SEDIMENT CORE IMAGES AS CLIMATE AND …
105
1 SEDIMENT CORE IMAGES AS CLIMATE AND PALEOENVIRONMENTAL INDICATORS: EVIDENCE FROM MODERN AND HISTORICAL SEDIMENTS OF LAKE PETÉN ITZÁ, GUATEMALA By DUSTIN AARON GRZESIK A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2010
Transcript of SEDIMENT CORE IMAGES AS CLIMATE AND …
1
SEDIMENT CORE IMAGES AS CLIMATE AND PALEOENVIRONMENTAL INDICATORS: EVIDENCE FROM MODERN AND HISTORICAL SEDIMENTS OF
LAKE PETÉN ITZÁ, GUATEMALA
By
DUSTIN AARON GRZESIK
A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE
UNIVERSITY OF FLORIDA
2010
2
© 2010 Dustin Aaron Grzesik
3
In loving memory of my grandfather Richard Paul Grzesik Sr.
4
ACKNOWLEDGMENTS
I am honored to thank my co-chairs Dr. Mark Brenner, and Dr. David Hodell for
welcoming me as a student, for their academic support, and for the opportunity to
participate in the Petén Itzá Scientific Drilling Project. I would also like to thank Dr. Jon
Martin, for his critical evaluation of this thesis.
I would like to recognize Dr. Jason Curtis for his help searching for gastropods,
time spent on the mass spectrometer, time spent discussing mechanical and electronic
systems, and for being such a great roommate at La Casa de Don David during the field
campaign. I owe thanks to Andreas Müller for doing a great job as a co-database
curator during the field campaign and for being an excellent friend. Ray G. Thomas was
critical in executing software deployment prior to the field campaign, and has inspired
me to learn more about hardware and software design and modification. William
Kenney provided assistance with measurements of surficial sediment samples. I would
also like to acknowledge Dr. Adrian Gilli, Jennifer Mays, and Jaime Escobar for their
support and collaboration on this project.
I express my sincere thanks and appreciation to the Drilling, Observation and
Sampling of the Earth’s Continental Crust (DOSECC) drilling team (Ed Brown, James
Cramer, John Joice, Beau Marshall, Dennis Nielson, Vance Hiatt, Doug
Schnurrenberger, Kent Thomas, and Ryan Wilson), the International Continental
Scientific Drilling Program Operational Support Group (ICDP-OSG) (Christian Carnein,
Ronald Conzé, Uli Harms, Jochem Kueck, Martin Toepfer), and Lacustrine Core
Repository (LRC/LacCore) (Kristina Brady, Amy Myrbo and Anders Noren) whose hard
work and good humor made the project a success. Tom Guilderson provided assistance
with radiocarbon dating. I am also grateful to the numerous agencies and individuals in
5
Guatemala who provided assistance to this project. They include: Universidad del Valle,
Universidad San Carlos, Ministerio de Ambiente y Recursos Naturales, Consejo
Nacional de Areas Protegidas, Instituto de Antropología e Historia, Autoridad Para el
Manejo y Desarrollo Sostenible de la Cuenca del Lago Petén-Itzá, Wildlife Conservation
Society, Alex Arrivillaga, Cathy Lopez, Margaret Dix, Michael Dix, Margarita Palmieri,
David, Rosita, & Kelsey Kuhn, and the staff at La Casa de Don David, Lico Godoy, Tony
Ortiz, Franz Sperisen, Luis Toruño, Julian Tesucún, Liseth Perez, Melissa Orozco, Silja
Ramirez, Gabriela Alfaro, and Jacobo Blijdenstein. This project was funded by grants
from the U.S. National Science Foundation, the International Continental Scientific
Drilling Program, the Swiss Federal Institute of Technology, and the Swiss National
Science Foundation. Chuang Xuan provided critical evaluation and support for spectral
analysis and signal processing. Michael Cammer, former director of the Analytical
Imaging Facility of the Albert Einstein College of Medicine was critical in improving my
understanding of image analysis and provided support during the early phases of
development of routines for normalizing the image dataset. I thank Dr. John Sum-Ping,
Jaime Escobar, and Andreas Müller for reviewing this thesis and for their critical
evaluation.
I would also like to thank the faculty and graduate students of the Department of
Geological Sciences for their open doors, endless patience, and enthusiasm. Jie Wang
and Ryan Francis were excellent colleagues as TAs. My officemates Mike Ritorto and
Joshua Richards made work fun and I’m thankful for the help they provided whenever I
needed it. I enjoyed numerous lunches in the Plaza of the Americas with Jaime
Escobar and Natalia Hoyos.
6
Finally, I thank my parents Richard and Janice Grzesik and parents in-law John
and Lynal Sum-Ping for their support and encouragement. And most importantly I
would like to thank Joanne Sum-Ping for her endless patience, support, and motivation,
which allowed me to complete this work and remain not only sane, but supremely
happy.
7
TABLE OF CONTENTS
ACKNOWLEDGMENTS ..................................................................................................
page
4
LIST OF TABLES .......................................................................................................... 10
LIST OF FIGURES ........................................................................................................ 11
ABSTRACT ................................................................................................................... 16
CHAPTER
1 INTRODUCTION .................................................................................................... 18
Analytical Challenges ....................................................................................... 19 Imaging............................................................................................................. 20
2 BACKGROUND ...................................................................................................... 21
Petén Paleolimnology ....................................................................................... 21 Petén Itzá Limnology and Setting ..................................................................... 21 Cultural Setting ................................................................................................. 22 Petén Itzá Scientific Drilling Project (PISDP) .................................................... 23
3 METHODS .............................................................................................................. 26
Field Methods ......................................................................................................... 26 Sediment Drilling .............................................................................................. 26 Core Compositing with SPLICER ..................................................................... 27 Modern Surface Sediment Transect ................................................................. 28
Analytical Methods .................................................................................................. 28 Subsampling ..................................................................................................... 29 Image Acquisition ............................................................................................. 29
Red, green, blue color-digital images ......................................................... 30 L*a*b* colorspace ...................................................................................... 30
Image Processing and Analysis ....................................................................... 31 Geochemical Analyses ..................................................................................... 32 Core Chronology .............................................................................................. 32
4 RESULTS AND DISCUSSION ............................................................................... 39
Site PI-6 Chronology ............................................................................................... 39 Surface Sediments from the Water-Depth Transect ............................................... 39
Transect Image Data ........................................................................................ 39 Transect Geochemistry .................................................................................... 40 Transect Lithology ............................................................................................ 41
8
Site PI-6 Sediment Cores ....................................................................................... 41 Site PI-6 Lithology ............................................................................................ 41 Core Logging Data ........................................................................................... 41
Density ....................................................................................................... 41 Magnetic susceptibility ............................................................................... 42 XRF data .................................................................................................... 43 Composite image RGB data ...................................................................... 43 L*a*b* colorspace ...................................................................................... 44 PI-6 granularity ........................................................................................... 44
5 INTERPRETATION OF SEDIMENTS FROM SITE PI-6 ......................................... 64
Modern Sediment Properties .................................................................................. 64 Modern Limnological Setting ............................................................................ 64 Lightness and E/P ............................................................................................ 64 Confounding Factors ........................................................................................ 65 Surface Sediment Transect Samples ............................................................... 65
Sediment Cores ...................................................................................................... 65 Core Images ..................................................................................................... 65 Variations in Sediment Lithology ...................................................................... 66 X-Ray Fluorescence (XRF) Data ...................................................................... 67 L*a*b* Data Evaluation and Interpretation ........................................................ 67 Magnetic Susceptibility ..................................................................................... 68 Granularity ........................................................................................................ 68
6 CORRELATION TO OTHER CLIMATE RECORDS ............................................... 79
Regional Paleoclimate ............................................................................................ 79 Global Paleoclimate and Timeseries Analysis ........................................................ 80
7 SUMMARY AND CONCLUSIONS .......................................................................... 89
Limitations of Images as Paleoclimatic Proxies ...................................................... 90 Future Work ............................................................................................................ 91
APPENDIX
A SOFTWARE USED................................................................................................. 92
Field Software ......................................................................................................... 92 Drilling Information System ............................................................................... 92 SPLICER .......................................................................................................... 92
Image Processing and Analysis .............................................................................. 93 ImageJ.............................................................................................................. 94 Matlab............................................................................................................... 95 Corelyzer and Corewall .................................................................................... 95 Analyseries ....................................................................................................... 96
9
B PALEOLAKE HIGH-STANDS ................................................................................. 97
C SITE PI-3 ................................................................................................................ 99
LIST OF REFERENCES ............................................................................................. 101
BIOGRAPHICAL SKETCH .......................................................................................... 105
10
LIST OF TABLES
Table
page
3-1 Drilling site summary. ......................................................................................... 33
3-2 Locations of surface sediment samples collected along a water-depth transect. .............................................................................................................. 33
4-1 Dates and tie points used to develop the chronology for site PI-6 including sample, depth, 14C dates, positions. ................................................................... 46
4-2 Composition of surface sediments along the water-depth transect. All samples were collected using a clamshell dredge. ............................................. 47
5-1 Multivariate correlations between elements of core sections from site PI-6. .................................................................................................................... 70
11
LIST OF FIGURES
Figure
page
1-1 Maps of the Americas showing the location of the study site. ............................ 24
1-2 Lake Petén Itzá climate regime .......................................................................... 25
3-1 Flowchart of field data and operations ................................................................ 34
3-2 Gretag-MacBeth mini checker calibration card.. ................................................. 35
3-3 Surface sediments collected along the water-depth transect ............................. 36
3-4 Normalized sediment core images. .................................................................... 37
3-5 Sediment core image granularity threshold images. ........................................... 38
4-1 PI-6 and PI-3 radiocarbon ages and composite PI-6 age-model. ....................... 48
4-2 Magnetic susceptibility of PI-3 and PI-6. ........................................................... 49
4-3 Surface sediment transect L*a*b* versus water depth meters (m). .................. 50
4-4 Surface sediment transect %CaCO3 and % organic carbon versus water depth. ................................................................................................................. 51
4-5 Surface sediment transect %N, and % organic carbon ...................................... 52
4-6 Surface sediment transect total phosphorus....................................................... 53
4-7 Digital images of various lithologic units from Lake Petén Itzá cores ................. 54
4-8 Stratigraphy of PI-6 (with permission from A. Müller). ........................................ 55
4-9 PI-6 composite Gamma Ray Attenuation Porosity Evaluation (GRAPE) density record (g/cm3)......................................................................................... 56
4-10 PI-6 composite magnetic susceptibility record (SI x E-06). .................................. 57
4-11 Selected X-Ray Fluorescence (XRF) data (kcps) from site PI-6-6C-4H-1 plotted versus depth in the core section (mm). ................................................... 58
4-12 PI-6 Composite green color reflectance data versus age raw data (ybp). .......... 59
4-13 L* image data versus age (ybp). ......................................................................... 60
4-14 a* image data versus age (ybp). ......................................................................... 61
12
4-15 b* image data versus age (ybp). ......................................................................... 62
4-16 Image-derived granularity versus age (ybp) ....................................................... 63
5-1 Cartoon depicting stratification of Lake Petén Itzá .............................................. 70
5-2 CaCO3 content of modern surface sediment samples. ....................................... 71
5-3 Surface sample L* and a* vs depth. ................................................................... 72
5-4 Multivariate correlation matrix of PI-6 XRF data ................................................. 73
5-5 PI-6 6C-4H-1 XRF Ca, Fe, Sr, and Ti data (kcps). ............................................. 74
5-6 PI-6 6C-4H-1 XRF S, Mn, K, and Si data (kcps). ................................................ 75
5-7 PI-6 6C-4H-1 L*a*b* data. . ............................................................................... 76
5-8 PI-6 composite oxygen isotopic data (blue) courtesy of Jaime Escobar (in prep), and a* color data. ..................................................................................... 77
5-9 PI-6 Composite estimate of grain size proxy (using the “find edges” based algorithm). Magnetic susceptibility composite data (courtesy of Dr. A. Gilli) ...... 78
6-1 Pollen accumulation, charcoal concentration, magnetic susceptibility from site PI-6, color reflectance, and δ18O of the Greenland Ice Core. ....................... 81
6-2 PI-6 a* data plotted vs time and the Cariaco 550 nm reflectance plotted on its independent time scale. ...................................................................................... 82
6-3 a* frequency spectrum, colored horizontal bars indicate significance (red= 99.9, yellow= 99, green= 95, teal=90, white=50). ............................................... 83
6-4 a* frequency spectrum following cubic spline filter to remove low frequency signal. ................................................................................................................. 84
6-5 a* wavelet surface plot, statistical envelope shown with curved lines and transparent mask. ............................................................................................... 85
6-6 L*a* plots of PI-6 and age-model fitted images showing the relation between composite sediment core images and the L* and a* variables. .......................... 86
6-7 Summer insolation at 17 degrees north (Analyseries) plotted vs age (kybp). ..... 87
6-8 Cross-correlation between filtered a* and summer insolation at 17 degrees north. .................................................................................................................. 88
B-1 Hydrobiidae gastropods (1mm ticks) .................................................................. 97
13
B-2 A cartoon illustrating the implications for a 20-meter rise in the level of Lake Petén Itzá. .......................................................................................................... 98
C-1 PI-3 composite lithological description by Andreas Müller. ............................... 100
14
LIST OF ABBREVIATIONS AMS Accelerator Mass Spectrometry
AMSL Above Mean Sea Level
DIS Drilling Information System
DOSECC Drilling, Observation and Sampling of the Earth’s Continental Crust
dpi dots per inch
GB Gigabyte
GPS Global Positioning System
GRAPE Gamma Ray Attenuation Porosity Evaluation
GUI Graphical User Interface
HPC Hydraulic Piston Corer
IC Inorganic Carbon
ICDP International Continental Scientific Drilling Program
IODP Integrated Ocean Drilling Program
L*a*b* Lightness, a, and b colorspace
LGM Late Glacial Maximum
m meter
MB Megabyte
MBLF Meters Below Lake Floor
MCD Mean Composite Depth
MIS Marine Isotope Stage
MSL Mean Sea Level
MSCL Multi-Sensor Core Logger
PISDP Petén Itzá Scientific Drilling Project
px pixel(s)
15
TC Total Carbon
TN Total Nitrogen
TP Total Phosphorus
XRF X-Ray Fluorescence
ybp years before present (1950)
zmax maximum depth in meters
%OC Percent Organic Carbon
16
Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the
Requirements for the Master of Science
SEDIMENT CORE IMAGES AS CLIMATE AND PALEOENVIRONMENTAL INDICATORS: EVIDENCE FROM MODERN AND HISTORICAL SEDIMENTS OF
LAKE PETÉN ITZÁ, GUATEMALA
By
Dustin Aaron Grzesik
May 2010
Chair: Mark Brenner Major: Geology
Drilling of Lake Petén Itzá in the lowlands of northern Guatemala during February
and March 2006 recovered 1,327 meters of sediment from seven sites. Accelerator
Mass Spectrometry (AMS) 14C dates, and a tephra layer from a volcanic event of known
age indicate that site (PI-6) has a sediment record spanning the past 84,000 years.
Sediment core images were used to interpret the paleoenvironmental and paleoclimatic
history of the region.
Sediment image profiles were used successfully for stratigraphic correlation
among cores collected from Lake Petén Itzá. A novel technique for image processing
and automation was developed using ImageJ, the premier free, open-source software
package funded by the National Institutes of Health. Semi-automatic routines for image
normalization were necessary for quantitative comparison of sediment color data due to
the large size of the dataset and the limitations of graphics processing engines of
current computer hardware.
Current hardware, software, and the utility of RGB analysis within a complex
lithologic setting were addressed by tailoring algorithms and a multi-variable approach.
17
A routine to determine the relative granularity of the sediments using the core images
was developed. The resulting record displayed an inverse relation to magnetic
susceptibility. L*a*b* colorspace data suggest the influence of precessional forcing.
The a* variable is coherent and in-phase with precessionally influenced summer
insolation. Changes in a composite record image clearly illustrate the transition from
the Last Glacial Maximum (LGM) to the deglacial and modern conditions. Sediment
data suggest that between 25,000 and 18,000 years before present the climate of
Petén, Guatemala was wet and cool.
Image data were used to aid interpretation of magnetic susceptibility data and
other preliminary core-logging data. The image and susceptibility data suggest
correlations with climate events during the termination of the Pleistocene, including
possible winter precipitation and response to a meltwater flux into the Gulf of Mexico.
The image data also suggest that lightness is influenced by various depositional and
diagenetic processes, as well as analytical methods. The lightness of sediments from
site PI-6 was not correlated directly to changes in paleoclimate, although sediment
lightness and carbonate content was related in the modern surface samples. The a*
variable was found to be correlated to shifts in solar insolation and to be related to the
oxidative regimes in the modern surface samples. The estimate of granularity illustrates
the utility of developing complex algorithms for rapid, quantitative identification of
features and characteristics of the sediment.
18
CHAPTER 1 INTRODUCTION
Paleolimnological research has provided high-resolution records of past
continental climate and environmental change (Smol, 1992). Several long-term climate
records have been recovered from the continents and oceans. However, because of
expanded glacial ice cover at high latitudes it has proved difficult to find ice-age climate
records from low-elevation tropical lakes (Deevey, 1957). Many of these lakes,
especially the shallower ones, dried completely during the glacial advance. Here I
examine high-resolution digital images of sediment cores from the deep Lake Petén Itzá
basin, Guatemala, and explore the potential for using them to infer paleoenvironmental
conditions over the past ~85,000 years. I also discuss the use of sediment images, and
the characteristics of Red Green Blue (RGB) analysis and other image processing
techniques for project planning and sampling strategy.
I argue that color reflectance of surface sediment samples collected along a water-
depth transect in Lake Petén Itzá reflect various properties of the sediment that are
indicative of environmental conditions. Modern surface sediment depth transect
samples vary in geochemical and optical properties. My hypothesis is that color
reflectance measured in Lake Petén Itzá sediment cores can be related to reflectance in
the surface sediment samples from the water-depth transect and used to infer changes
in paleoenvironmental conditions. Shifts in the location of the carbonate depositional
zone cause changes in sediment lightness and such changes in sediment lightness
within a core thus indicate past changes in water depth. The comparison of X-Ray
Fluorescence (XRF) data, core composite images and paleoclimate proxies are used to
test the hypothesis. Confounding factors include diagenetic changes, and strata that
19
are not represented in the modern lake sediments and limit the application of lightness
as a paleoclimatic proxy in the Petén Itzá sediment record. Hue-specific image
interpretation, such as a* image data provide a mechanism for paleoclimatic
interpretation from sediment core image data of Lake Petén Itzá.
The Petén Itzá Scientific Drilling Project (Hodell et al., 2006) recovered 1,327 m of
sediment core from seven sites in early 2006. These cores are the first such long
records from the lowland Neotropics. Operating procedures generally followed the
workflow standards of the Integrated Ocean Drilling Program (IODP). Sediments
recovered from Lake Petén Itzá provide the opportunity to study global and regional
climate change. Analyses of oxygen isotopes and pollen in cores collected in 2002
provided insight into the impact of climate change on Classic Maya culture (Hodell et al.,
2005; Brenner et al., 2002). Such studies are relatively expensive and time consuming,
and require destructive analyses of sediments.
Analytical Challenges
Recent developments of core logging methods and equipment such as the Geotek
Multi-Sensor Core Logger (MSCL) have allowed researchers to assess sediment
characteristics rapidly. Measures that can be made include Gamma Ray Attenuation
Porosity Evaluation (GRAPE) density, magnetic susceptibility, p-wave velocity, natural
gamma, and high-resolution digital imaging.
Automated logging of lake sediment cores and analysis with scanning XRF have
allowed scientists to generate highly detailed chemical profiles of sediment cores,
minimizing the need to perform time consuming “wet” chemical analyses. Scanning
XRF techniques, however, are still not widely available and require expensive,
specialized equipment.
20
Imaging
Imaging techniques, including microscopy and Munsell color charts, have long
played a role in sediment and geological description and interpretation. Images,
however, have traditionally been used in a descriptive, qualitative fashion, failing to take
advantage of the valuable quantitative information that may be extracted from these
analog and digital data. As imaging technologies have become cheaper and more
reliable, image analysis techniques have been used increasingly in sediment studies
(Francus, 2004).
Varved lake sediment cores provide an ideal subject for image analysis given their
dramatic changes in color/shade. They possess layers that are clearly distinguishable
both by eye and by the sensors of digital cameras. Varved deposits, however, are rare.
Non-laminated sediments, with more subtle changes in hue and intensity, may best
illustrate the quantitative powers of digital image analysis.
21
CHAPTER 2 BACKGROUND
Petén Paleolimnology
Climate change in the lowland tropics of northern Guatemala was first explored by
Dr. Edward S. Deevey, a pioneer in paleolimnological study of the region. The search
for Pleistocene-age sediments in the region began in 1940 when Deevey cored lakes in
Mexico (Deevey, 1955). This early “boring” expedition failed to recover Pleistocene lake
deposits, but in a prescient speculation, Deevey noted that such a record should be
“sought in the Petén, in Guatemala” (Deevey, 1957).
In May 1980 Deevey and others cored several relatively small lakes in Petén,
including Quexil, Salpeten, and Macanche. The region was selected because lake
basins in the area were thought to have been deep enough to have held water during
the Last Glacial Maximum (LGM) (Deevey et al., 1980). Deevey et. al. (1983), referring
to the 20-m core that was retrieved from Lake Quexil, noted that “Recovery of
Pleistocene lacustrine deposits culminates a forty-year search…” Indeed, sediments
spanning the Holocene and penetrating into the late Pleistocene had been collected.
The record, however, remained poorly dated and was discontinuous, in that there were
gaps between each of the sections recovered by the split-spoon corer. Furthermore,
there was some concern that the oldest deposits might contain hiatuses, or were
deposited in very shallow water. More than 25 years passed before further attempts
were made to raise Pleistocene sediments from a Petén lake.
Petén Itzá Limnology and Setting
Lake Petén Itzá (Figure 1-1) has a surface area of ~100 km2 (16°55’ North,
~89°50’ West, 110 meters AMSL). It was formed by a combination of tectonic activity
22
and solution processes (Vinson, 1962). The lake is sensitive to changes in climate and
lake level fluctuates in response to precipitation (Deevey et al., 1980).
The dominant cation and anion in Lake Petén Itzá’s water are, in terms of milli-
equivalents, calcium and sulfate, respectively, with magnesium and bicarbonate
following in abundance (Deevey et al., 1980: Hillesheim, 2005). The lake’s pH is basic
(8.0), and sediments in water depths <~30 m are rich in the remains of calcareous
organisms, such as snails and ostracods. High calcium carbonate (CaCO3) content of
shallow-water deposits forms a lighter “halo” in the littoral zone of the lake, particularly
in the gradually sloping southern basin, where sediments are replete with the remains of
lacustrine gastropods and other calcareous microfossils (Covich, 1976; Curtis et al.,
1998).
The modern lake level fluctuates in response to variations in
evaporation/precipitation (E/P). The water body has no major outlets and loss to
groundwater seepage is minor so that the basin is effectively “closed”. Lake Petén Itzá
is sensitive to global climate through shifts in the seasonal movement of the Inter-
Tropical Convergence Zone (ITCZ) (Hodell et al., 2001). The lake is located along a
steep north-south gradient of precipitation on the Yucatan Peninsula (Wilson, 1980)
(Figure 1-2). The climatologic setting makes it sensitive to changes in the ITCZ and
precipitation in the region.
Cultural Setting
Lake Petén Itzá is part of a chain of lakes that includes, from west to east,
Perdida, Sacpuy, Petenxil, Quexil, Salpeten, Macanche, Yaxha, and Sacnab. This
region, referred to as the Petén Lake District, constitutes part of the core region for
ancient lowland Maya settlement. Located approximately 50 km north are the
23
archaeological ruins of Tikal. Sediments from Lake Petén Itzá suggest that changes in
Holocene climate may have had an impact on Maya cultural evolution (Müller et al.
2009, 2010).
Petén Itzá Scientific Drilling Project (PISDP)
Lake Petén Itzá (zmax = 160 m) was selected as a potential candidate for lake
drilling after detailed seismic surveys were completed (Anselmetti et al. 2005), and
Kullenberg coring was unable to penetrate stiff gypsum–rich sediments about 6 m below
the lake floor, deposited during the late Glacial (Hillesheim, 2005). Seismic stratigraphy
provided evidence for up to 100 m of sediment in some basins (Anselmetti, 2006).
24
Figure 1-1. Maps of the Americas showing the location of the study site. (a) the Petén Lake District, northern Guatemala. (Modified from Anselmetti et al. 2006). (b) Lakes of the Petén Lake District (Modified from Hilleshiem et al. 2005). (c) Coring sites and bathymetry of Lake Petén Itzá (from Anselmetti et al. 2006). Drilling sites were selected after preliminary seismic surveys and Kullenberg coring provided evidence for a continuous sediment record.
25
Figure 1-2. Lake Petén Itzá lies in an area that may be described as a tropical monsoon
climate regime, along the steep precipitation gradient of the Yucatan peninsula (Modified from Wilson, 1980). Contours are mm of annual rainfall.
26
CHAPTER 3 METHODS
Field Methods
Sediment Drilling
Sediment cores were collected during the Lake Petén Itzá Scientific Drilling Project
(PISDP) in February and March 2006. The modified GLAD800 (Global Lake Drilling
System, aka RV Kerry Kelts) served as the drilling platform for core recovery. A team
from Drilling, Observation and Sampling of the Earth’s Continental Crust (DOSECC),
and scientists of the PISDP collected cores around the clock, during two twelve-hour
shifts. All tolled, 1,327 m of sediment was collected from seven sites (Table 3-1).
Drilling was accomplished using proprietary DOSECC coring tools including the
Hydraulic Piston Corer (HPC), Extended Nose, and Alien. The HPC was the most used
coring tool during the PISDP because of the tool’s suitability for recovering complete, 3-
m sections of undisturbed clay-rich sediments with each “drive.” Prior to each “drive,”
the coring tool was fitted with a polycarbonate liner. After a successful drive, the
sediment-filled liner was retrieved, cut into 1.5 m lengths, capped and labeled. All core
sections were measured and labeled with Expedition, Site, Hole, Core, Section, and
Type (tool type) in preparation for shipping and storage.
The 3-meter “stroke” of the coring tools left gaps or section breaks between drives.
To recover sediment from the missing intervals it was necessary to core secondary and
tertiary holes near the first hole. Drives in secondary hole B were offset vertically by
~1.5 m with respect to depths in primary hole A. Hole C was typically offset by 0.75 m
relative to holes A and B. At sites PI-1, PI-3, PI-4, PI-6, PI-7 the team first cored hole A,
27
and subsequently holes B and C. Coring at shallow-water site PI-9 was halted after
slumping prevented drilling progress. Five holes (A, B, C, D, E) were drilled at Site PI-2.
Following each work shift change, sediment cores were transported to the land-
based laboratory where science team members measured gamma ray attenuation
(GRAPE density), magnetic susceptibility, and p-wave velocity using a Geotek Multi-
Sensor Core Logger (MSCL) provided by the International Continental Drilling Program
(ICDP). Scientists also measured pore-fluid alkalinity and pH, and analyzed smear
slides to describe core lithology. Platform and lab-collected data were managed and
stored using the Drilling Information System (DIS) database developed and supported
by the ICDP. Sediment cores were bar-coded, boxed, and stored in an on-site
refrigerated shipping container. Upon completion of the field campaign, the sediment
cores were shipped to the Lacustrine Core Repository (LacCore) at the University of
Minnesota, Minneapolis.
Core Compositing with SPLICER
MSCL data were correlated stratigraphically with SPLICER software to enable
real-time drilling decisions for increased efficiency and to guarantee recovery of
complete sediment records at each site (Figure 3-1). SPLICER software is used often
on IODP drilling projects, but has not been used frequently in lake drilling projects
because of equipment and personnel limitations. SPLICER played an important role in
drilling decisions and the shore-based laboratory team members would often contact
the drilling platform via cellular phone to discuss drilling depth offsets for the final hole at
a site. SPLICER’s graphical user interface (GUI) allows movement of data from core
sections from different holes at a site in relation to one another, so that features may be
28
correlated. A real-time correlation coefficient is displayed. A composite record for each
site was created by “depth-shifting” cores from multiple holes at a site, using SPLICER.
During the field campaign, GRAPE density and magnetic susceptibility were used
for splicing. A key, called a “Splice table” was created by moving down-core, and
selecting tie points from one core section to another. Splicer allowed the science team
to identify gaps in recovery and to target sediment gaps that had to be recovered in the
final hole. Splicer was also used to construct the composite section for sampling and
analysis.
Modern Surface Sediment Transect
In addition to sediment coring, a suite of surface sediment samples was collected
along a water-depth transect from the shoreline to ~100 m water depth, using a
clamshell dredge (Table 3-2). Global Positioning System (GPS) coordinates and water
depth were recorded at each location.
Analytical Methods
PI-6 was selected as the primary site for study because of the continuity of the
record and the apparent lack of sediment slumping. Disturbances and slumping were
observed in the record of PI-3 (Appendix Figure C-1). In June 2006 the PISDP held a
sampling party at the LacCore facility. Sediment core liners were split lengthwise using
an oscillating cast-saw. Soft clay sediments were split with a metal plate and stiffer
deposits were cut with a metal guitar string. Each core was split into working and
archive halves. The core surfaces were cleaned and smoothed using glass slides. The
working halves were sampled for sediment analyses that are destructive. The archival
halves were used for non-destructive imaging, logging, and lithologic description.
29
Subsampling
Sediment subsamples were collected for analyses of pollen, diatoms, δ18O and
geomicrobiology. During this process, cores were described and a lithologic profile was
created. Visual inspection of cores improved the understanding of among-hole
correlations and the final, spliced composite core table was created. U-channels,
elongate plastic boxes 2 cm wide by 2 cm deep, were cut to length according to the
SPLICER table. The cut channels were placed on top of the working core half and
firmly pressed into the sediment. Polypropylene line was pulled along the open face of
the U-channell to cut the sediment-filled box from the remaining sediment in the core
tube. Next, the channels were capped and prepared for transport.
Image Acquisition
Sediment image data were collected using a Geotek multi-track core line scan
camera, which collects images free from parallax or barrel distortion. The camera
scans the core sections line by line as it moves down a track. Archival core halves were
placed one at a time on the core line scanner and a Gretag-MacBeth Mini Checker
calibration card was placed at the bottom of each core section (Figure 3-2). The light
source of the Geotek was polarized with a film. An additional polarizer on the camera
allowed for glare-free imaging. Images were saved as lossless RGB (.bmp or .tif) image
files with a full resolution of 10 pixels (px) per mm and average file size of ~40 MB per
1.5-m core section.
A sample tray was built to hold modern surface sediment samples collected along
the water depth transect. The tray was loaded with sediment samples and placed on
the Geotek multi-track line scanner and imaged under conditions identical to those used
for scanning the sediment cores (Figure 3-3).
30
Red, green, blue color-digital images
Red, Green, Blue (RGB) color-digital images typically consist of a matrix of
individual pixels with numerical values that represent the intensity of three hues; Red,
Green, and Blue. These pixels have discrete values assigned to them that are
dependent on the bit-depth of the camera. The range of intensity values that cameras
can record is known as the dynamic range. An 8-bit color image has RGB values that
range from 0 (black) to 255 (white). Conventionally, 0 indicates the darkest possible
pixels (pure black) and 255 represents the brightest (pure white) (Russ, 1999). In this
system, a black pixel has the values (0, 0, 0), a white pixel is coded (255, 255, 255), a
red one is (255, 0, 0), and so on. With this system, it is possible to quantify the relative
change in hue of an object’s surface from “reddish” to “greenish.” Similarly, with a
grayscale image, it is possible to determine the relative change in intensity from the
darkest to lightest areas.
L*a*b* colorspace
L*a*b* colorspace divides color images into three components. Lightness is
named the L* component where L* of 0 is black and L* of 100 is white. The ratio of
green to red is named a*. In the a* channel, positive (+) a* levels are indicative of red
and negative (-) values are greenish. Finally, the ratio of yellow to blue is named b*
where (-) values are blue and (+) are yellow. These ratios do not include the effect of
lightness (Francus, 2004). The combination of these two channels a*(+) and b*(+)
represents yellow. Sediments shifting from “greenish” to “reddish” can be identified by
the shift in a* values from (-) to (+).
Because the sediment core and transect sample images were calibrated, it was
possible to perform a transformation from RGB to L*a*b* colorspace for analysis. The
31
transformed L*a*b* images were analyzed using the ImageJ profile macro used for RGB
images.
Image Processing and Analysis
Post acquisition image processing was performed with ImageJ image analysis
software (Appendix A). Because of the limited range of the 8-bit sensors, post-
acquisition normalization was done to compare sections with highly variable sediment
lightness. Normalization of the image dataset was performed using an algorithm that
measured RGB values of the red, green and blue tiles from the Gretag-MacBeth Mini
Checker from the entire dataset (Figure 3-4). After normalization, the images were
cropped, labeled and added to a composite image dataset developed with the SPLICER
“splice” file.
Sulfur nodules were excluded from the composite image record because they
were thought to have been created by post-depositional diagenesis, and therefore are
not a proxy for paleoenvironmental or paleoclimatic conditions. Similarly, the image
dataset was scrutinized and gaps in the core sections, or disturbances due to coring,
were also excluded from the composite by filling the disturbed areas with pure black (0,
0, 0). These quality control decisions were made after examining the sediment core
images, and would not have been possible by visual examination or measurements
through the core liner.
A composite RGB profile was created using an ImageJ macro. The macro was
modified to include a “find edges” command, and the images were “thresholded” to
estimate granularity, i.e. shifts in grain-size (Figure 3-5). In addition, a L*a*b*
(Lightness, a and b) colorspace conversion and profile was created with a modified
version of the macro. Corelyzer was used to visualize the core image composite
32
dataset at full resolution. Images were used to aid stratigraphic correlation among holes
and to project 14C dates from site PI-3 to PI-6.
Geochemical Analyses
Samples from the surface sediment transect were freeze-dried and crushed with a
mortar and pestle. Total carbon (TC) and total nitrogen (TN) were analyzed using a
Carlo Erba NA 1500 C/N/S analyzer. Total phosphorus in sediments (TP) was analyzed
using a Bran-Luebbe auto analyzer. Inorganic carbon (IC) was measured by
coulometric titration using a UIC/Coulometrics 5011 coulometer and a UIC 5240-TIC
carbonate autosampler. Percent organic carbon by weight (%OC) was figured as TC-
IC. Select archival core sections were analyzed for element concentrations at the
University of Minnesota, Duluth by Dr. Erik Brown, using an ITRAX X-Ray Fluorescence
(XRF) core scanner.
Core Chronology
Samples of wood, leaves, and charcoal were taken from the sediments for 14C
dating. Samples were pre-treated (acid-base-acid treatment) and analyzed at the
Lawrence Livermore National Laboratory using Accelerator Mass Spectrometry (AMS).
Dates were calibrated using Fairbanks et al. (2005) using the on-line radiocarbon
calibration program. Stratigraphic correlation between the magnetic susceptibility
records from sites PI-6 and PI-3 permitted use of radiocarbon dates from the PI-3 core
in the final age model for the PI-6 composite section. Electron microprobe glass
analysis was used to identify known tephra layers and extended the chronology beyond
the limits of 14C dating.
33
Table 3-1. Drilling site summary. Drilling at site PI-9 was stopped after slumping caused high pressure on the drill-string and prevented advancement.
Water Depth Penetration
Depth (mblf) Average %
Site Latitude Longitude (m) Hole
A Hole B Hole C Holes D/E Recovery
PI-1 16° 59.9706' N 89° 47.7396' W 65 94.5 90.3 82.5 89.3 PI-2 16° 59.9712' N 89° 44.685' W 54 66.5 41.2 82.4 42/68.5 86.3 PI-3 17° 0.2016' N 89° 49.24' W 100 96.9 95.3 90 92.9 PI-4 17° 0.3342' N 89° 50.772' W 150 67.4 46.1 25.4 86.7 PI-6 17° 0.0162' N 89° 47.0868' W 71 75.9 66.4 66.8 94.9 PI-7 16° 59.7234' N 89° 47.6844' W 46 133.2 122.8 63.8 92.1 PI-9 16° 59.436' N 89° 47.646' W 30 16.4 91.8
Table 3-2. Locations of surface sediment samples collected along a water-depth
transect. Site Latitude Longitude
Water Depth (M)
TS06 89° 47.309' 16° 58.469' 1
TS05 89° 47.309' 16° 58.469' 1.2
TS07 89° 47.283' 16° 58.54' 1.5
TS08 89° 47.336' 16° 58.54' 1.5
TS09 89° 47.346' 16° 58.6' 2
TS10 89° 47.357' 16° 58.637' 3
TS24 89° 41.81' 16° 59.662' 5
TS11 89° 47.419' 16° 58.727' 6
TS12 89° 47.429' 16° 58.795' 7
TS13 89° 47.466' 16° 58.884' 8.3
TS14 89° 47.497' 16° 58.988' 10
TS15 89° 47.547' 16° 59.134' 12
TS16 89° 47.564' 16° 59.191' 14
TS17 89° 47.527' 16° 59.248' 17.7
TS18 89° 47.563' 16° 59.296' 20.8
TS03 89° 49.663' 17° 0.317' 23
TS19 89° 47.599' 16° 59.38' 25.9
TS20 89° 47.613' 16° 59.48' 32
TS02 89° 43.282' 16° 59.37' 40
TS21 89° 47.663' 16° 59.682' 40.6
TS22 89° 47.623' 16° 59.818' 50 TS23 89° 47.542' 17° 0.082' 70
TS04 89° 49.657' 17° 0.324' 108
34
Figure 3-1. Flowchart of field data and operations
35
Figure 3-2. Gretag-MacBeth mini checker calibration card. The lower right square (pure
white) was used for image recognition and calibration.
36
Figure 3-3. Surface sediments collected along the water-depth transect in trays arranged from shallowest water (upper-left) to deepest water (lower-right). Refer to Table 3-2 for locations and water depths.
37
Figure 3-4. Normalized sediment core images. Top 3 sections pre-normalization, bottom 3 are post-normalization. From top to bottom cores sections are PI-6-6A-24E-2, PI-6-6A-10H-2, and PI-6-6A-7H-2.
38
Figure 3-5. Sediment core image granularity threshold images. Top core images are original gypsum dominant, and clay
dominant sections characterized by coarse and relatively fine grainsize. The thresholded images below are red where edges indicate a roughness in image texture. Black vertical lines in the top grainsize threshold image indicate areas excluded from the composite record. From top to bottom are sections from cores and PI-6-6B-5H1, and PI-6-6B-2H-2
39
CHAPTER 4 RESULTS AND DISCUSSION
Site PI-6 Chronology
Twenty radiocarbon dates from the PI-6 composite section were used to establish
the core chronology (Table 4-1). Not all radiocarbon ages were in stratigraphic order.
A tephra layer near the base of the section has a geochemical signature consistent with
the Los Chocoyos eruption of Atitlan, which occurred approximately 84,000 ybp. An
age-depth model was generated by linear interpolation between points (Figure 4-1). At
depths with more than three 14C dates grouped closely together, a single age/depth
point was established using mean values for the ages/depths in the group. The date
from 6C 6H-1 120-121 cm was excluded because of contamination concerns. The date
from 6A 20-E-1 45 was beyond the range of 14C dating and was excluded from the
model.
As mentioned, stratigraphic correlation with site PI-3 allowed projection of 14C
dates from cores at PI-3 onto PI-6 for development of the PI-6 composite age model
(Figure 4-2). Several plateaus in the 14C calibration curve used to convert 14C years to
calendar years occur during the termination of the last glacial period and early
Holocene. There are two possible volcanic ashes in sediments deposited during the
last deglaciation (17,000-10,000 ybp) that may provide additional age constraints.
Surface Sediments from the Water-Depth Transect
Transect Image Data
Surface sediment samples from the water depth transect varied with respect to L*,
a* and b* color channels. Variations in lightness (L*) and the a* channel are more
closely correlated than is either to the b* variable (Figure 4-3). Sediments from depths
40
0-14 m have the highest L* values. Samples from depths >14 m are characterized by
lower L* values and are darker than sediments from shallower water. The a* values of
sediments from the shallowest water fluctuate between values of 5 and approximately -
3.5, indicating that shallowest-water sediments may be both reddish (+) and greenish
(-). Sediment sample a* values from depths of 5-14 m are in the range of 4-6,
representing generally reddish sediments. At depths >14 m, surface sediments range
from approximately 0.5 to -2 and are generally greenish. The b* values range from 21
to 31 for samples from 0-14 m. At greater depths, the sample b* values range from 18
to 25.
Transect Geochemistry
CaCO3 and organic carbon content were measured on samples from the surface
sediment transect. There is an inverse relationship between the two constituents
(Figure 4-4). CaCO3 content in sediments ranged from 50% to 80% of dry mass in
samples from depths of 0-20 m. The CaCO3 content dropped to ~40% in samples from
25-50 m, and to 5% at the 108-m site. Percent organic carbon was highest, ~11%, in
the sample from 110 m water depth, and ranged from 2 to 8% in samples from 0-40 m
water depth.
Percent nitrogen (%N) and percent organic carbon (%orgC) in surface sediments
varied similarly as a function of water depth (r = 0.91) (Figure 4-5). Total phosphorus
generally increased with water depth, ranging from 0.03 to 0.3 mg/g, with the highest
concentration, 0.6 mg/g recorded at a depth of ~26 m (Figure 4-6).
41
Transect Lithology
Modern surface sediments are dominated by organic-rich (gyttja) deposits in deep
water (>20 m) and carbonate-rich sediments in the shallow, littoral zone. Sediments
from shallowest-water sites contained organic debris, including root material from
aquatic macrophytes. Gastropods and coarse material were found in high
concentrations of winnowed sediments collected at a water depth of ~20 m.
Site PI-6 Sediment Cores
Site PI-6 Lithology
The sediment cores are characterized by several lithologic facies, including coarse
gypsum sands. In addition, gray clay, carbonate turbidites, and sulfur nodules are
present (Figure 4-7). A detailed lithologic description was constructed by Andreas
Müller (Figure 4-8) (Müller et al. 2010).
Coarse authigenic gypsum sands and crystals in some glacial-age sections
suggest a high evaporation to precipitation ratio (E/P) in Petén at the time of deposition.
Sulfur nodules observed in the sediment record are post-depositional, diagenetically
formed facies. Sulfur-reducing bacteria have been cultured from the sulfur nodules
(Vasconcelos et al. 2006). Abundant hydrogen sulfide at depths in the cores suggests
the absence of oxidative metals in the sediments.
Core Logging Data
Density
GRAPE density was measured on cores from sites PI-1, PI-2, PI-3, PI-4, PI-7, and
PI-9 on-site during the field campaign. PI-6 was also logged at the LRC. The PI-6
density record is reported in units of g/cm3 as determined on the Geotek MSCL (Figure
4-9). Data were processed to remove gaps. Total drive recovery was calculated by
42
subtracting the sum of all MSCL section lengths from the total drive lengths recorded in
the field-drilling database.
The oldest sediments are marked by density fluctuations on millennial to sub-
millennial scales, with values ranging from 1.5 to 1.9 g/cm3. Between ~24,000 and
18,000 ybp, sediment density is low (~1.5 g/cm3). A transition to higher density (~1.8
g/cm3) at 18,000 persists until ~9,000 ybp. The remaining record is marked by a shift to
lower-density Holocene sediments.
GRAPE density values reflect changes in lithology. Clays have density values
~1.5 g/cm3. The density of gypsum-rich sediment ranges between 1.8 and 2.0 g/cm3.
Changes in density also reflect gaps such as gas pockets, sediment breaks or other
disturbances. Because sediment density was measured on unopened core tubes, gaps
or other disturbances were not identified until core liners were opened and cores were
split. This limits the accuracy and utility of GRAPE density. Nevertheless, the measure
was still useful for on-site core correlation for making drilling decisions.
Magnetic susceptibility
Magnetic susceptibility data are presented as SI units (SI x E-06) (Figure 4-10). A
spike with a maximum of ~200 SI x E-06 in the susceptibility record is observed at
~83,000 ybp. It is not associated with a visible ash layer. Sediments with low, but
variable susceptibility persist through ~25,000 ybp. Between 25,000 ybp and 18,000
ybp, values are relatively high, ranging from 45 to 50 SI x E-06. At 18,000 ybp there is
an abrupt transition to lower susceptibility, which generally persists through 9,000 ybp.
Two intervals of slightly higher susceptibility are exceptions during this period. Values
ranging from 30 to 40 SI x E-06 persist from ~9,000 ybp until present.
43
XRF data
Scanning XRF data were collected for selected sections of PI-6, however these
data have not been prepared as a composite record and there is not yet a composite
XRF record for the site. A suite of XRF-derived metal concentrations is plotted (Figure
4-11) to illustrate the variability of key earth elements within a core section that displays
high variability. These data and their relation to lithologic and image characterizations,
are discussed further in Chapter 5.
Composite image RGB data
The Green component of the digital images was selected because green light is
similar in wavelength to widely cited 550 nm spectrometer reflectance data. Data were
normalized to a scale in which 0 represents the darkest pixels and 255 is the brightest
value. The green profile illustrates changes in sediment lithology throughout the record
(Figure 4-12).
The oldest part of the record is marked by fluctuations on millennial and sub-
millennial scales. The Green profile remains light from ~18,000 to 9,000 ybp. The gray
clay transitions to greenish gyttja, and at about 9,000 ybp, the sediment is darker. The
Holocene sediments lighten at about 5,000 ybp, with the shift to gray clay. Sediments
near the sediment surface are darker, similar to modern greenish-brown gyttja found in
surface sediments from the deeper parts of the lake.
RGB profiles of the surface sediment transect illustrate variability with water depth.
Sediments from the shallower littoral zone are light, with values of 150-200. At depths >
20 m, however, samples are darker, with values ranging from 50-100. The surface
sediment sample collected at 110 m water depth had a value <50.
44
L*a*b* colorspace
L*a*b* colorspace analyses generated profiles from the sediment core images of
Lake Petén Itza. The response of the lightness (L*) channel of the Lab colorspace of
the sediment record was very similar to the individual green reflectance channel (Figure
4-13). The oldest part of the a* record is characterized by millennial and sub-millenial
variations and displays a high degree of noise (Figure 4-14). Minima in the record occur
at approximately 41,000 and 21,000 ybp. From the 21,000 ybp minimum until
approximately 9,000 ybp, the record displays a trend toward more positive values, but
with significant variability during the climb. An abrupt decrease in a* is observed at
approximately 9,000 ybp, after which the record stabilizes and a* begins to increase.
The b* channel (Figure 4-15) shows patterns that are similar to those of green
reflectance and L*. The b* record, however, is less variable than either the L* or green
reflectance.
PI-6 granularity
Grain size analyses are presented as a proportion of a 1-cm wide cross-section of
the core composite image area that was thresholded after being processed with a “find
edges” command (Figure 4-16). The earliest part of record is characterized by mostly
smooth surfaces, except for 1-2-kyr intervals at 55,000 ybp, 53,000 ybp and 45,000 ybp.
A transition to sediments with high granularity occurs at ~40,000 ybp, after which there
is an increase in granularity through ~25,000 ybp. Between 25,000 and 18,000 ybp the
sediments were generally smooth. Between 18,000 and 15,000 ybp there is a shift in
values that suggests a moderate decrease in granularity. A minimum in granularity
occurs at about 15,000 ybp. Since then, granularity oscillates, with alternating smooth
45
and rough surfaces through ~9,000 ypb. Following the Pleistocene-Holocene transition,
there is a shift to much smoother surfaces, typical of the modern, deepwater sediment.
46
Table 4-1. Dates and tie points used to develop the chronology for site PI-6 including sample, depth, 14C dates, positions. AMS 14C ages were measured at the Lawrence Livermore National Laboratory.
Accession# Site-hole
core-type-section
Depth-interval
Depth (mcd)
14C yr BP ±(1σ)
Age (cal yr BP) (1σ)
125895 6C 3H-1 141.5 cm 8.076 3,920 35 4,376 53 125903 6C 3H-1 141.5 cm 8.076 3,905 35 4,355 61 125896 6B 4H-2 77.5 cm 10.406 8,655 30 9,579 31 125904 6B 4H-2 77.5 cm 10.406 8,630 60 9,572 52 131225 6C 4H-1 59 cm 10.436 7,985 40 8,882 102 131226 6C 4H-1 101 cm 10.856 9,040 35 10,213 16 128611 6B 5H-2 33 cm 13.354 11,290 60 13,125 72 128613 6B 5H-2 39 cm 13.414 11,380 140 13,229 154 131224 6B 5H-2 85 cm 13.874 11,390 50 13,236 74 128612 6C 5H-2 7 cm 14.448 12,460 60 14,396 150 131223 6C 6H-1 120 cm 17.12 12,280 60 14,079 95 125897 6A 7H-1 128 cm 20.026 14,130 120 16,537 212 125898 6C 8H-2 6 cm 23.636 17,650 240 20,896 321 125899 6C 9H-1 62 cm 25.729 19,990 180 23,880 217 125900 6A 11H-2 145 cm 33.733 30,700 3600 36,000 3663 125901 6B 12H-1 82.5 cm 33.828 29,120 170 34,536 224 125905 6B 12H-1 82.5 cm 33.828 29,010 170 34,421 222 126525 6B 12H-1 138.5 cm 34.388 28,040 470 33,406 517 126526 6C 14H-1 80.5 cm 41.303 39,000 700 43,855 634 126528 6B 15H-1 60 cm 42.762 35,900 1200 41,145 1103 126527 6A 15H-2 90.5 cm 45.307 38,100 1100 43,080 967 126529 6A 16H-2 85 cm 46.879 40,500 2600 126530 6A 20E-1 4 cm 57.912 54,000
47
Table 4-2. Composition of surface sediments along the water-depth transect. All samples were collected using a clamshell dredge.
Sample ID %N %C %CaCO3 %orgC
Total P (mg/g)
Water Depth (m)
TS06 0.15 8.56 55.92 1.85 0.064 1 TS05 0.06 7.58 51.83 1.36 0.031 1.2 TS07 0.12 9.25 60.42 2.00 0.039 1.5 TS08 0.16 8.17 50.08 2.16 0.056 1.5 TS09 0.26 9.99 54.67 3.43 0.050 2 TS10 0.29 10.90 66.92 2.87 0.080 3 TS24 0.46 14.17 70.17 5.75 0.171 5 TS11 0.71 15.88 68.50 7.66 0.152 6 TS12 0.21 11.24 69.42 2.91 0.068 7 TS13 0.32 13.82 80.17 4.20 0.086 8.3 TS14 0.39 14.19 71.58 5.60 0.204 10 TS15 0.20 13.08 70.58 4.61 0.047 12 TS16 0.18 12.91 64.83 5.13 0.061 14 TS17 0.39 13.60 72.33 4.92 0.240 17.7 TS18 0.28 12.80 76.58 3.61 0.250 20.8 TS03 0.32 11.23 57.46 4.34 0.160 23 TS19 0.60 13.41 57.42 6.52 0.599 25.9 TS20 0.25 8.45 39.17 3.75 0.213 32 TS02 0.55 9.87 32.67 5.95 0.305 40 TS21 0.53 12.43 40.17 7.61 0.264 40.6 TS22 0.51 11.53 20.75 9.04 0.124 50 TS23 0.69 12.37 19.92 9.98 0.312 70 TS04 0.73 11.96 6.58 11.17 0.323 108
48
Age (Years BP) Age (Years BP) Figure 4-1. PI-6 and PI-3 radiocarbon ages and composite PI-6 age-model. A. Radiocarbon dates calibrated using
Fairbanks et al. (2005) versus mcd for Sites PI-3 (blue) and PI-6 (red). A sample at 59.44 mcd yielded an infinite radiocarbon age (>54 kyr). Green dots represent the position of tephra layers. CGT = Congo Tephra; Guasal1 = Guatemala–El Salvador Tephra 1; ACT = Arce Tephra; LCY = Los Chocoyos Tephra B. Combined radiocarbon dates from Sites PI-3 (blue open circles) and PI-6 (red open circles) based upon correlation of the magnetic susceptibility records in Fig. 4-2. Age model (line) was derived using a weighted fit through selected age–depth points from Sites PI-3 and PI-6 and three ash layers (CGT, Guasal1 and LCY). After Hodell et al. (2008)
49
Figure 4-2. Magnetic susceptibility of PI-3 and PI-6. (Lower panels) Spliced magnetic susceptibility records from Sites PI-3 (red) and PI-6 (blue). (Upper panel) Comparison of magnetic susceptibility records after the PI-3 record was correlated to the mcd scale of PI-6.
50
0
20
40
60
80
100
12010 20 30 40 50 60 70 80 90
-4 -2 0 2 4 6
Sediment surface sample L*a*b* vs depth
L*a*b*
Dep
th (m
)
L*, b*
a*
Figure 4-3. Surface sediment transect L*a*b* versus water depth (m). Samples were
collected using a clamshell dredge. After Hodell et al. (2008)
51
0
20
40
60
80
100
1200 20 40 60 80 100
0 2 4 6 8 10 12
Surface sediment geochemistry (% CaCO3, % organic C)
% CaCO3
% organic carbon
Dep
th (m
)
% organic C
% CaCO3
Figure 4-4. Surface sediment transect %CaCO3 and % organic carbon versus water depth. Samples from the shallowest-water sites had a large amount of plant material, reducing the %CaCO3.
52
0
20
40
60
80
100
1200 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
0 2 4 6 8 10 12
Surface sediment transect (% N, % organic C)
% nitrogen% organic carbon
Dep
th (m
)
% N
% organic C
Figure 4-5. Surface sediment transect %N, and % organic carbon are highly correlated (R= 0.91) in surface sediment transect samples.
53
0
20
40
60
80
100
1200 0.1 0.2 0.3 0.4 0.5 0.6
Surface sediment geochemistry (total phosphorus (mg/g))
Total P (mg/g)
Dep
th (m
)
Total P (mg/g)
Figure 4-6. Surface sediment transect total phosphorus.
54
a. b. c.
d. e. f.
Figure 4-7. Digital images of various lithologic units from Lake Petén Itzá cores (scale
interval is 1 cm) a. brown organic-rich gyttja b. gray clay, c. gypsum interbanded gray clay, d. carbonate turbidites within greenish clay, e. sulfur nodules, f. limestone gravel.
55
Figure 4-8. Stratigraphy of PI-6 (with permission from A. Müller). A. PI-6 lithologic units and 14C ages. B. Magnetic susceptibility and lake level inference.
56
1
1.2
1.4
1.6
1.8
2
2.2
0 10000 20000 30000 40000 50000 60000 70000 80000
PI-6 density record vs time (ybp)
density
dens
ity (g
/cm
3 )
Age (y bp)
Figure 4-9. PI-6 composite GRAPE density record (g/cm3). Black line represents original data, the red line is a 100 point running mean. After Mueller et al. (2010)
57
-20
0
20
40
60
80
100
120
140
600
800
1000
0 20000 40000 60000 80000
magnetic susceptibility (SI x E-06)
magnetic susceptibility (SI x E-06)
mag
netic
sus
cept
ibili
ty (
SI
x E
-06)
Age (y bp) Figure 4-10. PI-6 composite magnetic susceptibility record (SI x E-06).
58
0
50,000
100,000
150,000
200,000
0
2,000
4,000
6,000
8,000
10,000
0 200 400 600 800 1000 1200 1400
PI-6-6C-4H1 XRF data vs depth in core (mm)
Ca
Fe
Sr
Si
S
K
Ti
Mn
Ca,
Fe,
Sr
Si, S
, K, Ti, M
n
depth in core (mm)
Figure 4-11. Selected XRF data (kcps) from site PI-6-6C-4H-1 plotted versus depth in the core section (mm).
59
0
50
100
150
200
250
0 20000 40000 60000 80000
PI-6 composite green color reflectance data
green Intensity
gree
n in
tens
ity
Age (ybp)
Figure 4-12. PI-6 Composite green color reflectance data versus age raw data (ybp).
60
Figure 4-13. L* image data versus age (ybp). Original values are light grey, and a 100 point running mean is plotted with a thick dark blue line.
61
Figure 4-14. a* image data versus age (ybp). Original values are light green, and a 100 point running mean is plotted with a thick dark green line.
62
Figure 4-15. b* image data versus age (ybp). Original values are gold, and a 100 point running mean is plotted with a thick black line.
63
0
10
20
30
40
50
60
70
80
0 20000 40000 60000 80000
Image-derived granularity vs time (ybp)
% area (granularity threshold)
Age (y bp)
% a
rea
Figure 4-16. Image-derived granularity versus age (ybp)
64
CHAPTER 5 INTERPRETATION OF SEDIMENTS FROM SITE PI-6
Modern Sediment Properties
Modern Limnological Setting
Aerial photographs of lakes in the Petén region, notably Petén Itzá, Quexil, and
Salpetén display a “halo” of lighter colored sediments in the shallow, littoral zones of the
lakes. This is consequence of accumulation of bio-induced carbonate in shallow water
(Brenner, 1994). Lake Petén Itzá’s north basin has a gradually sloping southern littoral
zone, whereas the north shore is steeper and has a narrow littoral zone. In the modern
lake, the southern, shallow-water zone is the major region of carbonate accumulation.
Gastropod and ostracod shells are abundant on the lake floor at a water depth of 20 m
in the southern basin. This shell accumulation may be explained by the winnowing of
fine sediments and the concentration of shell material at this water depth, associated
with epilimnetic water-column circulation (Figure 5-1).
Lightness and E/P
My hypothesis is that shifts in the location of the carbonate depositional zone
cause changes in sediment core lightness and that such changes in core lightness
indicate changes in water depth. Hodell et al. (2007) reported similar relations between
water depth and sediment lightness in the relatively shallow (zmax<25m) Lake Punta
Laguna in the northeastern Yucatan Peninsula, Mexico. Shallow-water sediments in
both Punta Laguna and Petén Itzá contain high amounts of bio-induced carbonate and
are lighter. In the deeper areas of these thermally-stratified lakes, dissolution of
carbonate under relatively lower pH conditions yields sediments that have higher
organic matter content and are hence darker.
65
My hypothesis is that when lake level dropped in the past, the zone of carbonate
deposition migrated toward deeper areas of the lake, changing the color and hue of the
deposited sediments. Thus, past lake level can be inferred using changes in lightness
and hue of the sediments. Lighter sediments are indicative of lower water levels, and
hence an increase in the E/P ratio.
Confounding Factors
The original hypothesis is confounded when using the composite sediment core record
lightness of PI-6. The variations in sediment lithology and the presence of gypsum,
carbonate turbidites and variations in the clay all influence sediment core lightness so
that sediment core lightness alone is not an indicator of water depth in Lake Petén Itzá.
Surface Sediment Transect Samples
The goal of collecting modern surface sediment samples along a water-depth
transect was to evaluate the hypothesized relation of changes in sediment lightness and
geochemistry to water depth, and to calibrate the hypothesized paleoclimatic signal in
changes of sediment core image lightness. The modern Lake Petén Itzá surface
sediment samples show a relation between water depth, carbonate content, and
lightness of the sediments (Figures 5-2, 5-3).
Sediment Cores
Core Images
Core images are representations of the sediment record. Image hue, brightness,
and structure are functions of the biological, geochemical, and lithological
characteristics of the sediment as well as the depositional environment and perhaps
diagenetic changes. Understanding the relation between an image and the
66
biogeochemical properties of sediments is the first step in using images as proxies for
paleoenvironmental conditions.
Variations in Sediment Lithology
The diversity of facies within the PI-6 sediment cores suggests that modern
surface sediment samples do not capture the whole range of depositional and lithologic
environments that existed in the lake through time. This finding makes sense in light of
the major climate changes during the last glacial period and termination, which lack
analogs in modern Lake Petén Itzá (e.g. deglacial gypsum deposition). Carbonate-rich
turbidite layers are observed in several cores, but the record does not suggest direct
climatic linkage to the turbidites. Holocene gray clays are believed to be largely a
consequence of anthropogenic erosion of detrital material into the lake (Hodell et al.,
2000), whereas glacial-age clays are thought to have been delivered by increased
precipitation.
Turbidites, gypsum layers and sulfur nodules are lighter than clays, but they do not
influence sediment lightness in a manner that corresponds to a direct relation with E/P.
Because of these complex depositional and post-depositional changes to sediments,
lightness alone was not linked to changes in E/P. There are no surface sediment
lithologic units with modern day analogs of gypsum sands or sulfur nodules.
Although the gypsum sands lack a modern day analog in Lake Petén Itzá,
precipitation of gypsum has been observed in nearby Lake Salpeten and in Lake
Chichancanab, farther north on the Mexican part of the Yucatan Peninsula (Hodell et al.
2001). Dryer conditions favor greater E/P and concentrate gypsum CaSO4•2H2O and
other salts in Lake Petén Itzá. The granular, gypsum-rich intervals are not significantly
lighter than the clay/carbonate or turbidite units in the cores. They are however, easily
67
identified using “find edges” as part of an algorithm to estimate granularity of the
sediments (i.e. roughness of the core surface).
XRF Data
Scanning XRF data collected from sections of the PI-6 sediment record provide
nearly continuous, non-destructive measurements of elemental concentrations. The
data have not yet been assembled into a composite record, but individual core section
data were used to evaluate changes in lithology and optical properties to understand the
elemental and depositional conditions influencing the sediment core record.
Multivariate analysis suggests that certain detrital elements are highly correlated within
the sediment record of PI-6 (Table 5-1, Figure 5-4). Titanium and iron are extremely
closely correlated, and both are correlated to silicon and manganese. Calcium is
slightly negatively correlated with the detrital elements.
L*a*b* Data Evaluation and Interpretation
L*a*b* data provide an opportunity to independently evaluate changes in lightness
and hue of sediments. Shifts in a* values are well suited to evaluating changes in
sediment redox conditions as reddish oxidative environments are more positive in a*
and reducing environments with iron are often greenish. Sediment redox conditions and
mineralogy have been studied using a* variations in marine cores (Nagao, 1992), and
pigments of lake sediments and marine sediments have been used to understand
changes in redox conditions (Lyle 1982, Engstrom, 1985).
In the surface sediments from the shallowest water, the a* data from surface
sediment samples suggest they are greenish. This may be due to mineralogy in the
littoral zone, or the presence of macrophytes. In water depths that range from 5 to 14
m, the samples have the highest a* values, and are reddish in color. This may indicate
68
that sediments in this depositional environment are within an oxidizing zone. Surface
sediment samples in water depths >14 m have more negative a* values and indicate
greener gyttja, which is suggestive of an anaerobic depositional setting. A selected
core section from site PI-6 is presented with XRF and L*a*b* image data to illustrate the
common elemental variations among the lithological units (Figure 5-5 to 5-7).
Using scanning XRF data from surface sediment samples, the following relations
were identified between the sediment and a* values:
1) Carbonates are characterized by higher a* values.
2) Clays are characterized by lower a* values under reducing conditions and
display more positive in a* values under oxidizing conditions. This may be
due to the presence of manganese or the oxidation state of Mn and Fe.
3) Gypsum layers are characterized by higher a* values.
4) Oxygen isotopic data, which are influenced by changes in E/P, and a* values
are correlated in PI-6 sediments (Figure 5-8).
Magnetic Susceptibility
Magnetic susceptibility may reflect changing climate conditions in that E/P
influences weathering and input of allochthonous clays to the lake. In addition,
magnetic susceptibility is known to vary with particle size. Susceptibility spikes coincide
with the presence of volcanic ash and accessory minerals. Gypsum-rich layers are
characterized by low magnetic susceptibility and relatively large particle size.
Granularity
The proxy for particle size (granularity) provides a rapid estimate of roughness of
the composite core-surface. This measure is presented as a ratio of the area, and is
not an absolute measure of the particle size or grainsize. Because of the relation
69
between susceptibility and particle size, the grainsize proxy and susceptibility are
inversely related (Figure 5-9). Because of the great difference in grain size between
gypsum-rich intervals and clay-dominated sediments, the granularity proxy was very
useful to classify the lithology of site PI-6 cores.
70
Table 5-1. Multivariate correlations between elements of core sections from site PI-6. Si S K Ca Ti Mn Fe Sr
Si 1 -0.6894 0.6933 -0.4403 0.8889 0.8111 0.8867 -0.7126 S -0.6894 1 -0.632 0.2804 -0.6809 -0.7305 -0.6869 0.5282 K 0.6933 -0.632 1 -0.2243 0.6535 0.7392 0.6676 -0.6044 Ca -0.4403 0.2804 -0.2243 1 -0.6662 -0.2672 -0.6725 0.4392 Ti 0.8889 -0.6809 0.6535 -0.6662 1 0.7603 0.9962 -0.7488 Mn 0.8111 -0.7305 0.7392 -0.2672 0.7603 1 0.7682 -0.6711 Fe 0.8867 -0.6869 0.6676 -0.6725 0.9962 0.7682 1 -0.7559 Sr -0.7126 0.5282 -0.6044 0.4392 -0.7488 -0.6711 -0.7559 1
Figure 5-1 Cartoon depicting stratification of Lake Petén Itzá (water temperature profile
taken on 13 August 2002 (Hillesheim et al., 2005)). The thermocline at approximately 30 m water depth is associated with an erosional feature due to subaquatic currents (Anselmetti et al., 2006, figure modified by A. Müller, 2009). Dissolved oxygen is plotted for May (solid line) and August (dashed line) 1980 (unpublished data).
71
0
20
40
60
80
100
1200 20 40 60 80 100
Surface sediment geochemistry (CaCO3) vs depth (m)
CaCO3
Dep
th (m
)
% CaCO3
Figure 5-2. CaCO3 content of modern surface sediment samples.
72
0
20
40
60
80
100
12010 20 30 40 50 60 70 80 90
-4 -2 0 2 4 6
Surface sample L*, a* vs depth
L*a*
Dep
th (m
)
L*
a*
Figure 5-3. Surface sample L* and a* vs depth. Highest (reddish) a* values circled. Right plate shows transect sample images in depth-order with lines showing select correlating points on plot.
73
Figure 5-4. Multivariate correlation matrix of PI-6 XRF data. Ellipses are calculated and plotted using JMP (by SAS) software for multivariate analysis.
74
0
50,000
100,000
150,000
200,000
-500
0
500
1,000
1,500
2,000
2,500
3,000
3,500
0 200 400 600 800 1000 1200 1400
PI-6 6C-4H1 XRF data vs depth in section (mm)
Ca
Fe
Sr
Ti
Ca,
Fe,
Sr (
mg/
kg)
Core position (mm)
Ti (mg/kg)
Figure 5-5. PI-6 6C-4H-1 XRF Ca, Fe, Sr, and Ti data (kcps). Clay, dominant in the core section from 0 to 1000 mm has
high concentrations of Ti, and Fe, which are detrital in Lake Petén Itzá. Carbonate turbidites are low in Fe and Ti but elevated in Ca between 300 and 400 mm section depth. Gypsiferous intervals from 1000 to 1400 mm depth are characterized by lower levels of detrital elements (Fe, Ti) but are elevated in Ca, and Sr.
75
0
2,000
4,000
6,000
8,000
10,000
0
200
400
600
800
1,000
1,200
1,400
0 200 400 600 800 1000 1200 1400
PI-6 6C-4H1 XRF data vs depth in section (mm)
SMn
K
Si
S, M
n (m
g/kg
)
Core position (mm)
K, S
i (mg/kg)
Figure 5-6. PI-6 6C-4H-1 XRF S, Mn, K, and Si data (kcps). The clays in the top of the core section are higher in Si, and
K, both indicative of detrital material. Clay between 500 and 1000 mm in the core section has higher levels of manganese. Gypsum in the bottom of the core section has elevated levels of sulfur.
76
0
20
40
60
80
100
-8
-6
-4
-2
0
0 200 400 600 800 1000 1200 1400
PI-6 6C-4H-1 L*a*b* data
L*
b*
a*
L*, b
* a*
Core position (mm) Figure 5-7. PI-6 6C-4H-1 L*a*b* data. Carbonate turbidites have elevated L* values and drops in a*. The greenish clays
from 500 to 900 mm in the core section have lower a* values.
77
3
4
5
6
7
8
-10
-5
0
5
10
15
12000 16000 20000 24000 28000 32000 36000 40000 44000
PI-6 δ18O and a* vs time (ybp)
δ18O
a*
δ18
O A*
Age (y bp) Figure 5-8. PI-6 composite oxygen isotopic data (blue) courtesy of Jaime Escobar (in prep), and a* color data.
78
-20
0
20
40
60
80
100
120
140
600
800
1000
0
20
40
60
80
0 20000 40000 60000 80000
Granularity and magnetic susceptibility (SI x E-06)
magnetic susceptibility (SI x E-06)
granularity
mag
netic
sus
cept
ibili
ty (
SI
x E
-06)
% A
rea
Age (y bp) Figure 5-9. PI-6 Composite estimate of grain size proxy (using the “find edges” based algorithm). Magnetic susceptibility
composite data (courtesy of Dr. A. Gilli)
79
CHAPTER 6 CORRELATION TO OTHER CLIMATE RECORDS
Regional Paleoclimate
Bradbury (1997) proposed that strong inputs of winter Pacific precipitation
accounted for wet conditions in the central Mexican highlands during the last glacial
period. The lowlands of Petén were thought to have been dry during the last glacial
maximum. Interpretation of pollen records from Lake Quexil, Guatemala suggested that
the climate in the region was cool and dry (Leyden et al. 1993). The Pleistocene portion
of the Quexil section, however, was poorly dated, leaving uncertainties about the timing
of vegetation shifts.
Magnetic susceptibility, grainsize, and the green reflectance of sediments from
Lake Petén Itzá suggest that the last glacial maximum (25-18 kybp) was cool and wet
(Figure 6-1). At ~18,000 ybp the sediment record suggests a change to cool, dry
climate. During the deglaciation, two periods of increased magnetic susceptibility
coincide with clay-dominant sediments (Bush et al. 2009).
The marine Cariaco Basin sediments, off Venezuela, have been studied to
examine the shifts in the position of the ITCZ. The basin sediments are in an anoxic
depositional environment and display changes associated with fluctuations in
productivity due to increased upwelling. The 550 nm spectral reflectance in sediment
cores is interpreted to reflect past climate change. Increases in the reflectance values
are linked to increased upwelling and a southerly shift in the annual mean position of
the ITCZ. The PI-6 a* channel is plotted on an independent time-scale with Cariaco 550
nm reflectance (Figure 6-2). Between ~48,000 and ~23,000 ybp the PI-6 record
alternates between gypsum- and clay-rich intervals during stadials and interstadials,
80
respectively, that correlate well with other records including Cariaco (Müller et al. 2009).
The a* variable illustrates the millennial fluctuations that are influenced by these
changes in lithology. The PI-6 lithology and a* values indicate that the
paleoenvironmental conditions of Lake Petén Itzá may be interpreted using the image
data, and that the a* variable may be influenced by shifts in the depositional regime and
redox conditions.
Global Paleoclimate and Timeseries Analysis
The PI-6 record exhibits orbitally-influenced precession-driven variability (Bush et
al. 2009), however, a suite of whole-record time-series analysis has not yet been
generated. The PI-6 image data provide a continuous, high-resolution record that
should prove useful for investigating such changes and conducting time-series analysis.
Spectral analysis was performed on linearly-spaced data using AutoSignal (Figure
6-3). A low frequency component was removed using the cubic spline normalization
function AutoSignal (Figure 6-4). The resulting spectrum exhibits statistically significant
signals at frequencies of 0.021, 0.043, 0.103, and 0.189, which correspond to periods of
~48, ~23, ~9 and ~5 ky respectively. Wavelet analysis also captures some of the
statistically significant components of the a* dataset (Figure 6-5)
The a* variable was filtered with a 20-kyr filter using Analyseries (Berger 1992,
Paillard et al. 1996) and was plotted along with L* and aged images. The images have
been stretched and compressed according to the age model so that they correspond
with the age-scaled data (Figure 6-6). Cross-correlation of the filtered 20-ky component
indicates that the a* variable is coherent and in-phase with the summer insolation at 17
degrees north (Figure 6-7). A cross-spectral analysis was performed to test the
correlation and the correlation ranges from 0.64 to 0.93 (Figure 6-8)
81
Figure 6-1. Pollen accumulation, charcoal concentration, magnetic susceptibility from site PI-6 (Bush et al., 2009), color
reflectance from site 1002 (Cariaco Basin – Peterson et al., 2000), and δ18O of the Greenland Ice Core (Grootes et al., 1993).
82
Figure 6-2. PI-6 a* data plotted vs time (light green are original data and the dark green line is a 100-point running mean) and the Cariaco 550 nm reflectance (Peterson et al., 2000) plotted on its independent time scale.
H0 H1 H2 H3
H4 H5 H6 H0 H1 H2 H3
H4 H5 H6
83
Figure 6-3. a* frequency spectrum, colored horizontal bars indicate significance (red= 99.9, yellow= 99, green= 95,
teal=90, white=50).
84
Figure 6-4. a* frequency spectrum following cubic spline filter to remove low frequency signal. Colored horizontal bars
indicate significance (red= 99.9, yellow= 99, green= 95, teal=90, white=50).
85
Figure 6-5. a* wavelet surface plot, statistical envelope shown with curved lines and transparent mask.
86
Figure 6-6. L*a* plots of PI-6 and age-model fitted images showing the relation between
composite sediment core images and the L* and a* variables. The a* data are plotted with a 20-kyr filtered component that corresponds with solar insolation.
87
Figure 6-7. Summer insolation at 17 degrees north (Analyseries) plotted vs age (kybp).
88
Figure 6-8. Cross-correlation between filtered a* and summer insolation at 17 degrees north.
89
CHAPTER 7 SUMMARY AND CONCLUSIONS
The sediment core lithology and image data suggest that Lake Petén Itzá
continuously held water throughout the past ~85,000 years. Data from several proxies,
including image data, provide evidence for a wetter climate regime between 25,000 and
18,000 ybp. Variations in lightness may be less sensitive to shifts in sediment type
between facies such as gypsum and carbonate-rich clay, but shifts in hue allow for the
comparison of these sediments. The hypothesis that lightness of sediments is
correlated to climatic signals was rejected. Image data are influenced by the
biogeochemistry and depositional environments of a lake, as well as post-depositional
diagenetic changes. Although lightness was not suitable for paleoclimatic interpretation,
the a* variable is correlated to shifts in geochemical properties of modern day
sediments and to historic changes in solar insolation, when measured in cores. Modern
surface sediments did show a relation between geochemical properties, lightness and
a*.
The a* image data are influenced by the oxidative state of the sediments, and
provide a mechanism for evaluating changes in the varied lithologies. Some clay layers
were more red and others more green. The shifts in hue of these clays may be due to
changes in redox conditions. These variations could be evaluated with a complete
series of XRF data for site PI-6.
This work suggests that sediment-core images from Lake Petén Itzá are well
suited to the investigation of paleoenvironmental conditions and that the a* variable is
well suited to understanding millennial-scale variations and orbital-influenced variability
in sediments. Image analysis of granularity is useful for classifying sediment lithology
90
and provides an approximation for grainsize in the sediments. Analyses of both
granularity and L*a*b* in sediments from Lake Petén Itzá were performed non-
destructively and at high resolution. In the future, other researchers will have the
opportunity to analyze the original images. The digital files will not degrade with time,
enabling new algorithms or techniques to be applied to the same images used in this
study.
Limitations of Images as Paleoclimatic Proxies
The image data collected for this project were limited by the dynamic range of the
Geotek hardware. The camera used in this study had a dynamic range of 8-bits (or 0-
255). Normalization of images with very light sediments led to “saturation” of some
pixels. The degree of lightness extended beyond the range of possible values.
Cameras with 16-bit and even 32-bit greyscale capabilities are currently available, but
such cameras are still expensive. Normalization allowed comparison of sediment
images in this study, but improvement in the dynamic range of imaging hardware would
allow direct comparison of extremely light and dark sediments on a continuous scale.
An additional limitation of the method in this study was that images were
influenced by a wide range of depositional and post-depositional processes. Because
the sediments are not composed solely of varying proportions of clay, carbonate,
gypsum, and organic matter, the assumption that lightness is controlled largely by lake
level and carbonate deposition is not valid, especially in the Pleistocene deposits.
Estimations of grainsize (granularity) were useful to identify regions of gypsum
deposition and would not have been as easily measured without automated image
processing.
91
Future Work
As the remaining sediment cores collected during the Petén Itzá Scientific Drilling
Project are processed, digital core images and image analysis will continue to provide a
rapid means of sediment characterization. With the integration of NCLIP, the successor
to SPLICER, the Corewall Suite images will provide the backbone for collaborative
interpretation of sediments. Additional radiocarbon dates will improve core chronology
and permit a more accurate interpretation of the data during the Pleistocene
deglaciation. With a more tightly constrained 14C profile, oxygen isotopic signatures
may be useful to determine source of the precipitation during events that occurred at
15,000 and 12,000 ybp.
92
APPENDIX A SOFTWARE USED
Field Software
Drilling Information System
The primary field database was designed and implemented by the Operational
Support Group (OSG) of the International Continental Scientific Drilling Program (ICDP)
and custom tailored for the drilling operations prior to field deployment. Two database
curators were trained by Ronald Conzé of the OSG in December of 2005. This
database was critical for keeping track of drilling information, Multi-Sensor Core Logger
(MSCL), and smear slide data. The science team also used the software to update the
project web page:
http://www.icdp-online.de/contenido/icdp/front_content.php?idart=1082
SPLICER
Splicer was used for stratigraphic correlation on site, using the GEOTEK MSCL
data. The software runs on Unix, and for this project much effort expended by
University of Florida Geosciences guru Ray G. Thomas. Luckily, Ray’s expertise
prevailed, and Splicer was successfully ported through the X environment on an Apple
iBook. Splicer was designed by Peter DeMenocal and Ann Esmay for the Ocean
Drilling Program (ODP). This software is currently being overhauled and the successor
is NCLIP, one part of the Corewall suite.
SPLICER- http://www.iagp.net/joi/index.html
-Please note that this is far from a “plug and play” installation. Proceed with caution!
NCLIP- http://www.iagp.net/NClip/Binaries/
93
Image Processing and Analysis
The large file sizes prevent current computers from opening core composite
images at full resolution. In the case of the Petén Itzá cores, a full-resolution composite
core image for site Pi-6 is well over 2 Gigabytes (GB). When combined with multiple
holes and sites, a dataset of full-resolution images cannot be rapidly transmitted over
current computer networks. One of the best ways of transferring such data locally is the
use of a “sneakernet.” A sneakernet describes moving external storage devices from
location to location by physically carrying them.
Many large image files necessitate the automation of tasks and batch processing
of individual section images. However, because the needs of these analyses are
unique, software packages such as Adobe Photoshop™ are not adequate for analysis
and processing. The first step in performing the digital image analysis was to normalize
the core images. Despite the assembly-line approach to core curation at the Lacustrine
Core Repository (LRC/LacCore), there is still variability in the imaging conditions over
time. The variability is most visible (pun intended) when imaging of cores is performed
over several days or by different users. One of the essential safeguards during the
image acquisition process is the scanning of the same, uniform, color calibration plate
with each core section. Acquiring an image of the tile is critical for understanding the
variability of the relative hue or color of the sediments and fundamental for creating an
accurate composite core image or image dataset.
Using the LRC standard core acquisition protocol, the calibration tile is placed near
the bottom of the core section (the right side) and it is placed so that the white – gray
tiles are located along the bottom of the core image. The calibration plate must be
clean, free of debris and on the same plane as the sample being analyzed.
94
ImageJ
ImageJ is an open-source Java™ based program funded by the National Institutes
of Health, and is freely available for download at: http://rsb.info.nih.gov/ij/. It was the
primary software used for image analysis and normalization.
Due to limitations of current computing hardware, it is not yet possible to
manipulate image datasets spanning several 10s of meters of sediment core. Because
images are scanned in individual sections they may be handled individually. The design
and use of macros (small scripts) to automate image normalization and analysis to
perform tasks requires an initial investment in time, but upon development, the macros
may be easily and rapidly used to perform the operations on other image datasets. All
image processing discussed in this manuscript was performed in ImageJ.
A series of macros was designed for image processing of large image datasets for
the Petén Itzá Scientific Drilling Project (PISDP). To measure the degree of variability
across the image set, a macro was written to measure the values of the Red, Green,
and Blue colored squares on the digital image. However, because each image is
different in length, the “image position” of the calibration color tiles was unknown.
The following ImageJ plug-ins were used and are necessary for the use of macros
included in this thesis:
Stack Combiner http://rsb.info.nih.gov/ij/plugins/combiner.html
stackreg http://bigwww.epfl.ch/thevenaz/stackreg/
turboreg http://bigwww.epfl.ch/thevenaz/turboreg/
RGB Measure http://rsb.info.nih.gov/ij/plugins/rgb-measure.html
Colorspace Converter http://rsb.info.nih.gov/ij/plugins/color-space-converter.html
95
Matlab
Matlab was used for data manipulation and plotting.
Corelyzer and Corewall
Several limnologists and sedimentologists who have worked in the field for a
number of years have seen the effects of a dry cold-room and poorly handled cores.
One of the great aspects of images, and especially properly collected digital core
images, is that the sediment image will remain exactly the same as when it was
collected, for as long as the data can be stored. This allows for maintenance of a virtual
“core,” which scientists may consult to review stratigraphy or select depths where
samples are to be collected.
Core images are poised to be the backbone of a suite of applications now known
as Corewall. This suite will allow users to view enormous image datasets and other
data, which will enable collaborators to comment and share images in real time.
Currently, the visualization tool is functioning and is known as the “Corelyzer”. This
software allows the science team to load an image dataset (at full resolution) and then
view the image dataset on a common personal computer, as opposed to the
considerably more costly commercial alternatives. Corewall will incorporate several
aspects of previous software packages into a tool for presenting data and images. The
software should also allow researchers to collaborate remotely.
url: http://www.evl.uic.edu/cavern/corewall/about.html
wiki: http://sqlcore.geo.umn.edu/CoreVault/cwWiki/index.php/Main_Page
96
Analyseries
Analyseries is a MacIntosh time series analysis tool by Didier Paillard, LSCE. The
latest version of AnalySeries is compatible with Mac OSX and is available at
the LSCE website.
url: http://www.lsce.ipsl.fr/logiciels/index.php
97
APPENDIX B PALEOLAKE HIGH-STANDS
During the field campaign of the Petén Itzá Scientific Drilling Project, lacustrine
deposits were identified by the presence of aquatic gastropods at elevations ranging
from 15 to 20 meters above the modern lake level (Grzesik, Curtis, Escobar–
unpublished data). These samples may provide the means for stratigraphic correlation
of extreme lake level highstands to the studied paleo record.
Initial analysis of gastropods from the high-elevation deposits revealed oxygen
isotopic signatures that are slightly more negative in δ18O than those of modern day
lake gastropods. This value may serve as a “fingerprint” to identify down-core
gastropods that are contemporaneous with the shells that were deposited above the
modern lake surface.
Results of this research increase understanding of an unknown aspect of the
history of Petén lakes, which have been studied during the past 30 years (Deevey
1978). Understanding the high-stand timing and magnitude is something that will allow
more accurate modeling of the regional and global climate. The Petén Itzá Scientific
Drilling Project will benefit from this research as interpretation and analysis of longer
cores begins.
Figure B-1. Hydrobiidae gastropods (1mm ticks)
98
Figure B-2. A cartoon illustrating the implications for a 20-meter rise in the level of Lake
Petén Itzá. The dark area shows the extent of the modern lake. Lighter shaded area shows areas that would be flooded by a 20-m rise in water level. (white dot illustrates sample locality)
99
APPENDIX C SITE PI-3
Site PI-3 was split and imaged at the LRC during the Summer 2006 sampling
party, but it was not presented in this thesis because of concern over slumping below 32
meters, and the lack of a well-defined chronology. Image data were used for
stratigraphic correlation with the Corelyzer application.
100
Figure C-1. PI-3 composite lithological description by Andreas Müller.
101
LIST OF REFERENCES
Anselmetti FS, Ariztegui D, Hodell DA, Hillesheim MB, Brenner M, Gilli A, McKenzie JA, Müller, AD (2006) Late Quaternary climate-induced lake level variations in Lake Petén Itzá, Guatemala, inferred from seismic stratigraphic analysis. Palaeogeogr Palaeoclimatol Palaeoecol 230: 52–69
Binford MW, Brenner M, Whitmore TJ, Higuera-Gundy A, Deevey ES, Leyden BW (1987) Ecosystems, paleoecology, and human disturbance insubtropical and tropical America. Quat Sci Rev 6: 115–128
Brenner M (1994) Lakes Salpetén and Quexil, Petén, Guatemala, Central America. In: Kelts K, Gierlowski- Kordesch E (eds). Global Geological Record of Lake Basins, Volume 1. Camb University Press. Cambridge, UK, pp 377–380
Brenner M, Rosenmeier MF, Hodell DA, Curtis JH (2002) Paleolimnology of the Maya Lowlands: long-term perspectives on interactions among climate, environment, and humans. Anc Mesoam 13: 141–157
Bond G, Kromer B, Beer J, Muscheler R, Evans MN, Showers W, Hoffman S, Lotti-Bond R, Hajdas I, Bonani G (2001) Persistent solar influence on North Atlantic climate during the Holocene. Science 294: 2130–2136
Bradbury JP (1997) Sources of glacial moisture in Mesoamerica. Quat Int 43/44: 97–110
Bronk Ramsey C (2006) OxCal 4.0 (Oxford Radiocarbon Accelerator Unit, Oxford, UK)
Bush MB, Correa-Metrio A, Hodell DA, Brenner M, Anselmetti FS, Ariztegui D, Müller A D, Curtis JH, Grzesik D, Burton C, Gilli A (2009) The Last Glacial Maximum in lowland Central America. In: Vimeux F, Sylvestre F, and Khodri M (eds) Past Climate Variability in South America and Surrounding Regions. Springer-Verlag, GmbH, pp 380
Covich AP, Stuiver M (1974) Changes in oxygen 18 as a measure of long-term fluctuations in tropical lake levels and molluscan populations. Limnol Oceanogr 19: 682–691
Deevey ES (1957) Limnologic studies in Middle America, with a chapter on Aztec limnology. Conn Acad Arts Sci Trans 39: 213–328
Deevey ES (1978) Holocene forests and Maya disturbance near Quexil Lake, Petén, Guatemala. Polskie Arch Hydrobiol 25, No. 1/2: 117–129
Deevey ES, Brenner M, Flannery MS, Yezdani GH (1980) Lakes Yaxha and Sacnab, Petén, Guatemala: Limnology and hydrology. Archaeol Hydrobiol 57: 419–460
102
Deevey ES, Brenner M, Binford MW(1983) Paleolimnology of the Petén Lake District, Guatemala III: Late Pleistocene and Gamblian environments of the Maya area. Hydrobiologia 103: 211–216
Delvaux D, Kervyn F, Vittori E, Kajara RSA, Kilembe E (1998) Late Quaternary tectonic activity and lake level change in the Rukwa Rift Basin. J Afr Earth Sci 26, 3: 397–421
Engstrom DR, Swain EB, Kingston JC (1985) A paleolimnological record of human disturbance from Harvey’s Lake, Vermont: geochemistry, pigments and diatoms. J Freshw Biol 15: 261–288
Flower BP, Hastings DW, Hill HW, Quinn TM (2004) Phasing of deglacial warming and Laurentide Ice Sheet meltwater in the Gulf of Mexico. Geology 32: 597–600
Francus P (2004) Image Analysis, Sediments and Paleoenvironments. Vol 7. Kluwer
Haug GH, Hughen KA, Peterson LC, Sigman DM, Röhl U (2001). Southward migration of the Intertropical Convergence Zone through the Holocene. Science 293: 1304–1308
Haug GH, Gunther D, Peterson LC, Sigman DM, Hughen KA, Aeschlimann B (2003) Climate and the collapse of Mayan civilization. Science 299: 1731–1735
Hillesheim MB, Hodell DA, Leyden BW, Brenner M, Curtis JH, Anselmetti FS, Ariztegui D, Buck DG, Guilderson TP, Rosenmeier MF, Schnurrenberger DW (2005) Climate change in lowland Central America during the late deglacial and early Holocene. J Quat Sci 20: 363–376
Hodell DA, Curtis JH, Brenner M (1995) Possible role of climate change in the collapse of the Maya civilization. Nature 375: 391–394
Hodell DA, Brenner M, Curtis JH, Guilderson T (2001). Solar forcing of drought frequency in the Maya Lowlands. Science 292: 1367–1369
Hodell DA, Brenner M, Curtis JH (2000) in: D. L. Lentz, (ed) Imperfect Balance: Landscape Transformations in the Precolumbian Americas, Columbia Univ Press, New York, pp 13–38
Hodell DA, Brenner M, Curtis JH (2005) Terminal Classic drought in the northern Maya Lowlands inferred from multiple sediment cores in Lake Chichancanab (Mexico). Quat Sci Rev 24: 1413–1427
Hodell DA, Brenner M, Curtis JH, Medina-Gonzalez R, Rosenmeier MF, Guilderson TP, Chan-Can EI, Albornaz-Pat A (2005). Climate change on the Yucatan Peninsula during the Little Ice Age. Quat Res 63: 109–121
103
Hodell DA, Brenner M, Curtis JH (2007) Climate and cultural history of the northeastern Yucatan Peninsula, Quintana Roo, Mexico. Climatic Change. DOI 10.1007/s10584-006-9177-4
Hodell DA, Anselmetti F, Ariztegui D, Brenner M, Curtis J, Gilli A, Müller A, Grzesik D, and members of the Peten Itza Scientific Drilling Party (2006) Preliminary Results of the Lake Peten Itza Scientific Drilling Project, 10th Int Paleolimnol Symp
Hughen K, Southon J, Lehman S, Bertrand C, Turnbull J (2006) Marine-derived 14C calibration and activity record for the past 50,000 years updated from the Cariaco Basin. Quat Sci Rev 25: 3216–3227
Johnnson TC, Scholz CA, Talbot MR, Kelts K, Ricketts RD, Ngobi G, Beuning K, Ssemmanda I, McGill JW (1996) Late Pleistocene dessication of Lake Victoria and rapid evolution of Cichlid fishes. Science 273: 1091–1093
Leyden B W (1984) Guatemalan Forest Synthesis after Pleistocene Aridity. Proc of the Natl Acad of Sci USA 81: 4856–4859
Leyden BW, Brenner M, Hodell DA, Curtis JH (1993) Orbital and internal forcing of Climate on the Yucatan Peninsula for the past ~36 kyr. Palaeogeogr Palaeoclimatol Palaeoecol 109: 193–210
Lyle M, (1982) The brown-green color transition in marine sediments: A marker of the FE(III)-Fe(II) redox boundary. Limnol Oceonogr 28(5): 1026–1033
Müller AD, Anselmetti FS, Ariztegui D, Brenner M, Hodell DA, Curtis JH, Escobar J, Gilli A, Grzesik DA, Guilderson TP, Kutterolf S, Plotze M (2010) Late Quaternary palaeoenvironment of northern Guatemala:evidence from deep drill cores and seismic stratigraphy of Lake Peten Itza. Sedimentology (In press)
Nagao S, Nakashima S (1992) The factors controlling vertical color variations of North Atlantic Madeira Abyssal Plain sediments. Geology 109: 83–94.
Peterson LC, Haug GH, Hughen KA, Rohl U (2000) Rapid Changes in the HydrologicalCycle of the Tropical Atlantic During the Last Glacial. Science 290: 1947–1951
Rosenmeier MF, Hodell DA, Brenner M, Curtis JH, (2002) A 4000-year lacustrine record of environmental change in the southern Maya lowlands, Petén, Guatemala. J Quat Res 57: 183–190
Rosenmeier MF, Hodell DA, Brenner M, Curtis JH, Martin JB, Anselmetti FS, Ariztegui D, Guilderson TP(2002) Influence of vegetation change on watershed hydrology: Implications for paleoclimatic interpretation of lacustrine δ18O records. J Paleolim 27: 117–131
104
Russ JC (1999) The Image Processing Handbook. CRC Press, Boca Raton, Florida, 771 pp
Smol JP, GlewJR (1992) Paleolim Encycl Earth Syst Sci Vol 3:551–564
Vasconcelos C, Postec A, Warthmann R, McKenzie JA, Anselmetti F, Ariztegui D, Hodell D (2006) Geomicrobiological Study of Sulfur Nodules Forming in Sediments of Lake Peten Itza, Guatemala, Eos Trans AGU, 87(52)
Vinson GL (1962) Upper Cretaceous and Tertiary stratigraphy of Guatemala. Bull of the Amer Assoc of Petrol Geol 46: 425–456
Weaver AJ, Saenko OA, Clark PU, Mitrovica JX(2003) Meltwater Pulse 1A from Antarctica as a Trigger of the Bølling-Allerød Warm Interval, Science 299: 1709
Wilson EM (1984). in Moseley ED, Terry ED (eds) Physical geography of the Yucatan Peninsula., Yucatan: A World Apart, second ed., University of Alabama Press, Tuscaloosa, pp 5–40
105
BIOGRAPHICAL SKETCH
Dustin Aaron Grzesik was born in Utica New York in 1980 to Richard and Janice
Grzesik. He grew up on the northeastern shore of Oneida Lake, and in third grade
wrote a prize-winning editorial addressing his concern with pollution of the lake. As a
high school junior, Dustin was a Rotary youth exchange student in Coffs Harbour, NSW,
Australia. After graduating from High School in June 1998, Dustin enrolled at Paul
Smith’s College, Paul Smiths NY where he studied ecology and environmental
technology. After graduating with an Associate of Applied Science (A.A.S.) Dustin
participated in a field campaign to collect sediments from Lake Victoria, in Kenya and
Uganda with Dr. Curt Stager and three fellow students. In September of 2000, Dustin
enrolled at Northern Arizona University. While at Northern Arizona University, Dr.
Lawrence Fritz welcomed Dustin to the Electron Microscopy Facility and he began
analysis of the diatoms in sediments recovered from Lake Victoria cores. In May of
2002 he graduated with a Bachelor of Science (B.S.) degree in zoology. From June
2002 to April 2003, Dustin served as a U.S. Peace Corps Volunteer in Chongqing
China, where he was a lecturer of environmental chemistry and environmental
engineering at Chongqing Polytechnic College. It was in China that Dustin met his wife
Joanne Sum-Ping. After returning from China, Dustin worked as a Research
Technician at the Albert Einstein College of Medicine’s Analytical Imaging Facility. At
the AIF, he learned the nuances of confocal microscopy and digital image analysis. In
August 2005, Dustin enrolled at the University of Florida in the Department of
Geological Sciences. Dustin moved to New York, New York, in May of 2007 where he’s
employed as a geochemist for Malcolm Pirnie Inc. specializing in contaminated
sediments and their remediation.
