the South-
ern Altiplano. Climate of the Past Discussions 11, 1821-1855. URL:
http: //www. clim-past-discuss.net/11/1821/2015/, doi:10.5194/cpd-11-1821-2015.
Mosblech, N.A.S., Bush, M.B., Gosling, W.D., Hodell, D., Thomas, L., van Calsteren,
P., Correa-Metrio, A., Valencia, B.G., Curtis, J., van Woesik, R., 2012. North Atlantic
forcing of Amazonian precipitation during the last ice age. Nature Geoscience 5, 817-820.
URL: http: //oro. open. ac.uk/34485/, doi:10.1038/ngeo588.
Miller, G., 1971. 'Gravitational' cement: An indicator for the vadose zone of the subaerial
diagenetic environment, in: Bricker, O.P. (Ed.), Carbonate Cements. Johns Hopkins
Press, Baltimore, MD, pp. 301-302.
Murphy, R.J., Lenhart, J.J., Honeyman, B.D., 1999. The sorption of thorium
(IV) and uranium (VI) to hematite in the presence of natural organic mat-
ter. Colloids and Surfaces A: Physicochemical and Engineering Aspects 157, 47-
62. URL: http://linkinghub.elsevier.com/retrieve/pii/S0927775799001156,
doi:10.1016/SO927-7757(99)00115-6.
Nagao, S., Nakashima, S., 1992. The factors controlling vertical color variations
of North Atlantic Madeira Abyssal Plain sediments. Marine Geology 109, 83-
94. URL: https://linkinghub.elsevier.com/retrieve/pii/0025322792902224,
doi:10.1016/0025-3227(92)90222-4.
Nester, P.L., Gay6, E., Latorre, C., Jordan, T.E., Blanco, N., 2007. Perennial stream dis-
charge in the hyperarid Atacama Desert of northern Chile during the latest Pleistocene.
Proceedings of the National Academy of Sciences of the United States of America 104,
19724-19729. doi:10.1073/pnas.0705373104.
Nishizawa, T., Tanaka, M., 1983. The Annual Change in the Tropospheric Circulation and
the Rainfall in South America. Archives for Meteorology, Geophysics, and Bioclimatology
B, 107-116.
Nogu6s-Paegle, J., Mechoso, C.R., Fu, R., Berbery, E.H., Chao,. W.C., Chen, T.C.,
Cook, K., Diaz, A.F., Enfield, D., Ferreira, R., Grimm, A.M., Kousky, V.E., Lieb-
mann, B., Marengo, J.A., Mo, K., Neelin, J.D., Paegle, J., Robertson, A.W.,
199
Seth, A., Vera, C.S., Zhou, J., 2002. Progress in Pan American CLIVAR Re-
search: Understanding the South American Monsoon. Meteorologica 27, 1-30.
doi:cptec.inpe.br/walmeida/2004/05.13.14.45.34-0.
Nowaczyk, N.R., Haltia, E.M., Ulbricht, D., Wennrich, V., Sauerbrey, M.A., Ros6n, P.,
Vogel, H., Francke, A., Meyer-Jacob, C., Andreev, A.A., Lozhkin, A.V., 2013. Chronol-
ogy of Lake El'gygytgyn sediments - A combined magnetostratigraphic, palaeoclimatic
and orbital tuning study based on multi-parameter analyses. Climate of the Past 9,
2413-2432. doi:10.5194/cp-9-2413-2013.
Oehlerich, M., Mayr, C., Griesshaber, E., Liicke, A., Oeckler, O.M., Qhlen-
dorf, C., Schmahl, W.W., Zolitschka, B., 2013. Ikaite precipitation in a la-
custrine environment - implications for palaeoclimatic studies using carbonates.
from Laguna Potrok Aike (Patagonia, Argentina). Quaternary Science Re-
views 71, 46-53. URL: http://dx.doi.org/10.1016/j.quascirev.2012.05.024,
doi:10.1016/j.quascirev.2012.05.024.
Oviatt, C.G., McCoy, W.D., Nash, W.P., 1994. Sequence stratigraphy
of lacustrine deposits: a Quaternary example from the Bonneville basin,
Utah. Geological Society of America Bulletin 106, 133-144. URL:
http://gsabulletin.gsapubs.org/content/106/1/133.short, doi:10.1130/0016-
7606(1994)106<0133.
Palacios-Fest, M.R., Alin, S.R., Cohen, A.S., Tanner, B., Heuser, H., 2005. Paleolimno-
logical investigations of anthropogenic environmental change in Lake Tanganyika: IV.
Lacustrine paleoecology. Journal of Paleolimnology 34, 51-71. doi:10.1007/s10933-005-
2397-1.
Peng, T.H., Goddard, J.G., Broecker, W.S., 1978. A Direct Comparison of 14C and 230Th
Ages at Searles Lake, California. Quaternary Research 9, 319-329.
Perez, L., Lorenschat, J., Bugja, R., Brenner, M., Scharf, B., Schwalb, A., 2010. Dis-
tribution, diversity and ecology of modern freshwater ostracodes (Crustacea), and hy-
drochemical characteristics of Lago Peten Itzi, Guatemala. Journal of Limnology 69,
146-159.
Perez-Sanz, A., Li, G., Gonzilez-Samperiz, P., Harrison, S.P., 2014. Evaluation of
modern and mid-Holocene seasonal precipitation of the Mediterranean and north-
ern Africa in the CMIP5 simulations. Climate of the Past 10, 551-5.68. URL:
http://www.'blim-past.net/10/551/2014/, doi:10.5194/cp-10-551-2014.
Peterson, L.C., Haug, G.H., 2006. Variability in the mean latitude of the Atlantic In-
tertropical Convergence Zone as recorded by riverine input of sediments to the Cariaco
200
Basin (Venezuela). Palaeogeography, Palaeoclimatology, Palaeoecology 234, 97-113.
doi:10.1016/j.palaeo.2005.10.021.
Peterson, L.C., Haug, G.H., Hughen, K.A., R6hl, U., 2000. Rapid changes in the hy-
drologic cycle of the tropical Atlantic during the last glacial. Science 290, 1947-1951.
doi:10.1126/science.290.5498.1947.
Petit, J., Jouzel, J., Raynaud, D., Barkov, N., Barnola, J.M., Basile, I., Bender, M.,
Chappellaz, J., Davis, M., Delaygue, G., Delmotte, M., Kotlyakov, V., Legrand, M.,
Lipenkov, V., Lorius, C., Pepin, L., Ritz, C., Saltzman, E., Stievenard, M., 1999. Climate
and atmospheric history of the past 420,000 years from the Vostok ice core, Antarctica.
Nature 399, 429-436. doi:10.1038/20859.
Phillips, F.M., Zreda, M.G., Ku, T.L., Luo, S., Huang, Q., Elmore, D., Kubik, P.W.,
Sharma, P., 1993. 230Th/234U and 36C1 dating of evaporite deposits from the western
Qaidam Basin, China: Implications for glacial-period dust export from Central Asia.
Geological Society of America Bulletin 105, 1606-1616.
Pigati, J.S., Springer, K.B., Honke, J.S., 2019. Desert wetlands
record hydrologic variability within the Younger Dryas chronozone,
Mojave Desert, USA. Quaternary Research 91, 51-62. URL:
https://www.cambridge.org/core/product/identifier/S0033589418000145/type/journal-article,
doi:10.1017/qua.2018.14.
Placzek, C., Patchett, P.J., Quade, J., Wagner, J.D.M., 2006a. Strategies for successful
U-Th dating of paleolake carbonates: An example from the Bolivian Altiplano. Geo-
chemistry, Geophysics, Geosystemns 7, Q05024. doi:10.1029/2005GC001157.
Placzek, C.,Quade, J., Betancourt, J.L., Jonathan, P., Rech, J.A., Latorre, C., Matmon,
A., Holmgren, C., English, N.B., Patchett, P.J., 2009. Climate in the Dry Central Andes
Over Geologic, Millenial, and Interannual Timescales. Annals of the Missouri Botanical
Garden 96, 386-397. doi:10.3417/2008019.
Placzek, C., Quade, J., Patchett, P.J., 2006b. Geochronology and stratigraphy of late
Pleistocene lake cycles on the southern Bolivian Altiplano: Implications for causes of
tropical climate change. Bulletin of the Geological Society of America 118, 515-532.
doi:10.1130/B25770.1.
Placzek, C.J., Quade, J., Patchett, P.J., 2013. A 130 ka reconstruction of rain-
fall on the Bolivian Altiplano. Earth and Planetary Science Letters 363, 97-108.
doi:10.1016/j.epsl.2012.12.017.
Plater, A.J., Ivanovich, M., Dugdale, R.E., 1992. Uranium series disequilibrium in river
sediments and waters: the significance of anomalous activity ratios. Applied Geochem-
istry 7, 101-110. doi:10.1016/0883-2927(92)90029-3.
201
Porcelli, D., Andersson, P., Wasserburg, G., Ingri, J., Baskaran, M., 1997. The im-
portance of colloids and mires for the transport of uranium isotopes through the
Kalix River watershed and Baltic Sea. Geochimica et Cosmochimica Acta 61, 4095-
4113. URL: http: //l1inkinghub. elsevier. com/retr ieve /pii/S0016703797002354,
doi:10.1016/S0016-7037(97)00235-4.
Powell, R., Hergt, J., Woodhead, J., 2002. Improving isochron calcula-
tions with robust statistics and the bootstrap. Chemical Geology 185, -191-
204. URL: http: //linkinghub. elsevier. com/retr ieve /pii/S000925410100403X,
doi:10.1016/S0009-2541(01)00403-X.
Prokopenko, A.A., Hinnov, L.A., Williams, D.F., Kuzmin, M.I., 2006. Orbital forcing
of continental climate during the Pleistocene: a complete astronomically tuned cli-
inatic record from Lake Baikal, SE Siberia. Quaternary Science Reviews 25, 3431-3457.
doi:10.1016/j.quascirev.2006.10.002.
Przybylowicz, W., Schwarcz, H.P., Latham, A.G., 1991. Dirty calcites 2. Uranium-series
dating of artificial calcite-detritus mixtures. Chemical Geology: Isotope Geoscience
Section 86, 161-178. doi:10.1016/0168-9622(91)90060-A.
Qin, B., Harrison, S.P., Kutzbach, J.E., 1998. Evaluation of modelled regional water
balance using lake status data: A comparison of 6 ka simulations with the NCAR CCM.
Quaternary Science Reviews 17, 535-548. doi:10.1016/S277-3791(98)00011-0.
Quade, J., Dettinger, M.P., Carrapa, B., DeCelles,- P., Murray, K.E., Huntington, K.A.,
Cartwright, A., Canavan, R.R., Gehrels, G., Clementz, M., 2015. The growth of the
central Andes, 22 S-26 S. Geodynamics of a Cordilleran Orogenic System: The Central
Andes of Argentina and Northern Chile: Geological Society of America Memoir 212.
Quade, J., Rech, J.A., Betancourt, J.L., Latorre, C., Quade, B., Rylander, K.A., Fisher,
T., 2008. Paleowetlands and regional climate change in the central Atacama Desert,
northern Chile. Quaternary Research 69, 343-360. doi:10.1016/j.yqres.2008.01.003.
Reeder, R.J., Nugent, M., Lamble, G.M., Tait, C.D., Morris, D.E., 2000.
Uranyl Incorporation into Calcite and Aragonite: XAFS and Lumines-
cence Studies. Environmental Science & Technology 34, 638-644. URL:
https: //pubs. acs. org/doi/10. 1021/es990981j, doi:10.1021/es990981j.
Reeder, R.J., Nugent, M., Tait, C.D., Morris, D.E., Heald, S.M., Beck, K.M., Hess,
W.P., Lanzirotti, A., 2001. Coprecipitation of Uranium(VI) with Calcite: XAFS, micro-
XAS, and luminescence characterization. Geochimica et Cosmochimica Acta'65 3491-
3503. URL: http: //linkinghub. e lsevier. c om/retr ieve /pii/S0016703701006470,
doi:10.1016/S0016-7037(01)00647-0.
202
Riding, R., 1979. Origin and diagenesis of lacustrine algal bioherms at the
margin of the Ries crater, Upper Miocene, southern Germany. Sedimentology
26, 645--680. URL: http://dx.doi.org/10.1111/j.1365-3091.1979.tb00936.x,
doi:10.1111/j.1365-3091.1979.tb00936.x.
Rissmann, C., Leybourne, M., Benn, C., Christenson, B., 2015. The origin of solutes
within the groundwaters of a high Andean aquifer. Chemical Geology 396, 164-
181. URL: http://linkinghub.elsevier.com/retrieve/pii/S009254114005877,
doi:10.1016/j.chemgeo.2014.11.029.
Roberts, H.M., Bryant, C.L., Huws, D.G., Lamb, H.F., 2018. Generat-
ing long chronologies for lacustrine sediments using luminescence dating: a
250,000 year record from Lake Tana, Ethiopia. Quaternary Science Re-
views 202, 66-77. URL: https://doi.org/10.1016/j.quascirev.2018.10.037,
doi:10.1016/j.quascirev.2018.10.037.
Robinson, L.F., Adkins, J.F., Fernandez, D.P., Burnett, D.S., Wang, S.L., Gagnon, A.C.,
Krakauer, N., 2006. Primary U distribution in scleractinian corals and its implications for
U series dating. Geochemistry, Geophysics, Geosystems 7. doi:10.1029/2005GC01138.
Robinson, L.F.. Henderson, G.M., Hall, L., Matthews, I., 2004. Climatic
Control of Riverine and Seawater Uranium-Isotope Ratios. Science 305,
851--854. URL: http://www.sciencemag.org/cgi/doi/10.1126/science.1099673,
doi:10.1126/science.1099673.
Robinson, L.F., Henderson, G.M., Slowey, N.C., 2002. U-Th dating of marine isotope
stage 7 in Bahamas slope sediments. Earth and Planetary Science Letters 196, 175-187.
doi:10.1016/S0012-821X(01)00610-0.
Rodbell, D.T., Abbott, M.B., the 2011 ICDP Lake Junin Working Group, 2012. Workshop
on drilling of Lake Junin, Peru: Potential for development of a continuous tropical
climate record. Scientific Drilling 3, 58-60. doi:10.2204/iodp.sd.13.10.2011.
Roehrig, R., Bouniol, D., Guichard, F., Hourdin, F., Redelsperger, J.L., 2013. The
Present and Future of the West African Monsoon: A Process-Oriented Assessment
of CMIP5 Simulations along the AMMA Transect. Journal of Climate 26, 6471-
6505. URL: http://journals.ametsoc.org/doi/abs/10.1175/JCLI-D-12-00505.1,
doi:10.1175/.JCLI-D-12-00505.1.
Rouchy, J.M., Servant, M., Fournier, M., Causse, C., 1996. Extensive carbonate algal bio-
herms in upper Pleistocene saline lakes of the central Altiplano of Bolivia. Sedimentology
43, 973-993. doi:10,1111/j.1365-3091.1996.tb01514.x.
203
Rowe, P.J., Richards, D.A., Atkinson, T.C., Bottrell, S.H., Cliff, R.A., 1997. Geochemistry
and radiometric dating of a Middle Pleistocene peat. Geochinica et Cosmochimica Acta
61, 4201-4211. doi:10.1016/S0016-7037(97)00213-5.
Rowell, D.P., 2012. Sources of uncertainty in future changes in
local precipitation. Climate Dynamics 39, 1929-1950. URL:
http://link.springer.com/10.1007/s00382-011-1210-2, doi:10.1007/s00382-
011-1210-2.
Rudnick, R., Gao, S., 2003. Composition of the Continental Crust, in: Treatise on Geo-
chemistry. Elsevier. volume 3, pp. 1-64. doi:10.1016/B-08-043751-6/03016-4.
Rutllant, J., Ulriksen, P., 1979. Boundary-layer dynamics of the extremely arid northern
part of Chile: The Antofagasta Field Experiment. Boundary Layer Meteorology 17,
41-55.
Sarin, M., Krishnaswami, S., Moore, W., et al., 1990. Chemistry of uranium, thorium,
and radium isotopes in the ganga-brahmaputra river system: Weathering processes and
fluxes to the bay of bengal. Geochimica et Cosmochimica Acta 54, 1387-1396.
Saulo, A.C., Nicolini, M., Chou, S.C., 2000. Model characterization of the South American
low-level flow during the 1997-1998 spring-summer season. Climate Dynamics 16, 867-
881.
Schmeide, K., Pompe, S., Bubner, M., Heise, K.H., Bernhard, G., Nitsche,
H., 2000. Uranium(VI) sorption onto phyllite and selected minerals in
the presence of humic acid. Radiochimica Acta 88, 723-728. URL:
http://www.degruyter.com/view/j/ract.2000.88.issue-9-11/ract.2000.88.9-11.723/ract.20C
doi:10.1524/ract.2000.88.9-11.723.
Schnurrenberger, D., Russell, J., Kelts, K., 2003. Classification of lacustrine sediments
based on sedimentary components. Journal of Paleolimnology 29, 141-154.
Schoelle, P.A., Ulmer-Schoelle, D.S., 2003. Syngenetic/Eogenetic Marine Diagenesis, in:
Lorenz, J.C. (Ed.), AAPG Memoir 77: A Color Guide to the Petrography of Carbonate
Rocks: Grains, textures, porosity, diagenesis. The American Association of Petroleum
Geologists, Tulsa, Oklahoma. chapter'Chapter 23, p. 328.
Schoene, B., Condon, D.J., Morgan, L., McLean, N., 2013. Precision and Accuracy in
Geochronology. Elements 9, 19-24. doi:10.2113/gselements.9.1.19.
Schwalb, A., Locke, S.M., Dean, W.E., 1995. Ostracode 6180 and 613C evidence of Holocene
environmental changes in the sediments of two Minnesota lakes., Journal of Paleolimnol-
ogy 14, 281-296. doi:10.1007/BF00682429.
204
Schwarcz, H.P., Latham, A.G., 1989. Dirty calcites 1. Uranium-series dating of contami-
nated calcite using leachates alone. Chemical Geology: Isotope Geoscience Section 80,
35-43. doi:10.1016/0168-9622(89)90046-8.
Selleck, B.W., Carr, P.F., Jones, B.G., 2007. A Review and Synthesis of Glendonites
(Pseudomorphs after Ikaite) with New Data: Assessing Applicability as Recorders
of Ancient Coldwater Conditions. Journal of Sedimentary Research 77, 980-991.
doi:10.2110/jsr.2007.087.
Seltzer, G., Rodbell, D., Burns, S., 2000. Isotopic evidence for late Quater-
nary climatic change in tropical South America. Geology 28, 35. URL:
https://pubs.geoscienceworld.org/geology/article/28/1/35-38/187921,
doi:10.1130/0091-7613(2000)28<35:IEFLQC>2.0.CO;2.
Seluchi, M.E., Saulo, A.C., Nicolini, M., Satyamurty, P., 2003. The Northwestern Argen-
tinean Low: A Study of Two Typical Events. Monthly Weather Review 131, 2361-2378.
doi:10.1175/1520-0493(2003)131<2361:TNALAS>2.0.CO;2.
Shakun, J.D., Clark, P.U., He, F., Marcott, S.A., Mix, A.C., Liu, Z., Otto-Bliesner, B.,
Schmnittner, A., Bard, E., 2012. Global warming preceded by increasing carbon dioxide
concentrations during the last deglaciation. Nature 484, 49-54. doi:10.1038/nature10915.
Shanahan, T.M., Peck, J.A., McKay, N., Heil, C.W., King, J., Forman,
S.L., Hoffmann, D.L., Richards, D.A., Overpeck, J.T., Scholz, C., 2013.
Age models for long lacustrine sediment records using multiple dating ap-
proaches - An example from Lake Bosumtwi, Ghana. Quaternary Geochronol-
ogy 15, 47-60. URL: http://dx.doi.org/10.1016/j.quageo.2012.12.001
https://linkinghub.elsevier.com/retrieve/pii/S1871101412001872,
doi:10.1016/j.quageo.2012.12.001.
Shearman, D.J., McGugan, A., Stein, C., Smith, A.J., 1989. Ikaite, CaCO3 6H20, pre-
cursor of the thinolites in the Quaternary tufas and tufa mounds of the Lahontan and
Mono Lake Basins, western United States. Geological Society of America Bulletin 101,
913-917. doi10.1130/0016-7606(1989)101<0913.
Shearman, D.J., Smith, A.J., 1985. Ikaite, the parent mineral of jarrowite-type
pseudomorphs. Proceedings of the Geologists' Association 96, 305-314. URL:
http://www.scopus. com/inward/record.url?eid=2-s2.0-0022214193&partnerID=tZOtx3yl,
doi:10.1016/S0016-7878(85)80019-5.
Shen, C.C., Lawrence Edwards, R., Cheng, H., Dorale, J.A., Thomas, R.B.,,Bradley Moran,
S., Weinstein, S.E., Edmonds, H.N., 2002. Uranium and thorium isotopic and concentra-
tion measurements by magnetic sector inductively coupled plasma mass spectrometry.
Chemical Geology 185, 165-178. doi:10.1016/S0009-2541(01)00404-1.
205
Shen, C.C., Wu, C.C., Cheng, H., Edwards, R.L., Hsieh, Y.T., Gallet,
S., Chang, C.C., Li, T.Y., Lam, D.D., Kano, A., Hori, M., Sp6tl, C.,
2012. High-precision and high-resolution carbonate 230Th dating by MC-ICP-
MS with SEM protocols. G.eochimica et Cosmochimica Acta 99, 71 86.
67 3 20 0524 8
URL: http://www.sciencedirect.com/science/article/pii/S01 0 71 ,
doi:https://doi.org/-10.1016/j.gca.2012.09.018.
Shepherd, T.G., 2014. Atmospheric circulation as a source of uncertainty
in climate change projections. Nature Geoscience 7, 703-708. URL:
http://www.nature. com/doif inder/10.1038/ngeo2253, doi:10.1038/ngeo2253.
Silva, V.B.S., Kousky, V.E., 2012. The South American Mon-
soon System: Climatology and Variability, in: Wang, S.Y., - Gillies,
R.R. (Eds.), Modern Climatology. InTech, pp. 123-152. URL:
http://www.intechopen.com/books/modern-climatology/the-south-american-monsoon-system
Skrivanek, A., Li, J., Dutton, A., 2018. Relative sea-level change during the Last In-
terglacial as recorded in Bahamian fossil reefs. Quaternary Science Reviews 200, 160-
4
177. URL: https://linkinghub.elsevier.com/retrieve/pii/S27737911830438 ,
doi:10.1016/j.quascirev.2018.09.033.
Smith, A.J., Delorme, L.D., 2010. Ostracoda. Ecology and Classification of North American
Freshwater Invertebrates , 725-771doi:10.1016/B978-0-12-374855-3.00019-4.
Smith, J.A., Finkel, R.C., Farber, D.L., Rodbell, D.T., Seltzer, G.O., 2005a. Moraine
preservation and boulder erosion in the tropical Andes: Interpreting old surface exposure
ages in glaciated valleys. Journal of Quaternary Science 20, 735-758. doi:10.1002/jqs.981.
Smith, J.A., Rodbell, D.T., 2010. Cross-cutting moraines reveal evidence for north atlantic
influence on glaciers in the tropical Andes. Journal of Quaternary Science 25, 243-248.
doi:10.1002/jqs.1393.
Smith, J.A., Seltzer, G.O., Farber, D.L., Rodbell, D.T., Finkel, R.C., 2005b.
Early local last glacial maximum in the tropical Andes. Science 308, 678-681.
doi:10.1126/science.1107075.
Smith, W., 1816. Strata identified by organized fossils: containing prints on colored paper
of the most characteristic specimens in each stratum.
Stein, M., Goldstein, S.L., 2006. U-Th and radiocarbon chronologies of late
Quaternary lacustrine records of the Dead Sea basin: Methods and applica-
tions, in: Enzel, Y., Agnon, A., Stein, M. (Eds.), New frontiers in Dead
Sea paleoenvironmental research: Geological Society of America -Special Pa-
per 401. Geological Society of America. volume 401, pp. 141-154. URL:
206
https://pubs.geoscienceworld.org/books/book/558/chapter/3802608/,
doi:10.1130/2006.2401(09).
Steno, N., 1669. De Solido Intra Solidum Naturaliter Contento Dissertationis Prodromus.
Ph.D. thesis.
Steponaitis, E., Andrews, A., McGee, D., Quade, J., Hsieh, Y.T., Broecker, W.S.,
Shuman, B.N., Burns, S.J., Cheng, H., 2015. Mid-Holocene drying of the
U.S. Great Basin recorded in Nevada speleothems. Quaternary Science Reviews
, 1-12URL: http://linkinghub.elsevier.com/retrieve/pii/S0277379115001535,
doi:10.1016/j.quascirev.2015.04.011.
Stirling, C., Lee, D.C., Christensen, J., Halliday, A., 2000. High-precision
in situ 238U-234U-230Th isotopic analysis using laser ablation multiple-
collector ICPMS. Geochimica et Cosmochimica Acta 64, 3737-3750. URL:
https://linkinghub.elsevier.com/retrieve/pii/S001670370000 4 579,
doi:10.1016/S0016-7037(00)00457-9.
Stockhecke, M., Kwiecien, 0., Vigliotti, L., Anselmetti, F.S., Beer, J., Qagatay, M.N.,
Channell, J.E., Kipfer, R., Lachner, J., Litt, T., Pickarski, N., Sturm, M., 2014. Chronos-
tratigraphy of the 600,000 year old continental record of Lake Van (Turkey). Quaternary
Science Reviews 104, 8-17. doi:10.1016/j.quascirev.2014.04.008.
Strikis, N.M., Chiessi, C.M., Cruz, F.W., Vuille, M., Cheng, H., Barreto, E.A.S., Mollen-
hauer, G., Kasten, S., Karmann, I., Edwards, R.L., Bernal, J.P., Sales, R., 2015. Timing
and structure of Mega-SACZ events during Heinrich Stadial 1. Geophysical Research
Letters 42, 1-8. doi:10.1002/2015GL64048.Received.
Suess, E., Balzer, W., Hesse, K.F., Muller, P.J., Ungerer, C.A., Wefer, G.,
1982. Calcium Carbonate Hexahydrate from Organic-Rich Sediments of
the Antarctic Shelf: Precursors of Glendonites. Science 216, 1128-1131.
URL: http://www.sciencemag.org/cgi/doi/10.1126/science.216.4550.1128,
doi:10.1126/science.216.4550.1128.
Swainson, I.P., Hammond, R.P., 2001. Ikaite, CaCO3 - 6H20: Cold comfort for glendonites
as paleothermometers. American Mineralogist 86, 1530-1533.
Sylvestre, F., Servant, M., Servant-Vildary, S., Causse, C., Fournier,
M., Ybert, J.P., 1999. Lake-Level Chronology on the Southern Bo-
livian Altiplano (18°-23°S) During Late-Glacial Time and the Early
Holocene. Quaternary Research 51, 54-66. URL: isi:000078140200005
https://www.cambridge.org/core/product/identifier/S033589400033937/type/journal_-article,
doi:10.1006/qres.1998,2017.
207
Thornalley, D.J.R., Bauch, H.A., Gebbie, G., Guo, W., Ziegler, M., Bernasconi, S.M.,
Barker, S., Skinner, L.C., Yu, J., 2015. A warm and poorly ventilated deep Arctic
Mediterranean during the last glacial period. Science 349, 706-710.
Thurber, D.L., 1962. Anomalous U 234 /U 238 in nature. Journal of Geophysical Research
67, 4518-4520.
Tomlinson, C.W., 1916. The origin of red beds: a study of the conditions of origin of the
Permo-Carboniferous and Triassic red beds of the western United States. The Journal
of Geology.24, 1537179.
Torfstein, A., Goldstein, S.L., Kushnir, Y., Enzel, Y., Haug, G:, Stein, M.,
2015. Dead Sea drawdown and monsoonal impacts in the Levant during the
last interglacial. Earth and Planetary Science Letters 412, 235-244. URL:
http://dx.doi.org/10.1016/j.epsl.2014.12.013, doi:10.1016/j.epsl.2014.12.013.
Trachsel, M., Telford, R.J., 2017. All age-depth models are
wrong, but are getting better. The Holocene 27, 860-869.
URL: http://journals.sagepub.com/doi/10.1177/0959683616675939,
doi:10.1177/0959683616675939.
Valero-Garces, B., Grosjean, M., Schwalb, A., Geyh, M., Messerli, B., Kelts, K., 1996.
Limnogeology of Laguna Miscanti: evidence for mid to late Holocene moisture changes
in the Atacama Altiplano (Northern Chile). Journal of Paleolimnology 16, 1-21.
doi:10.1007/BFOO173268.
Valero-Garc's, B.L., Grosjean, M., Kelts, K., Schreier, H., Messerli, B., 1999. Holocene
lacustrine deposition in the Atacama Altiplano: Facies models, climate and tec-
tonic forcing. Palaeogeography, Palaeoclimatology, Palaeoecology 151, 101-125.
doi:10.1016/S0031-0182(99)00018-8.
Van Der Wijk, A., El-Daoushy, F., Arends, A.R., Mook, W.G., 1986. Dating peat with
U/Th disequilibrium: Some geochemical considerations. Chemical Geology: Isotope
Geoscience Section 59, 283-292. doi:10.1016/0168-9622(86)90077-1.
Vera, C., Higgins, W., Amador, J., Ambrizzi, T., Garreaud, R., Gochis, D., Gutzler, D.,
Lettenmaier, D., Marengo, J., Mechoso, C.R., Nogues-Paegle, J., Silva Dias, P.L.,Zhang,
C., 2006. Toward a unified view of the American monsoon systems. Journal of Climate
19, 4977-5000. doi:10.1175/JCLI3896.1.
Vermeesch, P., 2018. IsoplotR: A free and open toolbox for geochronology. Geoscience
Frontiers 9, 1479-1493. doi:10.1016/j.gsf.2018.04.001.
208
Vicufia, S., McPhee, J., Garreaud, R.D., 2012. Agriculture vulnerability to climate change
in a snowmelt driven basin in semiarid Chile. J. Water Resour. Plann. Manage. 138,
431-441. doi:10.1061/(ASCE)WR.1943-5452.0000202.
Vuille, M., 1999. Atmospheric circulation over the Bolivian Altiplano during dry and
wet periods and extreme phases of the Southern Oscillation. International Journal of
Climatology 19, 1579-1600. doi:10.1002/(SICI)1097-0088(19991130)19:14<1579::AID-
JOC441>3.0.CO;2-N.
Vuille, M., Baumgartner, M.F., 1993. Hydrologic Investigations in the North Chilean
Altiplano Using Landsat - MSS and- TM data. Geocarto International 3, 35-45.
doi:10.1080/10106049309354418.
Vuille, M., Bradley, R.S.. Keimig, F., 2000. Interannual climate variability in the Central
Andes and its relation to tropical Pacific and Atlantic forcing. Journal of Geophysical
Research 105, 12447. doi:10.1029/2000JD900134.
Vuille, M., Keimig, F., 2004. Interannual variability of summertime convective cloudiness
and precipitation in the central Andes derived from ISCCP-B3 data. Journal of Climate
17, 3334-3348. doi:10.1175/1520-0442(2004)017<3334:IVOSCC>2.0.CO;2.
Wade, B.L., 2014. Chasing South America's monsoon. Science 346, 1042-1043.
Wang, J., Fang, X., Appel, E., Zhang, W., 2013. Magnetostratigraphic and radiometric con-
straints on salt formation in the Qaidam Basin, NE Tibetan Plateau. Quaternary Science
Reviews 78, 53-64. URL: http://dx.doi.org/10.1016/j.quascirev.2013.07.017,
doi:10.1016/j.quascirev.2013.07.017.
Wang, X., -Auler, A.S., Edwards, R.L., Cheng, H., Cristalli, P.S., Smart, P.L., Richards,
D.a., Shen, C.C., 2004. Wet periods in northeastern Brazil over the past 210 kyr linked
to distant climate anomalies. Nature 432, 740-743. doi:10.1038/nature03067.
Wang, X., Auler, A.S.,-Edwards, R.L., Cheng, H., Ito, E., Solheid, M., 2006.. Interhemi-
spheric anti-phasing of rainfall during the last glacial period. Quaternary Science Reviews
25, 3391-3403. doi:10.1016/j.quascirev.2006.02.009.
Wang, X., Auler, A.S., Edwards, R.L., Cheng, H., Ito, E., Wang, Y., Kong, X., Solheid,
M., 2007. Millennial-scale precipitation changes in southern Brazil over the past 90,000
years. Geophysical Research Letters 34, 1-5. doi:10.1029/2007GL31149.
Wang, Y.J., Cheng, H., Edwards, R.L., An, Z.S., Wu, J.Y., Shen, C.C., Dorale, J.A.,
2001. A high-resolution absolute-dated late Pleistocene Monsoon record from Hulu Cave,
China. Science 294, 2345-2348. doi:10.1126/science.1064618.
209
Wendt, I., Carl, C., 1991. The statistical distribution of the mean squared
weighted deviation. Chemical Geology: Isotope Geoscience section. 86, 275-
285. URL: https://linkinghub.elsevier.com/retrieve/pii/016896229190010T,
doi:10.1016/0168-9622(91)90010-T.
Winograd, I.J., Coplen, T.B., Landwehr, J.M., Riggs, A.C., Ludwig, K.R., Szabo, B.J.,
Kolesar, P.T., Revesz, K.M., 1992. Continuous 500,000-Year Climate Record From Vein
Calcite in Devils-Hole, Nevada. Science 258, 255-260.
Wood, R.D., Imahori, W., 1959. Geographical Distribution of Characeae. Bulletin of the
Torrey Botanical Club 86, 172-183.
Woods, A., Rodbell, D., Abbott, M., Hatfield, R.G., Chen, C.Y., Lehmann, S., McGee, D.,
Weidhaas, N., Tapia, P.M., Valero-Garc6s, B.L., Bush, M., Stoner, J.S., 2019. Abrupt
arctic warming repeatedly led to prolonged drought and glacial retreat in the tropical
andes during the last glacial cycle. EarthArXiv doi:10.31223/osf.io/7hd3a.
Wright, H.E., 1983., Late-Pleistocene Glaciation and Climate around the Junin Plain,
Central Peruvian Highlands. Geografiska Annaler. Series A, Physical Geography 65,
35-43. URL: http://www.jstor.org/stable/520719.
Wrozyna, C., Frenzel, P., Daut, G., Miusbacher, R., Zhu, L., Schwalb, A., 2012. Holocene
Lake-Level Changes of Lake Nam Co, Tibetan Plateau, Deduced from Ostracod As-
semblages and j180 and 6 13 C Signatures of Their Valves. Developments in Quaternary
Science 17, 281-295. doi:10.1016/B978-0-444-53636-5.00016-0.
Xuan, C., Channell, J.E., Hodell, D.A., 2016. Quaternary magnetic and oxygen isotope
stratigraphy in diatom-rich sediments of the southern Gardar Drift (IODP Site U1304,
North Atlantic). Quaternary Science Reviews 142, 74-89.
Yliruokanen, I., 1980. The occurence of uranium in some finnish peat bogs. Kemia-Kemi
7, 213-217.
York, D., 1968. Least squares fitting of a straight line with corre-
lated errors. Earth and Planetary Science Letters 5, 320-324. URL:
http: //www. sciencedirect. c om/s c ience/art icle/pii/S12821X688005971
doi:https://doi.org/10.1016/S0012-821X(68)80059-7.
Yu, G., Harrison, S.P.,_ Xue, B., 2001.Lake status records from China: Data Base Docu-
mentation. MPI-BGC Technical Report 4, 240.
Zhou, J., Lau, K.M., 1998. Does a monsoon climate exist over South America? Journal of
Climate.11, 1020-1040. doi:10.1175/1520-0442(1998)011<1020:DAMCEO>20.CO;2.
210
Ziegler, L., Constable, C., Johnson, C., Tauxe, L., 2011. PADM2M: a penalized maximum
likelihood model of the 0-2 Ma palaeomagnetic axial dipole moment. Geophysical Journal
International 184, 1069-1089.
211